| Age | Commit message (Collapse) | Author |
|---|
|
The DXIL validator version option(/validator-version) decide the validator version when compile hlsl.
The format is major.minor like 1.0.
In normal case, the value of validator version should be got from DXIL validator. Before we got DXIL validator ready for llvm/main, DXIL validator version option is added first to set validator version.
It will affect code generation for DXIL, so it is treated as a code gen option.
A new member std::string DxilValidatorVersion is added to clang::CodeGenOptions.
Then CGHLSLRuntime is added to clang::CodeGenModule.
It is used to translate clang::CodeGenOptions::DxilValidatorVersion into a ModuleFlag under key "dx.valver" at end of clang code generation.
Reviewed By: beanz
Differential Revision: https://reviews.llvm.org/D123884
|
|
Different TU's may have this globl var. appending linkage can
only be used with lld recognized special variables.
Change it to internal linkage.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D124466
|
|
This patch is a continuation of https://reviews.llvm.org/D123353.
Not only kernels in anonymous namespace, but also template
kernels with template arguments in anonymous namespace
need to be externalized.
To be more generic, this patch checks the linkage of a kernel
assuming the kernel does not have __global__ attribute. If
the linkage is internal then clang will externalize it.
This patch also fixes the postfix for externalized symbol
since nvptx does not allow '.' in symbol name.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D124189
Fixes: https://github.com/llvm/llvm-project/issues/54560
|
|
|
|
Rename gpu.used.external as __clang_gpu_used_external as ptxas does not
allow . in global variable name.
Fixes: https://github.com/llvm/llvm-project/issues/54934
Reviewed by: Joseph Huber, Artem Belevich
Differential Revision: https://reviews.llvm.org/D123946
|
|
In D123649, I got the formula for getFlexibleArrayInitChars slightly
wrong: the flexible array elements can be contained in the tail padding
of the struct. Fix the formula to account for that.
With the fixed formula, we run into another issue: in some cases, we
were emitting extra padding for flexible arrray initializers. Fix
CGExprConstant so it uses a packed struct when necessary, to avoid this
extra padding.
Differential Revision: https://reviews.llvm.org/D123826
|
|
They're causing failures in LLVM test-suite. Added some regression
tests that explain the issue.
|
|
Flexible array initialization is a C/C++ extension implemented in many
compilers to allow initializing the flexible array tail of a struct type
that contains a flexible array. In clang, this is currently restricted
to C. But this construct is used in the Microsoft SDK headers, so I'd
like to extend it to C++.
For now, this doesn't handle dynamic initialization; probably not hard
to implement, but it's extra code, and I don't think it's necessary for
the expected uses. And we explicitly fail out of constant evaluation.
I've added some additional code to assert that initializers have the
correct size, with or without flexible array init. This might catch
issues unrelated to flexible array init.
Differential Revision: https://reviews.llvm.org/D123649
|
|
For -fgpu-rdc, a host function may call an external kernel
which is defined in an archive of bitcode. Since this external
kernel is only referenced in host function, the device
bitcode does not contain reference to this external
kernel, then the linker will not try to resolve this external
kernel in the archive.
To fix this issue, host-used external kernels and device
variables are tracked. A global array containing pointers
to these external kernels and variables is emitted which
serves as an artificial references to the external kernels
and variables used by host.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D123441
|
|
LTO objects might compiled with different `mbranch-protection` flags which will cause an error in the linker.
Such a setup is allowed in the normal build with this change that is possible.
Reviewed By: pcc
Differential Revision: https://reviews.llvm.org/D123493
|
|
kernels in anonymous name space needs to have unique name
to avoid duplicate symbols.
Fixes: https://github.com/llvm/llvm-project/issues/54560
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D123353
|
|
Differential Revision: https://reviews.llvm.org/D118948
|
|
This reverts commit 8aa1490513f111afd407d87c3f07d26f65c8a686.
Broke testing: https://lab.llvm.org/buildbot/#/builders/109/builds/36233
|
|
Currently, enablement of heap MTE on Android is specified by an ELF note, which
signals to the linker to enable heap MTE. This change allows
-fsanitize=memtag-heap to synthesize these notes, rather than adding them
through the build system. We need to extend this feature to also signal the
linker to do special work for MTE globals (in future) and MTE stack (currently
implemented in the toolchain, but not implemented in the loader).
Current Android uses a non-backwards-compatible ELF note, called
".note.android.memtag". Stack MTE is an ABI break anyway, so we don't mind that
we won't be able to run executables with stack MTE on Android 11/12 devices.
