[PATCH] D31428: [x86] fix AVX FP cmp intrinsic documentation (PR28110)

spatel updated this revision to Diff 93366. spatel added a comment. Patch updated: Standardize the spelling of "non-signaling" and don't abbreviate "unordered". https://reviews.llvm.org/D31428 Files: lib/Headers/avxintrin.h Index: lib/Headers/avxintrin.h

[PATCH] D30643: [OpenCL] Extended diagnostics for atomic initialization

echuraev updated this revision to Diff 93368. echuraev marked 2 inline comments as done. https://reviews.llvm.org/D30643 Files: include/clang/Basic/DiagnosticSemaKinds.td lib/Sema/SemaExpr.cpp lib/Sema/SemaInit.cpp test/Parser/opencl-atomics-cl20.cl test/SemaOpenCL/atomic-init.cl

[Driver] Add option to print the resource directory

Hi, I tried to submit this patch via Phabricator yesterday, but I don't think it made it to the list: https://reviews.llvm.org/D31447 Thanks, Catherine ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D31460: [coroutines] Add cleanup for compiler injected objects/allocations in coroutine body

GorNishanov created this revision. - Use pushCleanup to emit freeing coroutine memory on normal and EH exits. - Surround emitted code with CodeGenFunction::RunCleanupsScope. https://reviews.llvm.org/D31460 Files: lib/CodeGen/CGCoroutine.cpp test/CodeGenCoroutines/coro-cleanup.cpp

Re: r298913 - Added `applyAtomicChanges` function.

Hi Juergen, thanks for taking care of this, but I'm wondering if this build bot is using a different set of build rules? The error message says "Clang_Tooling -> Clang_Format -> Clang_Tooling"; however, the actual dependency is clangToolingRefactor -> clangFormat -> clangToolingCore, which seems

[PATCH] D31417: [OpenMP] Add support for omp simd pragmas without runtime

rengolin added a comment. Hi Graham, I don't know much about Clang's machinery, but would it be possible to have `-fopenmp-simd` generate the same handler, but with restrictions? I fear this slight duplication could get considerably worse as we support more and more "non-RT" OMP pragmas.

[PATCH] D31417: [OpenMP] Add support for omp simd pragmas without runtime

huntergr added a comment. Hi Renato, In https://reviews.llvm.org/D31417#713162, @rengolin wrote: > I don't know much about Clang's machinery, but would it be possible to have > `-fopenmp-simd` generate the same handler, but with restrictions? I fear this > slight duplication could get

[PATCH] D31417: [OpenMP] Add support for omp simd pragmas without runtime

rengolin added a comment. In https://reviews.llvm.org/D31417#713171, @huntergr wrote: > The other alternative I thought of was to perform the filtering in > ParseOpenMP.cpp instead, but I need to figure out how to delete or skip > tokens there without cluttering up the rest of the OpenMP

[PATCH] D31378: [PCH] Attach instance's dependency collectors to PCH external AST sources.

bruno accepted this revision. bruno added a comment. Thanks for working on this! LGTM too https://reviews.llvm.org/D31378 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D31422: Add builder for libunwind docs

jroelofs closed this revision. jroelofs added a comment. r299003 https://reviews.llvm.org/D31422 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D31378: [PCH] Attach instance's dependency collectors to PCH external AST sources.

doug.gregor accepted this revision. doug.gregor added a comment. This revision is now accepted and ready to land. LGTM! https://reviews.llvm.org/D31378 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D31308: [clang-tidy] new check readability-no-alternative-tokens

aaron.ballman added inline comments. Comment at: clang-tidy/readability/OperatorsRepresentationCheck.cpp:34 + + if (const auto *B = Result.Nodes.getNodeAs("binary")) { +switch (B->getOpcode()) { mgehre wrote: > aaron.ballman wrote: > > alexfh wrote: > > >

[PATCH] D25157: [compiler-rt] [cmake] Respect COMPILER_RT_BUILD_* for libs, headers and tests

weimingz added a comment. Looks good to me but I'm not very familiar with the build of sanitizer and xray. https://reviews.llvm.org/D25157 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D31241: [Modules][PCH] Serialize #pragma pack

bruno accepted this revision. bruno added a comment. This revision is now accepted and ready to land. Thanks Alex. LGTM Repository: rL LLVM https://reviews.llvm.org/D31241 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

Re: r298956 - Default enable the rtm feature only on skylake and later for now because Intel disabled the feature on some haswell and broadwell processors:

