[PATCH] D39505: [OpenMP] Show error if VLAs are not supported

Hahnfeld updated this revision to Diff 123364. Hahnfeld added a comment. Update changes to be generic. https://reviews.llvm.org/D39505 Files: include/clang/Basic/DiagnosticSemaKinds.td include/clang/Basic/TargetInfo.h include/clang/Sema/Sema.h lib/Basic/TargetInfo.cpp lib/Basic/Target

[PATCH] D39505: [OpenMP] Show error if VLAs are not supported

Hahnfeld marked an inline comment as done. Hahnfeld added inline comments. Comment at: include/clang/Basic/TargetInfo.h:944 + /// \brief Whether target supports variable-length arrays. + bool isVLASupported() const { return VLASupported; } + rjmccall wrote: > H

[PATCH] D39505: [OpenMP] Show error if VLAs are not supported

This revision was automatically updated to reflect the committed changes. Hahnfeld marked an inline comment as done. Closed by commit rL318601: [OpenMP] Show error if VLAs are not supported (authored by Hahnfeld). Changed prior to commit: https://reviews.llvm.org/D39505?vs=123364&id=123473#toc

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld added a comment. A different approach would be to completely replace the type in `Driver::ConstructPhaseAction` but 1. this was previously considered a too radical change, 2. we currently don't have the necessary information (ToolChain, OffloadKind) to take that decision, 3. this might

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld created this revision. This was previously done in some places, but for example not for bundling so that single object compilation with -c failed. In addition cubin was used for all file types during unbundling which is incorrect for assembly files that are passed to ptxas. Tighten up the

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld marked 2 inline comments as done. Hahnfeld added inline comments. Comment at: lib/Driver/ToolChains/Cuda.cpp:431 -SmallString<256> Name(II.getFilename()); -llvm::sys::path::replace_extension(Name, "cubin"); - -const char *CubinF = -C.addTempFile(C.g

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld marked 2 inline comments as done. Hahnfeld added inline comments. Comment at: lib/Driver/ToolChains/Cuda.cpp:431 -SmallString<256> Name(II.getFilename()); -llvm::sys::path::replace_extension(Name, "cubin"); - -const char *CubinF = -C.addTempFile(C.g

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld added inline comments. Comment at: lib/Driver/ToolChains/Clang.cpp:5341-5344 +if (const auto *OA = dyn_cast(JA.getInputs()[I])) { + OA->doOnEachDependence( + [&](Action *, const ToolChain *TC, const char *) { CurTC = TC; }); +} tra

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

Hahnfeld updated this revision to Diff 123649. Hahnfeld marked 13 inline comments as done. Hahnfeld added a comment. Address reviewers' comments. https://reviews.llvm.org/D40250 Files: include/clang/Driver/ToolChain.h lib/Driver/ToolChain.cpp lib/Driver/ToolChains/Clang.cpp lib/Driver/T

[PATCH] D40275: [CUDA] Report "unsupported VLA" errors only on device side.

Hahnfeld requested changes to this revision. Hahnfeld added a comment. This revision now requires changes to proceed. In https://reviews.llvm.org/D39505 @rjmccall requested that the check should be made independent of the language. To preserve this, I think the CUDA specific checks should be add

[PATCH] D40275: [CUDA] Report "unsupported VLA" errors only on device side.

Hahnfeld added a comment. And please add a regression test which is apparently missing for the case that a VLA is **NOT** diagnosed in CUDA mode https://reviews.llvm.org/D40275 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.ll

[PATCH] D40275: [CUDA] Report "unsupported VLA" errors only on device side.

Hahnfeld added a comment. In https://reviews.llvm.org/D40275#930981, @tra wrote: > In https://reviews.llvm.org/D40275#930948, @Hahnfeld wrote: > > > In https://reviews.llvm.org/D39505 @rjmccall requested that the check > > should be made independent of the language. To preserve this, I think the

[PATCH] D40250: [OpenMP] Consistently use cubin extension for nvlink

This revision was automatically updated to reflect the committed changes. Hahnfeld marked an inline comment as done. Closed by commit rL318763: [OpenMP] Consistently use cubin extension for nvlink (authored by Hahnfeld). Changed prior to commit: https://reviews.llvm.org/D40250?vs=123649&id=1237

[PATCH] D146497: libclang: Pass Clang install directory to driver via argv[0].

