Page MenuHomePhabricator

vtjnash (Jameson Nash)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 21 2013, 6:34 AM (371 w, 5 d)

Recent Activity

Thu, Dec 3

vtjnash committed rGc01fb6449e17: repair cygwin build (authored by vtjnash).
repair cygwin build
Thu, Dec 3, 8:50 AM
vtjnash closed D92336: repair cygwin build.
Thu, Dec 3, 8:50 AM · Restricted Project
vtjnash added a comment to D92336: repair cygwin build.

I think with the move to newer C++ standards, some of the missing functionality in C could be replicated (or use the _WIN32 code paths for __CYGWIN__ in more places). I tried that for a bit, but it was pieces like struct sigaction, realpath, and getsid missing, which are very platform-specific.

Thu, Dec 3, 7:21 AM · Restricted Project

Mon, Nov 30

vtjnash updated the summary of D92336: repair cygwin build.
Mon, Nov 30, 11:16 AM · Restricted Project
vtjnash added a reviewer for D92336: repair cygwin build: compnerd.
Mon, Nov 30, 11:15 AM · Restricted Project
vtjnash requested review of D92336: repair cygwin build.
Mon, Nov 30, 11:14 AM · Restricted Project

Mon, Nov 23

vtjnash added a reviewer for D90705: [GVN] permit forwarding to non-byte-sized loads: kuba.
Mon, Nov 23, 5:32 PM · Restricted Project
vtjnash added reviewers for D90969: Renovate CMake file for the `llvm-cfi-verify` tool: stephenneuendorffer, tstellar.
Mon, Nov 23, 5:30 PM · Restricted Project
vtjnash added reviewers for D90970: Renovate CMake files in the `llvm-exegesis` tool.: tstellar, stephenneuendorffer.
Mon, Nov 23, 5:29 PM · Restricted Project
vtjnash committed rG8f1e0dcbb090: fix some Wundef warnings in public headers (authored by vtjnash).
fix some Wundef warnings in public headers
Mon, Nov 23, 5:19 PM
vtjnash closed D91094: fix some Wundef warnings in public headers.
Mon, Nov 23, 5:19 PM · Restricted Project

Mon, Nov 16

vtjnash added a reverting change for rG105ed27ed80d: Revert "[AsmPrinter] fix -disable-debug-info option": rGbf6ed355c8d1: Reland "[AsmPrinter] fix -disable-debug-info option".
Mon, Nov 16, 10:35 AM
vtjnash committed rGbf6ed355c8d1: Reland "[AsmPrinter] fix -disable-debug-info option" (authored by vtjnash).
Reland "[AsmPrinter] fix -disable-debug-info option"
Mon, Nov 16, 10:35 AM
vtjnash added a comment to D88138: [NPM] Add target specific hook to add passes for New Pass Manager.

I think this, and similar recent commits, are causing the shared library builds to fail some tests if this code gets linked into libLLVM.so: https://bugs.llvm.org/show_bug.cgi?id=48181. I assume it might actually a bug in ld (GNU Binutils for Ubuntu 2.34), as I don't understand the linker behavior there?

Mon, Nov 16, 10:24 AM · Restricted Project, Restricted Project

Sat, Nov 14

vtjnash added a comment to D91083: [AsmPrinter] fix -disable-debug-info option.

Thanks for identifying that. I think I can re-land just without that part of the test. It is not essential, I added it just to try to check all of the dwarf-related output.

Sat, Nov 14, 10:59 AM · Restricted Project

Thu, Nov 12

vtjnash committed rG9606ef03f039: [AsmPrinter] fix -disable-debug-info option (authored by vtjnash).
[AsmPrinter] fix -disable-debug-info option
Thu, Nov 12, 9:59 PM
vtjnash closed D91083: [AsmPrinter] fix -disable-debug-info option.
Thu, Nov 12, 9:59 PM · Restricted Project

Mon, Nov 9

vtjnash added a reviewer for D91094: fix some Wundef warnings in public headers: bob.wilson.
Mon, Nov 9, 11:01 AM · Restricted Project
vtjnash requested review of D91094: fix some Wundef warnings in public headers.
Mon, Nov 9, 11:01 AM · Restricted Project
vtjnash added a reviewer for D90969: Renovate CMake file for the `llvm-cfi-verify` tool: chapuni.
Mon, Nov 9, 10:25 AM · Restricted Project
vtjnash removed a reviewer for D88995: Support vectors in CastInst::isBitOrNoopPointerCastable: vtjnash.
Mon, Nov 9, 10:20 AM · Restricted Project
vtjnash added reviewers for D91083: [AsmPrinter] fix -disable-debug-info option: uabelho, dstenb.
Mon, Nov 9, 8:01 AM · Restricted Project
vtjnash requested review of D91083: [AsmPrinter] fix -disable-debug-info option.
Mon, Nov 9, 8:00 AM · Restricted Project

Fri, Nov 6

vtjnash added a comment to D89613: make the AsmPrinterHandler array public.

