diff options
Diffstat (limited to 'llvm/test/CodeGen/X86')
82 files changed, 9487 insertions, 4268 deletions
diff --git a/llvm/test/CodeGen/X86/abds-neg.ll b/llvm/test/CodeGen/X86/abds-neg.ll index f6d66ab47ce0..2911edfbfd40 100644 --- a/llvm/test/CodeGen/X86/abds-neg.ll +++ b/llvm/test/CodeGen/X86/abds-neg.ll @@ -367,44 +367,49 @@ define i128 @abd_ext_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: subl %ecx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %edx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: sbbl %edi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %esi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll %eax, %esi -; X86-NEXT: cmovll %ebx, %edi -; X86-NEXT: cmovll %ebp, %edx -; X86-NEXT: cmovll (%esp), %ecx # 4-byte Folded Reload -; X86-NEXT: xorl %ebx, %ebx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %edx +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: sbbl %edx, %ebx +; X86-NEXT: movl 52(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %eax +; X86-NEXT: sbbl 32(%ebp), %edx +; X86-NEXT: sbbl 36(%ebp), %esi +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %ebx, %edx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: xorl %edi, %edi ; X86-NEXT: negl %ecx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %edx, %ebp -; X86-NEXT: movl $0, %edx -; X86-NEXT: sbbl %edi, %edx -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %edx, 8(%eax) -; X86-NEXT: movl %ebx, 12(%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %eax, %ebx +; X86-NEXT: movl $0, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 8(%ebp), %edx +; X86-NEXT: movl %ecx, (%edx) +; X86-NEXT: movl %ebx, 4(%edx) +; X86-NEXT: movl %eax, 8(%edx) +; X86-NEXT: movl %edi, 12(%edx) +; X86-NEXT: movl %edx, %eax +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -438,44 +443,49 @@ define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: subl %ecx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %edx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: sbbl %edi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %esi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll %eax, %esi -; X86-NEXT: cmovll %ebx, %edi -; X86-NEXT: cmovll %ebp, %edx -; X86-NEXT: cmovll (%esp), %ecx # 4-byte Folded Reload -; X86-NEXT: xorl %ebx, %ebx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %edx +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: sbbl %edx, %ebx +; X86-NEXT: movl 52(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %eax +; X86-NEXT: sbbl 32(%ebp), %edx +; X86-NEXT: sbbl 36(%ebp), %esi +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %ebx, %edx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: xorl %edi, %edi ; X86-NEXT: negl %ecx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %edx, %ebp -; X86-NEXT: movl $0, %edx -; X86-NEXT: sbbl %edi, %edx -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %edx, 8(%eax) -; X86-NEXT: movl %ebx, 12(%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %eax, %ebx +; X86-NEXT: movl $0, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 8(%ebp), %edx +; X86-NEXT: movl %ecx, (%edx) +; X86-NEXT: movl %ebx, 4(%edx) +; X86-NEXT: movl %eax, 8(%edx) +; X86-NEXT: movl %edi, 12(%edx) +; X86-NEXT: movl %edx, %eax +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -639,55 +649,59 @@ define i128 @abd_minmax_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_minmax_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: cmpl %eax, %esi -; X86-NEXT: sbbl %ebx, %ecx -; X86-NEXT: movl %edx, %ecx -; X86-NEXT: sbbl %ebp, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl %edx, %ecx -; X86-NEXT: sbbl %edi, %ecx -; X86-NEXT: movl %edi, %ecx -; X86-NEXT: cmovll %edx, %ecx -; X86-NEXT: movl %ecx, (%esp) # 4-byte Spill -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl %ebx, %ecx -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %eax, %edx -; X86-NEXT: cmovll %esi, %edx -; X86-NEXT: cmpl %esi, %eax -; X86-NEXT: movl %ebx, %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl %edi, %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %ebx -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %eax -; X86-NEXT: subl %eax, %edx -; X86-NEXT: sbbl %ebx, %ecx -; X86-NEXT: sbbl %esi, %ebp -; X86-NEXT: movl (%esp), %esi # 4-byte Reload -; X86-NEXT: sbbl %edi, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: movl %ecx, 4(%eax) -; X86-NEXT: movl %ebp, 8(%eax) -; X86-NEXT: movl %esi, 12(%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %esi +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: cmpl %esi, %edi +; X86-NEXT: sbbl 44(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: movl %ebx, %eax +; X86-NEXT: cmovll %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %edx, %eax +; X86-NEXT: cmovll 32(%ebp), %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 44(%ebp), %eax +; X86-NEXT: cmovll 28(%ebp), %eax +; X86-NEXT: movl %esi, %ecx +; X86-NEXT: cmovll %edi, %ecx +; X86-NEXT: cmpl %edi, %esi +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: sbbl 28(%ebp), %edi +; X86-NEXT: movl %edx, %edi +; X86-NEXT: sbbl 32(%ebp), %edi +; X86-NEXT: movl %ebx, %edi +; X86-NEXT: sbbl 36(%ebp), %edi +; X86-NEXT: cmovll 36(%ebp), %ebx +; X86-NEXT: cmovll 32(%ebp), %edx +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: cmovll 28(%ebp), %edi +; X86-NEXT: cmovll 24(%ebp), %esi +; X86-NEXT: subl %esi, %ecx +; X86-NEXT: sbbl %edi, %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-NEXT: sbbl %edx, %edi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: sbbl %ebx, %esi +; X86-NEXT: movl 8(%ebp), %edx +; X86-NEXT: movl %ecx, (%edx) +; X86-NEXT: movl %eax, 4(%edx) +; X86-NEXT: movl %edi, 8(%edx) +; X86-NEXT: movl %esi, 12(%edx) +; X86-NEXT: movl %edx, %eax +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -848,37 +862,41 @@ define i128 @abd_cmp_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_cmp_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovgel (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovgel %ebx, %esi -; X86-NEXT: cmovgel %ebp, %ecx -; X86-NEXT: cmovgel %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: movl 40(%ebp), %eax +; X86-NEXT: movl 44(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 48(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 36(%ebp), %ebx +; X86-NEXT: movl 52(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 40(%ebp), %ecx +; X86-NEXT: sbbl 44(%ebp), %edx +; X86-NEXT: sbbl 48(%ebp), %esi +; X86-NEXT: sbbl 52(%ebp), %ebx +; X86-NEXT: cmovgel {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovgel {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovgel %edi, %esi +; X86-NEXT: cmovgel %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -1118,35 +1136,39 @@ define i128 @abd_subnsw_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_subnsw_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: movl %ebx, %ebp -; X86-NEXT: subl %edi, %ebp -; X86-NEXT: movl %ebx, %edi -; X86-NEXT: sbbl %esi, %edi -; X86-NEXT: movl %ebx, %esi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: subl 40(%ebp), %esi +; X86-NEXT: sbbl 44(%ebp), %edx +; X86-NEXT: sbbl 48(%ebp), %ecx +; X86-NEXT: sbbl 52(%ebp), %eax +; X86-NEXT: movl %eax, %edi +; X86-NEXT: sarl $31, %edi +; X86-NEXT: xorl %edi, %eax +; X86-NEXT: xorl %edi, %ecx +; X86-NEXT: xorl %edi, %edx +; X86-NEXT: xorl %edi, %esi +; X86-NEXT: movl %edi, %ebx +; X86-NEXT: subl %esi, %ebx +; X86-NEXT: movl %edi, %esi ; X86-NEXT: sbbl %edx, %esi -; X86-NEXT: sbbl %ecx, %ebx -; X86-NEXT: movl %ebp, (%eax) -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: movl %esi, 8(%eax) -; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %edi, %edx +; X86-NEXT: sbbl %ecx, %edx +; X86-NEXT: sbbl %eax, %edi +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: movl %esi, 4(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -1175,35 +1197,39 @@ define i128 @abd_subnsw_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_subnsw_i128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: movl %ebx, %ebp -; X86-NEXT: subl %edi, %ebp -; X86-NEXT: movl %ebx, %edi -; X86-NEXT: sbbl %esi, %edi -; X86-NEXT: movl %ebx, %esi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: subl 40(%ebp), %esi +; X86-NEXT: sbbl 44(%ebp), %edx +; X86-NEXT: sbbl 48(%ebp), %ecx +; X86-NEXT: sbbl 52(%ebp), %eax +; X86-NEXT: movl %eax, %edi +; X86-NEXT: sarl $31, %edi +; X86-NEXT: xorl %edi, %eax +; X86-NEXT: xorl %edi, %ecx +; X86-NEXT: xorl %edi, %edx +; X86-NEXT: xorl %edi, %esi +; X86-NEXT: movl %edi, %ebx +; X86-NEXT: subl %esi, %ebx +; X86-NEXT: movl %edi, %esi ; X86-NEXT: sbbl %edx, %esi -; X86-NEXT: sbbl %ecx, %ebx -; X86-NEXT: movl %ebp, (%eax) -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: movl %esi, 8(%eax) -; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %edi, %edx +; X86-NEXT: sbbl %ecx, %edx +; X86-NEXT: sbbl %eax, %edi +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: movl %esi, 4(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/abds.ll b/llvm/test/CodeGen/X86/abds.ll index 0356c2702a41..a1a4ba81ae49 100644 --- a/llvm/test/CodeGen/X86/abds.ll +++ b/llvm/test/CodeGen/X86/abds.ll @@ -343,37 +343,41 @@ define i128 @abd_ext_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovll %ebx, %esi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: cmovll %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: sbbl 36(%ebp), %ebx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -404,37 +408,41 @@ define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovll %ebx, %esi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: cmovll %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: sbbl 36(%ebp), %ebx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -585,37 +593,41 @@ define i128 @abd_minmax_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_minmax_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovll %ebx, %esi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: cmovll %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: sbbl 36(%ebp), %ebx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -768,37 +780,41 @@ define i128 @abd_cmp_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_cmp_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovll %ebx, %esi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: cmovll %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: sbbl 36(%ebp), %ebx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -1027,35 +1043,38 @@ define i64 @abd_subnsw_i64_undef(i64 %a, i64 %b) nounwind { define i128 @abd_subnsw_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_subnsw_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_subnsw_i128: @@ -1079,35 +1098,38 @@ define i128 @abd_subnsw_i128(i128 %a, i128 %b) nounwind { define i128 @abd_subnsw_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_subnsw_i128_undef: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_subnsw_i128_undef: @@ -1282,37 +1304,41 @@ define i128 @abd_select_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_select_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %edi, %eax -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll (%esp), %edx # 4-byte Folded Reload -; X86-NEXT: cmovll %ebx, %esi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: cmovll %eax, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: subl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: subl 24(%ebp), %ecx +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: sbbl 36(%ebp), %ebx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovll %edi, %esi +; X86-NEXT: cmovll %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/abdu-neg.ll b/llvm/test/CodeGen/X86/abdu-neg.ll index 6bda99c89a37..b7c34070f1af 100644 --- a/llvm/test/CodeGen/X86/abdu-neg.ll +++ b/llvm/test/CodeGen/X86/abdu-neg.ll @@ -355,39 +355,43 @@ define i128 @abd_ext_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %edi, %edi -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebp, %ebp -; X86-NEXT: xorl %ebp, %ecx -; X86-NEXT: xorl %ebp, %esi -; X86-NEXT: xorl %ebp, %ebx -; X86-NEXT: xorl %ebp, %edx -; X86-NEXT: subl %ebp, %edx -; X86-NEXT: sbbl %ebp, %ebx -; X86-NEXT: sbbl %ebp, %esi -; X86-NEXT: sbbl %ebp, %ecx -; X86-NEXT: negl %edx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebx, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: xorl %edx, %edx +; X86-NEXT: subl 40(%ebp), %ecx +; X86-NEXT: sbbl 44(%ebp), %edi +; X86-NEXT: sbbl 48(%ebp), %esi +; X86-NEXT: sbbl 52(%ebp), %eax ; X86-NEXT: movl $0, %ebx -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: sbbl %ecx, %edi -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %ebx, 8(%eax) -; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: sbbl %ebx, %ebx +; X86-NEXT: xorl %ebx, %eax +; X86-NEXT: xorl %ebx, %esi +; X86-NEXT: xorl %ebx, %edi +; X86-NEXT: xorl %ebx, %ecx +; X86-NEXT: subl %ebx, %ecx +; X86-NEXT: sbbl %ebx, %edi +; X86-NEXT: sbbl %ebx, %esi +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: negl %ecx +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl $0, %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: movl %ebx, 4(%eax) +; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -423,39 +427,43 @@ define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %edi, %edi -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebp, %ebp -; X86-NEXT: xorl %ebp, %ecx -; X86-NEXT: xorl %ebp, %esi -; X86-NEXT: xorl %ebp, %ebx -; X86-NEXT: xorl %ebp, %edx -; X86-NEXT: subl %ebp, %edx -; X86-NEXT: sbbl %ebp, %ebx -; X86-NEXT: sbbl %ebp, %esi -; X86-NEXT: sbbl %ebp, %ecx -; X86-NEXT: negl %edx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebx, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: xorl %edx, %edx +; X86-NEXT: subl 40(%ebp), %ecx +; X86-NEXT: sbbl 44(%ebp), %edi +; X86-NEXT: sbbl 48(%ebp), %esi +; X86-NEXT: sbbl 52(%ebp), %eax +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %ebx, %ebx +; X86-NEXT: xorl %ebx, %eax +; X86-NEXT: xorl %ebx, %esi +; X86-NEXT: xorl %ebx, %edi +; X86-NEXT: xorl %ebx, %ecx +; X86-NEXT: subl %ebx, %ecx +; X86-NEXT: sbbl %ebx, %edi +; X86-NEXT: sbbl %ebx, %esi +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: negl %ecx ; X86-NEXT: movl $0, %ebx -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: sbbl %ecx, %edi -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %ebx, 8(%eax) -; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl $0, %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: movl %ebx, 4(%eax) +; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -621,55 +629,59 @@ define i128 @abd_minmax_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_minmax_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: cmpl %eax, %esi -; X86-NEXT: sbbl %ebx, %ecx -; X86-NEXT: movl %edx, %ecx -; X86-NEXT: sbbl %ebp, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl %edx, %ecx -; X86-NEXT: sbbl %edi, %ecx -; X86-NEXT: movl %edi, %ecx -; X86-NEXT: cmovbl %edx, %ecx -; X86-NEXT: movl %ecx, (%esp) # 4-byte Spill -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl %ebx, %ecx -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %eax, %edx -; X86-NEXT: cmovbl %esi, %edx -; X86-NEXT: cmpl %esi, %eax -; X86-NEXT: movl %ebx, %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl %edi, %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %eax -; X86-NEXT: subl %eax, %edx -; X86-NEXT: sbbl %ebx, %ecx -; X86-NEXT: sbbl %esi, %ebp -; X86-NEXT: movl (%esp), %esi # 4-byte Reload -; X86-NEXT: sbbl %edi, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: movl %ecx, 4(%eax) -; X86-NEXT: movl %ebp, 8(%eax) -; X86-NEXT: movl %esi, 12(%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %esi +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: cmpl %esi, %edi +; X86-NEXT: sbbl 44(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %eax +; X86-NEXT: sbbl %edx, %eax +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: movl %ebx, %eax +; X86-NEXT: cmovbl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %edx, %eax +; X86-NEXT: cmovbl 32(%ebp), %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 44(%ebp), %eax +; X86-NEXT: cmovbl 28(%ebp), %eax +; X86-NEXT: movl %esi, %ecx +; X86-NEXT: cmovbl %edi, %ecx +; X86-NEXT: cmpl %edi, %esi +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: sbbl 28(%ebp), %edi +; X86-NEXT: movl %edx, %edi +; X86-NEXT: sbbl 32(%ebp), %edi +; X86-NEXT: movl %ebx, %edi +; X86-NEXT: sbbl 36(%ebp), %edi +; X86-NEXT: cmovbl 36(%ebp), %ebx +; X86-NEXT: cmovbl 32(%ebp), %edx +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: cmovbl 28(%ebp), %edi +; X86-NEXT: cmovbl 24(%ebp), %esi +; X86-NEXT: subl %esi, %ecx +; X86-NEXT: sbbl %edi, %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-NEXT: sbbl %edx, %edi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: sbbl %ebx, %esi +; X86-NEXT: movl 8(%ebp), %edx +; X86-NEXT: movl %ecx, (%edx) +; X86-NEXT: movl %eax, 4(%edx) +; X86-NEXT: movl %edi, 8(%edx) +; X86-NEXT: movl %esi, 12(%edx) +; X86-NEXT: movl %edx, %eax +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -827,39 +839,43 @@ define i128 @abd_cmp_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_cmp_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %edi, %edi -; X86-NEXT: subl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebp, %ebp -; X86-NEXT: xorl %ebp, %ecx -; X86-NEXT: xorl %ebp, %esi -; X86-NEXT: xorl %ebp, %ebx -; X86-NEXT: xorl %ebp, %edx -; X86-NEXT: subl %ebp, %edx -; X86-NEXT: sbbl %ebp, %ebx -; X86-NEXT: sbbl %ebp, %esi -; X86-NEXT: sbbl %ebp, %ecx -; X86-NEXT: negl %edx -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %ebx, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: xorl %edx, %edx +; X86-NEXT: subl 40(%ebp), %ecx +; X86-NEXT: sbbl 44(%ebp), %edi +; X86-NEXT: sbbl 48(%ebp), %esi +; X86-NEXT: sbbl 52(%ebp), %eax +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %ebx, %ebx +; X86-NEXT: xorl %ebx, %eax +; X86-NEXT: xorl %ebx, %esi +; X86-NEXT: xorl %ebx, %edi +; X86-NEXT: xorl %ebx, %ecx +; X86-NEXT: subl %ebx, %ecx +; X86-NEXT: sbbl %ebx, %edi +; X86-NEXT: sbbl %ebx, %esi +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: negl %ecx ; X86-NEXT: movl $0, %ebx -; X86-NEXT: sbbl %esi, %ebx -; X86-NEXT: sbbl %ecx, %edi -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %ebx, 8(%eax) -; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl $0, %edi +; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: movl %ebx, 4(%eax) +; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/abdu.ll b/llvm/test/CodeGen/X86/abdu.ll index 27acec32fd34..043c9155f52f 100644 --- a/llvm/test/CodeGen/X86/abdu.ll +++ b/llvm/test/CodeGen/X86/abdu.ll @@ -326,35 +326,38 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind { define i128 @abd_ext_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl %ebx, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: sbbl %eax, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_ext_i128: @@ -381,35 +384,38 @@ define i128 @abd_ext_i128(i128 %a, i128 %b) nounwind { define i128 @abd_ext_i128_undef(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_ext_i128_undef: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl %ebx, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: sbbl %eax, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_ext_i128_undef: @@ -548,35 +554,38 @@ define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind { define i128 @abd_minmax_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_minmax_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl %ebx, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: sbbl %eax, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_minmax_i128: @@ -717,35 +726,38 @@ define i64 @abd_cmp_i64(i64 %a, i64 %b) nounwind { define i128 @abd_cmp_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_cmp_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl %ebx, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: sbbl %eax, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_cmp_i128: @@ -887,35 +899,38 @@ define i64 @abd_select_i64(i64 %a, i64 %b) nounwind { define i128 @abd_select_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: abd_select_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: subl {{[0-9]+}}(%esp), %edi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: sbbl %ebx, %ebx -; X86-NEXT: xorl %ebx, %ecx -; X86-NEXT: xorl %ebx, %edx -; X86-NEXT: xorl %ebx, %esi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: subl %ebx, %edi -; X86-NEXT: sbbl %ebx, %esi -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl %ebx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: subl 40(%ebp), %edi +; X86-NEXT: sbbl 44(%ebp), %esi +; X86-NEXT: sbbl 48(%ebp), %edx +; X86-NEXT: sbbl 52(%ebp), %ecx +; X86-NEXT: sbbl %eax, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: abd_select_i128: diff --git a/llvm/test/CodeGen/X86/abs.ll b/llvm/test/CodeGen/X86/abs.ll index bae140abdf6b..e252d5953e60 100644 --- a/llvm/test/CodeGen/X86/abs.ll +++ b/llvm/test/CodeGen/X86/abs.ll @@ -144,31 +144,34 @@ define i128 @test_i128(i128 %a) nounwind { ; ; X86-LABEL: test_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %edx -; X86-NEXT: sarl $31, %edx -; X86-NEXT: xorl %edx, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: xorl %edx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: xorl %edx, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: xorl %edx, %ebx -; X86-NEXT: subl %edx, %ebx -; X86-NEXT: sbbl %edx, %edi -; X86-NEXT: sbbl %edx, %esi -; X86-NEXT: sbbl %edx, %ecx -; X86-NEXT: movl %ebx, (%eax) -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edi, (%eax) +; X86-NEXT: movl %esi, 4(%eax) +; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %r = call i128 @llvm.abs.i128(i128 %a, i1 false) ret i128 %r @@ -688,13 +691,17 @@ define i128 @test_sextinreg_i128(i128 %a) nounwind { ; ; X86-LABEL: test_sextinreg_i128: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %ecx ; X86-NEXT: movl %ecx, %edx ; X86-NEXT: sarl $31, %edx ; X86-NEXT: xorl %edx, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl 24(%ebp), %esi ; X86-NEXT: xorl %edx, %esi ; X86-NEXT: subl %edx, %esi ; X86-NEXT: sbbl %edx, %ecx @@ -702,7 +709,9 @@ define i128 @test_sextinreg_i128(i128 %a) nounwind { ; X86-NEXT: movl %ecx, 4(%eax) ; X86-NEXT: movl $0, 12(%eax) ; X86-NEXT: movl $0, 8(%eax) +; X86-NEXT: leal -4(%ebp), %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %shl = shl i128 %a, 64 %ashr = ashr exact i128 %shl, 64 diff --git a/llvm/test/CodeGen/X86/add-sub-bool.ll b/llvm/test/CodeGen/X86/add-sub-bool.ll index c2bfcf57185e..1df284fb9fe2 100644 --- a/llvm/test/CodeGen/X86/add-sub-bool.ll +++ b/llvm/test/CodeGen/X86/add-sub-bool.ll @@ -104,18 +104,21 @@ define i24 @test_i24_add_add_idx(i24 %x, i24 %y, i24 %z) nounwind { define i128 @test_i128_add_add_idx(i128 %x, i128 %y, i128 %z) nounwind { ; X86-LABEL: test_i128_add_add_idx: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: addl {{[0-9]+}}(%esp), %esi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %edi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: adcl {{[0-9]+}}(%esp), %edx -; X86-NEXT: btl $5, {{[0-9]+}}(%esp) +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: movl 40(%ebp), %esi +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: addl 24(%ebp), %esi +; X86-NEXT: adcl 28(%ebp), %edi +; X86-NEXT: adcl 32(%ebp), %ecx +; X86-NEXT: adcl 36(%ebp), %edx +; X86-NEXT: btl $5, 64(%ebp) ; X86-NEXT: adcl $0, %esi ; X86-NEXT: adcl $0, %edi ; X86-NEXT: adcl $0, %ecx @@ -124,8 +127,10 @@ define i128 @test_i128_add_add_idx(i128 %x, i128 %y, i128 %z) nounwind { ; X86-NEXT: movl %esi, (%eax) ; X86-NEXT: movl %ecx, 8(%eax) ; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: test_i128_add_add_idx: diff --git a/llvm/test/CodeGen/X86/arg-copy-elide.ll b/llvm/test/CodeGen/X86/arg-copy-elide.ll index 0eb2c630e681..f13627b55856 100644 --- a/llvm/test/CodeGen/X86/arg-copy-elide.ll +++ b/llvm/test/CodeGen/X86/arg-copy-elide.ll @@ -188,11 +188,11 @@ define void @split_i128(ptr %sret, i128 %x) { ; CHECK-NEXT: pushl %esi ; CHECK-NEXT: andl $-16, %esp ; CHECK-NEXT: subl $48, %esp -; CHECK-NEXT: movl 12(%ebp), %eax +; CHECK-NEXT: movl 24(%ebp), %eax ; CHECK-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; CHECK-NEXT: movl 16(%ebp), %ebx -; CHECK-NEXT: movl 20(%ebp), %esi -; CHECK-NEXT: movl 24(%ebp), %edi +; CHECK-NEXT: movl 28(%ebp), %ebx +; CHECK-NEXT: movl 32(%ebp), %esi +; CHECK-NEXT: movl 36(%ebp), %edi ; CHECK-NEXT: movl %edi, {{[0-9]+}}(%esp) ; CHECK-NEXT: movl %esi, {{[0-9]+}}(%esp) ; CHECK-NEXT: movl %ebx, {{[0-9]+}}(%esp) diff --git a/llvm/test/CodeGen/X86/avx512fp16-cvt.ll b/llvm/test/CodeGen/X86/avx512fp16-cvt.ll index f66f0c0ceabc..cc58bc1e44f3 100644 --- a/llvm/test/CodeGen/X86/avx512fp16-cvt.ll +++ b/llvm/test/CodeGen/X86/avx512fp16-cvt.ll @@ -628,13 +628,19 @@ define half @s128_to_half(i128 %x) { ; ; X86-LABEL: s128_to_half: ; X86: # %bb.0: -; X86-NEXT: subl $16, %esp -; X86-NEXT: .cfi_def_cfa_offset 20 -; X86-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-NEXT: pushl %ebp +; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $32, %esp +; X86-NEXT: vmovups 8(%ebp), %xmm0 ; X86-NEXT: vmovups %xmm0, (%esp) ; X86-NEXT: calll __floattihf -; X86-NEXT: addl $16, %esp -; X86-NEXT: .cfi_def_cfa_offset 4 +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl %a = sitofp i128 %x to half ret half %a @@ -713,13 +719,19 @@ define half @u128_to_half(i128 %x) { ; ; X86-LABEL: u128_to_half: ; X86: # %bb.0: -; X86-NEXT: subl $16, %esp -; X86-NEXT: .cfi_def_cfa_offset 20 -; X86-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-NEXT: pushl %ebp +; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $32, %esp +; X86-NEXT: vmovups 8(%ebp), %xmm0 ; X86-NEXT: vmovups %xmm0, (%esp) ; X86-NEXT: calll __floatuntihf -; X86-NEXT: addl $16, %esp -; X86-NEXT: .cfi_def_cfa_offset 4 +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl %a = uitofp i128 %x to half ret half %a @@ -1020,11 +1032,15 @@ define half @f128_to_half(fp128 %x) nounwind { ; ; X86-LABEL: f128_to_half: ; X86: # %bb.0: -; X86-NEXT: subl $16, %esp -; X86-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $32, %esp +; X86-NEXT: vmovups 8(%ebp), %xmm0 ; X86-NEXT: vmovups %xmm0, (%esp) ; X86-NEXT: calll __trunctfhf2 -; X86-NEXT: addl $16, %esp +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl %a = fptrunc fp128 %x to half ret half %a diff --git a/llvm/test/CodeGen/X86/bitselect.ll b/llvm/test/CodeGen/X86/bitselect.ll index 4fc0827ac4dd..33381313d3c1 100644 --- a/llvm/test/CodeGen/X86/bitselect.ll +++ b/llvm/test/CodeGen/X86/bitselect.ll @@ -146,37 +146,40 @@ define i64 @bitselect_i64(i64 %a, i64 %b, i64 %m) nounwind { define i128 @bitselect_i128(i128 %a, i128 %b, i128 %m) nounwind { ; X86-LABEL: bitselect_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: xorl %edi, %ecx -; X86-NEXT: andl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: xorl %edi, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: andl {{[0-9]+}}(%esp), %edi -; X86-NEXT: xorl %ebx, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: xorl %esi, %ebx -; X86-NEXT: andl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: xorl %esi, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: xorl %edx, %esi -; X86-NEXT: andl {{[0-9]+}}(%esp), %esi -; X86-NEXT: xorl %edx, %esi -; X86-NEXT: movl %esi, 12(%eax) -; X86-NEXT: movl %ebx, 8(%eax) -; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: xorl %esi, %ecx +; X86-NEXT: andl 56(%ebp), %ecx +; X86-NEXT: xorl %esi, %ecx +; X86-NEXT: movl 44(%ebp), %esi +; X86-NEXT: xorl %edi, %esi +; X86-NEXT: andl 60(%ebp), %esi +; X86-NEXT: xorl %edi, %esi +; X86-NEXT: movl 48(%ebp), %edi +; X86-NEXT: xorl %edx, %edi +; X86-NEXT: andl 64(%ebp), %edi +; X86-NEXT: xorl %edx, %edi +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: andl 68(%ebp), %edx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-NOBMI-LABEL: bitselect_i128: diff --git a/llvm/test/CodeGen/X86/bsf.ll b/llvm/test/CodeGen/X86/bsf.ll index 312f94c04123..143e10e6909e 100644 --- a/llvm/test/CodeGen/X86/bsf.ll +++ b/llvm/test/CodeGen/X86/bsf.ll @@ -263,70 +263,78 @@ define i128 @cmov_bsf128(i128 %x, i128 %y) nounwind { ; X86-LABEL: cmov_bsf128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl %edi, %edx -; X86-NEXT: orl %ebp, %edx -; X86-NEXT: movl %ecx, %esi -; X86-NEXT: orl %eax, %esi -; X86-NEXT: orl %edx, %esi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %ebx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: movl %esi, %eax +; X86-NEXT: orl %ebx, %eax +; X86-NEXT: movl %ecx, %edx +; X86-NEXT: orl %edi, %edx +; X86-NEXT: orl %eax, %edx ; X86-NEXT: je .LBB8_1 ; X86-NEXT: # %bb.2: # %cond.false ; X86-NEXT: testl %ecx, %ecx ; X86-NEXT: jne .LBB8_3 ; X86-NEXT: # %bb.4: # %cond.false -; X86-NEXT: rep bsfl %edi, %esi -; X86-NEXT: addl $32, %esi -; X86-NEXT: testl %eax, %eax -; X86-NEXT: je .LBB8_7 -; X86-NEXT: .LBB8_6: -; X86-NEXT: rep bsfl %eax, %edx -; X86-NEXT: jmp .LBB8_8 +; X86-NEXT: rep bsfl %esi, %eax +; X86-NEXT: addl $32, %eax +; X86-NEXT: jmp .LBB8_5 ; X86-NEXT: .LBB8_1: -; X86-NEXT: movl %ebp, %eax -; X86-NEXT: xorl %ebp, %ebp -; X86-NEXT: movl $128, %esi +; X86-NEXT: movl %ebx, %eax +; X86-NEXT: xorl %edi, %edi +; X86-NEXT: movl $128, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill ; X86-NEXT: jmp .LBB8_11 ; X86-NEXT: .LBB8_3: -; X86-NEXT: rep bsfl %ecx, %esi -; X86-NEXT: testl %eax, %eax +; X86-NEXT: rep bsfl %ecx, %eax +; X86-NEXT: .LBB8_5: # %cond.false +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: testl %edi, %edi ; X86-NEXT: jne .LBB8_6 -; X86-NEXT: .LBB8_7: # %cond.false -; X86-NEXT: rep bsfl %ebp, %edx +; X86-NEXT: # %bb.7: # %cond.false +; X86-NEXT: rep bsfl %ebx, %edx ; X86-NEXT: addl $32, %edx +; X86-NEXT: jmp .LBB8_8 +; X86-NEXT: .LBB8_6: +; X86-NEXT: rep bsfl %edi, %edx ; X86-NEXT: .LBB8_8: # %cond.false -; X86-NEXT: movl %ebp, %eax -; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, %eax +; X86-NEXT: movl %ecx, %edi +; X86-NEXT: orl %esi, %edi ; X86-NEXT: jne .LBB8_10 ; X86-NEXT: # %bb.9: # %cond.false ; X86-NEXT: addl $64, %edx -; X86-NEXT: movl %edx, %esi +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: .LBB8_10: # %cond.false -; X86-NEXT: xorl %ebp, %ebp +; X86-NEXT: xorl %edi, %edi ; X86-NEXT: .LBB8_11: # %cond.end -; X86-NEXT: xorl %ebx, %ebx ; X86-NEXT: xorl %edx, %edx -; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: orl %eax, %edi -; X86-NEXT: orl %ecx, %edi -; X86-NEXT: jne .LBB8_13 -; X86-NEXT: # %bb.12: -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: .LBB8_13: # %cond.end -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edx, 12(%eax) -; X86-NEXT: movl %ebx, 8(%eax) -; X86-NEXT: movl %ebp, 4(%eax) -; X86-NEXT: movl %esi, (%eax) +; X86-NEXT: xorl %ebx, %ebx +; X86-NEXT: orl 32(%ebp), %ecx +; X86-NEXT: orl %eax, %esi +; X86-NEXT: orl %ecx, %esi +; X86-NEXT: je .LBB8_12 +; X86-NEXT: # %bb.13: # %cond.end +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: jmp .LBB8_14 +; X86-NEXT: .LBB8_12: +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: .LBB8_14: # %cond.end +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -361,46 +369,49 @@ define i128 @cmov_bsf128_undef(i128 %x, i128 %y) nounwind { ; X86-LABEL: cmov_bsf128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl %ecx, %edi -; X86-NEXT: orl %esi, %edi -; X86-NEXT: movl %edx, %ebp -; X86-NEXT: orl %ebx, %ebp -; X86-NEXT: orl %edi, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: orl %esi, %eax +; X86-NEXT: movl %edx, %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: orl %eax, %ebx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: je .LBB9_11 ; X86-NEXT: # %bb.1: # %select.true.sink ; X86-NEXT: testl %edx, %edx ; X86-NEXT: jne .LBB9_2 ; X86-NEXT: # %bb.3: # %select.true.sink -; X86-NEXT: rep bsfl %ecx, %edi -; X86-NEXT: addl $32, %edi -; X86-NEXT: testl %ebx, %ebx +; X86-NEXT: rep bsfl %ecx, %ebx +; X86-NEXT: addl $32, %ebx +; X86-NEXT: testl %edi, %edi ; X86-NEXT: je .LBB9_6 ; X86-NEXT: .LBB9_5: -; X86-NEXT: rep bsfl %ebx, %esi +; X86-NEXT: rep bsfl %edi, %esi ; X86-NEXT: orl %ecx, %edx ; X86-NEXT: je .LBB9_8 ; X86-NEXT: jmp .LBB9_9 ; X86-NEXT: .LBB9_11: # %select.end -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl 52(%ebp), %ecx +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 44(%ebp), %esi +; X86-NEXT: movl 40(%ebp), %edi ; X86-NEXT: movl %edi, (%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) ; X86-NEXT: jmp .LBB9_10 ; X86-NEXT: .LBB9_2: -; X86-NEXT: rep bsfl %edx, %edi -; X86-NEXT: testl %ebx, %ebx +; X86-NEXT: rep bsfl %edx, %ebx +; X86-NEXT: testl %edi, %edi ; X86-NEXT: jne .LBB9_5 ; X86-NEXT: .LBB9_6: # %select.true.sink ; X86-NEXT: rep bsfl %esi, %esi @@ -409,13 +420,14 @@ define i128 @cmov_bsf128_undef(i128 %x, i128 %y) nounwind { ; X86-NEXT: jne .LBB9_9 ; X86-NEXT: .LBB9_8: # %select.true.sink ; X86-NEXT: addl $64, %esi -; X86-NEXT: movl %esi, %edi +; X86-NEXT: movl %esi, %ebx ; X86-NEXT: .LBB9_9: # %select.true.sink -; X86-NEXT: movl %edi, (%eax) +; X86-NEXT: movl %ebx, (%eax) ; X86-NEXT: movl $0, 12(%eax) ; X86-NEXT: movl $0, 8(%eax) ; X86-NEXT: movl $0, 4(%eax) ; X86-NEXT: .LBB9_10: # %select.true.sink +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/bsr.ll b/llvm/test/CodeGen/X86/bsr.ll index fbca4af425ea..ab0478a4e944 100644 --- a/llvm/test/CodeGen/X86/bsr.ll +++ b/llvm/test/CodeGen/X86/bsr.ll @@ -291,79 +291,80 @@ define i128 @cmov_bsr128(i128 %x, i128 %y) nounwind { ; X86-LABEL: cmov_bsr128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: pushl %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl %edi, %edx -; X86-NEXT: orl %ebp, %edx -; X86-NEXT: movl %ecx, %esi -; X86-NEXT: orl %ebx, %esi -; X86-NEXT: orl %edx, %esi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: movl %edi, %eax +; X86-NEXT: orl %esi, %eax +; X86-NEXT: movl %ecx, %edx +; X86-NEXT: orl %ebx, %edx +; X86-NEXT: orl %eax, %edx ; X86-NEXT: je .LBB8_1 ; X86-NEXT: # %bb.2: # %cond.false -; X86-NEXT: testl %ebp, %ebp +; X86-NEXT: testl %esi, %esi ; X86-NEXT: jne .LBB8_3 ; X86-NEXT: # %bb.4: # %cond.false -; X86-NEXT: bsrl %ebx, %edx -; X86-NEXT: xorl $31, %edx -; X86-NEXT: orl $32, %edx +; X86-NEXT: bsrl %ebx, %esi +; X86-NEXT: xorl $31, %esi +; X86-NEXT: orl $32, %esi ; X86-NEXT: testl %edi, %edi ; X86-NEXT: je .LBB8_7 ; X86-NEXT: .LBB8_6: -; X86-NEXT: bsrl %edi, %esi -; X86-NEXT: xorl $31, %esi +; X86-NEXT: bsrl %edi, %eax +; X86-NEXT: xorl $31, %eax ; X86-NEXT: jmp .LBB8_8 ; X86-NEXT: .LBB8_1: -; X86-NEXT: movl %ebx, %eax -; X86-NEXT: movl $0, (%esp) # 4-byte Folded Spill -; X86-NEXT: movl $128, %edx +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: movl $128, %esi ; X86-NEXT: jmp .LBB8_11 ; X86-NEXT: .LBB8_3: -; X86-NEXT: bsrl %ebp, %edx -; X86-NEXT: xorl $31, %edx +; X86-NEXT: bsrl %esi, %esi +; X86-NEXT: xorl $31, %esi ; X86-NEXT: testl %edi, %edi ; X86-NEXT: jne .LBB8_6 ; X86-NEXT: .LBB8_7: # %cond.false -; X86-NEXT: bsrl %ecx, %esi -; X86-NEXT: xorl $31, %esi -; X86-NEXT: orl $32, %esi +; X86-NEXT: bsrl %ecx, %eax +; X86-NEXT: xorl $31, %eax +; X86-NEXT: orl $32, %eax ; X86-NEXT: .LBB8_8: # %cond.false -; X86-NEXT: movl %ebx, %eax -; X86-NEXT: orl %ebp, %ebx +; X86-NEXT: movl %ebx, %edx +; X86-NEXT: orl 36(%ebp), %edx ; X86-NEXT: jne .LBB8_10 ; X86-NEXT: # %bb.9: # %cond.false -; X86-NEXT: orl $64, %esi -; X86-NEXT: movl %esi, %edx +; X86-NEXT: orl $64, %eax +; X86-NEXT: movl %eax, %esi ; X86-NEXT: .LBB8_10: # %cond.false -; X86-NEXT: movl $0, (%esp) # 4-byte Folded Spill +; X86-NEXT: xorl %eax, %eax ; X86-NEXT: .LBB8_11: # %cond.end -; X86-NEXT: xorl %esi, %esi ; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: orl %eax, %ecx -; X86-NEXT: orl %ebp, %edi +; X86-NEXT: xorl %edx, %edx +; X86-NEXT: orl 32(%ebp), %ecx +; X86-NEXT: orl 36(%ebp), %edi ; X86-NEXT: orl %ecx, %edi ; X86-NEXT: je .LBB8_12 ; X86-NEXT: # %bb.13: # %cond.end -; X86-NEXT: xorl $127, %edx -; X86-NEXT: movl (%esp), %ecx # 4-byte Reload +; X86-NEXT: xorl $127, %esi +; X86-NEXT: movl %eax, %ecx ; X86-NEXT: jmp .LBB8_14 ; X86-NEXT: .LBB8_12: -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: movl 48(%ebp), %ebx +; X86-NEXT: movl 44(%ebp), %ecx +; X86-NEXT: movl 40(%ebp), %esi ; X86-NEXT: .LBB8_14: # %cond.end -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %ebx, 12(%eax) -; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %ebx, 8(%eax) ; X86-NEXT: movl %ecx, 4(%eax) -; X86-NEXT: movl %edx, (%eax) -; X86-NEXT: addl $4, %esp +; X86-NEXT: movl %esi, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -398,62 +399,67 @@ define i128 @cmov_bsr128_undef(i128 %x, i128 %y) nounwind { ; X86-LABEL: cmov_bsr128_undef: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: testl %edi, %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: testl %eax, %eax ; X86-NEXT: jne .LBB9_1 ; X86-NEXT: # %bb.2: -; X86-NEXT: bsrl %esi, %ecx -; X86-NEXT: xorl $31, %ecx -; X86-NEXT: orl $32, %ecx +; X86-NEXT: bsrl %edi, %esi +; X86-NEXT: xorl $31, %esi +; X86-NEXT: orl $32, %esi ; X86-NEXT: jmp .LBB9_3 ; X86-NEXT: .LBB9_1: -; X86-NEXT: bsrl %edi, %ecx -; X86-NEXT: xorl $31, %ecx +; X86-NEXT: bsrl %eax, %esi +; X86-NEXT: xorl $31, %esi ; X86-NEXT: .LBB9_3: -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl 24(%ebp), %ebx ; X86-NEXT: testl %edx, %edx ; X86-NEXT: jne .LBB9_4 ; X86-NEXT: # %bb.5: -; X86-NEXT: bsrl %ebx, %ebp -; X86-NEXT: xorl $31, %ebp -; X86-NEXT: orl $32, %ebp -; X86-NEXT: jmp .LBB9_6 +; X86-NEXT: bsrl %ebx, %ecx +; X86-NEXT: xorl $31, %ecx +; X86-NEXT: orl $32, %ecx +; X86-NEXT: orl %eax, %edi +; X86-NEXT: je .LBB9_7 +; X86-NEXT: jmp .LBB9_8 ; X86-NEXT: .LBB9_4: -; X86-NEXT: bsrl %edx, %ebp -; X86-NEXT: xorl $31, %ebp -; X86-NEXT: .LBB9_6: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: orl %edi, %esi +; X86-NEXT: bsrl %edx, %ecx +; X86-NEXT: xorl $31, %ecx +; X86-NEXT: orl %eax, %edi ; X86-NEXT: jne .LBB9_8 -; X86-NEXT: # %bb.7: -; X86-NEXT: orl $64, %ebp -; X86-NEXT: movl %ebp, %ecx +; X86-NEXT: .LBB9_7: +; X86-NEXT: orl $64, %ecx +; X86-NEXT: movl %ecx, %esi ; X86-NEXT: .LBB9_8: -; X86-NEXT: orl %edi, %edx -; X86-NEXT: orl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: orl %eax, %edx +; X86-NEXT: orl 32(%ebp), %ebx ; X86-NEXT: orl %edx, %ebx ; X86-NEXT: jne .LBB9_9 ; X86-NEXT: # %bb.10: -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl 52(%ebp), %edi +; X86-NEXT: movl 40(%ebp), %esi +; X86-NEXT: movl 44(%ebp), %ecx ; X86-NEXT: jmp .LBB9_11 ; X86-NEXT: .LBB9_9: -; X86-NEXT: xorl $127, %ecx +; X86-NEXT: xorl $127, %esi +; X86-NEXT: xorl %ecx, %ecx ; X86-NEXT: xorl %edx, %edx -; X86-NEXT: xorl %esi, %esi ; X86-NEXT: xorl %edi, %edi ; X86-NEXT: .LBB9_11: +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %esi, 8(%eax) -; X86-NEXT: movl %edx, 4(%eax) -; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %ecx, 4(%eax) +; X86-NEXT: movl %esi, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/bswap-wide-int.ll b/llvm/test/CodeGen/X86/bswap-wide-int.ll index 6d5e995a6d57..673b7f16de75 100644 --- a/llvm/test/CodeGen/X86/bswap-wide-int.ll +++ b/llvm/test/CodeGen/X86/bswap-wide-int.ll @@ -41,13 +41,16 @@ define i64 @bswap_i64(i64 %a0) nounwind { define i128 @bswap_i128(i128 %a0) nounwind { ; X86-LABEL: bswap_i128: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %edi ; X86-NEXT: bswapl %edi ; X86-NEXT: bswapl %esi ; X86-NEXT: bswapl %edx @@ -56,25 +59,32 @@ define i128 @bswap_i128(i128 %a0) nounwind { ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edi, (%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X86-MOVBE-LABEL: bswap_i128: ; X86-MOVBE: # %bb.0: +; X86-MOVBE-NEXT: pushl %ebp +; X86-MOVBE-NEXT: movl %esp, %ebp ; X86-MOVBE-NEXT: pushl %edi ; X86-MOVBE-NEXT: pushl %esi -; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-MOVBE-NEXT: andl $-16, %esp +; X86-MOVBE-NEXT: movl 8(%ebp), %eax +; X86-MOVBE-NEXT: movl 32(%ebp), %ecx +; X86-MOVBE-NEXT: movl 36(%ebp), %edx +; X86-MOVBE-NEXT: movl 24(%ebp), %esi +; X86-MOVBE-NEXT: movl 28(%ebp), %edi ; X86-MOVBE-NEXT: movbel %esi, 12(%eax) ; X86-MOVBE-NEXT: movbel %edi, 8(%eax) ; X86-MOVBE-NEXT: movbel %ecx, 4(%eax) ; X86-MOVBE-NEXT: movbel %edx, (%eax) +; X86-MOVBE-NEXT: leal -8(%ebp), %esp ; X86-MOVBE-NEXT: popl %esi ; X86-MOVBE-NEXT: popl %edi +; X86-MOVBE-NEXT: popl %ebp ; X86-MOVBE-NEXT: retl $4 ; ; X64-LABEL: bswap_i128: diff --git a/llvm/test/CodeGen/X86/constructor.ll b/llvm/test/CodeGen/X86/constructor.ll index f46325db8c19..dca62acff6ed 100644 --- a/llvm/test/CodeGen/X86/constructor.ll +++ b/llvm/test/CodeGen/X86/constructor.ll @@ -5,7 +5,6 @@ ; RUN: llc -mtriple x86_64-pc-linux < %s | FileCheck --check-prefix=INIT-ARRAY %s ; RUN: llc -mtriple x86_64-unknown-freebsd < %s | FileCheck --check-prefix=INIT-ARRAY %s ; RUN: llc -mtriple x86_64-pc-solaris2.11 < %s | FileCheck --check-prefix=INIT-ARRAY %s -; RUN: llc -mtriple x86_64-unknown-nacl < %s | FileCheck --check-prefix=NACL %s ; RUN: llc -mtriple i586-intel-elfiamcu -use-ctors < %s | FileCheck %s --check-prefix=MCU-CTORS ; RUN: llc -mtriple i586-intel-elfiamcu < %s | FileCheck %s --check-prefix=MCU-INIT-ARRAY ; RUN: llc -mtriple x86_64-win32-gnu < %s | FileCheck --check-prefix=COFF-CTOR %s @@ -62,18 +61,6 @@ entry: ; INIT-ARRAY-NEXT: .quad i ; INIT-ARRAY-NEXT: .quad j -; NACL: .section .init_array.15,"awG",@init_array,v,comdat -; NACL-NEXT: .p2align 2 -; NACL-NEXT: .long g -; NACL-NEXT: .section .init_array.55555,"awG",@init_array,v,comdat -; NACL-NEXT: .p2align 2 -; NACL-NEXT: .long h -; NACL-NEXT: .section .init_array,"aw",@init_array -; NACL-NEXT: .p2align 2 -; NACL-NEXT: .long f -; NACL-NEXT: .long i -; NACL-NEXT: .long j - ; MCU-CTORS: .section .ctors,"aw",@progbits ; MCU-INIT-ARRAY: .section .init_array,"aw",@init_array diff --git a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll index d869f8ec01a5..661e7bb19641 100644 --- a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll +++ b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll @@ -152,17 +152,17 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp ; X86-NEXT: subl $176, %esp -; X86-NEXT: movl 20(%ebp), %edx -; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 36(%ebp), %ecx ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: sarl $31, %eax ; X86-NEXT: xorl %eax, %ecx ; X86-NEXT: movl %ecx, %edi ; X86-NEXT: xorl %eax, %edx ; X86-NEXT: movl %edx, %esi -; X86-NEXT: movl 16(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %edx ; X86-NEXT: xorl %eax, %edx -; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %ecx ; X86-NEXT: xorl %eax, %ecx ; X86-NEXT: subl %eax, %ecx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill @@ -172,16 +172,16 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: sbbl %eax, %edi ; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %ecx ; X86-NEXT: movl %ecx, %edx ; X86-NEXT: sarl $31, %edx ; X86-NEXT: movl %ecx, %esi ; X86-NEXT: xorl %edx, %esi -; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 48(%ebp), %ecx ; X86-NEXT: xorl %edx, %ecx -; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: movl 44(%ebp), %ebx ; X86-NEXT: xorl %edx, %ebx -; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: movl 40(%ebp), %edi ; X86-NEXT: xorl %edx, %edi ; X86-NEXT: subl %edx, %edi ; X86-NEXT: sbbl %edx, %ebx @@ -488,13 +488,13 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: sbbl %ecx, %ebx ; X86-NEXT: sbbl %ecx, %esi ; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 44(%ebp), %ecx +; X86-NEXT: movl 56(%ebp), %ecx ; X86-NEXT: movl %edx, (%ecx) ; X86-NEXT: movl %eax, 4(%ecx) ; X86-NEXT: movl %ebx, 8(%ecx) ; X86-NEXT: movl %esi, 12(%ecx) ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 28(%ebp), %ecx +; X86-NEXT: movl 40(%ebp), %ecx ; X86-NEXT: movl %ebx, %edi ; X86-NEXT: movl %edx, %esi ; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill @@ -508,7 +508,7 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: addl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload ; X86-NEXT: adcl $0, %ebx ; X86-NEXT: movl %esi, %eax -; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 44(%ebp), %esi ; X86-NEXT: mull %esi ; X86-NEXT: addl %ecx, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill @@ -523,17 +523,17 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: adcl %eax, %edx ; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: movl 40(%ebp), %eax ; X86-NEXT: imull %eax, %ebx ; X86-NEXT: mull %edi ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: imull %esi, %edi ; X86-NEXT: addl %edx, %edi ; X86-NEXT: addl %ebx, %edi -; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %eax ; X86-NEXT: movl %eax, %esi ; X86-NEXT: imull {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload -; X86-NEXT: movl 40(%ebp), %ebx +; X86-NEXT: movl 52(%ebp), %ebx ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload ; X86-NEXT: imull %edx, %ebx ; X86-NEXT: mull %edx @@ -543,13 +543,13 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: adcl %edi, %ebx ; X86-NEXT: addl %ecx, %eax ; X86-NEXT: adcl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload -; X86-NEXT: movl 12(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %edx ; X86-NEXT: subl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload -; X86-NEXT: movl 16(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %ecx ; X86-NEXT: sbbl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload -; X86-NEXT: movl 20(%ebp), %edi +; X86-NEXT: movl 32(%ebp), %edi ; X86-NEXT: sbbl %eax, %edi -; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %esi ; X86-NEXT: sbbl %ebx, %esi ; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edx, (%eax) diff --git a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll index 7bbddefd8272..370e1c608e44 100644 --- a/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll +++ b/llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll @@ -152,26 +152,26 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp ; X86-NEXT: subl $160, %esp -; X86-NEXT: movl 28(%ebp), %ebx -; X86-NEXT: movl 40(%ebp), %esi -; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl 40(%ebp), %ebx +; X86-NEXT: movl 52(%ebp), %esi +; X86-NEXT: movl 44(%ebp), %edi ; X86-NEXT: movl %edi, %eax ; X86-NEXT: orl %esi, %eax ; X86-NEXT: movl %ebx, %ecx -; X86-NEXT: orl 36(%ebp), %ecx +; X86-NEXT: orl 48(%ebp), %ecx ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: sete %cl -; X86-NEXT: movl 16(%ebp), %eax -; X86-NEXT: orl 24(%ebp), %eax -; X86-NEXT: movl 12(%ebp), %edx -; X86-NEXT: orl 20(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: orl 36(%ebp), %eax +; X86-NEXT: movl 24(%ebp), %edx +; X86-NEXT: orl 32(%ebp), %edx ; X86-NEXT: orl %eax, %edx ; X86-NEXT: sete %al ; X86-NEXT: orb %cl, %al ; X86-NEXT: movb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill ; X86-NEXT: bsrl %esi, %edx ; X86-NEXT: xorl $31, %edx -; X86-NEXT: bsrl 36(%ebp), %ecx +; X86-NEXT: bsrl 48(%ebp), %ecx ; X86-NEXT: xorl $31, %ecx ; X86-NEXT: addl $32, %ecx ; X86-NEXT: testl %esi, %esi @@ -184,325 +184,310 @@ define i128 @scalar_i128(i128 %x, i128 %y, ptr %divdst) nounwind { ; X86-NEXT: testl %edi, %edi ; X86-NEXT: cmovnel %edx, %eax ; X86-NEXT: addl $64, %eax -; X86-NEXT: movl 36(%ebp), %edx +; X86-NEXT: movl 48(%ebp), %edx ; X86-NEXT: orl %esi, %edx ; X86-NEXT: cmovnel %ecx, %eax -; X86-NEXT: movl 24(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %ebx ; X86-NEXT: bsrl %ebx, %edx ; X86-NEXT: xorl $31, %edx -; X86-NEXT: movl 20(%ebp), %ecx +; X86-NEXT: movl 32(%ebp), %ecx ; X86-NEXT: bsrl %ecx, %ecx ; X86-NEXT: xorl $31, %ecx ; X86-NEXT: addl $32, %ecx ; X86-NEXT: testl %ebx, %ebx ; X86-NEXT: cmovnel %edx, %ecx -; X86-NEXT: movl 16(%ebp), %edi +; X86-NEXT: movl 28(%ebp), %edi ; X86-NEXT: bsrl %edi, %esi ; X86-NEXT: xorl $31, %esi -; X86-NEXT: bsrl 12(%ebp), %edx +; X86-NEXT: bsrl 24(%ebp), %edx ; X86-NEXT: xorl $31, %edx ; X86-NEXT: addl $32, %edx ; X86-NEXT: testl %edi, %edi ; X86-NEXT: cmovnel %esi, %edx ; X86-NEXT: addl $64, %edx -; X86-NEXT: movl 20(%ebp), %edi -; X86-NEXT: movl %edi, %esi +; X86-NEXT: movl 32(%ebp), %esi ; X86-NEXT: orl %ebx, %esi ; X86-NEXT: cmovnel %ecx, %edx +; X86-NEXT: xorl %edi, %edi ; X86-NEXT: subl %edx, %eax -; X86-NEXT: movl $0, %edx -; X86-NEXT: sbbl %edx, %edx ; X86-NEXT: movl $0, %ebx ; X86-NEXT: sbbl %ebx, %ebx +; X86-NEXT: movl $0, %ecx +; X86-NEXT: sbbl %ecx, %ecx ; X86-NEXT: movl $0, %esi ; X86-NEXT: sbbl %esi, %esi -; X86-NEXT: movl $127, %ecx -; X86-NEXT: cmpl %eax, %ecx -; X86-NEXT: movl $0, %ecx -; X86-NEXT: sbbl %edx, %ecx -; X86-NEXT: movl $0, %ecx -; X86-NEXT: sbbl %ebx, %ecx -; X86-NEXT: movl $0, %ecx -; X86-NEXT: sbbl %esi, %ecx -; X86-NEXT: setb %cl -; X86-NEXT: orb {{[-0-9]+}}(%e{{[sb]}}p), %cl # 1-byte Folded Reload +; X86-NEXT: movl $127, %edx ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: xorl $127, %eax +; X86-NEXT: cmpl %eax, %edx +; X86-NEXT: movl $0, %edx ; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: orl %ebx, %eax -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: orl %esi, %edx -; X86-NEXT: orl %eax, %edx -; X86-NEXT: sete %al -; X86-NEXT: testb %cl, %cl -; X86-NEXT: movb %cl, %ah -; X86-NEXT: movl 24(%ebp), %ebx -; X86-NEXT: movl $0, %esi -; X86-NEXT: cmovnel %esi, %ebx -; X86-NEXT: movl %edi, %ecx -; X86-NEXT: cmovnel %esi, %ecx +; X86-NEXT: sbbl %ebx, %edx ; X86-NEXT: movl $0, %edx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 16(%ebp), %esi -; X86-NEXT: cmovnel %edx, %esi -; X86-NEXT: movl 12(%ebp), %edi -; X86-NEXT: movl %edi, %ecx -; X86-NEXT: cmovnel %edx, %ecx -; X86-NEXT: orb %ah, %al -; X86-NEXT: movl 44(%ebp), %eax -; X86-NEXT: jne .LBB4_7 -; X86-NEXT: # %bb.1: # %udiv-bb1 -; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: sbbl %ecx, %edx +; X86-NEXT: movl $0, %edx +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: sbbl %esi, %edx +; X86-NEXT: setb %dl +; X86-NEXT: orb {{[-0-9]+}}(%e{{[sb]}}p), %dl # 1-byte Folded Reload +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: cmovnel %edi, %eax +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: cmovnel %edi, %esi +; X86-NEXT: movl 28(%ebp), %edx +; X86-NEXT: cmovnel %edi, %edx +; X86-NEXT: movl 24(%ebp), %ebx +; X86-NEXT: cmovnel %edi, %ebx +; X86-NEXT: movl 56(%ebp), %edi +; X86-NEXT: jne .LBB4_8 +; X86-NEXT: # %bb.1: # %_udiv-special-cases +; X86-NEXT: movl %eax, %edi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: xorl $127, %eax +; X86-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: orl %eax, %ecx +; X86-NEXT: movl %edi, %eax +; X86-NEXT: movl 56(%ebp), %edi +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: je .LBB4_8 +; X86-NEXT: # %bb.2: # %udiv-bb1 +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) ; X86-NEXT: xorps %xmm0, %xmm0 ; X86-NEXT: movaps %xmm0, {{[0-9]+}}(%esp) -; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %eax ; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) -; X86-NEXT: movl 20(%ebp), %edx -; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) -; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 32(%ebp), %eax ; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: movl %edx, %ecx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, %ecx ; X86-NEXT: xorb $127, %cl ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: shrb $3, %al ; X86-NEXT: andb $12, %al ; X86-NEXT: negb %al ; X86-NEXT: movsbl %al, %eax -; X86-NEXT: movl 136(%esp,%eax), %edi -; X86-NEXT: movl 140(%esp,%eax), %esi -; X86-NEXT: shldl %cl, %edi, %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 136(%esp,%eax), %esi +; X86-NEXT: movl 140(%esp,%eax), %edx +; X86-NEXT: shldl %cl, %esi, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl 128(%esp,%eax), %ebx -; X86-NEXT: movl 132(%esp,%eax), %eax -; X86-NEXT: shldl %cl, %eax, %edi -; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %eax, %edi -; X86-NEXT: shldl %cl, %ebx, %edi +; X86-NEXT: movl 132(%esp,%eax), %edx +; X86-NEXT: shldl %cl, %edx, %esi +; X86-NEXT: shldl %cl, %ebx, %edx ; X86-NEXT: shll %cl, %ebx -; X86-NEXT: movl %ebx, %ecx -; X86-NEXT: addl $1, %edx -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: addl $1, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-NEXT: adcl $0, %eax -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload -; X86-NEXT: adcl $0, %esi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: adcl $0, %ecx ; X86-NEXT: adcl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; X86-NEXT: movl 20(%ebp), %ebx -; X86-NEXT: jae .LBB4_2 -; X86-NEXT: # %bb.5: -; X86-NEXT: xorl %edx, %edx -; X86-NEXT: xorl %eax, %eax -; X86-NEXT: movl %edi, %esi -; X86-NEXT: jmp .LBB4_6 -; X86-NEXT: .LBB4_2: # %udiv-preheader +; X86-NEXT: jae .LBB4_3 +; X86-NEXT: # %bb.6: +; X86-NEXT: movl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: movl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: jmp .LBB4_7 +; X86-NEXT: .LBB4_3: # %udiv-preheader ; X86-NEXT: movaps %xmm0, {{[0-9]+}}(%esp) -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 12(%ebp), %edx -; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) -; X86-NEXT: movl 16(%ebp), %edx -; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) -; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl 36(%ebp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 24(%ebp), %eax -; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload -; X86-NEXT: # kill: def $al killed $al killed $eax +; X86-NEXT: movl %ecx, %eax ; X86-NEXT: shrb $3, %al ; X86-NEXT: andb $12, %al ; X86-NEXT: movzbl %al, %eax -; X86-NEXT: movl 92(%esp,%eax), %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 88(%esp,%eax), %edx -; X86-NEXT: movl %edx, %ebx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-NEXT: shrdl %cl, %esi, %ebx -; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 92(%esp,%eax), %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 88(%esp,%eax), %edi ; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 80(%esp,%eax), %edi -; X86-NEXT: movl 84(%esp,%eax), %eax -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: movl %eax, %esi -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: shrdl %cl, %edx, %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: shrl %cl, %edx -; X86-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-NEXT: shrdl %cl, %eax, %edi +; X86-NEXT: shrdl %cl, %edx, %edi ; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 80(%esp,%eax), %edx +; X86-NEXT: movl 84(%esp,%eax), %eax +; X86-NEXT: movl %eax, %ebx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-NEXT: shrdl %cl, %edi, %ebx +; X86-NEXT: shrl %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: shrdl %cl, %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 40(%ebp), %eax ; X86-NEXT: addl $-1, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 32(%ebp), %eax +; X86-NEXT: movl 44(%ebp), %eax ; X86-NEXT: adcl $-1, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 36(%ebp), %esi -; X86-NEXT: adcl $-1, %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 40(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %eax +; X86-NEXT: adcl $-1, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 52(%ebp), %eax ; X86-NEXT: adcl $-1, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: xorl %eax, %eax ; X86-NEXT: movl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: movl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: .p2align 4 -; X86-NEXT: .LBB4_3: # %udiv-do-while +; X86-NEXT: .LBB4_4: # %udiv-do-while ; X86-NEXT: # =>This Inner Loop Header: Depth=1 -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload -; X86-NEXT: shldl $1, %esi, %edx -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload -; X86-NEXT: shldl $1, %edi, %esi -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: shldl $1, %ebx, %edi ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: shldl $1, %edx, %ebx -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: shldl $1, %ecx, %edx -; X86-NEXT: orl %eax, %edx +; X86-NEXT: shldl $1, %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: shldl $1, %ebx, %edx ; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: shldl $1, %edx, %ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-NEXT: shldl $1, %edi, %ebx +; X86-NEXT: shldl $1, %ecx, %edi +; X86-NEXT: shldl $1, %esi, %ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: shldl $1, %ecx, %edx -; X86-NEXT: orl %eax, %edx -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: addl %ecx, %ecx -; X86-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: shldl $1, %ecx, %esi +; X86-NEXT: orl %eax, %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: shldl $1, %esi, %ecx +; X86-NEXT: orl %eax, %ecx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: cmpl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: addl %esi, %esi +; X86-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: cmpl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: sbbl %edi, %ecx +; X86-NEXT: sbbl %ebx, %ecx ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: sbbl %esi, %ecx +; X86-NEXT: sbbl %edx, %ecx ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: sbbl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload ; X86-NEXT: sarl $31, %ecx ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: andl $1, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %ecx, %esi +; X86-NEXT: andl 52(%ebp), %esi ; X86-NEXT: movl %ecx, %eax -; X86-NEXT: andl 40(%ebp), %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %ecx, %eax -; X86-NEXT: andl 36(%ebp), %eax +; X86-NEXT: andl 48(%ebp), %eax ; X86-NEXT: movl %ecx, %edx -; X86-NEXT: andl 32(%ebp), %edx -; X86-NEXT: andl 28(%ebp), %ecx -; X86-NEXT: subl %ecx, %ebx -; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: sbbl %edx, %edi +; X86-NEXT: andl 44(%ebp), %edx +; X86-NEXT: andl 40(%ebp), %ecx +; X86-NEXT: subl %ecx, %edi ; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: sbbl %eax, %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: sbbl %edx, %ebx +; X86-NEXT: sbbl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: sbbl %esi, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: sbbl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: addl $-1, %ecx ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-NEXT: adcl $-1, %eax -; X86-NEXT: adcl $-1, %ebx -; X86-NEXT: adcl $-1, %esi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: adcl $-1, %edx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-NEXT: adcl $-1, %edi ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: orl %esi, %eax +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: orl %edi, %eax ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: orl %ebx, %ecx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: orl %edx, %ecx ; X86-NEXT: orl %eax, %ecx -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload -; X86-NEXT: jne .LBB4_3 -; X86-NEXT: # %bb.4: -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: jne .LBB4_4 +; X86-NEXT: # %bb.5: ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: .LBB4_6: # %udiv-loop-exit ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload -; X86-NEXT: shldl $1, %edi, %ebx -; X86-NEXT: orl %eax, %ebx -; X86-NEXT: shldl $1, %esi, %edi -; X86-NEXT: orl %eax, %edi -; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: shldl $1, %ecx, %esi -; X86-NEXT: orl %eax, %esi -; X86-NEXT: addl %ecx, %ecx -; X86-NEXT: orl %edx, %ecx -; X86-NEXT: movl 44(%ebp), %eax -; X86-NEXT: .LBB4_7: # %udiv-end -; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: movl %edx, 8(%eax) -; X86-NEXT: movl %ebx, 12(%eax) -; X86-NEXT: movl %esi, %edx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: movl 56(%ebp), %edi +; X86-NEXT: .LBB4_7: # %udiv-loop-exit +; X86-NEXT: shldl $1, %esi, %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: orl %ecx, %eax +; X86-NEXT: shldl $1, %edx, %esi +; X86-NEXT: orl %ecx, %esi +; X86-NEXT: shldl $1, %ebx, %edx +; X86-NEXT: orl %ecx, %edx +; X86-NEXT: addl %ebx, %ebx +; X86-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload +; X86-NEXT: .LBB4_8: # %udiv-end ; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %ebx, (%edi) +; X86-NEXT: movl %edx, 4(%edi) +; X86-NEXT: movl %esi, 8(%edi) +; X86-NEXT: movl %eax, 12(%edi) +; X86-NEXT: movl %eax, %ecx +; X86-NEXT: movl 48(%ebp), %eax ; X86-NEXT: movl %eax, %esi ; X86-NEXT: imull %edx, %esi -; X86-NEXT: mull %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: mull %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: addl %esi, %edx -; X86-NEXT: movl 40(%ebp), %edi -; X86-NEXT: imull %ecx, %edi +; X86-NEXT: movl 52(%ebp), %edi +; X86-NEXT: imull %ebx, %edi ; X86-NEXT: addl %edx, %edi -; X86-NEXT: movl 28(%ebp), %eax -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload -; X86-NEXT: mull %esi -; X86-NEXT: movl %eax, %ecx -; X86-NEXT: imull 28(%ebp), %ebx -; X86-NEXT: addl %edx, %ebx -; X86-NEXT: movl 32(%ebp), %edx -; X86-NEXT: imull %edx, %esi -; X86-NEXT: addl %ebx, %esi -; X86-NEXT: addl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: adcl %edi, %esi +; X86-NEXT: movl 40(%ebp), %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: mull %ebx +; X86-NEXT: movl %eax, %esi +; X86-NEXT: imull 40(%ebp), %ecx +; X86-NEXT: addl %edx, %ecx +; X86-NEXT: movl 44(%ebp), %eax +; X86-NEXT: imull %eax, %ebx +; X86-NEXT: addl %ecx, %ebx +; X86-NEXT: addl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload ; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload -; X86-NEXT: movl %edi, %eax -; X86-NEXT: movl 28(%ebp), %ecx +; X86-NEXT: adcl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: movl %esi, %eax +; X86-NEXT: movl 40(%ebp), %ecx ; X86-NEXT: mull %ecx -; X86-NEXT: movl %edx, %esi +; X86-NEXT: movl %edx, %edi ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: movl %ebx, %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-NEXT: mull %ecx +; X86-NEXT: movl %edx, %ebx ; X86-NEXT: movl %eax, %ecx -; X86-NEXT: addl %esi, %ecx -; X86-NEXT: adcl $0, %edx -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl %edi, %eax -; X86-NEXT: mull 32(%ebp) -; X86-NEXT: movl 16(%ebp), %esi +; X86-NEXT: addl %edi, %ecx +; X86-NEXT: adcl $0, %ebx +; X86-NEXT: movl %esi, %eax +; X86-NEXT: mull 44(%ebp) +; X86-NEXT: movl 28(%ebp), %esi ; X86-NEXT: movl %edx, %edi ; X86-NEXT: addl %ecx, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: adcl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload +; X86-NEXT: adcl %ebx, %edi ; X86-NEXT: setb %cl -; X86-NEXT: movl %ebx, %eax -; X86-NEXT: mull 32(%ebp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: mull 44(%ebp) ; X86-NEXT: addl %edi, %eax ; X86-NEXT: movzbl %cl, %ecx ; X86-NEXT: adcl %ecx, %edx ; X86-NEXT: addl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload ; X86-NEXT: adcl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload -; X86-NEXT: movl 12(%ebp), %ebx +; X86-NEXT: movl 24(%ebp), %ebx ; X86-NEXT: subl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload ; X86-NEXT: sbbl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload -; X86-NEXT: movl 20(%ebp), %edi +; X86-NEXT: movl 32(%ebp), %edi ; X86-NEXT: sbbl %eax, %edi -; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %ecx ; X86-NEXT: sbbl %edx, %ecx ; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %ebx, (%eax) diff --git a/llvm/test/CodeGen/X86/dollar-name-asm.ll b/llvm/test/CodeGen/X86/dollar-name-asm.ll new file mode 100644 index 000000000000..cc649f24e09e --- /dev/null +++ b/llvm/test/CodeGen/X86/dollar-name-asm.ll @@ -0,0 +1,7 @@ +; RUN: llc < %s -mtriple=x86_64 | FileCheck %s --check-prefix=ATT +; RUN: llc < %s -mtriple=x86_64 -output-asm-variant=1 | FileCheck %s --check-prefix=INTEL + +module asm "mov ($foo), %eax" + +; ATT: movl ($foo), %eax +; INTEL: mov eax, dword ptr [$foo] diff --git a/llvm/test/CodeGen/X86/dollar-name.ll b/llvm/test/CodeGen/X86/dollar-name.ll index fc9d6a77f66e..b997b5107f01 100644 --- a/llvm/test/CodeGen/X86/dollar-name.ll +++ b/llvm/test/CodeGen/X86/dollar-name.ll @@ -1,18 +1,73 @@ -; RUN: llc < %s -mtriple=i386-linux | FileCheck %s -; PR1339 +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 +;; See also dollar-name-asm.ll for module asm tests, which update_llc_test_checks.py doesn't support. +; RUN: llc < %s -mtriple=i686 -relocation-model=static | FileCheck %s --check-prefix=STATIC +; RUN: llc < %s -mtriple=x86_64 -relocation-model=pic | FileCheck %s --check-prefix=PIC +; RUN: llc < %s -mtriple=x86_64 -relocation-model=pic -output-asm-variant=1 | FileCheck %s --check-prefix=INTEL-PIC -@"$bar" = global i32 zeroinitializer -@"$qux" = external dso_local global i32 +@"$arr" = global [2 x i32] zeroinitializer +@"$arr_h" = hidden global [2 x i32] zeroinitializer +@"$tls" = hidden thread_local global i32 0 define i32 @"$foo"() nounwind { -; CHECK: movl $bar, -; CHECK: addl $qux, -; CHECK: calll $hen - %m = load i32, ptr @"$bar" - %n = load i32, ptr @"$qux" - %t = add i32 %m, %n - %u = call i32 @"$hen"(i32 %t) +; STATIC-LABEL: $foo: +; STATIC: # %bb.0: +; STATIC-NEXT: movl ($arr), %eax +; STATIC-NEXT: movl %gs:0, %ecx +; STATIC-NEXT: addl ($arr+4), %eax +; STATIC-NEXT: addl ($tls@NTPOFF)(%ecx), %eax +; STATIC-NEXT: pushl ($arr_h) +; STATIC-NEXT: pushl %eax +; STATIC-NEXT: calll ($hen@PLT) +; STATIC-NEXT: addl $8, %esp +; STATIC-NEXT: retl +; +; PIC-LABEL: $foo: +; PIC: # %bb.0: +; PIC-NEXT: pushq %rbp +; PIC-NEXT: pushq %r14 +; PIC-NEXT: pushq %rbx +; PIC-NEXT: movq ($arr@GOTPCREL)(%rip), %r14 +; PIC-NEXT: movl (%r14), %ebx +; PIC-NEXT: movl ($arr_h)(%rip), %ebp +; PIC-NEXT: leaq ($tls@TLSLD)(%rip), %rdi +; PIC-NEXT: callq __tls_get_addr@PLT +; PIC-NEXT: addl 4(%r14), %ebx +; PIC-NEXT: addl ($tls@DTPOFF)(%rax), %ebx +; PIC-NEXT: movl %ebx, %edi +; PIC-NEXT: movl %ebp, %esi +; PIC-NEXT: callq ($hen@PLT) +; PIC-NEXT: popq %rbx +; PIC-NEXT: popq %r14 +; PIC-NEXT: popq %rbp +; PIC-NEXT: retq +; +; INTEL-PIC-LABEL: $foo: +; INTEL-PIC: # %bb.0: +; INTEL-PIC-NEXT: push rbp +; INTEL-PIC-NEXT: push r14 +; INTEL-PIC-NEXT: push rbx +; INTEL-PIC-NEXT: mov r14, qword ptr [rip + $arr@GOTPCREL] +; INTEL-PIC-NEXT: mov ebx, dword ptr [r14] +; INTEL-PIC-NEXT: mov ebp, dword ptr [rip + $arr_h] +; INTEL-PIC-NEXT: lea rdi, [rip + $tls@TLSLD] +; INTEL-PIC-NEXT: call __tls_get_addr@PLT +; INTEL-PIC-NEXT: add ebx, dword ptr [r14 + 4] +; INTEL-PIC-NEXT: add ebx, dword ptr [rax + $tls@DTPOFF] +; INTEL-PIC-NEXT: mov edi, ebx +; INTEL-PIC-NEXT: mov esi, ebp +; INTEL-PIC-NEXT: call $hen@PLT +; INTEL-PIC-NEXT: pop rbx +; INTEL-PIC-NEXT: pop r14 +; INTEL-PIC-NEXT: pop rbp +; INTEL-PIC-NEXT: ret + %m = load i32, ptr @"$arr" + %m1 = load i32, ptr getelementptr inbounds nuw (i32, ptr @"$arr", i23 1) + %m2 = load i32, ptr @"$arr_h" + %tls_v = load i32, ptr @"$tls" + %t0 = add i32 %m, %m1 + %t1 = add i32 %t0, %tls_v + %u = call i32 @"$hen"(i32 %t1, i32 %m2) ret i32 %u } -declare i32 @"$hen"(i32 %a) +declare i32 @"$hen"(i32 %a, i32 %b) diff --git a/llvm/test/CodeGen/X86/exp10-libcall.ll b/llvm/test/CodeGen/X86/exp10-libcall.ll index a6959ace073f..4abf6b360ca3 100644 --- a/llvm/test/CodeGen/X86/exp10-libcall.ll +++ b/llvm/test/CodeGen/X86/exp10-libcall.ll @@ -2,14 +2,12 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s ; RUN: llc < %s -mtriple=i386-pc-win32 | FileCheck %s -check-prefix=CHECK-WIN -define float @call_exp10f(float %a) { +define float @call_exp10f(float %a) nounwind { ; CHECK-LABEL: call_exp10f: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq exp10f@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_exp10f: @@ -24,14 +22,12 @@ define float @call_exp10f(float %a) { ret float %result } -define double @call_exp10(double %a) { +define double @call_exp10(double %a) nounwind { ; CHECK-LABEL: call_exp10: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq exp10@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_exp10: @@ -46,16 +42,14 @@ define double @call_exp10(double %a) { ret double %result } -define x86_fp80 @call_exp10l(x86_fp80 %a) { +define x86_fp80 @call_exp10l(x86_fp80 %a) nounwind { ; CHECK-LABEL: call_exp10l: ; CHECK: # %bb.0: ; CHECK-NEXT: subq $24, %rsp -; CHECK-NEXT: .cfi_def_cfa_offset 32 ; CHECK-NEXT: fldt {{[0-9]+}}(%rsp) ; CHECK-NEXT: fstpt (%rsp) ; CHECK-NEXT: callq exp10l@PLT ; CHECK-NEXT: addq $24, %rsp -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_exp10l: diff --git a/llvm/test/CodeGen/X86/fast-isel-x32.ll b/llvm/test/CodeGen/X86/fast-isel-x32.ll index 23f6304c88d2..e01cebecdbb0 100644 --- a/llvm/test/CodeGen/X86/fast-isel-x32.ll +++ b/llvm/test/CodeGen/X86/fast-isel-x32.ll @@ -1,5 +1,4 @@ ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 -fast-isel -fast-isel-abort=1 | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-nacl -fast-isel -fast-isel-abort=1 | FileCheck %s ; Test that alloca addresses are materialized with the right size instruction. diff --git a/llvm/test/CodeGen/X86/fp128-cast-strict.ll b/llvm/test/CodeGen/X86/fp128-cast-strict.ll index 707b05f3478d..bb5640aeb66f 100644 --- a/llvm/test/CodeGen/X86/fp128-cast-strict.ll +++ b/llvm/test/CodeGen/X86/fp128-cast-strict.ll @@ -481,18 +481,21 @@ define i128 @fptosi_i128(fp128 %x) nounwind strictfp { ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $20, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __fixtfti -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -501,7 +504,7 @@ define i128 @fptosi_i128(fp128 %x) nounwind strictfp { ; X86-NEXT: movl %eax, (%esi) ; X86-NEXT: movl %ecx, 4(%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $20, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: retl $4 @@ -620,18 +623,21 @@ define i128 @fptoui_i128(fp128 %x) nounwind strictfp { ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $20, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __fixunstfti -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -640,7 +646,7 @@ define i128 @fptoui_i128(fp128 %x) nounwind strictfp { ; X86-NEXT: movl %eax, (%esi) ; X86-NEXT: movl %ecx, 4(%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $20, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: retl $4 @@ -818,18 +824,21 @@ define fp128 @sitofp_i128(i128 %x) nounwind strictfp { ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $20, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __floattitf -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -838,7 +847,7 @@ define fp128 @sitofp_i128(i128 %x) nounwind strictfp { ; X86-NEXT: movl %eax, (%esi) ; X86-NEXT: movl %ecx, 4(%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $20, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: retl $4 @@ -1016,18 +1025,21 @@ define fp128 @uitofp_i128(i128 %x) nounwind strictfp { ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $20, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __floatuntitf -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1036,7 +1048,7 @@ define fp128 @uitofp_i128(i128 %x) nounwind strictfp { ; X86-NEXT: movl %eax, (%esi) ; X86-NEXT: movl %ecx, 4(%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $20, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: retl $4 diff --git a/llvm/test/CodeGen/X86/fp128-cast.ll b/llvm/test/CodeGen/X86/fp128-cast.ll index 1de2484d47ba..6d4ec063ccd4 100644 --- a/llvm/test/CodeGen/X86/fp128-cast.ll +++ b/llvm/test/CodeGen/X86/fp128-cast.ll @@ -415,16 +415,20 @@ define dso_local void @TestFPToSIF128_I128() nounwind { ; X86-LABEL: TestFPToSIF128_I128: ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %esi -; X86-NEXT: subl $36, %esp +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl vf128, %eax +; X86-NEXT: movl vf128+4, %ecx +; X86-NEXT: movl vf128+8, %edx +; X86-NEXT: movl vf128+12, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __fixtfti -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -432,7 +436,7 @@ define dso_local void @TestFPToSIF128_I128() nounwind { ; X86-NEXT: movl %edx, vi128+8 ; X86-NEXT: movl %ecx, vi128+4 ; X86-NEXT: movl %eax, vi128 -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $56, %esp ; X86-NEXT: popl %esi ; X86-NEXT: retl ; @@ -466,16 +470,20 @@ define dso_local void @TestFPToUIF128_U128() nounwind { ; X86-LABEL: TestFPToUIF128_U128: ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %esi -; X86-NEXT: subl $36, %esp +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl vf128, %eax +; X86-NEXT: movl vf128+4, %ecx +; X86-NEXT: movl vf128+8, %edx +; X86-NEXT: movl vf128+12, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __fixunstfti -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -483,7 +491,7 @@ define dso_local void @TestFPToUIF128_U128() nounwind { ; X86-NEXT: movl %edx, vu128+8 ; X86-NEXT: movl %ecx, vu128+4 ; X86-NEXT: movl %eax, vu128 -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $56, %esp ; X86-NEXT: popl %esi ; X86-NEXT: retl ; @@ -913,16 +921,20 @@ define dso_local void @TestSIToFPI128_F128() nounwind { ; X86-LABEL: TestSIToFPI128_F128: ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %esi -; X86-NEXT: subl $36, %esp +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl vi128, %eax +; X86-NEXT: movl vi128+4, %ecx +; X86-NEXT: movl vi128+8, %edx +; X86-NEXT: movl vi128+12, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl vi128+12 -; X86-NEXT: pushl vi128+8 -; X86-NEXT: pushl vi128+4 -; X86-NEXT: pushl vi128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __floattitf -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -930,7 +942,7 @@ define dso_local void @TestSIToFPI128_F128() nounwind { ; X86-NEXT: movl %edx, vf128+8 ; X86-NEXT: movl %ecx, vf128+4 ; X86-NEXT: movl %eax, vf128 -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $56, %esp ; X86-NEXT: popl %esi ; X86-NEXT: retl ; @@ -964,16 +976,20 @@ define dso_local void @TestUIToFPU128_F128() #2 { ; X86-LABEL: TestUIToFPU128_F128: ; X86: # %bb.0: # %entry ; X86-NEXT: pushl %esi -; X86-NEXT: subl $36, %esp +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl vu128, %eax +; X86-NEXT: movl vu128+4, %ecx +; X86-NEXT: movl vu128+8, %edx +; X86-NEXT: movl vu128+12, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl vu128+12 -; X86-NEXT: pushl vu128+8 -; X86-NEXT: pushl vu128+4 -; X86-NEXT: pushl vu128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __floatuntitf -; X86-NEXT: addl $28, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -981,7 +997,7 @@ define dso_local void @TestUIToFPU128_F128() #2 { ; X86-NEXT: movl %edx, vf128+8 ; X86-NEXT: movl %ecx, vf128+4 ; X86-NEXT: movl %eax, vf128 -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $56, %esp ; X86-NEXT: popl %esi ; X86-NEXT: retl ; @@ -1134,33 +1150,30 @@ define dso_local i32 @TestBits128(fp128 %ld) nounwind { ; ; X86-LABEL: TestBits128: ; X86: # %bb.0: # %entry -; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $20, %esp +; X86-NEXT: subl $72, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: subl $12, %esp -; X86-NEXT: leal {{[0-9]+}}(%esp), %edx -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %edx +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __multf3 -; X86-NEXT: addl $44, %esp +; X86-NEXT: subl $4, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: xorl %eax, %eax -; X86-NEXT: orl (%esp), %ecx +; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: sete %al -; X86-NEXT: addl $20, %esp +; X86-NEXT: addl $72, %esp ; X86-NEXT: popl %esi -; X86-NEXT: popl %edi ; X86-NEXT: retl ; ; X64-AVX-LABEL: TestBits128: @@ -1359,12 +1372,14 @@ define i1 @PR34866(i128 %x) nounwind { ; ; X86-LABEL: PR34866: ; X86: # %bb.0: +; X86-NEXT: subl $12, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl %ecx, %eax ; X86-NEXT: sete %al +; X86-NEXT: addl $12, %esp ; X86-NEXT: retl ; ; X64-AVX-LABEL: PR34866: @@ -1394,12 +1409,14 @@ define i1 @PR34866_commute(i128 %x) nounwind { ; ; X86-LABEL: PR34866_commute: ; X86: # %bb.0: +; X86-NEXT: subl $12, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl %ecx, %eax ; X86-NEXT: sete %al +; X86-NEXT: addl $12, %esp ; X86-NEXT: retl ; ; X64-AVX-LABEL: PR34866_commute: diff --git a/llvm/test/CodeGen/X86/fp128-libcalls-strict.ll b/llvm/test/CodeGen/X86/fp128-libcalls-strict.ll index a7eea04181f6..ad2d690fd7ed 100644 --- a/llvm/test/CodeGen/X86/fp128-libcalls-strict.ll +++ b/llvm/test/CodeGen/X86/fp128-libcalls-strict.ll @@ -41,27 +41,40 @@ define fp128 @add(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: add: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __addtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: add: @@ -81,24 +94,32 @@ define fp128 @add(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___addtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -107,9 +128,10 @@ define fp128 @add(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -141,27 +163,40 @@ define fp128 @sub(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: sub: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __subtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: sub: @@ -181,24 +216,32 @@ define fp128 @sub(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___subtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -207,9 +250,10 @@ define fp128 @sub(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -241,27 +285,40 @@ define fp128 @mul(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: mul: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __multf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: mul: @@ -281,24 +338,32 @@ define fp128 @mul(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___multf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -307,9 +372,10 @@ define fp128 @mul(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -341,27 +407,40 @@ define fp128 @div(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: div: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __divtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: div: @@ -381,24 +460,32 @@ define fp128 @div(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___divtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -407,9 +494,10 @@ define fp128 @div(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -434,31 +522,48 @@ define fp128 @fma(fp128 %x, fp128 %y, fp128 %z) nounwind strictfp { ; ; X86-LABEL: fma: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $92, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmaf128 -; X86-NEXT: addl $60, %esp -; X86-NEXT: movaps (%esp), %xmm0 -; X86-NEXT: movaps %xmm0, (%esi) -; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 +; X86-NEXT: movaps %xmm0, (%ebp) +; X86-NEXT: movl %ebp, %eax +; X86-NEXT: addl $92, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: fma: @@ -481,28 +586,40 @@ define fp128 @fma(fp128 %x, fp128 %y, fp128 %z) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $96, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 56(%ebp) -; WIN-X86-NEXT: pushl 52(%ebp) -; WIN-X86-NEXT: pushl 48(%ebp) -; WIN-X86-NEXT: pushl 44(%ebp) -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 52(%ebp), %ebx +; WIN-X86-NEXT: movl 56(%ebp), %edi +; WIN-X86-NEXT: movl 60(%ebp), %edx +; WIN-X86-NEXT: movl 64(%ebp), %ecx +; WIN-X86-NEXT: movl 68(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 48(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 44(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 40(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmal -; WIN-X86-NEXT: addl $52, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -511,9 +628,10 @@ define fp128 @fma(fp128 %x, fp128 %y, fp128 %z) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -538,27 +656,40 @@ define fp128 @frem(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: frem: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmodf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: frem: @@ -578,24 +709,32 @@ define fp128 @frem(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmodl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -604,9 +743,10 @@ define fp128 @frem(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -631,23 +771,28 @@ define fp128 @ceil(fp128 %x) nounwind strictfp { ; ; X86-LABEL: ceil: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll ceilf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: ceil: @@ -667,17 +812,20 @@ define fp128 @ceil(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _ceill -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -713,23 +861,28 @@ define fp128 @acos(fp128 %x) nounwind strictfp { ; ; X86-LABEL: acos: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll acosf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: acos: @@ -749,17 +902,20 @@ define fp128 @acos(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _acosl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -795,23 +951,28 @@ define fp128 @cos(fp128 %x) nounwind strictfp { ; ; X86-LABEL: cos: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll cosf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: cos: @@ -831,17 +992,20 @@ define fp128 @cos(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _cosl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -877,23 +1041,28 @@ define fp128 @cosh(fp128 %x) nounwind strictfp { ; ; X86-LABEL: cosh: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll coshf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: cosh: @@ -913,17 +1082,20 @@ define fp128 @cosh(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _coshl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -959,23 +1131,28 @@ define fp128 @exp(fp128 %x) nounwind strictfp { ; ; X86-LABEL: exp: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll expf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: exp: @@ -995,17 +1172,20 @@ define fp128 @exp(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _expl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1041,23 +1221,28 @@ define fp128 @exp2(fp128 %x) nounwind strictfp { ; ; X86-LABEL: exp2: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll exp2f128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: exp2: @@ -1077,17 +1262,20 @@ define fp128 @exp2(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _exp2l -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1123,23 +1311,28 @@ define fp128 @floor(fp128 %x) nounwind strictfp { ; ; X86-LABEL: floor: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll floorf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: floor: @@ -1159,17 +1352,20 @@ define fp128 @floor(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _floorl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1205,23 +1401,28 @@ define fp128 @log(fp128 %x) nounwind strictfp { ; ; X86-LABEL: log: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll logf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: log: @@ -1241,17 +1442,20 @@ define fp128 @log(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _logl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1287,23 +1491,28 @@ define fp128 @log10(fp128 %x) nounwind strictfp { ; ; X86-LABEL: log10: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll log10f128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: log10: @@ -1323,17 +1532,20 @@ define fp128 @log10(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _log10l -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1369,23 +1581,28 @@ define fp128 @log2(fp128 %x) nounwind strictfp { ; ; X86-LABEL: log2: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll log2f128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: log2: @@ -1405,17 +1622,20 @@ define fp128 @log2(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _log2l -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1451,27 +1671,40 @@ define fp128 @maxnum(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: maxnum: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmaxf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: maxnum: @@ -1491,24 +1724,32 @@ define fp128 @maxnum(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmaxl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1517,9 +1758,10 @@ define fp128 @maxnum(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1544,27 +1786,40 @@ define fp128 @minnum(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: minnum: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fminf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: minnum: @@ -1584,24 +1839,32 @@ define fp128 @minnum(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fminl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1610,9 +1873,10 @@ define fp128 @minnum(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1637,23 +1901,28 @@ define fp128 @nearbyint(fp128 %x) nounwind strictfp { ; ; X86-LABEL: nearbyint: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll nearbyintf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: nearbyint: @@ -1673,17 +1942,20 @@ define fp128 @nearbyint(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _nearbyintl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1719,27 +1991,40 @@ define fp128 @pow(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: pow: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll powf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: pow: @@ -1759,24 +2044,32 @@ define fp128 @pow(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _powl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1785,9 +2078,10 @@ define fp128 @pow(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1819,24 +2113,32 @@ define fp128 @powi(fp128 %x, i32 %y) nounwind strictfp { ; ; X86-LABEL: powi: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $64, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $8, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __powitf2 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $64, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx ; X86-NEXT: retl $4 ; ; WIN-LABEL: powi: @@ -1853,21 +2155,26 @@ define fp128 @powi(fp128 %x, i32 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___powitf2 -; WIN-X86-NEXT: addl $24, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1876,9 +2183,10 @@ define fp128 @powi(fp128 %x, i32 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1903,23 +2211,28 @@ define fp128 @rint(fp128 %x) nounwind strictfp { ; ; X86-LABEL: rint: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll rintf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: rint: @@ -1939,17 +2252,20 @@ define fp128 @rint(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _rintl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1985,23 +2301,28 @@ define fp128 @round(fp128 %x) nounwind strictfp { ; ; X86-LABEL: round: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll roundf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: round: @@ -2021,17 +2342,20 @@ define fp128 @round(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _roundl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2067,23 +2391,28 @@ define fp128 @roundeven(fp128 %x) nounwind strictfp { ; ; X86-LABEL: roundeven: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll roundevenf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: roundeven: @@ -2103,17 +2432,20 @@ define fp128 @roundeven(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _roundevenl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2149,23 +2481,28 @@ define fp128 @asin(fp128 %x) nounwind strictfp { ; ; X86-LABEL: asin: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll asinf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: asin: @@ -2185,17 +2522,20 @@ define fp128 @asin(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _asinl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2231,23 +2571,28 @@ define fp128 @sin(fp128 %x) nounwind strictfp { ; ; X86-LABEL: sin: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sinf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: sin: @@ -2267,17 +2612,20 @@ define fp128 @sin(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sinl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2313,23 +2661,28 @@ define fp128 @sinh(fp128 %x) nounwind strictfp { ; ; X86-LABEL: sinh: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sinhf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: sinh: @@ -2349,17 +2702,20 @@ define fp128 @sinh(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sinhl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2395,23 +2751,28 @@ define fp128 @sqrt(fp128 %x) nounwind strictfp { ; ; X86-LABEL: sqrt: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sqrtf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: sqrt: @@ -2431,17 +2792,20 @@ define fp128 @sqrt(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sqrtl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2477,23 +2841,28 @@ define fp128 @atan(fp128 %x) nounwind strictfp { ; ; X86-LABEL: atan: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll atanf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: atan: @@ -2513,17 +2882,20 @@ define fp128 @atan(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _atanl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2559,27 +2931,40 @@ define fp128 @atan2(fp128 %x, fp128 %y) nounwind strictfp { ; ; X86-LABEL: atan2: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll atan2f128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: atan2: @@ -2599,24 +2984,32 @@ define fp128 @atan2(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _atan2l -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2625,9 +3018,10 @@ define fp128 @atan2(fp128 %x, fp128 %y) nounwind strictfp { ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -2652,23 +3046,28 @@ define fp128 @tan(fp128 %x) nounwind strictfp { ; ; X86-LABEL: tan: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll tanf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: tan: @@ -2688,17 +3087,20 @@ define fp128 @tan(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _tanl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2734,23 +3136,28 @@ define fp128 @tanh(fp128 %x) nounwind strictfp { ; ; X86-LABEL: tanh: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll tanhf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: tanh: @@ -2770,17 +3177,20 @@ define fp128 @tanh(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _tanhl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2816,23 +3226,28 @@ define fp128 @trunc(fp128 %x) nounwind strictfp { ; ; X86-LABEL: trunc: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll truncf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: trunc: @@ -2852,17 +3267,20 @@ define fp128 @trunc(fp128 %x) nounwind strictfp { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _truncl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2919,12 +3337,18 @@ define i32 @lrint(fp128 %x) nounwind strictfp { ; ; WIN-X86-LABEL: lrint: ; WIN-X86: # %bb.0: # %entry -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 20(%ebp) +; WIN-X86-NEXT: pushl 16(%ebp) +; WIN-X86-NEXT: pushl 12(%ebp) +; WIN-X86-NEXT: pushl 8(%ebp) ; WIN-X86-NEXT: calll _lrintl ; WIN-X86-NEXT: addl $16, %esp +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: %rint = call i32 @llvm.experimental.constrained.lrint.i32.f128(fp128 %x, metadata !"round.dynamic", metadata !"fpexcept.strict") #0 @@ -2969,12 +3393,18 @@ define i64 @llrint(fp128 %x) nounwind strictfp { ; ; WIN-X86-LABEL: llrint: ; WIN-X86: # %bb.0: # %entry -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 20(%ebp) +; WIN-X86-NEXT: pushl 16(%ebp) +; WIN-X86-NEXT: pushl 12(%ebp) +; WIN-X86-NEXT: pushl 8(%ebp) ; WIN-X86-NEXT: calll _llrintl ; WIN-X86-NEXT: addl $16, %esp +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: %rint = call i64 @llvm.experimental.constrained.llrint.i64.f128(fp128 %x, metadata !"round.dynamic", metadata !"fpexcept.strict") #0 @@ -3019,12 +3449,18 @@ define i32 @lround(fp128 %x) nounwind strictfp { ; ; WIN-X86-LABEL: lround: ; WIN-X86: # %bb.0: # %entry -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 20(%ebp) +; WIN-X86-NEXT: pushl 16(%ebp) +; WIN-X86-NEXT: pushl 12(%ebp) +; WIN-X86-NEXT: pushl 8(%ebp) ; WIN-X86-NEXT: calll _lroundl ; WIN-X86-NEXT: addl $16, %esp +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: %round = call i32 @llvm.experimental.constrained.lround.i32.f128(fp128 %x, metadata !"fpexcept.strict") #0 @@ -3069,12 +3505,18 @@ define i64 @llround(fp128 %x) nounwind strictfp { ; ; WIN-X86-LABEL: llround: ; WIN-X86: # %bb.0: # %entry -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 20(%ebp) +; WIN-X86-NEXT: pushl 16(%ebp) +; WIN-X86-NEXT: pushl 12(%ebp) +; WIN-X86-NEXT: pushl 8(%ebp) ; WIN-X86-NEXT: calll _llroundl ; WIN-X86-NEXT: addl $16, %esp +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: %round = call i64 @llvm.experimental.constrained.llround.i64.f128(fp128 %x, metadata !"fpexcept.strict") #0 @@ -3176,26 +3618,32 @@ define i64 @cmp(i64 %a, i64 %b, fp128 %x, fp128 %y) #0 { ; ; WIN-X86-LABEL: cmp: ; WIN-X86: # %bb.0: -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) +; WIN-X86-NEXT: pushl 36(%ebp) +; WIN-X86-NEXT: pushl 32(%ebp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___eqtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: testl %eax, %eax ; WIN-X86-NEXT: je LBB37_1 ; WIN-X86-NEXT: # %bb.2: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 16(%ebp), %ecx ; WIN-X86-NEXT: jmp LBB37_3 ; WIN-X86-NEXT: LBB37_1: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 8(%ebp), %ecx ; WIN-X86-NEXT: LBB37_3: ; WIN-X86-NEXT: movl (%ecx), %eax ; WIN-X86-NEXT: movl 4(%ecx), %edx +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl %cond = call i1 @llvm.experimental.constrained.fcmp.f128( fp128 %x, fp128 %y, @@ -3300,26 +3748,32 @@ define i64 @cmps(i64 %a, i64 %b, fp128 %x, fp128 %y) #0 { ; ; WIN-X86-LABEL: cmps: ; WIN-X86: # %bb.0: -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) +; WIN-X86-NEXT: pushl 36(%ebp) +; WIN-X86-NEXT: pushl 32(%ebp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___eqtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: testl %eax, %eax ; WIN-X86-NEXT: je LBB38_1 ; WIN-X86-NEXT: # %bb.2: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 16(%ebp), %ecx ; WIN-X86-NEXT: jmp LBB38_3 ; WIN-X86-NEXT: LBB38_1: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 8(%ebp), %ecx ; WIN-X86-NEXT: LBB38_3: ; WIN-X86-NEXT: movl (%ecx), %eax ; WIN-X86-NEXT: movl 4(%ecx), %edx +; WIN-X86-NEXT: movl %ebp, %esp +; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl %cond = call i1 @llvm.experimental.constrained.fcmps.f128( fp128 %x, fp128 %y, @@ -3496,44 +3950,47 @@ define i64 @cmp_ueq_q(i64 %a, i64 %b, fp128 %x, fp128 %y) #0 { ; WIN-X86-LABEL: cmp_ueq_q: ; WIN-X86: # %bb.0: ; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: movl 32(%ebp), %edi +; WIN-X86-NEXT: movl 36(%ebp), %esi +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: pushl %edi -; WIN-X86-NEXT: pushl %ebp -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___eqtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: testl %eax, %eax ; WIN-X86-NEXT: sete %bl -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: pushl %edi -; WIN-X86-NEXT: pushl %ebp -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___unordtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: orb %bl, %al ; WIN-X86-NEXT: jne LBB39_1 ; WIN-X86-NEXT: # %bb.2: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 16(%ebp), %ecx ; WIN-X86-NEXT: jmp LBB39_3 ; WIN-X86-NEXT: LBB39_1: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 8(%ebp), %ecx ; WIN-X86-NEXT: LBB39_3: ; WIN-X86-NEXT: movl (%ecx), %eax ; WIN-X86-NEXT: movl 4(%ecx), %edx +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi ; WIN-X86-NEXT: popl %ebx @@ -3716,32 +4173,34 @@ define i64 @cmp_one_q(i64 %a, i64 %b, fp128 %x, fp128 %y) #0 { ; WIN-X86-LABEL: cmp_one_q: ; WIN-X86: # %bb.0: ; WIN-X86-NEXT: pushl %ebp +; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: andl $-16, %esp +; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: movl 32(%ebp), %edi +; WIN-X86-NEXT: movl 36(%ebp), %esi +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: pushl %edi -; WIN-X86-NEXT: pushl %ebp -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___eqtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: testl %eax, %eax ; WIN-X86-NEXT: setne %bl -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 52(%ebp) +; WIN-X86-NEXT: pushl 48(%ebp) +; WIN-X86-NEXT: pushl 44(%ebp) +; WIN-X86-NEXT: pushl 40(%ebp) ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: pushl %edi -; WIN-X86-NEXT: pushl %ebp -; WIN-X86-NEXT: pushl {{[0-9]+}}(%esp) +; WIN-X86-NEXT: pushl 28(%ebp) +; WIN-X86-NEXT: pushl 24(%ebp) ; WIN-X86-NEXT: calll ___unordtf2 ; WIN-X86-NEXT: addl $32, %esp ; WIN-X86-NEXT: testl %eax, %eax @@ -3749,13 +4208,14 @@ define i64 @cmp_one_q(i64 %a, i64 %b, fp128 %x, fp128 %y) #0 { ; WIN-X86-NEXT: testb %bl, %al ; WIN-X86-NEXT: jne LBB40_1 ; WIN-X86-NEXT: # %bb.2: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 16(%ebp), %ecx ; WIN-X86-NEXT: jmp LBB40_3 ; WIN-X86-NEXT: LBB40_1: -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx +; WIN-X86-NEXT: leal 8(%ebp), %ecx ; WIN-X86-NEXT: LBB40_3: ; WIN-X86-NEXT: movl (%ecx), %eax ; WIN-X86-NEXT: movl 4(%ecx), %edx +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi ; WIN-X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/fp128-libcalls.ll b/llvm/test/CodeGen/X86/fp128-libcalls.ll index f727a7907862..4b0449fd7502 100644 --- a/llvm/test/CodeGen/X86/fp128-libcalls.ll +++ b/llvm/test/CodeGen/X86/fp128-libcalls.ll @@ -42,22 +42,38 @@ define dso_local void @Test128Add(fp128 %d1, fp128 %d2) nounwind { ; ; X86-LABEL: Test128Add: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __addtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128Add: @@ -78,22 +94,31 @@ define dso_local void @Test128Add(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %edi +; WIN-X86-NEXT: movl 28(%ebp), %ebx +; WIN-X86-NEXT: movl 32(%ebp), %ecx +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___addtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -101,8 +126,10 @@ define dso_local void @Test128Add(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+8 ; WIN-X86-NEXT: movl %ecx, _vf128+4 ; WIN-X86-NEXT: movl %eax, _vf128 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -144,22 +171,38 @@ define dso_local void @Test128_1Add(fp128 %d1) nounwind { ; ; X86-LABEL: Test128_1Add: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl vf128, %edi +; X86-NEXT: movl vf128+4, %ebx +; X86-NEXT: movl vf128+8, %ebp +; X86-NEXT: movl vf128+12, %eax +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __addtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128_1Add: @@ -180,22 +223,31 @@ define dso_local void @Test128_1Add(fp128 %d1) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl _vf128+12 -; WIN-X86-NEXT: pushl _vf128+8 -; WIN-X86-NEXT: pushl _vf128+4 -; WIN-X86-NEXT: pushl _vf128 -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %esi +; WIN-X86-NEXT: movl 20(%ebp), %edi +; WIN-X86-NEXT: movl _vf128, %edx +; WIN-X86-NEXT: movl _vf128+4, %ebx +; WIN-X86-NEXT: movl _vf128+8, %ecx +; WIN-X86-NEXT: movl _vf128+12, %eax +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___addtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -203,8 +255,10 @@ define dso_local void @Test128_1Add(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+12 ; WIN-X86-NEXT: movl %eax, _vf128 ; WIN-X86-NEXT: movl %ecx, _vf128+4 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -241,22 +295,38 @@ define dso_local void @Test128Sub(fp128 %d1, fp128 %d2) nounwind { ; ; X86-LABEL: Test128Sub: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __subtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128Sub: @@ -277,22 +347,31 @@ define dso_local void @Test128Sub(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %edi +; WIN-X86-NEXT: movl 28(%ebp), %ebx +; WIN-X86-NEXT: movl 32(%ebp), %ecx +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___subtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -300,8 +379,10 @@ define dso_local void @Test128Sub(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+8 ; WIN-X86-NEXT: movl %ecx, _vf128+4 ; WIN-X86-NEXT: movl %eax, _vf128 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -343,22 +424,38 @@ define dso_local void @Test128_1Sub(fp128 %d1) nounwind { ; ; X86-LABEL: Test128_1Sub: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl vf128, %edi +; X86-NEXT: movl vf128+4, %ebx +; X86-NEXT: movl vf128+8, %ebp +; X86-NEXT: movl vf128+12, %eax +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __subtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128_1Sub: @@ -379,22 +476,31 @@ define dso_local void @Test128_1Sub(fp128 %d1) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl _vf128+12 -; WIN-X86-NEXT: pushl _vf128+8 -; WIN-X86-NEXT: pushl _vf128+4 -; WIN-X86-NEXT: pushl _vf128 -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %esi +; WIN-X86-NEXT: movl 20(%ebp), %edi +; WIN-X86-NEXT: movl _vf128, %edx +; WIN-X86-NEXT: movl _vf128+4, %ebx +; WIN-X86-NEXT: movl _vf128+8, %ecx +; WIN-X86-NEXT: movl _vf128+12, %eax +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___subtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -402,8 +508,10 @@ define dso_local void @Test128_1Sub(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+12 ; WIN-X86-NEXT: movl %eax, _vf128 ; WIN-X86-NEXT: movl %ecx, _vf128+4 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -440,22 +548,38 @@ define dso_local void @Test128Mul(fp128 %d1, fp128 %d2) nounwind { ; ; X86-LABEL: Test128Mul: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __multf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128Mul: @@ -476,22 +600,31 @@ define dso_local void @Test128Mul(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %edi +; WIN-X86-NEXT: movl 28(%ebp), %ebx +; WIN-X86-NEXT: movl 32(%ebp), %ecx +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___multf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -499,8 +632,10 @@ define dso_local void @Test128Mul(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+8 ; WIN-X86-NEXT: movl %ecx, _vf128+4 ; WIN-X86-NEXT: movl %eax, _vf128 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -542,22 +677,38 @@ define dso_local void @Test128_1Mul(fp128 %d1) nounwind { ; ; X86-LABEL: Test128_1Mul: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl vf128, %edi +; X86-NEXT: movl vf128+4, %ebx +; X86-NEXT: movl vf128+8, %ebp +; X86-NEXT: movl vf128+12, %eax +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __multf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128_1Mul: @@ -578,22 +729,31 @@ define dso_local void @Test128_1Mul(fp128 %d1) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl _vf128+12 -; WIN-X86-NEXT: pushl _vf128+8 -; WIN-X86-NEXT: pushl _vf128+4 -; WIN-X86-NEXT: pushl _vf128 -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %esi +; WIN-X86-NEXT: movl 20(%ebp), %edi +; WIN-X86-NEXT: movl _vf128, %edx +; WIN-X86-NEXT: movl _vf128+4, %ebx +; WIN-X86-NEXT: movl _vf128+8, %ecx +; WIN-X86-NEXT: movl _vf128+12, %eax +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___multf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -601,8 +761,10 @@ define dso_local void @Test128_1Mul(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+12 ; WIN-X86-NEXT: movl %eax, _vf128 ; WIN-X86-NEXT: movl %ecx, _vf128+4 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -639,22 +801,38 @@ define dso_local void @Test128Div(fp128 %d1, fp128 %d2) nounwind { ; ; X86-LABEL: Test128Div: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __divtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128Div: @@ -675,22 +853,31 @@ define dso_local void @Test128Div(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %edi +; WIN-X86-NEXT: movl 28(%ebp), %ebx +; WIN-X86-NEXT: movl 32(%ebp), %ecx +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___divtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -698,8 +885,10 @@ define dso_local void @Test128Div(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+8 ; WIN-X86-NEXT: movl %ecx, _vf128+4 ; WIN-X86-NEXT: movl %eax, _vf128 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -741,22 +930,38 @@ define dso_local void @Test128_1Div(fp128 %d1) nounwind { ; ; X86-LABEL: Test128_1Div: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl vf128, %edi +; X86-NEXT: movl vf128+4, %ebx +; X86-NEXT: movl vf128+8, %ebp +; X86-NEXT: movl vf128+12, %eax +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll __divtf3 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128_1Div: @@ -777,22 +982,31 @@ define dso_local void @Test128_1Div(fp128 %d1) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl _vf128+12 -; WIN-X86-NEXT: pushl _vf128+8 -; WIN-X86-NEXT: pushl _vf128+4 -; WIN-X86-NEXT: pushl _vf128 -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %esi +; WIN-X86-NEXT: movl 20(%ebp), %edi +; WIN-X86-NEXT: movl _vf128, %edx +; WIN-X86-NEXT: movl _vf128+4, %ebx +; WIN-X86-NEXT: movl _vf128+8, %ecx +; WIN-X86-NEXT: movl _vf128+12, %eax +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll ___divtf3 -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -800,8 +1014,10 @@ define dso_local void @Test128_1Div(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+12 ; WIN-X86-NEXT: movl %eax, _vf128 ; WIN-X86-NEXT: movl %ecx, _vf128+4 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -830,22 +1046,38 @@ define dso_local void @Test128Rem(fp128 %d1, fp128 %d2) nounwind { ; ; X86-LABEL: Test128Rem: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmodf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128Rem: @@ -866,22 +1098,31 @@ define dso_local void @Test128Rem(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %edi +; WIN-X86-NEXT: movl 28(%ebp), %ebx +; WIN-X86-NEXT: movl 32(%ebp), %ecx +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmodl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -889,8 +1130,10 @@ define dso_local void @Test128Rem(fp128 %d1, fp128 %d2) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+8 ; WIN-X86-NEXT: movl %ecx, _vf128+4 ; WIN-X86-NEXT: movl %eax, _vf128 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -922,22 +1165,38 @@ define dso_local void @Test128_1Rem(fp128 %d1) nounwind { ; ; X86-LABEL: Test128_1Rem: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $76, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl vf128, %edi +; X86-NEXT: movl vf128+4, %ebx +; X86-NEXT: movl vf128+8, %ebp +; X86-NEXT: movl vf128+12, %eax +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl vf128+12 -; X86-NEXT: pushl vf128+8 -; X86-NEXT: pushl vf128+4 -; X86-NEXT: pushl vf128 