OK, I went ahead and did this in the backend as well in: echristo@athyra ~/s/llvm> git svn dcommit Committing to https://llvm.org/svn/llvm-project/llvm/trunk ... M lib/Target/X86/X86.td Committed r298986 -eric On Tue, Mar 28, 2017 at 5:49 PM Jim Grosbach wrote: > SGTM > >

r299007 - Test Commit

Author: bkelley Date: Wed Mar 29 12:18:05 2017 New Revision: 299007 URL: http://llvm.org/viewvc/llvm-project?rev=299007=rev Log: Test Commit Remove trailing whitespace. Modified: cfe/trunk/lib/Sema/SemaDeclCXX.cpp Modified: cfe/trunk/lib/Sema/SemaDeclCXX.cpp URL:

r299008 - [Objective-C] C++ Classes with __weak Members non-POD Types when using -fobjc-weak

Author: bkelley Date: Wed Mar 29 12:31:42 2017 New Revision: 299008 URL: http://llvm.org/viewvc/llvm-project?rev=299008=rev Log: [Objective-C] C++ Classes with __weak Members non-POD Types when using -fobjc-weak Summary: When adding an Objective-C retainable type member to a C++ class, also

r299009 - [PCH] Attach instance's dependency collectors to PCH external AST sources.

Author: graydon Date: Wed Mar 29 12:33:09 2017 New Revision: 299009 URL: http://llvm.org/viewvc/llvm-project?rev=299009=rev Log: [PCH] Attach instance's dependency collectors to PCH external AST sources. Summary: When a PCH is included via -include-pch, clang should treat the current TU as

r299010 - [Objective-C] Fix __weak type traits with -fobjc-weak

Author: bkelley Date: Wed Mar 29 12:40:35 2017 New Revision: 299010 URL: http://llvm.org/viewvc/llvm-project?rev=299010=rev Log: [Objective-C] Fix __weak type traits with -fobjc-weak Summary: Similar to ARC, in ObjCWeak Objective-C object pointers qualified with a weak lifetime are not POD or

r299011 - [Objective-C] Fix "repeated use of weak" warning with -fobjc-weak

Author: bkelley Date: Wed Mar 29 12:55:11 2017 New Revision: 299011 URL: http://llvm.org/viewvc/llvm-project?rev=299011=rev Log: [Objective-C] Fix "repeated use of weak" warning with -fobjc-weak Summary: -Warc-repeated-use-of-weak should produce the same warnings with -fobjc-weak as it does

[PATCH] D31440: PR32382: Adapt to LLVM changes in DIExpression.

aprantl added a comment. In https://reviews.llvm.org/D31440#713308, @dblaikie wrote: > I'm a bit confused - the alloca was only emitted at -O0, by the looks of it. > Presumably it's pessimizing in some way at higher optimization levels? Or is > that not the case? I think it is really working

[PATCH] D29659: [OpenMP] Add flag for disabling the default generation of relocatable OpenMP target code for NVIDIA GPUs.

Hahnfeld added a comment. Having something with dashes behind other used prefixes (`CHK-PTXAS`, `CHK-PTXAS-C`, `CHK-PTXAS-C-RELO`) might not be optimal and break when other suffixes like `-NOT` or `-SAME` are added to lit. Please see inline about my suggestions but feel free to use others...

r299014 - [Objective-C] Fix "weak-unavailable" warning with -fobjc-weak

Author: bkelley Date: Wed Mar 29 13:09:02 2017 New Revision: 299014 URL: http://llvm.org/viewvc/llvm-project?rev=299014=rev Log: [Objective-C] Fix "weak-unavailable" warning with -fobjc-weak Summary: clang should produce the same errors Objective-C classes that cannot be assigned to weak

[PATCH] D29904: [OpenMP] Prevent emission of exception handling code when using OpenMP to offload to NVIDIA devices.

Hahnfeld added a comment. LGTM. Please run `clang-format` before committing! Comment at: test/OpenMP/target_parallel_no_exceptions.cpp:6-7 + +#define SIZE 100 +#define EPS 1e-10 + Not needed, please keep the test as small as possible Repository: rL

[PATCH] D29904: [OpenMP] Prevent emission of exception handling code when using OpenMP to offload to NVIDIA devices.

