Bump LLVM to v19.1.1+1 #56130

Zentrik · 2024-10-12T16:54:14Z

Including #55650 till that's merged.

Zentrik · 2024-10-12T17:47:34Z

The /home/rag/Documents/Code/julia/usr/include/llvm/ADT/StringRef.h:871:20: warning: 'int __builtin_memcmp_eq(const void*, const void*, long unsigned int)' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overread] warning seems to be a gcc bug, https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101361

EDIT: gcc is giving a slightly different warning on CI but likely a bug as well

giordano · 2024-10-12T17:54:21Z

In that case you can guard the offending line with something like

#pragma GCC diagnostic push
#if defined(_COMPILER_GCC_) && __GNUC__ >= 12 // if this is version-dependent
// Explain why this is being ignored...
#pragma GCC diagnostic ignored "-Wstringop-overflow"
#endif
...
#pragma GCC diagnostic pop

giordano · 2024-10-12T18:07:00Z

Also, all tests are passing already on aarch64-darwin 🥳

Zentrik · 2024-10-12T19:51:23Z

Looks like the only two issues are a incorrect warning during the build (mose's suggestion doesn't seem to work though #pragma GCC diagnostic ignored "-Wstringop-overflow" by itself does) and /cache/build/tester-demeter6-13/julialang/julia-master/tmp/test-asan/asan/usr/bin/julia: error while loading shared libraries: libclang_rt.asan-x86_64.so: cannot open shared object file: No such file or directory. (I'll push the analyzegc fix in a bit).

Zentrik · 2024-10-12T20:54:14Z

Guessing the asan problem is clang -print-runtime-dir is returning toolchain/usr/lib/clang/19/lib/x86_64-unknown-linux-gnu instead of toolchain/usr/lib/clang/19/lib/linux so we don't find libclang_rt.asan... to copy to the build folder.

giordano · 2024-10-13T08:50:21Z

though #pragma GCC diagnostic ignored "-Wstringop-overflow" by itself does

Without the push-and-pop, that becomes closer to just adding -Wstringop-overflow to

julia/src/Makefile

Lines 23 to 25 in 80e60c8

    
           ifeq ($(USEGCC),1) # GCC bug #25509 (void)__attribute__((warn_unused_result)) 
        
           FLAGS += -Wno-unused-result 
        
           endif

(please always add a comment to explain why the warning is being skipped) but it'd be nicer to keep the ignore as local as possible, if feasible.

Zentrik · 2024-10-13T11:54:41Z

Looks like remaining issue is some failures in analyzegc that I guess weren't caught by previous versions of clang.

Zentrik · 2024-10-13T15:03:25Z

Second analyzegc failure looks incorrect, it seems to incorrectly think uv_dup could return 0 when dupfd is set to -1 (minimal reproducer https://godbolt.org/z/4Wo99v1nv, llvm/llvm-project#43015 looks to be the same issue).

Zentrik · 2024-10-13T15:03:36Z

@nanosoldier runbenchmarks(ALL, vs=":master")

giordano · 2024-10-13T15:39:10Z

Need to rebase on master now that #56133 has been merged.

nanosoldier · 2024-10-13T23:58:34Z

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

vtjnash · 2024-10-14T20:05:27Z

First analyzegc failure looks incorrect, it seems to incorrectly think uv_dup could return 0 when dupfd is set to -1 (minimal reproducer https://godbolt.org/z/4Wo99v1nv, llvm/llvm-project#43015 looks to be the same issue).

FWIW, the standard way to deal with this is to add an assert and a comment, as that will make both reviewers and bots happy

Zentrik · 2024-10-16T20:51:33Z

The union.array.(perf_sum, Float32, 0) regression is due to SIMD not occurring. This is due to a change in the result from the instcombine pass. See https://godbolt.org/z/qqdzM7oxh,
The change from

%12 = select i1 %exactly_isa.not, float -0.000000e+00, float %immutable_union.sroa.0.0.copyload, !dbg !77
%value_phi6 = fadd float %value_phi, %12, !dbg !77

to

%12 = fadd float %value_phi, %immutable_union.sroa.0.0.copyload, !dbg !77
%value_phi6 = select i1 %exactly_isa.not, float %value_phi, float %12, !dbg !77

prevents the loopvectorize pass from SIMDing the code.

I suspect the other regression for the union.array benchmarks are similar.

Zentrik · 2024-10-17T06:40:36Z

I'll try rebasing on top of #52850 and see if that fixes the regressions.

Zentrik · 2024-10-17T06:40:45Z

@nanosoldier runtests()

nanosoldier · 2024-10-17T22:16:20Z

The package evaluation job you requested has completed - possible new issues were detected.
The full report is available.

oscardssmith · 2024-10-18T14:19:04Z

@nanosoldier runbenchmarks(ALL, vs=":gb/pipeline-fun")

nanosoldier · 2024-10-18T23:16:01Z

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

oscardssmith · 2024-10-19T02:09:20Z

looks like there are a few 8x (vectorization probably) regressions here.