Hahnfeld added a comment. FYI this commit breaks `clang/test/Index/index-file.cu` for me: warning: CUDA version 11.8 is only partially supported [-Wunknown-cuda-version] warning: CUDA version 11.8 is only partially supported [-Wunknown-cuda-version] /home/jhahnfel/LLVM/src/clang/test/Inde

[PATCH] D146497: libclang: Pass Clang install directory to driver via argv[0].

Hahnfeld added a comment. In D146497#4216226 , @aaron.ballman wrote: > The issue doesn't reproduce for me locally on my Windows machine, so it may > be something specific to the VE setup. FWIW I'm not doing anything specific for VE, just set(CMAKE_C

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld abandoned this revision. Hahnfeld added a comment. Ok, I understand that fixing `ODRHash` is the wrong approach for our needs - we'll likely need to implement a custom hashing of template arguments to work as a lookup for lazy loading. Repository: rG LLVM Github Monorepo CHANGES SI

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld added a comment. In D153003#4458323 , @ChuanqiXu wrote: > In D153003#4456595 , @rsmith wrote: > >> I think the behavior change for the testcase here is correct, though I'm not >> sure that the patch is g

[PATCH] D146389: [clang-repl][CUDA] Initial interactive CUDA support for clang-repl

Hahnfeld added inline comments. Comment at: clang/include/clang/Interpreter/Interpreter.h:43-44 public: static llvm::Expected> create(std::vector &ClangArgv); + static llvm::Expected> If I understand the change correctly, the "old" `create` function on it

[PATCH] D146389: [clang-repl][CUDA] Initial interactive CUDA support for clang-repl

Hahnfeld added inline comments. Comment at: clang/lib/CodeGen/CodeGenAction.cpp:296 } + LinkModules.clear(); return false; // success v.g.vassilev wrote: > argentite wrote: > > Hahnfeld wrote: > > > This looks like a change that has implication

[PATCH] D146389: [clang-repl][CUDA] Initial interactive CUDA support for clang-repl

Hahnfeld added a comment. Some comments, but otherwise LGTM Comment at: clang/include/clang/Interpreter/Interpreter.h:43 public: + IncrementalCompilerBuilder(){}; + and this should probably be run through `clang-format`... Comment at: clang

[PATCH] D149551: [Interpreter] Filter out RISC-V +relax feature

Hahnfeld abandoned this revision. Hahnfeld added a comment. Linker relaxation and proper alignment handling for RISC-V is implemented in https://reviews.llvm.org/D149526, so this isn't needed anymore. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D1

[PATCH] D150773: [clang][modules] Add features for recent C++ versions

Hahnfeld created this revision. Hahnfeld added reviewers: ChuanqiXu, Bigcheese, v.g.vassilev, aaron.ballman. Herald added a project: All. Hahnfeld requested review of this revision. Herald added a project: clang. Herald added a subscriber: cfe-commits. Add `cplusplus20`, `cplusplus23`, and `cplusp

[PATCH] D150773: [clang][modules] Add features for recent C++ versions

Hahnfeld updated this revision to Diff 523021. Hahnfeld added a comment. > LGTM, but please add a release note when landing. There wasn't a category specific to modules, so I just added two entries for C++20 and C++23 - ok? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D150773/new/ htt

[PATCH] D150773: [clang][modules] Add features for recent C++ versions

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG53c03a3db16c: [clang][modules] Add features for recent C++ versions (authored by Hahnfeld). Repository: rG LLVM Github Monorepo CHANGES SINCE LAS

[PATCH] D149551: [Interpreter] Filter out RISC-V +relax feature

Hahnfeld created this revision. Hahnfeld added reviewers: v.g.vassilev, sgraenitz, lhames, StephenFan. Herald added subscribers: VincentWu, vkmr, luismarques, sameer.abuasal, s.egerton, Jim, PkmX, rogfer01, shiva0217, kito-cheng, simoncook, arichardson. Herald added a project: All. Hahnfeld reques

[PATCH] D152995: Remove clang/ModuleInfo.txt

Hahnfeld created this revision. Hahnfeld added reviewers: rsmith, aaron.ballman. Herald added a project: All. Hahnfeld requested review of this revision. Herald added a project: clang. Herald added a subscriber: cfe-commits. The script `build-for-llvm-top.sh` and LLVM's `ModuleInfo.txt` are gone s

