Page MenuHomePhabricator

Today

Fznamznon added inline comments to D60455: [SYCL] Implement SYCL device code outlining.
Wed, Jun 19, 1:46 PM · Restricted Project
ychen updated the diff for D63280: [llvm-objdump] Use <first-symbol>-<offset> as the section start symbol.
  • update
Wed, Jun 19, 1:46 PM · Restricted Project
reames committed rGce53e2226cc4: [LFTR] Stylistic cleanup as suggested in last review comment of D62939 [NFC] (authored by reames).
[LFTR] Stylistic cleanup as suggested in last review comment of D62939 [NFC]
Wed, Jun 19, 1:43 PM
arsenm committed rG4d000d248896: AMDGPU: Fix folding immediate into readfirstlane through reg_sequence (authored by arsenm).
AMDGPU: Fix folding immediate into readfirstlane through reg_sequence
Wed, Jun 19, 1:42 PM
reames committed rL363877: [LFTR] Stylistic cleanup as suggested in last review comment of D62939 [NFC].
[LFTR] Stylistic cleanup as suggested in last review comment of D62939 [NFC]
Wed, Jun 19, 1:42 PM
reames committed rGf8104f01e6bf: [LFTR] Rename variable to minimize confusion [NFC] (authored by reames).
[LFTR] Rename variable to minimize confusion [NFC]
Wed, Jun 19, 1:42 PM
arsenm committed rL363876: AMDGPU: Fix folding immediate into readfirstlane through reg_sequence.
AMDGPU: Fix folding immediate into readfirstlane through reg_sequence
Wed, Jun 19, 1:42 PM
Fznamznon updated the diff for D60455: [SYCL] Implement SYCL device code outlining.

Appled part of comments from @aaron.ballman:

  • Fixed grammar and code style in all places except sycl_kernel docs
  • Added a lit test which checks that sycl_device attribute implicitly added to proper declarations
Wed, Jun 19, 1:42 PM · Restricted Project
arsenm closed D63566: AMDGPU: Fix folding immediate into readfirstlane through reg_sequence .

r363876

Wed, Jun 19, 1:42 PM
reames committed rL363875: [LFTR] Rename variable to minimize confusion [NFC].
[LFTR] Rename variable to minimize confusion [NFC]
Wed, Jun 19, 1:38 PM
pcc committed rG2742eeb78e3e: hwasan: Shrink outlined checks by 1 instruction. (authored by pcc).
hwasan: Shrink outlined checks by 1 instruction.
Wed, Jun 19, 1:37 PM
pcc committed rL363874: hwasan: Shrink outlined checks by 1 instruction..
hwasan: Shrink outlined checks by 1 instruction.
Wed, Jun 19, 1:37 PM
pcc closed D63568: hwasan: Shrink outlined checks by 1 instruction..
Wed, Jun 19, 1:37 PM · Restricted Project
rampitec accepted D63566: AMDGPU: Fix folding immediate into readfirstlane through reg_sequence .

LGTM

Wed, Jun 19, 1:36 PM
eugenis accepted D63568: hwasan: Shrink outlined checks by 1 instruction..

LGTM

Wed, Jun 19, 1:28 PM · Restricted Project
rnk created D63570: [COFF] Implement /safeseh:no and check @feat.00 flags by default.
Wed, Jun 19, 1:20 PM · Restricted Project
aaron.ballman committed rG07e6da933049: Print whether a generic selection expression is result dependent when dumping… (authored by aaron.ballman).
Print whether a generic selection expression is result dependent when dumping…
Wed, Jun 19, 1:15 PM
ldionne committed rG3491119f9ba4: [pstl] Remove warnings in tests and headers (authored by ldionne).
[pstl] Remove warnings in tests and headers
Wed, Jun 19, 1:15 PM
aaronballman committed rL363873: Print whether a generic selection expression is result dependent when dumping….
Print whether a generic selection expression is result dependent when dumping…
Wed, Jun 19, 1:15 PM
ldionne committed rL363872: [pstl] Remove warnings in tests and headers.
[pstl] Remove warnings in tests and headers
Wed, Jun 19, 1:14 PM
arsenm committed rGf46f41411b1f: Reapply "r363684: AMDGPU: Add GWS instruction builtins" (authored by arsenm).
Reapply "r363684: AMDGPU: Add GWS instruction builtins"
Wed, Jun 19, 12:53 PM
arsenm committed rG4d55d024be80: Reapply "AMDGPU: Add ds_gws_init / ds_gws_barrier intrinsics" (authored by arsenm).
Reapply "AMDGPU: Add ds_gws_init / ds_gws_barrier intrinsics"
Wed, Jun 19, 12:53 PM
arsenm committed rL363871: Reapply "r363684: AMDGPU: Add GWS instruction builtins".
Reapply "r363684: AMDGPU: Add GWS instruction builtins"
Wed, Jun 19, 12:52 PM
arsenm committed rL363870: Reapply "AMDGPU: Add ds_gws_init / ds_gws_barrier intrinsics".
Reapply "AMDGPU: Add ds_gws_init / ds_gws_barrier intrinsics"
Wed, Jun 19, 12:52 PM
SjoerdMeijer created D63569: Make TargetParserTest.ARMExtensionFeatures not run out of memory on 32-bit (PR42316).
Wed, Jun 19, 12:51 PM
spatel added a comment to D63294: [Analysis] enhance FP library function prototype checking to match types with name suffix .

