Page MenuHomePhabricator
Feed Advanced Search

Thu, Jul 2

dmajor added a comment to D76885: [lld][COFF][ELF][WebAssembly] Replace --[no-]threads /threads[:no] with --threads={1,2,...} /threads:{1,2,...}.

This commit breaks the build when -DLLVM_ENABLE_THREADS=OFF is used. Could you please take a look?

Thu, Jul 2, 9:10 AM · Restricted Project

Tue, Jun 30

dmajor added a comment to D65543: [Windows] Autolink with basenames and add libdir to libpath.

In that case, more specifically, the problem I'm seeing is that these flags are not making it through to a small handful of files: LLVMHello.dll, PrintFunctionNames.dll, AnnotateFunctions.dll, Attribute.dll. Let me dig in a little deeper...

Tue, Jun 30, 9:44 AM · Restricted Project, Restricted Project
dmajor created D82888: [cmake][Windows] Add libpath to CMAKE_MODULE_LINKER_FLAGS too.
Tue, Jun 30, 9:44 AM · Restricted Project
dmajor added a comment to D65543: [Windows] Autolink with basenames and add libdir to libpath.

After this change, PGO builds of clang on Windows can no longer find clang_rt.profile-x86_64.lib. Is there some cmake magic that could potentially make this seamless? Or is it expected that the burden is on the user to set up the paths manually during the LLVM_BUILD_INSTRUMENTED phase?

Tue, Jun 30, 9:44 AM · Restricted Project, Restricted Project
dmajor added a comment to D65543: [Windows] Autolink with basenames and add libdir to libpath.

After this change, PGO builds of clang on Windows can no longer find clang_rt.profile-x86_64.lib. Is there some cmake magic that could potentially make this seamless? Or is it expected that the burden is on the user to set up the paths manually during the LLVM_BUILD_INSTRUMENTED phase?

Tue, Jun 30, 8:39 AM · Restricted Project, Restricted Project

Mon, Jun 15

dmajor added a comment to D71786: RFC: [Support] On Windows, add optional support for {rpmalloc|snmalloc|mimalloc}.

That was on our c5.4xlarge AWS builders, which have 16 vCPUs according to the docs. I'm not sure offhand whether there is HT involved, but anyway I was testing with this patch applied to LLVM 10 which doesn't have the lldltojobs=all patch.

Mon, Jun 15, 9:46 AM · Restricted Project
dmajor added a comment to D71786: RFC: [Support] On Windows, add optional support for {rpmalloc|snmalloc|mimalloc}.

Very nice! This reduces the link time of Firefox's xul.dll from 13 minutes to 7 minutes.

Mon, Jun 15, 9:13 AM · Restricted Project

Thu, Jun 11

Herald updated subscribers of D71786: RFC: [Support] On Windows, add optional support for {rpmalloc|snmalloc|mimalloc}.
Thu, Jun 11, 12:40 PM · Restricted Project

May 12 2020

dmajor accepted D79787: [IndVarSimplify][LoopUtils] Avoid TOCTOU/ordering issues (PR45835).

Thanks a lot for taking this over! Looks good from my side but you may wish to wait for a more expert review.

May 12 2020, 9:39 AM · Restricted Project
dmajor added a comment to D79787: [IndVarSimplify][LoopUtils] Avoid TOCTOU/ordering issues (PR45835).

Is it a problem that the test case doesn't fail on trunk before the patch, since SCEVMinMaxExpr is no longer automatically high cost? I noted this in D79720 but wasn't sure how to nudge the code around it.

May 12 2020, 8:01 AM · Restricted Project
dmajor abandoned D79720: [IndVarSimplify][LoopUtils] Track rewrite cost per unique BB (PR45835).

Abandoning in favor of D79787. Thanks!

May 12 2020, 8:01 AM · Restricted Project

May 11 2020

dmajor added a comment to D79720: [IndVarSimplify][LoopUtils] Track rewrite cost per unique BB (PR45835).

Yes, but it is still not obvious to me as to why that happens?
It's the same PHI node, we are asking about the same value, for the same basic block.
Why do we not find expansion first time but do find it second time?
Did we perform some expansion inbetween?

May 11 2020, 11:50 AM · Restricted Project
dmajor added a comment to D79720: [IndVarSimplify][LoopUtils] Track rewrite cost per unique BB (PR45835).