The current expectation is to support the verbiage used by Android, in
that "SYNC" means MTE Synchronous mode, and "ASYNC" effectively means
"fast", using the Kernel auto-upgrade feature that allows
hardware-specific and core-specific configuration as to whether "ASYNC"
would end up being Asynchronous, Asymmetric, or Synchronous on that
particular core, whichever has a reasonable performance delta. Of
course, this is platform and loader-specific.
Differential Revision: https://reviews.llvm.org/D118948
|
|
This eliminates an unused-variable warning
|
|
clang to emit DWARF information for global alias variable as
DW_TAG_imported_declaration. This change also handles nested
(recursive) imported declarations.
Reviewed by: dblaikie, aprantl
Differential Revision: https://reviews.llvm.org/D120989
|
|
This change merges code for emit of target and target_clones multiversion
resolver functions and, in doing so, corrects handling of target_clones
functions that are declared but not defined. Previously, a use of such
a target_clones function would result in an attempted emit of an ifunc
that referenced an undefined resolver function. Ifunc references to
undefined resolver functions are not allowed and, when the LLVM verifier
is not disabled (via '-disable-llvm-verifier'), resulted in the verifier
issuing a "IFunc resolver must be a definition" error and aborting the
compilation. With this change, ifuncs and resolver function definitions
are always emitted for used target_clones functions regardless of whether
the target_clones function is defined (if the function is defined, then
the ifunc and resolver are emitted regardless of whether the function is
used).
This change has the side effect of causing target_clones variants and
resolver functions to be emitted in a different order than they were
previously. This is harmless and is reflected in the updated tests.
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D122958
|
|
resolvers.
This change modifies CodeGenModule::emitMultiVersionFunctions() in preparation
for a change that will merge support for emitting target_clones resolvers into
this function. This change mostly serves to isolate indentation changes from
later behavior modifying changes.
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D122957
|
|
CodeGenModule::GetOrCreateMultiVersionResolver().
Previously, GetOrCreateMultiVersionResolver() required the caller to provide
a GlobalDecl along with an llvm::type and FunctionDecl. The latter two can be
cheaply obtained from the first, and the llvm::type parameter is not always
used, so requiring the caller to provide them was unnecessary and created the
possibility that callers would pass an inconsistent set. This change simplifies
the interface to only require the GlobalDecl value.
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D122956
|
|
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D122955
|
|
This is required for converting function calls such as get_global_id()
into SPIR-V builtins.
Differential Revision: https://reviews.llvm.org/D123049
|
|
Previously, resolver functions synthesized for target_clones multiversion
functions were not emitted as COMDAT. Now fixed.
|
|
With opaque pointers, this is not necessarily a ConstantExpr. And
we don't need one here either, just Constant is sufficient.
|
|
We expect that `extern "C"` static functions to be usable in things like
inline assembly, as well as ifuncs:
See the bug report here: https://github.com/llvm/llvm-project/issues/54549
However, we were diagnosing this as 'not defined', because the
ifunc's attempt to look up its resolver would generate a declared IR
function.
Additionally, as background, the way we allow these static extern "C"
functions to work in inline assembly is by making an alias with the C
mangling in MOST situations to the version we emit with
internal-linkage/mangling.
The problem here was multi-fold: First- We generated the alias after the
ifunc was checked, so the function by that name didn't exist yet.
Second, the ifunc's generation caused a symbol to exist under the name
of the alias already (the declared function above), which suppressed the
alias generation.
This patch fixes all of this by moving the checking of ifuncs/CFE aliases
until AFTER we have generated the extern-C alias. Then, it does a
'fixup' around the GlobalIFunc to make sure we correct the reference.
Differential Revision: https://reviews.llvm.org/D122608
|
|
Adds DXContainer to switch statements in Clang and LLDB to silence
warnings.
|
|
This builtin returns the address of a global instance of the
`std::source_location::__impl` type, which must be defined (with an
appropriate shape) before calling the builtin.
It will be used to implement std::source_location in libc++ in a
future change. The builtin is compatible with GCC's implementation,
and libstdc++'s usage. An intentional divergence is that GCC declares
the builtin's return type to be `const void*` (for
ease-of-implementation reasons), while Clang uses the actual type,
`const std::source_location::__impl*`.
In order to support this new functionality, I've also added a new
'UnnamedGlobalConstantDecl'. This artificial Decl is modeled after
MSGuidDecl, and is used to represent a generic concept of an lvalue
constant with global scope, deduplicated by its value. It's possible
that MSGuidDecl itself, or some of the other similar sorts of things
in Clang might be able to be refactored onto this more-generic
concept, but there's enough special-case weirdness in MSGuidDecl that
I gave up attempting to share code there, at least for now.