ABataev added inline comments. Comment at: lib/Frontend/CompilerInvocation.cpp:2167-2172 +// Set the flag to prevent the implementation from emitting device exception +// handling code for those requiring so. +if (Opts.OpenMPIsDevice && T.isNVPTX()) { +

[PATCH] D29599: Clang Changes for alloc_align

aaron.ballman added inline comments. Comment at: include/clang/Basic/Attr.td:1230 + let Spellings = [GCC<"alloc_align">]; + let Subjects = SubjectList<[ Function]>; + let Args = [IntArgument<"ParamIndex">]; There's a spurious space between [ and Function. If

[PATCH] D29642: [OpenMP] Make OpenMP generated code for the NVIDIA device relocatable by default

Hahnfeld added inline comments. Comment at: test/Driver/openmp-offload.c:598 +// CHK-PTXAS: ptxas{{.*}}" "-c" +// CHK-PTXAS-NEXT: /bin/cp This path might not be correct on all systems. Do we really need this check? Repository: rL LLVM

Re: r298956 - Default enable the rtm feature only on skylake and later for now because Intel disabled the feature on some haswell and broadwell processors:

Thanks, Eric. ~Craig On Wed, Mar 29, 2017 at 10:19 AM, Eric Christopher wrote: > OK, I went ahead and did this in the backend as well in: > > echristo@athyra ~/s/llvm> git svn dcommit > Committing to https://llvm.org/svn/llvm-project/llvm/trunk ... > M

r299015 - [Objective-C] Miscellaneous -fobjc-weak Fixes

Author: bkelley Date: Wed Mar 29 13:16:38 2017 New Revision: 299015 URL: http://llvm.org/viewvc/llvm-project?rev=299015=rev Log: [Objective-C] Miscellaneous -fobjc-weak Fixes Summary: After examining the remaining uses of LangOptions.ObjCAutoRefCount, found a some additional places to also

[PATCH] D30547: [clang-tidy] Forwarding reference overload in constructors

aaron.ballman added inline comments. Comment at: clang-tidy/misc/ForwardingReferenceOverloadCheck.cpp:119 + DisabledMove = false; + for (const auto *OtherCtor : Ctor->getParent()->ctors()) { +if (OtherCtor->isCopyConstructor()) { leanil wrote: > This

[PATCH] D31440: PR32382: Adapt to LLVM changes in DIExpression.

dblaikie added a comment. I'm a bit confused - the alloca was only emitted at -O0, by the looks of it. Presumably it's pessimizing in some way at higher optimization levels? Or is that not the case? Also, it looks like this change lost the "if > gmlt" test, so might cause variable

r299012 - Unbreak windows bot.

Author: graydon Date: Wed Mar 29 12:58:41 2017 New Revision: 299012 URL: http://llvm.org/viewvc/llvm-project?rev=299012=rev Log: Unbreak windows bot. Modified: cfe/trunk/test/PCH/emit-dependencies.c Modified: cfe/trunk/test/PCH/emit-dependencies.c URL:

[PATCH] D25157: [compiler-rt] [cmake] Respect COMPILER_RT_BUILD_* for libs, headers and tests

compnerd added inline comments. Comment at: cmake/config-ix.cmake:438-441 if (SANITIZER_COMMON_SUPPORTED_ARCH AND NOT LLVM_USE_SANITIZER AND +(COMPILER_RT_BUILD_SANITIZERS OR COMPILER_RT_BUILD_XRAY) AND (OS_NAME MATCHES "Android|Darwin|Linux|FreeBSD" OR (OS_NAME

[PATCH] D25157: [compiler-rt] [cmake] Respect COMPILER_RT_BUILD_* for libs, headers and tests

mgorny added inline comments. Comment at: cmake/config-ix.cmake:438-441 if (SANITIZER_COMMON_SUPPORTED_ARCH AND NOT LLVM_USE_SANITIZER AND +(COMPILER_RT_BUILD_SANITIZERS OR COMPILER_RT_BUILD_XRAY) AND (OS_NAME MATCHES "Android|Darwin|Linux|FreeBSD" OR (OS_NAME

[PATCH] D31378: [PCH] Attach instance's dependency collectors to PCH external AST sources.

This revision was automatically updated to reflect the committed changes. Closed by commit rL299009: [PCH] Attach instance's dependency collectors to PCH external AST sources. (authored by graydon). Changed prior to commit: https://reviews.llvm.org/D31378?vs=93089=93387#toc Repository: rL

[PATCH] D29644: [OpenMP] Pass -v to PTXAS if it was passed to the driver.

Hahnfeld added inline comments. Comment at: test/Driver/openmp-offload.c:607 +// CHK-VERBOSE: ptxas{{.*}}" "-v" +// CHK-VERBOSE-NEXT: /bin/cp This path might not be correct on all systems. Do we really need this check? Repository: rL LLVM

[PATCH] D29599: Clang Changes for alloc_align

erichkeane marked 9 inline comments as done. erichkeane added inline comments. Comment at: lib/Sema/SemaDeclAttr.cpp:1608-1612 + IndexVal += 1 + isInstanceMethod(FuncDecl); + + if (!checkParamIsIntegerType(*this, FuncDecl, TmpAttr, ParamExpr, IndexVal, +