Ah, okay, yes that is more clearly needing to be fixed. I think I just need to rearrange how that command line option works.

Fri, Nov 6, 2:32 PM · Restricted Project
vtjnash updated the summary of D90969: Renovate CMake file for the `llvm-cfi-verify` tool.
Fri, Nov 6, 1:39 PM · Restricted Project
vtjnash requested review of D90970: Renovate CMake files in the `llvm-exegesis` tool..
Fri, Nov 6, 1:38 PM · Restricted Project
vtjnash requested review of D90969: Renovate CMake file for the `llvm-cfi-verify` tool.
Fri, Nov 6, 1:12 PM · Restricted Project
vtjnash updated the diff for D34078: Enable support for floating-point division reductions.

add null test as requested, rebase

Fri, Nov 6, 11:58 AM · Restricted Project, Unknown Object (Project)
vtjnash committed rG18805ea951be: Fix unwind info relocation with large code model on AArch64 (authored by vchuravy).
Fix unwind info relocation with large code model on AArch64
Fri, Nov 6, 11:44 AM
vtjnash closed D27629: Fix unwind info relocation with large code model on AArch64.
Fri, Nov 6, 11:43 AM · Restricted Project, Unknown Object (Project)
vtjnash committed rG85f4be09124c: [RTDYLD] support absolute relocations where needed (authored by vchuravy).
[RTDYLD] support absolute relocations where needed
Fri, Nov 6, 11:24 AM
vtjnash closed D90722: [RTDYLD] support absolute relocations where needed.
Fri, Nov 6, 11:24 AM · Restricted Project

Thu, Nov 5

vtjnash added a comment to D90722: [RTDYLD] support absolute relocations where needed.

If it's too fiddly I don't mind this going in without a test case right now -- I think extra cycles would be just as well spent bringing up / testing JITLink for ELF. We should make sure we don't replicate this issue by using empty strings as sentinels there.

Thu, Nov 5, 11:31 AM · Restricted Project

Nov 5 2020

vtjnash added a comment to D89613: make the AsmPrinterHandler array public.

The .cfi_startproc directive is part of the unwind information (eh_frame), and is not debug info, so that seems like a possible bugfix even? I'd expect that observation to be consistent with the goal of this PR, though I didn't see specifically where that pass was being affected in a cursory glance back through the PR.

Nov 5 2020, 6:28 AM · Restricted Project

Nov 3 2020

vtjnash updated the summary of D90722: [RTDYLD] support absolute relocations where needed.
Nov 3 2020, 2:49 PM · Restricted Project
vtjnash added reviewers for D90722: [RTDYLD] support absolute relocations where needed: lhames, vchuravy.
Nov 3 2020, 2:48 PM · Restricted Project
vtjnash requested review of D90722: [RTDYLD] support absolute relocations where needed.
Nov 3 2020, 2:47 PM · Restricted Project
vtjnash updated the summary of D90705: [GVN] permit forwarding to non-byte-sized loads.
Nov 3 2020, 1:11 PM · Restricted Project
vtjnash updated the summary of D90703: [GVN] Make handling of zeros more consistent.
Nov 3 2020, 12:39 PM · Restricted Project
vtjnash updated the diff for D90705: [GVN] permit forwarding to non-byte-sized loads.

use helper method

Nov 3 2020, 12:01 PM · Restricted Project
vtjnash updated the diff for D90705: [GVN] permit forwarding to non-byte-sized loads.

fix test

Nov 3 2020, 11:55 AM · Restricted Project
vtjnash requested review of D90705: [GVN] permit forwarding to non-byte-sized loads.
Nov 3 2020, 11:50 AM · Restricted Project
vtjnash requested review of D90703: [GVN] Make handling of zeros more consistent.
Nov 3 2020, 11:49 AM · Restricted Project
vtjnash committed rG59a6ab28c42b: [GVN] small improvements to comments (authored by vtjnash).
[GVN] small improvements to comments
Nov 3 2020, 10:22 AM
vtjnash abandoned D59661: [GVN] Try to be more careful about handling constant zero and SVE.
Nov 3 2020, 10:06 AM · Restricted Project
vtjnash abandoned D69910: [GVN] Remove duplicated canCoerce tests.
Nov 3 2020, 10:06 AM · Restricted Project
vtjnash added a comment to D89241: [RTDYLD] be more defensive about relocations with empty symbol names.