Zentrik · 2024-10-27T16:56:25Z

PkgEval failure for McCormick is ~~likely llvm/llvm-project#113810~~ llvm/llvm-project#101213, fixed by llvm/llvm-project@f70f122.

giordano · 2024-11-03T23:41:15Z

Analyzegc is still failing:

/cache/build/builder-demeter6-5/julialang/julia-master/usr/include/llvm/ADT/APInt.h:172:7: error: Attempt to free released memory [cplusplus.NewDelete]
  172 |       delete[] U.pVal;
      |       ^

`clang -print-runtime-dir` reports a non-existent directory as we build with `LLVM_ENABLE_PER_TARGET_RUNTIME_DIR=OFF`. See llvm/llvm-project#102834. I suspect llvm/llvm-project@b6a1473 caused the change by chaning the code in Driver.cpp to not check whether the printed directory existed.

Locally I get: ``` In file included from /home/rag/Documents/Code/julia/usr/include/llvm/Object/ObjectFile.h:18, from /home/rag/Documents/Code/julia/usr/include/llvm/DebugInfo/DIContext.h:18, from /home/rag/Documents/Code/julia/src/debuginfo.cpp:6: In function 'bool llvm::operator==(llvm::StringRef, llvm::StringRef)', inlined from 'bool llvm::operator!=(llvm::StringRef, llvm::StringRef)' at /home/rag/Documents/Code/julia/usr/include/llvm/ADT/StringRef.h:874:71, inlined from 'objfileentry_t find_object_file(uint64_t, llvm::StringRef)' at /home/rag/Documents/Code/julia/src/debuginfo.cpp:948:43, inlined from 'bool jl_dylib_DI_for_fptr(size_t, llvm::object::SectionRef*, int64_t*, llvm::DIContext**, bool, bool*, uint64_t*, void**, char**, char**)' at /home/rag/Documents/Code/julia/src/debuginfo.cpp:1135:34: /home/rag/Documents/Code/julia/usr/include/llvm/ADT/StringRef.h:871:20: warning: 'int __builtin_memcmp_eq(const void*, const void*, long unsigned int)' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overread] 871 | return ::memcmp(LHS.data(), RHS.data(), LHS.size()) == 0; | ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /home/rag/Documents/Code/julia/src/debuginfo.cpp: In function 'bool jl_dylib_DI_for_fptr(size_t, llvm::object::SectionRef*, int64_t*, llvm::DIContext**, bool, bool*, uint64_t*, void**, char**, char**)': /home/rag/Documents/Code/julia/src/debuginfo.cpp:1133:11: note: source object allocated here 1133 | fname = dlinfo.dli_fname; ``` On CI: ``` In file included from /cache/build/builder-amdci4-2/julialang/julia-master/usr/include/llvm/Object/ObjectFile.h:18, from /cache/build/builder-amdci4-2/julialang/julia-master/usr/include/llvm/DebugInfo/DIContext.h:18, from /cache/build/builder-amdci4-2/julialang/julia-master/src/debuginfo.cpp:6: In function 'bool llvm::operator==(llvm::StringRef, llvm::StringRef)', inlined from 'bool llvm::operator!=(llvm::StringRef, llvm::StringRef)' at /cache/build/builder-amdci4-2/julialang/julia-master/usr/include/llvm/ADT/StringRef.h:874:71, inlined from 'objfileentry_t find_object_file(uint64_t, llvm::StringRef)' at /cache/build/builder-amdci4-2/julialang/julia-master/src/debuginfo.cpp:943:43, inlined from 'bool jl_dylib_DI_for_fptr(size_t, llvm::object::SectionRef*, int64_t*, llvm::DIContext**, bool, bool*, uint64_t*, void**, char**, char**)' at /cache/build/builder-amdci4-2/julialang/julia-master/src/debuginfo.cpp:1126:47: /cache/build/builder-amdci4-2/julialang/julia-master/usr/include/llvm/ADT/StringRef.h:871:20: error: 'int __builtin_memcmp_eq(const void*, const void*, long unsigned int)' specified size 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Werror=stringop-overflow=] 871 | return ::memcmp(LHS.data(), RHS.data(), LHS.size()) == 0; | ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ```

The change to `GCChecker.cpp` is so the static analyzer doesn't see e.g. a call to a function pointer in llvm and then complain that it might be a safepoint.

e7698a13e319a9919af04d3d693a6f6ea7168a44 isn't in llvm 19

Zentrik · 2024-11-24T14:32:05Z

@nanosoldier runtests()