[PATCH] D152995: Remove clang/ModuleInfo.txt

Hahnfeld added a comment. Just checking to be sure there is no other use case that the file is used for, but I doubt it... Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D152995/new/ https://reviews.llvm.org/D152995

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld created this revision. Hahnfeld added reviewers: rtrieu, vsapsai, ChuanqiXu, Bigcheese. Herald added a project: All. Hahnfeld requested review of this revision. Herald added a project: clang. Herald added a subscriber: cfe-commits. For hashing, we should not differentiate between template

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld added a comment. In D153003#4423926 , @ChuanqiXu wrote: > This looks not so good. In this way, we can't disambiguate other template > types. At least we added the kind and a TODO here. Which template name types would we need to disambiguate? We

[PATCH] D152995: Remove clang/ModuleInfo.txt

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGce8ff3facc63: Remove clang/ModuleInfo.txt (authored by Hahnfeld). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld added a comment. > For example, the template name with QualifiedTemplate kind has different hash > name with the name with DependentTemplate kind. But it is not true after the > patch. Yes, I do understand. However, what I'm asking is why we need to differentiate between the two, ie i

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld added a comment. In D153003#4427412 , @ChuanqiXu wrote: > An important node here is that ODRHash is used to check the AST Nodes are > keeping the same across compilations. There is gap to use ODRHash to check > the semantical equality. Oh ok..

[PATCH] D153003: [ODRHash] Fix ODR hashing of template names

Hahnfeld added a comment. > We can't say this abstractly. It should be fine to work in ODRHash for C++20 > modules issues as long as we don't lose informations. I honestly don't understand this: For the approach to work, we would need to hash the two cases of an unqualified and a qualified temp

[PATCH] D138792: [AArch64] Improve TargetParser API

Hahnfeld added inline comments. Comment at: llvm/include/llvm/Support/AArch64TargetParser.h:154-160 +// Create ArchInfo structs named +#define AARCH64_ARCH(MAJOR, MINOR, PROFILE, NAME, ID, ARCH_FEATURE, \ + ARCH_BASE_EXT)

[PATCH] D138792: [AArch64] Improve TargetParser API

Hahnfeld added a comment. In D138792#3969310 , @mgorny wrote: > In D138792#3966920 , @Hahnfeld > wrote: > >> Hi, I bisected this change to lead to a couple of test failures when >> building with `LLVM_LINK_LLVM_

[PATCH] D138792: [AArch64] Improve TargetParser API

Hahnfeld added a comment. What about @bkramer's suggestion to move the definitions out of the header? Based on what you write (comparison by address), this should also solve the issue. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D138792/new/ htt

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld added a comment. ping https://reviews.llvm.org/D26244 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

Hahnfeld updated this revision to Diff 82102. Hahnfeld added a comment. Version that works for me https://reviews.llvm.org/D22452 Files: test/std/experimental/filesystem/fs.op.funcs/fs.op.last_write_time/last_write_time.pass.cpp Index: test/std/experimental/filesystem/fs.op.funcs/fs.op.last

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld added a reviewer: chandlerc. Hahnfeld added a comment. Ping. Looks like Chandler originally added this logic... https://reviews.llvm.org/D26244 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/li

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

Hahnfeld updated this revision to Diff 82803. Hahnfeld added a comment. Resolve errors because of recently enabled warnings https://reviews.llvm.org/D22452 Files: test/std/experimental/filesystem/fs.op.funcs/fs.op.last_write_time/last_write_time.pass.cpp Index: test/std/experimental/filesys

[PATCH] D28289: Fix tests with CLANG_DEFAULT_LINKER

Hahnfeld created this revision. Hahnfeld added reviewers: phosek, mgorny. Hahnfeld added a subscriber: cfe-commits. Herald added subscribers: srhines, danalbert. I originally requested this to be tested in https://reviews.llvm.org/D25263 but in the end forgot to make sure that it was done. http

[PATCH] D28298: [OpenMP] Add fields for flags in the offload entry descriptor.