I think these need to be handled in these places, given how clang emits relocations in .dwarf sections. For some contextual history, this was added as a sentinel for absolute symbols in https://github.com/llvm/llvm-project/commit/ad6d349fbcb2f9ced3c57b5b61231315175bcb47 (and later adjusted in https://github.com/llvm/llvm-project/commit/8f1f87c73487dcb83cf5d6d44b062f1e9796fae6). If I compare readelf --relocations of clang vs. gcc, I see GCC emit a name for these symbols for x86 inside the .dwarf sections, but that clang uses the name NULL, which gets translated there into the empty string.

Nov 3 2020, 7:30 AM · Restricted Project
vtjnash committed rGa0ad066ce490: make the AsmPrinterHandler array public (authored by vtjnash).
make the AsmPrinterHandler array public
Nov 3 2020, 7:03 AM
vtjnash closed D89613: make the AsmPrinterHandler array public.
Nov 3 2020, 7:03 AM · Restricted Project

Oct 16 2020

vtjnash added a comment to D89613: make the AsmPrinterHandler array public.

In my first attempt, I missed that it's intentional we're claiming MMI->hasDebugInfo() if there's information available for DWARF output if using fast-isel, even if we don't need that DebugInfo because the output for it is disabled (CodeView information is still somewhat separate, as my attempt to bring them back in line is what failed the CI testing). Fortunately, that's what tests are for when this change was added in 8763c0c5b7a566d2fe9476259aa346640b0a9deb (https://reviews.llvm.org/D53885). I think this should pass all tests now, as I've now fully removed that part of the change (it passes locally anyways).

Oct 16 2020, 10:35 PM · Restricted Project
vtjnash added a reviewer for D89613: make the AsmPrinterHandler array public: rnk.
Oct 16 2020, 10:05 PM · Restricted Project
vtjnash updated the summary of D89613: make the AsmPrinterHandler array public.
Oct 16 2020, 10:05 PM · Restricted Project
vtjnash requested review of D89613: make the AsmPrinterHandler array public.
Oct 16 2020, 10:04 PM · Restricted Project
vtjnash added a comment to D74158: make the AsmPrinterHandler array public.

Yep, I saw that too. I pushed a revert commit while I wait for the tests locally to rerun and confirm my fix for that.

Oct 16 2020, 2:36 PM · Restricted Project
vtjnash committed rG4242df14708c: Revert "make the AsmPrinterHandler array public" (authored by vtjnash).
Revert "make the AsmPrinterHandler array public"
Oct 16 2020, 2:23 PM
vtjnash committed rGac2def2d8d8c: make the AsmPrinterHandler array public (authored by vtjnash).
make the AsmPrinterHandler array public
Oct 16 2020, 1:28 PM
vtjnash closed D74158: make the AsmPrinterHandler array public.
Oct 16 2020, 1:28 PM · Restricted Project
vtjnash added a comment to D74158: make the AsmPrinterHandler array public.

@rnk I've generalized the API to your suggestion and added a test. Let me know what you think, thanks!

Oct 16 2020, 9:46 AM · Restricted Project
vtjnash updated the diff for D74158: make the AsmPrinterHandler array public.

update

Oct 16 2020, 9:33 AM · Restricted Project

Oct 15 2020

vtjnash updated the diff for D74158: make the AsmPrinterHandler array public.
Oct 15 2020, 11:48 PM · Restricted Project
vtjnash committed rG122d92dfc31f: fix symbol printing on windows (authored by vtjnash).
fix symbol printing on windows
Oct 15 2020, 2:15 PM
vtjnash closed D87099: fix symbol printing on windows.
Oct 15 2020, 2:15 PM · Restricted Project
vtjnash added a comment to D69910: [GVN] Remove duplicated canCoerce tests.