[PATCH] D29599: Clang Changes for alloc_align

rjmccall added a comment. I see that GCC is up to its same parameter-indexing shenanigans again. Comment at: include/clang/Basic/AttrDocs.td:252 +declaration to specify that the return value of the function (which must be a +pointer type) has an alignment specified by the

[PATCH] D31272: Do not pass an explicit reexported symbol list when building libc++ dylib if also defining new/delete

mehdi_amini added inline comments. Comment at: lib/CMakeLists.txt:155 +# We can't use the "-reexported_symbols_list" when we build the +# new/delete operators as part of the dylib: the linker would fail. +set(OSX_RE_EXPORT_LINE

[PATCH] D17053: [libcxx]: vector: Use < instead of != to improve failure mode

EricWF resigned from this revision. EricWF added a comment. This review seems stuck and dead. Resigning as a reviewer. Please re-add me if you want this revision to proceed. https://reviews.llvm.org/D17053 ___ cfe-commits mailing list

[libcxx] r299052 - Do not pass an explicit reexported symbol list when building libc++ dylib if also defining new/delete

Author: mehdi_amini Date: Wed Mar 29 23:40:56 2017 New Revision: 299052 URL: http://llvm.org/viewvc/llvm-project?rev=299052=rev Log: Do not pass an explicit reexported symbol list when building libc++ dylib if also defining new/delete Modified: libcxx/trunk/lib/CMakeLists.txt Modified:

[libcxx] r299053 - libc++ testing: allow to provide a path for `use_system_cxx_lib`

Author: mehdi_amini Date: Wed Mar 29 23:45:33 2017 New Revision: 299053 URL: http://llvm.org/viewvc/llvm-project?rev=299053=rev Log: libc++ testing: allow to provide a path for `use_system_cxx_lib` As we're trying to setup testing / bots for all shipping version of libc++ on macOS/iOS, we'll

[PATCH] D31486: libc++ testing: allow to provide a path for `use_system_cxx_lib`

This revision was automatically updated to reflect the committed changes. Closed by commit rL299053: libc++ testing: allow to provide a path for `use_system_cxx_lib` (authored by mehdi_amini). Changed prior to commit: https://reviews.llvm.org/D31486?vs=93432=93444#toc Repository: rL LLVM

[PATCH] D31272: Do not pass an explicit reexported symbol list when building libc++ dylib if also defining new/delete

mehdi_amini closed this revision. mehdi_amini added a comment. r299052 ; let me know if you want to improve the comment in any way. https://reviews.llvm.org/D31272 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D30765: Reexport operator new / delete from libc++abi

This revision was automatically updated to reflect the committed changes. Closed by commit rL299054: Reexport operator new / delete from libc++abi (authored by mehdi_amini). Changed prior to commit: https://reviews.llvm.org/D30765?vs=91091=93445#toc Repository: rL LLVM

[libcxx] r299054 - Reexport operator new / delete from libc++abi

Author: mehdi_amini Date: Wed Mar 29 23:47:19 2017 New Revision: 299054 URL: http://llvm.org/viewvc/llvm-project?rev=299054=rev Log: Reexport operator new / delete from libc++abi Both libc++ and libc++abi export a weak definition of operator new/delete. On Darwin, this can often cause dirty

[libcxx] r299055 - libc++ config testing: `use_system_cxx_lib` can be a bool

Author: mehdi_amini Date: Wed Mar 29 23:51:19 2017 New Revision: 299055 URL: http://llvm.org/viewvc/llvm-project?rev=299055=rev Log: libc++ config testing: `use_system_cxx_lib` can be a bool Fix the libc++ Green Dragon bot. Modified: libcxx/trunk/utils/libcxx/test/config.py Modified:

[PATCH] D17469: [libcxx] Add deployment knobs to tests (for Apple platforms)

mehdi_amini closed this revision. mehdi_amini added a comment. Was committed a while back in r297798 https://reviews.llvm.org/D17469 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[libcxx] r299056 - libc++ testing: fix assertion around `use_system_cxx_lib`

Author: mehdi_amini Date: Thu Mar 30 00:07:18 2017 New Revision: 299056 URL: http://llvm.org/viewvc/llvm-project?rev=299056=rev Log: libc++ testing: fix assertion around `use_system_cxx_lib` Actually fix (hopefully) the assertions about `use_system_cxx_lib`, the previous attempt failed because I

r299033 - Use FPContractModeKind universally

Author: anemet Date: Wed Mar 29 16:54:24 2017 New Revision: 299033 URL: http://llvm.org/viewvc/llvm-project?rev=299033=rev Log: Use FPContractModeKind universally FPContractModeKind is the codegen option flag which is already ternary (off, on, fast). This makes it universally the type for the

[PATCH] D31460: [coroutines] Add cleanup for compiler injected objects/allocations in coroutine body

GorNishanov added inline comments. Comment at: lib/CodeGen/CGCoroutine.cpp:225 + void Emit(CodeGenFunction , Flags) override { +CGF.EmitStmt(Deallocate); + } rnk wrote: > This will be called twice: once for a normal exit and once for exceptional > exit.