Finally, for compatibility with libstdc++'s <source_location> header,
I've added a second exception to the "cannot cast from void* to T* in
constant evaluation" rule. This seems a bit distasteful, but feels
like the best available option.
Reviewers: aaron.ballman, erichkeane
Differential Revision: https://reviews.llvm.org/D120159
|
|
Remove two stray uses in CodeGenModule and CGCUDANV.
|
|
|
|
Currently we use the `-fembed-offload-object` option to embed a binary
file into the host as a named section. This is currently only used as a
codegen action, meaning we only handle this option correctly when the
input is a bitcode file. This patch adds the same handling to embed an
offloading object after we complete code generation. This allows us to
embed the object correctly if the input file is source or bitcode.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D120270
|
|
Due to various implementation constraints, despite the programmer
choosing a 'processor' cpu_dispatch/cpu_specific needs to use the
'feature' list of a processor to identify it. This results in the
identified processor in source-code not being propogated to the
optimizer, and thus, not able to be tuned for.
This patch changes to use the actual cpu as written for tune-cpu so that
opt can make decisions based on the cpu-as-spelled, which should better
match the behavior expected by the programmer.
Note that the 'valid' list of processors for x86 is in
llvm/include/llvm/Support/X86TargetParser.def. At the moment, this list
contains only Intel processors, but other vendors may wish to add their
own entries as 'alias'es (or with different feature lists!).
If this is not done, there is two potential performance issues with the
patch, but I believe them to be worth it in light of the improvements to
behavior and performance.
1- In the event that the user spelled "ProcessorB", but we only have the
features available to test for "ProcessorA" (where A is B minus
features),
AND there is an optimization opportunity for "B" that negatively affects
"A", the optimizer will likely choose to do so.
2- In the event that the user spelled VendorI's processor, and the
feature
list allows it to run on VendorA's processor of similar features, AND
there
is an optimization opportunity for VendorIs that negatively affects
"A"s,
the optimizer will likely choose to do so. This can be fixed by adding
an
alias to X86TargetParser.def.
Differential Revision: https://reviews.llvm.org/D121410
|
|
The purpose of this change is to fix the following codegen bug:
```
// main.c
__attribute__((cpu_specific(generic)))
int *foo(void) { static int z; return &z;}
int main() { return *foo() = 5; }
// other.c
__attribute__((cpu_dispatch(generic))) int *foo(void);
// run:
clang main.c other.c -o main; ./main
```
This will segfault prior to the change, and return the correct
exit code 5 after the change.
The underlying cause is that when a translation unit contains
a cpu_specific function without the corresponding cpu_dispatch
the generated code binds the reference to foo() against a
GlobalIFunc whose resolver is undefined. This is invalid: the
resolver must be defined in the same translation unit as the
ifunc, but historically the LLVM bitcode verifier did not check
that. The generated code then binds against the resolver rather
than the ifunc, so it ends up calling the resolver rather than
the resolvee. In the example above it treats its return value as
an int *, therefore trying to write to program text.
The root issue at the representation level is that GlobalIFunc,
like GlobalAlias, does not support a "declaration" state. The
object which provides the correct semantics in these cases
is a Function declaration, but unlike Functions, changing a
declaration to a definition in the GlobalIFunc case constitutes
a change of the object type, as opposed to simply emitting code
into a Function.
I think this limitation is unlikely to change, so I implemented
the fix by returning a function declaration rather than an ifunc
when encountering cpu_specific, and upgrading it to an ifunc
when emitting cpu_dispatch.
This uses `takeName` + `replaceAllUsesWith` in similar vein to
other places where the correct IR object type cannot be known
locally/up-front, like in `CodeGenModule::EmitAliasDefinition`.
Previous discussion in: https://reviews.llvm.org/D112349
Signed-off-by: Itay Bookstein <ibookstein@gmail.com>
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D120266
|
|
To make uses of the deprecated constructor easier to spot, and to
ensure that no new uses are introduced, rename it to
Address::deprecated().
While doing the rename, I've filled in element types in cases
where it was relatively obvious, but we're still left with 135
calls to the deprecated constructor.
|
|
We have the `clang -cc1` command-line option `-funwind-tables=1|2` and
the codegen option `VALUE_CODEGENOPT(UnwindTables, 2, 0) ///< Unwind
tables (1) or asynchronous unwind tables (2)`. However, this is
encoded in LLVM IR by the presence or the absence of the `uwtable`
attribute, i.e. we lose the information whether to generate want just
some unwind tables or asynchronous unwind tables.