Removed NFCI, since as I look back, I think this also enabled optimization of constant structures, and I was misusing the NFCI term

Oct 15 2020, 11:43 AM · Restricted Project
vtjnash retitled D69910: [GVN] Remove duplicated canCoerce tests from [GVN] Remove duplicated canCoerce tests [NFCI] to [GVN] Remove duplicated canCoerce tests.
Oct 15 2020, 11:42 AM · Restricted Project
vtjnash retitled D69910: [GVN] Remove duplicated canCoerce tests from [GVN] Remove duplicated canCoerce tests, NFCI to [GVN] Remove duplicated canCoerce tests [NFCI].
Oct 15 2020, 11:42 AM · Restricted Project
vtjnash added a comment to D59661: [GVN] Try to be more careful about handling constant zero and SVE.

I've removed NFCI, since I ended up changing some of the SVE handling when attempting to rebase it

Oct 15 2020, 11:21 AM · Restricted Project
vtjnash retitled D59661: [GVN] Try to be more careful about handling constant zero and SVE from [GVN] Try to be more careful about handling constant zero, NFCI to [GVN] Try to be more careful about handling constant zero and SVE.
Oct 15 2020, 11:20 AM · Restricted Project

Oct 7 2020

vtjnash accepted D88995: Support vectors in CastInst::isBitOrNoopPointerCastable.

Makes sense to me

Oct 7 2020, 11:53 AM · Restricted Project
vtjnash updated the diff for D69910: [GVN] Remove duplicated canCoerce tests.

rebase

Oct 7 2020, 11:31 AM · Restricted Project
vtjnash updated the diff for D59661: [GVN] Try to be more careful about handling constant zero and SVE.

rebase

Oct 7 2020, 11:29 AM · Restricted Project

Sep 22 2020

vtjnash updated the diff for D74158: make the AsmPrinterHandler array public.

clang-format patch

Sep 22 2020, 9:43 AM · Restricted Project
vtjnash updated subscribers of D74158: make the AsmPrinterHandler array public.

@rnk would you be an appropriate reviewer for this idea and willing to look it over? I see @timurrrr originally created this generalization that I'm proposing exposing (r196288 / 1cd1444449cb68e2091e8f36785403288ebd3326), but that he's not working with the llvm project anymore. Thanks!

Sep 22 2020, 9:40 AM · Restricted Project
vtjnash updated the diff for D74158: make the AsmPrinterHandler array public.

rebase

Sep 22 2020, 9:01 AM · Restricted Project
vtjnash updated the diff for D87099: fix symbol printing on windows.

fixing commit diff

Sep 22 2020, 8:41 AM · Restricted Project
vtjnash updated the diff for D87099: fix symbol printing on windows.

rename helper functions to canBeUnquotedInDirective

Sep 22 2020, 8:40 AM · Restricted Project

Sep 3 2020

vtjnash added a reviewer for D87099: fix symbol printing on windows: compnerd.
Sep 3 2020, 10:32 AM · Restricted Project
vtjnash requested review of D87099: fix symbol printing on windows.
Sep 3 2020, 10:28 AM · Restricted Project

Sep 1 2020

vtjnash added a comment to D84677: [libunwind] Remove old keymgr related logic.

I'm glad to see this is gone. I was just about to file a bug report that this code used to incur a deadlock risk when unw_step was used from a signal handler since macOS 10.9.0 on x86_64, since Apple had removed the other side of this code circa keymgr-28, so it now had been attempting to call malloc. (I'd observed this deadlock occur in testing, and implemented a work around for it downstream, in https://github.com/JuliaLang/julia/pull/37101/commits/fad04d39d592d8e0fcbfba439e8157f582bbc850#diff-86a37833b03dbe7f0874e30469cb1d46R93-R102)

Sep 1 2020, 9:42 AM · Restricted Project, Restricted Project

Jul 20 2020

vtjnash added a comment to D83870: [ConstantFolding] check applicability of AllOnes constant creation first.

Does that mean I should commit it there also, or how does that process work?

