rivanvx updated this revision to Diff 64920.
rivanvx marked 2 inline comments as done.
https://reviews.llvm.org/D20168
Files:
lib/CodeGen/TargetInfo.cpp
test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
Index: test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
==
rivanvx added a comment.
Addressed both concerns.
https://reviews.llvm.org/D20168
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
rivanvx updated this revision to Diff 64417.
rivanvx added a comment.
Specifically handle only kernels.
https://reviews.llvm.org/D20168
Files:
lib/CodeGen/TargetInfo.cpp
test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
Index: test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
===
rivanvx updated this revision to Diff 58920.
rivanvx added a comment.
Updated patch. Single element structs are coerced to its element, and there are
tests for structs of different sizes, structs of arrays, structs containing
structs. Arrays of structs are disallowed by clang in kernels.
Non-ker
rivanvx added a comment.
@Anastasia I looked into introducing a separate getOpenCLVersion() function (or
perhaps three - major version, minor version and version string). This would
have to be used in lib/CodeGen/TargetInfo.cpp and lib/Parse/ParseDecl.cpp, and
I am undecided on where should one
rivanvx updated this revision to Diff 57023.
rivanvx added a comment.
Now with 100% more tests.
http://reviews.llvm.org/D20168
Files:
lib/CodeGen/TargetInfo.cpp
test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
Index: test/CodeGenOpenCL/amdgpu-abi-struct-coerce.cl
rivanvx added a comment.
Could we solve that at a later point? There is one more place where such code
is already used, but this would enlarge the scope of this patch.
If yes, I am wiling to factor it out after this is merged.
http://reviews.llvm.org/D19780
_
rivanvx created this revision.
rivanvx added reviewers: arsenm, tstellarAMD.
rivanvx added a subscriber: cfe-commits.
Herald added a subscriber: kzhuravl.
Structs are currently handled as pointer + byval, which makes AMDGPU LLVM
backend generate incorrect code when structs are used. This patch ch
rivanvx added a comment.
Please, can anyone push this?
http://reviews.llvm.org/D19780
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
rivanvx added a comment.
Thanks for the reviews!
http://reviews.llvm.org/D19780
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
rivanvx updated this revision to Diff 56479.
rivanvx added a comment.
Make that int const as well.
http://reviews.llvm.org/D19780
Files:
include/clang/Basic/DiagnosticSemaKinds.td
lib/Parse/ParseDecl.cpp
test/Parser/opencl-cl20.cl
test/Parser/opencl-storage-class.cl
test/SemaOpenCL/in
rivanvx added inline comments.
Comment at: lib/Parse/ParseDecl.cpp:3519
@@ +3518,3 @@
+ / 100);
+const char *VerSpec = (VerMajor + std::string (".") +
VerMinor).c_str();
+Diag(Tok, DiagID) << VerSpec << PrevSpec << isS
rivanvx updated this revision to Diff 56478.
rivanvx marked 3 inline comments as done.
rivanvx added a comment.
I am neither aware how to convert ints to StringRef nor how to concatenate
StringRefs. Apologies if I missed something in the API.
In any case, this approach looks pretty clean to me.
rivanvx added a comment.
@Anastasia would you still prefer to make VerSpec a std::string?
http://reviews.llvm.org/D19780
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
14 matches
Mail list logo