Hahnfeld added a subscriber: cfe-commits. Hahnfeld added a comment. Otherwise SGTM Comment at: lib/CodeGen/CGOpenMPRuntime.h:250-252 + // \brief Flags associated the device global. + int32_t Flags; + Is that intentionally not in the `protected` sectio

[PATCH] D28298: [OpenMP] Add fields for flags in the offload entry descriptor.

Hahnfeld added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.h:250-252 + // \brief Flags associated the device global. + int32_t Flags; + sfantao wrote: > Hahnfeld wrote: > > Is that intentionally not in the `protected` section below? > Yes,

[PATCH] D28298: [OpenMP] Add fields for flags in the offload entry descriptor.

Hahnfeld accepted this revision. Hahnfeld added a comment. This revision is now accepted and ready to land. LGTM https://reviews.llvm.org/D28298 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/c

[PATCH] D28289: Fix tests with CLANG_DEFAULT_LINKER

This revision was automatically updated to reflect the committed changes. Closed by commit rL291389: Fix tests with CLANG_DEFAULT_LINKER (authored by Hahnfeld). Changed prior to commit: https://reviews.llvm.org/D28289?vs=83036&id=83553#toc Repository: rL LLVM https://reviews.llvm.org/D28289

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

Hahnfeld added a comment. Ping. Would this be ok to land before branching for 4.0 this week? https://reviews.llvm.org/D22452 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld added a comment. Ping. Could I get comments before the branching this week? https://reviews.llvm.org/D26244 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld added a comment. This change doesn't break existing tests but maybe I should extend `linux-ld.c` and so on? About the impact: Which libraries are there next to clang that could conflict with the system? When I look at my current installation dir: - `libLLVM*`, `libclang*`, `liblld*` l

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld updated this revision to Diff 83784. Hahnfeld added a comment. Add test case to check that - `bin/../lib` is the first library path by default - user-specified `-L` preceeds https://reviews.llvm.org/D26244 Files: lib/Driver/ToolChains.cpp test/Driver/clang-libraries.c Index: tes

[PATCH] D26244: [Driver] Prefer libraries installed next to Clang over those from GCC

Hahnfeld added a comment. In https://reviews.llvm.org/D26244#641099, @chandlerc wrote: > Those are just the libriaries LLVM is installing. > > If it is installed into a prefix, say, /opt/mumble/{bin,lib,...} along with > other libraries, then suddenly that directory is searched in a very differe

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

Hahnfeld added a comment. @EricWF ? https://reviews.llvm.org/D22452 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime

Hahnfeld added a comment. Another ping https://reviews.llvm.org/D30087 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D30739: [OpenMP] "declare simd" for AArch64 Advanced SIMD.

Hahnfeld added a comment. In principal looks good to me although I'm not really familiar with this part. Does that work for you if you have the `declare simd` in a header file and the implementation in another file? On x86_64 I currently get: remark: loop not vectorized: call instruction cann

[PATCH] D30739: [OpenMP] "declare simd" for AArch64 Advanced SIMD.

Hahnfeld accepted this revision. Hahnfeld added a comment. This revision is now accepted and ready to land. In https://reviews.llvm.org/D30739#706304, @fpetrogalli wrote: > AFAIK, none of the machinery required in LLVM to expose the mangled names in > the vectorizer is present in trunk. There is

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime

Hahnfeld updated this revision to Diff 93213. Hahnfeld added a comment. Herald added a subscriber: rengolin. Rebase and ping. https://reviews.llvm.org/D30087 Files: lib/Driver/ToolChains/CommonArgs.cpp lib/Driver/ToolChains/CommonArgs.h lib/Driver/ToolChains/Darwin.cpp lib/Driver/ToolCh

[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 https://reviews.llvm.or

[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 https://reviews.llv

[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...