[PATCH] D30760: Record command lines in objects built by clang, Clang part

zhizhouy updated this revision to Diff 93401. zhizhouy marked 2 inline comments as done. zhizhouy edited the summary of this revision. zhizhouy added a comment. Added two more testcases, one is options with both grecord-gcc-switches and gno-record-gcc-switches; the other one is testing if "-o -"

[PATCH] D31167: Use FPContractModeKind universally

aaron.ballman accepted this revision. aaron.ballman added a comment. This revision is now accepted and ready to land. Aside from a minor comment nit, LGTM Comment at: include/clang/Basic/LangOptions.h:92 + enum FPContractModeKind { +FPC_Off,// Form fused FP ops

r299029 - Revert "Use FPContractModeKind universally"

Author: anemet Date: Wed Mar 29 16:24:19 2017 New Revision: 299029 URL: http://llvm.org/viewvc/llvm-project?rev=299029=rev Log: Revert "Use FPContractModeKind universally" This reverts commit r299027. It's causing a test failure in clang's CodeGenCUDE/fp-contract.cu Modified:

[PATCH] D29599: Clang Changes for alloc_align

erichkeane updated this revision to Diff 93409. https://reviews.llvm.org/D29599 Files: include/clang/Basic/Attr.td include/clang/Basic/AttrDocs.td include/clang/Sema/Sema.h lib/CodeGen/CGCall.cpp lib/CodeGen/CodeGenFunction.h lib/Sema/SemaDeclAttr.cpp

[PATCH] D29599: Clang Changes for alloc_align

erichkeane marked 2 inline comments as done. erichkeane added inline comments. Comment at: include/clang/Basic/AttrDocs.td:252 +declaration to specify that the return value of the function (which must be a +pointer type) has an alignment specified by the indicated parameter. The

[PATCH] D31460: [coroutines] Add cleanup for compiler injected objects/allocations in coroutine body

rnk added inline comments. Comment at: lib/CodeGen/CGCoroutine.cpp:225 + void Emit(CodeGenFunction , Flags) override { +CGF.EmitStmt(Deallocate); + } GorNishanov wrote: > rnk wrote: > > This will be called twice: once for a normal exit and once for

[PATCH] D31482: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set

kzhuravl created this revision. Herald added subscribers: t-tye, tpr, dstuttard, yaxunl, wdng. https://reviews.llvm.org/D31482 Files: llvm/tools/clang/lib/Basic/Targets.cpp llvm/tools/clang/test/CodeGenOpenCL/gfx9-fp32-denorms.cl Index:

[PATCH] D31482: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set

arsenm added inline comments. Comment at: llvm/tools/clang/lib/Basic/Targets.cpp:2114-2116 + static bool hasFullSpeedFP32Denorms(StringRef GPUName) { +return parseAMDGCNName(GPUName) >= GK_GFX9; + } This is misleading since it was true on VI as well. I

[PATCH] D30760: Record command lines in objects built by clang, Clang part

zhizhouy updated this revision to Diff 93403. zhizhouy marked 2 inline comments as done. zhizhouy added a comment. Added testcase for recording other useful options. https://reviews.llvm.org/D30760 Files: lib/Driver/ToolChains/Clang.cpp test/Driver/debug-options.c Index:

[PATCH] D29599: Clang Changes for alloc_align

erichkeane marked 2 inline comments as done. erichkeane added inline comments. Comment at: include/clang/Basic/AttrDocs.td:252 +declaration to specify that the return value of the function (which must be a +pointer type) has an alignment specified by the indicated parameter,

[PATCH] D31167: Use FPContractModeKind universally

anemet added inline comments. Comment at: include/clang/Basic/LangOptions.h:92 + enum FPContractModeKind { +FPC_Off,// Form fused FP ops only where result will not be affected. +FPC_On, // Form fused FP ops according to FP_CONTRACT rules.