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmodf128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $76, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; WIN-LABEL: Test128_1Rem: @@ -958,22 +1217,31 @@ define dso_local void @Test128_1Rem(fp128 %d1) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx +; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl _vf128+12 -; WIN-X86-NEXT: pushl _vf128+8 -; WIN-X86-NEXT: pushl _vf128+4 -; WIN-X86-NEXT: pushl _vf128 -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $80, %esp +; WIN-X86-NEXT: movl 16(%ebp), %esi +; WIN-X86-NEXT: movl 20(%ebp), %edi +; WIN-X86-NEXT: movl _vf128, %edx +; WIN-X86-NEXT: movl _vf128+4, %ebx +; WIN-X86-NEXT: movl _vf128+8, %ecx +; WIN-X86-NEXT: movl _vf128+12, %eax +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 12(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmodl -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -981,8 +1249,10 @@ define dso_local void @Test128_1Rem(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %edx, _vf128+12 ; WIN-X86-NEXT: movl %eax, _vf128 ; WIN-X86-NEXT: movl %ecx, _vf128+4 -; WIN-X86-NEXT: leal -4(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi +; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1011,18 +1281,24 @@ define dso_local void @Test128Sqrt(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Sqrt: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sqrtf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Sqrt: @@ -1042,16 +1318,19 @@ define dso_local void @Test128Sqrt(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sqrtl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1089,18 +1368,24 @@ define dso_local void @Test128Sin(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Sin: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sinf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Sin: @@ -1120,16 +1405,19 @@ define dso_local void @Test128Sin(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sinl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1167,18 +1455,24 @@ define dso_local void @Test128Cos(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Cos: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll cosf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Cos: @@ -1198,16 +1492,19 @@ define dso_local void @Test128Cos(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _cosl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1245,18 +1542,24 @@ define dso_local void @Test128Ceil(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Ceil: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll ceilf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Ceil: @@ -1276,16 +1579,19 @@ define dso_local void @Test128Ceil(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _ceill -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1323,18 +1629,24 @@ define dso_local void @Test128Floor(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Floor: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll floorf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Floor: @@ -1354,16 +1666,19 @@ define dso_local void @Test128Floor(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _floorl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1401,18 +1716,24 @@ define dso_local void @Test128Trunc(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Trunc: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll truncf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Trunc: @@ -1432,16 +1753,19 @@ define dso_local void @Test128Trunc(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _truncl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1479,18 +1803,24 @@ define dso_local void @Test128Nearbyint(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Nearbyint: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll nearbyintf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Nearbyint: @@ -1510,16 +1840,19 @@ define dso_local void @Test128Nearbyint(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _nearbyintl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1557,18 +1890,24 @@ define dso_local void @Test128Rint(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Rint: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll rintf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Rint: @@ -1588,16 +1927,19 @@ define dso_local void @Test128Rint(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _rintl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1635,18 +1977,24 @@ define dso_local void @Test128Round(fp128 %d1) nounwind { ; ; X86-LABEL: Test128Round: ; X86: # %bb.0: # %entry -; X86-NEXT: subl $40, %esp +; X86-NEXT: pushl %esi +; X86-NEXT: subl $56, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll roundf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, vf128 -; X86-NEXT: addl $28, %esp +; X86-NEXT: addl $56, %esp +; X86-NEXT: popl %esi ; X86-NEXT: retl ; ; WIN-LABEL: Test128Round: @@ -1666,16 +2014,19 @@ define dso_local void @Test128Round(fp128 %d1) nounwind { ; WIN-X86-NEXT: movl %esp, %ebp ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $32, %esp -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl 8(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: movl 8(%ebp), %eax +; WIN-X86-NEXT: movl 12(%ebp), %ecx +; WIN-X86-NEXT: movl 16(%ebp), %edx +; WIN-X86-NEXT: movl 20(%ebp), %esi +; WIN-X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _roundl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %esi @@ -1705,31 +2056,48 @@ define fp128 @Test128FMA(fp128 %a, fp128 %b, fp128 %c) nounwind { ; ; X86-LABEL: Test128FMA: ; X86: # %bb.0: # %entry +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $92, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll fmaf128 -; X86-NEXT: addl $60, %esp -; X86-NEXT: movaps (%esp), %xmm0 -; X86-NEXT: movaps %xmm0, (%esi) -; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 +; X86-NEXT: movaps %xmm0, (%ebp) +; X86-NEXT: movl %ebp, %eax +; X86-NEXT: addl $92, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128FMA: @@ -1752,28 +2120,40 @@ define fp128 @Test128FMA(fp128 %a, fp128 %b, fp128 %c) nounwind { ; WIN-X86: # %bb.0: # %entry ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $96, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 56(%ebp) -; WIN-X86-NEXT: pushl 52(%ebp) -; WIN-X86-NEXT: pushl 48(%ebp) -; WIN-X86-NEXT: pushl 44(%ebp) -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 52(%ebp), %ebx +; WIN-X86-NEXT: movl 56(%ebp), %edi +; WIN-X86-NEXT: movl 60(%ebp), %edx +; WIN-X86-NEXT: movl 64(%ebp), %ecx +; WIN-X86-NEXT: movl 68(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 48(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 44(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 40(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 36(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _fmal -; WIN-X86-NEXT: addl $52, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1782,9 +2162,10 @@ define fp128 @Test128FMA(fp128 %a, fp128 %b, fp128 %c) nounwind { ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl entry: @@ -1804,23 +2185,28 @@ define fp128 @Test128Acos(fp128 %a) nounwind { ; ; X86-LABEL: Test128Acos: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll acosf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Acos: @@ -1840,17 +2226,20 @@ define fp128 @Test128Acos(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _acosl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1879,23 +2268,28 @@ define fp128 @Test128Asin(fp128 %a) nounwind { ; ; X86-LABEL: Test128Asin: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll asinf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Asin: @@ -1915,17 +2309,20 @@ define fp128 @Test128Asin(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _asinl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -1954,23 +2351,28 @@ define fp128 @Test128Atan(fp128 %a) nounwind { ; ; X86-LABEL: Test128Atan: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll atanf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Atan: @@ -1990,17 +2392,20 @@ define fp128 @Test128Atan(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _atanl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2029,27 +2434,40 @@ define fp128 @Test128Atan2(fp128 %a, fp128 %b) nounwind { ; ; X86-LABEL: Test128Atan2: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $76, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebp, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll atan2f128 -; X86-NEXT: addl $44, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $76, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Atan2: @@ -2069,24 +2487,32 @@ define fp128 @Test128Atan2(fp128 %a, fp128 %b) nounwind { ; WIN-X86: # %bb.0: ; WIN-X86-NEXT: pushl %ebp ; WIN-X86-NEXT: movl %esp, %ebp +; WIN-X86-NEXT: pushl %ebx ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $80, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 40(%ebp) -; WIN-X86-NEXT: pushl 36(%ebp) -; WIN-X86-NEXT: pushl 32(%ebp) -; WIN-X86-NEXT: pushl 28(%ebp) -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl 40(%ebp), %ebx +; WIN-X86-NEXT: movl 44(%ebp), %edx +; WIN-X86-NEXT: movl 48(%ebp), %ecx +; WIN-X86-NEXT: movl 52(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 32(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 28(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _atan2l -; WIN-X86-NEXT: addl $36, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2095,9 +2521,10 @@ define fp128 @Test128Atan2(fp128 %a, fp128 %b) nounwind { ; WIN-X86-NEXT: movl %ecx, 4(%esi) ; WIN-X86-NEXT: movl %eax, (%esi) ; WIN-X86-NEXT: movl %esi, %eax -; WIN-X86-NEXT: leal -8(%ebp), %esp +; WIN-X86-NEXT: leal -12(%ebp), %esp ; WIN-X86-NEXT: popl %esi ; WIN-X86-NEXT: popl %edi +; WIN-X86-NEXT: popl %ebx ; WIN-X86-NEXT: popl %ebp ; WIN-X86-NEXT: retl %x = call fp128 @llvm.atan2.f128(fp128 %a, fp128 %b) @@ -2115,23 +2542,28 @@ define fp128 @Test128Cosh(fp128 %a) nounwind { ; ; X86-LABEL: Test128Cosh: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll coshf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Cosh: @@ -2151,17 +2583,20 @@ define fp128 @Test128Cosh(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _coshl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2190,23 +2625,28 @@ define fp128 @Test128Sinh(fp128 %a) nounwind { ; ; X86-LABEL: Test128Sinh: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll sinhf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Sinh: @@ -2226,17 +2666,20 @@ define fp128 @Test128Sinh(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _sinhl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2265,23 +2708,28 @@ define fp128 @Test128Tan(fp128 %a) nounwind { ; ; X86-LABEL: Test128Tan: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll tanf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Tan: @@ -2301,17 +2749,20 @@ define fp128 @Test128Tan(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _tanl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2340,23 +2791,28 @@ define fp128 @Test128Tanh(fp128 %a) nounwind { ; ; X86-LABEL: Test128Tanh: ; X86: # %bb.0: +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $24, %esp +; X86-NEXT: subl $52, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $12, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll tanhf128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $52, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Tanh: @@ -2376,17 +2832,20 @@ define fp128 @Test128Tanh(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $16, %esp +; WIN-X86-NEXT: subl $48, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi -; WIN-X86-NEXT: movl %esp, %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %eax +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _tanhl -; WIN-X86-NEXT: addl $20, %esp -; WIN-X86-NEXT: movl (%esp), %eax +; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %edi @@ -2425,27 +2884,34 @@ define { fp128, fp128 } @Test128Modf(fp128 %a) nounwind { ; ; X86-LABEL: Test128Modf: ; X86: # %bb.0: +; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: subl $40, %esp +; X86-NEXT: subl $80, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: subl $8, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: leal {{[0-9]+}}(%esp), %ebx +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: leal {{[0-9]+}}(%esp), %ecx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl {{[0-9]+}}(%esp) -; X86-NEXT: pushl %ecx +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: calll modff128 -; X86-NEXT: addl $28, %esp -; X86-NEXT: movaps (%esp), %xmm0 +; X86-NEXT: subl $4, %esp +; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm0 ; X86-NEXT: movaps {{[0-9]+}}(%esp), %xmm1 ; X86-NEXT: movaps %xmm1, 16(%esi) ; X86-NEXT: movaps %xmm0, (%esi) ; X86-NEXT: movl %esi, %eax -; X86-NEXT: addl $40, %esp +; X86-NEXT: addl $80, %esp ; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: popl %ebx ; X86-NEXT: retl $4 ; ; WIN-LABEL: Test128Modf: @@ -2468,18 +2934,21 @@ define { fp128, fp128 } @Test128Modf(fp128 %a) nounwind { ; WIN-X86-NEXT: pushl %edi ; WIN-X86-NEXT: pushl %esi ; WIN-X86-NEXT: andl $-16, %esp -; WIN-X86-NEXT: subl $64, %esp +; WIN-X86-NEXT: subl $112, %esp ; WIN-X86-NEXT: movl 8(%ebp), %esi +; WIN-X86-NEXT: movl 24(%ebp), %eax +; WIN-X86-NEXT: movl 28(%ebp), %ecx +; WIN-X86-NEXT: movl 32(%ebp), %edx +; WIN-X86-NEXT: movl 36(%ebp), %edi +; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ebx +; WIN-X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; WIN-X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; WIN-X86-NEXT: leal {{[0-9]+}}(%esp), %ecx -; WIN-X86-NEXT: pushl %eax -; WIN-X86-NEXT: pushl 24(%ebp) -; WIN-X86-NEXT: pushl 20(%ebp) -; WIN-X86-NEXT: pushl 16(%ebp) -; WIN-X86-NEXT: pushl 12(%ebp) -; WIN-X86-NEXT: pushl %ecx +; WIN-X86-NEXT: movl %eax, (%esp) ; WIN-X86-NEXT: calll _modfl -; WIN-X86-NEXT: addl $24, %esp ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; WIN-X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; WIN-X86-NEXT: movl {{[0-9]+}}(%esp), %eax diff --git a/llvm/test/CodeGen/X86/frameaddr.ll b/llvm/test/CodeGen/X86/frameaddr.ll index 7f2f9b8bd6fd..33c9e3855c10 100644 --- a/llvm/test/CodeGen/X86/frameaddr.ll +++ b/llvm/test/CodeGen/X86/frameaddr.ll @@ -5,8 +5,6 @@ ; RUN: llc < %s -mtriple=x86_64-unknown -fast-isel -fast-isel-abort=1 | FileCheck %s --check-prefix=CHECK-64 ; RUN: llc < %s -mtriple=x86_64-gnux32 | FileCheck %s --check-prefix=CHECK-X32ABI ; RUN: llc < %s -mtriple=x86_64-gnux32 -fast-isel -fast-isel-abort=1 | FileCheck %s --check-prefix=CHECK-X32ABI -; RUN: llc < %s -mtriple=x86_64-nacl | FileCheck %s --check-prefix=CHECK-NACL64 -; RUN: llc < %s -mtriple=x86_64-nacl -fast-isel -fast-isel-abort=1 | FileCheck %s --check-prefix=CHECK-NACL64 define ptr @test1() nounwind { entry: @@ -34,10 +32,6 @@ entry: ; CHECK-X32ABI-NEXT: movl %ebp, %eax ; CHECK-X32ABI-NEXT: popq %rbp ; CHECK-X32ABI-NEXT: ret -; CHECK-NACL64-LABEL: test1 -; CHECK-NACL64: pushq %rbp -; CHECK-NACL64-NEXT: movq %rsp, %rbp -; CHECK-NACL64-NEXT: movl %ebp, %eax %0 = tail call ptr @llvm.frameaddress(i32 0) ret ptr %0 } @@ -71,11 +65,6 @@ entry: ; CHECK-X32ABI-NEXT: movl (%eax), %eax ; CHECK-X32ABI-NEXT: popq %rbp ; CHECK-X32ABI-NEXT: ret -; CHECK-NACL64-LABEL: test2 -; CHECK-NACL64: pushq %rbp -; CHECK-NACL64-NEXT: movq %rsp, %rbp -; CHECK-NACL64-NEXT: movl (%ebp), %eax -; CHECK-NACL64-NEXT: movl (%eax), %eax %0 = tail call ptr @llvm.frameaddress(i32 2) ret ptr %0 } diff --git a/llvm/test/CodeGen/X86/freeze-unary.ll b/llvm/test/CodeGen/X86/freeze-unary.ll index 8602c385af83..bc9e29957c74 100644 --- a/llvm/test/CodeGen/X86/freeze-unary.ll +++ b/llvm/test/CodeGen/X86/freeze-unary.ll @@ -70,6 +70,86 @@ define <2 x i64> @freeze_zext_vec(<2 x i16> %a0) nounwind { ret <2 x i64> %z } +define i32 @freeze_abs(i32 %a0) nounwind { +; X86-LABEL: freeze_abs: +; X86: # %bb.0: +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: negl %eax +; X86-NEXT: cmovsl %ecx, %eax +; X86-NEXT: retl +; +; X64-LABEL: freeze_abs: +; X64: # %bb.0: +; X64-NEXT: movl %edi, %eax +; X64-NEXT: negl %eax +; X64-NEXT: cmovsl %edi, %eax +; X64-NEXT: retq + %x = call i32 @llvm.abs.i32(i32 %a0, i1 0) + %f = freeze i32 %x + %r = call i32 @llvm.abs.i32(i32 %f, i1 0) + ret i32 %r +} + +define <4 x i32> @freeze_abs_vec(<4 x i32> %a0) nounwind { +; X86-LABEL: freeze_abs_vec: +; X86: # %bb.0: +; X86-NEXT: movdqa %xmm0, %xmm1 +; X86-NEXT: psrad $31, %xmm1 +; X86-NEXT: pxor %xmm1, %xmm0 +; X86-NEXT: psubd %xmm1, %xmm0 +; X86-NEXT: retl +; +; X64-LABEL: freeze_abs_vec: +; X64: # %bb.0: +; X64-NEXT: pabsd %xmm0, %xmm0 +; X64-NEXT: retq + %x = call <4 x i32> @llvm.abs.v4i32(<4 x i32> %a0, i1 0) + %f = freeze <4 x i32> %x + %r = call <4 x i32> @llvm.abs.v4i32(<4 x i32> %f, i1 0) + ret <4 x i32> %r +} + +define i32 @freeze_abs_undef(i32 %a0) nounwind { +; X86-LABEL: freeze_abs_undef: +; X86: # %bb.0: +; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: negl %eax +; X86-NEXT: cmovsl %ecx, %eax +; X86-NEXT: retl +; +; X64-LABEL: freeze_abs_undef: +; X64: # %bb.0: +; X64-NEXT: movl %edi, %eax +; X64-NEXT: negl %eax +; X64-NEXT: cmovsl %edi, %eax +; X64-NEXT: retq + %x = call i32 @llvm.abs.i32(i32 %a0, i1 -1) + %f = freeze i32 %x + %r = call i32 @llvm.abs.i32(i32 %f, i1 -1) + ret i32 %r +} + +define <4 x i32> @freeze_abs_undef_vec(<4 x i32> %a0) nounwind { +; X86-LABEL: freeze_abs_undef_vec: +; X86: # %bb.0: +; X86-NEXT: movdqa %xmm0, %xmm1 +; X86-NEXT: psrad $31, %xmm1 +; X86-NEXT: pxor %xmm1, %xmm0 +; X86-NEXT: psubd %xmm1, %xmm0 +; X86-NEXT: retl +; +; X64-LABEL: freeze_abs_undef_vec: +; X64: # %bb.0: +; X64-NEXT: pabsd %xmm0, %xmm0 +; X64-NEXT: retq + %x = call <4 x i32> @llvm.abs.v4i32(<4 x i32> %a0, i1 -1) + %f = freeze <4 x i32> %x + %r = call <4 x i32> @llvm.abs.v4i32(<4 x i32> %f, i1 -1) + ret <4 x i32> %r +} + define i32 @freeze_bswap(i32 %a0) nounwind { ; X86-LABEL: freeze_bswap: ; X86: # %bb.0: diff --git a/llvm/test/CodeGen/X86/fshl.ll b/llvm/test/CodeGen/X86/fshl.ll index e8c8ccfa8d37..ec1b8a3c8d6d 100644 --- a/llvm/test/CodeGen/X86/fshl.ll +++ b/llvm/test/CodeGen/X86/fshl.ll @@ -264,53 +264,62 @@ define i128 @var_shift_i128(i128 %x, i128 %y, i128 %z) nounwind { ; X86-FAST-LABEL: var_shift_i128: ; X86-FAST: # %bb.0: ; X86-FAST-NEXT: pushl %ebp +; X86-FAST-NEXT: movl %esp, %ebp ; X86-FAST-NEXT: pushl %ebx ; X86-FAST-NEXT: pushl %edi ; X86-FAST-NEXT: pushl %esi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-FAST-NEXT: andl $-16, %esp +; X86-FAST-NEXT: subl $16, %esp +; X86-FAST-NEXT: movl 24(%ebp), %edi +; X86-FAST-NEXT: movl 28(%ebp), %edx +; X86-FAST-NEXT: movl 48(%ebp), %esi +; X86-FAST-NEXT: movl 56(%ebp), %ecx ; X86-FAST-NEXT: testb $64, %cl +; X86-FAST-NEXT: movl 52(%ebp), %eax ; X86-FAST-NEXT: jne .LBB6_1 ; X86-FAST-NEXT: # %bb.2: -; X86-FAST-NEXT: movl %ebx, %ebp ; X86-FAST-NEXT: movl %esi, %ebx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-FAST-NEXT: movl %edi, %eax -; X86-FAST-NEXT: movl %edx, %edi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-FAST-NEXT: movl %edi, %esi +; X86-FAST-NEXT: movl 32(%ebp), %edi +; X86-FAST-NEXT: movl %eax, (%esp) # 4-byte Spill +; X86-FAST-NEXT: movl %edx, %eax +; X86-FAST-NEXT: movl 36(%ebp), %edx ; X86-FAST-NEXT: testb $32, %cl ; X86-FAST-NEXT: je .LBB6_5 ; X86-FAST-NEXT: .LBB6_4: -; X86-FAST-NEXT: movl %esi, %edx -; X86-FAST-NEXT: movl %edi, %esi -; X86-FAST-NEXT: movl %ebx, %edi -; X86-FAST-NEXT: movl %eax, %ebx +; X86-FAST-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-FAST-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-FAST-NEXT: movl %esi, %eax +; X86-FAST-NEXT: movl (%esp), %esi # 4-byte Reload ; X86-FAST-NEXT: jmp .LBB6_6 ; X86-FAST-NEXT: .LBB6_1: -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-FAST-NEXT: movl 44(%ebp), %ebx +; X86-FAST-NEXT: movl %ebx, (%esp) # 4-byte Spill +; X86-FAST-NEXT: movl 40(%ebp), %ebx ; X86-FAST-NEXT: testb $32, %cl ; X86-FAST-NEXT: jne .LBB6_4 ; X86-FAST-NEXT: .LBB6_5: -; X86-FAST-NEXT: movl %eax, %ebp +; X86-FAST-NEXT: movl (%esp), %ebx # 4-byte Reload +; X86-FAST-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-FAST-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-FAST-NEXT: .LBB6_6: -; X86-FAST-NEXT: movl %ebx, %eax -; X86-FAST-NEXT: shldl %cl, %ebp, %eax -; X86-FAST-NEXT: movl %edi, %ebp -; X86-FAST-NEXT: shldl %cl, %ebx, %ebp -; X86-FAST-NEXT: movl %esi, %ebx -; X86-FAST-NEXT: shldl %cl, %edi, %ebx +; X86-FAST-NEXT: movl %esi, %edi +; X86-FAST-NEXT: shldl %cl, %ebx, %edi +; X86-FAST-NEXT: movl %eax, %edx +; X86-FAST-NEXT: movl %eax, %ebx +; X86-FAST-NEXT: shldl %cl, %esi, %ebx +; X86-FAST-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-FAST-NEXT: movl %eax, %esi +; X86-FAST-NEXT: shldl %cl, %edx, %esi ; X86-FAST-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-FAST-NEXT: shldl %cl, %esi, %edx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-FAST-NEXT: movl %edx, 12(%ecx) -; X86-FAST-NEXT: movl %ebx, 8(%ecx) -; X86-FAST-NEXT: movl %ebp, 4(%ecx) -; X86-FAST-NEXT: movl %eax, (%ecx) -; X86-FAST-NEXT: movl %ecx, %eax +; X86-FAST-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-FAST-NEXT: shldl %cl, %eax, %edx +; X86-FAST-NEXT: movl 8(%ebp), %eax +; X86-FAST-NEXT: movl %edx, 12(%eax) +; X86-FAST-NEXT: movl %esi, 8(%eax) +; X86-FAST-NEXT: movl %ebx, 4(%eax) +; X86-FAST-NEXT: movl %edi, (%eax) +; X86-FAST-NEXT: leal -12(%ebp), %esp ; X86-FAST-NEXT: popl %esi ; X86-FAST-NEXT: popl %edi ; X86-FAST-NEXT: popl %ebx @@ -320,77 +329,91 @@ define i128 @var_shift_i128(i128 %x, i128 %y, i128 %z) nounwind { ; X86-SLOW-LABEL: var_shift_i128: ; X86-SLOW: # %bb.0: ; X86-SLOW-NEXT: pushl %ebp +; X86-SLOW-NEXT: movl %esp, %ebp ; X86-SLOW-NEXT: pushl %ebx ; X86-SLOW-NEXT: pushl %edi ; X86-SLOW-NEXT: pushl %esi -; X86-SLOW-NEXT: pushl %eax -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-SLOW-NEXT: testb $64, %al +; X86-SLOW-NEXT: andl $-16, %esp +; X86-SLOW-NEXT: subl $32, %esp +; X86-SLOW-NEXT: movl 24(%ebp), %esi +; X86-SLOW-NEXT: movl 28(%ebp), %eax +; X86-SLOW-NEXT: movl 48(%ebp), %edx +; X86-SLOW-NEXT: movl 56(%ebp), %ecx +; X86-SLOW-NEXT: testb $64, %cl +; X86-SLOW-NEXT: movl 52(%ebp), %edi ; X86-SLOW-NEXT: jne .LBB6_1 ; X86-SLOW-NEXT: # %bb.2: -; X86-SLOW-NEXT: movl %edx, %ebp -; X86-SLOW-NEXT: movl %ebx, %edx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-SLOW-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %esi, %edx +; X86-SLOW-NEXT: movl 32(%ebp), %esi ; X86-SLOW-NEXT: movl %edi, %ecx -; X86-SLOW-NEXT: movl %esi, %edi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-SLOW-NEXT: testb $32, %al -; X86-SLOW-NEXT: je .LBB6_5 -; X86-SLOW-NEXT: .LBB6_4: -; X86-SLOW-NEXT: movl %ebx, (%esp) # 4-byte Spill -; X86-SLOW-NEXT: movl %edi, %ebx -; X86-SLOW-NEXT: movl %edx, %edi -; X86-SLOW-NEXT: movl %ecx, %edx -; X86-SLOW-NEXT: jmp .LBB6_6 +; X86-SLOW-NEXT: movl %eax, %edi +; X86-SLOW-NEXT: movl 36(%ebp), %eax +; X86-SLOW-NEXT: jmp .LBB6_3 ; X86-SLOW-NEXT: .LBB6_1: -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-SLOW-NEXT: testb $32, %al +; X86-SLOW-NEXT: movl 40(%ebp), %ecx +; X86-SLOW-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl 44(%ebp), %ecx +; X86-SLOW-NEXT: .LBB6_3: +; X86-SLOW-NEXT: movl 56(%ebp), %ebx +; X86-SLOW-NEXT: testb $32, %bl ; X86-SLOW-NEXT: jne .LBB6_4 -; X86-SLOW-NEXT: .LBB6_5: -; X86-SLOW-NEXT: movl %ecx, %ebp -; X86-SLOW-NEXT: movl %esi, (%esp) # 4-byte Spill +; X86-SLOW-NEXT: # %bb.5: +; X86-SLOW-NEXT: movl %ecx, %ebx +; X86-SLOW-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: jmp .LBB6_6 +; X86-SLOW-NEXT: .LBB6_4: +; X86-SLOW-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %ecx, %edx +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload ; X86-SLOW-NEXT: .LBB6_6: ; X86-SLOW-NEXT: movl %edx, %esi -; X86-SLOW-NEXT: movl %eax, %ecx -; X86-SLOW-NEXT: shll %cl, %esi -; X86-SLOW-NEXT: shrl %ebp -; X86-SLOW-NEXT: movb %al, %ch -; X86-SLOW-NEXT: notb %ch -; X86-SLOW-NEXT: movb %ch, %cl -; X86-SLOW-NEXT: shrl %cl, %ebp -; X86-SLOW-NEXT: orl %esi, %ebp -; X86-SLOW-NEXT: movl %edi, %esi -; X86-SLOW-NEXT: movb %al, %cl -; X86-SLOW-NEXT: shll %cl, %esi -; X86-SLOW-NEXT: shrl %edx -; X86-SLOW-NEXT: movb %ch, %cl -; X86-SLOW-NEXT: shrl %cl, %edx -; X86-SLOW-NEXT: orl %esi, %edx -; X86-SLOW-NEXT: movl %ebx, %esi -; X86-SLOW-NEXT: movb %al, %cl +; X86-SLOW-NEXT: movl 56(%ebp), %ecx ; X86-SLOW-NEXT: shll %cl, %esi +; X86-SLOW-NEXT: movl %ebx, %edi ; X86-SLOW-NEXT: shrl %edi -; X86-SLOW-NEXT: movb %ch, %cl +; X86-SLOW-NEXT: movl %ecx, %ebx +; X86-SLOW-NEXT: notb %bl +; X86-SLOW-NEXT: movl %ebx, %ecx +; X86-SLOW-NEXT: movb %bl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill ; X86-SLOW-NEXT: shrl %cl, %edi ; X86-SLOW-NEXT: orl %esi, %edi -; X86-SLOW-NEXT: movb %al, %cl -; X86-SLOW-NEXT: movl (%esp), %eax # 4-byte Reload +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-SLOW-NEXT: movl %esi, %eax +; X86-SLOW-NEXT: movl 56(%ebp), %ecx +; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-SLOW-NEXT: shll %cl, %eax +; X86-SLOW-NEXT: shrl %edx +; X86-SLOW-NEXT: movl %ebx, %ecx +; X86-SLOW-NEXT: shrl %cl, %edx +; X86-SLOW-NEXT: orl %eax, %edx +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-SLOW-NEXT: movl %ebx, %eax +; X86-SLOW-NEXT: movl 56(%ebp), %ecx +; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-SLOW-NEXT: shll %cl, %eax +; X86-SLOW-NEXT: shrl %esi +; X86-SLOW-NEXT: movzbl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 1-byte Folded Reload +; X86-SLOW-NEXT: shrl %cl, %esi +; X86-SLOW-NEXT: orl %eax, %esi +; X86-SLOW-NEXT: movl 56(%ebp), %ecx +; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-SLOW-NEXT: shll %cl, %eax ; X86-SLOW-NEXT: shrl %ebx -; X86-SLOW-NEXT: movb %ch, %cl +; X86-SLOW-NEXT: movzbl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 1-byte Folded Reload ; X86-SLOW-NEXT: shrl %cl, %ebx ; X86-SLOW-NEXT: orl %eax, %ebx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SLOW-NEXT: movl 8(%ebp), %eax ; X86-SLOW-NEXT: movl %ebx, 12(%eax) -; X86-SLOW-NEXT: movl %edi, 8(%eax) +; X86-SLOW-NEXT: movl %esi, 8(%eax) ; X86-SLOW-NEXT: movl %edx, 4(%eax) -; X86-SLOW-NEXT: movl %ebp, (%eax) -; X86-SLOW-NEXT: addl $4, %esp +; X86-SLOW-NEXT: movl %edi, (%eax) +; X86-SLOW-NEXT: leal -12(%ebp), %esp ; X86-SLOW-NEXT: popl %esi ; X86-SLOW-NEXT: popl %edi ; X86-SLOW-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/fshr.ll b/llvm/test/CodeGen/X86/fshr.ll index 4340f8fd484a..544ab7fc7737 100644 --- a/llvm/test/CodeGen/X86/fshr.ll +++ b/llvm/test/CodeGen/X86/fshr.ll @@ -258,51 +258,53 @@ define i128 @var_shift_i128(i128 %x, i128 %y, i128 %z) nounwind { ; X86-FAST-LABEL: var_shift_i128: ; X86-FAST: # %bb.0: ; X86-FAST-NEXT: pushl %ebp +; X86-FAST-NEXT: movl %esp, %ebp ; X86-FAST-NEXT: pushl %ebx ; X86-FAST-NEXT: pushl %edi ; X86-FAST-NEXT: pushl %esi -; X86-FAST-NEXT: pushl %eax -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-FAST-NEXT: andl $-16, %esp +; X86-FAST-NEXT: subl $16, %esp +; X86-FAST-NEXT: movl 24(%ebp), %esi +; X86-FAST-NEXT: movl 28(%ebp), %eax +; X86-FAST-NEXT: movl 48(%ebp), %edx +; X86-FAST-NEXT: movl 56(%ebp), %ecx ; X86-FAST-NEXT: testb $64, %cl +; X86-FAST-NEXT: movl 52(%ebp), %ebx ; X86-FAST-NEXT: je .LBB6_1 ; X86-FAST-NEXT: # %bb.2: -; X86-FAST-NEXT: movl %edx, (%esp) # 4-byte Spill -; X86-FAST-NEXT: movl %edi, %edx -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-FAST-NEXT: movl %esi, %ebp -; X86-FAST-NEXT: movl %ebx, %esi -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-FAST-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-FAST-NEXT: movl %esi, %edx +; X86-FAST-NEXT: movl 32(%ebp), %esi +; X86-FAST-NEXT: movl %ebx, %edi +; X86-FAST-NEXT: movl %eax, %ebx +; X86-FAST-NEXT: movl 36(%ebp), %eax ; X86-FAST-NEXT: testb $32, %cl ; X86-FAST-NEXT: je .LBB6_4 ; X86-FAST-NEXT: jmp .LBB6_5 ; X86-FAST-NEXT: .LBB6_1: -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-FAST-NEXT: movl %ebp, (%esp) # 4-byte Spill -; X86-FAST-NEXT: movl {{[0-9]+}}(%esp), %ebp +; X86-FAST-NEXT: movl 40(%ebp), %edi +; X86-FAST-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-FAST-NEXT: movl 44(%ebp), %edi ; X86-FAST-NEXT: testb $32, %cl ; X86-FAST-NEXT: jne .LBB6_5 ; X86-FAST-NEXT: .LBB6_4: -; X86-FAST-NEXT: movl %edi, %ebx -; X86-FAST-NEXT: movl %esi, %edi -; X86-FAST-NEXT: movl %edx, %esi -; X86-FAST-NEXT: movl %ebp, %edx -; X86-FAST-NEXT: movl (%esp), %ebp # 4-byte Reload +; X86-FAST-NEXT: movl %esi, %eax +; X86-FAST-NEXT: movl %ebx, %esi +; X86-FAST-NEXT: movl %edx, %ebx +; X86-FAST-NEXT: movl %edi, %edx +; X86-FAST-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload ; X86-FAST-NEXT: .LBB6_5: -; X86-FAST-NEXT: shrdl %cl, %edx, %ebp -; X86-FAST-NEXT: shrdl %cl, %esi, %edx -; X86-FAST-NEXT: shrdl %cl, %edi, %esi +; X86-FAST-NEXT: shrdl %cl, %edx, %edi +; X86-FAST-NEXT: shrdl %cl, %ebx, %edx +; X86-FAST-NEXT: shrdl %cl, %esi, %ebx ; X86-FAST-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-FAST-NEXT: shrdl %cl, %ebx, %edi -; X86-FAST-NEXT: movl %edi, 12(%eax) -; X86-FAST-NEXT: movl %esi, 8(%eax) +; X86-FAST-NEXT: shrdl %cl, %eax, %esi +; X86-FAST-NEXT: movl 8(%ebp), %eax +; X86-FAST-NEXT: movl %esi, 12(%eax) +; X86-FAST-NEXT: movl %ebx, 8(%eax) ; X86-FAST-NEXT: movl %edx, 4(%eax) -; X86-FAST-NEXT: movl %ebp, (%eax) -; X86-FAST-NEXT: addl $4, %esp +; X86-FAST-NEXT: movl %edi, (%eax) +; X86-FAST-NEXT: leal -12(%ebp), %esp ; X86-FAST-NEXT: popl %esi ; X86-FAST-NEXT: popl %edi ; X86-FAST-NEXT: popl %ebx @@ -312,78 +314,88 @@ define i128 @var_shift_i128(i128 %x, i128 %y, i128 %z) nounwind { ; X86-SLOW-LABEL: var_shift_i128: ; X86-SLOW: # %bb.0: ; X86-SLOW-NEXT: pushl %ebp +; X86-SLOW-NEXT: movl %esp, %ebp ; X86-SLOW-NEXT: pushl %ebx ; X86-SLOW-NEXT: pushl %edi ; X86-SLOW-NEXT: pushl %esi -; X86-SLOW-NEXT: subl $8, %esp -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-SLOW-NEXT: testb $64, %cl +; X86-SLOW-NEXT: andl $-16, %esp +; X86-SLOW-NEXT: subl $16, %esp +; X86-SLOW-NEXT: movl 24(%ebp), %edx +; X86-SLOW-NEXT: movl 28(%ebp), %esi +; X86-SLOW-NEXT: movl 48(%ebp), %ebx +; X86-SLOW-NEXT: movl 56(%ebp), %eax +; X86-SLOW-NEXT: testb $64, %al +; X86-SLOW-NEXT: movl 52(%ebp), %edi ; X86-SLOW-NEXT: je .LBB6_1 ; X86-SLOW-NEXT: # %bb.2: -; X86-SLOW-NEXT: movl %ebp, %eax -; X86-SLOW-NEXT: movl %ebx, %ebp -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-SLOW-NEXT: movl %edi, %edx +; X86-SLOW-NEXT: movl %ebx, (%esp) # 4-byte Spill +; X86-SLOW-NEXT: movl %edx, %ebx +; X86-SLOW-NEXT: movl 32(%ebp), %edx +; X86-SLOW-NEXT: movl %edi, %eax ; X86-SLOW-NEXT: movl %esi, %edi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-SLOW-NEXT: testb $32, %cl -; X86-SLOW-NEXT: jne .LBB6_5 -; X86-SLOW-NEXT: .LBB6_4: -; X86-SLOW-NEXT: movl %ebx, %esi -; X86-SLOW-NEXT: movl %edi, (%esp) # 4-byte Spill -; X86-SLOW-NEXT: movl %ebp, %edi -; X86-SLOW-NEXT: movl %edx, %ebp -; X86-SLOW-NEXT: movl %eax, %edx -; X86-SLOW-NEXT: jmp .LBB6_6 +; X86-SLOW-NEXT: movl 36(%ebp), %esi +; X86-SLOW-NEXT: jmp .LBB6_3 ; X86-SLOW-NEXT: .LBB6_1: -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-SLOW-NEXT: movl 40(%ebp), %eax +; X86-SLOW-NEXT: movl %eax, (%esp) # 4-byte Spill +; X86-SLOW-NEXT: movl 44(%ebp), %eax +; X86-SLOW-NEXT: .LBB6_3: +; X86-SLOW-NEXT: movl 56(%ebp), %ecx ; X86-SLOW-NEXT: testb $32, %cl ; X86-SLOW-NEXT: je .LBB6_4 -; X86-SLOW-NEXT: .LBB6_5: -; X86-SLOW-NEXT: movl %ebx, (%esp) # 4-byte Spill +; X86-SLOW-NEXT: # %bb.5: +; X86-SLOW-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: jmp .LBB6_6 +; X86-SLOW-NEXT: .LBB6_4: +; X86-SLOW-NEXT: movl %edx, %esi +; X86-SLOW-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl %eax, %ebx +; X86-SLOW-NEXT: movl (%esp), %eax # 4-byte Reload ; X86-SLOW-NEXT: .LBB6_6: -; X86-SLOW-NEXT: shrl %cl, %edx -; X86-SLOW-NEXT: movl %ecx, %ebx -; X86-SLOW-NEXT: notb %bl -; X86-SLOW-NEXT: leal (%ebp,%ebp), %eax -; X86-SLOW-NEXT: movl %ebx, %ecx -; X86-SLOW-NEXT: shll %cl, %eax -; X86-SLOW-NEXT: orl %edx, %eax -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-SLOW-NEXT: shrl %cl, %eax +; X86-SLOW-NEXT: movl %eax, %edx +; X86-SLOW-NEXT: movl %ecx, %eax +; X86-SLOW-NEXT: notb %al +; X86-SLOW-NEXT: movl %ebx, %edi +; X86-SLOW-NEXT: addl %ebx, %ebx +; X86-SLOW-NEXT: movl %eax, %ecx +; X86-SLOW-NEXT: shll %cl, %ebx +; X86-SLOW-NEXT: orl %edx, %ebx +; X86-SLOW-NEXT: movl %ebx, (%esp) # 4-byte Spill +; X86-SLOW-NEXT: movl 56(%ebp), %ecx ; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-SLOW-NEXT: shrl %cl, %ebp -; X86-SLOW-NEXT: leal (%edi,%edi), %edx -; X86-SLOW-NEXT: movl %ebx, %ecx +; X86-SLOW-NEXT: shrl %cl, %edi +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-SLOW-NEXT: leal (%ebx,%ebx), %edx +; X86-SLOW-NEXT: movl %eax, %ecx ; X86-SLOW-NEXT: shll %cl, %edx -; X86-SLOW-NEXT: orl %ebp, %edx -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-SLOW-NEXT: orl %edi, %edx +; X86-SLOW-NEXT: movl 56(%ebp), %ecx ; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-SLOW-NEXT: shrl %cl, %edi -; X86-SLOW-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-SLOW-NEXT: movl (%esp), %edi # 4-byte Reload -; X86-SLOW-NEXT: leal (%edi,%edi), %ebp -; X86-SLOW-NEXT: movl %ebx, %ecx -; X86-SLOW-NEXT: shll %cl, %ebp -; X86-SLOW-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %ebp # 4-byte Folded Reload -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-SLOW-NEXT: shrl %cl, %ebx +; X86-SLOW-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SLOW-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload +; X86-SLOW-NEXT: leal (%edi,%edi), %ebx +; X86-SLOW-NEXT: movl %eax, %ecx +; X86-SLOW-NEXT: shll %cl, %ebx +; X86-SLOW-NEXT: orl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload +; X86-SLOW-NEXT: movl 56(%ebp), %ecx ; X86-SLOW-NEXT: # kill: def $cl killed $cl killed $ecx ; X86-SLOW-NEXT: shrl %cl, %edi ; X86-SLOW-NEXT: addl %esi, %esi -; X86-SLOW-NEXT: movl %ebx, %ecx +; X86-SLOW-NEXT: movl %eax, %ecx ; X86-SLOW-NEXT: shll %cl, %esi ; X86-SLOW-NEXT: orl %edi, %esi -; X86-SLOW-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-SLOW-NEXT: movl 8(%ebp), %ecx ; X86-SLOW-NEXT: movl %esi, 12(%ecx) -; X86-SLOW-NEXT: movl %ebp, 8(%ecx) +; X86-SLOW-NEXT: movl %ebx, 8(%ecx) ; X86-SLOW-NEXT: movl %edx, 4(%ecx) +; X86-SLOW-NEXT: movl (%esp), %eax # 4-byte Reload ; X86-SLOW-NEXT: movl %eax, (%ecx) ; X86-SLOW-NEXT: movl %ecx, %eax -; X86-SLOW-NEXT: addl $8, %esp +; X86-SLOW-NEXT: leal -12(%ebp), %esp ; X86-SLOW-NEXT: popl %esi ; X86-SLOW-NEXT: popl %edi ; X86-SLOW-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/funnel-shift.ll b/llvm/test/CodeGen/X86/funnel-shift.ll index a464d78f9af3..df97f49440f7 100644 --- a/llvm/test/CodeGen/X86/funnel-shift.ll +++ b/llvm/test/CodeGen/X86/funnel-shift.ll @@ -74,43 +74,57 @@ define i128 @fshl_i128(i128 %x, i128 %y, i128 %z) nounwind { ; X86-SSE2-LABEL: fshl_i128: ; X86-SSE2: # %bb.0: ; X86-SSE2-NEXT: pushl %ebp +; X86-SSE2-NEXT: movl %esp, %ebp ; X86-SSE2-NEXT: pushl %ebx ; X86-SSE2-NEXT: pushl %edi ; X86-SSE2-NEXT: pushl %esi -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-SSE2-NEXT: andl $-16, %esp +; X86-SSE2-NEXT: subl $16, %esp +; X86-SSE2-NEXT: movl 48(%ebp), %edi +; X86-SSE2-NEXT: movl 52(%ebp), %eax +; X86-SSE2-NEXT: movl 24(%ebp), %edx +; X86-SSE2-NEXT: movl 56(%ebp), %ecx ; X86-SSE2-NEXT: testb $64, %cl -; X86-SSE2-NEXT: movl %esi, %eax -; X86-SSE2-NEXT: cmovnel %ebx, %eax -; X86-SSE2-NEXT: movl %edx, %ebp -; X86-SSE2-NEXT: cmovnel %edi, %ebp -; X86-SSE2-NEXT: cmovnel {{[0-9]+}}(%esp), %edi -; X86-SSE2-NEXT: cmovnel {{[0-9]+}}(%esp), %ebx -; X86-SSE2-NEXT: cmovel {{[0-9]+}}(%esp), %edx -; X86-SSE2-NEXT: cmovel {{[0-9]+}}(%esp), %esi +; X86-SSE2-NEXT: movl %edx, %ecx +; X86-SSE2-NEXT: cmovnel %edi, %ecx +; X86-SSE2-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SSE2-NEXT: movl 28(%ebp), %esi +; X86-SSE2-NEXT: movl %esi, %ebx +; X86-SSE2-NEXT: cmovnel %eax, %ebx +; X86-SSE2-NEXT: cmovnel 44(%ebp), %eax +; X86-SSE2-NEXT: cmovnel 40(%ebp), %edi +; X86-SSE2-NEXT: cmovel 36(%ebp), %esi +; X86-SSE2-NEXT: cmovel 32(%ebp), %edx +; X86-SSE2-NEXT: movl 56(%ebp), %ecx ; X86-SSE2-NEXT: testb $32, %cl -; X86-SSE2-NEXT: cmovnel %esi, %edx -; X86-SSE2-NEXT: cmovnel %ebp, %esi -; X86-SSE2-NEXT: cmovnel %eax, %ebp -; X86-SSE2-NEXT: cmovel %edi, %ebx +; X86-SSE2-NEXT: cmovnel %edx, %esi +; X86-SSE2-NEXT: cmovnel %ebx, %edx +; X86-SSE2-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-SSE2-NEXT: cmovnel %ecx, %ebx ; X86-SSE2-NEXT: cmovel %eax, %edi -; X86-SSE2-NEXT: movl %edi, %eax -; X86-SSE2-NEXT: shldl %cl, %ebx, %eax -; X86-SSE2-NEXT: movl %ebp, %ebx -; X86-SSE2-NEXT: shldl %cl, %edi, %ebx -; X86-SSE2-NEXT: movl %esi, %edi -; X86-SSE2-NEXT: shldl %cl, %ebp, %edi +; X86-SSE2-NEXT: cmovel %ecx, %eax +; X86-SSE2-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SSE2-NEXT: movl 56(%ebp), %ecx ; X86-SSE2-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-SSE2-NEXT: shldl %cl, %esi, %edx -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-SSE2-NEXT: movl %edx, 12(%ecx) -; X86-SSE2-NEXT: movl %edi, 8(%ecx) -; X86-SSE2-NEXT: movl %ebx, 4(%ecx) -; X86-SSE2-NEXT: movl %eax, (%ecx) -; X86-SSE2-NEXT: movl %ecx, %eax +; X86-SSE2-NEXT: shldl %cl, %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-SSE2-NEXT: movl %ebx, %edi +; X86-SSE2-NEXT: movl 56(%ebp), %ecx +; X86-SSE2-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-SSE2-NEXT: shldl %cl, %eax, %edi +; X86-SSE2-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-SSE2-NEXT: movl %edx, %edi +; X86-SSE2-NEXT: movl 56(%ebp), %ecx +; X86-SSE2-NEXT: shldl %cl, %ebx, %edi +; X86-SSE2-NEXT: movl 8(%ebp), %eax +; X86-SSE2-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-SSE2-NEXT: shldl %cl, %edx, %esi +; X86-SSE2-NEXT: movl %esi, 12(%eax) +; X86-SSE2-NEXT: movl %edi, 8(%eax) +; X86-SSE2-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-SSE2-NEXT: movl %ecx, 4(%eax) +; X86-SSE2-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-SSE2-NEXT: movl %ecx, (%eax) +; X86-SSE2-NEXT: leal -12(%ebp), %esp ; X86-SSE2-NEXT: popl %esi ; X86-SSE2-NEXT: popl %edi ; X86-SSE2-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/i128-add.ll b/llvm/test/CodeGen/X86/i128-add.ll index 2849e448a053..b4546c1e983c 100644 --- a/llvm/test/CodeGen/X86/i128-add.ll +++ b/llvm/test/CodeGen/X86/i128-add.ll @@ -5,17 +5,20 @@ define i128 @add_i128(i128 %x, i128 %y) nounwind { ; X86-LABEL: add_i128: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: addl {{[0-9]+}}(%esp), %esi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %edi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: adcl {{[0-9]+}}(%esp), %edx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: addl 40(%ebp), %esi +; X86-NEXT: adcl 44(%ebp), %edi +; X86-NEXT: adcl 48(%ebp), %ecx +; X86-NEXT: adcl 52(%ebp), %edx ; X86-NEXT: addl $1, %esi ; X86-NEXT: adcl $0, %edi ; X86-NEXT: adcl $0, %ecx @@ -24,8 +27,10 @@ define i128 @add_i128(i128 %x, i128 %y) nounwind { ; X86-NEXT: movl %esi, (%eax) ; X86-NEXT: movl %ecx, 8(%eax) ; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: add_i128: diff --git a/llvm/test/CodeGen/X86/i128-fp128-abi.ll b/llvm/test/CodeGen/X86/i128-fp128-abi.ll index be8f7923b8f9..2174d5056e6c 100644 --- a/llvm/test/CodeGen/X86/i128-fp128-abi.ll +++ b/llvm/test/CodeGen/X86/i128-fp128-abi.ll @@ -1,4 +1,4 @@ -; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --no_x86_scrub_sp ; Combined ABI tests for fp128 and i128 @@ -15,6 +15,167 @@ ; RUN: sed 's/PrimTy/fp128/g' %s | sed 's/Prim0/0xL0/g' | llc -mtriple=i686-pc-windows-msvc -verify-machineinstrs | FileCheck %s --check-prefix=CHECK-MSVC32 ; RUN: sed 's/PrimTy/i128/g' %s | sed 's/Prim0/0/g' | llc -mtriple=i686-pc-windows-msvc -verify-machineinstrs | FileCheck %s --check-prefix=CHECK-MSVC32 +define void @store(PrimTy %x, ptr %p) nounwind { +; CHECK-X64-F128-LABEL: store: +; CHECK-X64-F128: # %bb.0: +; CHECK-X64-F128-NEXT: movaps %xmm0, (%rdi) +; CHECK-X64-F128-NEXT: retq +; +; CHECK-X64-I128-LABEL: store: +; CHECK-X64-I128: # %bb.0: +; CHECK-X64-I128-NEXT: movq %rsi, 8(%rdx) +; CHECK-X64-I128-NEXT: movq %rdi, (%rdx) +; CHECK-X64-I128-NEXT: retq +; +; CHECK-MSVC64-F128-LABEL: store: +; CHECK-MSVC64-F128: # %bb.0: +; CHECK-MSVC64-F128-NEXT: movaps (%rcx), %xmm0 +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, (%rdx) +; CHECK-MSVC64-F128-NEXT: retq +; +; CHECK-MSVC64-I128-LABEL: store: +; CHECK-MSVC64-I128: # %bb.0: +; CHECK-MSVC64-I128-NEXT: movq %rdx, 8(%r8) +; CHECK-MSVC64-I128-NEXT: movq %rcx, (%r8) +; CHECK-MSVC64-I128-NEXT: retq +; +; CHECK-MINGW-F128-LABEL: store: +; CHECK-MINGW-F128: # %bb.0: +; CHECK-MINGW-F128-NEXT: movaps (%rcx), %xmm0 +; CHECK-MINGW-F128-NEXT: movaps %xmm0, (%rdx) +; CHECK-MINGW-F128-NEXT: retq +; +; CHECK-MINGW-I128-LABEL: store: +; CHECK-MINGW-I128: # %bb.0: +; CHECK-MINGW-I128-NEXT: movq %rdx, 8(%r8) +; CHECK-MINGW-I128-NEXT: movq %rcx, (%r8) +; CHECK-MINGW-I128-NEXT: retq +; +; CHECK-X86-LABEL: store: +; CHECK-X86: # %bb.0: +; CHECK-X86-NEXT: pushl %edi +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 16(%esp), %eax +; CHECK-X86-NEXT: movl 20(%esp), %ecx +; CHECK-X86-NEXT: movl 24(%esp), %edx +; CHECK-X86-NEXT: movl 28(%esp), %esi +; CHECK-X86-NEXT: movl 32(%esp), %edi +; CHECK-X86-NEXT: movl %esi, 12(%edi) +; CHECK-X86-NEXT: movl %edx, 8(%edi) +; CHECK-X86-NEXT: movl %ecx, 4(%edi) +; CHECK-X86-NEXT: movl %eax, (%edi) +; CHECK-X86-NEXT: addl $4, %esp +; CHECK-X86-NEXT: popl %esi +; CHECK-X86-NEXT: popl %edi +; CHECK-X86-NEXT: retl +; +; CHECK-MSVC32-LABEL: store: +; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %edi +; CHECK-MSVC32-NEXT: pushl %esi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 12(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 16(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 20(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 24(%ebp), %edi +; CHECK-MSVC32-NEXT: movl %esi, 12(%edi) +; CHECK-MSVC32-NEXT: movl %edx, 8(%edi) +; CHECK-MSVC32-NEXT: movl %ecx, 4(%edi) +; CHECK-MSVC32-NEXT: movl %eax, (%edi) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi +; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp +; CHECK-MSVC32-NEXT: retl + store PrimTy %x, ptr %p + ret void +} + +; Illustrate stack alignment +define void @store_perturbed(i8 %_0, PrimTy %x, ptr %p) nounwind { +; CHECK-X64-F128-LABEL: store_perturbed: +; CHECK-X64-F128: # %bb.0: +; CHECK-X64-F128-NEXT: movaps %xmm0, (%rsi) +; CHECK-X64-F128-NEXT: retq +; +; CHECK-X64-I128-LABEL: store_perturbed: +; CHECK-X64-I128: # %bb.0: +; CHECK-X64-I128-NEXT: movq %rdx, 8(%rcx) +; CHECK-X64-I128-NEXT: movq %rsi, (%rcx) +; CHECK-X64-I128-NEXT: retq +; +; CHECK-MSVC64-F128-LABEL: store_perturbed: +; CHECK-MSVC64-F128: # %bb.0: +; CHECK-MSVC64-F128-NEXT: movaps (%rdx), %xmm0 +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, (%r8) +; CHECK-MSVC64-F128-NEXT: retq +; +; CHECK-MSVC64-I128-LABEL: store_perturbed: +; CHECK-MSVC64-I128: # %bb.0: +; CHECK-MSVC64-I128-NEXT: movq %r8, 8(%r9) +; CHECK-MSVC64-I128-NEXT: movq %rdx, (%r9) +; CHECK-MSVC64-I128-NEXT: retq +; +; CHECK-MINGW-F128-LABEL: store_perturbed: +; CHECK-MINGW-F128: # %bb.0: +; CHECK-MINGW-F128-NEXT: movaps (%rdx), %xmm0 +; CHECK-MINGW-F128-NEXT: movaps %xmm0, (%r8) +; CHECK-MINGW-F128-NEXT: retq +; +; CHECK-MINGW-I128-LABEL: store_perturbed: +; CHECK-MINGW-I128: # %bb.0: +; CHECK-MINGW-I128-NEXT: movq %r8, 8(%r9) +; CHECK-MINGW-I128-NEXT: movq %rdx, (%r9) +; CHECK-MINGW-I128-NEXT: retq +; +; CHECK-X86-LABEL: store_perturbed: +; CHECK-X86: # %bb.0: +; CHECK-X86-NEXT: pushl %edi +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 32(%esp), %eax +; CHECK-X86-NEXT: movl 36(%esp), %ecx +; CHECK-X86-NEXT: movl 40(%esp), %edx +; CHECK-X86-NEXT: movl 44(%esp), %esi +; CHECK-X86-NEXT: movl 48(%esp), %edi +; CHECK-X86-NEXT: movl %esi, 12(%edi) +; CHECK-X86-NEXT: movl %edx, 8(%edi) +; CHECK-X86-NEXT: movl %ecx, 4(%edi) +; CHECK-X86-NEXT: movl %eax, (%edi) +; CHECK-X86-NEXT: addl $4, %esp +; CHECK-X86-NEXT: popl %esi +; CHECK-X86-NEXT: popl %edi +; CHECK-X86-NEXT: retl +; +; CHECK-MSVC32-LABEL: store_perturbed: +; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %edi +; CHECK-MSVC32-NEXT: pushl %esi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 24(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 28(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 32(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 36(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 40(%ebp), %edi +; CHECK-MSVC32-NEXT: movl %esi, 12(%edi) +; CHECK-MSVC32-NEXT: movl %edx, 8(%edi) +; CHECK-MSVC32-NEXT: movl %ecx, 4(%edi) +; CHECK-MSVC32-NEXT: movl %eax, (%edi) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi +; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp +; CHECK-MSVC32-NEXT: retl + store PrimTy %x, ptr %p + ret void +} + define PrimTy @return(ptr %p) nounwind { ; CHECK-X64-F128-LABEL: return: ; CHECK-X64-F128: # %bb.0: @@ -53,8 +214,8 @@ define PrimTy @return(ptr %p) nounwind { ; CHECK-X86: # %bb.0: ; CHECK-X86-NEXT: pushl %edi ; CHECK-X86-NEXT: pushl %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; CHECK-X86-NEXT: movl 12(%esp), %eax +; CHECK-X86-NEXT: movl 16(%esp), %ecx ; CHECK-X86-NEXT: movl (%ecx), %edx ; CHECK-X86-NEXT: movl 4(%ecx), %esi ; CHECK-X86-NEXT: movl 8(%ecx), %edi @@ -71,8 +232,8 @@ define PrimTy @return(ptr %p) nounwind { ; CHECK-MSVC32: # %bb.0: ; CHECK-MSVC32-NEXT: pushl %edi ; CHECK-MSVC32-NEXT: pushl %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %ecx +; CHECK-MSVC32-NEXT: movl 12(%esp), %eax +; CHECK-MSVC32-NEXT: movl 16(%esp), %ecx ; CHECK-MSVC32-NEXT: movl (%ecx), %edx ; CHECK-MSVC32-NEXT: movl 4(%ecx), %esi ; CHECK-MSVC32-NEXT: movl 8(%ecx), %edi @@ -123,38 +284,47 @@ define PrimTy @first_arg(PrimTy %x) nounwind { ; CHECK-X86: # %bb.0: ; CHECK-X86-NEXT: pushl %edi ; CHECK-X86-NEXT: pushl %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 16(%esp), %eax +; CHECK-X86-NEXT: movl 32(%esp), %ecx +; CHECK-X86-NEXT: movl 36(%esp), %edx +; CHECK-X86-NEXT: movl 40(%esp), %esi +; CHECK-X86-NEXT: movl 44(%esp), %edi ; CHECK-X86-NEXT: movl %edi, 12(%eax) ; CHECK-X86-NEXT: movl %esi, 8(%eax) ; CHECK-X86-NEXT: movl %edx, 4(%eax) ; CHECK-X86-NEXT: movl %ecx, (%eax) +; CHECK-X86-NEXT: addl $4, %esp ; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: popl %edi ; CHECK-X86-NEXT: retl $4 ; ; CHECK-MSVC32-LABEL: first_arg: ; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp ; CHECK-MSVC32-NEXT: pushl %edi ; CHECK-MSVC32-NEXT: pushl %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 24(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 28(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 32(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 36(%ebp), %edi ; CHECK-MSVC32-NEXT: movl %edi, 12(%eax) ; CHECK-MSVC32-NEXT: movl %esi, 8(%eax) ; CHECK-MSVC32-NEXT: movl %edx, 4(%eax) ; CHECK-MSVC32-NEXT: movl %ecx, (%eax) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp ; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl ret PrimTy %x } +; Leading args such that i128 is the last possible position where it still +; gets passed in registers. define PrimTy @leading_args(i64 %_0, i64 %_1, i64 %_2, i64 %_3, PrimTy %x) nounwind { ; CHECK-X64-F128-LABEL: leading_args: ; CHECK-X64-F128: # %bb.0: @@ -168,64 +338,72 @@ define PrimTy @leading_args(i64 %_0, i64 %_1, i64 %_2, i64 %_3, PrimTy %x) nounw ; ; CHECK-MSVC64-F128-LABEL: leading_args: ; CHECK-MSVC64-F128: # %bb.0: -; CHECK-MSVC64-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq 40(%rsp), %rax ; CHECK-MSVC64-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MSVC64-F128-NEXT: retq ; ; CHECK-MSVC64-I128-LABEL: leading_args: ; CHECK-MSVC64-I128: # %bb.0: -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MSVC64-I128-NEXT: movq 40(%rsp), %rax +; CHECK-MSVC64-I128-NEXT: movq 48(%rsp), %rdx ; CHECK-MSVC64-I128-NEXT: retq ; ; CHECK-MINGW-F128-LABEL: leading_args: ; CHECK-MINGW-F128: # %bb.0: -; CHECK-MINGW-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq 40(%rsp), %rax ; CHECK-MINGW-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MINGW-F128-NEXT: retq ; ; CHECK-MINGW-I128-LABEL: leading_args: ; CHECK-MINGW-I128: # %bb.0: -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MINGW-I128-NEXT: movq 40(%rsp), %rax +; CHECK-MINGW-I128-NEXT: movq 48(%rsp), %rdx ; CHECK-MINGW-I128-NEXT: retq ; ; CHECK-X86-LABEL: leading_args: ; CHECK-X86: # %bb.0: ; CHECK-X86-NEXT: pushl %edi ; CHECK-X86-NEXT: pushl %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 16(%esp), %eax +; CHECK-X86-NEXT: movl 64(%esp), %ecx +; CHECK-X86-NEXT: movl 68(%esp), %edx +; CHECK-X86-NEXT: movl 72(%esp), %esi +; CHECK-X86-NEXT: movl 76(%esp), %edi ; CHECK-X86-NEXT: movl %edi, 12(%eax) ; CHECK-X86-NEXT: movl %esi, 8(%eax) ; CHECK-X86-NEXT: movl %edx, 4(%eax) ; CHECK-X86-NEXT: movl %ecx, (%eax) +; CHECK-X86-NEXT: addl $4, %esp ; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: popl %edi ; CHECK-X86-NEXT: retl $4 ; ; CHECK-MSVC32-LABEL: leading_args: ; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp ; CHECK-MSVC32-NEXT: pushl %edi ; CHECK-MSVC32-NEXT: pushl %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 56(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 60(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 64(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 68(%ebp), %edi ; CHECK-MSVC32-NEXT: movl %edi, 12(%eax) ; CHECK-MSVC32-NEXT: movl %esi, 8(%eax) ; CHECK-MSVC32-NEXT: movl %edx, 4(%eax) ; CHECK-MSVC32-NEXT: movl %ecx, (%eax) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp ; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl ret PrimTy %x } +; The i128 of interest must be in memory. define PrimTy @many_leading_args(i64 %_0, i64 %_1, i64 %_2, i64 %_3, i64 %_4, PrimTy %_5, PrimTy %x) nounwind { ; CHECK-X64-F128-LABEL: many_leading_args: ; CHECK-X64-F128: # %bb.0: @@ -234,66 +412,73 @@ define PrimTy @many_leading_args(i64 %_0, i64 %_1, i64 %_2, i64 %_3, i64 %_4, Pr ; ; CHECK-X64-I128-LABEL: many_leading_args: ; CHECK-X64-I128: # %bb.0: -; CHECK-X64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-X64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-X64-I128-NEXT: movq 24(%rsp), %rax +; CHECK-X64-I128-NEXT: movq 32(%rsp), %rdx ; CHECK-X64-I128-NEXT: retq ; ; CHECK-MSVC64-F128-LABEL: many_leading_args: ; CHECK-MSVC64-F128: # %bb.0: -; CHECK-MSVC64-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq 56(%rsp), %rax ; CHECK-MSVC64-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MSVC64-F128-NEXT: retq ; ; CHECK-MSVC64-I128-LABEL: many_leading_args: ; CHECK-MSVC64-I128: # %bb.0: -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MSVC64-I128-NEXT: movq 64(%rsp), %rax +; CHECK-MSVC64-I128-NEXT: movq 72(%rsp), %rdx ; CHECK-MSVC64-I128-NEXT: retq ; ; CHECK-MINGW-F128-LABEL: many_leading_args: ; CHECK-MINGW-F128: # %bb.0: -; CHECK-MINGW-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq 56(%rsp), %rax ; CHECK-MINGW-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MINGW-F128-NEXT: retq ; ; CHECK-MINGW-I128-LABEL: many_leading_args: ; CHECK-MINGW-I128: # %bb.0: -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MINGW-I128-NEXT: movq 64(%rsp), %rax +; CHECK-MINGW-I128-NEXT: movq 72(%rsp), %rdx ; CHECK-MINGW-I128-NEXT: retq ; ; CHECK-X86-LABEL: many_leading_args: ; CHECK-X86: # %bb.0: ; CHECK-X86-NEXT: pushl %edi ; CHECK-X86-NEXT: pushl %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 16(%esp), %eax +; CHECK-X86-NEXT: movl 80(%esp), %ecx +; CHECK-X86-NEXT: movl 84(%esp), %edx +; CHECK-X86-NEXT: movl 88(%esp), %esi +; CHECK-X86-NEXT: movl 92(%esp), %edi ; CHECK-X86-NEXT: movl %edi, 12(%eax) ; CHECK-X86-NEXT: movl %esi, 8(%eax) ; CHECK-X86-NEXT: movl %edx, 4(%eax) ; CHECK-X86-NEXT: movl %ecx, (%eax) +; CHECK-X86-NEXT: addl $4, %esp ; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: popl %edi ; CHECK-X86-NEXT: retl $4 ; ; CHECK-MSVC32-LABEL: many_leading_args: ; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp ; CHECK-MSVC32-NEXT: pushl %edi ; CHECK-MSVC32-NEXT: pushl %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 72(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 76(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 80(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 84(%ebp), %edi ; CHECK-MSVC32-NEXT: movl %edi, 12(%eax) ; CHECK-MSVC32-NEXT: movl %esi, 8(%eax) ; CHECK-MSVC32-NEXT: movl %edx, 4(%eax) ; CHECK-MSVC32-NEXT: movl %ecx, (%eax) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp ; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl ret PrimTy %x } @@ -305,66 +490,73 @@ define PrimTy @trailing_arg(i64 %_0, i64 %_1, i64 %_2, i64 %_3, i64 %_4, PrimTy ; ; CHECK-X64-I128-LABEL: trailing_arg: ; CHECK-X64-I128: # %bb.0: -; CHECK-X64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-X64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-X64-I128-NEXT: movq 8(%rsp), %rax +; CHECK-X64-I128-NEXT: movq 16(%rsp), %rdx ; CHECK-X64-I128-NEXT: retq ; ; CHECK-MSVC64-F128-LABEL: trailing_arg: ; CHECK-MSVC64-F128: # %bb.0: -; CHECK-MSVC64-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq 48(%rsp), %rax ; CHECK-MSVC64-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MSVC64-F128-NEXT: retq ; ; CHECK-MSVC64-I128-LABEL: trailing_arg: ; CHECK-MSVC64-I128: # %bb.0: -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MSVC64-I128-NEXT: movq 48(%rsp), %rax +; CHECK-MSVC64-I128-NEXT: movq 56(%rsp), %rdx ; CHECK-MSVC64-I128-NEXT: retq ; ; CHECK-MINGW-F128-LABEL: trailing_arg: ; CHECK-MINGW-F128: # %bb.0: -; CHECK-MINGW-F128-NEXT: movq {{[0-9]+}}(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq 48(%rsp), %rax ; CHECK-MINGW-F128-NEXT: movaps (%rax), %xmm0 ; CHECK-MINGW-F128-NEXT: retq ; ; CHECK-MINGW-I128-LABEL: trailing_arg: ; CHECK-MINGW-I128: # %bb.0: -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-I128-NEXT: movq {{[0-9]+}}(%rsp), %rdx +; CHECK-MINGW-I128-NEXT: movq 48(%rsp), %rax +; CHECK-MINGW-I128-NEXT: movq 56(%rsp), %rdx ; CHECK-MINGW-I128-NEXT: retq ; ; CHECK-X86-LABEL: trailing_arg: ; CHECK-X86: # %bb.0: ; CHECK-X86-NEXT: pushl %edi ; CHECK-X86-NEXT: pushl %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: movl 16(%esp), %eax +; CHECK-X86-NEXT: movl 64(%esp), %ecx +; CHECK-X86-NEXT: movl 68(%esp), %edx +; CHECK-X86-NEXT: movl 72(%esp), %esi +; CHECK-X86-NEXT: movl 76(%esp), %edi ; CHECK-X86-NEXT: movl %edi, 12(%eax) ; CHECK-X86-NEXT: movl %esi, 8(%eax) ; CHECK-X86-NEXT: movl %edx, 4(%eax) ; CHECK-X86-NEXT: movl %ecx, (%eax) +; CHECK-X86-NEXT: addl $4, %esp ; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: popl %edi ; CHECK-X86-NEXT: retl $4 ; ; CHECK-MSVC32-LABEL: trailing_arg: ; CHECK-MSVC32: # %bb.0: +; CHECK-MSVC32-NEXT: pushl %ebp +; CHECK-MSVC32-NEXT: movl %esp, %ebp ; CHECK-MSVC32-NEXT: pushl %edi ; CHECK-MSVC32-NEXT: pushl %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %eax -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edx -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %esi -; CHECK-MSVC32-NEXT: movl {{[0-9]+}}(%esp), %edi +; CHECK-MSVC32-NEXT: andl $-16, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 56(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 60(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 64(%ebp), %esi +; CHECK-MSVC32-NEXT: movl 68(%ebp), %edi ; CHECK-MSVC32-NEXT: movl %edi, 12(%eax) ; CHECK-MSVC32-NEXT: movl %esi, 8(%eax) ; CHECK-MSVC32-NEXT: movl %edx, 4(%eax) ; CHECK-MSVC32-NEXT: movl %ecx, (%eax) +; CHECK-MSVC32-NEXT: leal -8(%ebp), %esp ; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %edi +; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl ret PrimTy %x } @@ -388,8 +580,8 @@ define void @call_first_arg(PrimTy %x) nounwind { ; CHECK-MSVC64-F128: # %bb.0: ; CHECK-MSVC64-F128-NEXT: subq $56, %rsp ; CHECK-MSVC64-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MSVC64-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rcx +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, 32(%rsp) +; CHECK-MSVC64-F128-NEXT: leaq 32(%rsp), %rcx ; CHECK-MSVC64-F128-NEXT: callq first_arg ; CHECK-MSVC64-F128-NEXT: addq $56, %rsp ; CHECK-MSVC64-F128-NEXT: retq @@ -405,8 +597,8 @@ define void @call_first_arg(PrimTy %x) nounwind { ; CHECK-MINGW-F128: # %bb.0: ; CHECK-MINGW-F128-NEXT: subq $56, %rsp ; CHECK-MINGW-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MINGW-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rcx +; CHECK-MINGW-F128-NEXT: movaps %xmm0, 32(%rsp) +; CHECK-MINGW-F128-NEXT: leaq 32(%rsp), %rcx ; CHECK-MINGW-F128-NEXT: callq first_arg ; CHECK-MINGW-F128-NEXT: addq $56, %rsp ; CHECK-MINGW-F128-NEXT: retq @@ -420,32 +612,43 @@ define void @call_first_arg(PrimTy %x) nounwind { ; ; CHECK-X86-LABEL: call_first_arg: ; CHECK-X86: # %bb.0: -; CHECK-X86-NEXT: subl $40, %esp -; CHECK-X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: subl $56, %esp +; CHECK-X86-NEXT: movl 64(%esp), %eax +; CHECK-X86-NEXT: movl 68(%esp), %ecx +; CHECK-X86-NEXT: movl 72(%esp), %edx +; CHECK-X86-NEXT: movl 76(%esp), %esi +; CHECK-X86-NEXT: movl %esi, 28(%esp) +; CHECK-X86-NEXT: movl %edx, 24(%esp) +; CHECK-X86-NEXT: movl %ecx, 20(%esp) +; CHECK-X86-NEXT: movl %eax, 16(%esp) +; CHECK-X86-NEXT: leal 32(%esp), %eax +; CHECK-X86-NEXT: movl %eax, (%esp) ; CHECK-X86-NEXT: calll first_arg@PLT -; CHECK-X86-NEXT: addl $56, %esp +; CHECK-X86-NEXT: addl $52, %esp +; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: retl ; ; CHECK-MSVC32-LABEL: call_first_arg: ; CHECK-MSVC32: # %bb.0: ; CHECK-MSVC32-NEXT: pushl %ebp ; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %esi ; CHECK-MSVC32-NEXT: andl $-16, %esp -; CHECK-MSVC32-NEXT: subl $32, %esp -; CHECK-MSVC32-NEXT: movl %esp, %eax -; CHECK-MSVC32-NEXT: pushl 20(%ebp) -; CHECK-MSVC32-NEXT: pushl 16(%ebp) -; CHECK-MSVC32-NEXT: pushl 12(%ebp) -; CHECK-MSVC32-NEXT: pushl 8(%ebp) -; CHECK-MSVC32-NEXT: pushl %eax +; CHECK-MSVC32-NEXT: subl $64, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 12(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 16(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 20(%ebp), %esi +; CHECK-MSVC32-NEXT: movl %esi, 28(%esp) +; CHECK-MSVC32-NEXT: movl %edx, 24(%esp) +; CHECK-MSVC32-NEXT: movl %ecx, 20(%esp) +; CHECK-MSVC32-NEXT: movl %eax, 16(%esp) +; CHECK-MSVC32-NEXT: leal 32(%esp), %eax +; CHECK-MSVC32-NEXT: movl %eax, (%esp) ; CHECK-MSVC32-NEXT: calll _first_arg -; CHECK-MSVC32-NEXT: addl $20, %esp -; CHECK-MSVC32-NEXT: movl %ebp, %esp +; CHECK-MSVC32-NEXT: leal -4(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl call PrimTy @first_arg(PrimTy %x) @@ -481,9 +684,9 @@ define void @call_leading_args(PrimTy %x) nounwind { ; CHECK-MSVC64-F128: # %bb.0: ; CHECK-MSVC64-F128-NEXT: subq $72, %rsp ; CHECK-MSVC64-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MSVC64-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MSVC64-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MSVC64-F128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-F128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-F128-NEXT: xorl %r8d, %r8d @@ -495,8 +698,8 @@ define void @call_leading_args(PrimTy %x) nounwind { ; CHECK-MSVC64-I128-LABEL: call_leading_args: ; CHECK-MSVC64-I128: # %bb.0: ; CHECK-MSVC64-I128-NEXT: subq $56, %rsp -; CHECK-MSVC64-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rdx, 40(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rcx, 32(%rsp) ; CHECK-MSVC64-I128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-I128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-I128-NEXT: xorl %r8d, %r8d @@ -509,9 +712,9 @@ define void @call_leading_args(PrimTy %x) nounwind { ; CHECK-MINGW-F128: # %bb.0: ; CHECK-MINGW-F128-NEXT: subq $72, %rsp ; CHECK-MINGW-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MINGW-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MINGW-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MINGW-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MINGW-F128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-F128-NEXT: xorl %edx, %edx ; CHECK-MINGW-F128-NEXT: xorl %r8d, %r8d @@ -523,8 +726,8 @@ define void @call_leading_args(PrimTy %x) nounwind { ; CHECK-MINGW-I128-LABEL: call_leading_args: ; CHECK-MINGW-I128: # %bb.0: ; CHECK-MINGW-I128-NEXT: subq $56, %rsp -; CHECK-MINGW-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MINGW-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rdx, 40(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rcx, 32(%rsp) ; CHECK-MINGW-I128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-I128-NEXT: xorl %edx, %edx ; CHECK-MINGW-I128-NEXT: xorl %r8d, %r8d @@ -535,48 +738,59 @@ define void @call_leading_args(PrimTy %x) nounwind { ; ; CHECK-X86-LABEL: call_leading_args: ; CHECK-X86: # %bb.0: -; CHECK-X86-NEXT: subl $40, %esp -; CHECK-X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: subl $88, %esp +; CHECK-X86-NEXT: movl 96(%esp), %eax +; CHECK-X86-NEXT: movl 100(%esp), %ecx +; CHECK-X86-NEXT: movl 104(%esp), %edx +; CHECK-X86-NEXT: movl 108(%esp), %esi +; CHECK-X86-NEXT: movl %esi, 60(%esp) +; CHECK-X86-NEXT: movl %edx, 56(%esp) +; CHECK-X86-NEXT: movl %ecx, 52(%esp) +; CHECK-X86-NEXT: movl %eax, 48(%esp) +; CHECK-X86-NEXT: leal 64(%esp), %eax +; CHECK-X86-NEXT: movl %eax, (%esp) +; CHECK-X86-NEXT: movl $0, 32(%esp) +; CHECK-X86-NEXT: movl $0, 28(%esp) +; CHECK-X86-NEXT: movl $0, 24(%esp) +; CHECK-X86-NEXT: movl $0, 20(%esp) +; CHECK-X86-NEXT: movl $0, 16(%esp) +; CHECK-X86-NEXT: movl $0, 12(%esp) +; CHECK-X86-NEXT: movl $0, 8(%esp) +; CHECK-X86-NEXT: movl $0, 4(%esp) ; CHECK-X86-NEXT: calll leading_args@PLT -; CHECK-X86-NEXT: addl $88, %esp +; CHECK-X86-NEXT: addl $84, %esp +; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: retl ; ; CHECK-MSVC32-LABEL: call_leading_args: ; CHECK-MSVC32: # %bb.0: ; CHECK-MSVC32-NEXT: pushl %ebp ; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %esi ; CHECK-MSVC32-NEXT: andl $-16, %esp -; CHECK-MSVC32-NEXT: subl $32, %esp -; CHECK-MSVC32-NEXT: movl %esp, %eax -; CHECK-MSVC32-NEXT: pushl 20(%ebp) -; CHECK-MSVC32-NEXT: pushl 16(%ebp) -; CHECK-MSVC32-NEXT: pushl 12(%ebp) -; CHECK-MSVC32-NEXT: pushl 8(%ebp) -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl %eax +; CHECK-MSVC32-NEXT: subl $96, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 12(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 16(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 20(%ebp), %esi +; CHECK-MSVC32-NEXT: movl %esi, 60(%esp) +; CHECK-MSVC32-NEXT: movl %edx, 56(%esp) +; CHECK-MSVC32-NEXT: movl %ecx, 52(%esp) +; CHECK-MSVC32-NEXT: movl %eax, 48(%esp) +; CHECK-MSVC32-NEXT: leal 64(%esp), %eax +; CHECK-MSVC32-NEXT: movl %eax, (%esp) +; CHECK-MSVC32-NEXT: movl $0, 32(%esp) +; CHECK-MSVC32-NEXT: movl $0, 28(%esp) +; CHECK-MSVC32-NEXT: movl $0, 24(%esp) +; CHECK-MSVC32-NEXT: movl $0, 20(%esp) +; CHECK-MSVC32-NEXT: movl $0, 16(%esp) +; CHECK-MSVC32-NEXT: movl $0, 12(%esp) +; CHECK-MSVC32-NEXT: movl $0, 8(%esp) +; CHECK-MSVC32-NEXT: movl $0, 4(%esp) ; CHECK-MSVC32-NEXT: calll _leading_args -; CHECK-MSVC32-NEXT: addl $52, %esp -; CHECK-MSVC32-NEXT: movl %ebp, %esp +; CHECK-MSVC32-NEXT: leal -4(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl call PrimTy @leading_args(i64 0, i64 0, i64 0, i64 0, PrimTy %x) @@ -620,12 +834,12 @@ define void @call_many_leading_args(PrimTy %x) nounwind { ; CHECK-MSVC64-F128-NEXT: subq $88, %rsp ; CHECK-MSVC64-F128-NEXT: movaps (%rcx), %xmm0 ; CHECK-MSVC64-F128-NEXT: xorps %xmm1, %xmm1 -; CHECK-MSVC64-F128-NEXT: movaps %xmm1, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-F128-NEXT: movaps %xmm1, 64(%rsp) +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MSVC64-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq %rax, 40(%rsp) +; CHECK-MSVC64-F128-NEXT: leaq 64(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MSVC64-F128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-F128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-F128-NEXT: xorl %r8d, %r8d @@ -637,10 +851,10 @@ define void @call_many_leading_args(PrimTy %x) nounwind { ; CHECK-MSVC64-I128-LABEL: call_many_leading_args: ; CHECK-MSVC64-I128: # %bb.0: ; CHECK-MSVC64-I128-NEXT: subq $72, %rsp -; CHECK-MSVC64-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-I128-NEXT: movq $0, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-I128-NEXT: movq $0, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rdx, 56(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rcx, 48(%rsp) +; CHECK-MSVC64-I128-NEXT: movq $0, 40(%rsp) +; CHECK-MSVC64-I128-NEXT: movq $0, 32(%rsp) ; CHECK-MSVC64-I128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-I128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-I128-NEXT: xorl %r8d, %r8d @@ -654,12 +868,12 @@ define void @call_many_leading_args(PrimTy %x) nounwind { ; CHECK-MINGW-F128-NEXT: subq $88, %rsp ; CHECK-MINGW-F128-NEXT: movaps (%rcx), %xmm0 ; CHECK-MINGW-F128-NEXT: xorps %xmm1, %xmm1 -; CHECK-MINGW-F128-NEXT: movaps %xmm1, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MINGW-F128-NEXT: movaps %xmm1, 64(%rsp) +; CHECK-MINGW-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MINGW-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq %rax, 40(%rsp) +; CHECK-MINGW-F128-NEXT: leaq 64(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MINGW-F128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-F128-NEXT: xorl %edx, %edx ; CHECK-MINGW-F128-NEXT: xorl %r8d, %r8d @@ -671,10 +885,10 @@ define void @call_many_leading_args(PrimTy %x) nounwind { ; CHECK-MINGW-I128-LABEL: call_many_leading_args: ; CHECK-MINGW-I128: # %bb.0: ; CHECK-MINGW-I128-NEXT: subq $72, %rsp -; CHECK-MINGW-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MINGW-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) -; CHECK-MINGW-I128-NEXT: movq $0, {{[0-9]+}}(%rsp) -; CHECK-MINGW-I128-NEXT: movq $0, {{[0-9]+}}(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rdx, 56(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rcx, 48(%rsp) +; CHECK-MINGW-I128-NEXT: movq $0, 40(%rsp) +; CHECK-MINGW-I128-NEXT: movq $0, 32(%rsp) ; CHECK-MINGW-I128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-I128-NEXT: xorl %edx, %edx ; CHECK-MINGW-I128-NEXT: xorl %r8d, %r8d @@ -685,56 +899,67 @@ define void @call_many_leading_args(PrimTy %x) nounwind { ; ; CHECK-X86-LABEL: call_many_leading_args: ; CHECK-X86: # %bb.0: -; CHECK-X86-NEXT: subl $40, %esp -; CHECK-X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: subl $104, %esp +; CHECK-X86-NEXT: movl 112(%esp), %eax +; CHECK-X86-NEXT: movl 116(%esp), %ecx +; CHECK-X86-NEXT: movl 120(%esp), %edx +; CHECK-X86-NEXT: movl 124(%esp), %esi +; CHECK-X86-NEXT: movl %esi, 76(%esp) +; CHECK-X86-NEXT: movl %edx, 72(%esp) +; CHECK-X86-NEXT: movl %ecx, 68(%esp) +; CHECK-X86-NEXT: movl %eax, 64(%esp) +; CHECK-X86-NEXT: leal 80(%esp), %eax +; CHECK-X86-NEXT: movl %eax, (%esp) +; CHECK-X86-NEXT: movl $0, 60(%esp) +; CHECK-X86-NEXT: movl $0, 56(%esp) +; CHECK-X86-NEXT: movl $0, 52(%esp) +; CHECK-X86-NEXT: movl $0, 48(%esp) +; CHECK-X86-NEXT: movl $0, 32(%esp) +; CHECK-X86-NEXT: movl $0, 28(%esp) +; CHECK-X86-NEXT: movl $0, 24(%esp) +; CHECK-X86-NEXT: movl $0, 20(%esp) +; CHECK-X86-NEXT: movl $0, 16(%esp) +; CHECK-X86-NEXT: movl $0, 12(%esp) +; CHECK-X86-NEXT: movl $0, 8(%esp) +; CHECK-X86-NEXT: movl $0, 4(%esp) ; CHECK-X86-NEXT: calll many_leading_args@PLT -; CHECK-X86-NEXT: addl $104, %esp +; CHECK-X86-NEXT: addl $100, %esp +; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: retl ; ; CHECK-MSVC32-LABEL: call_many_leading_args: ; CHECK-MSVC32: # %bb.0: ; CHECK-MSVC32-NEXT: pushl %ebp ; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %esi ; CHECK-MSVC32-NEXT: andl $-16, %esp -; CHECK-MSVC32-NEXT: subl $32, %esp -; CHECK-MSVC32-NEXT: movl %esp, %eax -; CHECK-MSVC32-NEXT: pushl 20(%ebp) -; CHECK-MSVC32-NEXT: pushl 16(%ebp) -; CHECK-MSVC32-NEXT: pushl 12(%ebp) -; CHECK-MSVC32-NEXT: pushl 8(%ebp) -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl %eax +; CHECK-MSVC32-NEXT: subl $112, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 12(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 16(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 20(%ebp), %esi +; CHECK-MSVC32-NEXT: movl %esi, 76(%esp) +; CHECK-MSVC32-NEXT: movl %edx, 72(%esp) +; CHECK-MSVC32-NEXT: movl %ecx, 68(%esp) +; CHECK-MSVC32-NEXT: movl %eax, 64(%esp) +; CHECK-MSVC32-NEXT: leal 80(%esp), %eax +; CHECK-MSVC32-NEXT: movl %eax, (%esp) +; CHECK-MSVC32-NEXT: movl $0, 60(%esp) +; CHECK-MSVC32-NEXT: movl $0, 56(%esp) +; CHECK-MSVC32-NEXT: movl $0, 52(%esp) +; CHECK-MSVC32-NEXT: movl $0, 48(%esp) +; CHECK-MSVC32-NEXT: movl $0, 32(%esp) +; CHECK-MSVC32-NEXT: movl $0, 28(%esp) +; CHECK-MSVC32-NEXT: movl $0, 24(%esp) +; CHECK-MSVC32-NEXT: movl $0, 20(%esp) +; CHECK-MSVC32-NEXT: movl $0, 16(%esp) +; CHECK-MSVC32-NEXT: movl $0, 12(%esp) +; CHECK-MSVC32-NEXT: movl $0, 8(%esp) +; CHECK-MSVC32-NEXT: movl $0, 4(%esp) ; CHECK-MSVC32-NEXT: calll _many_leading_args -; CHECK-MSVC32-NEXT: addl $68, %esp -; CHECK-MSVC32-NEXT: movl %ebp, %esp +; CHECK-MSVC32-NEXT: leal -4(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl call PrimTy @many_leading_args(i64 0, i64 0, i64 0, i64 0, PrimTy Prim0, PrimTy %x) @@ -770,9 +995,9 @@ define void @call_trailing_arg(PrimTy %x) nounwind { ; CHECK-MSVC64-F128: # %bb.0: ; CHECK-MSVC64-F128-NEXT: subq $72, %rsp ; CHECK-MSVC64-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MSVC64-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MSVC64-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MSVC64-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MSVC64-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MSVC64-F128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-F128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-F128-NEXT: xorl %r8d, %r8d @@ -784,8 +1009,8 @@ define void @call_trailing_arg(PrimTy %x) nounwind { ; CHECK-MSVC64-I128-LABEL: call_trailing_arg: ; CHECK-MSVC64-I128: # %bb.0: ; CHECK-MSVC64-I128-NEXT: subq $56, %rsp -; CHECK-MSVC64-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MSVC64-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rdx, 40(%rsp) +; CHECK-MSVC64-I128-NEXT: movq %rcx, 32(%rsp) ; CHECK-MSVC64-I128-NEXT: xorl %ecx, %ecx ; CHECK-MSVC64-I128-NEXT: xorl %edx, %edx ; CHECK-MSVC64-I128-NEXT: xorl %r8d, %r8d @@ -798,9 +1023,9 @@ define void @call_trailing_arg(PrimTy %x) nounwind { ; CHECK-MINGW-F128: # %bb.0: ; CHECK-MINGW-F128-NEXT: subq $72, %rsp ; CHECK-MINGW-F128-NEXT: movaps (%rcx), %xmm0 -; CHECK-MINGW-F128-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) -; CHECK-MINGW-F128-NEXT: leaq {{[0-9]+}}(%rsp), %rax -; CHECK-MINGW-F128-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; CHECK-MINGW-F128-NEXT: movaps %xmm0, 48(%rsp) +; CHECK-MINGW-F128-NEXT: leaq 48(%rsp), %rax +; CHECK-MINGW-F128-NEXT: movq %rax, 32(%rsp) ; CHECK-MINGW-F128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-F128-NEXT: xorl %edx, %edx ; CHECK-MINGW-F128-NEXT: xorl %r8d, %r8d @@ -812,8 +1037,8 @@ define void @call_trailing_arg(PrimTy %x) nounwind { ; CHECK-MINGW-I128-LABEL: call_trailing_arg: ; CHECK-MINGW-I128: # %bb.0: ; CHECK-MINGW-I128-NEXT: subq $56, %rsp -; CHECK-MINGW-I128-NEXT: movq %rdx, {{[0-9]+}}(%rsp) -; CHECK-MINGW-I128-NEXT: movq %rcx, {{[0-9]+}}(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rdx, 40(%rsp) +; CHECK-MINGW-I128-NEXT: movq %rcx, 32(%rsp) ; CHECK-MINGW-I128-NEXT: xorl %ecx, %ecx ; CHECK-MINGW-I128-NEXT: xorl %edx, %edx ; CHECK-MINGW-I128-NEXT: xorl %r8d, %r8d @@ -824,48 +1049,59 @@ define void @call_trailing_arg(PrimTy %x) nounwind { ; ; CHECK-X86-LABEL: call_trailing_arg: ; CHECK-X86: # %bb.0: -; CHECK-X86-NEXT: subl $40, %esp -; CHECK-X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl $0 -; CHECK-X86-NEXT: pushl %eax +; CHECK-X86-NEXT: pushl %esi +; CHECK-X86-NEXT: subl $88, %esp +; CHECK-X86-NEXT: movl 96(%esp), %eax +; CHECK-X86-NEXT: movl 100(%esp), %ecx +; CHECK-X86-NEXT: movl 104(%esp), %edx +; CHECK-X86-NEXT: movl 108(%esp), %esi +; CHECK-X86-NEXT: movl %esi, 60(%esp) +; CHECK-X86-NEXT: movl %edx, 56(%esp) +; CHECK-X86-NEXT: movl %ecx, 52(%esp) +; CHECK-X86-NEXT: movl %eax, 48(%esp) +; CHECK-X86-NEXT: leal 64(%esp), %eax +; CHECK-X86-NEXT: movl %eax, (%esp) +; CHECK-X86-NEXT: movl $0, 32(%esp) +; CHECK-X86-NEXT: movl $0, 28(%esp) +; CHECK-X86-NEXT: movl $0, 24(%esp) +; CHECK-X86-NEXT: movl $0, 20(%esp) +; CHECK-X86-NEXT: movl $0, 16(%esp) +; CHECK-X86-NEXT: movl $0, 12(%esp) +; CHECK-X86-NEXT: movl $0, 8(%esp) +; CHECK-X86-NEXT: movl $0, 4(%esp) ; CHECK-X86-NEXT: calll trailing_arg@PLT -; CHECK-X86-NEXT: addl $88, %esp +; CHECK-X86-NEXT: addl $84, %esp +; CHECK-X86-NEXT: popl %esi ; CHECK-X86-NEXT: retl ; ; CHECK-MSVC32-LABEL: call_trailing_arg: ; CHECK-MSVC32: # %bb.0: ; CHECK-MSVC32-NEXT: pushl %ebp ; CHECK-MSVC32-NEXT: movl %esp, %ebp +; CHECK-MSVC32-NEXT: pushl %esi ; CHECK-MSVC32-NEXT: andl $-16, %esp -; CHECK-MSVC32-NEXT: subl $32, %esp -; CHECK-MSVC32-NEXT: movl %esp, %eax -; CHECK-MSVC32-NEXT: pushl 20(%ebp) -; CHECK-MSVC32-NEXT: pushl 16(%ebp) -; CHECK-MSVC32-NEXT: pushl 12(%ebp) -; CHECK-MSVC32-NEXT: pushl 8(%ebp) -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl $0 -; CHECK-MSVC32-NEXT: pushl %eax +; CHECK-MSVC32-NEXT: subl $96, %esp +; CHECK-MSVC32-NEXT: movl 8(%ebp), %eax +; CHECK-MSVC32-NEXT: movl 12(%ebp), %ecx +; CHECK-MSVC32-NEXT: movl 16(%ebp), %edx +; CHECK-MSVC32-NEXT: movl 20(%ebp), %esi +; CHECK-MSVC32-NEXT: movl %esi, 60(%esp) +; CHECK-MSVC32-NEXT: movl %edx, 56(%esp) +; CHECK-MSVC32-NEXT: movl %ecx, 52(%esp) +; CHECK-MSVC32-NEXT: movl %eax, 48(%esp) +; CHECK-MSVC32-NEXT: leal 64(%esp), %eax +; CHECK-MSVC32-NEXT: movl %eax, (%esp) +; CHECK-MSVC32-NEXT: movl $0, 32(%esp) +; CHECK-MSVC32-NEXT: movl $0, 28(%esp) +; CHECK-MSVC32-NEXT: movl $0, 24(%esp) +; CHECK-MSVC32-NEXT: movl $0, 20(%esp) +; CHECK-MSVC32-NEXT: movl $0, 16(%esp) +; CHECK-MSVC32-NEXT: movl $0, 12(%esp) +; CHECK-MSVC32-NEXT: movl $0, 8(%esp) +; CHECK-MSVC32-NEXT: movl $0, 4(%esp) ; CHECK-MSVC32-NEXT: calll _trailing_arg -; CHECK-MSVC32-NEXT: addl $52, %esp -; CHECK-MSVC32-NEXT: movl %ebp, %esp +; CHECK-MSVC32-NEXT: leal -4(%ebp), %esp +; CHECK-MSVC32-NEXT: popl %esi ; CHECK-MSVC32-NEXT: popl %ebp ; CHECK-MSVC32-NEXT: retl call PrimTy @trailing_arg(i64 0, i64 0, i64 0, i64 0, PrimTy %x) diff --git a/llvm/test/CodeGen/X86/i128-sdiv.ll b/llvm/test/CodeGen/X86/i128-sdiv.ll index 717f52f198ee..7d5757392c98 100644 --- a/llvm/test/CodeGen/X86/i128-sdiv.ll +++ b/llvm/test/CodeGen/X86/i128-sdiv.ll @@ -8,18 +8,21 @@ define i128 @test1(i128 %x) nounwind { ; X86-LABEL: test1: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 36(%ebp), %ecx ; X86-NEXT: movl %ecx, %esi ; X86-NEXT: sarl $31, %esi ; X86-NEXT: movl %esi, %edx ; X86-NEXT: shrl $30, %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl 24(%ebp), %edi ; X86-NEXT: addl %esi, %edi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %esi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %edx +; X86-NEXT: adcl 28(%ebp), %esi +; X86-NEXT: adcl 32(%ebp), %edx ; X86-NEXT: adcl $0, %ecx ; X86-NEXT: shrdl $2, %ecx, %edx ; X86-NEXT: movl %ecx, %esi @@ -29,8 +32,10 @@ define i128 @test1(i128 %x) nounwind { ; X86-NEXT: movl %ecx, 8(%eax) ; X86-NEXT: movl %esi, 4(%eax) ; X86-NEXT: movl %edx, (%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: test1: @@ -52,38 +57,44 @@ define i128 @test1(i128 %x) nounwind { define i128 @test2(i128 %x) nounwind { ; X86-LABEL: test2: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl %edx, %esi -; X86-NEXT: sarl $31, %esi -; X86-NEXT: movl %esi, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl %eax, %edx +; X86-NEXT: sarl $31, %edx +; X86-NEXT: movl %edx, %ecx ; X86-NEXT: shrl $30, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: addl %esi, %edi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %esi -; X86-NEXT: adcl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: adcl $0, %edx -; X86-NEXT: shrdl $2, %edx, %ecx -; X86-NEXT: movl %edx, %esi +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: addl %edx, %esi +; X86-NEXT: adcl 28(%ebp), %edx +; X86-NEXT: adcl 32(%ebp), %ecx +; X86-NEXT: adcl $0, %eax +; X86-NEXT: shrdl $2, %eax, %ecx +; X86-NEXT: movl %eax, %esi ; X86-NEXT: sarl $31, %esi -; X86-NEXT: sarl $2, %edx -; X86-NEXT: xorl %edi, %edi +; X86-NEXT: sarl $2, %eax +; X86-NEXT: xorl %edx, %edx ; X86-NEXT: negl %ecx +; X86-NEXT: movl $0, %edi +; X86-NEXT: sbbl %eax, %edi ; X86-NEXT: movl $0, %ebx -; X86-NEXT: sbbl %edx, %ebx -; X86-NEXT: movl $0, %edx +; X86-NEXT: sbbl %esi, %ebx ; X86-NEXT: sbbl %esi, %edx -; X86-NEXT: sbbl %esi, %edi +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %ebx, 4(%eax) -; X86-NEXT: movl %edx, 8(%eax) -; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %ebx, 8(%eax) +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: test2: diff --git a/llvm/test/CodeGen/X86/i128-udiv.ll b/llvm/test/CodeGen/X86/i128-udiv.ll index 3f890b7f2443..901183242132 100644 --- a/llvm/test/CodeGen/X86/i128-udiv.ll +++ b/llvm/test/CodeGen/X86/i128-udiv.ll @@ -8,15 +8,21 @@ define i128 @test1(i128 %x) nounwind { ; X86-LABEL: test1: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %edx ; X86-NEXT: shrdl $2, %edx, %ecx ; X86-NEXT: shrl $2, %edx ; X86-NEXT: movl %edx, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) ; X86-NEXT: movl $0, 12(%eax) ; X86-NEXT: movl $0, 8(%eax) +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: test1: diff --git a/llvm/test/CodeGen/X86/iabs.ll b/llvm/test/CodeGen/X86/iabs.ll index 55c318e87a5a..bdceeefbcfab 100644 --- a/llvm/test/CodeGen/X86/iabs.ll +++ b/llvm/test/CodeGen/X86/iabs.ll @@ -123,31 +123,34 @@ define i64 @test_i64(i64 %a) nounwind { define i128 @test_i128(i128 %a) nounwind { ; X86-LABEL: test_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %edx -; X86-NEXT: sarl $31, %edx -; X86-NEXT: xorl %edx, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: xorl %edx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: xorl %edx, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: xorl %edx, %ebx -; X86-NEXT: subl %edx, %ebx -; X86-NEXT: sbbl %edx, %edi -; X86-NEXT: sbbl %edx, %esi -; X86-NEXT: sbbl %edx, %ecx -; X86-NEXT: movl %ebx, (%eax) -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: xorl %eax, %ecx +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: xorl %eax, %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: xorl %eax, %esi +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: xorl %eax, %edi +; X86-NEXT: subl %eax, %edi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %eax, %edx +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edi, (%eax) +; X86-NEXT: movl %esi, 4(%eax) +; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: test_i128: diff --git a/llvm/test/CodeGen/X86/icmp-shift-opt.ll b/llvm/test/CodeGen/X86/icmp-shift-opt.ll index c52b3ed6c926..4a6c1d0ae5de 100644 --- a/llvm/test/CodeGen/X86/icmp-shift-opt.ll +++ b/llvm/test/CodeGen/X86/icmp-shift-opt.ll @@ -10,33 +10,39 @@ define i128 @opt_setcc_lt_power_of_2(i128 %a) nounwind { ; X86-LABEL: opt_setcc_lt_power_of_2: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %edi +; X86-NEXT: movl 24(%ebp), %esi ; X86-NEXT: .p2align 4 ; X86-NEXT: .LBB0_1: # %loop ; X86-NEXT: # =>This Inner Loop Header: Depth=1 -; X86-NEXT: addl $1, %edi -; X86-NEXT: adcl $0, %esi +; X86-NEXT: addl $1, %esi +; X86-NEXT: adcl $0, %edi ; X86-NEXT: adcl $0, %edx ; X86-NEXT: adcl $0, %ecx ; X86-NEXT: movl %edx, %ebx ; X86-NEXT: orl %ecx, %ebx -; X86-NEXT: movl %esi, %ebp -; X86-NEXT: orl %edx, %ebp -; X86-NEXT: orl %ecx, %ebp -; X86-NEXT: shrdl $28, %ebx, %ebp +; X86-NEXT: movl %esi, %eax +; X86-NEXT: movl %edi, %esi +; X86-NEXT: orl %edx, %esi +; X86-NEXT: orl %ecx, %esi +; X86-NEXT: shrdl $28, %ebx, %esi +; X86-NEXT: movl %eax, %esi ; X86-NEXT: jne .LBB0_1 ; X86-NEXT: # %bb.2: # %exit -; X86-NEXT: movl %edi, (%eax) -; X86-NEXT: movl %esi, 4(%eax) +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %esi, (%eax) +; X86-NEXT: movl %edi, 4(%eax) ; X86-NEXT: movl %edx, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -73,15 +79,21 @@ exit: define i1 @opt_setcc_srl_eq_zero(i128 %a) nounwind { ; X86-LABEL: opt_setcc_srl_eq_zero: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: orl 20(%ebp), %ecx +; X86-NEXT: movl 8(%ebp), %edx ; X86-NEXT: orl %eax, %edx ; X86-NEXT: orl %ecx, %edx ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: shldl $15, %edx, %ecx ; X86-NEXT: sete %al +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; X64-LABEL: opt_setcc_srl_eq_zero: @@ -98,15 +110,21 @@ define i1 @opt_setcc_srl_eq_zero(i128 %a) nounwind { define i1 @opt_setcc_srl_ne_zero(i128 %a) nounwind { ; X86-LABEL: opt_setcc_srl_ne_zero: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: orl 20(%ebp), %ecx +; X86-NEXT: movl 8(%ebp), %edx ; X86-NEXT: orl %eax, %edx ; X86-NEXT: orl %ecx, %edx ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: shldl $15, %edx, %ecx ; X86-NEXT: setne %al +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; X64-LABEL: opt_setcc_srl_ne_zero: @@ -123,13 +141,19 @@ define i1 @opt_setcc_srl_ne_zero(i128 %a) nounwind { define i1 @opt_setcc_shl_eq_zero(i128 %a) nounwind { ; X86-LABEL: opt_setcc_shl_eq_zero: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl 20(%ebp), %ecx ; X86-NEXT: shll $17, %ecx -; X86-NEXT: orl {{[0-9]+}}(%esp), %eax -; X86-NEXT: orl {{[0-9]+}}(%esp), %eax +; X86-NEXT: orl 8(%ebp), %eax +; X86-NEXT: orl 12(%ebp), %eax ; X86-NEXT: orl %ecx, %eax ; X86-NEXT: sete %al +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; X64-LABEL: opt_setcc_shl_eq_zero: @@ -146,13 +170,19 @@ define i1 @opt_setcc_shl_eq_zero(i128 %a) nounwind { define i1 @opt_setcc_shl_ne_zero(i128 %a) nounwind { ; X86-LABEL: opt_setcc_shl_ne_zero: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl 20(%ebp), %ecx ; X86-NEXT: shll $17, %ecx -; X86-NEXT: orl {{[0-9]+}}(%esp), %eax -; X86-NEXT: orl {{[0-9]+}}(%esp), %eax +; X86-NEXT: orl 8(%ebp), %eax +; X86-NEXT: orl 12(%ebp), %eax ; X86-NEXT: orl %ecx, %eax ; X86-NEXT: setne %al +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; X64-LABEL: opt_setcc_shl_ne_zero: @@ -170,13 +200,17 @@ define i1 @opt_setcc_shl_ne_zero(i128 %a) nounwind { define i1 @opt_setcc_shl_eq_zero_multiple_shl_users(i128 %a) nounwind { ; X86-LABEL: opt_setcc_shl_eq_zero_multiple_shl_users: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movl 16(%ebp), %edx +; X86-NEXT: movl 20(%ebp), %esi ; X86-NEXT: shldl $17, %edx, %esi ; X86-NEXT: shldl $17, %ecx, %edx ; X86-NEXT: shldl $17, %eax, %ecx @@ -194,9 +228,11 @@ define i1 @opt_setcc_shl_eq_zero_multiple_shl_users(i128 %a) nounwind { ; X86-NEXT: calll use@PLT ; X86-NEXT: addl $16, %esp ; X86-NEXT: movl %ebx, %eax +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; X64-LABEL: opt_setcc_shl_eq_zero_multiple_shl_users: diff --git a/llvm/test/CodeGen/X86/invalid-operand-bundle-call.ll b/llvm/test/CodeGen/X86/invalid-operand-bundle-call.ll index 1f3a458ecf3a..17065a4a61c2 100644 --- a/llvm/test/CodeGen/X86/invalid-operand-bundle-call.ll +++ b/llvm/test/CodeGen/X86/invalid-operand-bundle-call.ll @@ -1,10 +1,10 @@ ; RUN: not llc -mtriple=x86_64-unknown-linux-gnu < %s 2>&1 | FileCheck %s -; CHECK: LLVM ERROR: cannot lower calls with arbitrary operand bundles! +; CHECK: LLVM ERROR: cannot lower calls with arbitrary operand bundles: foo, bar, baz declare void @g() define void @f(i32 %arg) { - call void @g() [ "foo"(i32 %arg) ] + call void @g() [ "foo"(i32 %arg), "bar"(i32 %arg), "baz"(i32 %arg) ] ret void } diff --git a/llvm/test/CodeGen/X86/invalid-operand-bundle-callbr.ll b/llvm/test/CodeGen/X86/invalid-operand-bundle-callbr.ll index 56157d205b1c..79bddc075541 100644 --- a/llvm/test/CodeGen/X86/invalid-operand-bundle-callbr.ll +++ b/llvm/test/CodeGen/X86/invalid-operand-bundle-callbr.ll @@ -1,6 +1,6 @@ ; RUN: not llc -mtriple=x86_64-unknown-linux-gnu < %s 2>&1 | FileCheck %s -; CHECK: LLVM ERROR: cannot lower callbrs with arbitrary operand bundles! +; CHECK: LLVM ERROR: cannot lower callbrs with arbitrary operand bundles: foo define void @f(i32 %arg) { callbr void asm "", ""() [ "foo"(i32 %arg) ] diff --git a/llvm/test/CodeGen/X86/invalid-operand-bundle-invoke.ll b/llvm/test/CodeGen/X86/invalid-operand-bundle-invoke.ll index 8091a220a44c..1da41aeab68b 100644 --- a/llvm/test/CodeGen/X86/invalid-operand-bundle-invoke.ll +++ b/llvm/test/CodeGen/X86/invalid-operand-bundle-invoke.ll @@ -1,6 +1,6 @@ ; RUN: not llc -mtriple=x86_64-unknown-linux-gnu < %s 2>&1 | FileCheck %s -; CHECK: LLVM ERROR: cannot lower invokes with arbitrary operand bundles! +; CHECK: LLVM ERROR: cannot lower invokes with arbitrary operand bundles: foo declare void @g() declare i32 @__gxx_personality_v0(...) diff --git a/llvm/test/CodeGen/X86/kcfi.ll b/llvm/test/CodeGen/X86/kcfi.ll index 059efcc71b0e..fd93b8e3d418 100644 --- a/llvm/test/CodeGen/X86/kcfi.ll +++ b/llvm/test/CodeGen/X86/kcfi.ll @@ -138,6 +138,29 @@ define void @f8() { ret void } +declare i32 @__gxx_personality_v0(...) + +define void @f9() personality ptr @__gxx_personality_v0 { +; MIR-LABEL: name: f9 +; MIR: body: +; ISEL: CALL64m killed %0, 1, $noreg, 0, $noreg, csr_64, implicit $rsp, implicit $ssp, implicit-def $rsp, implicit-def $ssp, cfi-type 12345678 +; KCFI: $r11 = MOV64rm killed renamable $rax, 1, $noreg, 0, $noreg +; KCFI-NEXT: BUNDLE{{.*}} { +; KCFI-NEXT: KCFI_CHECK $r11, 12345678, implicit-def $r10, implicit-def $r11, implicit-def $eflags +; KCFI-NEXT: CALL64r internal $r11, csr_64, implicit $rsp, implicit $ssp, implicit-def $rsp, implicit-def $ssp +; KCFI-NEXT: } + %1 = load ptr, ptr @g, align 8 + invoke void %1() [ "kcfi"(i32 12345678) ] + to label %cont + unwind label %err +cont: + ret void +err: + %exn = landingpad { i8*, i32 } + catch i8* null + resume { i8*, i32 } %exn +} + attributes #0 = { "target-features"="+retpoline-indirect-branches,+retpoline-indirect-calls" } !llvm.module.flags = !{!0} diff --git a/llvm/test/CodeGen/X86/ldexp-f80.ll b/llvm/test/CodeGen/X86/ldexp-f80.ll index 3a10eab2f47c..84d18d5888cf 100644 --- a/llvm/test/CodeGen/X86/ldexp-f80.ll +++ b/llvm/test/CodeGen/X86/ldexp-f80.ll @@ -3,33 +3,29 @@ ; FIXME: Expansion without libcall ; XUN: llc -mtriple=i386-pc-win32 < %s | FileCheck -check-prefix=WIN32 %s -define x86_fp80 @ldexp_f80(x86_fp80 %arg0, i32 %arg1) { +define x86_fp80 @ldexp_f80(x86_fp80 %arg0, i32 %arg1) nounwind { ; X64-LABEL: ldexp_f80: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: callq ldexpl@PLT ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %ldexp = call x86_fp80 @llvm.ldexp.f80.i32(x86_fp80 %arg0, i32 %arg1) ret x86_fp80 %ldexp } -define x86_fp80 @test_strict_ldexp_f80_i32(ptr addrspace(1) %out, x86_fp80 %a, i32 %b) #2 { +define x86_fp80 @test_strict_ldexp_f80_i32(ptr addrspace(1) %out, x86_fp80 %a, i32 %b) nounwind #2 { ; X64-LABEL: test_strict_ldexp_f80_i32: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: movl %esi, %edi ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: wait ; X64-NEXT: callq ldexpl@PLT ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call x86_fp80 @llvm.experimental.constrained.ldexp.f80.i32(x86_fp80 %a, i32 %b, metadata !"round.dynamic", metadata !"fpexcept.strict") ret x86_fp80 %result diff --git a/llvm/test/CodeGen/X86/ldexp-libcall.ll b/llvm/test/CodeGen/X86/ldexp-libcall.ll index 3aec4960383f..74256c801d02 100644 --- a/llvm/test/CodeGen/X86/ldexp-libcall.ll +++ b/llvm/test/CodeGen/X86/ldexp-libcall.ll @@ -2,7 +2,7 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s ; RUN: llc < %s -mtriple=i386-pc-win32 | FileCheck %s -check-prefix=CHECK-WIN -define float @call_ldexpf(float %a, i32 %b) { +define float @call_ldexpf(float %a, i32 %b) nounwind { ; CHECK-LABEL: call_ldexpf: ; CHECK: # %bb.0: ; CHECK-NEXT: jmp ldexpf@PLT # TAILCALL @@ -21,7 +21,7 @@ define float @call_ldexpf(float %a, i32 %b) { ret float %result } -define double @call_ldexp(double %a, i32 %b) { +define double @call_ldexp(double %a, i32 %b) nounwind { ; CHECK-LABEL: call_ldexp: ; CHECK: # %bb.0: ; CHECK-NEXT: jmp ldexp@PLT # TAILCALL @@ -40,16 +40,14 @@ define double @call_ldexp(double %a, i32 %b) { ret double %result } -define x86_fp80 @call_ldexpl(x86_fp80 %a, i32 %b) { +define x86_fp80 @call_ldexpl(x86_fp80 %a, i32 %b) nounwind { ; CHECK-LABEL: call_ldexpl: ; CHECK: # %bb.0: ; CHECK-NEXT: subq $24, %rsp -; CHECK-NEXT: .cfi_def_cfa_offset 32 ; CHECK-NEXT: fldt {{[0-9]+}}(%rsp) ; CHECK-NEXT: fstpt (%rsp) ; CHECK-NEXT: callq ldexpl@PLT ; CHECK-NEXT: addq $24, %rsp -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_ldexpl: diff --git a/llvm/test/CodeGen/X86/ldexp-not-readonly.ll b/llvm/test/CodeGen/X86/ldexp-not-readonly.ll index aec777375925..9a67cf8b3171 100644 --- a/llvm/test/CodeGen/X86/ldexp-not-readonly.ll +++ b/llvm/test/CodeGen/X86/ldexp-not-readonly.ll @@ -2,14 +2,12 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s ; RUN: llc < %s -mtriple=i386-pc-win32 | FileCheck %s -check-prefix=CHECK-WIN -define float @call_ldexpf(float %a, i32 %b) { +define float @call_ldexpf(float %a, i32 %b) nounwind { ; CHECK-LABEL: call_ldexpf: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexpf@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_ldexpf: @@ -26,14 +24,12 @@ define float @call_ldexpf(float %a, i32 %b) { ret float %result } -define double @call_ldexp(double %a, i32 %b) { +define double @call_ldexp(double %a, i32 %b) nounwind { ; CHECK-LABEL: call_ldexp: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexp@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: call_ldexp: diff --git a/llvm/test/CodeGen/X86/ldexp-strict.ll b/llvm/test/CodeGen/X86/ldexp-strict.ll index 67e348cb95ff..f13c59da46c2 100644 --- a/llvm/test/CodeGen/X86/ldexp-strict.ll +++ b/llvm/test/CodeGen/X86/ldexp-strict.ll @@ -4,45 +4,40 @@ ; FIXME: Expansion support without libcalls ; FIXME: Implement f16->f32 promotion for strictfp -; define half @test_strict_ldexp_f16_i32(ptr addrspace(1) %out, half %a, i32 %b) #2 { +; define half @test_strict_ldexp_f16_i32(ptr addrspace(1) %out, half %a, i32 %b) nounwind #2 { ; %result = call half @llvm.experimental.constrained.ldexp.f16.i32(half %a, i32 %b, metadata !"round.dynamic", metadata !"fpexcept.strict") ; ret half %result ; } -define float @test_strict_ldexp_f32_i32(ptr addrspace(1) %out, float %a, i32 %b) #2 { +define float @test_strict_ldexp_f32_i32(ptr addrspace(1) %out, float %a, i32 %b) nounwind #2 { ; X64-LABEL: test_strict_ldexp_f32_i32: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: movl %esi, %edi ; X64-NEXT: callq ldexpf@PLT ; X64-NEXT: popq %rax -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call float @llvm.experimental.constrained.ldexp.f32.i32(float %a, i32 %b, metadata !"round.dynamic", metadata !"fpexcept.strict") ret float %result } -define double @test_strict_ldexp_f64_i32(ptr addrspace(1) %out, double %a, i32 %b) #2 { +define double @test_strict_ldexp_f64_i32(ptr addrspace(1) %out, double %a, i32 %b) nounwind #2 { ; X64-LABEL: test_strict_ldexp_f64_i32: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: movl %esi, %edi ; X64-NEXT: callq ldexp@PLT ; X64-NEXT: popq %rax -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call double @llvm.experimental.constrained.ldexp.f64.i32(double %a, i32 %b, metadata !"round.dynamic", metadata !"fpexcept.strict") ret double %result } -define <2 x float> @test_strict_ldexp_v2f32_v2i32(ptr addrspace(1) %out, <2 x float> %a, <2 x i32> %b) #2 { +define <2 x float> @test_strict_ldexp_v2f32_v2i32(ptr addrspace(1) %out, <2 x float> %a, <2 x i32> %b) nounwind #2 { ; X64-LABEL: test_strict_ldexp_v2f32_v2i32: ; X64: # %bb.0: ; X64-NEXT: subq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 64 ; X64-NEXT: movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: movd %xmm1, %edi @@ -58,7 +53,6 @@ define <2 x float> @test_strict_ldexp_v2f32_v2i32(ptr addrspace(1) %out, <2 x fl ; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] ; X64-NEXT: movaps %xmm1, %xmm0 ; X64-NEXT: addq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call <2 x float> @llvm.experimental.constrained.ldexp.v2f32.v2i32(<2 x float> %a, <2 x i32> %b, metadata !"round.dynamic", metadata !"fpexcept.strict") ret <2 x float> %result diff --git a/llvm/test/CodeGen/X86/ldexp-wrong-signature.ll b/llvm/test/CodeGen/X86/ldexp-wrong-signature.ll index ac58bb53b174..b4ba53f80ad5 100644 --- a/llvm/test/CodeGen/X86/ldexp-wrong-signature.ll +++ b/llvm/test/CodeGen/X86/ldexp-wrong-signature.ll @@ -2,14 +2,12 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s ; RUN: llc < %s -mtriple=i386-pc-win32 | FileCheck %s -check-prefix=CHECK-WIN -define float @ldexpf_too_many_args(float %a, i32 %b, i32 %c) { +define float @ldexpf_too_many_args(float %a, i32 %b, i32 %c) nounwind { ; CHECK-LABEL: ldexpf_too_many_args: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexpf@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: ldexpf_too_many_args: @@ -28,14 +26,12 @@ define float @ldexpf_too_many_args(float %a, i32 %b, i32 %c) { ret float %result } -define float @ldexp_wrong_fp_type(float %a, i32 %b) { +define float @ldexp_wrong_fp_type(float %a, i32 %b) nounwind { ; CHECK-LABEL: ldexp_wrong_fp_type: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexp@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: ldexp_wrong_fp_type: diff --git a/llvm/test/CodeGen/X86/ldexp-wrong-signature2.ll b/llvm/test/CodeGen/X86/ldexp-wrong-signature2.ll index ac79973106ac..d48c1c2e0a9a 100644 --- a/llvm/test/CodeGen/X86/ldexp-wrong-signature2.ll +++ b/llvm/test/CodeGen/X86/ldexp-wrong-signature2.ll @@ -2,14 +2,12 @@ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s ; RUN: llc < %s -mtriple=i386-pc-win32 | FileCheck %s -check-prefix=CHECK-WIN -define i32 @ldexpf_not_fp(i32 %a, i32 %b) { +define i32 @ldexpf_not_fp(i32 %a, i32 %b) nounwind { ; CHECK-LABEL: ldexpf_not_fp: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexpf@PLT ; CHECK-NEXT: popq %rcx -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: ldexpf_not_fp: @@ -23,14 +21,12 @@ define i32 @ldexpf_not_fp(i32 %a, i32 %b) { ret i32 %result } -define float @ldexp_not_int(float %a, float %b) { +define float @ldexp_not_int(float %a, float %b) nounwind { ; CHECK-LABEL: ldexp_not_int: ; CHECK: # %bb.0: ; CHECK-NEXT: pushq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: callq ldexp@PLT ; CHECK-NEXT: popq %rax -; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: retq ; ; CHECK-WIN-LABEL: ldexp_not_int: diff --git a/llvm/test/CodeGen/X86/ldexp.ll b/llvm/test/CodeGen/X86/ldexp.ll index 859139463b7e..59ec7bfcaa91 100644 --- a/llvm/test/CodeGen/X86/ldexp.ll +++ b/llvm/test/CodeGen/X86/ldexp.ll @@ -3,7 +3,7 @@ ; RUN: llc -mtriple=x86_64-pc-win32 -verify-machineinstrs < %s | FileCheck -check-prefixes=WIN64 %s ; RUN: llc -mtriple=i386-pc-win32 -verify-machineinstrs < %s | FileCheck -check-prefix=WIN32 %s -define float @ldexp_f32(i8 zeroext %x) { +define float @ldexp_f32(i8 zeroext %x) nounwind { ; X64-LABEL: ldexp_f32: ; X64: # %bb.0: ; X64-NEXT: movss {{.*#+}} xmm0 = [1.0E+0,0.0E+0,0.0E+0,0.0E+0] @@ -12,17 +12,12 @@ define float @ldexp_f32(i8 zeroext %x) { ; WIN64-LABEL: ldexp_f32: ; WIN64: # %bb.0: ; WIN64-NEXT: subq $40, %rsp -; WIN64-NEXT: .seh_stackalloc 40 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movzbl %cl, %edx ; WIN64-NEXT: movsd {{.*#+}} xmm0 = [1.0E+0,0.0E+0] ; WIN64-NEXT: callq ldexp ; WIN64-NEXT: cvtsd2ss %xmm0, %xmm0 -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $40, %rsp -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_f32: ; WIN32: # %bb.0: @@ -41,7 +36,7 @@ define float @ldexp_f32(i8 zeroext %x) { ret float %ldexp } -define double @ldexp_f64(i8 zeroext %x) { +define double @ldexp_f64(i8 zeroext %x) nounwind { ; X64-LABEL: ldexp_f64: ; X64: # %bb.0: ; X64-NEXT: movsd {{.*#+}} xmm0 = [1.0E+0,0.0E+0] @@ -68,11 +63,10 @@ define double @ldexp_f64(i8 zeroext %x) { ret double %ldexp } -define <2 x float> @ldexp_v2f32(<2 x float> %val, <2 x i32> %exp) { +define <2 x float> @ldexp_v2f32(<2 x float> %val, <2 x i32> %exp) nounwind { ; X64-LABEL: ldexp_v2f32: ; X64: # %bb.0: ; X64-NEXT: subq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 64 ; X64-NEXT: movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: movd %xmm1, %edi @@ -88,22 +82,15 @@ define <2 x float> @ldexp_v2f32(<2 x float> %val, <2 x i32> %exp) { ; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] ; X64-NEXT: movaps %xmm1, %xmm0 ; X64-NEXT: addq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN64-LABEL: ldexp_v2f32: ; WIN64: # %bb.0: ; WIN64-NEXT: pushq %rsi -; WIN64-NEXT: .seh_pushreg %rsi ; WIN64-NEXT: subq $80, %rsp -; WIN64-NEXT: .seh_stackalloc 80 ; WIN64-NEXT: movaps %xmm8, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm8, 64 ; WIN64-NEXT: movaps %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm7, 48 ; WIN64-NEXT: movaps %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm6, 32 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movq %rdx, %rsi ; WIN64-NEXT: movaps (%rcx), %xmm7 ; WIN64-NEXT: movl 12(%rdx), %edx @@ -140,12 +127,9 @@ define <2 x float> @ldexp_v2f32(<2 x float> %val, <2 x i32> %exp) { ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm7 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm8 # 16-byte Reload -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $80, %rsp ; WIN64-NEXT: popq %rsi -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_v2f32: ; WIN32: # %bb.0: @@ -172,11 +156,10 @@ define <2 x float> @ldexp_v2f32(<2 x float> %val, <2 x i32> %exp) { ret <2 x float> %1 } -define <4 x float> @ldexp_v4f32(<4 x float> %val, <4 x i32> %exp) { +define <4 x float> @ldexp_v4f32(<4 x float> %val, <4 x i32> %exp) nounwind { ; X64-LABEL: ldexp_v4f32: ; X64: # %bb.0: ; X64-NEXT: subq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 80 ; X64-NEXT: movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] @@ -210,22 +193,15 @@ define <4 x float> @ldexp_v4f32(<4 x float> %val, <4 x i32> %exp) { ; X64-NEXT: # xmm1 = xmm1[0],mem[0] ; X64-NEXT: movaps %xmm1, %xmm0 ; X64-NEXT: addq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN64-LABEL: ldexp_v4f32: ; WIN64: # %bb.0: ; WIN64-NEXT: pushq %rsi -; WIN64-NEXT: .seh_pushreg %rsi ; WIN64-NEXT: subq $80, %rsp -; WIN64-NEXT: .seh_stackalloc 80 ; WIN64-NEXT: movaps %xmm8, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm8, 64 ; WIN64-NEXT: movaps %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm7, 48 ; WIN64-NEXT: movaps %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm6, 32 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movq %rdx, %rsi ; WIN64-NEXT: movaps (%rcx), %xmm7 ; WIN64-NEXT: movl 12(%rdx), %edx @@ -262,12 +238,9 @@ define <4 x float> @ldexp_v4f32(<4 x float> %val, <4 x i32> %exp) { ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm7 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm8 # 16-byte Reload -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $80, %rsp ; WIN64-NEXT: popq %rsi -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_v4f32: ; WIN32: # %bb.0: @@ -324,11 +297,10 @@ define <4 x float> @ldexp_v4f32(<4 x float> %val, <4 x i32> %exp) { ret <4 x float> %1 } -define <2 x double> @ldexp_v2f64(<2 x double> %val, <2 x i32> %exp) { +define <2 x double> @ldexp_v2f64(<2 x double> %val, <2 x i32> %exp) nounwind { ; X64-LABEL: ldexp_v2f64: ; X64: # %bb.0: ; X64-NEXT: subq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 64 ; X64-NEXT: movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: movd %xmm1, %edi @@ -344,20 +316,14 @@ define <2 x double> @ldexp_v2f64(<2 x double> %val, <2 x i32> %exp) { ; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] ; X64-NEXT: movaps %xmm1, %xmm0 ; X64-NEXT: addq $56, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN64-LABEL: ldexp_v2f64: ; WIN64: # %bb.0: ; WIN64-NEXT: pushq %rsi -; WIN64-NEXT: .seh_pushreg %rsi ; WIN64-NEXT: subq $64, %rsp -; WIN64-NEXT: .seh_stackalloc 64 ; WIN64-NEXT: movaps %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm7, 48 ; WIN64-NEXT: movaps %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm6, 32 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movaps (%rcx), %xmm6 ; WIN64-NEXT: movl (%rdx), %eax ; WIN64-NEXT: movl 4(%rdx), %esi @@ -373,12 +339,9 @@ define <2 x double> @ldexp_v2f64(<2 x double> %val, <2 x i32> %exp) { ; WIN64-NEXT: movaps %xmm7, %xmm0 ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm7 # 16-byte Reload -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $64, %rsp ; WIN64-NEXT: popq %rsi -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_v2f64: ; WIN32: # %bb.0: @@ -406,17 +369,12 @@ define <2 x double> @ldexp_v2f64(<2 x double> %val, <2 x i32> %exp) { ret <2 x double> %1 } -define <4 x double> @ldexp_v4f64(<4 x double> %val, <4 x i32> %exp) { +define <4 x double> @ldexp_v4f64(<4 x double> %val, <4 x i32> %exp) nounwind { ; X64-LABEL: ldexp_v4f64: ; X64: # %bb.0: ; X64-NEXT: pushq %rbp -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: pushq %rbx -; X64-NEXT: .cfi_def_cfa_offset 24 ; X64-NEXT: subq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 96 -; X64-NEXT: .cfi_offset %rbx, -24 -; X64-NEXT: .cfi_offset %rbp, -16 ; X64-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: movdqa %xmm2, (%rsp) # 16-byte Spill @@ -449,30 +407,19 @@ define <4 x double> @ldexp_v4f64(<4 x double> %val, <4 x i32> %exp) { ; X64-NEXT: # xmm1 = xmm1[0],mem[0] ; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload ; X64-NEXT: addq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 24 ; X64-NEXT: popq %rbx -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: popq %rbp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN64-LABEL: ldexp_v4f64: ; WIN64: # %bb.0: ; WIN64-NEXT: pushq %rsi -; WIN64-NEXT: .seh_pushreg %rsi ; WIN64-NEXT: pushq %rdi -; WIN64-NEXT: .seh_pushreg %rdi ; WIN64-NEXT: pushq %rbx -; WIN64-NEXT: .seh_pushreg %rbx ; WIN64-NEXT: subq $80, %rsp -; WIN64-NEXT: .seh_stackalloc 80 ; WIN64-NEXT: movaps %xmm8, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm8, 64 ; WIN64-NEXT: movaps %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm7, 48 ; WIN64-NEXT: movaps %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill -; WIN64-NEXT: .seh_savexmm %xmm6, 32 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movl 12(%r8), %esi ; WIN64-NEXT: movl 8(%r8), %edi ; WIN64-NEXT: movaps (%rdx), %xmm6 @@ -501,14 +448,11 @@ define <4 x double> @ldexp_v4f64(<4 x double> %val, <4 x i32> %exp) { ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm7 # 16-byte Reload ; WIN64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm8 # 16-byte Reload -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $80, %rsp ; WIN64-NEXT: popq %rbx ; WIN64-NEXT: popq %rdi ; WIN64-NEXT: popq %rsi -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_v4f64: ; WIN32: # %bb.0: @@ -565,41 +509,31 @@ define <4 x double> @ldexp_v4f64(<4 x double> %val, <4 x i32> %exp) { ret <4 x double> %1 } -define half @ldexp_f16(half %arg0, i32 %arg1) { +define half @ldexp_f16(half %arg0, i32 %arg1) nounwind { ; X64-LABEL: ldexp_f16: ; X64: # %bb.0: ; X64-NEXT: pushq %rbx -; X64-NEXT: .cfi_def_cfa_offset 16 -; X64-NEXT: .cfi_offset %rbx, -16 ; X64-NEXT: movl %edi, %ebx ; X64-NEXT: callq __extendhfsf2@PLT ; X64-NEXT: movl %ebx, %edi ; X64-NEXT: callq ldexpf@PLT ; X64-NEXT: callq __truncsfhf2@PLT ; X64-NEXT: popq %rbx -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN64-LABEL: ldexp_f16: ; WIN64: # %bb.0: ; WIN64-NEXT: pushq %rsi -; WIN64-NEXT: .seh_pushreg %rsi ; WIN64-NEXT: subq $32, %rsp -; WIN64-NEXT: .seh_stackalloc 32 -; WIN64-NEXT: .seh_endprologue ; WIN64-NEXT: movl %edx, %esi ; WIN64-NEXT: callq __extendhfsf2 ; WIN64-NEXT: cvtss2sd %xmm0, %xmm0 ; WIN64-NEXT: movl %esi, %edx ; WIN64-NEXT: callq ldexp ; WIN64-NEXT: callq __truncdfhf2 -; WIN64-NEXT: nop -; WIN64-NEXT: .seh_startepilogue ; WIN64-NEXT: addq $32, %rsp ; WIN64-NEXT: popq %rsi -; WIN64-NEXT: .seh_endepilogue ; WIN64-NEXT: retq -; WIN64-NEXT: .seh_endproc ; ; WIN32-LABEL: ldexp_f16: ; WIN32: # %bb.0: diff --git a/llvm/test/CodeGen/X86/lea-2.ll b/llvm/test/CodeGen/X86/lea-2.ll index a48c02ff3e0b..0883a8e726e2 100644 --- a/llvm/test/CodeGen/X86/lea-2.ll +++ b/llvm/test/CodeGen/X86/lea-2.ll @@ -2,7 +2,6 @@ ; RUN: llc < %s -mtriple=i686-linux | FileCheck %s --check-prefix=X86 ; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s --check-prefix=X64 ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 | FileCheck %s --check-prefix=X64 -; RUN: llc < %s -mtriple=x86_64-nacl | FileCheck %s --check-prefix=X64 ; The computation of %t4 should match a single lea, without using actual add instructions. diff --git a/llvm/test/CodeGen/X86/lea-3.ll b/llvm/test/CodeGen/X86/lea-3.ll index b7f1c4ae1154..2cbefc0689c1 100644 --- a/llvm/test/CodeGen/X86/lea-3.ll +++ b/llvm/test/CodeGen/X86/lea-3.ll @@ -1,7 +1,6 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-nacl | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-win32 | FileCheck %s --check-prefix=WIN32 define i64 @test2(i64 %a) { diff --git a/llvm/test/CodeGen/X86/lea-4.ll b/llvm/test/CodeGen/X86/lea-4.ll index e1f0b73c33ff..c33697e0abf3 100644 --- a/llvm/test/CodeGen/X86/lea-4.ll +++ b/llvm/test/CodeGen/X86/lea-4.ll @@ -1,7 +1,6 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 | FileCheck %s -; RUN: llc < %s -mtriple=x86_64-nacl | FileCheck %s define zeroext i16 @t1(i32 %on_off) nounwind { ; CHECK-LABEL: t1: diff --git a/llvm/test/CodeGen/X86/lea-5.ll b/llvm/test/CodeGen/X86/lea-5.ll index 908ec3eae7f6..39051eac45d7 100644 --- a/llvm/test/CodeGen/X86/lea-5.ll +++ b/llvm/test/CodeGen/X86/lea-5.ll @@ -4,7 +4,6 @@ ; RUN: llc < %s -mtriple=x86_64-linux -O2 | FileCheck %s ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 -O2 | FileCheck %s -check-prefix=X32 -; RUN: llc < %s -mtriple=x86_64-nacl -O2 | FileCheck %s -check-prefix=X32 ; Function Attrs: nounwind readnone uwtable define void @foo(i32 %x, i32 %d) #0 { diff --git a/llvm/test/CodeGen/X86/lea.ll b/llvm/test/CodeGen/X86/lea.ll index 33d121f6849b..28c66b94a69e 100644 --- a/llvm/test/CodeGen/X86/lea.ll +++ b/llvm/test/CodeGen/X86/lea.ll @@ -2,7 +2,6 @@ ; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s --check-prefixes=LINUX ; RUN: llc < %s -mtriple=x86_64-win32 | FileCheck %s --check-prefixes=WIN ; RUN: llc < %s -mtriple=x86_64-linux-gnux32 | FileCheck %s --check-prefixes=LINUX -; RUN: llc < %s -mtriple=x86_64-nacl | FileCheck %s --check-prefixes=LINUX define i32 @test1(i32 %x) nounwind { ; LINUX-LABEL: test1: diff --git a/llvm/test/CodeGen/X86/llvm.frexp.f80.ll b/llvm/test/CodeGen/X86/llvm.frexp.f80.ll index 92d8c53c8182..61e2bba82201 100644 --- a/llvm/test/CodeGen/X86/llvm.frexp.f80.ll +++ b/llvm/test/CodeGen/X86/llvm.frexp.f80.ll @@ -3,52 +3,46 @@ ; RUN: llc -mtriple=x86_64-unknown-unknown < %s | FileCheck -check-prefixes=X64 %s ; XUN: llc -mtriple=i386-pc-win32 < %s | FileCheck -check-prefix=X64 %s -define { x86_fp80, i32 } @test_frexp_f80_i32(x86_fp80 %a) { +define { x86_fp80, i32 } @test_frexp_f80_i32(x86_fp80 %a) nounwind { ; X64-LABEL: test_frexp_f80_i32: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpl@PLT ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call { x86_fp80, i32 } @llvm.frexp.f80.i32(x86_fp80 %a) ret { x86_fp80, i32 } %result } -define x86_fp80 @test_frexp_f80_i32_only_use_fract(x86_fp80 %a) { +define x86_fp80 @test_frexp_f80_i32_only_use_fract(x86_fp80 %a) nounwind { ; X64-LABEL: test_frexp_f80_i32_only_use_fract: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpl@PLT ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call { x86_fp80, i32 } @llvm.frexp.f80.i32(x86_fp80 %a) %result.0 = extractvalue { x86_fp80, i32 } %result, 0 ret x86_fp80 %result.0 } -define x86_fp80 @test_frexp_f80_i32_only_use_fract_math(x86_fp80 %a) { +define x86_fp80 @test_frexp_f80_i32_only_use_fract_math(x86_fp80 %a) nounwind { ; X64-LABEL: test_frexp_f80_i32_only_use_fract_math: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpl@PLT ; X64-NEXT: fadd %st, %st(0) ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call { x86_fp80, i32 } @llvm.frexp.f80.i32(x86_fp80 %a) %result.0 = extractvalue { x86_fp80, i32 } %result, 0 @@ -56,11 +50,10 @@ define x86_fp80 @test_frexp_f80_i32_only_use_fract_math(x86_fp80 %a) { ret x86_fp80 %add } -define i32 @test_frexp_f80_i32_only_use_exp(x86_fp80 %a) { +define i32 @test_frexp_f80_i32_only_use_exp(x86_fp80 %a) nounwind { ; X64-LABEL: test_frexp_f80_i32_only_use_exp: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: fldt {{[0-9]+}}(%rsp) ; X64-NEXT: fstpt (%rsp) ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi @@ -68,7 +61,6 @@ define i32 @test_frexp_f80_i32_only_use_exp(x86_fp80 %a) { ; X64-NEXT: fstp %st(0) ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq %result = call { x86_fp80, i32 } @llvm.frexp.f80.i32(x86_fp80 %a) %result.0 = extractvalue { x86_fp80, i32 } %result, 1 diff --git a/llvm/test/CodeGen/X86/llvm.frexp.ll b/llvm/test/CodeGen/X86/llvm.frexp.ll index 8436c1052552..83840dd85c53 100644 --- a/llvm/test/CodeGen/X86/llvm.frexp.ll +++ b/llvm/test/CodeGen/X86/llvm.frexp.ll @@ -2,11 +2,10 @@ ; RUN: llc -mtriple=x86_64-unknown-unknown < %s | FileCheck -check-prefixes=X64 %s ; RUN: llc -mtriple=i386-pc-win32 < %s | FileCheck -check-prefix=WIN32 %s -define { half, i32 } @test_frexp_f16_i32(half %a) { +define { half, i32 } @test_frexp_f16_i32(half %a) nounwind { ; X64-LABEL: test_frexp_f16_i32: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: callq __extendhfsf2@PLT ; X64-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 @@ -36,7 +35,6 @@ define { half, i32 } @test_frexp_f16_i32(half %a) { ; X64-NEXT: cmovbel %edx, %ecx ; X64-NEXT: pinsrw $0, %ecx, %xmm0 ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f16_i32: @@ -63,11 +61,10 @@ define { half, i32 } @test_frexp_f16_i32(half %a) { ret { half, i32 } %result } -define half @test_frexp_f16_i32_only_use_fract(half %a) { +define half @test_frexp_f16_i32_only_use_fract(half %a) nounwind { ; X64-LABEL: test_frexp_f16_i32_only_use_fract: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: callq __extendhfsf2@PLT ; X64-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 @@ -87,7 +84,6 @@ define half @test_frexp_f16_i32_only_use_fract(half %a) { ; X64-NEXT: cmovbel %ecx, %eax ; X64-NEXT: pinsrw $0, %eax, %xmm0 ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f16_i32_only_use_fract: @@ -111,11 +107,10 @@ define half @test_frexp_f16_i32_only_use_fract(half %a) { ret half %result.0 } -define i32 @test_frexp_f16_i32_only_use_exp(half %a) { +define i32 @test_frexp_f16_i32_only_use_exp(half %a) nounwind { ; X64-LABEL: test_frexp_f16_i32_only_use_exp: ; X64: # %bb.0: ; X64-NEXT: subq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 32 ; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill ; X64-NEXT: callq __extendhfsf2@PLT ; X64-NEXT: mulss {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0 @@ -138,7 +133,6 @@ define i32 @test_frexp_f16_i32_only_use_exp(half %a) { ; X64-NEXT: cmpl $33792, %ecx # imm = 0x8400 ; X64-NEXT: cmoval %edx, %eax ; X64-NEXT: addq $24, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f16_i32_only_use_exp: @@ -161,33 +155,31 @@ define i32 @test_frexp_f16_i32_only_use_exp(half %a) { } ; FIXME -; define { <2 x half>, <2 x i32> } @test_frexp_v2f16_v2i32(<2 x half> %a) { +; define { <2 x half>, <2 x i32> } @test_frexp_v2f16_v2i32(<2 x half> %a) nounwind { ; %result = call { <2 x half>, <2 x i32> } @llvm.frexp.v2f16.v2i32(<2 x half> %a) ; ret { <2 x half>, <2 x i32> } %result ; } -; define <2 x half> @test_frexp_v2f16_v2i32_only_use_fract(<2 x half> %a) { +; define <2 x half> @test_frexp_v2f16_v2i32_only_use_fract(<2 x half> %a) nounwind { ; %result = call { <2 x half>, <2 x i32> } @llvm.frexp.v2f16.v2i32(<2 x half> %a) ; %result.0 = extractvalue { <2 x half>, <2 x i32> } %result, 0 ; ret <2 x half> %result.0 ; } -; define <2 x i32> @test_frexp_v2f16_v2i32_only_use_exp(<2 x half> %a) { +; define <2 x i32> @test_frexp_v2f16_v2i32_only_use_exp(<2 x half> %a) nounwind { ; %result = call { <2 x half>, <2 x i32> } @llvm.frexp.v2f16.v2i32(<2 x half> %a) ; %result.1 = extractvalue { <2 x half>, <2 x i32> } %result, 1 ; ret <2 x i32> %result.1 ; } -define { float, i32 } @test_frexp_f32_i32(float %a) { +define { float, i32 } @test_frexp_f32_i32(float %a) nounwind { ; X64-LABEL: test_frexp_f32_i32: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpf@PLT ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: popq %rcx -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f32_i32: @@ -207,15 +199,13 @@ define { float, i32 } @test_frexp_f32_i32(float %a) { ret { float, i32 } %result } -define float @test_frexp_f32_i32_only_use_fract(float %a) { +define float @test_frexp_f32_i32_only_use_fract(float %a) nounwind { ; X64-LABEL: test_frexp_f32_i32_only_use_fract: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpf@PLT ; X64-NEXT: popq %rax -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f32_i32_only_use_fract: @@ -235,16 +225,14 @@ define float @test_frexp_f32_i32_only_use_fract(float %a) { ret float %result.0 } -define i32 @test_frexp_f32_i32_only_use_exp(float %a) { +define i32 @test_frexp_f32_i32_only_use_exp(float %a) nounwind { ; X64-LABEL: test_frexp_f32_i32_only_use_exp: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexpf@PLT ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: popq %rcx -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f32_i32_only_use_exp: @@ -265,28 +253,27 @@ define i32 @test_frexp_f32_i32_only_use_exp(float %a) { } ; FIXME: Widen vector result -; define { <2 x float>, <2 x i32> } @test_frexp_v2f32_v2i32(<2 x float> %a) { +; define { <2 x float>, <2 x i32> } @test_frexp_v2f32_v2i32(<2 x float> %a) nounwind { ; %result = call { <2 x float>, <2 x i32> } @llvm.frexp.v2f32.v2i32(<2 x float> %a) ; ret { <2 x float>, <2 x i32> } %result ; } -; define <2 x float> @test_frexp_v2f32_v2i32_only_use_fract(<2 x float> %a) { +; define <2 x float> @test_frexp_v2f32_v2i32_only_use_fract(<2 x float> %a) nounwind { ; %result = call { <2 x float>, <2 x i32> } @llvm.frexp.v2f32.v2i32(<2 x float> %a) ; %result.0 = extractvalue { <2 x float>, <2 x i32> } %result, 0 ; ret <2 x float> %result.0 ; } -; define <2 x i32> @test_frexp_v2f32_v2i32_only_use_exp(<2 x float> %a) { +; define <2 x i32> @test_frexp_v2f32_v2i32_only_use_exp(<2 x float> %a) nounwind { ; %result = call { <2 x float>, <2 x i32> } @llvm.frexp.v2f32.v2i32(<2 x float> %a) ; %result.1 = extractvalue { <2 x float>, <2 x i32> } %result, 1 ; ret <2 x i32> %result.1 ; } -define { <4 x float>, <4 x i32> } @test_frexp_v4f32_v4i32(<4 x float> %a) { +define { <4 x float>, <4 x i32> } @test_frexp_v4f32_v4i32(<4 x float> %a) nounwind { ; X64-LABEL: test_frexp_v4f32_v4i32: ; X64: # %bb.0: ; X64-NEXT: subq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 80 ; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] ; X64-NEXT: movq %rsp, %rdi @@ -320,7 +307,6 @@ define { <4 x float>, <4 x i32> } @test_frexp_v4f32_v4i32(<4 x float> %a) { ; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1] ; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0] ; X64-NEXT: addq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_v4f32_v4i32: @@ -372,11 +358,10 @@ define { <4 x float>, <4 x i32> } @test_frexp_v4f32_v4i32(<4 x float> %a) { ret { <4 x float>, <4 x i32> } %result } -define <4 x float> @test_frexp_v4f32_v4i32_only_use_fract(<4 x float> %a) { +define <4 x float> @test_frexp_v4f32_v4i32_only_use_fract(<4 x float> %a) nounwind { ; X64-LABEL: test_frexp_v4f32_v4i32_only_use_fract: ; X64: # %bb.0: ; X64-NEXT: subq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 80 ; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi @@ -403,7 +388,6 @@ define <4 x float> @test_frexp_v4f32_v4i32_only_use_fract(<4 x float> %a) { ; X64-NEXT: # xmm1 = xmm1[0],mem[0] ; X64-NEXT: movaps %xmm1, %xmm0 ; X64-NEXT: addq $72, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_v4f32_v4i32_only_use_fract: @@ -456,11 +440,10 @@ define <4 x float> @test_frexp_v4f32_v4i32_only_use_fract(<4 x float> %a) { ret <4 x float> %result.0 } -define <4 x i32> @test_frexp_v4f32_v4i32_only_use_exp(<4 x float> %a) { +define <4 x i32> @test_frexp_v4f32_v4i32_only_use_exp(<4 x float> %a) nounwind { ; X64-LABEL: test_frexp_v4f32_v4i32_only_use_exp: ; X64: # %bb.0: ; X64-NEXT: subq $40, %rsp -; X64-NEXT: .cfi_def_cfa_offset 48 ; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill ; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] ; X64-NEXT: movq %rsp, %rdi @@ -484,7 +467,6 @@ define <4 x i32> @test_frexp_v4f32_v4i32_only_use_exp(<4 x float> %a) { ; X64-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] ; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0] ; X64-NEXT: addq $40, %rsp -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_v4f32_v4i32_only_use_exp: @@ -524,16 +506,14 @@ define <4 x i32> @test_frexp_v4f32_v4i32_only_use_exp(<4 x float> %a) { ret <4 x i32> %result.1 } -define { double, i32 } @test_frexp_f64_i32(double %a) { +define { double, i32 } @test_frexp_f64_i32(double %a) nounwind { ; X64-LABEL: test_frexp_f64_i32: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexp@PLT ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: popq %rcx -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f64_i32: @@ -551,15 +531,13 @@ define { double, i32 } @test_frexp_f64_i32(double %a) { ret { double, i32 } %result } -define double @test_frexp_f64_i32_only_use_fract(double %a) { +define double @test_frexp_f64_i32_only_use_fract(double %a) nounwind { ; X64-LABEL: test_frexp_f64_i32_only_use_fract: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexp@PLT ; X64-NEXT: popq %rax -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f64_i32_only_use_fract: @@ -577,16 +555,14 @@ define double @test_frexp_f64_i32_only_use_fract(double %a) { ret double %result.0 } -define i32 @test_frexp_f64_i32_only_use_exp(double %a) { +define i32 @test_frexp_f64_i32_only_use_exp(double %a) nounwind { ; X64-LABEL: test_frexp_f64_i32_only_use_exp: ; X64: # %bb.0: ; X64-NEXT: pushq %rax -; X64-NEXT: .cfi_def_cfa_offset 16 ; X64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi ; X64-NEXT: callq frexp@PLT ; X64-NEXT: movl {{[0-9]+}}(%rsp), %eax ; X64-NEXT: popq %rcx -; X64-NEXT: .cfi_def_cfa_offset 8 ; X64-NEXT: retq ; ; WIN32-LABEL: test_frexp_f64_i32_only_use_exp: @@ -607,18 +583,18 @@ define i32 @test_frexp_f64_i32_only_use_exp(double %a) { } ; FIXME: Widen vector result -; define { <2 x double>, <2 x i32> } @test_frexp_v2f64_v2i32(<2 x double> %a) { +; define { <2 x double>, <2 x i32> } @test_frexp_v2f64_v2i32(<2 x double> %a) nounwind { ; %result = call { <2 x double>, <2 x i32> } @llvm.frexp.v2f64.v2i32(<2 x double> %a) ; ret { <2 x double>, <2 x i32> } %result ; } -; define <2 x double> @test_frexp_v2f64_v2i32_only_use_fract(<2 x double> %a) { +; define <2 x double> @test_frexp_v2f64_v2i32_only_use_fract(<2 x double> %a) nounwind { ; %result = call { <2 x double>, <2 x i32> } @llvm.frexp.v2f64.v2i32(<2 x double> %a) ; %result.0 = extractvalue { <2 x double>, <2 x i32> } %result, 0 ; ret <2 x double> %result.0 ; } -; define <2 x i32> @test_frexp_v2f64_v2i32_only_use_exp(<2 x double> %a) { +; define <2 x i32> @test_frexp_v2f64_v2i32_only_use_exp(<2 x double> %a) nounwind { ; %result = call { <2 x double>, <2 x i32> } @llvm.frexp.v2f64.v2i32(<2 x double> %a) ; %result.1 = extractvalue { <2 x double>, <2 x i32> } %result, 1 ; ret <2 x i32> %result.1 diff --git a/llvm/test/CodeGen/X86/mul128.ll b/llvm/test/CodeGen/X86/mul128.ll index fc1cc1f65627..e10e48f9aea0 100644 --- a/llvm/test/CodeGen/X86/mul128.ll +++ b/llvm/test/CodeGen/X86/mul128.ll @@ -18,85 +18,80 @@ define i128 @foo(i128 %t, i128 %u) { ; X86: # %bb.0: ; X86-NEXT: pushl %ebp ; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp ; X86-NEXT: pushl %ebx -; X86-NEXT: .cfi_def_cfa_offset 12 ; X86-NEXT: pushl %edi -; X86-NEXT: .cfi_def_cfa_offset 16 ; X86-NEXT: pushl %esi -; X86-NEXT: .cfi_def_cfa_offset 20 -; X86-NEXT: subl $8, %esp -; X86-NEXT: .cfi_def_cfa_offset 28 +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp ; X86-NEXT: .cfi_offset %esi, -20 ; X86-NEXT: .cfi_offset %edi, -16 ; X86-NEXT: .cfi_offset %ebx, -12 -; X86-NEXT: .cfi_offset %ebp, -8 -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: imull %ecx, %ebp -; X86-NEXT: movl %eax, %esi +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl 40(%ebp), %edi +; X86-NEXT: movl 44(%ebp), %esi +; X86-NEXT: imull %ecx, %esi +; X86-NEXT: movl %edi, %eax ; X86-NEXT: mull %ecx -; X86-NEXT: movl %eax, %edi +; X86-NEXT: movl %eax, (%esp) # 4-byte Spill ; X86-NEXT: movl %edx, %ebx -; X86-NEXT: addl %ebp, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: imull %esi, %eax +; X86-NEXT: addl %esi, %ebx +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: imull %edi, %eax ; X86-NEXT: addl %eax, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %eax, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: imull %ecx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: mull %ebp -; X86-NEXT: addl %esi, %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: imull %ebp, %esi +; X86-NEXT: movl 48(%ebp), %eax +; X86-NEXT: movl %eax, %ecx +; X86-NEXT: imull 28(%ebp), %ecx +; X86-NEXT: movl 24(%ebp), %edi +; X86-NEXT: mull %edi +; X86-NEXT: addl %ecx, %edx +; X86-NEXT: movl 52(%ebp), %esi +; X86-NEXT: imull %edi, %esi ; X86-NEXT: addl %edx, %esi -; X86-NEXT: addl %edi, %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: addl (%esp), %eax # 4-byte Folded Reload +; X86-NEXT: movl %eax, (%esp) # 4-byte Spill ; X86-NEXT: adcl %ebx, %esi -; X86-NEXT: movl %ebp, %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: mull %edi +; X86-NEXT: movl %edi, %eax +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: mull %ecx ; X86-NEXT: movl %edx, %ebx -; X86-NEXT: movl %eax, (%esp) # 4-byte Spill -; X86-NEXT: movl %ecx, %eax -; X86-NEXT: mull %edi +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: mull %ecx ; X86-NEXT: movl %edx, %edi ; X86-NEXT: movl %eax, %ecx ; X86-NEXT: addl %ebx, %ecx ; X86-NEXT: adcl $0, %edi -; X86-NEXT: movl %ebp, %eax -; X86-NEXT: mull {{[0-9]+}}(%esp) -; X86-NEXT: movl %edx, %ebp -; X86-NEXT: movl %eax, %ebx -; X86-NEXT: addl %ecx, %ebx -; X86-NEXT: adcl %edi, %ebp +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: mull 44(%ebp) +; X86-NEXT: movl %edx, %ebx +; X86-NEXT: addl %ecx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: adcl %edi, %ebx ; X86-NEXT: setb %cl -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: mull {{[0-9]+}}(%esp) -; X86-NEXT: addl %ebp, %eax +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: mull 44(%ebp) +; X86-NEXT: addl %ebx, %eax ; X86-NEXT: movzbl %cl, %ecx ; X86-NEXT: adcl %ecx, %edx -; X86-NEXT: addl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload +; X86-NEXT: addl (%esp), %eax # 4-byte Folded Reload ; X86-NEXT: adcl %esi, %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ebx, 4(%ecx) -; X86-NEXT: movl (%esp), %esi # 4-byte Reload +; X86-NEXT: movl 8(%ebp), %ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: movl %esi, 4(%ecx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload ; X86-NEXT: movl %esi, (%ecx) ; X86-NEXT: movl %eax, 8(%ecx) ; X86-NEXT: movl %edx, 12(%ecx) ; X86-NEXT: movl %ecx, %eax -; X86-NEXT: addl $8, %esp -; X86-NEXT: .cfi_def_cfa_offset 20 +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi -; X86-NEXT: .cfi_def_cfa_offset 16 ; X86-NEXT: popl %edi -; X86-NEXT: .cfi_def_cfa_offset 12 ; X86-NEXT: popl %ebx -; X86-NEXT: .cfi_def_cfa_offset 8 ; X86-NEXT: popl %ebp -; X86-NEXT: .cfi_def_cfa_offset 4 +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl $4 %k = mul i128 %t, %u ret i128 %k diff --git a/llvm/test/CodeGen/X86/neg-abs.ll b/llvm/test/CodeGen/X86/neg-abs.ll index 961205c50d97..724b2dc4c431 100644 --- a/llvm/test/CodeGen/X86/neg-abs.ll +++ b/llvm/test/CodeGen/X86/neg-abs.ll @@ -105,31 +105,35 @@ define i128 @neg_abs_i128(i128 %x) nounwind { ; X86-LABEL: neg_abs_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edx, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl %eax, %ecx ; X86-NEXT: sarl $31, %ecx +; X86-NEXT: xorl %ecx, %eax +; X86-NEXT: movl 32(%ebp), %edx ; X86-NEXT: xorl %ecx, %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl 28(%ebp), %esi ; X86-NEXT: xorl %ecx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl 24(%ebp), %edi ; X86-NEXT: xorl %ecx, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: xorl %ecx, %ebx -; X86-NEXT: movl %ecx, %ebp -; X86-NEXT: subl %ebx, %ebp ; X86-NEXT: movl %ecx, %ebx -; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: subl %edi, %ebx ; X86-NEXT: movl %ecx, %edi ; X86-NEXT: sbbl %esi, %edi -; X86-NEXT: sbbl %edx, %ecx -; X86-NEXT: movl %ebp, (%eax) -; X86-NEXT: movl %ebx, 4(%eax) -; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: movl %ecx, %esi +; X86-NEXT: sbbl %edx, %esi +; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %esi, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -259,37 +263,42 @@ define i64 @sub_abs_i64(i64 %x, i64 %y) nounwind { define i128 @sub_abs_i128(i128 %x, i128 %y) nounwind { ; X86-LABEL: sub_abs_i128: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 36(%ebp), %eax ; X86-NEXT: movl %eax, %edx ; X86-NEXT: sarl $31, %edx ; X86-NEXT: xorl %edx, %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl 32(%ebp), %ecx ; X86-NEXT: xorl %edx, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl 28(%ebp), %esi ; X86-NEXT: xorl %edx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl 24(%ebp), %edi ; X86-NEXT: xorl %edx, %edi ; X86-NEXT: subl %edx, %edi ; X86-NEXT: sbbl %edx, %esi ; X86-NEXT: sbbl %edx, %ecx ; X86-NEXT: sbbl %edx, %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: movl 40(%ebp), %edx ; X86-NEXT: subl %edi, %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl 44(%ebp), %edi ; X86-NEXT: sbbl %esi, %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl 48(%ebp), %esi ; X86-NEXT: sbbl %ecx, %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: movl 52(%ebp), %ecx ; X86-NEXT: sbbl %eax, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %edx, (%eax) ; X86-NEXT: movl %edi, 4(%eax) ; X86-NEXT: movl %esi, 8(%eax) ; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 ; ; X64-LABEL: sub_abs_i128: diff --git a/llvm/test/CodeGen/X86/pcsections-atomics.ll b/llvm/test/CodeGen/X86/pcsections-atomics.ll index 672ebc1ec727..69ae1f19f320 100644 --- a/llvm/test/CodeGen/X86/pcsections-atomics.ll +++ b/llvm/test/CodeGen/X86/pcsections-atomics.ll @@ -9,6 +9,7 @@ ; RUN: llc -O1 -mattr=cx16 < %s | FileCheck %s --check-prefixes=O1 ; RUN: llc -O2 -mattr=cx16 < %s | FileCheck %s --check-prefixes=O2 ; RUN: llc -O3 -mattr=cx16 < %s | FileCheck %s --check-prefixes=O3 +; RUN: llc -O3 -mcpu=haswell -mattr=cx16 < %s | FileCheck %s --check-prefixes=HASWELL-O3 target triple = "x86_64-unknown-linux-gnu" @@ -50,6 +51,14 @@ define void @mixed_atomic_non_atomic(ptr %a) { ; O3-NEXT: movl $1, (%rdi) ; O3-NEXT: decl (%rdi) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: mixed_atomic_non_atomic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: incl (%rdi) +; HASWELL-O3-NEXT: .Lpcsection0: +; HASWELL-O3-NEXT: movl $1, (%rdi) +; HASWELL-O3-NEXT: decl (%rdi) +; HASWELL-O3-NEXT: retq entry: ; Accesses the same location atomically and non-atomically. %0 = load volatile i32, ptr %a, align 4 @@ -107,6 +116,17 @@ define i64 @mixed_complex_atomic_non_atomic(ptr %a, ptr %b) { ; O3-NEXT: movq %rdx, (%rsi) ; O3-NEXT: addq %rcx, %rax ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: mixed_complex_atomic_non_atomic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movl $1, %eax +; HASWELL-O3-NEXT: .Lpcsection1: +; HASWELL-O3-NEXT: lock xaddq %rax, (%rdi) +; HASWELL-O3-NEXT: movq (%rsi), %rcx +; HASWELL-O3-NEXT: leaq 1(%rcx), %rdx +; HASWELL-O3-NEXT: movq %rdx, (%rsi) +; HASWELL-O3-NEXT: addq %rcx, %rax +; HASWELL-O3-NEXT: retq entry: %0 = atomicrmw add ptr %a, i64 1 monotonic, align 8, !pcsections !0 %1 = load i64, ptr %b, align 8 @@ -148,6 +168,14 @@ define i8 @atomic8_load_unordered(ptr %a) { ; O3-NEXT: movzbl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_load_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection2: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i8, ptr %a unordered, align 1, !pcsections !0 @@ -187,6 +215,14 @@ define i8 @atomic8_load_monotonic(ptr %a) { ; O3-NEXT: movzbl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_load_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection3: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i8, ptr %a monotonic, align 1, !pcsections !0 @@ -226,6 +262,14 @@ define i8 @atomic8_load_acquire(ptr %a) { ; O3-NEXT: movzbl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_load_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection4: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i8, ptr %a acquire, align 1, !pcsections !0 @@ -265,6 +309,14 @@ define i8 @atomic8_load_seq_cst(ptr %a) { ; O3-NEXT: movzbl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_load_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection5: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i8, ptr %a seq_cst, align 1, !pcsections !0 @@ -304,6 +356,14 @@ define void @atomic8_store_unordered(ptr %a) { ; O3-NEXT: movb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_store_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection6: +; HASWELL-O3-NEXT: movb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i8 42, ptr %a unordered, align 1, !pcsections !0 @@ -343,6 +403,14 @@ define void @atomic8_store_monotonic(ptr %a) { ; O3-NEXT: movb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_store_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection7: +; HASWELL-O3-NEXT: movb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i8 42, ptr %a monotonic, align 1, !pcsections !0 @@ -382,6 +450,14 @@ define void @atomic8_store_release(ptr %a) { ; O3-NEXT: movb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_store_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection8: +; HASWELL-O3-NEXT: movb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i8 42, ptr %a release, align 1, !pcsections !0 @@ -425,6 +501,15 @@ define void @atomic8_store_seq_cst(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_store_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection9: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i8 42, ptr %a seq_cst, align 1, !pcsections !0 @@ -468,6 +553,15 @@ define void @atomic8_xchg_monotonic(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xchg_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection10: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -507,6 +601,14 @@ define void @atomic8_add_monotonic(ptr %a) { ; O3-NEXT: lock addb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_add_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection11: +; HASWELL-O3-NEXT: lock addb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -546,6 +648,14 @@ define void @atomic8_sub_monotonic(ptr %a) { ; O3-NEXT: lock subb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_sub_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection12: +; HASWELL-O3-NEXT: lock subb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -585,6 +695,14 @@ define void @atomic8_and_monotonic(ptr %a) { ; O3-NEXT: lock andb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_and_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection13: +; HASWELL-O3-NEXT: lock andb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -624,6 +742,14 @@ define void @atomic8_or_monotonic(ptr %a) { ; O3-NEXT: lock orb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_or_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection14: +; HASWELL-O3-NEXT: lock orb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -663,6 +789,14 @@ define void @atomic8_xor_monotonic(ptr %a) { ; O3-NEXT: lock xorb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xor_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection15: +; HASWELL-O3-NEXT: lock xorb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -763,6 +897,27 @@ define void @atomic8_nand_monotonic(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_nand_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection16: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB16_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection17: +; HASWELL-O3-NEXT: notb %cl +; HASWELL-O3-NEXT: .Lpcsection18: +; HASWELL-O3-NEXT: orb $-43, %cl +; HASWELL-O3-NEXT: .Lpcsection19: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection20: +; HASWELL-O3-NEXT: jne .LBB16_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i8 42 monotonic, align 1, !pcsections !0 @@ -806,6 +961,15 @@ define void @atomic8_xchg_acquire(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xchg_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection21: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -845,6 +1009,14 @@ define void @atomic8_add_acquire(ptr %a) { ; O3-NEXT: lock addb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_add_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection22: +; HASWELL-O3-NEXT: lock addb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -884,6 +1056,14 @@ define void @atomic8_sub_acquire(ptr %a) { ; O3-NEXT: lock subb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_sub_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection23: +; HASWELL-O3-NEXT: lock subb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -923,6 +1103,14 @@ define void @atomic8_and_acquire(ptr %a) { ; O3-NEXT: lock andb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_and_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection24: +; HASWELL-O3-NEXT: lock andb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -962,6 +1150,14 @@ define void @atomic8_or_acquire(ptr %a) { ; O3-NEXT: lock orb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_or_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection25: +; HASWELL-O3-NEXT: lock orb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -1001,6 +1197,14 @@ define void @atomic8_xor_acquire(ptr %a) { ; O3-NEXT: lock xorb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xor_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection26: +; HASWELL-O3-NEXT: lock xorb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -1101,6 +1305,27 @@ define void @atomic8_nand_acquire(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_nand_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection27: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB23_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection28: +; HASWELL-O3-NEXT: notb %cl +; HASWELL-O3-NEXT: .Lpcsection29: +; HASWELL-O3-NEXT: orb $-43, %cl +; HASWELL-O3-NEXT: .Lpcsection30: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection31: +; HASWELL-O3-NEXT: jne .LBB23_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i8 42 acquire, align 1, !pcsections !0 @@ -1144,6 +1369,15 @@ define void @atomic8_xchg_release(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xchg_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection32: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1183,6 +1417,14 @@ define void @atomic8_add_release(ptr %a) { ; O3-NEXT: lock addb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_add_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection33: +; HASWELL-O3-NEXT: lock addb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1222,6 +1464,14 @@ define void @atomic8_sub_release(ptr %a) { ; O3-NEXT: lock subb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_sub_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection34: +; HASWELL-O3-NEXT: lock subb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1261,6 +1511,14 @@ define void @atomic8_and_release(ptr %a) { ; O3-NEXT: lock andb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_and_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection35: +; HASWELL-O3-NEXT: lock andb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1300,6 +1558,14 @@ define void @atomic8_or_release(ptr %a) { ; O3-NEXT: lock orb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_or_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection36: +; HASWELL-O3-NEXT: lock orb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1339,6 +1605,14 @@ define void @atomic8_xor_release(ptr %a) { ; O3-NEXT: lock xorb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xor_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection37: +; HASWELL-O3-NEXT: lock xorb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1439,6 +1713,27 @@ define void @atomic8_nand_release(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_nand_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection38: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB30_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection39: +; HASWELL-O3-NEXT: notb %cl +; HASWELL-O3-NEXT: .Lpcsection40: +; HASWELL-O3-NEXT: orb $-43, %cl +; HASWELL-O3-NEXT: .Lpcsection41: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection42: +; HASWELL-O3-NEXT: jne .LBB30_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i8 42 release, align 1, !pcsections !0 @@ -1482,6 +1777,15 @@ define void @atomic8_xchg_acq_rel(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xchg_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection43: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1521,6 +1825,14 @@ define void @atomic8_add_acq_rel(ptr %a) { ; O3-NEXT: lock addb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_add_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection44: +; HASWELL-O3-NEXT: lock addb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1560,6 +1872,14 @@ define void @atomic8_sub_acq_rel(ptr %a) { ; O3-NEXT: lock subb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_sub_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection45: +; HASWELL-O3-NEXT: lock subb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1599,6 +1919,14 @@ define void @atomic8_and_acq_rel(ptr %a) { ; O3-NEXT: lock andb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_and_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection46: +; HASWELL-O3-NEXT: lock andb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1638,6 +1966,14 @@ define void @atomic8_or_acq_rel(ptr %a) { ; O3-NEXT: lock orb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_or_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection47: +; HASWELL-O3-NEXT: lock orb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1677,6 +2013,14 @@ define void @atomic8_xor_acq_rel(ptr %a) { ; O3-NEXT: lock xorb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xor_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection48: +; HASWELL-O3-NEXT: lock xorb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1777,6 +2121,27 @@ define void @atomic8_nand_acq_rel(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_nand_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection49: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB37_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection50: +; HASWELL-O3-NEXT: notb %cl +; HASWELL-O3-NEXT: .Lpcsection51: +; HASWELL-O3-NEXT: orb $-43, %cl +; HASWELL-O3-NEXT: .Lpcsection52: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection53: +; HASWELL-O3-NEXT: jne .LBB37_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i8 42 acq_rel, align 1, !pcsections !0 @@ -1820,6 +2185,15 @@ define void @atomic8_xchg_seq_cst(ptr %a) { ; O3-NEXT: xchgb %al, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xchg_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection54: +; HASWELL-O3-NEXT: xchgb %al, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -1859,6 +2233,14 @@ define void @atomic8_add_seq_cst(ptr %a) { ; O3-NEXT: lock addb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_add_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection55: +; HASWELL-O3-NEXT: lock addb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -1898,6 +2280,14 @@ define void @atomic8_sub_seq_cst(ptr %a) { ; O3-NEXT: lock subb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_sub_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection56: +; HASWELL-O3-NEXT: lock subb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -1937,6 +2327,14 @@ define void @atomic8_and_seq_cst(ptr %a) { ; O3-NEXT: lock andb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_and_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection57: +; HASWELL-O3-NEXT: lock andb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -1976,6 +2374,14 @@ define void @atomic8_or_seq_cst(ptr %a) { ; O3-NEXT: lock orb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_or_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection58: +; HASWELL-O3-NEXT: lock orb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -2015,6 +2421,14 @@ define void @atomic8_xor_seq_cst(ptr %a) { ; O3-NEXT: lock xorb $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_xor_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection59: +; HASWELL-O3-NEXT: lock xorb $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -2115,6 +2529,27 @@ define void @atomic8_nand_seq_cst(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_nand_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection60: +; HASWELL-O3-NEXT: movzbl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB44_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection61: +; HASWELL-O3-NEXT: notb %cl +; HASWELL-O3-NEXT: .Lpcsection62: +; HASWELL-O3-NEXT: orb $-43, %cl +; HASWELL-O3-NEXT: .Lpcsection63: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection64: +; HASWELL-O3-NEXT: jne .LBB44_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i8 42 seq_cst, align 1, !pcsections !0 @@ -2200,6 +2635,25 @@ define void @atomic8_cas_monotonic(ptr %a) { ; O3-NEXT: lock cmpxchgb %cl, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_cas_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $1, %cl +; HASWELL-O3-NEXT: .Lpcsection65: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection66: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection67: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection68: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection69: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection70: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i8 42, i8 1 monotonic monotonic, align 1, !pcsections !0 @@ -2287,6 +2741,25 @@ define void @atomic8_cas_acquire(ptr %a) { ; O3-NEXT: lock cmpxchgb %cl, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_cas_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $1, %cl +; HASWELL-O3-NEXT: .Lpcsection71: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection72: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection73: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection74: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection75: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection76: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i8 42, i8 1 acquire monotonic, align 1, !pcsections !0 @@ -2374,6 +2847,25 @@ define void @atomic8_cas_release(ptr %a) { ; O3-NEXT: lock cmpxchgb %cl, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_cas_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $1, %cl +; HASWELL-O3-NEXT: .Lpcsection77: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection78: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection79: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection80: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection81: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection82: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i8 42, i8 1 release monotonic, align 1, !pcsections !0 @@ -2461,6 +2953,25 @@ define void @atomic8_cas_acq_rel(ptr %a) { ; O3-NEXT: lock cmpxchgb %cl, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_cas_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $1, %cl +; HASWELL-O3-NEXT: .Lpcsection83: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection84: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection85: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection86: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection87: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection88: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i8 42, i8 1 acq_rel monotonic, align 1, !pcsections !0 @@ -2548,6 +3059,25 @@ define void @atomic8_cas_seq_cst(ptr %a) { ; O3-NEXT: lock cmpxchgb %cl, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic8_cas_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movb $1, %cl +; HASWELL-O3-NEXT: .Lpcsection89: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection90: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection91: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection92: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection93: +; HASWELL-O3-NEXT: movb $42, %al +; HASWELL-O3-NEXT: .Lpcsection94: +; HASWELL-O3-NEXT: lock cmpxchgb %cl, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i8 42, i8 1 seq_cst monotonic, align 1, !pcsections !0 @@ -2589,6 +3119,14 @@ define i16 @atomic16_load_unordered(ptr %a) { ; O3-NEXT: movzwl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_load_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection95: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i16, ptr %a unordered, align 2, !pcsections !0 @@ -2628,6 +3166,14 @@ define i16 @atomic16_load_monotonic(ptr %a) { ; O3-NEXT: movzwl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_load_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection96: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i16, ptr %a monotonic, align 2, !pcsections !0 @@ -2667,6 +3213,14 @@ define i16 @atomic16_load_acquire(ptr %a) { ; O3-NEXT: movzwl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_load_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection97: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i16, ptr %a acquire, align 2, !pcsections !0 @@ -2706,6 +3260,14 @@ define i16 @atomic16_load_seq_cst(ptr %a) { ; O3-NEXT: movzwl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_load_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection98: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i16, ptr %a seq_cst, align 2, !pcsections !0 @@ -2745,6 +3307,14 @@ define void @atomic16_store_unordered(ptr %a) { ; O3-NEXT: movw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_store_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection99: +; HASWELL-O3-NEXT: movw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i16 42, ptr %a unordered, align 2, !pcsections !0 @@ -2784,6 +3354,14 @@ define void @atomic16_store_monotonic(ptr %a) { ; O3-NEXT: movw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_store_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection100: +; HASWELL-O3-NEXT: movw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i16 42, ptr %a monotonic, align 2, !pcsections !0 @@ -2823,6 +3401,14 @@ define void @atomic16_store_release(ptr %a) { ; O3-NEXT: movw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_store_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection101: +; HASWELL-O3-NEXT: movw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i16 42, ptr %a release, align 2, !pcsections !0 @@ -2866,6 +3452,15 @@ define void @atomic16_store_seq_cst(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_store_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection102: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i16 42, ptr %a seq_cst, align 2, !pcsections !0 @@ -2909,6 +3504,15 @@ define void @atomic16_xchg_monotonic(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xchg_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection103: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -2948,6 +3552,14 @@ define void @atomic16_add_monotonic(ptr %a) { ; O3-NEXT: lock addw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_add_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection104: +; HASWELL-O3-NEXT: lock addw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -2987,6 +3599,14 @@ define void @atomic16_sub_monotonic(ptr %a) { ; O3-NEXT: lock subw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_sub_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection105: +; HASWELL-O3-NEXT: lock subw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -3026,6 +3646,14 @@ define void @atomic16_and_monotonic(ptr %a) { ; O3-NEXT: lock andw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_and_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection106: +; HASWELL-O3-NEXT: lock andw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -3065,6 +3693,14 @@ define void @atomic16_or_monotonic(ptr %a) { ; O3-NEXT: lock orw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_or_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection107: +; HASWELL-O3-NEXT: lock orw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -3104,6 +3740,14 @@ define void @atomic16_xor_monotonic(ptr %a) { ; O3-NEXT: lock xorw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xor_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection108: +; HASWELL-O3-NEXT: lock xorw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -3220,6 +3864,31 @@ define void @atomic16_nand_monotonic(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_nand_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection109: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB64_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection110: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection111: +; HASWELL-O3-NEXT: orl $65493, %ecx # imm = 0xFFD5 +; HASWELL-O3-NEXT: .Lpcsection112: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax killed $eax +; HASWELL-O3-NEXT: .Lpcsection113: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection114: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax def $eax +; HASWELL-O3-NEXT: .Lpcsection115: +; HASWELL-O3-NEXT: jne .LBB64_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i16 42 monotonic, align 2, !pcsections !0 @@ -3263,6 +3932,15 @@ define void @atomic16_xchg_acquire(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xchg_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection116: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3302,6 +3980,14 @@ define void @atomic16_add_acquire(ptr %a) { ; O3-NEXT: lock addw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_add_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection117: +; HASWELL-O3-NEXT: lock addw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3341,6 +4027,14 @@ define void @atomic16_sub_acquire(ptr %a) { ; O3-NEXT: lock subw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_sub_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection118: +; HASWELL-O3-NEXT: lock subw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3380,6 +4074,14 @@ define void @atomic16_and_acquire(ptr %a) { ; O3-NEXT: lock andw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_and_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection119: +; HASWELL-O3-NEXT: lock andw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3419,6 +4121,14 @@ define void @atomic16_or_acquire(ptr %a) { ; O3-NEXT: lock orw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_or_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection120: +; HASWELL-O3-NEXT: lock orw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3458,6 +4168,14 @@ define void @atomic16_xor_acquire(ptr %a) { ; O3-NEXT: lock xorw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xor_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection121: +; HASWELL-O3-NEXT: lock xorw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3574,6 +4292,31 @@ define void @atomic16_nand_acquire(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_nand_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection122: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB71_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection123: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection124: +; HASWELL-O3-NEXT: orl $65493, %ecx # imm = 0xFFD5 +; HASWELL-O3-NEXT: .Lpcsection125: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax killed $eax +; HASWELL-O3-NEXT: .Lpcsection126: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection127: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax def $eax +; HASWELL-O3-NEXT: .Lpcsection128: +; HASWELL-O3-NEXT: jne .LBB71_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i16 42 acquire, align 2, !pcsections !0 @@ -3617,6 +4360,15 @@ define void @atomic16_xchg_release(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xchg_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection129: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3656,6 +4408,14 @@ define void @atomic16_add_release(ptr %a) { ; O3-NEXT: lock addw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_add_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection130: +; HASWELL-O3-NEXT: lock addw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3695,6 +4455,14 @@ define void @atomic16_sub_release(ptr %a) { ; O3-NEXT: lock subw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_sub_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection131: +; HASWELL-O3-NEXT: lock subw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3734,6 +4502,14 @@ define void @atomic16_and_release(ptr %a) { ; O3-NEXT: lock andw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_and_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection132: +; HASWELL-O3-NEXT: lock andw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3773,6 +4549,14 @@ define void @atomic16_or_release(ptr %a) { ; O3-NEXT: lock orw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_or_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection133: +; HASWELL-O3-NEXT: lock orw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3812,6 +4596,14 @@ define void @atomic16_xor_release(ptr %a) { ; O3-NEXT: lock xorw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xor_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection134: +; HASWELL-O3-NEXT: lock xorw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3928,6 +4720,31 @@ define void @atomic16_nand_release(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_nand_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection135: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB78_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection136: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection137: +; HASWELL-O3-NEXT: orl $65493, %ecx # imm = 0xFFD5 +; HASWELL-O3-NEXT: .Lpcsection138: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax killed $eax +; HASWELL-O3-NEXT: .Lpcsection139: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection140: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax def $eax +; HASWELL-O3-NEXT: .Lpcsection141: +; HASWELL-O3-NEXT: jne .LBB78_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i16 42 release, align 2, !pcsections !0 @@ -3971,6 +4788,15 @@ define void @atomic16_xchg_acq_rel(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xchg_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection142: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4010,6 +4836,14 @@ define void @atomic16_add_acq_rel(ptr %a) { ; O3-NEXT: lock addw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_add_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection143: +; HASWELL-O3-NEXT: lock addw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4049,6 +4883,14 @@ define void @atomic16_sub_acq_rel(ptr %a) { ; O3-NEXT: lock subw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_sub_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection144: +; HASWELL-O3-NEXT: lock subw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4088,6 +4930,14 @@ define void @atomic16_and_acq_rel(ptr %a) { ; O3-NEXT: lock andw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_and_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection145: +; HASWELL-O3-NEXT: lock andw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4127,6 +4977,14 @@ define void @atomic16_or_acq_rel(ptr %a) { ; O3-NEXT: lock orw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_or_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection146: +; HASWELL-O3-NEXT: lock orw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4166,6 +5024,14 @@ define void @atomic16_xor_acq_rel(ptr %a) { ; O3-NEXT: lock xorw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xor_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection147: +; HASWELL-O3-NEXT: lock xorw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4282,6 +5148,31 @@ define void @atomic16_nand_acq_rel(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_nand_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection148: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB85_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection149: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection150: +; HASWELL-O3-NEXT: orl $65493, %ecx # imm = 0xFFD5 +; HASWELL-O3-NEXT: .Lpcsection151: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax killed $eax +; HASWELL-O3-NEXT: .Lpcsection152: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection153: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax def $eax +; HASWELL-O3-NEXT: .Lpcsection154: +; HASWELL-O3-NEXT: jne .LBB85_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i16 42 acq_rel, align 2, !pcsections !0 @@ -4325,6 +5216,15 @@ define void @atomic16_xchg_seq_cst(ptr %a) { ; O3-NEXT: xchgw %ax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xchg_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection155: +; HASWELL-O3-NEXT: xchgw %ax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4364,6 +5264,14 @@ define void @atomic16_add_seq_cst(ptr %a) { ; O3-NEXT: lock addw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_add_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection156: +; HASWELL-O3-NEXT: lock addw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4403,6 +5311,14 @@ define void @atomic16_sub_seq_cst(ptr %a) { ; O3-NEXT: lock subw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_sub_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection157: +; HASWELL-O3-NEXT: lock subw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4442,6 +5358,14 @@ define void @atomic16_and_seq_cst(ptr %a) { ; O3-NEXT: lock andw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_and_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection158: +; HASWELL-O3-NEXT: lock andw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4481,6 +5405,14 @@ define void @atomic16_or_seq_cst(ptr %a) { ; O3-NEXT: lock orw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_or_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection159: +; HASWELL-O3-NEXT: lock orw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4520,6 +5452,14 @@ define void @atomic16_xor_seq_cst(ptr %a) { ; O3-NEXT: lock xorw $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_xor_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection160: +; HASWELL-O3-NEXT: lock xorw $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4636,6 +5576,31 @@ define void @atomic16_nand_seq_cst(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_nand_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection161: +; HASWELL-O3-NEXT: movzwl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB92_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection162: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection163: +; HASWELL-O3-NEXT: orl $65493, %ecx # imm = 0xFFD5 +; HASWELL-O3-NEXT: .Lpcsection164: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax killed $eax +; HASWELL-O3-NEXT: .Lpcsection165: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection166: +; HASWELL-O3-NEXT: # kill: def $ax killed $ax def $eax +; HASWELL-O3-NEXT: .Lpcsection167: +; HASWELL-O3-NEXT: jne .LBB92_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i16 42 seq_cst, align 2, !pcsections !0 @@ -4712,6 +5677,22 @@ define void @atomic16_cas_monotonic(ptr %a) { ; O3-NEXT: lock cmpxchgw %cx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_cas_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $1, %cx +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection168: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection169: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection170: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i16 42, i16 1 monotonic monotonic, align 2, !pcsections !0 @@ -4790,6 +5771,22 @@ define void @atomic16_cas_acquire(ptr %a) { ; O3-NEXT: lock cmpxchgw %cx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_cas_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $1, %cx +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection171: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection172: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection173: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i16 42, i16 1 acquire monotonic, align 2, !pcsections !0 @@ -4868,6 +5865,22 @@ define void @atomic16_cas_release(ptr %a) { ; O3-NEXT: lock cmpxchgw %cx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_cas_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $1, %cx +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection174: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection175: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection176: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i16 42, i16 1 release monotonic, align 2, !pcsections !0 @@ -4946,6 +5959,22 @@ define void @atomic16_cas_acq_rel(ptr %a) { ; O3-NEXT: lock cmpxchgw %cx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_cas_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $1, %cx +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection177: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection178: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection179: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i16 42, i16 1 acq_rel monotonic, align 2, !pcsections !0 @@ -5024,6 +6053,22 @@ define void @atomic16_cas_seq_cst(ptr %a) { ; O3-NEXT: lock cmpxchgw %cx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic16_cas_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movw $1, %cx +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection180: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection181: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movw $42, %ax +; HASWELL-O3-NEXT: .Lpcsection182: +; HASWELL-O3-NEXT: lock cmpxchgw %cx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i16 42, i16 1 seq_cst monotonic, align 2, !pcsections !0 @@ -5065,6 +6110,14 @@ define i32 @atomic32_load_unordered(ptr %a) { ; O3-NEXT: movl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_load_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection183: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i32, ptr %a unordered, align 4, !pcsections !0 @@ -5104,6 +6157,14 @@ define i32 @atomic32_load_monotonic(ptr %a) { ; O3-NEXT: movl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_load_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection184: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i32, ptr %a monotonic, align 4, !pcsections !0 @@ -5143,6 +6204,14 @@ define i32 @atomic32_load_acquire(ptr %a) { ; O3-NEXT: movl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_load_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection185: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i32, ptr %a acquire, align 4, !pcsections !0 @@ -5182,6 +6251,14 @@ define i32 @atomic32_load_seq_cst(ptr %a) { ; O3-NEXT: movl (%rdi), %eax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_load_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection186: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i32, ptr %a seq_cst, align 4, !pcsections !0 @@ -5221,6 +6298,14 @@ define void @atomic32_store_unordered(ptr %a) { ; O3-NEXT: movl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_store_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection187: +; HASWELL-O3-NEXT: movl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i32 42, ptr %a unordered, align 4, !pcsections !0 @@ -5260,6 +6345,14 @@ define void @atomic32_store_monotonic(ptr %a) { ; O3-NEXT: movl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_store_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection188: +; HASWELL-O3-NEXT: movl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i32 42, ptr %a monotonic, align 4, !pcsections !0 @@ -5299,6 +6392,14 @@ define void @atomic32_store_release(ptr %a) { ; O3-NEXT: movl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_store_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection189: +; HASWELL-O3-NEXT: movl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i32 42, ptr %a release, align 4, !pcsections !0 @@ -5342,6 +6443,15 @@ define void @atomic32_store_seq_cst(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_store_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection190: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i32 42, ptr %a seq_cst, align 4, !pcsections !0 @@ -5385,6 +6495,15 @@ define void @atomic32_xchg_monotonic(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xchg_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection191: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5424,6 +6543,14 @@ define void @atomic32_add_monotonic(ptr %a) { ; O3-NEXT: lock addl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_add_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection192: +; HASWELL-O3-NEXT: lock addl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5463,6 +6590,14 @@ define void @atomic32_sub_monotonic(ptr %a) { ; O3-NEXT: lock subl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_sub_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection193: +; HASWELL-O3-NEXT: lock subl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5502,6 +6637,14 @@ define void @atomic32_and_monotonic(ptr %a) { ; O3-NEXT: lock andl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_and_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection194: +; HASWELL-O3-NEXT: lock andl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5541,6 +6684,14 @@ define void @atomic32_or_monotonic(ptr %a) { ; O3-NEXT: lock orl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_or_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection195: +; HASWELL-O3-NEXT: lock orl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5580,6 +6731,14 @@ define void @atomic32_xor_monotonic(ptr %a) { ; O3-NEXT: lock xorl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xor_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection196: +; HASWELL-O3-NEXT: lock xorl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5680,6 +6839,27 @@ define void @atomic32_nand_monotonic(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_nand_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection197: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB112_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection198: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection199: +; HASWELL-O3-NEXT: orl $-43, %ecx +; HASWELL-O3-NEXT: .Lpcsection200: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection201: +; HASWELL-O3-NEXT: jne .LBB112_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i32 42 monotonic, align 4, !pcsections !0 @@ -5723,6 +6903,15 @@ define void @atomic32_xchg_acquire(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xchg_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection202: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -5762,6 +6951,14 @@ define void @atomic32_add_acquire(ptr %a) { ; O3-NEXT: lock addl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_add_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection203: +; HASWELL-O3-NEXT: lock addl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -5801,6 +6998,14 @@ define void @atomic32_sub_acquire(ptr %a) { ; O3-NEXT: lock subl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_sub_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection204: +; HASWELL-O3-NEXT: lock subl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -5840,6 +7045,14 @@ define void @atomic32_and_acquire(ptr %a) { ; O3-NEXT: lock andl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_and_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection205: +; HASWELL-O3-NEXT: lock andl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -5879,6 +7092,14 @@ define void @atomic32_or_acquire(ptr %a) { ; O3-NEXT: lock orl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_or_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection206: +; HASWELL-O3-NEXT: lock orl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -5918,6 +7139,14 @@ define void @atomic32_xor_acquire(ptr %a) { ; O3-NEXT: lock xorl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xor_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection207: +; HASWELL-O3-NEXT: lock xorl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -6018,6 +7247,27 @@ define void @atomic32_nand_acquire(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_nand_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection208: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB119_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection209: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection210: +; HASWELL-O3-NEXT: orl $-43, %ecx +; HASWELL-O3-NEXT: .Lpcsection211: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection212: +; HASWELL-O3-NEXT: jne .LBB119_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i32 42 acquire, align 4, !pcsections !0 @@ -6061,6 +7311,15 @@ define void @atomic32_xchg_release(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xchg_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection213: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6100,6 +7359,14 @@ define void @atomic32_add_release(ptr %a) { ; O3-NEXT: lock addl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_add_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection214: +; HASWELL-O3-NEXT: lock addl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6139,6 +7406,14 @@ define void @atomic32_sub_release(ptr %a) { ; O3-NEXT: lock subl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_sub_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection215: +; HASWELL-O3-NEXT: lock subl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6178,6 +7453,14 @@ define void @atomic32_and_release(ptr %a) { ; O3-NEXT: lock andl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_and_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection216: +; HASWELL-O3-NEXT: lock andl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6217,6 +7500,14 @@ define void @atomic32_or_release(ptr %a) { ; O3-NEXT: lock orl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_or_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection217: +; HASWELL-O3-NEXT: lock orl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6256,6 +7547,14 @@ define void @atomic32_xor_release(ptr %a) { ; O3-NEXT: lock xorl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xor_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection218: +; HASWELL-O3-NEXT: lock xorl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6356,6 +7655,27 @@ define void @atomic32_nand_release(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_nand_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection219: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB126_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection220: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection221: +; HASWELL-O3-NEXT: orl $-43, %ecx +; HASWELL-O3-NEXT: .Lpcsection222: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection223: +; HASWELL-O3-NEXT: jne .LBB126_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i32 42 release, align 4, !pcsections !0 @@ -6399,6 +7719,15 @@ define void @atomic32_xchg_acq_rel(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xchg_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection224: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6438,6 +7767,14 @@ define void @atomic32_add_acq_rel(ptr %a) { ; O3-NEXT: lock addl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_add_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection225: +; HASWELL-O3-NEXT: lock addl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6477,6 +7814,14 @@ define void @atomic32_sub_acq_rel(ptr %a) { ; O3-NEXT: lock subl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_sub_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection226: +; HASWELL-O3-NEXT: lock subl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6516,6 +7861,14 @@ define void @atomic32_and_acq_rel(ptr %a) { ; O3-NEXT: lock andl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_and_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection227: +; HASWELL-O3-NEXT: lock andl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6555,6 +7908,14 @@ define void @atomic32_or_acq_rel(ptr %a) { ; O3-NEXT: lock orl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_or_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection228: +; HASWELL-O3-NEXT: lock orl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6594,6 +7955,14 @@ define void @atomic32_xor_acq_rel(ptr %a) { ; O3-NEXT: lock xorl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xor_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection229: +; HASWELL-O3-NEXT: lock xorl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6694,6 +8063,27 @@ define void @atomic32_nand_acq_rel(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_nand_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection230: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB133_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection231: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection232: +; HASWELL-O3-NEXT: orl $-43, %ecx +; HASWELL-O3-NEXT: .Lpcsection233: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection234: +; HASWELL-O3-NEXT: jne .LBB133_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i32 42 acq_rel, align 4, !pcsections !0 @@ -6737,6 +8127,15 @@ define void @atomic32_xchg_seq_cst(ptr %a) { ; O3-NEXT: xchgl %eax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xchg_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection235: +; HASWELL-O3-NEXT: xchgl %eax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -6776,6 +8175,14 @@ define void @atomic32_add_seq_cst(ptr %a) { ; O3-NEXT: lock addl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_add_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection236: +; HASWELL-O3-NEXT: lock addl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -6815,6 +8222,14 @@ define void @atomic32_sub_seq_cst(ptr %a) { ; O3-NEXT: lock subl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_sub_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection237: +; HASWELL-O3-NEXT: lock subl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -6854,6 +8269,14 @@ define void @atomic32_and_seq_cst(ptr %a) { ; O3-NEXT: lock andl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_and_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection238: +; HASWELL-O3-NEXT: lock andl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -6893,6 +8316,14 @@ define void @atomic32_or_seq_cst(ptr %a) { ; O3-NEXT: lock orl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_or_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection239: +; HASWELL-O3-NEXT: lock orl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -6932,6 +8363,14 @@ define void @atomic32_xor_seq_cst(ptr %a) { ; O3-NEXT: lock xorl $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_xor_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection240: +; HASWELL-O3-NEXT: lock xorl $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -7032,6 +8471,27 @@ define void @atomic32_nand_seq_cst(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_nand_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection241: +; HASWELL-O3-NEXT: movl (%rdi), %eax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB140_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection242: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection243: +; HASWELL-O3-NEXT: orl $-43, %ecx +; HASWELL-O3-NEXT: .Lpcsection244: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection245: +; HASWELL-O3-NEXT: jne .LBB140_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i32 42 seq_cst, align 4, !pcsections !0 @@ -7117,6 +8577,25 @@ define void @atomic32_cas_monotonic(ptr %a) { ; O3-NEXT: lock cmpxchgl %ecx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_cas_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection246: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection247: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection248: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection249: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection250: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection251: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i32 42, i32 1 monotonic monotonic, align 4, !pcsections !0 @@ -7204,6 +8683,25 @@ define void @atomic32_cas_acquire(ptr %a) { ; O3-NEXT: lock cmpxchgl %ecx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_cas_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection252: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection253: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection254: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection255: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection256: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection257: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i32 42, i32 1 acquire monotonic, align 4, !pcsections !0 @@ -7291,6 +8789,25 @@ define void @atomic32_cas_release(ptr %a) { ; O3-NEXT: lock cmpxchgl %ecx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_cas_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection258: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection259: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection260: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection261: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection262: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection263: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i32 42, i32 1 release monotonic, align 4, !pcsections !0 @@ -7378,6 +8895,25 @@ define void @atomic32_cas_acq_rel(ptr %a) { ; O3-NEXT: lock cmpxchgl %ecx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_cas_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection264: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection265: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection266: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection267: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection268: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection269: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i32 42, i32 1 acq_rel monotonic, align 4, !pcsections !0 @@ -7465,6 +9001,25 @@ define void @atomic32_cas_seq_cst(ptr %a) { ; O3-NEXT: lock cmpxchgl %ecx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic32_cas_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection270: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection271: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection272: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection273: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection274: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection275: +; HASWELL-O3-NEXT: lock cmpxchgl %ecx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i32 42, i32 1 seq_cst monotonic, align 4, !pcsections !0 @@ -7506,6 +9061,14 @@ define i64 @atomic64_load_unordered(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_load_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection276: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i64, ptr %a unordered, align 8, !pcsections !0 @@ -7545,6 +9108,14 @@ define i64 @atomic64_load_monotonic(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_load_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection277: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i64, ptr %a monotonic, align 8, !pcsections !0 @@ -7584,6 +9155,14 @@ define i64 @atomic64_load_acquire(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_load_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection278: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i64, ptr %a acquire, align 8, !pcsections !0 @@ -7623,6 +9202,14 @@ define i64 @atomic64_load_seq_cst(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_load_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection279: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i64, ptr %a seq_cst, align 8, !pcsections !0 @@ -7662,6 +9249,14 @@ define ptr @atomic64_load_seq_cst_ptr_ty(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_load_seq_cst_ptr_ty: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection280: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic ptr, ptr %a seq_cst, align 8, !pcsections !0 @@ -7701,6 +9296,14 @@ define void @atomic64_store_unordered(ptr %a) { ; O3-NEXT: movq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_store_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection281: +; HASWELL-O3-NEXT: movq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i64 42, ptr %a unordered, align 8, !pcsections !0 @@ -7740,6 +9343,14 @@ define void @atomic64_store_monotonic(ptr %a) { ; O3-NEXT: movq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_store_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection282: +; HASWELL-O3-NEXT: movq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i64 42, ptr %a monotonic, align 8, !pcsections !0 @@ -7779,6 +9390,14 @@ define void @atomic64_store_release(ptr %a) { ; O3-NEXT: movq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_store_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection283: +; HASWELL-O3-NEXT: movq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i64 42, ptr %a release, align 8, !pcsections !0 @@ -7822,6 +9441,15 @@ define void @atomic64_store_seq_cst(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_store_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection284: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i64 42, ptr %a seq_cst, align 8, !pcsections !0 @@ -7861,6 +9489,14 @@ define void @atomic64_store_seq_cst_ptr_ty(ptr %a, ptr %v) { ; O3-NEXT: xchgq %rsi, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_store_seq_cst_ptr_ty: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection285: +; HASWELL-O3-NEXT: xchgq %rsi, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic ptr %v, ptr %a seq_cst, align 8, !pcsections !0 @@ -7904,6 +9540,15 @@ define void @atomic64_xchg_monotonic(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xchg_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection286: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -7943,6 +9588,14 @@ define void @atomic64_add_monotonic(ptr %a) { ; O3-NEXT: lock addq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_add_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection287: +; HASWELL-O3-NEXT: lock addq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -7982,6 +9635,14 @@ define void @atomic64_sub_monotonic(ptr %a) { ; O3-NEXT: lock subq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_sub_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection288: +; HASWELL-O3-NEXT: lock subq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -8021,6 +9682,14 @@ define void @atomic64_and_monotonic(ptr %a) { ; O3-NEXT: lock andq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_and_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection289: +; HASWELL-O3-NEXT: lock andq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -8060,6 +9729,14 @@ define void @atomic64_or_monotonic(ptr %a) { ; O3-NEXT: lock orq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_or_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection290: +; HASWELL-O3-NEXT: lock orq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -8099,6 +9776,14 @@ define void @atomic64_xor_monotonic(ptr %a) { ; O3-NEXT: lock xorq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xor_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection291: +; HASWELL-O3-NEXT: lock xorq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -8202,6 +9887,27 @@ define void @atomic64_nand_monotonic(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_nand_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection292: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB162_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection293: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection294: +; HASWELL-O3-NEXT: orq $-43, %rcx +; HASWELL-O3-NEXT: .Lpcsection295: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection296: +; HASWELL-O3-NEXT: jne .LBB162_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i64 42 monotonic, align 8, !pcsections !0 @@ -8245,6 +9951,15 @@ define void @atomic64_xchg_acquire(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xchg_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection297: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8284,6 +9999,14 @@ define void @atomic64_add_acquire(ptr %a) { ; O3-NEXT: lock addq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_add_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection298: +; HASWELL-O3-NEXT: lock addq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8323,6 +10046,14 @@ define void @atomic64_sub_acquire(ptr %a) { ; O3-NEXT: lock subq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_sub_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection299: +; HASWELL-O3-NEXT: lock subq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8362,6 +10093,14 @@ define void @atomic64_and_acquire(ptr %a) { ; O3-NEXT: lock andq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_and_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection300: +; HASWELL-O3-NEXT: lock andq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8401,6 +10140,14 @@ define void @atomic64_or_acquire(ptr %a) { ; O3-NEXT: lock orq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_or_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection301: +; HASWELL-O3-NEXT: lock orq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8440,6 +10187,14 @@ define void @atomic64_xor_acquire(ptr %a) { ; O3-NEXT: lock xorq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xor_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection302: +; HASWELL-O3-NEXT: lock xorq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8543,6 +10298,27 @@ define void @atomic64_nand_acquire(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_nand_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection303: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB169_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection304: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection305: +; HASWELL-O3-NEXT: orq $-43, %rcx +; HASWELL-O3-NEXT: .Lpcsection306: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection307: +; HASWELL-O3-NEXT: jne .LBB169_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i64 42 acquire, align 8, !pcsections !0 @@ -8586,6 +10362,15 @@ define void @atomic64_xchg_release(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xchg_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection308: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8625,6 +10410,14 @@ define void @atomic64_add_release(ptr %a) { ; O3-NEXT: lock addq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_add_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection309: +; HASWELL-O3-NEXT: lock addq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8664,6 +10457,14 @@ define void @atomic64_sub_release(ptr %a) { ; O3-NEXT: lock subq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_sub_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection310: +; HASWELL-O3-NEXT: lock subq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8703,6 +10504,14 @@ define void @atomic64_and_release(ptr %a) { ; O3-NEXT: lock andq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_and_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection311: +; HASWELL-O3-NEXT: lock andq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8742,6 +10551,14 @@ define void @atomic64_or_release(ptr %a) { ; O3-NEXT: lock orq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_or_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection312: +; HASWELL-O3-NEXT: lock orq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8781,6 +10598,14 @@ define void @atomic64_xor_release(ptr %a) { ; O3-NEXT: lock xorq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xor_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection313: +; HASWELL-O3-NEXT: lock xorq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8884,6 +10709,27 @@ define void @atomic64_nand_release(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_nand_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection314: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB176_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection315: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection316: +; HASWELL-O3-NEXT: orq $-43, %rcx +; HASWELL-O3-NEXT: .Lpcsection317: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection318: +; HASWELL-O3-NEXT: jne .LBB176_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i64 42 release, align 8, !pcsections !0 @@ -8927,6 +10773,15 @@ define void @atomic64_xchg_acq_rel(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xchg_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection319: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -8966,6 +10821,14 @@ define void @atomic64_add_acq_rel(ptr %a) { ; O3-NEXT: lock addq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_add_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection320: +; HASWELL-O3-NEXT: lock addq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9005,6 +10868,14 @@ define void @atomic64_sub_acq_rel(ptr %a) { ; O3-NEXT: lock subq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_sub_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection321: +; HASWELL-O3-NEXT: lock subq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9044,6 +10915,14 @@ define void @atomic64_and_acq_rel(ptr %a) { ; O3-NEXT: lock andq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_and_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection322: +; HASWELL-O3-NEXT: lock andq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9083,6 +10962,14 @@ define void @atomic64_or_acq_rel(ptr %a) { ; O3-NEXT: lock orq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_or_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection323: +; HASWELL-O3-NEXT: lock orq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9122,6 +11009,14 @@ define void @atomic64_xor_acq_rel(ptr %a) { ; O3-NEXT: lock xorq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xor_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection324: +; HASWELL-O3-NEXT: lock xorq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9225,6 +11120,27 @@ define void @atomic64_nand_acq_rel(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_nand_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection325: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB183_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection326: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection327: +; HASWELL-O3-NEXT: orq $-43, %rcx +; HASWELL-O3-NEXT: .Lpcsection328: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection329: +; HASWELL-O3-NEXT: jne .LBB183_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i64 42 acq_rel, align 8, !pcsections !0 @@ -9268,6 +11184,15 @@ define void @atomic64_xchg_seq_cst(ptr %a) { ; O3-NEXT: xchgq %rax, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xchg_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection330: +; HASWELL-O3-NEXT: xchgq %rax, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9307,6 +11232,14 @@ define void @atomic64_add_seq_cst(ptr %a) { ; O3-NEXT: lock addq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_add_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection331: +; HASWELL-O3-NEXT: lock addq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9346,6 +11279,14 @@ define void @atomic64_sub_seq_cst(ptr %a) { ; O3-NEXT: lock subq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_sub_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection332: +; HASWELL-O3-NEXT: lock subq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9385,6 +11326,14 @@ define void @atomic64_and_seq_cst(ptr %a) { ; O3-NEXT: lock andq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_and_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection333: +; HASWELL-O3-NEXT: lock andq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9424,6 +11373,14 @@ define void @atomic64_or_seq_cst(ptr %a) { ; O3-NEXT: lock orq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_or_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection334: +; HASWELL-O3-NEXT: lock orq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9463,6 +11420,14 @@ define void @atomic64_xor_seq_cst(ptr %a) { ; O3-NEXT: lock xorq $42, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_xor_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection335: +; HASWELL-O3-NEXT: lock xorq $42, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9566,6 +11531,27 @@ define void @atomic64_nand_seq_cst(ptr %a) { ; O3-NEXT: # %bb.2: # %atomicrmw.end ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_nand_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection336: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB190_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ecx +; HASWELL-O3-NEXT: .Lpcsection337: +; HASWELL-O3-NEXT: notl %ecx +; HASWELL-O3-NEXT: .Lpcsection338: +; HASWELL-O3-NEXT: orq $-43, %rcx +; HASWELL-O3-NEXT: .Lpcsection339: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection340: +; HASWELL-O3-NEXT: jne .LBB190_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i64 42 seq_cst, align 8, !pcsections !0 @@ -9651,6 +11637,25 @@ define void @atomic64_cas_monotonic(ptr %a) { ; O3-NEXT: lock cmpxchgq %rcx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection341: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection342: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection343: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection344: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection345: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection346: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i64 42, i64 1 monotonic monotonic, align 8, !pcsections !0 @@ -9738,6 +11743,25 @@ define void @atomic64_cas_acquire(ptr %a) { ; O3-NEXT: lock cmpxchgq %rcx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection347: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection348: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection349: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection350: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection351: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection352: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i64 42, i64 1 acquire monotonic, align 8, !pcsections !0 @@ -9825,6 +11849,25 @@ define void @atomic64_cas_release(ptr %a) { ; O3-NEXT: lock cmpxchgq %rcx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection353: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection354: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection355: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection356: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection357: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection358: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i64 42, i64 1 release monotonic, align 8, !pcsections !0 @@ -9912,6 +11955,25 @@ define void @atomic64_cas_acq_rel(ptr %a) { ; O3-NEXT: lock cmpxchgq %rcx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection359: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection360: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection361: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection362: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection363: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection364: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i64 42, i64 1 acq_rel monotonic, align 8, !pcsections !0 @@ -9999,6 +12061,25 @@ define void @atomic64_cas_seq_cst(ptr %a) { ; O3-NEXT: lock cmpxchgq %rcx, (%rdi) ; O3-NEXT: movq $3, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: movl $1, %ecx +; HASWELL-O3-NEXT: .Lpcsection365: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection366: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection367: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection368: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection369: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection370: +; HASWELL-O3-NEXT: lock cmpxchgq %rcx, (%rdi) +; HASWELL-O3-NEXT: movq $3, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i64 42, i64 1 seq_cst monotonic, align 8, !pcsections !0 @@ -10044,6 +12125,15 @@ define void @atomic64_cas_seq_cst_ptr_ty(ptr %a, ptr %v1, ptr %v2) { ; O3-NEXT: lock cmpxchgq %rdx, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic64_cas_seq_cst_ptr_ty: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq %rsi, %rax +; HASWELL-O3-NEXT: movq foo(%rip), %rcx +; HASWELL-O3-NEXT: .Lpcsection371: +; HASWELL-O3-NEXT: lock cmpxchgq %rdx, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, ptr %v1, ptr %v2 seq_cst seq_cst, align 8, !pcsections !0 @@ -10102,6 +12192,18 @@ define i64 @atomic_use_cond(ptr %a) { ; O3-NEXT: .LBB197_2: # %else ; O3-NEXT: movl $2, %eax ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic_use_cond: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: .Lpcsection372: +; HASWELL-O3-NEXT: lock decq (%rdi) +; HASWELL-O3-NEXT: jne .LBB197_2 +; HASWELL-O3-NEXT: # %bb.1: # %then +; HASWELL-O3-NEXT: movl $1, %eax +; HASWELL-O3-NEXT: retq +; HASWELL-O3-NEXT: .LBB197_2: # %else +; HASWELL-O3-NEXT: movl $2, %eax +; HASWELL-O3-NEXT: retq entry: %x = atomicrmw sub ptr %a, i64 1 seq_cst, align 8, !pcsections !0 %y = icmp eq i64 %x, 1 @@ -10196,6 +12298,18 @@ define i128 @atomic128_load_unordered(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_load_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection373: +; HASWELL-O3-NEXT: vmovdqa (%rdi), %xmm0 +; HASWELL-O3-NEXT: .Lpcsection374: +; HASWELL-O3-NEXT: vmovq %xmm0, %rax +; HASWELL-O3-NEXT: .Lpcsection375: +; HASWELL-O3-NEXT: vpextrq $1, %xmm0, %rdx +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i128, ptr %a unordered, align 16, !pcsections !0 @@ -10285,6 +12399,18 @@ define i128 @atomic128_load_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_load_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection376: +; HASWELL-O3-NEXT: vmovdqa (%rdi), %xmm0 +; HASWELL-O3-NEXT: .Lpcsection377: +; HASWELL-O3-NEXT: vmovq %xmm0, %rax +; HASWELL-O3-NEXT: .Lpcsection378: +; HASWELL-O3-NEXT: vpextrq $1, %xmm0, %rdx +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i128, ptr %a monotonic, align 16, !pcsections !0 @@ -10374,6 +12500,18 @@ define i128 @atomic128_load_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_load_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection379: +; HASWELL-O3-NEXT: vmovdqa (%rdi), %xmm0 +; HASWELL-O3-NEXT: .Lpcsection380: +; HASWELL-O3-NEXT: vmovq %xmm0, %rax +; HASWELL-O3-NEXT: .Lpcsection381: +; HASWELL-O3-NEXT: vpextrq $1, %xmm0, %rdx +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i128, ptr %a acquire, align 16, !pcsections !0 @@ -10463,6 +12601,18 @@ define i128 @atomic128_load_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_load_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection382: +; HASWELL-O3-NEXT: vmovdqa (%rdi), %xmm0 +; HASWELL-O3-NEXT: .Lpcsection383: +; HASWELL-O3-NEXT: vmovq %xmm0, %rax +; HASWELL-O3-NEXT: .Lpcsection384: +; HASWELL-O3-NEXT: vpextrq $1, %xmm0, %rdx +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic i128, ptr %a seq_cst, align 16, !pcsections !0 @@ -10502,6 +12652,14 @@ define ptr @atomic128_load_seq_cst_ptr_ty(ptr %a) { ; O3-NEXT: movq (%rdi), %rax ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_load_seq_cst_ptr_ty: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection385: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = load atomic ptr, ptr %a seq_cst, align 16, !pcsections !0 @@ -10629,6 +12787,16 @@ define void @atomic128_store_unordered(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_store_unordered: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection386: +; HASWELL-O3-NEXT: vmovss {{.*#+}} xmm0 = [42,0,0,0] +; HASWELL-O3-NEXT: .Lpcsection387: +; HASWELL-O3-NEXT: vmovaps %xmm0, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i128 42, ptr %a unordered, align 16, !pcsections !0 @@ -10756,6 +12924,16 @@ define void @atomic128_store_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_store_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection388: +; HASWELL-O3-NEXT: vmovss {{.*#+}} xmm0 = [42,0,0,0] +; HASWELL-O3-NEXT: .Lpcsection389: +; HASWELL-O3-NEXT: vmovaps %xmm0, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i128 42, ptr %a monotonic, align 16, !pcsections !0 @@ -10883,6 +13061,16 @@ define void @atomic128_store_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_store_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection390: +; HASWELL-O3-NEXT: vmovss {{.*#+}} xmm0 = [42,0,0,0] +; HASWELL-O3-NEXT: .Lpcsection391: +; HASWELL-O3-NEXT: vmovaps %xmm0, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i128 42, ptr %a release, align 16, !pcsections !0 @@ -11010,6 +13198,18 @@ define void @atomic128_store_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_store_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection392: +; HASWELL-O3-NEXT: vmovss {{.*#+}} xmm0 = [42,0,0,0] +; HASWELL-O3-NEXT: .Lpcsection393: +; HASWELL-O3-NEXT: vmovaps %xmm0, (%rdi) +; HASWELL-O3-NEXT: .Lpcsection394: +; HASWELL-O3-NEXT: lock orl $0, -{{[0-9]+}}(%rsp) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic i128 42, ptr %a seq_cst, align 16, !pcsections !0 @@ -11049,6 +13249,14 @@ define void @atomic128_store_seq_cst_ptr_ty(ptr %a, ptr %v) { ; O3-NEXT: xchgq %rsi, (%rdi) ; O3-NEXT: movq $1, foo(%rip) ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_store_seq_cst_ptr_ty: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection395: +; HASWELL-O3-NEXT: xchgq %rsi, (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 store atomic ptr %v, ptr %a seq_cst, align 16, !pcsections !0 @@ -11176,6 +13384,33 @@ define void @atomic128_xchg_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xchg_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection396: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection397: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection398: +; HASWELL-O3-NEXT: movl $42, %ebx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB208_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: .Lpcsection399: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection400: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection401: +; HASWELL-O3-NEXT: jne .LBB208_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11309,6 +13544,35 @@ define void @atomic128_add_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_add_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection402: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection403: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB209_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection404: +; HASWELL-O3-NEXT: addq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection405: +; HASWELL-O3-NEXT: adcq $0, %rcx +; HASWELL-O3-NEXT: .Lpcsection406: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection407: +; HASWELL-O3-NEXT: jne .LBB209_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11442,6 +13706,35 @@ define void @atomic128_sub_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_sub_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection408: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection409: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB210_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection410: +; HASWELL-O3-NEXT: addq $-42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection411: +; HASWELL-O3-NEXT: adcq $-1, %rcx +; HASWELL-O3-NEXT: .Lpcsection412: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection413: +; HASWELL-O3-NEXT: jne .LBB210_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11574,6 +13867,34 @@ define void @atomic128_and_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_and_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection414: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection415: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB211_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection416: +; HASWELL-O3-NEXT: andl $42, %ebx +; HASWELL-O3-NEXT: .Lpcsection417: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection418: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection419: +; HASWELL-O3-NEXT: jne .LBB211_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11699,6 +14020,33 @@ define void @atomic128_or_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_or_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection420: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection421: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB212_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection422: +; HASWELL-O3-NEXT: orq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection423: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection424: +; HASWELL-O3-NEXT: jne .LBB212_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11824,6 +14172,33 @@ define void @atomic128_xor_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xor_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection425: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection426: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB213_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection427: +; HASWELL-O3-NEXT: xorq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection428: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection429: +; HASWELL-O3-NEXT: jne .LBB213_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -11964,6 +14339,36 @@ define void @atomic128_nand_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_nand_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection430: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection431: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection432: +; HASWELL-O3-NEXT: movq $-1, %rcx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB214_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection433: +; HASWELL-O3-NEXT: notl %ebx +; HASWELL-O3-NEXT: .Lpcsection434: +; HASWELL-O3-NEXT: orq $-43, %rbx +; HASWELL-O3-NEXT: .Lpcsection435: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection436: +; HASWELL-O3-NEXT: jne .LBB214_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i128 42 monotonic, align 16, !pcsections !0 @@ -12091,6 +14496,33 @@ define void @atomic128_xchg_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xchg_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection437: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection438: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection439: +; HASWELL-O3-NEXT: movl $42, %ebx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB215_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: .Lpcsection440: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection441: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection442: +; HASWELL-O3-NEXT: jne .LBB215_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12224,6 +14656,35 @@ define void @atomic128_add_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_add_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection443: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection444: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB216_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection445: +; HASWELL-O3-NEXT: addq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection446: +; HASWELL-O3-NEXT: adcq $0, %rcx +; HASWELL-O3-NEXT: .Lpcsection447: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection448: +; HASWELL-O3-NEXT: jne .LBB216_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12357,6 +14818,35 @@ define void @atomic128_sub_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_sub_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection449: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection450: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB217_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection451: +; HASWELL-O3-NEXT: addq $-42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection452: +; HASWELL-O3-NEXT: adcq $-1, %rcx +; HASWELL-O3-NEXT: .Lpcsection453: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection454: +; HASWELL-O3-NEXT: jne .LBB217_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12489,6 +14979,34 @@ define void @atomic128_and_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_and_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection455: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection456: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB218_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection457: +; HASWELL-O3-NEXT: andl $42, %ebx +; HASWELL-O3-NEXT: .Lpcsection458: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection459: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection460: +; HASWELL-O3-NEXT: jne .LBB218_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12614,6 +15132,33 @@ define void @atomic128_or_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_or_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection461: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection462: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB219_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection463: +; HASWELL-O3-NEXT: orq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection464: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection465: +; HASWELL-O3-NEXT: jne .LBB219_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12739,6 +15284,33 @@ define void @atomic128_xor_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xor_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection466: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection467: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB220_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection468: +; HASWELL-O3-NEXT: xorq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection469: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection470: +; HASWELL-O3-NEXT: jne .LBB220_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -12879,6 +15451,36 @@ define void @atomic128_nand_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_nand_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection471: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection472: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection473: +; HASWELL-O3-NEXT: movq $-1, %rcx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB221_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection474: +; HASWELL-O3-NEXT: notl %ebx +; HASWELL-O3-NEXT: .Lpcsection475: +; HASWELL-O3-NEXT: orq $-43, %rbx +; HASWELL-O3-NEXT: .Lpcsection476: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection477: +; HASWELL-O3-NEXT: jne .LBB221_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i128 42 acquire, align 16, !pcsections !0 @@ -13006,6 +15608,33 @@ define void @atomic128_xchg_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xchg_release: +; HASWELL-O3: # %bb.0: +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection478: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection479: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection480: +; HASWELL-O3-NEXT: movl $42, %ebx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB222_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: .Lpcsection481: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection482: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection483: +; HASWELL-O3-NEXT: jne .LBB222_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i128 