Asynchronous unwind tables take more space in the runtime image, I'd
estimate something like 80-90% more, as the difference is adding
roughly the same number of CFI directives as for prologues, only a bit
simpler (e.g. `.cfi_offset reg, off` vs. `.cfi_restore reg`). Or even
more, if you consider tail duplication of epilogue blocks.
Asynchronous unwind tables could also restrict code generation to
having only a finite number of frame pointer adjustments (an example
of *not* having a finite number of `SP` adjustments is on AArch64 when
untagging the stack (MTE) in some cases the compiler can modify `SP`
in a loop).
Having the CFI precise up to an instruction generally also means one
cannot bundle together CFI instructions once the prologue is done,
they need to be interspersed with ordinary instructions, which means
extra `DW_CFA_advance_loc` commands, further increasing the unwind
tables size.
That is to say, async unwind tables impose a non-negligible overhead,
yet for the most common use cases (like C++ exceptions), they are not
even needed.
This patch extends the `uwtable` attribute with an optional
value:
- `uwtable` (default to `async`)
- `uwtable(sync)`, synchronous unwind tables
- `uwtable(async)`, asynchronous (instruction precise) unwind tables
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D114543
|
|
CodeGenModule::GetAddrOfGlobalTemporary()
|
|
The module flag to indicate use of hostcall is insufficient to catch
all cases where hostcall might be in use by a kernel. This is now
replaced by a function attribute that gets propagated to top-level
kernel functions via their respective call-graph.
If the attribute "amdgpu-no-hostcall-ptr" is absent on a kernel, the
default behaviour is to emit kernel metadata indicating that the
kernel uses the hostcall buffer pointer passed as an implicit
argument.
The attribute may be placed explicitly by the user, or inferred by the
AMDGPU attributor by examining the call-graph. The attribute is
inferred only if the function is not being sanitized, and the
implictarg_ptr does not result in a load of any byte in the hostcall
pointer argument.
Reviewed By: jdoerfert, arsenm, kpyzhov
Differential Revision: https://reviews.llvm.org/D119216
|
|
code object version determines ABI, therefore should not be mixed.
This patch emits amdgpu_code_object_version module flag in LLVM IR
based on code object version (default 4).
The amdgpu_code_object_version value is code object version times 100.
LLVM IR with different amdgpu_code_object_version module flag cannot
be linked.
The -cc1 option -mcode-object-version=none is for ROCm device library use
only, which supports multiple ABI.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D119026
|
|
This issue is an oversight in D108621.
Literals in HIP are emitted as global constant variables with default
address space which maps to Generic address space for HIPSPV. In
SPIR-V such variables translate to OpVariable instructions with
Generic storage class which are not legal. Fix by mapping literals
to CrossWorkGroup address space.
The literals are not mapped to UniformConstant because the “flat”
pointers in HIP may reference them and “flat” pointers are modeled
as Generic pointers in SPIR-V. In SPIR-V/OpenCL UniformConstant
pointers may not be casted to Generic.
Patch by: Henry Linjamäki
Reviewed by: Yaxun Liu
Differential Revision: https://reviews.llvm.org/D118876
|
|
Even if the reference itself is dllexport, the temporary should not be.
In fact, we're already giving it internal linkage, so dllexporting it
is not just wasteful, but will fail to link, as in the example below:
$ cat /tmp/a.cc
void _DllMainCRTStartup() {}
const int __declspec(dllexport) &foo = 42;
$ clang-cl -fuse-ld=lld /tmp/a.cc /Zl /link /dll /out:a.dll
lld-link: error: <root>: undefined symbol: int const &foo::$RT1
Differential revision: https://reviews.llvm.org/D118980
|
|
Branch protection in M-class is supported by
- Armv8.1-M.Main
- Armv8-M.Main
- Armv7-M
Attempting to enable this for other architectures, either by
command-line (e.g -mbranch-protection=bti) or by target attribute
in source code (e.g. __attribute__((target("branch-protection=..."))) )
will generate a warning.
In both cases function attributes related to branch protection will not
be emitted. Regardless of the warning, module level attributes related to
branch protection will be emitted when it is enabled via the command-line.
The following people also contributed to this patch:
- Victor Campos
Reviewed By: chill
Differential Revision: https://reviews.llvm.org/D115501
|
|
Intel's CET/IBT requires every indirect branch target to be an ENDBR instruction. Because of that, the compiler needs to correctly emit these instruction on function's prologues. Because this is a security feature, it is desirable that only actual indirect-branch-targeted functions are emitted with ENDBRs. While it is possible to identify address-taken functions through LTO, minimizing these ENDBR instructions remains a hard task for user-space binaries because exported functions may end being reachable through PLT entries, that will use an indirect branch for such. Because this cannot be determined during compilation-time, the compiler currently emits ENDBRs to every non-local-linkage function.