[PATCH] D31460: [coroutines] Add cleanup for compiler injected objects/allocations in coroutine body

rnk accepted this revision. rnk added a comment. This revision is now accepted and ready to land. lgtm Comment at: lib/CodeGen/CGCoroutine.cpp:225 + void Emit(CodeGenFunction , Flags) override { +CGF.EmitStmt(Deallocate); + } This will be called twice:

[PATCH] D30760: Record command lines in objects built by clang, Clang part

echristo added inline comments. Comment at: test/Driver/debug-options.c:201-202 // +// GRECORD: "-dwarf-debug-flags" +// GRECORD: -### -c -grecord-gcc-switches +// george.burgess.iv wrote: > echristo wrote: > > This seems a little light on the testing, would

Re: [PATCH] D30760: Record command lines in objects built by clang, Clang part

Not really. Would you please help me commit it? Also there is another diff for LLVM part of this bug. Thanks. On Wed, Mar 29, 2017 at 1:16 PM, Eric Christopher via Phabricator < revi...@reviews.llvm.org> wrote: > echristo accepted this revision. > echristo added a comment. > This revision is

[PATCH] D31167: Use FPContractModeKind universally

aaron.ballman added inline comments. Comment at: include/clang/Basic/LangOptions.h:92 + enum FPContractModeKind { +FPC_Off,// Form fused FP ops only where result will not be affected. +FPC_On, // Form fused FP ops according to FP_CONTRACT rules.

[PATCH] D30760: Record command lines in objects built by clang, Clang part

echristo accepted this revision. echristo added a comment. This revision is now accepted and ready to land. Sounds good. Do you have commit access? https://reviews.llvm.org/D30760 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D29599: Clang Changes for alloc_align

erichkeane updated this revision to Diff 93404. erichkeane added a comment. Made the changes as requested. checkFunctionOrMethodParameterIndex corrects for 1->0 index and implicit this, which requires undoing, otherwise templates create a big hassle. Additionally, please note the AttrDocs

r299027 - Use FPContractModeKind universally

Author: anemet Date: Wed Mar 29 15:39:49 2017 New Revision: 299027 URL: http://llvm.org/viewvc/llvm-project?rev=299027=rev Log: Use FPContractModeKind universally FPContractModeKind is the codegen option flag which is already ternary (off, on, fast). This makes it universally the type for the

[PATCH] D31167: Use FPContractModeKind universally

This revision was automatically updated to reflect the committed changes. Closed by commit rL299027: Use FPContractModeKind universally (authored by anemet). Changed prior to commit: https://reviews.llvm.org/D31167?vs=92423=93406#toc Repository: rL LLVM https://reviews.llvm.org/D31167

[PATCH] D29599: Clang Changes for alloc_align

aaron.ballman added inline comments. Comment at: include/clang/Basic/AttrDocs.td:252 +declaration to specify that the return value of the function (which must be a +pointer type) has an alignment specified by the indicated parameter. The +alignment parameter is one-indexed. In

[PATCH] D31482: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set

kzhuravl added inline comments. Comment at: llvm/tools/clang/lib/Basic/Targets.cpp:2114-2116 + static bool hasFullSpeedFP32Denorms(StringRef GPUName) { +return parseAMDGCNName(GPUName) >= GK_GFX9; + } arsenm wrote: > This is misleading since it was true on

[PATCH] D29599: Clang Changes for alloc_align

erichkeane updated this revision to Diff 93421. erichkeane marked 2 inline comments as done. erichkeane added a comment. Fixes based on John's comments. A little bit of extra work was required to get the correct Value from the attribute, but impact was minimal. https://reviews.llvm.org/D29599

[PATCH] D30388: [XRay] Add -fxray-{always, never}-instrument= flags to clang

This revision was automatically updated to reflect the committed changes. dberris marked an inline comment as done. Closed by commit rL299041: [XRay] Add -fxray-{always,never}-instrument= flags to clang (authored by dberris). Changed prior to commit:

r299041 - [XRay] Add -fxray-{always, never}-instrument= flags to clang

Author: dberris Date: Wed Mar 29 19:29:36 2017 New Revision: 299041 URL: http://llvm.org/viewvc/llvm-project?rev=299041=rev Log: [XRay] Add -fxray-{always,never}-instrument= flags to clang Summary: The -fxray-always-instrument= and -fxray-never-instrument= flags take filenames that are used to

[PATCH] D31486: libc++ testing: allow to provide a path for `use_system_cxx_lib`

mehdi_amini created this revision. As we're trying to setup testing / bots for all shipping version of libc++ on macOS/iOS, we'll need to be able to pass a path to where to find the dylib for each previous version of the OS. https://reviews.llvm.org/D31486 Files: utils/libcxx/test/config.py

[PATCH] D29599: Clang Changes for alloc_align

rjmccall added inline comments. Comment at: include/clang/Basic/AttrDocs.td:252 +declaration to specify that the return value of the function (which must be a +pointer type) is at least as aligned as the value indicated parameter. The +parameter is given by its index in the