Jul 20 2020, 9:16 PM · Restricted Project

Jul 19 2020

vtjnash committed rG8b354cc8db41: [ConstantFolding] check applicability of AllOnes constant creation first (authored by vtjnash).
[ConstantFolding] check applicability of AllOnes constant creation first
Jul 19 2020, 10:15 AM
vtjnash closed D83870: [ConstantFolding] check applicability of AllOnes constant creation first.
Jul 19 2020, 10:15 AM · Restricted Project

Jul 16 2020

vtjnash added inline comments to D83870: [ConstantFolding] check applicability of AllOnes constant creation first.
Jul 16 2020, 9:05 PM · Restricted Project
vtjnash updated the diff for D83870: [ConstantFolding] check applicability of AllOnes constant creation first.

review

Jul 16 2020, 9:04 PM · Restricted Project

Jul 15 2020

vtjnash added a reviewer for D83870: [ConstantFolding] check applicability of AllOnes constant creation first: nathanchance.
Jul 15 2020, 6:59 AM · Restricted Project
Herald added a project to D83870: [ConstantFolding] check applicability of AllOnes constant creation first: Restricted Project.
Jul 15 2020, 6:57 AM · Restricted Project
vtjnash added a comment to D59730: [GVN] teach ConstantFolding correct handling of non-integral addrspace casts.

Thanks, that's a great reduction. I'll fix this shortly.

Jul 15 2020, 6:08 AM · Restricted Project

Jul 13 2020

vtjnash updated the diff for D69910: [GVN] Remove duplicated canCoerce tests.
Jul 13 2020, 10:09 PM · Restricted Project
vtjnash updated the diff for D59661: [GVN] Try to be more careful about handling constant zero and SVE.
Jul 13 2020, 10:06 PM · Restricted Project
vtjnash added a comment to D59730: [GVN] teach ConstantFolding correct handling of non-integral addrspace casts.

Merged as 19f01a484760d9a29f4f414b30da110dd6550191, e244f86f4dfd9e8982940d09294af522d3809d7f, and 2c7a07b59d5da54eba8e3e030e1cc040a88ecf58

Jul 13 2020, 7:07 PM · Restricted Project
vtjnash committed rG2c7a07b59d5d: [GVN] teach ConstantFolding correct handling of non-integral addrspace casts (authored by vtjnash).
[GVN] teach ConstantFolding correct handling of non-integral addrspace casts
Jul 13 2020, 7:06 PM
vtjnash committed rGe244f86f4dfd: [VNCoercion] avoid creating bitcast for zero offsets [NFCI] (authored by vtjnash).
[VNCoercion] avoid creating bitcast for zero offsets [NFCI]
Jul 13 2020, 7:06 PM
vtjnash committed rG19f01a484760: [GVN] add early exit to ConstantFoldLoadThroughBitcast [NFC] (authored by vtjnash).
[GVN] add early exit to ConstantFoldLoadThroughBitcast [NFC]
Jul 13 2020, 7:06 PM
vtjnash closed D59730: [GVN] teach ConstantFolding correct handling of non-integral addrspace casts.
Jul 13 2020, 7:06 PM · Restricted Project

Apr 29 2020

vtjnash added a comment to D44485: [MC] Always emit relocations for same-section function references.

We emit an offset table of the form:

ptrdiff_t symbol_offset = &f2 - &f1;

since we know this should normally be folded to a constant in the object file. This is normally rejected by a C compiler ("error: initializer element is not a compile-time constant", pointing to the f2 symbol) but LLVM accepts it. It used to be handled here (since the difference is a constant known here), though it seems the mingw-w64 linker later corrupts this expression.

Apr 29 2020, 10:11 AM · Restricted Project

Mar 19 2020

vtjnash added a comment to D74920: emit bitcode to file when Verifier fails.

I thought about that, but I didn't think it felt like it needed to. If the cwd is readonly, it'll just silently turn itself off.

Mar 19 2020, 7:31 AM · Restricted Project

Feb 20 2020

vtjnash updated the summary of D74920: emit bitcode to file when Verifier fails.
Feb 20 2020, 12:10 PM · Restricted Project
vtjnash updated the diff for D74920: emit bitcode to file when Verifier fails.

formatting

Feb 20 2020, 12:02 PM · Restricted Project
vtjnash updated the summary of D74920: emit bitcode to file when Verifier fails.
Feb 20 2020, 12:01 PM · Restricted Project
vtjnash created D74920: emit bitcode to file when Verifier fails.
Feb 20 2020, 11:33 AM · Restricted Project

Feb 6 2020

vtjnash updated subscribers of D74158: make the AsmPrinterHandler array public.
Feb 6 2020, 1:29 PM · Restricted Project