Note: the test case does not currently fail on trunk before this change (although it does fail on the 10.0 release branch, where I would eventually like to get this merged) because after D73744 it is no longer blanket assumed that SCEVMinMaxExpr is high cost. Any pointers on how I could modify the test to get around that? (I'm not familiar with why this IR is considered to have a SCEVMinMaxExpr or what it would take to make it high cost.)

May 11 2020, 9:39 AM · Restricted Project
dmajor created D79720: [IndVarSimplify][LoopUtils] Track rewrite cost per unique BB (PR45835).
May 11 2020, 9:39 AM · Restricted Project

May 5 2020

dmajor accepted D79451: [libclang] Remove duplicate dependency on LLVMSupport.

Thanks for the fast response! lgtm

May 5 2020, 5:17 PM · Restricted Project

Apr 24 2020

dmajor added a comment to D54370: [codeview] Expose -gcodeview-ghash for global type hashing.

Should the directions in http://blog.llvm.org/2018/01/improving-link-time-on-windows-with.html be updated with the new flags?

Apr 24 2020, 8:04 AM

Feb 26 2020

dmajor added a comment to D74784: [driver][darwin] Don't use -platform_version flag by default.

Thanks! I'm still working on getting commit access, would one of you be able to submit for me?

Feb 26 2020, 2:47 PM · Restricted Project, Restricted Project
dmajor updated the diff for D74784: [driver][darwin] Don't use -platform_version flag by default.

Updated the tests: three cases for {default, old, new} linkers in the platform-version tests; left alone the tests not specifically targeting this path.

Feb 26 2020, 9:50 AM · Restricted Project, Restricted Project

Feb 25 2020

dmajor added a comment to D74784: [driver][darwin] Don't use -platform_version flag by default.

@steven_wu, ping, could you clarify about the tests please?

Feb 25 2020, 10:00 AM · Restricted Project, Restricted Project

Feb 19 2020

dmajor added a comment to D74784: [driver][darwin] Don't use -platform_version flag by default.

I forgot if there is reason to use the option by default at all time (I did ask that in the previous review but Alex might have given more context offline).

Feb 19 2020, 9:52 AM · Restricted Project, Restricted Project

Feb 18 2020

dmajor created D74784: [driver][darwin] Don't use -platform_version flag by default.
Feb 18 2020, 12:08 PM · Restricted Project, Restricted Project

Feb 17 2020

dmajor accepted D74501: [TSan] Add CMake check for minimal SDK requirements on Darwin.
In D74501#1879277, @yln wrote:

@dmajor: are you okay with this?

Feb 17 2020, 1:25 PM · Restricted Project, Restricted Project

Feb 12 2020

dmajor added inline comments to D71579: [driver][darwin] Pass -platform_version flag to the linker instead of the -<platform>_version_min flag.
Feb 12 2020, 8:33 AM · Restricted Project, Restricted Project

Feb 10 2020

dmajor added a comment to D71060: [LLD][ELF] Add time-trace to ELF LLD (2/2).

Should this go into the docs / man page / release notes / etc.?

Feb 10 2020, 7:02 AM · Restricted Project

Feb 5 2020

dmajor added a comment to D74059: [TSan] Ensure we can compile the runtime with older SDKs.

Thanks @yln! For posterity, the commit was rGc32d809e9cae8da7d3016b6cb30e2a2a9c9e2762 / https://github.com/llvm/llvm-project/commit/c32d809e9cae8da7d3016b6cb30e2a2a9c9e2762.

Feb 5 2020, 11:52 AM · Restricted Project, Restricted Project
dmajor accepted D74059: [TSan] Ensure we can compile the runtime with older SDKs.

The build was successful, thank you!

Feb 5 2020, 10:43 AM · Restricted Project, Restricted Project

Feb 4 2020

dmajor added a comment to D73973: Revert "[TSan] #include header instead of forward declaring interceptees".

Incidentally I wrote up this same patch and applied it locally earlier today, and our build still fails with the error below. Any ideas?

Feb 4 2020, 9:35 AM · Restricted Project, Restricted Project

Jan 22 2020

dmajor added a comment to D67678: PR17164: Change clang's default behavior from -flax-vector-conversions=all to -flax-vector-conversions=integer..