Re: [PATCH] D30388: [XRay] Add -fxray-{always,never}-instrument= flags to clang

Oops, thanks -- yes, I'll make that change now. On Thu, Mar 30, 2017 at 12:03 PM Andrew Ford via Phabricator < revi...@reviews.llvm.org> wrote: > andrewford added a comment. > > Hi, this breaks the android build due to use of std::to_string. > llvm::to_string should be used instead. I would fix

Re: [PATCH] D30388: [XRay] Add -fxray-{always,never}-instrument= flags to clang

Fix committed in r299044. On Thu, Mar 30, 2017 at 12:12 PM Dean Michael Berris wrote: > Oops, thanks -- yes, I'll make that change now. > > On Thu, Mar 30, 2017 at 12:03 PM Andrew Ford via Phabricator < > revi...@reviews.llvm.org> wrote: > > andrewford added a comment. > >

r299044 - [XRay][clang] Use llvm::to_string instead of std::string

Author: dberris Date: Wed Mar 29 20:05:09 2017 New Revision: 299044 URL: http://llvm.org/viewvc/llvm-project?rev=299044=rev Log: [XRay][clang] Use llvm::to_string instead of std::string This should unbreak some bots. Follow-up on D30388. Modified: cfe/trunk/lib/Driver/XRayArgs.cpp

r299045 - Use 'unsigned' for enum bitfields

Author: rnk Date: Wed Mar 29 20:12:08 2017 New Revision: 299045 URL: http://llvm.org/viewvc/llvm-project?rev=299045=rev Log: Use 'unsigned' for enum bitfields Fixes this clang warning on Windows: warning: implicit truncation from 'clang::LangOptions::FPContractModeKind' to bit-field changes

[PATCH] D31487: [coroutines] Fix rebuilding of implicit and dependent coroutine statements.

EricWF created this revision. Certain implicitly generated coroutine statements, such as the calls to 'return_value()' or `return_void()` or `get_return_object_on_allocation_failure()`, cannot be built until the promise type is no longer dependent. This means they are not built until after the

[PATCH] D31272: Do not pass an explicit reexported symbol list when building libc++ dylib if also defining new/delete

EricWF accepted this revision. EricWF added inline comments. This revision is now accepted and ready to land. Comment at: lib/CMakeLists.txt:155 +# We can't use the "-reexported_symbols_list" when we build the +# new/delete operators as part of the dylib: the

Re: [libcxx] r281681 - [libc++] Add _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY to support GCC ABI compatibility

On Wed, Mar 29, 2017 at 9:00 PM, Duncan P. N. Exon Smith < dexonsm...@apple.com> wrote: > Why are we propagating the use of always_inline? > The intent of this change wasn't to propagate or remove always_inline from any functions. It was to fix incompatible dylibs built by GCC. The way it did

[PATCH] D31486: libc++ testing: allow to provide a path for `use_system_cxx_lib`

EricWF accepted this revision. EricWF added a comment. This revision is now accepted and ready to land. LGTM. https://reviews.llvm.org/D31486 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

Re: [libcxx] r281681 - [libc++] Add _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY to support GCC ABI compatibility

On Wed, Mar 29, 2017 at 9:30 PM, Duncan P. N. Exon Smith < dexonsm...@apple.com> wrote: > > On Mar 29, 2017, at 20:16, Eric Fiselier wrote: > > > > On Wed, Mar 29, 2017 at 9:00 PM, Duncan P. N. Exon Smith < > dexonsm...@apple.com> wrote: > >> Why are we propagating the use of

Re: [libcxx] r281681 - [libc++] Add _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY to support GCC ABI compatibility

Hi, I don't understand clang's logic here, seems like a bug to me. I changed slightly the test-case, and I'm wondering why only foo() is emitted as hidden in the following: #define INLINE_VISIBILITY __attribute__((visibility("hidden"), always_inline)) template struct Foo { void

Re: [libcxx] r281681 - [libc++] Add _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY to support GCC ABI compatibility

2017-03-29 20:30 GMT-07:00 Duncan P. N. Exon Smith : > > On Mar 29, 2017, at 20:16, Eric Fiselier wrote: > > > > On Wed, Mar 29, 2017 at 9:00 PM, Duncan P. N. Exon Smith < > dexonsm...@apple.com> wrote: > >> Why are we propagating the use of always_inline? >>

[PATCH] D31487: [coroutines] Fix rebuilding of implicit and dependent coroutine statements.