Not sure what you're talking about with clang; it does the right thing, as far as I can tell.

When compiling for AVR, clang will turn the libcall that uses doubles in C source into:

%call = call addrspace(1) float @sqrt(float 0.000000e+00) #2

Should clang translate that call into "sqrtf" to be more accurate? Similarly, if the source was written with "sqrtl" and "long double", we'd get this IR out of clang:

%call = call addrspace(1) float @sqrtl(float 0.000000e+00) #2
Wed, Jun 19, 12:50 PM · Restricted Project
pcc created D63568: hwasan: Shrink outlined checks by 1 instruction..
Wed, Jun 19, 12:44 PM · Restricted Project
aaron.ballman committed rG709a769cdd4e: Print out the union field being initialized by an InitListExpr when dumping the… (authored by aaron.ballman).
Print out the union field being initialized by an InitListExpr when dumping the…
Wed, Jun 19, 12:39 PM
aaronballman committed rL363869: Print out the union field being initialized by an InitListExpr when dumping the….
Print out the union field being initialized by an InitListExpr when dumping the…
Wed, Jun 19, 12:39 PM
spatel added a comment to D63294: [Analysis] enhance FP library function prototype checking to match types with name suffix .

Not sure what you're talking about with clang; it does the right thing, as far as I can tell.

Wed, Jun 19, 12:36 PM · Restricted Project
dyaroshev added a comment to D63063: Bug 42208: speeding up std::merge.

Reported this as an optimiser bug - maybe they'll have a suggestion: https://bugs.llvm.org/show_bug.cgi?id=42334

Wed, Jun 19, 12:33 PM
ychen committed rG40a156b79179: [llvm-readobj] Match GNU output for DT_RPATH and DT_RUNPATH when dumping… (authored by ychen).
[llvm-readobj] Match GNU output for DT_RPATH and DT_RUNPATH when dumping…
Wed, Jun 19, 12:32 PM
tra accepted D63277: [CUDA][HIP] Don't set "comdat" attribute for CUDA device stub functions..
Wed, Jun 19, 12:32 PM · Restricted Project, Restricted Project
JohnLLVM added a comment to D57450: [RISCV] Set MaxAtomicInlineWidth and MaxAtomicPromoteWidth for RV32/RV64 targets with atomics.

@asb: Gentle ping on this.

Wed, Jun 19, 12:32 PM · Restricted Project
gtbercea closed D63009: [OpenMP] Add target task alloc function with device ID.
Wed, Jun 19, 12:32 PM · Restricted Project, Restricted Project, Restricted Project
rjmccall added a comment to D63277: [CUDA][HIP] Don't set "comdat" attribute for CUDA device stub functions..

This optimization is disabled for functions not in COMDAT sections? Is that documented somewhere?

Wed, Jun 19, 12:32 PM · Restricted Project, Restricted Project
yuanfang committed rL363868: [llvm-readobj] Match GNU output for DT_RPATH and DT_RUNPATH when dumping….
[llvm-readobj] Match GNU output for DT_RPATH and DT_RUNPATH when dumping…
Wed, Jun 19, 12:27 PM
Diffusion closed D63347: [llvm-readobj] Match GNU output for DT_RPATH and DT_RUNPATH when dumping dynamic symbol table..
Wed, Jun 19, 12:27 PM · Restricted Project
rnk added a comment to D63453: [Support] Move llvm::MemoryBuffer to sys::fs::file_t.

FWIW, we do use lld-link to cross-compile from Darwin to Windows, and it would be awesome if that kept working :)

Wed, Jun 19, 12:27 PM · Restricted Project
clayborg added a comment to D63540: Fix lookup of symbols at the same address with no size vs. size.

my fix assumes we would have zero sized symbols first in the m_file_addr_to_index map, and the symbols with the same size would be ordered after the ones with zero size. Also, I am assuming when we do:

const FileRangeToIndexMap::Entry *entry = m_file_addr_to_index.FindEntryStartsAt(file_addr);

that it finds the first symbol that starts with "file_addr". If this isn't the case we might need to be able to back up a few symbols to ensure we get consistent results

Wed, Jun 19, 12:27 PM · Restricted Project