I noticed that this was merged to the 10.0 release branch. Should the merge be reverted while the dust settles on the trunk implementation?

Jan 22 2020, 12:42 PM · Restricted Project

Jan 8 2020

dmajor added a comment to D72167: Add support for __declspec(guard(nocf)).

I've confirmed that the current patch fixes our CFG failures. Thanks again!

Jan 8 2020, 1:27 PM · Restricted Project, Restricted Project

Jan 7 2020

dmajor added a comment to D72167: Add support for __declspec(guard(nocf)).

Is the current patch an interdiff? It would be helpful to have the diff against the master repo; Phabricator can take care of showing interdiffs if necessary.

Jan 7 2020, 3:24 PM · Restricted Project, Restricted Project

Jan 3 2020

dmajor added a comment to D72167: Add support for __declspec(guard(nocf)).

Thanks for doing this!

Jan 3 2020, 10:53 AM · Restricted Project, Restricted Project

Dec 5 2019

dmajor added a comment to D65761: Add Windows Control Flow Guard checks (/guard:cf)..
In D65761#1772031, @rnk wrote:

I think -fcf-protection and __attribute__((nocf_check)) have to do with CET and Intel's endbranch instruction or what have you. Similar goals, different implementation. I think at this point it's "patches welcome". =S

Dec 5 2019, 6:26 PM · Restricted Project, Restricted Project

Dec 3 2019

dmajor added a comment to D65761: Add Windows Control Flow Guard checks (/guard:cf)..

Are there any plans to implement __declspec(guard(nocf)) or an equivalent mechanism? __attribute__((nocf_check)) doesn't do anything without the -fcf-protection flag. (https://bugs.llvm.org/show_bug.cgi?id=44096)

Dec 3 2019, 6:13 AM · Restricted Project, Restricted Project
dmajor accepted D70910: [compiler-rt] Add a critical section when flushing gcov counters.

This lgtm from the Windows side but I'd feel better if you got an additional review from someone who knows gcov better than me.

Dec 3 2019, 6:04 AM · Restricted Project, Restricted Project

Dec 2 2019

dmajor updated subscribers of D70910: [compiler-rt] Add a critical section when flushing gcov counters.

The extra setup/teardown work for Windows relative to Linux is unfortunate. How would you feel about using an SRWLock instead?

Are there minimum Windows version requirements for compiler-rt? SRWLock is only available on Win7 forward, I think.

Dec 2 2019, 12:27 PM · Restricted Project, Restricted Project
dmajor added a comment to D70910: [compiler-rt] Add a critical section when flushing gcov counters.

The extra setup/teardown work for Windows relative to Linux is unfortunate. How would you feel about using an SRWLock instead? They accept a static initializer. (https://docs.microsoft.com/en-us/windows/win32/sync/slim-reader-writer--srw--locks)

Dec 2 2019, 10:43 AM · Restricted Project, Restricted Project

Nov 19 2019

dmajor added inline comments to D68351: [profile] Add a mode to continuously sync counter updates to a file.
Nov 19 2019, 12:24 PM · Restricted Project, Restricted Project, Restricted Project

Nov 18 2019

dmajor added a comment to D65761: Add Windows Control Flow Guard checks (/guard:cf)..

Hi, I filed https://bugs.llvm.org/show_bug.cgi?id=44049 for some strange crashes that we're seeing because the CFG code overwrites the lower byte of function pointers before jumping to them. (Commenting separately here because I was unable to CC @ajpaverd in Bugzilla)

Nov 18 2019, 12:48 PM · Restricted Project, Restricted Project

Oct 1 2019

dmajor added a comment to D55056: [CMake] Default options for faster executables on MSVC.

For posterity: this was reverted in r361837.

Oct 1 2019, 6:25 AM · Restricted Project

Aug 2 2019

Herald added a reviewer for D56326: [OpenMP 5.0] Parsing/sema support for "omp declare mapper" directive: jdoerfert.

While debugging something else, I noticed that with this patch, Decl now has 33 bits worth of bitfields, so it has gained an extra word. Is that ok? Just want to make sure it wasn't unintentional.