Despite the challenge presented for user-space, the kernel landscape is different as no PLTs are used. With the intent of providing the most fit ENDBR emission for the kernel, kernel developers proposed an optimization named "ibt-seal" which replaces the ENDBRs for NOPs directly in the binary. The discussion of this feature can be seen in [1].
This diff brings the enablement of the flag -mibt-seal, which in combination with LTO enforces a different policy for ENDBR placement in when the code-model is set to "kernel". In this scenario, the compiler will only emit ENDBRs to address taken functions, ignoring non-address taken functions that are don't have local linkage.
A comparison between an LTO-compiled kernel binaries without and with the -mibt-seal feature enabled shows that when -mibt-seal was used, the number of ENDBRs in the vmlinux.o binary patched by objtool decreased from 44383 to 33192, and that the number of superfluous ENDBR instructions nopped-out decreased from 11730 to 540.
The 540 missed superfluous ENDBRs need to be investigated further, but hypotheses are: assembly code not being taken care of by the compiler, kernel exported symbols mechanisms creating bogus address taken situations or even these being removed due to other binary optimizations like kernel's static_calls. For now, I assume that the large drop in the number of ENDBR instructions already justifies the feature being merged.
[1] - https://lkml.org/lkml/2021/11/22/591
Reviewed By: xiangzhangllvm
Differential Revision: https://reviews.llvm.org/D116070
|
|
HIP program with printf call fails to compile with -fsanitize=address
option, because of appending module flag - amdgpu_hostcall twice, one
for printf and one for sanitize option. This patch fixes that issue.
Patch by: Praveen Velliengiri
Reviewed by: Yaxun Liu, Roman Lebedev
Differential Revision: https://reviews.llvm.org/D116216
|
|
Use the AttributeSet constructor instead. There's no good reason
why AttrBuilder itself should exact the AttributeSet from the
AttributeList. Moving this out of the AttrBuilder generally results
in cleaner code.
|
|
Cases where there is a mangling of a cpu-dispatch/cpu-specific function
before the function becomes 'multiversion' (such as a member function)
causes the wrong name to be emitted for one of the variants/resolver,
since the name is cached. Make sure we invalidate the cache in
cpu-dispatch/cpu-specific modes, like we previously did for just target
multiversioning.
|
|
to repro error.
As mentioned yesterday, I've got a problem that I can only reproduce on
Godbolt (none of the build configs on my local machine!), so this is at
least somewhat usable until I figure out a cause.
|
|
I'm attempting to debug an issue that I can only get to happen on
godbolt, where the cpu-dispatch resolver for an out of line member
function is generated with the wrong name, causing a link failure.
|
|
Using and std::map<SmallString, SmallString> for target dependent attributes is
inefficient: it makes its constructor slightly heavier, and involves extra
allocation for each new string attribute. Storing the attribute key/value as
strings implies extra allocation/copy step.
Use a sorted vector instead. Given the low number of attributes generally
involved, this is cheaper, as showcased by
https://llvm-compile-time-tracker.com/compare.php?from=5de322295f4ade692dc4f1823ae4450ad3c48af2&to=05bc480bf641a9e3b466619af43a2d123ee3f71d&stat=instructions
Differential Revision: https://reviews.llvm.org/D116599
|
|
Identified with modernize-use-bool-literals.
|
|
This class is solely used as a lightweight and clean way to build a set of
attributes to be removed from an AttrBuilder. Previously AttrBuilder was used
both for building and removing, which introduced odd situation like creation of
Attribute with dummy value because the only relevant part was the attribute
kind.
Differential Revision: https://reviews.llvm.org/D116110
|
|
Control-Flow Integrity (CFI) replaces references to address-taken
functions with pointers to the CFI jump table. This is a problem
for low-level code, such as operating system kernels, which may
need the address of an actual function body without the jump table
indirection.
This change adds the __builtin_function_start() builtin, which
accepts an argument that can be constant-evaluated to a function,
and returns the address of the function body.
Link: https://github.com/ClangBuiltLinux/linux/issues/1353
Depends on D108478
Reviewed By: pcc, rjmccall
Differential Revision: https://reviews.llvm.org/D108479
|
|
Change all uses of the deprecated constructor to pass the
element type explicitly and drop it.
For cases where the correct element type was not immediately
obvious to me or would require a slightly larger change I'm
falling back to explicitly calling getPointerElementType() for now.
|