[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 LLVM

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

Hahnfeld accepted this revision. Hahnfeld added a comment. This revision is now accepted and ready to land. LGTM after fixing the check-prefix Comment at: test/Driver/openmp-offload.c:595 +// RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -save-temps -

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

Hahnfeld added a comment. Just noticed that this isn't specific to OpenMP. Could you move the test to `Driver/cuda-external-tools.cu` where the rest of `ptxas` is tested? Repository: rL LLVM https://reviews.llvm.org/D29644 ___ cfe-commits mailin

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

Hahnfeld requested changes to this revision. Hahnfeld added a comment. This revision now requires changes to proceed. Out of interest: This is a limition of the implementation, right? Because the standard only says: `A throw executed inside a target region must cause execution to resume within t

[PATCH] D29339: [OpenMP] Add support for auxiliary triple specification

Hahnfeld accepted this revision. Hahnfeld added a comment. This revision is now accepted and ready to land. LGTM Comment at: lib/Driver/ToolChains/Clang.cpp:1967-1968 +C.getSingleOffloadToolChain() +->getTriple() +.normalize(); +CmdArgs.push_

[PATCH] D29654: [OpenMP] Integrate OpenMP target region cubin into host binary

Hahnfeld added a reviewer: rnk. Hahnfeld added a subscriber: rnk. Hahnfeld added a comment. Please format all comments as full sentences. Comment at: lib/Driver/ToolChains/Cuda.cpp:338 +Args.getAllArgValues(options::OPT_march_EQ); +assert(GPUArchs.size() == 1 && "Ex

[PATCH] D29658: [OpenMP] Customize CUDA-based tool chain selection

Hahnfeld requested changes to this revision. Hahnfeld added inline comments. This revision now requires changes to proceed. Comment at: lib/Frontend/CompilerInvocation.cpp:2181-2190 + if (!(TT.getArch() == llvm::Triple::ppc || +TT.getArch() == llvm::Triple::ppc64

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime

Hahnfeld added a comment. Ping! https://reviews.llvm.org/D30087 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime. NFCI.

Hahnfeld added inline comments. Comment at: lib/Driver/ToolChains/CommonArgs.cpp:430 +bool tools::addOpenMPRuntime(ArgStringList &CmdArgs, const ToolChain &TC, + const ArgList &Args, const JobAction &JA, + bool GompNeedsRT)

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime. NFCI.

Hahnfeld updated this revision to Diff 94372. Hahnfeld marked 2 inline comments as done. Hahnfeld retitled this revision from "[Driver] Unify linking of OpenMP runtime" to "[Driver] Unify linking of OpenMP runtime. NFCI.". Hahnfeld edited the summary of this revision. https://reviews.llvm.org/D30

[PATCH] D29658: [OpenMP] Customize CUDA-based tool chain selection

Hahnfeld added a comment. Is there a way to actually test the changed code? The current test changes seem unrelated... Comment at: include/clang/Basic/DiagnosticDriverKinds.td:165-166 "The target '%0' is not a supported OpenMP host target.">; +def err_drv_omp_device_target_

[PATCH] D30087: [Driver] Unify linking of OpenMP runtime. NFCI.

Hahnfeld added a comment. Ping https://reviews.llvm.org/D30087 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

This revision was automatically updated to reflect the committed changes. Closed by commit rL292013: Fix last_write_time tests for filesystems that don't support negative and very… (authored by Hahnfeld). Changed prior to commit: https://reviews.llvm.org/D22452?vs=82803&id=84445#toc Repository

[PATCH] D22452: [libcxx] Fix last_write_time tests for filesystems that don't support negative and very large times.

Hahnfeld added a comment. In https://reviews.llvm.org/D22452#646165, @EricWF wrote: > Let me know once you've committed it and I'll merge it into the 4.0 branch. I've replied to the commit mail and asked Hans. Repository: rL LLVM https://reviews.llvm.org/D22452 _

[PATCH] D28529: [test] Port clang tests to canonicalized booleans

Hahnfeld accepted this revision. Hahnfeld added a comment. This revision is now accepted and ready to land. SGTM https://reviews.llvm.org/D28529 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/c

[PATCH] D26244: [Driver] Add CLANG_PREFER_GCC_LIBRARIES which can be disabled

Hahnfeld updated this revision to Diff 86396. Hahnfeld retitled this revision from "[Driver] Prefer libraries installed next to Clang over those from GCC" to "[Driver] Add CLANG_PREFER_GCC_LIBRARIES which can be disabled". Hahnfeld edited the summary of this revision. Hahnfeld added a reviewer: h

[PATCH] D26244: [Driver] Add CLANG_PREFER_GCC_LIBRARIES which can be disabled

Hahnfeld added a comment. gentle ping https://reviews.llvm.org/D26244 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D34158: to support gcc 4.8 (and newer) compatibility on Linux, preinclude