Aug 2 2019, 11:36 AM · Restricted Project, Restricted Project

Jul 30 2019

dmajor committed rG027bb52790ed: [COFF][ARM64] Reorder handling of aarch64 MSVC builtins (authored by dmajor).
[COFF][ARM64] Reorder handling of aarch64 MSVC builtins
Jul 30 2019, 8:35 AM
dmajor committed rL367323: [COFF][ARM64] Reorder handling of aarch64 MSVC builtins.
[COFF][ARM64] Reorder handling of aarch64 MSVC builtins
Jul 30 2019, 8:35 AM
dmajor closed D65403: [COFF, ARM64] Reorder handling of aarch64 MSVC builtins.
Jul 30 2019, 8:35 AM · Restricted Project, Restricted Project

Jul 29 2019

dmajor added inline comments to D65403: [COFF, ARM64] Reorder handling of aarch64 MSVC builtins.
Jul 29 2019, 12:26 PM · Restricted Project, Restricted Project
dmajor created D65403: [COFF, ARM64] Reorder handling of aarch64 MSVC builtins.
Jul 29 2019, 9:46 AM · Restricted Project, Restricted Project

May 16 2019

dmajor added inline comments to D62004: [LLD] [MinGW] Allow requesting PDB output without giving a file name.
May 16 2019, 6:49 AM · Restricted Project

May 13 2019

dmajor added a comment to D47916: [asan, myriad] Use local pool for new/delete when ASan run-time is not up.

No rush, I don't have any pressing need for a fix. I was just curious.

May 13 2019, 11:31 AM · Restricted Project
dmajor added inline comments to D47916: [asan, myriad] Use local pool for new/delete when ASan run-time is not up.
May 13 2019, 10:16 AM · Restricted Project

Apr 26 2019

dmajor added a comment to D61095: [AArch64][Windows] Compute function length correctly in unwind tables..

I've confirmed that this fixes the Firefox repro, thanks!

Apr 26 2019, 9:46 AM · Restricted Project

Mar 12 2019

dmajor added a comment to D56675: [lld] add options for context-sensitive PGO..

Are there any bugs/reviews/etc. that I can subscribe to for the implementation of this in COFF?

Mar 12 2019, 10:18 AM · Restricted Project

Feb 27 2019

dmajor accepted D58739: [COFF] Add address-taken import thunks to the fid table.

Oh you know what, I bet it's https://reviews.llvm.org/D53540#1326473. Rust has an LLVM roll in progress as I write this, so this should just sort itself out.

Feb 27 2019, 6:17 PM · Restricted Project
dmajor added inline comments to D58739: [COFF] Add address-taken import thunks to the fid table.
Feb 27 2019, 3:27 PM · Restricted Project
dmajor added a comment to D58739: [COFF] Add address-taken import thunks to the fid table.

The function that I'm interested in (a Rust function, on arm64, if it matters) is apparently failing the IMAGE_SYM_DTYPE_FUNCTION test. Is there any other category of thing that it could be?

Feb 27 2019, 3:08 PM · Restricted Project
dmajor added a comment to D58739: [COFF] Add address-taken import thunks to the fid table.

So after some further digging I was wrong, and my particular bug isn't to do with S->kind(). :(

Feb 27 2019, 2:43 PM · Restricted Project

Feb 25 2019

dmajor committed rGb4e16e6ad152: [winasan] Unpoison stack memory when threads exit (redux) (authored by dmajor).
[winasan] Unpoison stack memory when threads exit (redux)
Feb 25 2019, 5:38 PM
dmajor committed rCRT354836: [winasan] Unpoison stack memory when threads exit (redux).
[winasan] Unpoison stack memory when threads exit (redux)
Feb 25 2019, 5:35 PM
dmajor committed rL354836: [winasan] Unpoison stack memory when threads exit (redux).
[winasan] Unpoison stack memory when threads exit (redux)
Feb 25 2019, 5:35 PM
dmajor closed D58641: [winasan] Unpoison stack memory when threads exit (redux).
Feb 25 2019, 5:35 PM · Restricted Project, Restricted Project
dmajor added a comment to D58641: [winasan] Unpoison stack memory when threads exit (redux).

@rnk would you be able to test this out on the nacl browser_tests before landing?