42 release, align 16, !pcsections !0 store volatile i64 1, ptr @foo, align 8 @@ -13138,6 +15767,35 @@ define void @atomic128_add_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_add_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection484: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection485: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB223_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection486: +; HASWELL-O3-NEXT: addq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection487: +; HASWELL-O3-NEXT: adcq $0, %rcx +; HASWELL-O3-NEXT: .Lpcsection488: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection489: +; HASWELL-O3-NEXT: jne .LBB223_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13271,6 +15929,35 @@ define void @atomic128_sub_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_sub_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection490: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection491: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB224_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection492: +; HASWELL-O3-NEXT: addq $-42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection493: +; HASWELL-O3-NEXT: adcq $-1, %rcx +; HASWELL-O3-NEXT: .Lpcsection494: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection495: +; HASWELL-O3-NEXT: jne .LBB224_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13403,6 +16090,34 @@ define void @atomic128_and_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_and_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection496: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection497: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB225_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection498: +; HASWELL-O3-NEXT: andl $42, %ebx +; HASWELL-O3-NEXT: .Lpcsection499: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection500: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection501: +; HASWELL-O3-NEXT: jne .LBB225_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13528,6 +16243,33 @@ define void @atomic128_or_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_or_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection502: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection503: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB226_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection504: +; HASWELL-O3-NEXT: orq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection505: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection506: +; HASWELL-O3-NEXT: jne .LBB226_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13653,6 +16395,33 @@ define void @atomic128_xor_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xor_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection507: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection508: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB227_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection509: +; HASWELL-O3-NEXT: xorq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection510: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection511: +; HASWELL-O3-NEXT: jne .LBB227_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13793,6 +16562,36 @@ define void @atomic128_nand_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_nand_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection512: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection513: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection514: +; HASWELL-O3-NEXT: movq $-1, %rcx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB228_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection515: +; HASWELL-O3-NEXT: notl %ebx +; HASWELL-O3-NEXT: .Lpcsection516: +; HASWELL-O3-NEXT: orq $-43, %rbx +; HASWELL-O3-NEXT: .Lpcsection517: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection518: +; HASWELL-O3-NEXT: jne .LBB228_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i128 42 release, align 16, !pcsections !0 @@ -13920,6 +16719,33 @@ define void @atomic128_xchg_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xchg_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection519: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection520: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection521: +; HASWELL-O3-NEXT: movl $42, %ebx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB229_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: .Lpcsection522: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection523: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection524: +; HASWELL-O3-NEXT: jne .LBB229_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14053,6 +16879,35 @@ define void @atomic128_add_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_add_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection525: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection526: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB230_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection527: +; HASWELL-O3-NEXT: addq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection528: +; HASWELL-O3-NEXT: adcq $0, %rcx +; HASWELL-O3-NEXT: .Lpcsection529: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection530: +; HASWELL-O3-NEXT: jne .LBB230_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14186,6 +17041,35 @@ define void @atomic128_sub_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_sub_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection531: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection532: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB231_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection533: +; HASWELL-O3-NEXT: addq $-42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection534: +; HASWELL-O3-NEXT: adcq $-1, %rcx +; HASWELL-O3-NEXT: .Lpcsection535: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection536: +; HASWELL-O3-NEXT: jne .LBB231_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14318,6 +17202,34 @@ define void @atomic128_and_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_and_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection537: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection538: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB232_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection539: +; HASWELL-O3-NEXT: andl $42, %ebx +; HASWELL-O3-NEXT: .Lpcsection540: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection541: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection542: +; HASWELL-O3-NEXT: jne .LBB232_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14443,6 +17355,33 @@ define void @atomic128_or_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_or_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection543: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection544: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB233_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection545: +; HASWELL-O3-NEXT: orq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection546: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection547: +; HASWELL-O3-NEXT: jne .LBB233_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14568,6 +17507,33 @@ define void @atomic128_xor_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xor_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection548: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection549: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB234_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection550: +; HASWELL-O3-NEXT: xorq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection551: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection552: +; HASWELL-O3-NEXT: jne .LBB234_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14708,6 +17674,36 @@ define void @atomic128_nand_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_nand_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection553: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection554: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection555: +; HASWELL-O3-NEXT: movq $-1, %rcx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB235_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection556: +; HASWELL-O3-NEXT: notl %ebx +; HASWELL-O3-NEXT: .Lpcsection557: +; HASWELL-O3-NEXT: orq $-43, %rbx +; HASWELL-O3-NEXT: .Lpcsection558: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection559: +; HASWELL-O3-NEXT: jne .LBB235_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i128 42 acq_rel, align 16, !pcsections !0 @@ -14835,6 +17831,33 @@ define void @atomic128_xchg_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xchg_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection560: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection561: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection562: +; HASWELL-O3-NEXT: movl $42, %ebx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB236_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: .Lpcsection563: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection564: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection565: +; HASWELL-O3-NEXT: jne .LBB236_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xchg ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -14968,6 +17991,35 @@ define void @atomic128_add_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_add_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection566: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection567: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB237_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection568: +; HASWELL-O3-NEXT: addq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection569: +; HASWELL-O3-NEXT: adcq $0, %rcx +; HASWELL-O3-NEXT: .Lpcsection570: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection571: +; HASWELL-O3-NEXT: jne .LBB237_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw add ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15101,6 +18153,35 @@ define void @atomic128_sub_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_sub_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection572: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection573: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB238_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection574: +; HASWELL-O3-NEXT: addq $-42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection575: +; HASWELL-O3-NEXT: adcq $-1, %rcx +; HASWELL-O3-NEXT: .Lpcsection576: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection577: +; HASWELL-O3-NEXT: jne .LBB238_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw sub ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15233,6 +18314,34 @@ define void @atomic128_and_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_and_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection578: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection579: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB239_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection580: +; HASWELL-O3-NEXT: andl $42, %ebx +; HASWELL-O3-NEXT: .Lpcsection581: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection582: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection583: +; HASWELL-O3-NEXT: jne .LBB239_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw and ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15358,6 +18467,33 @@ define void @atomic128_or_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_or_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection584: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection585: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB240_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection586: +; HASWELL-O3-NEXT: orq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection587: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection588: +; HASWELL-O3-NEXT: jne .LBB240_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw or ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15483,6 +18619,33 @@ define void @atomic128_xor_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_xor_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection589: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection590: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB241_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movq %rax, %rbx +; HASWELL-O3-NEXT: .Lpcsection591: +; HASWELL-O3-NEXT: xorq $42, %rbx +; HASWELL-O3-NEXT: movq %rdx, %rcx +; HASWELL-O3-NEXT: .Lpcsection592: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection593: +; HASWELL-O3-NEXT: jne .LBB241_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw xor ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15623,6 +18786,36 @@ define void @atomic128_nand_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_nand_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection594: +; HASWELL-O3-NEXT: movq (%rdi), %rax +; HASWELL-O3-NEXT: .Lpcsection595: +; HASWELL-O3-NEXT: movq 8(%rdi), %rdx +; HASWELL-O3-NEXT: .Lpcsection596: +; HASWELL-O3-NEXT: movq $-1, %rcx +; HASWELL-O3-NEXT: .p2align 4 +; HASWELL-O3-NEXT: .LBB242_1: # %atomicrmw.start +; HASWELL-O3-NEXT: # =>This Inner Loop Header: Depth=1 +; HASWELL-O3-NEXT: movl %eax, %ebx +; HASWELL-O3-NEXT: .Lpcsection597: +; HASWELL-O3-NEXT: notl %ebx +; HASWELL-O3-NEXT: .Lpcsection598: +; HASWELL-O3-NEXT: orq $-43, %rbx +; HASWELL-O3-NEXT: .Lpcsection599: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection600: +; HASWELL-O3-NEXT: jne .LBB242_1 +; HASWELL-O3-NEXT: # %bb.2: # %atomicrmw.end +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = atomicrmw nand ptr %a, i128 42 seq_cst, align 16, !pcsections !0 @@ -15781,6 +18974,43 @@ define void @atomic128_cas_monotonic(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_cas_monotonic: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection601: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection602: +; HASWELL-O3-NEXT: movl $1, %ebx +; HASWELL-O3-NEXT: .Lpcsection603: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection604: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection605: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection606: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection607: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection608: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection609: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection610: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection611: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection612: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection613: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i128 42, i128 1 monotonic monotonic, align 16, !pcsections !0 @@ -15941,6 +19171,43 @@ define void @atomic128_cas_acquire(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_cas_acquire: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection614: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection615: +; HASWELL-O3-NEXT: movl $1, %ebx +; HASWELL-O3-NEXT: .Lpcsection616: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection617: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection618: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection619: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection620: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection621: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection622: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection623: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection624: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection625: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection626: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i128 42, i128 1 acquire monotonic, align 16, !pcsections !0 @@ -16101,6 +19368,43 @@ define void @atomic128_cas_release(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_cas_release: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection627: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection628: +; HASWELL-O3-NEXT: movl $1, %ebx +; HASWELL-O3-NEXT: .Lpcsection629: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection630: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection631: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection632: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection633: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection634: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection635: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection636: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection637: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection638: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection639: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i128 42, i128 1 release monotonic, align 16, !pcsections !0 @@ -16261,6 +19565,43 @@ define void @atomic128_cas_acq_rel(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_cas_acq_rel: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection640: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection641: +; HASWELL-O3-NEXT: movl $1, %ebx +; HASWELL-O3-NEXT: .Lpcsection642: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection643: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection644: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection645: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection646: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection647: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection648: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection649: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection650: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection651: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection652: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: movq $1, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i128 42, i128 1 acq_rel monotonic, align 16, !pcsections !0 @@ -16421,6 +19762,43 @@ define void @atomic128_cas_seq_cst(ptr %a) { ; O3-NEXT: popq %rbx ; O3-NEXT: .cfi_def_cfa_offset 8 ; O3-NEXT: retq +; +; HASWELL-O3-LABEL: atomic128_cas_seq_cst: +; HASWELL-O3: # %bb.0: # %entry +; HASWELL-O3-NEXT: pushq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 16 +; HASWELL-O3-NEXT: .cfi_offset %rbx, -16 +; HASWELL-O3-NEXT: movq foo(%rip), %rax +; HASWELL-O3-NEXT: .Lpcsection653: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection654: +; HASWELL-O3-NEXT: movl $1, %ebx +; HASWELL-O3-NEXT: .Lpcsection655: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection656: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection657: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection658: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection659: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection660: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection661: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: .Lpcsection662: +; HASWELL-O3-NEXT: movl $42, %eax +; HASWELL-O3-NEXT: .Lpcsection663: +; HASWELL-O3-NEXT: xorl %edx, %edx +; HASWELL-O3-NEXT: .Lpcsection664: +; HASWELL-O3-NEXT: xorl %ecx, %ecx +; HASWELL-O3-NEXT: .Lpcsection665: +; HASWELL-O3-NEXT: lock cmpxchg16b (%rdi) +; HASWELL-O3-NEXT: movq $3, foo(%rip) +; HASWELL-O3-NEXT: popq %rbx +; HASWELL-O3-NEXT: .cfi_def_cfa_offset 8 +; HASWELL-O3-NEXT: retq entry: load volatile i64, ptr @foo, align 8 %x = cmpxchg ptr %a, i128 42, i128 1 seq_cst monotonic, align 16, !pcsections !0 diff --git a/llvm/test/CodeGen/X86/popcnt.ll b/llvm/test/CodeGen/X86/popcnt.ll index 35c7c0e09f39..3004b8b72fcc 100644 --- a/llvm/test/CodeGen/X86/popcnt.ll +++ b/llvm/test/CodeGen/X86/popcnt.ll @@ -340,84 +340,87 @@ define i64 @cnt64(i64 %x) nounwind readnone { define i128 @cnt128(i128 %x) nounwind readnone { ; X86-NOSSE-LABEL: cnt128: ; X86-NOSSE: # %bb.0: -; X86-NOSSE-NEXT: pushl %ebx +; X86-NOSSE-NEXT: pushl %ebp +; X86-NOSSE-NEXT: movl %esp, %ebp ; X86-NOSSE-NEXT: pushl %edi ; X86-NOSSE-NEXT: pushl %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: shrl %ebx -; X86-NOSSE-NEXT: andl $1431655765, %ebx # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %ebx, %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: andl $858993459, %ebx # imm = 0x33333333 -; X86-NOSSE-NEXT: shrl $2, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %ebx, %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: shrl $4, %ebx -; X86-NOSSE-NEXT: addl %edi, %ebx -; X86-NOSSE-NEXT: andl $252645135, %ebx # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %ebx, %edi # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: shrl %ebx -; X86-NOSSE-NEXT: andl $1431655765, %ebx # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %ebx, %esi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: andl $858993459, %ebx # imm = 0x33333333 +; X86-NOSSE-NEXT: andl $-16, %esp +; X86-NOSSE-NEXT: movl 24(%ebp), %eax +; X86-NOSSE-NEXT: movl 32(%ebp), %ecx +; X86-NOSSE-NEXT: movl 36(%ebp), %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl %edx +; X86-NOSSE-NEXT: andl $1431655765, %edx # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 ; X86-NOSSE-NEXT: shrl $2, %esi ; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %ebx, %esi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: shrl $4, %ebx -; X86-NOSSE-NEXT: addl %esi, %ebx -; X86-NOSSE-NEXT: andl $252645135, %ebx # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %ebx, %esi # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %esi -; X86-NOSSE-NEXT: addl %edi, %esi -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: shrl %edi -; X86-NOSSE-NEXT: andl $1431655765, %edi # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %edi, %edx -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 -; X86-NOSSE-NEXT: shrl $2, %edx -; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %edi, %edx -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: shrl $4, %edi -; X86-NOSSE-NEXT: addl %edx, %edi -; X86-NOSSE-NEXT: andl $252645135, %edi # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %edi, %edx # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edx -; X86-NOSSE-NEXT: movl %ecx, %edi -; X86-NOSSE-NEXT: shrl %edi -; X86-NOSSE-NEXT: andl $1431655765, %edi # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %edi, %ecx -; X86-NOSSE-NEXT: movl %ecx, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl $4, %edx +; X86-NOSSE-NEXT: addl %esi, %edx +; X86-NOSSE-NEXT: movl %ecx, %esi +; X86-NOSSE-NEXT: shrl %esi +; X86-NOSSE-NEXT: andl $1431655765, %esi # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %esi, %ecx +; X86-NOSSE-NEXT: movl %ecx, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 ; X86-NOSSE-NEXT: shrl $2, %ecx ; X86-NOSSE-NEXT: andl $858993459, %ecx # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %edi, %ecx +; X86-NOSSE-NEXT: addl %esi, %ecx ; X86-NOSSE-NEXT: movl %ecx, %edi ; X86-NOSSE-NEXT: shrl $4, %edi ; X86-NOSSE-NEXT: addl %ecx, %edi +; X86-NOSSE-NEXT: movl 28(%ebp), %esi +; X86-NOSSE-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %edx, %edx # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edx ; X86-NOSSE-NEXT: andl $252645135, %edi # imm = 0xF0F0F0F ; X86-NOSSE-NEXT: imull $16843009, %edi, %ecx # imm = 0x1010101 ; X86-NOSSE-NEXT: shrl $24, %ecx ; X86-NOSSE-NEXT: addl %edx, %ecx -; X86-NOSSE-NEXT: addl %esi, %ecx -; X86-NOSSE-NEXT: movl %ecx, (%eax) +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl %edx +; X86-NOSSE-NEXT: andl $1431655765, %edx # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 +; X86-NOSSE-NEXT: shrl $2, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl $4, %edx +; X86-NOSSE-NEXT: addl %esi, %edx +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: shrl %esi +; X86-NOSSE-NEXT: andl $1431655765, %esi # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %esi, %eax +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 +; X86-NOSSE-NEXT: shrl $2, %eax +; X86-NOSSE-NEXT: andl $858993459, %eax # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %esi, %eax +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: shrl $4, %esi +; X86-NOSSE-NEXT: addl %eax, %esi +; X86-NOSSE-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %edx, %eax # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %eax +; X86-NOSSE-NEXT: andl $252645135, %esi # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %esi, %edx # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edx +; X86-NOSSE-NEXT: addl %eax, %edx +; X86-NOSSE-NEXT: addl %ecx, %edx +; X86-NOSSE-NEXT: movl 8(%ebp), %eax +; X86-NOSSE-NEXT: movl %edx, (%eax) ; X86-NOSSE-NEXT: movl $0, 12(%eax) ; X86-NOSSE-NEXT: movl $0, 8(%eax) ; X86-NOSSE-NEXT: movl $0, 4(%eax) +; X86-NOSSE-NEXT: leal -8(%ebp), %esp ; X86-NOSSE-NEXT: popl %esi ; X86-NOSSE-NEXT: popl %edi -; X86-NOSSE-NEXT: popl %ebx +; X86-NOSSE-NEXT: popl %ebp ; X86-NOSSE-NEXT: retl $4 ; ; X64-BASE-LABEL: cnt128: @@ -462,20 +465,26 @@ define i128 @cnt128(i128 %x) nounwind readnone { ; ; X86-POPCNT-LABEL: cnt128: ; X86-POPCNT: # %bb.0: +; X86-POPCNT-NEXT: pushl %ebp +; X86-POPCNT-NEXT: movl %esp, %ebp ; X86-POPCNT-NEXT: pushl %esi -; X86-POPCNT-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %edx +; X86-POPCNT-NEXT: andl $-16, %esp +; X86-POPCNT-NEXT: subl $16, %esp +; X86-POPCNT-NEXT: movl 8(%ebp), %eax +; X86-POPCNT-NEXT: popcntl 36(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 32(%ebp), %edx ; X86-POPCNT-NEXT: addl %ecx, %edx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %esi +; X86-POPCNT-NEXT: popcntl 28(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 24(%ebp), %esi ; X86-POPCNT-NEXT: addl %ecx, %esi ; X86-POPCNT-NEXT: addl %edx, %esi ; X86-POPCNT-NEXT: movl %esi, (%eax) ; X86-POPCNT-NEXT: movl $0, 12(%eax) ; X86-POPCNT-NEXT: movl $0, 8(%eax) ; X86-POPCNT-NEXT: movl $0, 4(%eax) +; X86-POPCNT-NEXT: leal -4(%ebp), %esp ; X86-POPCNT-NEXT: popl %esi +; X86-POPCNT-NEXT: popl %ebp ; X86-POPCNT-NEXT: retl $4 ; ; X64-POPCNT-LABEL: cnt128: @@ -522,7 +531,11 @@ define i128 @cnt128(i128 %x) nounwind readnone { ; ; X86-SSE2-LABEL: cnt128: ; X86-SSE2: # %bb.0: -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSE2-NEXT: pushl %ebp +; X86-SSE2-NEXT: movl %esp, %ebp +; X86-SSE2-NEXT: andl $-16, %esp +; X86-SSE2-NEXT: subl $16, %esp +; X86-SSE2-NEXT: movl 8(%ebp), %eax ; X86-SSE2-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSE2-NEXT: movdqa %xmm2, %xmm0 ; X86-SSE2-NEXT: psrlw $1, %xmm0 @@ -564,11 +577,17 @@ define i128 @cnt128(i128 %x) nounwind readnone { ; X86-SSE2-NEXT: movl $0, 12(%eax) ; X86-SSE2-NEXT: movl $0, 8(%eax) ; X86-SSE2-NEXT: movl $0, 4(%eax) +; X86-SSE2-NEXT: movl %ebp, %esp +; X86-SSE2-NEXT: popl %ebp ; X86-SSE2-NEXT: retl $4 ; ; X86-SSSE3-LABEL: cnt128: ; X86-SSSE3: # %bb.0: -; X86-SSSE3-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSSE3-NEXT: pushl %ebp +; X86-SSSE3-NEXT: movl %esp, %ebp +; X86-SSSE3-NEXT: andl $-16, %esp +; X86-SSSE3-NEXT: subl $16, %esp +; X86-SSSE3-NEXT: movl 8(%ebp), %eax ; X86-SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] ; X86-SSSE3-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSSE3-NEXT: movdqa %xmm2, %xmm3 @@ -600,6 +619,8 @@ define i128 @cnt128(i128 %x) nounwind readnone { ; X86-SSSE3-NEXT: movl $0, 12(%eax) ; X86-SSSE3-NEXT: movl $0, 8(%eax) ; X86-SSSE3-NEXT: movl $0, 4(%eax) +; X86-SSSE3-NEXT: movl %ebp, %esp +; X86-SSSE3-NEXT: popl %ebp ; X86-SSSE3-NEXT: retl $4 %cnt = tail call i128 @llvm.ctpop.i128(i128 %x) ret i128 %cnt @@ -928,87 +949,92 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; X86-NOSSE-LABEL: cnt128_optsize: ; X86-NOSSE: # %bb.0: ; X86-NOSSE-NEXT: pushl %ebp +; X86-NOSSE-NEXT: movl %esp, %ebp ; X86-NOSSE-NEXT: pushl %ebx ; X86-NOSSE-NEXT: pushl %edi ; X86-NOSSE-NEXT: pushl %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NOSSE-NEXT: movl %ebx, %ecx -; X86-NOSSE-NEXT: shrl %ecx -; X86-NOSSE-NEXT: movl $1431655765, %edi # imm = 0x55555555 -; X86-NOSSE-NEXT: andl %edi, %ecx -; X86-NOSSE-NEXT: subl %ecx, %ebx +; X86-NOSSE-NEXT: andl $-16, %esp +; X86-NOSSE-NEXT: subl $16, %esp +; X86-NOSSE-NEXT: movl 32(%ebp), %edx +; X86-NOSSE-NEXT: movl 36(%ebp), %esi +; X86-NOSSE-NEXT: movl %esi, %eax +; X86-NOSSE-NEXT: shrl %eax +; X86-NOSSE-NEXT: movl $1431655765, %ecx # imm = 0x55555555 +; X86-NOSSE-NEXT: andl %ecx, %eax +; X86-NOSSE-NEXT: subl %eax, %esi ; X86-NOSSE-NEXT: movl $858993459, %ecx # imm = 0x33333333 -; X86-NOSSE-NEXT: movl %ebx, %ebp -; X86-NOSSE-NEXT: andl %ecx, %ebp +; X86-NOSSE-NEXT: movl %esi, %edi +; X86-NOSSE-NEXT: andl %ecx, %edi +; X86-NOSSE-NEXT: shrl $2, %esi +; X86-NOSSE-NEXT: andl %ecx, %esi +; X86-NOSSE-NEXT: addl %edi, %esi +; X86-NOSSE-NEXT: movl %esi, %edi +; X86-NOSSE-NEXT: shrl $4, %edi +; X86-NOSSE-NEXT: addl %esi, %edi +; X86-NOSSE-NEXT: movl %edx, %esi +; X86-NOSSE-NEXT: shrl %esi +; X86-NOSSE-NEXT: movl $1431655765, %eax # imm = 0x55555555 +; X86-NOSSE-NEXT: andl %eax, %esi +; X86-NOSSE-NEXT: subl %esi, %edx +; X86-NOSSE-NEXT: movl %edx, %esi +; X86-NOSSE-NEXT: andl %ecx, %esi +; X86-NOSSE-NEXT: shrl $2, %edx +; X86-NOSSE-NEXT: andl %ecx, %edx +; X86-NOSSE-NEXT: addl %esi, %edx +; X86-NOSSE-NEXT: movl %edx, %ebx +; X86-NOSSE-NEXT: shrl $4, %ebx +; X86-NOSSE-NEXT: addl %edx, %ebx +; X86-NOSSE-NEXT: movl $252645135, %edx # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: andl %edx, %edi +; X86-NOSSE-NEXT: imull $16843009, %edi, %edi # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edi +; X86-NOSSE-NEXT: andl %edx, %ebx +; X86-NOSSE-NEXT: imull $16843009, %ebx, %edx # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edx +; X86-NOSSE-NEXT: addl %edi, %edx +; X86-NOSSE-NEXT: movl 28(%ebp), %ebx +; X86-NOSSE-NEXT: movl %ebx, %edi +; X86-NOSSE-NEXT: shrl %edi +; X86-NOSSE-NEXT: andl %eax, %edi +; X86-NOSSE-NEXT: subl %edi, %ebx +; X86-NOSSE-NEXT: movl %ebx, %edi +; X86-NOSSE-NEXT: andl %ecx, %edi ; X86-NOSSE-NEXT: shrl $2, %ebx ; X86-NOSSE-NEXT: andl %ecx, %ebx -; X86-NOSSE-NEXT: addl %ebp, %ebx -; X86-NOSSE-NEXT: movl %ebx, %ebp -; X86-NOSSE-NEXT: shrl $4, %ebp -; X86-NOSSE-NEXT: addl %ebx, %ebp +; X86-NOSSE-NEXT: addl %edi, %ebx +; X86-NOSSE-NEXT: movl %ebx, %edi +; X86-NOSSE-NEXT: shrl $4, %edi +; X86-NOSSE-NEXT: addl %ebx, %edi +; X86-NOSSE-NEXT: movl 24(%ebp), %eax ; X86-NOSSE-NEXT: movl %eax, %ebx ; X86-NOSSE-NEXT: shrl %ebx -; X86-NOSSE-NEXT: andl %edi, %ebx +; X86-NOSSE-NEXT: movl $1431655765, %esi # imm = 0x55555555 +; X86-NOSSE-NEXT: andl %esi, %ebx ; X86-NOSSE-NEXT: subl %ebx, %eax ; X86-NOSSE-NEXT: movl %eax, %ebx ; X86-NOSSE-NEXT: andl %ecx, %ebx ; X86-NOSSE-NEXT: shrl $2, %eax ; X86-NOSSE-NEXT: andl %ecx, %eax ; X86-NOSSE-NEXT: addl %ebx, %eax -; X86-NOSSE-NEXT: movl %eax, %edi -; X86-NOSSE-NEXT: shrl $4, %edi -; X86-NOSSE-NEXT: addl %eax, %edi -; X86-NOSSE-NEXT: movl $252645135, %ebx # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: andl %ebx, %ebp -; X86-NOSSE-NEXT: imull $16843009, %ebp, %eax # imm = 0x1010101 +; X86-NOSSE-NEXT: movl %eax, %ecx +; X86-NOSSE-NEXT: shrl $4, %ecx +; X86-NOSSE-NEXT: addl %eax, %ecx +; X86-NOSSE-NEXT: movl $252645135, %eax # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: andl %eax, %edi +; X86-NOSSE-NEXT: andl %eax, %ecx +; X86-NOSSE-NEXT: imull $16843009, %edi, %eax # imm = 0x1010101 ; X86-NOSSE-NEXT: shrl $24, %eax -; X86-NOSSE-NEXT: andl %ebx, %edi -; X86-NOSSE-NEXT: imull $16843009, %edi, %edi # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edi -; X86-NOSSE-NEXT: addl %eax, %edi -; X86-NOSSE-NEXT: movl %esi, %eax -; X86-NOSSE-NEXT: shrl %eax -; X86-NOSSE-NEXT: movl $1431655765, %ebp # imm = 0x55555555 -; X86-NOSSE-NEXT: andl %ebp, %eax -; X86-NOSSE-NEXT: subl %eax, %esi -; X86-NOSSE-NEXT: movl %esi, %eax -; X86-NOSSE-NEXT: andl %ecx, %eax -; X86-NOSSE-NEXT: shrl $2, %esi -; X86-NOSSE-NEXT: andl %ecx, %esi -; X86-NOSSE-NEXT: addl %eax, %esi -; X86-NOSSE-NEXT: movl %esi, %ebp -; X86-NOSSE-NEXT: shrl $4, %ebp -; X86-NOSSE-NEXT: addl %esi, %ebp -; X86-NOSSE-NEXT: movl %edx, %eax -; X86-NOSSE-NEXT: shrl %eax -; X86-NOSSE-NEXT: movl $1431655765, %esi # imm = 0x55555555 -; X86-NOSSE-NEXT: andl %esi, %eax -; X86-NOSSE-NEXT: subl %eax, %edx -; X86-NOSSE-NEXT: movl %edx, %eax -; X86-NOSSE-NEXT: andl %ecx, %eax -; X86-NOSSE-NEXT: shrl $2, %edx -; X86-NOSSE-NEXT: andl %ecx, %edx -; X86-NOSSE-NEXT: addl %eax, %edx -; X86-NOSSE-NEXT: movl %edx, %eax -; X86-NOSSE-NEXT: shrl $4, %eax -; X86-NOSSE-NEXT: addl %edx, %eax -; X86-NOSSE-NEXT: andl %ebx, %ebp -; X86-NOSSE-NEXT: andl %ebx, %eax -; X86-NOSSE-NEXT: imull $16843009, %ebp, %ecx # imm = 0x1010101 +; X86-NOSSE-NEXT: imull $16843009, %ecx, %ecx # imm = 0x1010101 ; X86-NOSSE-NEXT: shrl $24, %ecx -; X86-NOSSE-NEXT: imull $16843009, %eax, %edx # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edx -; X86-NOSSE-NEXT: addl %ecx, %edx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NOSSE-NEXT: addl %edi, %edx -; X86-NOSSE-NEXT: xorl %ecx, %ecx -; X86-NOSSE-NEXT: movl %ecx, 12(%eax) -; X86-NOSSE-NEXT: movl %ecx, 8(%eax) -; X86-NOSSE-NEXT: movl %ecx, 4(%eax) -; X86-NOSSE-NEXT: movl %edx, (%eax) +; X86-NOSSE-NEXT: addl %eax, %ecx +; X86-NOSSE-NEXT: movl 8(%ebp), %eax +; X86-NOSSE-NEXT: addl %edx, %ecx +; X86-NOSSE-NEXT: xorl %edx, %edx +; X86-NOSSE-NEXT: movl %edx, 12(%eax) +; X86-NOSSE-NEXT: movl %edx, 8(%eax) +; X86-NOSSE-NEXT: movl %edx, 4(%eax) +; X86-NOSSE-NEXT: movl %ecx, (%eax) +; X86-NOSSE-NEXT: leal -12(%ebp), %esp ; X86-NOSSE-NEXT: popl %esi ; X86-NOSSE-NEXT: popl %edi ; X86-NOSSE-NEXT: popl %ebx @@ -1057,13 +1083,17 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; ; X86-POPCNT-LABEL: cnt128_optsize: ; X86-POPCNT: # %bb.0: +; X86-POPCNT-NEXT: pushl %ebp +; X86-POPCNT-NEXT: movl %esp, %ebp ; X86-POPCNT-NEXT: pushl %esi -; X86-POPCNT-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %edx +; X86-POPCNT-NEXT: andl $-16, %esp +; X86-POPCNT-NEXT: subl $16, %esp +; X86-POPCNT-NEXT: movl 8(%ebp), %eax +; X86-POPCNT-NEXT: popcntl 36(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 32(%ebp), %edx ; X86-POPCNT-NEXT: addl %ecx, %edx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %esi +; X86-POPCNT-NEXT: popcntl 28(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 24(%ebp), %esi ; X86-POPCNT-NEXT: addl %ecx, %esi ; X86-POPCNT-NEXT: addl %edx, %esi ; X86-POPCNT-NEXT: xorl %ecx, %ecx @@ -1071,7 +1101,9 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; X86-POPCNT-NEXT: movl %ecx, 8(%eax) ; X86-POPCNT-NEXT: movl %ecx, 4(%eax) ; X86-POPCNT-NEXT: movl %esi, (%eax) +; X86-POPCNT-NEXT: leal -4(%ebp), %esp ; X86-POPCNT-NEXT: popl %esi +; X86-POPCNT-NEXT: popl %ebp ; X86-POPCNT-NEXT: retl $4 ; ; X64-POPCNT-LABEL: cnt128_optsize: @@ -1118,7 +1150,11 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; ; X86-SSE2-LABEL: cnt128_optsize: ; X86-SSE2: # %bb.0: -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSE2-NEXT: pushl %ebp +; X86-SSE2-NEXT: movl %esp, %ebp +; X86-SSE2-NEXT: andl $-16, %esp +; X86-SSE2-NEXT: subl $16, %esp +; X86-SSE2-NEXT: movl 8(%ebp), %eax ; X86-SSE2-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSE2-NEXT: movdqa %xmm2, %xmm0 ; X86-SSE2-NEXT: psrlw $1, %xmm0 @@ -1161,11 +1197,17 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; X86-SSE2-NEXT: movl %ecx, 8(%eax) ; X86-SSE2-NEXT: movl %ecx, 4(%eax) ; X86-SSE2-NEXT: movl %edx, (%eax) +; X86-SSE2-NEXT: movl %ebp, %esp +; X86-SSE2-NEXT: popl %ebp ; X86-SSE2-NEXT: retl $4 ; ; X86-SSSE3-LABEL: cnt128_optsize: ; X86-SSSE3: # %bb.0: -; X86-SSSE3-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSSE3-NEXT: pushl %ebp +; X86-SSSE3-NEXT: movl %esp, %ebp +; X86-SSSE3-NEXT: andl $-16, %esp +; X86-SSSE3-NEXT: subl $16, %esp +; X86-SSSE3-NEXT: movl 8(%ebp), %eax ; X86-SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] ; X86-SSSE3-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSSE3-NEXT: movdqa %xmm2, %xmm3 @@ -1198,6 +1240,8 @@ define i128 @cnt128_optsize(i128 %x) nounwind readnone optsize { ; X86-SSSE3-NEXT: movl %ecx, 8(%eax) ; X86-SSSE3-NEXT: movl %ecx, 4(%eax) ; X86-SSSE3-NEXT: movl %edx, (%eax) +; X86-SSSE3-NEXT: movl %ebp, %esp +; X86-SSSE3-NEXT: popl %ebp ; X86-SSSE3-NEXT: retl $4 %cnt = tail call i128 @llvm.ctpop.i128(i128 %x) ret i128 %cnt @@ -1415,85 +1459,88 @@ define i64 @cnt64_pgso(i64 %x) nounwind readnone !prof !14 { define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; X86-NOSSE-LABEL: cnt128_pgso: ; X86-NOSSE: # %bb.0: -; X86-NOSSE-NEXT: pushl %ebx +; X86-NOSSE-NEXT: pushl %ebp +; X86-NOSSE-NEXT: movl %esp, %ebp ; X86-NOSSE-NEXT: pushl %edi ; X86-NOSSE-NEXT: pushl %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NOSSE-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: shrl %ebx -; X86-NOSSE-NEXT: andl $1431655765, %ebx # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %ebx, %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: andl $858993459, %ebx # imm = 0x33333333 -; X86-NOSSE-NEXT: shrl $2, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %ebx, %edi -; X86-NOSSE-NEXT: movl %edi, %ebx -; X86-NOSSE-NEXT: shrl $4, %ebx -; X86-NOSSE-NEXT: addl %edi, %ebx -; X86-NOSSE-NEXT: andl $252645135, %ebx # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %ebx, %edi # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: shrl %ebx -; X86-NOSSE-NEXT: andl $1431655765, %ebx # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %ebx, %esi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: andl $858993459, %ebx # imm = 0x33333333 +; X86-NOSSE-NEXT: andl $-16, %esp +; X86-NOSSE-NEXT: movl 24(%ebp), %eax +; X86-NOSSE-NEXT: movl 32(%ebp), %ecx +; X86-NOSSE-NEXT: movl 36(%ebp), %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl %edx +; X86-NOSSE-NEXT: andl $1431655765, %edx # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 ; X86-NOSSE-NEXT: shrl $2, %esi ; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %ebx, %esi -; X86-NOSSE-NEXT: movl %esi, %ebx -; X86-NOSSE-NEXT: shrl $4, %ebx -; X86-NOSSE-NEXT: addl %esi, %ebx -; X86-NOSSE-NEXT: andl $252645135, %ebx # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %ebx, %esi # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %esi -; X86-NOSSE-NEXT: addl %edi, %esi -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: shrl %edi -; X86-NOSSE-NEXT: andl $1431655765, %edi # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %edi, %edx -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 -; X86-NOSSE-NEXT: shrl $2, %edx -; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %edi, %edx -; X86-NOSSE-NEXT: movl %edx, %edi -; X86-NOSSE-NEXT: shrl $4, %edi -; X86-NOSSE-NEXT: addl %edx, %edi -; X86-NOSSE-NEXT: andl $252645135, %edi # imm = 0xF0F0F0F -; X86-NOSSE-NEXT: imull $16843009, %edi, %edx # imm = 0x1010101 -; X86-NOSSE-NEXT: shrl $24, %edx -; X86-NOSSE-NEXT: movl %ecx, %edi -; X86-NOSSE-NEXT: shrl %edi -; X86-NOSSE-NEXT: andl $1431655765, %edi # imm = 0x55555555 -; X86-NOSSE-NEXT: subl %edi, %ecx -; X86-NOSSE-NEXT: movl %ecx, %edi -; X86-NOSSE-NEXT: andl $858993459, %edi # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl $4, %edx +; X86-NOSSE-NEXT: addl %esi, %edx +; X86-NOSSE-NEXT: movl %ecx, %esi +; X86-NOSSE-NEXT: shrl %esi +; X86-NOSSE-NEXT: andl $1431655765, %esi # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %esi, %ecx +; X86-NOSSE-NEXT: movl %ecx, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 ; X86-NOSSE-NEXT: shrl $2, %ecx ; X86-NOSSE-NEXT: andl $858993459, %ecx # imm = 0x33333333 -; X86-NOSSE-NEXT: addl %edi, %ecx +; X86-NOSSE-NEXT: addl %esi, %ecx ; X86-NOSSE-NEXT: movl %ecx, %edi ; X86-NOSSE-NEXT: shrl $4, %edi ; X86-NOSSE-NEXT: addl %ecx, %edi +; X86-NOSSE-NEXT: movl 28(%ebp), %esi +; X86-NOSSE-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %edx, %edx # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edx ; X86-NOSSE-NEXT: andl $252645135, %edi # imm = 0xF0F0F0F ; X86-NOSSE-NEXT: imull $16843009, %edi, %ecx # imm = 0x1010101 ; X86-NOSSE-NEXT: shrl $24, %ecx ; X86-NOSSE-NEXT: addl %edx, %ecx -; X86-NOSSE-NEXT: addl %esi, %ecx -; X86-NOSSE-NEXT: xorl %edx, %edx -; X86-NOSSE-NEXT: movl %edx, 12(%eax) -; X86-NOSSE-NEXT: movl %edx, 8(%eax) -; X86-NOSSE-NEXT: movl %edx, 4(%eax) -; X86-NOSSE-NEXT: movl %ecx, (%eax) +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl %edx +; X86-NOSSE-NEXT: andl $1431655765, %edx # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: andl $858993459, %edx # imm = 0x33333333 +; X86-NOSSE-NEXT: shrl $2, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %edx, %esi +; X86-NOSSE-NEXT: movl %esi, %edx +; X86-NOSSE-NEXT: shrl $4, %edx +; X86-NOSSE-NEXT: addl %esi, %edx +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: shrl %esi +; X86-NOSSE-NEXT: andl $1431655765, %esi # imm = 0x55555555 +; X86-NOSSE-NEXT: subl %esi, %eax +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: andl $858993459, %esi # imm = 0x33333333 +; X86-NOSSE-NEXT: shrl $2, %eax +; X86-NOSSE-NEXT: andl $858993459, %eax # imm = 0x33333333 +; X86-NOSSE-NEXT: addl %esi, %eax +; X86-NOSSE-NEXT: movl %eax, %esi +; X86-NOSSE-NEXT: shrl $4, %esi +; X86-NOSSE-NEXT: addl %eax, %esi +; X86-NOSSE-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %edx, %eax # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %eax +; X86-NOSSE-NEXT: andl $252645135, %esi # imm = 0xF0F0F0F +; X86-NOSSE-NEXT: imull $16843009, %esi, %edx # imm = 0x1010101 +; X86-NOSSE-NEXT: shrl $24, %edx +; X86-NOSSE-NEXT: addl %eax, %edx +; X86-NOSSE-NEXT: movl 8(%ebp), %eax +; X86-NOSSE-NEXT: addl %ecx, %edx +; X86-NOSSE-NEXT: xorl %ecx, %ecx +; X86-NOSSE-NEXT: movl %ecx, 12(%eax) +; X86-NOSSE-NEXT: movl %ecx, 8(%eax) +; X86-NOSSE-NEXT: movl %ecx, 4(%eax) +; X86-NOSSE-NEXT: movl %edx, (%eax) +; X86-NOSSE-NEXT: leal -8(%ebp), %esp ; X86-NOSSE-NEXT: popl %esi ; X86-NOSSE-NEXT: popl %edi -; X86-NOSSE-NEXT: popl %ebx +; X86-NOSSE-NEXT: popl %ebp ; X86-NOSSE-NEXT: retl $4 ; ; X64-BASE-LABEL: cnt128_pgso: @@ -1538,13 +1585,17 @@ define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; ; X86-POPCNT-LABEL: cnt128_pgso: ; X86-POPCNT: # %bb.0: +; X86-POPCNT-NEXT: pushl %ebp +; X86-POPCNT-NEXT: movl %esp, %ebp ; X86-POPCNT-NEXT: pushl %esi -; X86-POPCNT-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %edx +; X86-POPCNT-NEXT: andl $-16, %esp +; X86-POPCNT-NEXT: subl $16, %esp +; X86-POPCNT-NEXT: movl 8(%ebp), %eax +; X86-POPCNT-NEXT: popcntl 36(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 32(%ebp), %edx ; X86-POPCNT-NEXT: addl %ecx, %edx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %ecx -; X86-POPCNT-NEXT: popcntl {{[0-9]+}}(%esp), %esi +; X86-POPCNT-NEXT: popcntl 28(%ebp), %ecx +; X86-POPCNT-NEXT: popcntl 24(%ebp), %esi ; X86-POPCNT-NEXT: addl %ecx, %esi ; X86-POPCNT-NEXT: addl %edx, %esi ; X86-POPCNT-NEXT: xorl %ecx, %ecx @@ -1552,7 +1603,9 @@ define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; X86-POPCNT-NEXT: movl %ecx, 8(%eax) ; X86-POPCNT-NEXT: movl %ecx, 4(%eax) ; X86-POPCNT-NEXT: movl %esi, (%eax) +; X86-POPCNT-NEXT: leal -4(%ebp), %esp ; X86-POPCNT-NEXT: popl %esi +; X86-POPCNT-NEXT: popl %ebp ; X86-POPCNT-NEXT: retl $4 ; ; X64-POPCNT-LABEL: cnt128_pgso: @@ -1599,7 +1652,11 @@ define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; ; X86-SSE2-LABEL: cnt128_pgso: ; X86-SSE2: # %bb.0: -; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSE2-NEXT: pushl %ebp +; X86-SSE2-NEXT: movl %esp, %ebp +; X86-SSE2-NEXT: andl $-16, %esp +; X86-SSE2-NEXT: subl $16, %esp +; X86-SSE2-NEXT: movl 8(%ebp), %eax ; X86-SSE2-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSE2-NEXT: movdqa %xmm2, %xmm0 ; X86-SSE2-NEXT: psrlw $1, %xmm0 @@ -1642,11 +1699,17 @@ define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; X86-SSE2-NEXT: movl %ecx, 8(%eax) ; X86-SSE2-NEXT: movl %ecx, 4(%eax) ; X86-SSE2-NEXT: movl %edx, (%eax) +; X86-SSE2-NEXT: movl %ebp, %esp +; X86-SSE2-NEXT: popl %ebp ; X86-SSE2-NEXT: retl $4 ; ; X86-SSSE3-LABEL: cnt128_pgso: ; X86-SSSE3: # %bb.0: -; X86-SSSE3-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-SSSE3-NEXT: pushl %ebp +; X86-SSSE3-NEXT: movl %esp, %ebp +; X86-SSSE3-NEXT: andl $-16, %esp +; X86-SSSE3-NEXT: subl $16, %esp +; X86-SSSE3-NEXT: movl 8(%ebp), %eax ; X86-SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] ; X86-SSSE3-NEXT: movq {{.*#+}} xmm2 = mem[0],zero ; X86-SSSE3-NEXT: movdqa %xmm2, %xmm3 @@ -1679,6 +1742,8 @@ define i128 @cnt128_pgso(i128 %x) nounwind readnone !prof !14 { ; X86-SSSE3-NEXT: movl %ecx, 8(%eax) ; X86-SSSE3-NEXT: movl %ecx, 4(%eax) ; X86-SSSE3-NEXT: movl %edx, (%eax) +; X86-SSSE3-NEXT: movl %ebp, %esp +; X86-SSSE3-NEXT: popl %ebp ; X86-SSSE3-NEXT: retl $4 %cnt = tail call i128 @llvm.ctpop.i128(i128 %x) ret i128 %cnt diff --git a/llvm/test/CodeGen/X86/pr46004.ll b/llvm/test/CodeGen/X86/pr46004.ll index f7c7da089c36..829d6dfceba3 100644 --- a/llvm/test/CodeGen/X86/pr46004.ll +++ b/llvm/test/CodeGen/X86/pr46004.ll @@ -6,7 +6,17 @@ define void @fuzz22357(i128 %a0) { ; X86-LABEL: fuzz22357: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp ; X86-NEXT: movb $0, (%eax) +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl ; ; X64-LABEL: fuzz22357: @@ -24,6 +34,15 @@ define void @fuzz22357(i128 %a0) { define void @fuzz22723(i128 %a0) { ; X86-LABEL: fuzz22723: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl ; ; X64-LABEL: fuzz22723: diff --git a/llvm/test/CodeGen/X86/scalar-fp-to-i32.ll b/llvm/test/CodeGen/X86/scalar-fp-to-i32.ll index 50a967e1c2a1..ce9723b3a84b 100644 --- a/llvm/test/CodeGen/X86/scalar-fp-to-i32.ll +++ b/llvm/test/CodeGen/X86/scalar-fp-to-i32.ll @@ -762,11 +762,15 @@ define i32 @x_to_s32(x86_fp80 %a) nounwind { define i32 @t_to_u32(fp128 %a) nounwind { ; X86-AVX512-WIN-LABEL: t_to_u32: ; X86-AVX512-WIN: # %bb.0: -; X86-AVX512-WIN-NEXT: subl $16, %esp -; X86-AVX512-WIN-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-AVX512-WIN-NEXT: pushl %ebp +; X86-AVX512-WIN-NEXT: movl %esp, %ebp +; X86-AVX512-WIN-NEXT: andl $-16, %esp +; X86-AVX512-WIN-NEXT: subl $32, %esp +; X86-AVX512-WIN-NEXT: vmovups 8(%ebp), %xmm0 ; X86-AVX512-WIN-NEXT: vmovups %xmm0, (%esp) ; X86-AVX512-WIN-NEXT: calll ___fixunstfsi -; X86-AVX512-WIN-NEXT: addl $16, %esp +; X86-AVX512-WIN-NEXT: movl %ebp, %esp +; X86-AVX512-WIN-NEXT: popl %ebp ; X86-AVX512-WIN-NEXT: retl ; ; X86-AVX512-LIN-LABEL: t_to_u32: @@ -797,12 +801,18 @@ define i32 @t_to_u32(fp128 %a) nounwind { ; ; X86-SSE-WIN-LABEL: t_to_u32: ; X86-SSE-WIN: # %bb.0: -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X86-SSE-WIN-NEXT: pushl %ebp +; X86-SSE-WIN-NEXT: movl %esp, %ebp +; X86-SSE-WIN-NEXT: andl $-16, %esp +; X86-SSE-WIN-NEXT: subl $16, %esp +; X86-SSE-WIN-NEXT: pushl 20(%ebp) +; X86-SSE-WIN-NEXT: pushl 16(%ebp) +; X86-SSE-WIN-NEXT: pushl 12(%ebp) +; X86-SSE-WIN-NEXT: pushl 8(%ebp) ; X86-SSE-WIN-NEXT: calll ___fixunstfsi ; X86-SSE-WIN-NEXT: addl $16, %esp +; X86-SSE-WIN-NEXT: movl %ebp, %esp +; X86-SSE-WIN-NEXT: popl %ebp ; X86-SSE-WIN-NEXT: retl ; ; X86-SSE-LIN-LABEL: t_to_u32: @@ -835,12 +845,18 @@ define i32 @t_to_u32(fp128 %a) nounwind { ; ; X87-WIN-LABEL: t_to_u32: ; X87-WIN: # %bb.0: -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X87-WIN-NEXT: pushl %ebp +; X87-WIN-NEXT: movl %esp, %ebp +; X87-WIN-NEXT: andl $-16, %esp +; X87-WIN-NEXT: subl $16, %esp +; X87-WIN-NEXT: pushl 20(%ebp) +; X87-WIN-NEXT: pushl 16(%ebp) +; X87-WIN-NEXT: pushl 12(%ebp) +; X87-WIN-NEXT: pushl 8(%ebp) ; X87-WIN-NEXT: calll ___fixunstfsi ; X87-WIN-NEXT: addl $16, %esp +; X87-WIN-NEXT: movl %ebp, %esp +; X87-WIN-NEXT: popl %ebp ; X87-WIN-NEXT: retl ; ; X87-LIN-LABEL: t_to_u32: @@ -860,11 +876,15 @@ define i32 @t_to_u32(fp128 %a) nounwind { define i32 @t_to_s32(fp128 %a) nounwind { ; X86-AVX512-WIN-LABEL: t_to_s32: ; X86-AVX512-WIN: # %bb.0: -; X86-AVX512-WIN-NEXT: subl $16, %esp -; X86-AVX512-WIN-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-AVX512-WIN-NEXT: pushl %ebp +; X86-AVX512-WIN-NEXT: movl %esp, %ebp +; X86-AVX512-WIN-NEXT: andl $-16, %esp +; X86-AVX512-WIN-NEXT: subl $32, %esp +; X86-AVX512-WIN-NEXT: vmovups 8(%ebp), %xmm0 ; X86-AVX512-WIN-NEXT: vmovups %xmm0, (%esp) ; X86-AVX512-WIN-NEXT: calll ___fixtfsi -; X86-AVX512-WIN-NEXT: addl $16, %esp +; X86-AVX512-WIN-NEXT: movl %ebp, %esp +; X86-AVX512-WIN-NEXT: popl %ebp ; X86-AVX512-WIN-NEXT: retl ; ; X86-AVX512-LIN-LABEL: t_to_s32: @@ -895,12 +915,18 @@ define i32 @t_to_s32(fp128 %a) nounwind { ; ; X86-SSE-WIN-LABEL: t_to_s32: ; X86-SSE-WIN: # %bb.0: -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X86-SSE-WIN-NEXT: pushl %ebp +; X86-SSE-WIN-NEXT: movl %esp, %ebp +; X86-SSE-WIN-NEXT: andl $-16, %esp +; X86-SSE-WIN-NEXT: subl $16, %esp +; X86-SSE-WIN-NEXT: pushl 20(%ebp) +; X86-SSE-WIN-NEXT: pushl 16(%ebp) +; X86-SSE-WIN-NEXT: pushl 12(%ebp) +; X86-SSE-WIN-NEXT: pushl 8(%ebp) ; X86-SSE-WIN-NEXT: calll ___fixtfsi ; X86-SSE-WIN-NEXT: addl $16, %esp +; X86-SSE-WIN-NEXT: movl %ebp, %esp +; X86-SSE-WIN-NEXT: popl %ebp ; X86-SSE-WIN-NEXT: retl ; ; X86-SSE-LIN-LABEL: t_to_s32: @@ -933,12 +959,18 @@ define i32 @t_to_s32(fp128 %a) nounwind { ; ; X87-WIN-LABEL: t_to_s32: ; X87-WIN: # %bb.0: -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X87-WIN-NEXT: pushl %ebp +; X87-WIN-NEXT: movl %esp, %ebp +; X87-WIN-NEXT: andl $-16, %esp +; X87-WIN-NEXT: subl $16, %esp +; X87-WIN-NEXT: pushl 20(%ebp) +; X87-WIN-NEXT: pushl 16(%ebp) +; X87-WIN-NEXT: pushl 12(%ebp) +; X87-WIN-NEXT: pushl 8(%ebp) ; X87-WIN-NEXT: calll ___fixtfsi ; X87-WIN-NEXT: addl $16, %esp +; X87-WIN-NEXT: movl %ebp, %esp +; X87-WIN-NEXT: popl %ebp ; X87-WIN-NEXT: retl ; ; X87-LIN-LABEL: t_to_s32: diff --git a/llvm/test/CodeGen/X86/scalar-fp-to-i64.ll b/llvm/test/CodeGen/X86/scalar-fp-to-i64.ll index f516db8b30ff..3287869f2c60 100644 --- a/llvm/test/CodeGen/X86/scalar-fp-to-i64.ll +++ b/llvm/test/CodeGen/X86/scalar-fp-to-i64.ll @@ -1417,11 +1417,15 @@ define i64 @x_to_s64(x86_fp80 %a) nounwind { define i64 @t_to_u64(fp128 %a) nounwind { ; X86-AVX512-WIN-LABEL: t_to_u64: ; X86-AVX512-WIN: # %bb.0: -; X86-AVX512-WIN-NEXT: subl $16, %esp -; X86-AVX512-WIN-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-AVX512-WIN-NEXT: pushl %ebp +; X86-AVX512-WIN-NEXT: movl %esp, %ebp +; X86-AVX512-WIN-NEXT: andl $-16, %esp +; X86-AVX512-WIN-NEXT: subl $32, %esp +; X86-AVX512-WIN-NEXT: vmovups 8(%ebp), %xmm0 ; X86-AVX512-WIN-NEXT: vmovups %xmm0, (%esp) ; X86-AVX512-WIN-NEXT: calll ___fixunstfdi -; X86-AVX512-WIN-NEXT: addl $16, %esp +; X86-AVX512-WIN-NEXT: movl %ebp, %esp +; X86-AVX512-WIN-NEXT: popl %ebp ; X86-AVX512-WIN-NEXT: retl ; ; X86-AVX512-LIN-LABEL: t_to_u64: @@ -1452,12 +1456,18 @@ define i64 @t_to_u64(fp128 %a) nounwind { ; ; X86-SSE-WIN-LABEL: t_to_u64: ; X86-SSE-WIN: # %bb.0: -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X86-SSE-WIN-NEXT: pushl %ebp +; X86-SSE-WIN-NEXT: movl %esp, %ebp +; X86-SSE-WIN-NEXT: andl $-16, %esp +; X86-SSE-WIN-NEXT: subl $16, %esp +; X86-SSE-WIN-NEXT: pushl 20(%ebp) +; X86-SSE-WIN-NEXT: pushl 16(%ebp) +; X86-SSE-WIN-NEXT: pushl 12(%ebp) +; X86-SSE-WIN-NEXT: pushl 8(%ebp) ; X86-SSE-WIN-NEXT: calll ___fixunstfdi ; X86-SSE-WIN-NEXT: addl $16, %esp +; X86-SSE-WIN-NEXT: movl %ebp, %esp +; X86-SSE-WIN-NEXT: popl %ebp ; X86-SSE-WIN-NEXT: retl ; ; X86-SSE-LIN-LABEL: t_to_u64: @@ -1490,12 +1500,18 @@ define i64 @t_to_u64(fp128 %a) nounwind { ; ; X87-WIN-LABEL: t_to_u64: ; X87-WIN: # %bb.0: -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X87-WIN-NEXT: pushl %ebp +; X87-WIN-NEXT: movl %esp, %ebp +; X87-WIN-NEXT: andl $-16, %esp +; X87-WIN-NEXT: subl $16, %esp +; X87-WIN-NEXT: pushl 20(%ebp) +; X87-WIN-NEXT: pushl 16(%ebp) +; X87-WIN-NEXT: pushl 12(%ebp) +; X87-WIN-NEXT: pushl 8(%ebp) ; X87-WIN-NEXT: calll ___fixunstfdi ; X87-WIN-NEXT: addl $16, %esp +; X87-WIN-NEXT: movl %ebp, %esp +; X87-WIN-NEXT: popl %ebp ; X87-WIN-NEXT: retl ; ; X87-LIN-LABEL: t_to_u64: @@ -1515,11 +1531,15 @@ define i64 @t_to_u64(fp128 %a) nounwind { define i64 @t_to_s64(fp128 %a) nounwind { ; X86-AVX512-WIN-LABEL: t_to_s64: ; X86-AVX512-WIN: # %bb.0: -; X86-AVX512-WIN-NEXT: subl $16, %esp -; X86-AVX512-WIN-NEXT: vmovups {{[0-9]+}}(%esp), %xmm0 +; X86-AVX512-WIN-NEXT: pushl %ebp +; X86-AVX512-WIN-NEXT: movl %esp, %ebp +; X86-AVX512-WIN-NEXT: andl $-16, %esp +; X86-AVX512-WIN-NEXT: subl $32, %esp +; X86-AVX512-WIN-NEXT: vmovups 8(%ebp), %xmm0 ; X86-AVX512-WIN-NEXT: vmovups %xmm0, (%esp) ; X86-AVX512-WIN-NEXT: calll ___fixtfdi -; X86-AVX512-WIN-NEXT: addl $16, %esp +; X86-AVX512-WIN-NEXT: movl %ebp, %esp +; X86-AVX512-WIN-NEXT: popl %ebp ; X86-AVX512-WIN-NEXT: retl ; ; X86-AVX512-LIN-LABEL: t_to_s64: @@ -1550,12 +1570,18 @@ define i64 @t_to_s64(fp128 %a) nounwind { ; ; X86-SSE-WIN-LABEL: t_to_s64: ; X86-SSE-WIN: # %bb.0: -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X86-SSE-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X86-SSE-WIN-NEXT: pushl %ebp +; X86-SSE-WIN-NEXT: movl %esp, %ebp +; X86-SSE-WIN-NEXT: andl $-16, %esp +; X86-SSE-WIN-NEXT: subl $16, %esp +; X86-SSE-WIN-NEXT: pushl 20(%ebp) +; X86-SSE-WIN-NEXT: pushl 16(%ebp) +; X86-SSE-WIN-NEXT: pushl 12(%ebp) +; X86-SSE-WIN-NEXT: pushl 8(%ebp) ; X86-SSE-WIN-NEXT: calll ___fixtfdi ; X86-SSE-WIN-NEXT: addl $16, %esp +; X86-SSE-WIN-NEXT: movl %ebp, %esp +; X86-SSE-WIN-NEXT: popl %ebp ; X86-SSE-WIN-NEXT: retl ; ; X86-SSE-LIN-LABEL: t_to_s64: @@ -1588,12 +1614,18 @@ define i64 @t_to_s64(fp128 %a) nounwind { ; ; X87-WIN-LABEL: t_to_s64: ; X87-WIN: # %bb.0: -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) -; X87-WIN-NEXT: pushl {{[0-9]+}}(%esp) +; X87-WIN-NEXT: pushl %ebp +; X87-WIN-NEXT: movl %esp, %ebp +; X87-WIN-NEXT: andl $-16, %esp +; X87-WIN-NEXT: subl $16, %esp +; X87-WIN-NEXT: pushl 20(%ebp) +; X87-WIN-NEXT: pushl 16(%ebp) +; X87-WIN-NEXT: pushl 12(%ebp) +; X87-WIN-NEXT: pushl 8(%ebp) ; X87-WIN-NEXT: calll ___fixtfdi ; X87-WIN-NEXT: addl $16, %esp +; X87-WIN-NEXT: movl %ebp, %esp +; X87-WIN-NEXT: popl %ebp ; X87-WIN-NEXT: retl ; ; X87-LIN-LABEL: t_to_s64: diff --git a/llvm/test/CodeGen/X86/scmp.ll b/llvm/test/CodeGen/X86/scmp.ll index 874913629e9e..8a287229a1cb 100644 --- a/llvm/test/CodeGen/X86/scmp.ll +++ b/llvm/test/CodeGen/X86/scmp.ll @@ -118,30 +118,33 @@ define i8 @scmp.8.128(i128 %x, i128 %y) nounwind { ; X86-LABEL: scmp.8.128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: cmpl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl %ebx, %ebp -; X86-NEXT: sbbl %edx, %ebp -; X86-NEXT: movl %ecx, %ebp -; X86-NEXT: sbbl %eax, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl %esi, %ecx -; X86-NEXT: sbbl %ebp, %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 24(%ebp), %ecx +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %edi +; X86-NEXT: cmpl %ecx, 8(%ebp) +; X86-NEXT: sbbl %eax, %edi +; X86-NEXT: movl 32(%ebp), %edi +; X86-NEXT: movl 16(%ebp), %ebx +; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl 36(%ebp), %ebx +; X86-NEXT: movl 20(%ebp), %edx +; X86-NEXT: movl %edx, %ecx +; X86-NEXT: sbbl %ebx, %ecx ; X86-NEXT: setl %cl -; X86-NEXT: cmpl %edi, {{[0-9]+}}(%esp) -; X86-NEXT: sbbl %ebx, %edx -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %eax -; X86-NEXT: sbbl %esi, %ebp +; X86-NEXT: movl 24(%ebp), %esi +; X86-NEXT: cmpl 8(%ebp), %esi +; X86-NEXT: sbbl 12(%ebp), %eax +; X86-NEXT: sbbl 16(%ebp), %edi +; X86-NEXT: sbbl %edx, %ebx ; X86-NEXT: setl %al ; X86-NEXT: subb %cl, %al +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/sdiv_fix.ll b/llvm/test/CodeGen/X86/sdiv_fix.ll index 4925f8bc6c8b..392bc83d9d5d 100644 --- a/llvm/test/CodeGen/X86/sdiv_fix.ll +++ b/llvm/test/CodeGen/X86/sdiv_fix.ll @@ -307,69 +307,70 @@ define i64 @func5(i64 %x, i64 %y) nounwind { ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp -; X86-NEXT: subl $80, %esp +; X86-NEXT: subl $112, %esp ; X86-NEXT: movl 8(%ebp), %ecx -; X86-NEXT: movl 12(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %edi +; X86-NEXT: movl 16(%ebp), %eax ; X86-NEXT: movl 20(%ebp), %edx -; X86-NEXT: movl %edx, %esi +; X86-NEXT: movl %edx, %ebx +; X86-NEXT: sarl $31, %ebx +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, %esi ; X86-NEXT: sarl $31, %esi -; X86-NEXT: movl %eax, %edi -; X86-NEXT: sarl $31, %edi -; X86-NEXT: movl %edi, %ebx -; X86-NEXT: shldl $31, %eax, %ebx -; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: shldl $31, %ecx, %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: shldl $31, %edi, %esi +; X86-NEXT: shldl $31, %ecx, %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) ; X86-NEXT: shll $31, %ecx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %edx -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %ecx -; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %eax +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: subl $1, %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: sbbl $0, %ebx -; X86-NEXT: testl %esi, %esi -; X86-NEXT: sets %al -; X86-NEXT: testl %edi, %edi -; X86-NEXT: sets %cl -; X86-NEXT: xorb %al, %cl -; X86-NEXT: movb %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 20(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl 20(%ebp) -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl %edi -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: subl $1, %esi +; X86-NEXT: sbbl $0, %edi +; X86-NEXT: testl %ebx, %ebx +; X86-NEXT: sets %al +; X86-NEXT: testl %ecx, %ecx +; X86-NEXT: sets %bl +; X86-NEXT: xorb %al, %bl ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl {{[0-9]+}}(%esp), %ecx ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: setne %al -; X86-NEXT: testb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Folded Reload -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload -; X86-NEXT: movl %ebx, %edx +; X86-NEXT: testb %bl, %al +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload +; X86-NEXT: movl %esi, %eax +; X86-NEXT: movl %edi, %edx ; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi diff --git a/llvm/test/CodeGen/X86/sdiv_fix_sat.ll b/llvm/test/CodeGen/X86/sdiv_fix_sat.ll index e7727a0ab617..7df490f98492 100644 --- a/llvm/test/CodeGen/X86/sdiv_fix_sat.ll +++ b/llvm/test/CodeGen/X86/sdiv_fix_sat.ll @@ -370,67 +370,68 @@ define i64 @func5(i64 %x, i64 %y) nounwind { ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp -; X86-NEXT: subl $96, %esp -; X86-NEXT: movl 8(%ebp), %ecx -; X86-NEXT: movl 12(%ebp), %eax -; X86-NEXT: movl 20(%ebp), %esi -; X86-NEXT: movl %esi, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: movl %eax, %edx -; X86-NEXT: sarl $31, %edx -; X86-NEXT: movl %edx, %edi -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: shldl $31, %eax, %edi -; X86-NEXT: shldl $31, %ecx, %eax +; X86-NEXT: subl $128, %esp +; X86-NEXT: movl 8(%ebp), %esi +; X86-NEXT: movl 12(%ebp), %edi +; X86-NEXT: movl 16(%ebp), %ecx +; X86-NEXT: movl 20(%ebp), %edx +; X86-NEXT: movl %edx, %eax +; X86-NEXT: sarl $31, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: shll $31, %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %esi -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %ecx +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, %ebx +; X86-NEXT: sarl $31, %ebx +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: shldl $31, %edi, %ebx +; X86-NEXT: shldl $31, %esi, %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: shll $31, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 20(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl 16(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: subl $1, %esi -; X86-NEXT: sbbl $0, %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: subl $1, %edi ; X86-NEXT: sbbl $0, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: sbbl $0, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: testl %ebx, %ebx -; X86-NEXT: sets %al -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, %ebx +; X86-NEXT: sbbl $0, %ebx ; X86-NEXT: testl %ecx, %ecx -; X86-NEXT: sets %dl -; X86-NEXT: xorb %al, %dl -; X86-NEXT: movb %dl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill -; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl 20(%ebp) -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %edi -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl %eax +; X86-NEXT: sets %al +; X86-NEXT: testl %edx, %edx +; X86-NEXT: sets %cl +; X86-NEXT: xorb %al, %cl +; X86-NEXT: movb %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx @@ -438,41 +439,38 @@ define i64 @func5(i64 %x, i64 %y) nounwind { ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: setne %al ; X86-NEXT: testb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Folded Reload -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload +; X86-NEXT: cmovel %esi, %ebx ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload ; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload -; X86-NEXT: cmpl $-1, %esi -; X86-NEXT: movl %eax, %ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload +; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload +; X86-NEXT: cmpl $-1, %edi ; X86-NEXT: sbbl $2147483647, %ecx # imm = 0x7FFFFFFF -; X86-NEXT: movl %edx, %ecx -; X86-NEXT: movl %edx, %ebx +; X86-NEXT: movl %eax, %ecx ; X86-NEXT: sbbl $0, %ecx -; X86-NEXT: movl %edi, %ecx +; X86-NEXT: movl %ebx, %ecx ; X86-NEXT: sbbl $0, %ecx ; X86-NEXT: movl $2147483647, %edx # imm = 0x7FFFFFFF -; X86-NEXT: cmovll %eax, %edx +; X86-NEXT: cmovll {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload ; X86-NEXT: movl $0, %ecx -; X86-NEXT: cmovgel %ecx, %edi -; X86-NEXT: movl %edi, %eax ; X86-NEXT: cmovgel %ecx, %ebx +; X86-NEXT: cmovgel %ecx, %eax ; X86-NEXT: movl $-1, %ecx -; X86-NEXT: cmovgel %ecx, %esi -; X86-NEXT: movl %esi, %edi -; X86-NEXT: negl %edi -; X86-NEXT: movl $-2147483648, %edi # imm = 0x80000000 -; X86-NEXT: sbbl %edx, %edi -; X86-NEXT: movl $-1, %edi -; X86-NEXT: sbbl %ebx, %edi -; X86-NEXT: sbbl %eax, %ecx +; X86-NEXT: cmovgel %ecx, %edi +; X86-NEXT: movl %edi, %esi +; X86-NEXT: negl %esi +; X86-NEXT: movl $-2147483648, %esi # imm = 0x80000000 +; X86-NEXT: sbbl %edx, %esi +; X86-NEXT: movl $-1, %esi +; X86-NEXT: sbbl %eax, %esi +; X86-NEXT: sbbl %ebx, %ecx ; X86-NEXT: movl $0, %eax -; X86-NEXT: cmovgel %eax, %esi +; X86-NEXT: cmovgel %eax, %edi ; X86-NEXT: movl $-2147483648, %eax # imm = 0x80000000 ; X86-NEXT: cmovgel %eax, %edx -; X86-NEXT: movl %esi, %eax +; X86-NEXT: movl %edi, %eax ; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi @@ -805,137 +803,155 @@ define <4 x i32> @vec(<4 x i32> %x, <4 x i32> %y) nounwind { ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp -; X86-NEXT: subl $208, %esp -; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: subl $240, %esp +; X86-NEXT: movl 12(%ebp), %esi +; X86-NEXT: movl 20(%ebp), %edi ; X86-NEXT: movl 16(%ebp), %ebx -; X86-NEXT: movl 32(%ebp), %eax -; X86-NEXT: movl %eax, %edi -; X86-NEXT: movl %eax, %ecx -; X86-NEXT: sarl $31, %edi -; X86-NEXT: movl %ebx, %edx -; X86-NEXT: sarl $31, %edx -; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 32(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) ; X86-NEXT: leal (%ebx,%ebx), %eax ; X86-NEXT: shrl $31, %ebx ; X86-NEXT: shldl $31, %eax, %ebx -; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp -; X86-NEXT: sarl $31, %esi -; X86-NEXT: movl 20(%ebp), %ecx +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 36(%ebp), %ecx ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: sarl $31, %eax -; X86-NEXT: leal (%ecx,%ecx), %edx -; X86-NEXT: shrl $31, %ecx -; X86-NEXT: shldl $31, %edx, %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: leal {{[0-9]+}}(%esp), %edx -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl 36(%ebp) ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %edx +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: leal (%edi,%edi), %eax +; X86-NEXT: shrl $31, %edi +; X86-NEXT: shldl $31, %eax, %edi +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl 32(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl 32(%ebp) -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl 36(%ebp) -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp -; X86-NEXT: movl 28(%ebp), %edx -; X86-NEXT: movl %edx, %ebx -; X86-NEXT: sarl $31, %ebx -; X86-NEXT: movl 12(%ebp), %ecx -; X86-NEXT: movl %ecx, %edi +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: movl %eax, %edi ; X86-NEXT: sarl $31, %edi -; X86-NEXT: leal (%ecx,%ecx), %eax -; X86-NEXT: shrl $31, %ecx -; X86-NEXT: shldl $31, %eax, %ecx -; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %esi, %ebx +; X86-NEXT: sarl $31, %ebx +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: leal (%esi,%esi), %eax +; X86-NEXT: shrl $31, %esi +; X86-NEXT: shldl $31, %eax, %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp -; X86-NEXT: movl 40(%ebp), %esi -; X86-NEXT: sarl $31, %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: sarl $31, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) ; X86-NEXT: movl 24(%ebp), %ecx ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: sarl $31, %eax -; X86-NEXT: leal (%ecx,%ecx), %edx +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: leal (%ecx,%ecx), %eax ; X86-NEXT: shrl $31, %ecx -; X86-NEXT: shldl $31, %edx, %ecx +; X86-NEXT: shldl $31, %eax, %ecx ; X86-NEXT: movl %ecx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: leal {{[0-9]+}}(%esp), %edx -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl %esi -; X86-NEXT: pushl 40(%ebp) -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %edx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edi, {{[0-9]+}}(%esp) +; X86-NEXT: movl 28(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, {{[0-9]+}}(%esp) ; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl %ebx -; X86-NEXT: pushl 28(%ebp) -; X86-NEXT: pushl %edi -; X86-NEXT: pushl %edi -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __divti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl 40(%ebp), %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: subl $1, %eax @@ -949,18 +965,18 @@ define <4 x i32> @vec(<4 x i32> %x, <4 x i32> %y) nounwind { ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: sbbl $0, %edx +; X86-NEXT: testl %edi, %edi +; X86-NEXT: sets {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Folded Spill ; X86-NEXT: testl %ebx, %ebx ; X86-NEXT: sets %bl -; X86-NEXT: testl %edi, %edi -; X86-NEXT: sets %bh -; X86-NEXT: xorb %bl, %bh +; X86-NEXT: xorb {{[-0-9]+}}(%e{{[sb]}}p), %bl # 1-byte Folded Reload ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi ; X86-NEXT: orl {{[0-9]+}}(%esp), %edi ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi ; X86-NEXT: orl {{[0-9]+}}(%esp), %esi ; X86-NEXT: orl %edi, %esi -; X86-NEXT: setne %bl -; X86-NEXT: testb %bh, %bl +; X86-NEXT: setne %bh +; X86-NEXT: testb %bl, %bh ; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Folded Reload ; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Reload @@ -1107,36 +1123,24 @@ define <4 x i32> @vec(<4 x i32> %x, <4 x i32> %y) nounwind { ; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx ; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: subl $1, %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: sbbl $0, %edi +; X86-NEXT: movl {{[0-9]+}}(%esp), %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: sbbl $0, %esi ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill ; X86-NEXT: sbbl $0, %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: sbbl $0, %esi -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload -; X86-NEXT: testl %ecx, %ecx +; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: sbbl $0, %edi +; X86-NEXT: cmpl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload ; X86-NEXT: sets %al -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload -; X86-NEXT: testl %edx, %edx -; X86-NEXT: sets %ah -; X86-NEXT: xorb %al, %ah -; X86-NEXT: movb %ah, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill -; X86-NEXT: leal {{[0-9]+}}(%esp), %eax -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl 40(%ebp) -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %edx -; X86-NEXT: pushl {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %eax +; X86-NEXT: cmpl $0, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: sets %cl +; X86-NEXT: xorb %al, %cl +; X86-NEXT: movb %cl, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Spill ; X86-NEXT: calll __modti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx @@ -1144,38 +1148,38 @@ define <4 x i32> @vec(<4 x i32> %x, <4 x i32> %y) nounwind { ; X86-NEXT: orl %eax, %ecx ; X86-NEXT: setne %al ; X86-NEXT: testb %al, {{[-0-9]+}}(%e{{[sb]}}p) # 1-byte Folded Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Folded Reload -; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %edi # 4-byte Folded Reload +; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Folded Reload ; X86-NEXT: cmovel {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Folded Reload ; X86-NEXT: cmpl $-1, %ebx -; X86-NEXT: movl %edi, %eax +; X86-NEXT: movl %esi, %eax ; X86-NEXT: sbbl $0, %eax ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: sbbl $0, %eax -; X86-NEXT: movl %esi, %eax +; X86-NEXT: movl %edi, %eax ; X86-NEXT: sbbl $0, %eax ; X86-NEXT: movl $0, %eax -; X86-NEXT: cmovgel %eax, %esi -; X86-NEXT: cmovgel %eax, %ecx ; X86-NEXT: cmovgel %eax, %edi +; X86-NEXT: cmovgel %eax, %ecx +; X86-NEXT: cmovgel %eax, %esi ; X86-NEXT: movl $-1, %edx ; X86-NEXT: cmovgel %edx, %ebx ; X86-NEXT: movl %ebx, %eax ; X86-NEXT: negl %eax ; X86-NEXT: movl $-1, %eax -; X86-NEXT: sbbl %edi, %eax +; X86-NEXT: sbbl %esi, %eax ; X86-NEXT: movl $-1, %eax ; X86-NEXT: sbbl %ecx, %eax ; X86-NEXT: movl $-1, %eax -; X86-NEXT: sbbl %esi, %eax +; X86-NEXT: sbbl %edi, %eax ; X86-NEXT: movl $0, %eax ; X86-NEXT: cmovgel %eax, %ebx -; X86-NEXT: cmovgel %edx, %edi -; X86-NEXT: shldl $31, %ebx, %edi +; X86-NEXT: cmovgel %edx, %esi +; X86-NEXT: shldl $31, %ebx, %esi ; X86-NEXT: movl 8(%ebp), %eax -; X86-NEXT: movl %edi, 12(%eax) +; X86-NEXT: movl %esi, 12(%eax) ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload ; X86-NEXT: movl %ecx, 8(%eax) ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload diff --git a/llvm/test/CodeGen/X86/shift-combine.ll b/llvm/test/CodeGen/X86/shift-combine.ll index 76cb4e87bae1..dfeef48897e0 100644 --- a/llvm/test/CodeGen/X86/shift-combine.ll +++ b/llvm/test/CodeGen/X86/shift-combine.ll @@ -792,14 +792,24 @@ define <4 x i32> @or_tree_with_mismatching_shifts_vec_i32(<4 x i32> %a, <4 x i32 define void @combineShiftOfShiftedLogic(i128 %a1, i32 %a2, ptr %p) { ; X86-LABEL: combineShiftOfShiftedLogic: ; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: pushl %ebp +; X86-NEXT: .cfi_def_cfa_offset 8 +; X86-NEXT: .cfi_offset %ebp, -8 +; X86-NEXT: movl %esp, %ebp +; X86-NEXT: .cfi_def_cfa_register %ebp +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 24(%ebp), %eax +; X86-NEXT: movl 28(%ebp), %ecx ; X86-NEXT: movl %eax, 20(%ecx) ; X86-NEXT: movl $0, 16(%ecx) ; X86-NEXT: movl $0, 12(%ecx) ; X86-NEXT: movl $0, 8(%ecx) ; X86-NEXT: movl $0, 4(%ecx) ; X86-NEXT: movl $0, (%ecx) +; X86-NEXT: movl %ebp, %esp +; X86-NEXT: popl %ebp +; X86-NEXT: .cfi_def_cfa %esp, 4 ; X86-NEXT: retl ; ; X64-LABEL: combineShiftOfShiftedLogic: diff --git a/llvm/test/CodeGen/X86/shift-i128.ll b/llvm/test/CodeGen/X86/shift-i128.ll index 767bd772ab7a..9323cd5b1917 100644 --- a/llvm/test/CodeGen/X86/shift-i128.ll +++ b/llvm/test/CodeGen/X86/shift-i128.ll @@ -212,9 +212,18 @@ entry: } define void @test_lshr_i128_outofrange(i128 %x, ptr nocapture %r) nounwind { -; ALL-LABEL: test_lshr_i128_outofrange: -; ALL: # %bb.0: # %entry -; ALL-NEXT: ret{{[l|q]}} +; i686-LABEL: test_lshr_i128_outofrange: +; i686: # %bb.0: # %entry +; i686-NEXT: pushl %ebp +; i686-NEXT: movl %esp, %ebp +; i686-NEXT: andl $-16, %esp +; i686-NEXT: movl %ebp, %esp +; i686-NEXT: popl %ebp +; i686-NEXT: retl +; +; x86_64-LABEL: test_lshr_i128_outofrange: +; x86_64: # %bb.0: # %entry +; x86_64-NEXT: retq entry: %0 = lshr i128 %x, -1 store i128 %0, ptr %r, align 16 @@ -222,9 +231,18 @@ entry: } define void @test_ashr_i128_outofrange(i128 %x, ptr nocapture %r) nounwind { -; ALL-LABEL: test_ashr_i128_outofrange: -; ALL: # %bb.0: # %entry -; ALL-NEXT: ret{{[l|q]}} +; i686-LABEL: test_ashr_i128_outofrange: +; i686: # %bb.0: # %entry +; i686-NEXT: pushl %ebp +; i686-NEXT: movl %esp, %ebp +; i686-NEXT: andl $-16, %esp +; i686-NEXT: movl %ebp, %esp +; i686-NEXT: popl %ebp +; i686-NEXT: retl +; +; x86_64-LABEL: test_ashr_i128_outofrange: +; x86_64: # %bb.0: # %entry +; x86_64-NEXT: retq entry: %0 = ashr i128 %x, -1 store i128 %0, ptr %r, align 16 @@ -232,9 +250,18 @@ entry: } define void @test_shl_i128_outofrange(i128 %x, ptr nocapture %r) nounwind { -; ALL-LABEL: test_shl_i128_outofrange: -; ALL: # %bb.0: # %entry -; ALL-NEXT: ret{{[l|q]}} +; i686-LABEL: test_shl_i128_outofrange: +; i686: # %bb.0: # %entry +; i686-NEXT: pushl %ebp +; i686-NEXT: movl %esp, %ebp +; i686-NEXT: andl $-16, %esp +; i686-NEXT: movl %ebp, %esp +; i686-NEXT: popl %ebp +; i686-NEXT: retl +; +; x86_64-LABEL: test_shl_i128_outofrange: +; x86_64: # %bb.0: # %entry +; x86_64-NEXT: retq entry: %0 = shl i128 %x, -1 store i128 %0, ptr %r, align 16 @@ -874,26 +901,31 @@ define <2 x i256> @shl_zext_lshr_outofrange(<2 x i128> %a0) { define i128 @lshr_shl_mask(i128 %a0) { ; i686-LABEL: lshr_shl_mask: ; i686: # %bb.0: -; i686-NEXT: pushl %edi +; i686-NEXT: pushl %ebp ; i686-NEXT: .cfi_def_cfa_offset 8 +; i686-NEXT: .cfi_offset %ebp, -8 +; i686-NEXT: movl %esp, %ebp +; i686-NEXT: .cfi_def_cfa_register %ebp +; i686-NEXT: pushl %edi ; i686-NEXT: pushl %esi -; i686-NEXT: .cfi_def_cfa_offset 12 -; i686-NEXT: .cfi_offset %esi, -12 -; i686-NEXT: .cfi_offset %edi, -8 -; i686-NEXT: movl {{[0-9]+}}(%esp), %eax -; i686-NEXT: movl {{[0-9]+}}(%esp), %ecx -; i686-NEXT: movl {{[0-9]+}}(%esp), %edx -; i686-NEXT: movl {{[0-9]+}}(%esp), %esi +; i686-NEXT: andl $-16, %esp +; i686-NEXT: .cfi_offset %esi, -16 +; i686-NEXT: .cfi_offset %edi, -12 +; i686-NEXT: movl 8(%ebp), %eax +; i686-NEXT: movl 24(%ebp), %ecx +; i686-NEXT: movl 28(%ebp), %edx +; i686-NEXT: movl 32(%ebp), %esi ; i686-NEXT: movl $2147483647, %edi # imm = 0x7FFFFFFF -; i686-NEXT: andl {{[0-9]+}}(%esp), %edi +; i686-NEXT: andl 36(%ebp), %edi ; i686-NEXT: movl %edi, 12(%eax) ; i686-NEXT: movl %esi, 8(%eax) ; i686-NEXT: movl %edx, 4(%eax) ; i686-NEXT: movl %ecx, (%eax) +; i686-NEXT: leal -8(%ebp), %esp ; i686-NEXT: popl %esi -; i686-NEXT: .cfi_def_cfa_offset 8 ; i686-NEXT: popl %edi -; i686-NEXT: .cfi_def_cfa_offset 4 +; i686-NEXT: popl %ebp +; i686-NEXT: .cfi_def_cfa %esp, 4 ; i686-NEXT: retl $4 ; ; x86_64-LABEL: lshr_shl_mask: diff --git a/llvm/test/CodeGen/X86/sincos-stack-args.ll b/llvm/test/CodeGen/X86/sincos-stack-args.ll index fe1533083333..42c05a3e7a9b 100644 --- a/llvm/test/CodeGen/X86/sincos-stack-args.ll +++ b/llvm/test/CodeGen/X86/sincos-stack-args.ll @@ -7,11 +7,10 @@ declare double @g(double, double) ; Though not visible within the IR, this will lower to an FSINCOS node, with ; store users, that are within a (callseq_start, callseq_end) pair. In this ; case, the stores cannot be folded into the sincos call. -define double @negative_sincos_with_stores_within_call_sequence(double %a) { +define double @negative_sincos_with_stores_within_call_sequence(double %a) nounwind { ; CHECK-LABEL: negative_sincos_with_stores_within_call_sequence: ; CHECK: # %bb.0: # %entry ; CHECK-NEXT: subl $44, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 48 ; CHECK-NEXT: fldl 48(%esp) ; CHECK-NEXT: leal 24(%esp), %eax ; CHECK-NEXT: movl %eax, 12(%esp) @@ -27,7 +26,6 @@ define double @negative_sincos_with_stores_within_call_sequence(double %a) { ; CHECK-NEXT: fstpl (%esp) ; CHECK-NEXT: calll g@PLT ; CHECK-NEXT: addl $44, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 4 ; CHECK-NEXT: retl entry: %0 = tail call double @llvm.sin.f64(double %a) diff --git a/llvm/test/CodeGen/X86/sincos.ll b/llvm/test/CodeGen/X86/sincos.ll index 42b1555e78b1..79034077f183 100644 --- a/llvm/test/CodeGen/X86/sincos.ll +++ b/llvm/test/CodeGen/X86/sincos.ll @@ -9,11 +9,10 @@ declare double @sin(double) readonly declare x86_fp80 @sinl(x86_fp80) readonly -define float @test1(float %X) { +define float @test1(float %X) nounwind { ; CHECK-LABEL: test1: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $12, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: flds {{[0-9]+}}(%esp) ; CHECK-NEXT: fstps (%esp) ; CHECK-NEXT: calll _sinf @@ -23,11 +22,10 @@ define float @test1(float %X) { ret float %Y } -define double @test2(double %X) { +define double @test2(double %X) nounwind { ; CHECK-LABEL: test2: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $12, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: fldl {{[0-9]+}}(%esp) ; CHECK-NEXT: fstpl (%esp) ; CHECK-NEXT: calll _sin @@ -37,11 +35,10 @@ define double @test2(double %X) { ret double %Y } -define x86_fp80 @test3(x86_fp80 %X) { +define x86_fp80 @test3(x86_fp80 %X) nounwind { ; CHECK-LABEL: test3: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $28, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 32 ; CHECK-NEXT: fldt {{[0-9]+}}(%esp) ; CHECK-NEXT: fstpt (%esp) ; CHECK-NEXT: calll _sinl @@ -58,11 +55,10 @@ declare double @cos(double) readonly declare x86_fp80 @cosl(x86_fp80) readonly -define float @test4(float %X) { +define float @test4(float %X) nounwind { ; CHECK-LABEL: test4: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $12, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: flds {{[0-9]+}}(%esp) ; CHECK-NEXT: fstps (%esp) ; CHECK-NEXT: calll _cosf @@ -72,11 +68,10 @@ define float @test4(float %X) { ret float %Y } -define double @test5(double %X) { +define double @test5(double %X) nounwind { ; CHECK-LABEL: test5: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $12, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: fldl {{[0-9]+}}(%esp) ; CHECK-NEXT: fstpl (%esp) ; CHECK-NEXT: calll _cos @@ -86,11 +81,10 @@ define double @test5(double %X) { ret double %Y } -define x86_fp80 @test6(x86_fp80 %X) { +define x86_fp80 @test6(x86_fp80 %X) nounwind { ; CHECK-LABEL: test6: ; CHECK: ## %bb.0: ; CHECK-NEXT: subl $28, %esp -; CHECK-NEXT: .cfi_def_cfa_offset 32 ; CHECK-NEXT: fldt {{[0-9]+}}(%esp) ; CHECK-NEXT: fstpt (%esp) ; CHECK-NEXT: calll _cosl diff --git a/llvm/test/CodeGen/X86/smax.ll b/llvm/test/CodeGen/X86/smax.ll index 86891e964d96..509d4443e930 100644 --- a/llvm/test/CodeGen/X86/smax.ll +++ b/llvm/test/CodeGen/X86/smax.ll @@ -151,31 +151,34 @@ define i128 @test_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: test_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmpl %ebx, %edx -; X86-NEXT: movl %esi, %ebp -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl %edi, %ebp -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ebx +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: cmpl 24(%ebp), %ebx +; X86-NEXT: movl %edi, %esi +; X86-NEXT: sbbl 28(%ebp), %esi +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl %edx, %esi +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: movl 52(%ebp), %ecx ; X86-NEXT: movl %ecx, %eax -; X86-NEXT: sbbl %ebp, %eax -; X86-NEXT: cmovll %ebx, %edx -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovll %ebp, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: sbbl %esi, %eax +; X86-NEXT: cmovll 24(%ebp), %ebx +; X86-NEXT: cmovll 28(%ebp), %edi +; X86-NEXT: cmovll 32(%ebp), %edx +; X86-NEXT: cmovll %esi, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %ecx, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -717,29 +720,32 @@ define i128 @test_signbits_i128(i128 %a, i128 %b) nounwind { ; ; X86-LABEL: test_signbits_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: shrdl $28, %edi, %ecx -; X86-NEXT: sarl $28, %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: shrdl $28, %edx, %ecx +; X86-NEXT: sarl $28, %edx ; X86-NEXT: cmpl %esi, %ecx -; X86-NEXT: movl %edi, %ebx -; X86-NEXT: sbbl %edx, %ebx +; X86-NEXT: movl %edx, %edi +; X86-NEXT: sbbl %eax, %edi ; X86-NEXT: cmovll %esi, %ecx -; X86-NEXT: cmovll %edx, %edi -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: sarl $31, %edi +; X86-NEXT: cmovll %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: sarl $31, %edx +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %ax = ashr i128 %a, 64 %bx = ashr i128 %b, 92 diff --git a/llvm/test/CodeGen/X86/smin.ll b/llvm/test/CodeGen/X86/smin.ll index 8907f6c4cd59..5e9fe27b41d2 100644 --- a/llvm/test/CodeGen/X86/smin.ll +++ b/llvm/test/CodeGen/X86/smin.ll @@ -151,32 +151,34 @@ define i128 @test_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: test_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: cmpl %edx, %ebx -; X86-NEXT: sbbl %esi, %ebp -; X86-NEXT: movl %eax, %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: cmpl %ecx, 24(%ebp) +; X86-NEXT: sbbl %edx, %esi +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: sbbl %esi, %ebx +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %edi ; X86-NEXT: movl %edi, %eax -; X86-NEXT: sbbl %ebp, %eax -; X86-NEXT: cmovll %ebx, %edx -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovll {{[0-9]+}}(%esp), %ecx -; X86-NEXT: cmovll %edi, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %ebp, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: cmovll 24(%ebp), %ecx +; X86-NEXT: cmovll 28(%ebp), %edx +; X86-NEXT: cmovll 32(%ebp), %esi +; X86-NEXT: cmovll %edi, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -718,29 +720,32 @@ define i128 @test_signbits_i128(i128 %a, i128 %b) nounwind { ; ; X86-LABEL: test_signbits_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: shrdl $28, %edi, %ecx -; X86-NEXT: sarl $28, %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: shrdl $28, %edx, %ecx +; X86-NEXT: sarl $28, %edx ; X86-NEXT: cmpl %ecx, %esi -; X86-NEXT: movl %edx, %ebx -; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl %eax, %edi +; X86-NEXT: sbbl %edx, %edi ; X86-NEXT: cmovll %esi, %ecx -; X86-NEXT: cmovll %edx, %edi -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: sarl $31, %edi +; X86-NEXT: cmovll %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: sarl $31, %edx +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %ax = ashr i128 %a, 64 %bx = ashr i128 %b, 92 diff --git a/llvm/test/CodeGen/X86/stack-align2.ll b/llvm/test/CodeGen/X86/stack-align2.ll index 99f36d2ca8b7..095a9090ed08 100644 --- a/llvm/test/CodeGen/X86/stack-align2.ll +++ b/llvm/test/CodeGen/X86/stack-align2.ll @@ -3,13 +3,11 @@ ; RUN: llc < %s -mcpu=generic -mtriple=i386-netbsd | FileCheck %s -check-prefix=NETBSD-I386 ; RUN: llc < %s -mcpu=generic -mtriple=i686-apple-darwin8 | FileCheck %s -check-prefix=DARWIN-I386 ; RUN: llc < %s -mcpu=generic -mtriple=i386-pc-solaris2.11 | FileCheck %s -check-prefix=SOLARIS-I386 -; RUN: llc < %s -mcpu=generic -mtriple=i386-nacl | FileCheck %s -check-prefix=NACL-I386 ; RUN: llc < %s -mcpu=generic -mtriple=x86_64-linux | FileCheck %s -check-prefix=LINUX-X86_64 ; RUN: llc < %s -mcpu=generic -mtriple=x86_64-kfreebsd | FileCheck %s -check-prefix=KFREEBSD-X86_64 ; RUN: llc < %s -mcpu=generic -mtriple=x86_64-netbsd | FileCheck %s -check-prefix=NETBSD-X86_64 ; RUN: llc < %s -mcpu=generic -mtriple=x86_64-apple-darwin8 | FileCheck %s -check-prefix=DARWIN-X86_64 ; RUN: llc < %s -mcpu=generic -mtriple=x86_64-pc-solaris2.11 | FileCheck %s -check-prefix=SOLARIS-X86_64 -; RUN: llc < %s -mcpu=generic -mtriple=x86_64-nacl | FileCheck %s -check-prefix=NACL-X86_64 define i32 @test() nounwind { entry: @@ -19,7 +17,6 @@ entry: ; LINUX-I386: subl $12, %esp ; KFREEBSD-I386: subl $12, %esp ; DARWIN-I386: subl $12, %esp -; NACL-I386: subl $12, %esp ; NETBSD-I386-NOT: subl {{.*}}, %esp ; SOLARIS-I386-NOT: subl {{.*}}, %esp @@ -27,8 +24,6 @@ entry: ; LINUX-X86_64-NOT: subq {{.*}}, %rsp ; DARWIN-X86_64: pushq %{{.*}} ; DARWIN-X86_64-NOT: subq {{.*}}, %rsp -; NACL-X86_64: pushq %{{.*}} -; NACL-X86_64-NOT: subq {{.*}}, %rsp ; NETBSD-X86_64: pushq %{{.*}} ; NETBSD-X86_64-NOT: subq {{.*}}, %rsp ; SOLARIS-X86_64: pushq %{{.*}} diff --git a/llvm/test/CodeGen/X86/stack-protector-target-openbsd.ll b/llvm/test/CodeGen/X86/stack-protector-target-openbsd.ll new file mode 100644 index 000000000000..06382c6bbbbe --- /dev/null +++ b/llvm/test/CodeGen/X86/stack-protector-target-openbsd.ll @@ -0,0 +1,81 @@ +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 +; RUN: llc -mtriple=i386--linux < %s | FileCheck -check-prefix=LINUX32 %s +; RUN: llc -mtriple=x86_64--linux < %s | FileCheck -check-prefix=LINUX64 %s +; RUN: llc -mtriple=i386--openbsd < %s | FileCheck -check-prefix=OPENBSD32 %s +; RUN: llc -mtriple=x86_64--openbsd < %s | FileCheck -check-prefix=OPENBSD64 %s + +define void @func() sspreq nounwind { +; LINUX32-LABEL: func: +; LINUX32: # %bb.0: +; LINUX32-NEXT: subl $12, %esp +; LINUX32-NEXT: movl %gs:20, %eax +; LINUX32-NEXT: movl %eax, {{[0-9]+}}(%esp) +; LINUX32-NEXT: leal {{[0-9]+}}(%esp), %eax +; LINUX32-NEXT: movl %eax, (%esp) +; LINUX32-NEXT: calll capture@PLT +; LINUX32-NEXT: movl %gs:20, %eax +; LINUX32-NEXT: cmpl {{[0-9]+}}(%esp), %eax +; LINUX32-NEXT: jne .LBB0_2 +; LINUX32-NEXT: # %bb.1: # %SP_return +; LINUX32-NEXT: addl $12, %esp +; LINUX32-NEXT: retl +; LINUX32-NEXT: .LBB0_2: # %CallStackCheckFailBlk +; LINUX32-NEXT: calll __stack_chk_fail@PLT +; +; LINUX64-LABEL: func: +; LINUX64: # %bb.0: +; LINUX64-NEXT: subq $24, %rsp +; LINUX64-NEXT: movq %fs:40, %rax +; LINUX64-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; LINUX64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi +; LINUX64-NEXT: callq capture@PLT +; LINUX64-NEXT: movq %fs:40, %rax +; LINUX64-NEXT: cmpq {{[0-9]+}}(%rsp), %rax +; LINUX64-NEXT: jne .LBB0_2 +; LINUX64-NEXT: # %bb.1: # %SP_return +; LINUX64-NEXT: addq $24, %rsp +; LINUX64-NEXT: retq +; LINUX64-NEXT: .LBB0_2: # %CallStackCheckFailBlk +; LINUX64-NEXT: callq __stack_chk_fail@PLT +; +; OPENBSD32-LABEL: func: +; OPENBSD32: # %bb.0: +; OPENBSD32-NEXT: subl $8, %esp +; OPENBSD32-NEXT: movl __guard_local, %eax +; OPENBSD32-NEXT: movl %eax, {{[0-9]+}}(%esp) +; OPENBSD32-NEXT: movl %esp, %eax +; OPENBSD32-NEXT: pushl %eax +; OPENBSD32-NEXT: calll capture@PLT +; OPENBSD32-NEXT: addl $4, %esp +; OPENBSD32-NEXT: movl __guard_local, %eax +; OPENBSD32-NEXT: cmpl {{[0-9]+}}(%esp), %eax +; OPENBSD32-NEXT: jne .LBB0_2 +; OPENBSD32-NEXT: # %bb.1: # %SP_return +; OPENBSD32-NEXT: addl $8, %esp +; OPENBSD32-NEXT: retl +; OPENBSD32-NEXT: .LBB0_2: # %CallStackCheckFailBlk +; OPENBSD32-NEXT: pushl $.LSSH +; OPENBSD32-NEXT: calll __stack_smash_handler@PLT +; +; OPENBSD64-LABEL: func: +; OPENBSD64: # %bb.0: +; OPENBSD64-NEXT: subq $24, %rsp +; OPENBSD64-NEXT: movq __guard_local(%rip), %rax +; OPENBSD64-NEXT: movq %rax, {{[0-9]+}}(%rsp) +; OPENBSD64-NEXT: leaq {{[0-9]+}}(%rsp), %rdi +; OPENBSD64-NEXT: callq capture@PLT +; OPENBSD64-NEXT: movq __guard_local(%rip), %rax +; OPENBSD64-NEXT: cmpq {{[0-9]+}}(%rsp), %rax +; OPENBSD64-NEXT: jne .LBB0_2 +; OPENBSD64-NEXT: # %bb.1: # %SP_return +; OPENBSD64-NEXT: addq $24, %rsp +; OPENBSD64-NEXT: retq +; OPENBSD64-NEXT: .LBB0_2: # %CallStackCheckFailBlk +; OPENBSD64-NEXT: movl $.LSSH, %edi +; OPENBSD64-NEXT: callq __stack_smash_handler@PLT + %alloca = alloca i32, align 4 + call void @capture(ptr %alloca) + ret void +} + +declare void @capture(ptr) diff --git a/llvm/test/CodeGen/X86/test-shrink-bug.ll b/llvm/test/CodeGen/X86/test-shrink-bug.ll index 953a0d65c538..ab28a3b4a2b6 100644 --- a/llvm/test/CodeGen/X86/test-shrink-bug.ll +++ b/llvm/test/CodeGen/X86/test-shrink-bug.ll @@ -48,18 +48,17 @@ define dso_local void @fail(i16 %a, <2 x i8> %b) { ; CHECK-X86: ## %bb.0: ; CHECK-X86-NEXT: subl $12, %esp ; CHECK-X86-NEXT: .cfi_def_cfa_offset 16 -; CHECK-X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax +; CHECK-X86-NEXT: movzwl {{[0-9]+}}(%esp), %ecx ; CHECK-X86-NEXT: cmpb $123, {{[0-9]+}}(%esp) -; CHECK-X86-NEXT: setne %cl -; CHECK-X86-NEXT: testl $263, %eax ## imm = 0x107 -; CHECK-X86-NEXT: setne %al -; CHECK-X86-NEXT: testb %cl, %al -; CHECK-X86-NEXT: jne LBB1_2 -; CHECK-X86-NEXT: ## %bb.1: ## %yes -; CHECK-X86-NEXT: addl $12, %esp -; CHECK-X86-NEXT: retl -; CHECK-X86-NEXT: LBB1_2: ## %no +; CHECK-X86-NEXT: sete %al +; CHECK-X86-NEXT: testl $263, %ecx ## imm = 0x107 +; CHECK-X86-NEXT: je LBB1_3 +; CHECK-X86-NEXT: ## %bb.1: +; CHECK-X86-NEXT: testb %al, %al +; CHECK-X86-NEXT: jne LBB1_3 +; CHECK-X86-NEXT: ## %bb.2: ## %no ; CHECK-X86-NEXT: calll _bar +; CHECK-X86-NEXT: LBB1_3: ## %yes ; CHECK-X86-NEXT: addl $12, %esp ; CHECK-X86-NEXT: retl ; diff --git a/llvm/test/CodeGen/X86/ucmp.ll b/llvm/test/CodeGen/X86/ucmp.ll index 6a52acfe2fb3..7f17299b39e3 100644 --- a/llvm/test/CodeGen/X86/ucmp.ll +++ b/llvm/test/CodeGen/X86/ucmp.ll @@ -107,29 +107,33 @@ define i8 @ucmp.8.128(i128 %x, i128 %y) nounwind { ; X86-LABEL: ucmp.8.128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: cmpl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl %ebp, %eax -; X86-NEXT: sbbl %esi, %eax +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl 12(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %ecx +; X86-NEXT: cmpl %eax, 24(%ebp) ; X86-NEXT: movl %ecx, %eax ; X86-NEXT: sbbl %edx, %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl %edi, %eax +; X86-NEXT: movl 16(%ebp), %ebx +; X86-NEXT: movl 32(%ebp), %eax +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: movl 20(%ebp), %ecx +; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: movl %esi, %eax ; X86-NEXT: sbbl %ecx, %eax ; X86-NEXT: setb %al -; X86-NEXT: cmpl %ebx, {{[0-9]+}}(%esp) -; X86-NEXT: sbbl %ebp, %esi -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %edx -; X86-NEXT: sbbl %edi, %ecx +; X86-NEXT: movl 8(%ebp), %edi +; X86-NEXT: cmpl 24(%ebp), %edi +; X86-NEXT: sbbl 28(%ebp), %edx +; X86-NEXT: sbbl 32(%ebp), %ebx +; X86-NEXT: sbbl %esi, %ecx ; X86-NEXT: sbbb $0, %al +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx diff --git a/llvm/test/CodeGen/X86/udiv_fix.ll b/llvm/test/CodeGen/X86/udiv_fix.ll index 5b1e0545502b..82dfeeee1329 100644 --- a/llvm/test/CodeGen/X86/udiv_fix.ll +++ b/llvm/test/CodeGen/X86/udiv_fix.ll @@ -153,26 +153,28 @@ define i64 @func5(i64 %x, i64 %y) nounwind { ; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp -; X86-NEXT: subl $32, %esp +; X86-NEXT: subl $80, %esp ; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movl 16(%ebp), %edx +; X86-NEXT: movl 20(%ebp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) ; X86-NEXT: movl %ecx, %edx ; X86-NEXT: shrl %edx +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) ; X86-NEXT: shldl $31, %eax, %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) ; X86-NEXT: shll $31, %eax -; X86-NEXT: movl %esp, %esi -; X86-NEXT: pushl $0 -; X86-NEXT: pushl $0 -; X86-NEXT: pushl 20(%ebp) -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %esi +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __udivti3 -; X86-NEXT: addl $32, %esp -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: subl $4, %esp +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: leal -4(%ebp), %esp ; X86-NEXT: popl %esi diff --git a/llvm/test/CodeGen/X86/udiv_fix_sat.ll b/llvm/test/CodeGen/X86/udiv_fix_sat.ll index 30a7f80b2315..3da5973f9f90 100644 --- a/llvm/test/CodeGen/X86/udiv_fix_sat.ll +++ b/llvm/test/CodeGen/X86/udiv_fix_sat.ll @@ -194,32 +194,34 @@ define i64 @func5(i64 %x, i64 %y) nounwind { ; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %esi ; X86-NEXT: andl $-16, %esp -; X86-NEXT: subl $32, %esp +; X86-NEXT: subl $80, %esp ; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movl 16(%ebp), %edx +; X86-NEXT: movl 20(%ebp), %esi +; X86-NEXT: movl %esi, {{[0-9]+}}(%esp) +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) ; X86-NEXT: movl %ecx, %edx ; X86-NEXT: shrl %edx +; X86-NEXT: movl %edx, {{[0-9]+}}(%esp) ; X86-NEXT: shldl $31, %eax, %ecx +; X86-NEXT: movl %ecx, {{[0-9]+}}(%esp) ; X86-NEXT: shll $31, %eax -; X86-NEXT: movl %esp, %esi -; X86-NEXT: pushl $0 -; X86-NEXT: pushl $0 -; X86-NEXT: pushl 20(%ebp) -; X86-NEXT: pushl 16(%ebp) -; X86-NEXT: pushl $0 -; X86-NEXT: pushl %edx -; X86-NEXT: pushl %ecx -; X86-NEXT: pushl %eax -; X86-NEXT: pushl %esi +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: movl %eax, (%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) ; X86-NEXT: calll __udivti3 -; X86-NEXT: addl $32, %esp +; X86-NEXT: subl $4, %esp ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: orl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl $-1, %eax ; X86-NEXT: movl $-1, %edx ; X86-NEXT: jne .LBB4_2 ; X86-NEXT: # %bb.1: -; X86-NEXT: movl (%esp), %eax +; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: .LBB4_2: ; X86-NEXT: leal -4(%ebp), %esp diff --git a/llvm/test/CodeGen/X86/umax.ll b/llvm/test/CodeGen/X86/umax.ll index f589d4a7b04a..7ef859978cdb 100644 --- a/llvm/test/CodeGen/X86/umax.ll +++ b/llvm/test/CodeGen/X86/umax.ll @@ -232,31 +232,34 @@ define i128 @test_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: test_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmpl %ebx, %edx -; X86-NEXT: movl %esi, %ebp -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl %edi, %ebp -; X86-NEXT: sbbl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ebx +; X86-NEXT: movl 44(%ebp), %edi +; X86-NEXT: cmpl 24(%ebp), %ebx +; X86-NEXT: movl %edi, %esi +; X86-NEXT: sbbl 28(%ebp), %esi +; X86-NEXT: movl 48(%ebp), %edx +; X86-NEXT: movl %edx, %esi +; X86-NEXT: sbbl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %esi +; X86-NEXT: movl 52(%ebp), %ecx ; X86-NEXT: movl %ecx, %eax -; X86-NEXT: sbbl %ebp, %eax -; X86-NEXT: cmovbl %ebx, %edx -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %edi -; X86-NEXT: cmovbl %ebp, %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax +; X86-NEXT: sbbl %esi, %eax +; X86-NEXT: cmovbl 24(%ebp), %ebx +; X86-NEXT: cmovbl 28(%ebp), %edi +; X86-NEXT: cmovbl 32(%ebp), %edx +; X86-NEXT: cmovbl %esi, %ecx +; X86-NEXT: movl 8(%ebp), %eax ; X86-NEXT: movl %ecx, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -282,37 +285,40 @@ define i128 @test_i128_1(i128 %a) nounwind { ; X86-LABEL: test_i128_1: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 32(%ebp), %edx +; X86-NEXT: movl 24(%ebp), %eax ; X86-NEXT: cmpl $1, %eax -; X86-NEXT: movl %eax, %ebx -; X86-NEXT: adcl $0, %ebx -; X86-NEXT: testl %edx, %edx -; X86-NEXT: movl $1, %edi -; X86-NEXT: cmovnel %eax, %edi -; X86-NEXT: cmovel %ebx, %edi -; X86-NEXT: xorl %ebx, %ebx -; X86-NEXT: movl %ecx, %ebp -; X86-NEXT: negl %ebp -; X86-NEXT: movl $0, %ebp -; X86-NEXT: sbbl %esi, %ebp -; X86-NEXT: movl $1, %ebp -; X86-NEXT: cmovbl %eax, %ebp -; X86-NEXT: movl %ecx, %eax -; X86-NEXT: cmovbl %edx, %ebx -; X86-NEXT: orl %esi, %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %esi, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: cmovel %edi, %ebp -; X86-NEXT: cmovel %edx, %ebx -; X86-NEXT: movl %ebx, 4(%eax) -; X86-NEXT: movl %ebp, (%eax) +; X86-NEXT: movl %eax, %ecx +; X86-NEXT: adcl $0, %ecx +; X86-NEXT: cmpl $0, 28(%ebp) +; X86-NEXT: movl $1, %esi +; X86-NEXT: cmovnel %eax, %esi +; X86-NEXT: cmovel %ecx, %esi +; X86-NEXT: xorl %edi, %edi +; X86-NEXT: movl %edx, %ecx +; X86-NEXT: negl %ecx +; X86-NEXT: movl 36(%ebp), %ecx +; X86-NEXT: movl $0, %ebx +; X86-NEXT: sbbl %ecx, %ebx +; X86-NEXT: movl $1, %ebx +; X86-NEXT: cmovbl %eax, %ebx +; X86-NEXT: cmovbl 28(%ebp), %edi +; X86-NEXT: movl %edx, %eax +; X86-NEXT: orl %ecx, %eax +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: cmovel %esi, %ebx +; X86-NEXT: cmovel 28(%ebp), %edi +; X86-NEXT: movl %edi, 4(%eax) +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -1312,29 +1318,32 @@ define i128 @test_signbits_i128(i128 %a, i128 %b) nounwind { ; ; X86-LABEL: test_signbits_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: shrdl $28, %edi, %ecx -; X86-NEXT: sarl $28, %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: shrdl $28, %edx, %ecx +; X86-NEXT: sarl $28, %edx ; X86-NEXT: cmpl %esi, %ecx -; X86-NEXT: movl %edi, %ebx -; X86-NEXT: sbbl %edx, %ebx +; X86-NEXT: movl %edx, %edi +; X86-NEXT: sbbl %eax, %edi ; X86-NEXT: cmovbl %esi, %ecx -; X86-NEXT: cmovbl %edx, %edi -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: sarl $31, %edi +; X86-NEXT: cmovbl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: sarl $31, %edx +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %ax = ashr i128 %a, 64 %bx = ashr i128 %b, 92 diff --git a/llvm/test/CodeGen/X86/umin.ll b/llvm/test/CodeGen/X86/umin.ll index 7a5cdbb9ce75..c927abf3a426 100644 --- a/llvm/test/CodeGen/X86/umin.ll +++ b/llvm/test/CodeGen/X86/umin.ll @@ -147,32 +147,34 @@ define i128 @test_i128(i128 %a, i128 %b) nounwind { ; X86-LABEL: test_i128: ; X86: # %bb.0: ; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: cmpl %edx, %ebx -; X86-NEXT: sbbl %esi, %ebp -; X86-NEXT: movl %eax, %ebp -; X86-NEXT: sbbl %ecx, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $16, %esp +; X86-NEXT: movl 40(%ebp), %ecx +; X86-NEXT: movl 44(%ebp), %edx +; X86-NEXT: movl 28(%ebp), %esi +; X86-NEXT: cmpl %ecx, 24(%ebp) +; X86-NEXT: sbbl %edx, %esi +; X86-NEXT: movl 48(%ebp), %esi +; X86-NEXT: movl 32(%ebp), %ebx +; X86-NEXT: sbbl %esi, %ebx +; X86-NEXT: movl 52(%ebp), %ebx +; X86-NEXT: movl 36(%ebp), %edi ; X86-NEXT: movl %edi, %eax -; X86-NEXT: sbbl %ebp, %eax -; X86-NEXT: cmovbl %ebx, %edx -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %esi -; X86-NEXT: cmovbl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: cmovbl %edi, %ebp -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl %ebp, 12(%eax) -; X86-NEXT: movl %ecx, 8(%eax) -; X86-NEXT: movl %esi, 4(%eax) -; X86-NEXT: movl %edx, (%eax) +; X86-NEXT: sbbl %ebx, %eax +; X86-NEXT: cmovbl 24(%ebp), %ecx +; X86-NEXT: cmovbl 28(%ebp), %edx +; X86-NEXT: cmovbl 32(%ebp), %esi +; X86-NEXT: cmovbl %edi, %ebx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %ebx, 12(%eax) +; X86-NEXT: movl %esi, 8(%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: movl %ecx, (%eax) +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -727,29 +729,32 @@ define i128 @test_signbits_i128(i128 %a, i128 %b) nounwind { ; ; X86-LABEL: test_signbits_i128: ; X86: # %bb.0: -; X86-NEXT: pushl %ebx +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: movl {{[0-9]+}}(%esp), %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl {{[0-9]+}}(%esp), %edi -; X86-NEXT: shrdl $28, %edi, %ecx -; X86-NEXT: sarl $28, %edi +; X86-NEXT: andl $-16, %esp +; X86-NEXT: movl 32(%ebp), %esi +; X86-NEXT: movl 36(%ebp), %eax +; X86-NEXT: movl 48(%ebp), %ecx +; X86-NEXT: movl 52(%ebp), %edx +; X86-NEXT: shrdl $28, %edx, %ecx +; X86-NEXT: sarl $28, %edx ; X86-NEXT: cmpl %ecx, %esi -; X86-NEXT: movl %edx, %ebx -; X86-NEXT: sbbl %edi, %ebx +; X86-NEXT: movl %eax, %edi +; X86-NEXT: sbbl %edx, %edi ; X86-NEXT: cmovbl %esi, %ecx -; X86-NEXT: cmovbl %edx, %edi -; X86-NEXT: movl %edi, 4(%eax) -; X86-NEXT: sarl $31, %edi +; X86-NEXT: cmovbl %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: movl %edx, 4(%eax) ; X86-NEXT: movl %ecx, (%eax) -; X86-NEXT: movl %edi, 12(%eax) -; X86-NEXT: movl %edi, 8(%eax) +; X86-NEXT: sarl $31, %edx +; X86-NEXT: movl %edx, 12(%eax) +; X86-NEXT: movl %edx, 8(%eax) +; X86-NEXT: leal -8(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi -; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl $4 %ax = ashr i128 %a, 64 %bx = ashr i128 %b, 92 diff --git a/llvm/test/CodeGen/X86/umulo-128-legalisation-lowering.ll b/llvm/test/CodeGen/X86/umulo-128-legalisation-lowering.ll index 4c3170304b98..89afd1b00444 100644 --- a/llvm/test/CodeGen/X86/umulo-128-legalisation-lowering.ll +++ b/llvm/test/CodeGen/X86/umulo-128-legalisation-lowering.ll @@ -38,8 +38,8 @@ define { i128, i8 } @muloti_test(i128 %l, i128 %r) unnamed_addr #0 { ; X86-NEXT: .cfi_def_cfa_offset 16 ; X86-NEXT: pushl %esi ; X86-NEXT: .cfi_def_cfa_offset 20 -; X86-NEXT: subl $24, %esp -; X86-NEXT: .cfi_def_cfa_offset 44 +; X86-NEXT: subl $28, %esp +; X86-NEXT: .cfi_def_cfa_offset 48 ; X86-NEXT: .cfi_offset %esi, -20 ; X86-NEXT: .cfi_offset %edi, -16 ; X86-NEXT: .cfi_offset %ebx, -12 @@ -147,7 +147,7 @@ define { i128, i8 } @muloti_test(i128 %l, i128 %r) unnamed_addr #0 { ; X86-NEXT: andb $1, %al ; X86-NEXT: movb %al, 16(%ecx) ; X86-NEXT: movl %ecx, %eax -; X86-NEXT: addl $24, %esp +; X86-NEXT: addl $28, %esp ; X86-NEXT: .cfi_def_cfa_offset 20 ; X86-NEXT: popl %esi ; X86-NEXT: .cfi_def_cfa_offset 16 diff --git a/llvm/test/CodeGen/X86/unreachable-mbb-undef-phi.mir b/llvm/test/CodeGen/X86/unreachable-mbb-undef-phi.mir index 1bdbc5386266..232a5e3353b2 100644 --- a/llvm/test/CodeGen/X86/unreachable-mbb-undef-phi.mir +++ b/llvm/test/CodeGen/X86/unreachable-mbb-undef-phi.mir @@ -1,4 +1,5 @@ # RUN: llc -mtriple=x86_64-- %s -o - -run-pass=processimpdefs -run-pass=unreachable-mbb-elimination | FileCheck %s +# RUN: llc -mtriple=x86_64-- %s -o - -passes=process-imp-defs,unreachable-mbb-elimination | FileCheck %s --- name: f tracksRegLiveness: true diff --git a/llvm/test/CodeGen/X86/wide-integer-cmp.ll b/llvm/test/CodeGen/X86/wide-integer-cmp.ll index a15d633d8538..12dccca76eb1 100644 --- a/llvm/test/CodeGen/X86/wide-integer-cmp.ll +++ b/llvm/test/CodeGen/X86/wide-integer-cmp.ll @@ -92,6 +92,8 @@ define i32 @test_wide(i128 %a, i128 %b) { ; CHECK: # %bb.0: # %entry ; CHECK-NEXT: pushl %esi ; CHECK-NEXT: .cfi_def_cfa_offset 8 +; CHECK-NEXT: subl $8, %esp +; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: .cfi_offset %esi, -8 ; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax ; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx @@ -101,15 +103,15 @@ define i32 @test_wide(i128 %a, i128 %b) { ; CHECK-NEXT: sbbl {{[0-9]+}}(%esp), %esi ; CHECK-NEXT: sbbl {{[0-9]+}}(%esp), %ecx ; CHECK-NEXT: sbbl {{[0-9]+}}(%esp), %eax -; CHECK-NEXT: jge .LBB4_2 +; CHECK-NEXT: jge .LBB4_3 ; CHECK-NEXT: # %bb.1: # %bb1 ; CHECK-NEXT: movl $1, %eax -; CHECK-NEXT: popl %esi -; CHECK-NEXT: .cfi_def_cfa_offset 4 -; CHECK-NEXT: retl -; CHECK-NEXT: .LBB4_2: # %bb2 -; CHECK-NEXT: .cfi_def_cfa_offset 8 +; CHECK-NEXT: jmp .LBB4_2 +; CHECK-NEXT: .LBB4_3: # %bb2 ; CHECK-NEXT: movl $2, %eax +; CHECK-NEXT: .LBB4_2: # %bb1 +; CHECK-NEXT: addl $8, %esp +; CHECK-NEXT: .cfi_def_cfa_offset 8 ; CHECK-NEXT: popl %esi ; CHECK-NEXT: .cfi_def_cfa_offset 4 ; CHECK-NEXT: retl diff --git a/llvm/test/CodeGen/X86/win32-int-runtime-libcalls.ll b/llvm/test/CodeGen/X86/win32-int-runtime-libcalls.ll new file mode 100644 index 000000000000..5ac90a0af2e5 --- /dev/null +++ b/llvm/test/CodeGen/X86/win32-int-runtime-libcalls.ll @@ -0,0 +1,113 @@ +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 +; RUN: llc -mtriple=i686-windows-msvc < %s | FileCheck -check-prefix=CHECK32 %s +; RUN: llc -mtriple=x86_64-windows-msvc < %s | FileCheck -check-prefix=CHECK64 %s + +define i64 @test_sdiv_i64(i64 %a, i64 %b) { +; CHECK32-LABEL: test_sdiv_i64: +; CHECK32: # %bb.0: +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: calll __alldiv +; CHECK32-NEXT: retl +; +; CHECK64-LABEL: test_sdiv_i64: +; CHECK64: # %bb.0: +; CHECK64-NEXT: movq %rdx, %r8 +; CHECK64-NEXT: movq %rcx, %rax +; CHECK64-NEXT: cqto +; CHECK64-NEXT: idivq %r8 +; CHECK64-NEXT: retq + %ret = sdiv i64 %a, %b + ret i64 %ret +} + +define i64 @test_srem_i64(i64 %a, i64 %b) { +; CHECK32-LABEL: test_srem_i64: +; CHECK32: # %bb.0: +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: calll __allrem +; CHECK32-NEXT: retl +; +; CHECK64-LABEL: test_srem_i64: +; CHECK64: # %bb.0: +; CHECK64-NEXT: movq %rdx, %r8 +; CHECK64-NEXT: movq %rcx, %rax +; CHECK64-NEXT: cqto +; CHECK64-NEXT: idivq %r8 +; CHECK64-NEXT: movq %rdx, %rax +; CHECK64-NEXT: retq + %ret = srem i64 %a, %b + ret i64 %ret +} + +define i64 @test_udiv_i64(i64 %a, i64 %b) { +; CHECK32-LABEL: test_udiv_i64: +; CHECK32: # %bb.0: +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: calll __aulldiv +; CHECK32-NEXT: retl +; +; CHECK64-LABEL: test_udiv_i64: +; CHECK64: # %bb.0: +; CHECK64-NEXT: movq %rdx, %r8 +; CHECK64-NEXT: movq %rcx, %rax +; CHECK64-NEXT: xorl %edx, %edx +; CHECK64-NEXT: divq %r8 +; CHECK64-NEXT: retq + %ret = udiv i64 %a, %b + ret i64 %ret +} + +define i64 @test_urem_i64(i64 %a, i64 %b) { +; CHECK32-LABEL: test_urem_i64: +; CHECK32: # %bb.0: +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: pushl {{[0-9]+}}(%esp) +; CHECK32-NEXT: calll __aullrem +; CHECK32-NEXT: retl +; +; CHECK64-LABEL: test_urem_i64: +; CHECK64: # %bb.0: +; CHECK64-NEXT: movq %rdx, %r8 +; CHECK64-NEXT: movq %rcx, %rax +; CHECK64-NEXT: xorl %edx, %edx +; CHECK64-NEXT: divq %r8 +; CHECK64-NEXT: movq %rdx, %rax +; CHECK64-NEXT: retq + %ret = urem i64 %a, %b + ret i64 %ret +} + +define i64 @test_mul_i64(i64 %a, i64 %b) { +; CHECK32-LABEL: test_mul_i64: +; CHECK32: # %bb.0: +; CHECK32-NEXT: pushl %esi +; CHECK32-NEXT: movl {{[0-9]+}}(%esp), %ecx +; CHECK32-NEXT: movl {{[0-9]+}}(%esp), %esi +; CHECK32-NEXT: movl %ecx, %eax +; CHECK32-NEXT: mull %esi +; CHECK32-NEXT: imull {{[0-9]+}}(%esp), %ecx +; CHECK32-NEXT: addl %ecx, %edx +; CHECK32-NEXT: imull {{[0-9]+}}(%esp), %esi +; CHECK32-NEXT: addl %esi, %edx +; CHECK32-NEXT: popl %esi +; CHECK32-NEXT: retl +; +; CHECK64-LABEL: test_mul_i64: +; CHECK64: # %bb.0: +; CHECK64-NEXT: movq %rcx, %rax +; CHECK64-NEXT: imulq %rdx, %rax +; CHECK64-NEXT: retq + %ret = mul i64 %a, %b + ret i64 %ret +} diff --git a/llvm/test/CodeGen/X86/x86-64-baseptr.ll b/llvm/test/CodeGen/X86/x86-64-baseptr.ll index 020004def6e7..62c63d5defe6 100644 --- a/llvm/test/CodeGen/X86/x86-64-baseptr.ll +++ b/llvm/test/CodeGen/X86/x86-64-baseptr.ll @@ -2,14 +2,6 @@ ; RUN: llc -mtriple=x86_64-pc-linux -stackrealign -verify-machineinstrs < %s | FileCheck %s ; RUN: llc -mtriple=x86_64-pc-linux-gnux32 -stackrealign -verify-machineinstrs < %s | FileCheck -check-prefix=X32ABI %s -; This should run with NaCl as well ( -mtriple=x86_64-pc-nacl ) but currently doesn't due to PR22655 - -; Make sure the correct register gets set up as the base pointer -; This should be rbx for x64 and 64-bit NaCl and ebx for x32 -; NACL-LABEL: base -; NACL: subq $32, %rsp -; NACL: movq %rsp, %rbx - declare i32 @helper() nounwind define void @base() #0 { ; CHECK-LABEL: base: diff --git a/llvm/test/CodeGen/X86/x86-64-stack-and-frame-ptr.ll b/llvm/test/CodeGen/X86/x86-64-stack-and-frame-ptr.ll index bceebdc9ad7d..26be80ea5894 100644 --- a/llvm/test/CodeGen/X86/x86-64-stack-and-frame-ptr.ll +++ b/llvm/test/CodeGen/X86/x86-64-stack-and-frame-ptr.ll @@ -1,6 +1,5 @@ ; RUN: llc -verify-machineinstrs -mtriple=x86_64-pc-linux < %s | FileCheck %s ; RUN: llc -verify-machineinstrs -mtriple=x86_64-pc-linux-gnux32 < %s | FileCheck -check-prefix=X32ABI %s -; RUN: llc -verify-machineinstrs -mtriple=x86_64-pc-nacl < %s | FileCheck -check-prefix=NACL %s ; x32 uses %esp, %ebp as stack and frame pointers @@ -14,12 +13,6 @@ ; X32ABI: movl %esp, %ebp ; X32ABI: movl %edi, -4(%ebp) ; X32ABI: popq %rbp -; NACL-LABEL: foo -; NACL: pushq %rbp -; NACL: movq %rsp, %rbp -; NACL: movl %edi, -4(%rbp) -; NACL: popq %rbp - define void @foo(ptr %a) #0 { entry: @@ -30,5 +23,3 @@ entry: } attributes #0 = { nounwind uwtable "frame-pointer"="all"} - - |