Hahnfeld added a subscriber: cfe-commits. Hahnfeld edited reviewers, added: rsmith, rengolin; removed: cfe-commits. Hahnfeld added a comment. Some comments inline. In general you should consider posting an RFC on cfe-dev because this change will basically affect all compilations on GNU/Linux if t

[PATCH] D34888: [OpenMP] Fix mapping of scalars for combined directives

Hahnfeld created this revision. Combined directives like 'target parallel' have two captured statements. Sema has to check the right one from the right direction. Previously, Sema::IsOpenMPCapturedByRef would return false for mapped scalars on combined directives. This results in a wrong signat

[PATCH] D34888: [OpenMP] Fix mapping of scalars for combined directives

This revision was automatically updated to reflect the committed changes. Closed by commit rL306956: [OpenMP] Fix mapping of scalars for combined directives (authored by Hahnfeld). Changed prior to commit: https://reviews.llvm.org/D34888?vs=104863&id=104986#toc Repository: rL LLVM https://r

[PATCH] D34158: to support gcc 4.8 (and newer) compatibility on Linux, preinclude

Hahnfeld added a comment. In https://reviews.llvm.org/D34158#797170, @mibintc wrote: > The other test that fails is my own new test! It fails because I don't know > how to set it up so the test thinks it has a gcc toolchain with version > > 4.8. I tried using gcc-toolchain set to various other

[PATCH] D29905: [OpenMP] Pass argument to device kernel by reference when map is used.

Hahnfeld added a comment. In https://reviews.llvm.org/D34888#799576, @gtbercea wrote: > Does this also include the fixes in the following revision? > > https://reviews.llvm.org/D29905 Sorry, I wasn't aware of this revision and thought that it had long been committed. I just verified that the b

[PATCH] D29658: [OpenMP] Customize CUDA-based tool chain selection

Hahnfeld added inline comments. Comment at: lib/Driver/Driver.cpp:564 + auto &CudaTC = + ToolChains[TT.str() + "/" + HostTC->getTriple().normalize()]; + if (!CudaTC) gtbercea wrote: > Hahnfeld wrote: > > The code above u

[PATCH] D42168: [OpenMP] Generate unique name for offloading entries

Hahnfeld created this revision. Hahnfeld added reviewers: gtbercea, ABataev, carlo.bertolli. Herald added subscribers: cfe-commits, guansong. If there are multiple files with target regions, each file and region must generate a unique name or the linker will complain. Repository: rC Clang htt

[PATCH] D42168: [OpenMP] Generate unique name for offloading entries

Hahnfeld added a comment. In https://reviews.llvm.org/D42168#978593, @ABataev wrote: > After some investigation, I think this whole code is incorrect. We should not > rely on the fact that there is no padding on some architectures and should > mix the type generated by the Clang and initial val

[PATCH] D42168: [OpenMP] Generate unique name for offloading entries

Hahnfeld marked an inline comment as done. Hahnfeld added a comment. In https://reviews.llvm.org/D42168#978634, @ABataev wrote: > 1. Please, mark the record as packed. Where and how do I need to do this? I found that `PackedAttr` is checked on `FieldDecl`s but I'm not sure if we go through tha

[PATCH] D42168: [OpenMP] Correct generation of offloading entries

Hahnfeld updated this revision to Diff 130399. Hahnfeld retitled this revision from "[OpenMP] Generate unique name for offloading entries" to "[OpenMP] Correct generation of offloading entries". Hahnfeld edited the summary of this revision. Hahnfeld added a comment. Use `llvm::Twine`, add `Packed

[PATCH] D42168: [OpenMP] Correct generation of offloading entries

This revision was automatically updated to reflect the committed changes. Closed by commit rC322858: [OpenMP] Correct generation of offloading entries (authored by Hahnfeld, committed by ). Repository: rC Clang https://reviews.llvm.org/D42168 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/Op

[PATCH] D42349: [DOCS] Mention OpenMP Tools Interface in runtime library

Hahnfeld created this revision. Hahnfeld added reviewers: ABataev, hans. Herald added subscribers: cfe-commits, guansong. Also list supported configurations (architectures + operating systems). Repository: rC Clang https://reviews.llvm.org/D42349 Files: docs/OpenMPSupport.rst Index: docs