Feb 25 2019, 11:44 AM · Restricted Project, Restricted Project
dmajor created D58641: [winasan] Unpoison stack memory when threads exit (redux).
Feb 25 2019, 11:42 AM · Restricted Project, Restricted Project
Herald added a project to D55056: [CMake] Default options for faster executables on MSVC: Restricted Project.
Feb 25 2019, 7:58 AM · Restricted Project

Feb 12 2019

dmajor committed rG5b07e3040885: [gn build] Separate debug and optimization settings (authored by dmajor).
[gn build] Separate debug and optimization settings
Feb 12 2019, 2:25 PM
dmajor committed rL353888: [gn build] Separate debug and optimization settings.
[gn build] Separate debug and optimization settings
Feb 12 2019, 2:24 PM
dmajor closed D58075: [gn build] Separate debug and optimization settings.
Feb 12 2019, 2:24 PM · Restricted Project

Feb 11 2019

dmajor created D58075: [gn build] Separate debug and optimization settings.
Feb 11 2019, 2:17 PM · Restricted Project

Feb 4 2019

dmajor committed rG1137fce9e9e5: gn build: Windows: use a more standard format for PDB filenames (authored by dmajor).
gn build: Windows: use a more standard format for PDB filenames
Feb 4 2019, 1:29 PM
dmajor committed rL353099: gn build: Windows: use a more standard format for PDB filenames.
gn build: Windows: use a more standard format for PDB filenames
Feb 4 2019, 1:29 PM
dmajor committed rGd1934853a8b1: gn build: Revert r353094 (bad merge) (authored by dmajor).
gn build: Revert r353094 (bad merge)
Feb 4 2019, 1:25 PM
dmajor committed rL353098: gn build: Revert r353094 (bad merge).
gn build: Revert r353094 (bad merge)
Feb 4 2019, 1:25 PM
dmajor committed rG1469ff417bba: gn build: Windows: use a more standard format for PDB filenames (authored by dmajor).
gn build: Windows: use a more standard format for PDB filenames
Feb 4 2019, 1:24 PM
dmajor committed rL353094: gn build: Windows: use a more standard format for PDB filenames.
gn build: Windows: use a more standard format for PDB filenames
Feb 4 2019, 1:20 PM
dmajor closed D57613: gn build: Windows: use a more standard format for PDB filenames.
Feb 4 2019, 1:20 PM · Restricted Project
dmajor committed rG3c659cb267dd: gn build: Windows: write PDBs when is_debug (authored by dmajor).
gn build: Windows: write PDBs when is_debug
Feb 4 2019, 1:15 PM
dmajor committed rL353093: gn build: Windows: write PDBs when is_debug.
gn build: Windows: write PDBs when is_debug
Feb 4 2019, 1:13 PM
dmajor closed D57612: gn build: Windows: write PDBs when is_debug.
Feb 4 2019, 1:13 PM · Restricted Project
dmajor added a comment to D57613: gn build: Windows: use a more standard format for PDB filenames.

Do you care much about this? The naming here might be somewhat unconventional, but it's more flexible (consider the completely hypothetical case of a build having both "chrome.exe" and "chorme.dll" – if you call the pdb for both "chrome.dll" they'll get in each other's way, but with this scheme it works fine) and ever so slightly less code. So I weakly prefer what we currently have, but I almost never do debug builds so I don't care all that much. I'll accept this so that you can land it if you want, but I'd weakly prefer if you didn't :-)

Feb 4 2019, 8:19 AM · Restricted Project

Feb 1 2019

dmajor added a comment to D57575: [LLD] [COFF] Create range extension thunks for ARM64.

@ruiu - What do you think about this and D57574 wrt backporting to 8.0? They should be pretty safe, but on the other hand I don't think they're something that many random users will run into either.

I wouldn't personally request cherrypicking that change at this point because it's not a regression from 7.0 and doesn't seem like a big issue. But if you feel otherwise, feel free to submit a cherrypick request.

Feb 1 2019, 7:01 PM · Restricted Project
dmajor added inline comments to D57613: gn build: Windows: use a more standard format for PDB filenames.
Feb 1 2019, 12:36 PM · Restricted Project
dmajor created D57613: gn build: Windows: use a more standard format for PDB filenames.
Feb 1 2019, 12:35 PM · Restricted Project
dmajor created D57612: gn build: Windows: write PDBs when is_debug.
Feb 1 2019, 12:33 PM · Restricted Project

Jan 22 2019

dmajor added a comment to D56986: COFF, ELF: ICF: Perform 2 rounds of relocation hash propagation..
In D56986#1366874, @pcc wrote:

Is this worth picking over to 8.0 after it lands?