EricWF updated this revision to Diff 93441. EricWF added a comment. - Remove unneeded asserts. https://reviews.llvm.org/D31487 Files: include/clang/AST/StmtCXX.h lib/AST/StmtCXX.cpp lib/Sema/CoroutineBuilder.h lib/Sema/SemaCoroutine.cpp lib/Sema/TreeTransform.h

Re: [libcxx] r281681 - [libc++] Add _LIBCPP_EXTERN_TEMPLATE_INLINE_VISIBILITY to support GCC ABI compatibility

Why are we propagating the use of always_inline? In other places, we use always_inline to avoid affecting ABI (a visibility hack). But you're not removing basic_string from the dylib here. It has major downsides: - It causes inlining at -O0, which causes major regressions in debugging

[PATCH] D31167: Use FPContractModeKind universally

anemet added inline comments. Comment at: cfe/trunk/include/clang/Basic/LangOptions.h:217 /// Adjust BinaryOperator::FPFeatures to match the bit-field size of this. - unsigned fp_contract : 1; + LangOptions::FPContractModeKind fp_contract : 2; }; rnk

[libcxx] r299057 - libc++ testing: fix invalid python syntax

Author: mehdi_amini Date: Thu Mar 30 00:21:33 2017 New Revision: 299057 URL: http://llvm.org/viewvc/llvm-project?rev=299057=rev Log: libc++ testing: fix invalid python syntax Modified: libcxx/trunk/utils/libcxx/test/config.py Modified: libcxx/trunk/utils/libcxx/test/config.py URL:

[PATCH] D31460: [coroutines] Add cleanup for compiler injected objects/allocations in coroutine body

GorNishanov added inline comments. Comment at: lib/CodeGen/CGCoroutine.cpp:225 + void Emit(CodeGenFunction , Flags) override { +CGF.EmitStmt(Deallocate); + } rnk wrote: > GorNishanov wrote: > > rnk wrote: > > > This will be called twice: once for a normal

[PATCH] D31482: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set

kzhuravl updated this revision to Diff 93422. kzhuravl marked an inline comment as done. kzhuravl added a comment. Address review feedback. https://reviews.llvm.org/D31482 Files: lib/Basic/Targets.cpp test/CodeGenOpenCL/gfx9-fp32-denorms.cl Index: test/CodeGenOpenCL/gfx9-fp32-denorms.cl

r299037 - Add the -grecord-gcc-switches option and pass the flags down on the compile unit.

Author: echristo Date: Wed Mar 29 18:34:20 2017 New Revision: 299037 URL: http://llvm.org/viewvc/llvm-project?rev=299037=rev Log: Add the -grecord-gcc-switches option and pass the flags down on the compile unit. Patch by Zhizhou Yang Modified: cfe/trunk/lib/Driver/ToolChains/Clang.cpp

[PATCH] D30760: Record command lines in objects built by clang, Clang part

echristo added a comment. Committed thusly: echristo@athyra ~/s/l/t/clang> git svn dcommit Committing to https://llvm.org/svn/llvm-project/cfe/trunk ... M lib/Driver/ToolChains/Clang.cpp M test/Driver/debug-options.c Committed r299037 https://reviews.llvm.org/D30760

[PATCH] D30388: [XRay] Add -fxray-{always, never}-instrument= flags to clang

rnk accepted this revision. rnk added a comment. This revision is now accepted and ready to land. lgtm Comment at: include/clang/Driver/XRayArgs.h:22 +class XRayArgs { + std::vector AlwaysInstrumenFiles; + std::vector NeverInstrumentFiles; Any reason to omit

[PATCH] D30388: [XRay] Add -fxray-{always, never}-instrument= flags to clang

dberris added inline comments. Comment at: include/clang/Driver/XRayArgs.h:22 +class XRayArgs { + std::vector AlwaysInstrumenFiles; + std::vector NeverInstrumentFiles; rnk wrote: > Any reason to omit the 't' in "InstrumentFiles"? Wow, good catch -- no good

[PATCH] D31482: [AMDGPU][GFX9] Set +fp32-denormals for >=gfx900 unless -cl-denorms-are-zero is set

arsenm added inline comments. Comment at: llvm/tools/clang/lib/Basic/Targets.cpp:2114-2116 + static bool hasFullSpeedFP32Denorms(StringRef GPUName) { +return parseAMDGCNName(GPUName) >= GK_GFX9; + } kzhuravl wrote: > arsenm wrote: > > This is misleading

[PATCH] D31167: Use FPContractModeKind universally

rnk added inline comments. Comment at: cfe/trunk/include/clang/Basic/LangOptions.h:217 /// Adjust BinaryOperator::FPFeatures to match the bit-field size of this. - unsigned fp_contract : 1; + LangOptions::FPContractModeKind fp_contract : 2; }; Please do

  1   2   >