[PATCH] D42349: [DOCS] Mention OpenMP Tools Interface in runtime library

Hahnfeld added a comment. @hans I'd also like to merge this to `release_60` because the support landed before the branching. Is this ok? Repository: rC Clang https://reviews.llvm.org/D42349 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

Hahnfeld created this revision. Hahnfeld added reviewers: rsmith, ABataev, hans. Herald added subscribers: cfe-commits, guansong. The OpenMP runtime has no dedicated Release Notes, so add it to Clang's section about OpenMP. Repository: rC Clang https://reviews.llvm.org/D42352 Files: docs/

[PATCH] D42349: [DOCS] Mention OpenMP Tools Interface in runtime library

This revision was automatically updated to reflect the committed changes. Closed by commit rC323123: [DOCS] Mention OpenMP Tools Interface in runtime library (authored by Hahnfeld, committed by ). Repository: rC Clang https://reviews.llvm.org/D42349 Files: docs/OpenMPSupport.rst Index: do

[PATCH] D42349: [DOCS] Mention OpenMP Tools Interface in runtime library

Hahnfeld added a comment. In https://reviews.llvm.org/D42349#983611, @hans wrote: > In https://reviews.llvm.org/D42349#983180, @Hahnfeld wrote: > > > @hans I'd also like to merge this to `release_60` because the support > > landed before the branching. Is this ok? > > > Sounds good to me. Go ahe

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

Hahnfeld updated this revision to Diff 130900. Hahnfeld marked 2 inline comments as done. Hahnfeld added a comment. Address review comments. https://reviews.llvm.org/D42352 Files: docs/ReleaseNotes.rst Index: docs/ReleaseNotes.rst

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

Hahnfeld added inline comments. Comment at: docs/ReleaseNotes.rst:230 + applications, please rebuild the runtime library with `-DLIBOMP_OMPT_SUPPORT=OFF` + and file a bug at `LLVM's Bugzilla

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

Hahnfeld added inline comments. Comment at: docs/ReleaseNotes.rst:226 +- The LLVM OpenMP runtime `libomp` now supports the OpenMP Tools Interface (OMPT) + on x86, x86_64, AArch64, and PPC64 on Linux, Windows, and mac OS. If you observe + a measurable performance impact on one

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

Hahnfeld updated this revision to Diff 131004. Hahnfeld marked 3 inline comments as done. Hahnfeld added a comment. Fix `macOS`. https://reviews.llvm.org/D42352 Files: docs/ReleaseNotes.rst Index: docs/ReleaseNotes.rst === ---

[PATCH] D42352: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library

This revision was automatically updated to reflect the committed changes. Closed by commit rL323179: [ReleaseNotes] Mention OpenMP Tools Interface in runtime library (authored by Hahnfeld, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llvm.org

[PATCH] D42589: [git-clang-format] Process CUDA files

This revision was automatically updated to reflect the committed changes. Closed by commit rL323615: [git-clang-format] Process CUDA files (authored by Hahnfeld, committed by ). Herald added a subscriber: llvm-commits. Changed prior to commit: https://reviews.llvm.org/D42589?vs=131625&id=131701

[PATCH] D42642: [CUDA] Detect installation in PATH

Hahnfeld created this revision. Hahnfeld added reviewers: jlebar, tra. Herald added a subscriber: cfe-commits. If the CUDA toolkit is not installed to its default locations in /usr/local/cuda, the user is forced to specify --cuda-path. This is tedious and the driver can be smarter if well-known to

[PATCH] D42642: [CUDA] Detect installation in PATH

Hahnfeld added a comment. In https://reviews.llvm.org/D42642#990976, @tra wrote: > Some linux distributions integrate CUDA into the standard directory > structure. I.e. binaries go into /usr/bin, headers into /usr/include, bitcode > goes somewhere else, etc. ptxas will be found, but we would st

[PATCH] D42642: [CUDA] Detect installation in PATH

Hahnfeld added a comment. In https://reviews.llvm.org/D42642#991154, @tra wrote: > In https://reviews.llvm.org/D42642#991127, @Hahnfeld wrote: > > > In https://reviews.llvm.org/D42642#990976, @tra wrote: > > > > > Some linux distributions integrate CUDA into the standard directory > > > structur

<    1   2   3   4   5   6   >