Zentrik added compiler:llvm For issues that relate to LLVM building Build system, or building Julia or its dependencies external dependencies Involves LLVM, OpenBLAS, or other linked libraries JLLs labels Oct 12, 2024

Zentrik force-pushed the llvm-19-actual branch from c1900d0 to adc7db4 Compare October 12, 2024 16:55

Zentrik force-pushed the llvm-19-actual branch 2 times, most recently from 5c3c307 to 24f4d93 Compare October 13, 2024 09:48

giordano added the needs pkgeval Tests for all registered packages should be run with this change label Oct 13, 2024

Zentrik force-pushed the llvm-19-actual branch 2 times, most recently from f6b2376 to e12026f Compare October 13, 2024 20:43

giordano mentioned this pull request Oct 26, 2024

Fix make install for asan build #56347

Merged

Zentrik force-pushed the llvm-19-actual branch from c537718 to dfec0aa Compare November 3, 2024 10:28

Zentrik added 10 commits November 24, 2024 12:33

Bump LLVM to v19.1.1+0

ab20062

Update checksums

65af02f

Only throw error when installing sanitizers

e9590ca

Fix previous commit on gcc 9 and 10

7d0a77d

Get analyzegc pass passing

0581f88

The change to `GCChecker.cpp` is so the static analyzer doesn't see e.g. a call to a function pointer in llvm and then complain that it might be a safepoint.

Make comment clearer

02cab3e

Bump version bound

1fc382a

e7698a13e319a9919af04d3d693a6f6ea7168a44 isn't in llvm 19

Bump to v19.1.1+1

71f9f18

Zentrik force-pushed the llvm-19-actual branch from dd978af to 71f9f18 Compare November 24, 2024 12:33

Zentrik changed the title ~~Bump LLVM to v19.1.1+0~~ Bump LLVM to v19.1.1+1 Nov 24, 2024

Bump LLVM tag to v19.1.1+1

128354f

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bump LLVM to v19.1.1+1 #56130

Bump LLVM to v19.1.1+1 #56130

Zentrik commented Oct 12, 2024 •

edited

Loading

Zentrik commented Oct 12, 2024 •

edited

Loading

giordano commented Oct 12, 2024

giordano commented Oct 12, 2024

Zentrik commented Oct 12, 2024 •

edited

Loading

Zentrik commented Oct 12, 2024

giordano commented Oct 13, 2024

Zentrik commented Oct 13, 2024

Zentrik commented Oct 13, 2024 •

edited

Loading

Zentrik commented Oct 13, 2024

giordano commented Oct 13, 2024

nanosoldier commented Oct 13, 2024

vtjnash commented Oct 14, 2024

Zentrik commented Oct 16, 2024 •

edited by giordano

Loading

Zentrik commented Oct 17, 2024

Zentrik commented Oct 17, 2024

nanosoldier commented Oct 17, 2024

oscardssmith commented Oct 18, 2024

nanosoldier commented Oct 18, 2024

oscardssmith commented Oct 19, 2024 •

edited

Loading

Zentrik commented Oct 27, 2024 •

edited

Loading

giordano commented Nov 3, 2024

Zentrik commented Nov 24, 2024

Bump LLVM to v19.1.1+1 #56130

Are you sure you want to change the base?

Bump LLVM to v19.1.1+1 #56130

Conversation

Zentrik commented Oct 12, 2024 • edited Loading

Zentrik commented Oct 12, 2024 • edited Loading

giordano commented Oct 12, 2024

giordano commented Oct 12, 2024

Zentrik commented Oct 12, 2024 • edited Loading

Zentrik commented Oct 12, 2024

giordano commented Oct 13, 2024

Zentrik commented Oct 13, 2024

Zentrik commented Oct 13, 2024 • edited Loading

Zentrik commented Oct 13, 2024

giordano commented Oct 13, 2024

nanosoldier commented Oct 13, 2024

vtjnash commented Oct 14, 2024

Zentrik commented Oct 16, 2024 • edited by giordano Loading

Zentrik commented Oct 17, 2024

Zentrik commented Oct 17, 2024

nanosoldier commented Oct 17, 2024

oscardssmith commented Oct 18, 2024

nanosoldier commented Oct 18, 2024

oscardssmith commented Oct 19, 2024 • edited Loading

Zentrik commented Oct 27, 2024 • edited Loading

giordano commented Nov 3, 2024

Zentrik commented Nov 24, 2024

Zentrik commented Oct 12, 2024 •

edited

Loading

Zentrik commented Oct 12, 2024 •

edited

Loading

Zentrik commented Oct 12, 2024 •

edited

Loading

Zentrik commented Oct 13, 2024 •

edited

Loading

Zentrik commented Oct 16, 2024 •

edited by giordano

Loading

oscardssmith commented Oct 19, 2024 •

edited

Loading

Zentrik commented Oct 27, 2024 •

edited

Loading