I'm more ambivalent about this one since it's more of a perf improvement than a fix for a perf regression like D56955 is. But I have no strong objections to doing so.

That's fair; we can just cherry-pick it to our 8.0 branch internally if it doesn't end up merged here :) It's early in the release cycle though, so I figured it was okay for small-ish changes like this which result in big improvements to go in.

Jan 22 2019, 2:07 PM

Jan 18 2019

dmajor abandoned D56953: [COFF] Limit PR39667 ICF partitioning changes to ARM64.

Even better -- thanks!

Jan 18 2019, 10:30 PM
dmajor created D56953: [COFF] Limit PR39667 ICF partitioning changes to ARM64.
Jan 18 2019, 6:01 PM

Jan 16 2019

dmajor added a comment to D56813: [AArch64] [Windows] Share unwind codes between epilogues with identical unwind codes.

The updated patch works for Firefox too.

Jan 16 2019, 7:48 PM
dmajor added a comment to D56813: [AArch64] [Windows] Share unwind codes between epilogues with identical unwind codes.

I've confirmed that I can build Firefox successfully with this patch on top of latest trunk. Thank you @ssijaric!

Jan 16 2019, 4:11 PM

Jan 9 2019

dmajor committed rL350774: Don't require a null terminator when loading objects.
Don't require a null terminator when loading objects
Jan 9 2019, 3:40 PM
dmajor closed D56475: Don't require a null terminator when loading objects.
Jan 9 2019, 3:40 PM

Jan 8 2019

dmajor added a comment to D56419: [gn build] Move .gn file to the root of the monorepo.

I'd be interested in feedback from more people if anyone has any (e.g. "don't care").

Jan 8 2019, 3:13 PM

Nov 19 2018

dmajor added a comment to D54723: [COFF] Add exported functions to gfids table for /guard:cf.

Thanks! I applied the patch and confirmed that it fixes the problematic Firefox DLL. (I don't know if I'm allowed to formally approve it though...)

Nov 19 2018, 5:21 PM

Nov 9 2018

dmajor added inline comments to D53540: [COFF, ARM64] Implement support for SEH extensions __try/__except/__finally.
Nov 9 2018, 7:08 PM

Nov 8 2018

dmajor added inline comments to D53540: [COFF, ARM64] Implement support for SEH extensions __try/__except/__finally.
Nov 8 2018, 12:29 PM
dmajor added inline comments to D53540: [COFF, ARM64] Implement support for SEH extensions __try/__except/__finally.
Nov 8 2018, 7:57 AM

Nov 2 2018

dmajor added inline comments to D53988: Close file mapping handle on Windows, so flushed gcda files can be removed while the process is in execution.
Nov 2 2018, 12:59 PM

Oct 22 2018

dmajor added a comment to D51204: [COFF, ARM64] Add MS intrinsics: __getReg, _ReadStatusReg, _WriteStatusReg.

Will abandon this patch since I have implementations of these which I will upstream soon.

Oct 22 2018, 3:37 PM
Herald updated subscribers of D50166: [ARM64] [Windows] MCLayer support for exception handling.
Oct 22 2018, 2:48 PM
dmajor added a comment to D53431: clang appears not to respect __attribute__((noinline)).

The intent of noinline in LLVM's IR is to block inlining, not all interprocedural optimizations. So I don't think this is actually a bug

Oct 22 2018, 2:17 PM

Oct 15 2018

dmajor created P8111 aarch64 crash 2.
Oct 15 2018, 10:13 AM

Oct 8 2018

dmajor changed the visibility for P8110 aarch64 crash 1.
Oct 8 2018, 7:55 AM
dmajor created P8110 aarch64 crash 1.
Oct 8 2018, 7:53 AM

Oct 3 2018

dmajor added a comment to D52091: [winasan] Unpoison the stack in NtTerminateThread.

Relanded in r343606.

Oct 3 2018, 2:55 PM