In GFX10, a number of VOP3 instructions should allow opsel, including
V_MAX_I16, V_MAX_U16, V_MIN_I16, V_MIN_U16, V_MUL_LO_U16, V_LSHLREV_B16,
V_LSHRREV_B16, and V_ASHRREV_I16.

In an earlier state of this code, it was possible for an existing SDWA
MI to reach the code in the "createSDWAversion" function. This is no
longer possible; see assert at the top of the function.

Remove code that tries to handle operands on pre-existing SDWA
instructions from the function.

The sext modifier on an operand of V_CNDMASK_B32_sdwa gets erroneously
turned into a neg modifier in the assembly output.

As a workaround, to avoid miscompilation, this patch disables the
conversion of V_CNDMASK_B32 to the SDWA form if any operand uses an sext
modifier.

Fixes #138766.

---------

Co-authored-by: Matt Arsenault <arsenm2@gmail.com>

Before V_CNDMASK_B32_e64 gets converted to SDWA form, a conversion to
V_CNDMASK_B32_e32 occurs.
The vcc use of this instruction must be fixed into a vcc_lo use for wave32.
This fix only happens after the final conversion to the SDWA form. This led
to a compiler error in situations where the conversion to SDWA aborts.

Make sure that the vcc-fix gets applied even if the SDWA conversion is
not completed.

---------

Co-authored-by: Matt Arsenault <arsenm2@gmail.com>

This patch fixes:

  llvm/lib/Target/AMDGPU/SIPeepholeSDWA.cpp:1102:8: error: unused
  variable 'Converted' [-Werror,-Wunused-variable]

The VOP3 form of the V_CNDMASK_B32 instruction takes a carry-in
operand. The conversion to SDWA implies a conversion to VOP2 form
which reads from VCC instead.

Convert V_CNDMASK_B32_e64 instructions that might be converted to SDWA
to V_CNDMASK_B32_e32 first and introduce a copy of the carry-in operand
to VCC.

Closes #133431.

---------

Co-authored-by: Matt Arsenault <arsenm2@gmail.com>

The si-peephole-sdwa pass adjusts the selections on sdwa instructions to
the selections on their operands during its conversions. For instance,
if an instruction selects `BYTE_0` and its operand selects `WORD_1`, the
combined selection should be `BYTE_2`, i.e. "`BYTE_0` of `WORD_1`". The
existing implementation does not always handle this correctly in some
complex situations with instructions across different basic blocks as
demonstrated by the test cases included in this PR.

This PR adds an additional selection combination step to the conversion
to fix this issue. It reverts the changes made by PR #123942 which had
disabled the conversion of preexisting SDWA instructions completely as a
quick fix.

---------

Co-authored-by: Jeffrey Byrnes <Jeffrey.Byrnes@amd.com>
Co-authored-by: Matt Arsenault <arsenm2@gmail.com>

This PR reapplies the changes from PR #123942 which had to be reverted
because of a test failure. The test has been adjusted.

This reverts commit 6fdaaafd89d7cbc15dafe3ebf1aa3235d148aaab.
Breaks check-llvm, see
https://github.com/llvm/llvm-project/pull/123942#issuecomment-2609861953

This is meant as a short-term workaround for an invalid conversion in
this pass that occurs because existing SDWA selections are not correctly
taken into account during the conversion.

See the draft PR #123221 for an attempt to fix the actual issue.

---------

Co-authored-by: Frederik Harwath <fharwath@amd.com>