[PATCH] D44747: Set calling convention for CUDA kernel

tra added a comment. In https://reviews.llvm.org/D44747#1055894, @yaxunl wrote: > Let's revert it for now. I will create a review after fixing it and commit it > again with the fix. > > Thanks. > > Sam reverted in r329099. Repository: rL LLVM https://reviews.llvm.org/D44747 ___

RE: [PATCH] D44747: Set calling convention for CUDA kernel

l.com; Arsenault, Matthew Cc: jle...@google.com; llvm-comm...@lists.llvm.org; t...@google.com; Zhuravlyov, Konstantin ; wei.di...@amd.com; Stuttard, David ; tpr.l...@botech.co.uk; Tye, Tony ; cfe-commits@lists.llvm.org Subject: [PATCH] D44747: Set calling convention for CUDA kernel tra added a co

[PATCH] D44747: Set calling convention for CUDA kernel

tra added a comment. In https://reviews.llvm.org/D44747#1055877, @yaxunl wrote: > I will try fixing that. > > The CUDA kernel calling convention should be dropped in all DRE's since it is > invisible to the user. > > Sam Apparently it's not always the case. Here's a bit more elaborate example

RE: [PATCH] D44747: Set calling convention for CUDA kernel

rjmcc...@gmail.com; Arsenault, Matthew Cc: llvm-comm...@lists.llvm.org; t...@google.com; Zhuravlyov, Konstantin ; wei.di...@amd.com; Stuttard, David ; tpr.l...@botech.co.uk; Tye, Tony ; cfe-commits@lists.llvm.org; jle...@google.com Subject: [PATCH] D44747: Set calling convention for CUDA kernel

[PATCH] D44747: Set calling convention for CUDA kernel

tra added inline comments. Comment at: lib/Sema/SemaType.cpp:3319-3330 + // Attribute AT_CUDAGlobal affects the calling convention for AMDGPU targets. + // This is the simplest place to infer calling convention for CUDA kernels. + if (S.getLangOpts().CUDA && S.getLangOpts().CU

[PATCH] D44747: Set calling convention for CUDA kernel

tra added inline comments. Comment at: lib/Sema/SemaType.cpp:3319-3330 + // Attribute AT_CUDAGlobal affects the calling convention for AMDGPU targets. + // This is the simplest place to infer calling convention for CUDA kernels. + if (S.getLangOpts().CUDA && S.getLangOpts().CU

[PATCH] D44747: Set calling convention for CUDA kernel

This revision was automatically updated to reflect the committed changes. Closed by commit rL328795: Set calling convention for CUDA kernel (authored by yaxunl, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llvm.org/D44747?vs=140131&id=140256#

[PATCH] D44747: Set calling convention for CUDA kernel

This revision was automatically updated to reflect the committed changes. Closed by commit rC328795: Set calling convention for CUDA kernel (authored by yaxunl, committed by ). Repository: rL LLVM https://reviews.llvm.org/D44747 Files: include/clang/Basic/Specifiers.h lib/AST/ItaniumMangl

[PATCH] D44747: Set calling convention for CUDA kernel

rjmccall accepted this revision. rjmccall added a comment. This revision is now accepted and ready to land. LGTM. https://reviews.llvm.org/D44747 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/

[PATCH] D44747: Set calling convention for CUDA kernel

yaxunl updated this revision to Diff 140131. yaxunl marked 2 inline comments as done. yaxunl added a comment. Use getAs instead of dyn_cast as John suggested. https://reviews.llvm.org/D44747 Files: include/clang/Basic/Specifiers.h lib/AST/ItaniumMangle.cpp lib/AST/Type.cpp lib/AST/TypeP

[PATCH] D44747: Set calling convention for CUDA kernel

yaxunl added a comment. Matt, are you OK with the change from amdgcn backend point of view? Thanks. Comment at: lib/Sema/SemaExpr.cpp:1669 +} + } + rjmccall wrote: > You should use `getAs` here (no `const` necessary). It's > possible to declare a functio

[PATCH] D44747: Set calling convention for CUDA kernel

rjmccall added a comment. LGTM. If `__global__` is supported in C++ structures, you might also need to make sure that member function constants (`&A::kernel_function`) drop the CC. And it might be a good idea to make sure that `decltype(kernel_function)` doesn't have a problem with it, either

[PATCH] D44747: Set calling convention for CUDA kernel

yaxunl updated this revision to Diff 140076. yaxunl marked 4 inline comments as done. yaxunl added a comment. Revised by John's comments. Drop CUDA kernel calling convention in DRE. https://reviews.llvm.org/D44747 Files: include/clang/Basic/Specifiers.h lib/AST/ItaniumMangle.cpp lib/AST/T

[PATCH] D44747: Set calling convention for CUDA kernel

rjmccall added inline comments. Comment at: lib/Sema/SemaOverload.cpp:1492 + Changed = true; +} + yaxunl wrote: > rjmccall wrote: > > It's cheaper not to check the CUDA language mode here; pulling the CC out > > of the FPT is easy. > > > > Why is this

[PATCH] D44747: Set calling convention for CUDA kernel

yaxunl marked 3 inline comments as done. yaxunl added inline comments. Comment at: lib/Sema/SemaOverload.cpp:1492 + Changed = true; +} + rjmccall wrote: > It's cheaper not to check the CUDA language mode here; pulling the CC out of > the FPT is easy. >

[PATCH] D44747: Set calling convention for CUDA kernel

rjmccall added inline comments. Comment at: lib/AST/Type.cpp:2762 + case CC_CUDAKernel: +return "cuda_kernel"; } For consistency with the rest of this switch, please put the return on the same line as its case. Comment at: lib/AST/Type

[PATCH] D44747: Set calling convention for CUDA kernel

yaxunl updated this revision to Diff 139625. yaxunl retitled this revision from "[AMDGPU] Set calling convention for CUDA kernel" to "Set calling convention for CUDA kernel". yaxunl edited the summary of this revision. yaxunl added a comment. Revised by John's comments. Introduce CC_CUDAKernel ca