[PATCH] D50882: [ThinLTO] Correct documentation on default number of threads

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: dexonsmith, steven_wu, eraman, inglorion, mehdi_amini. The number of threads used for ThinLTO backend parallelism was dropped to the number of cores in r284618 to avoid oversubscribing physical cores due to

[PATCH] D50882: [ThinLTO] Correct documentation on default number of threads

This revision was automatically updated to reflect the committed changes. Closed by commit rC340021: [ThinLTO] Correct documentation on default number of threads (authored by tejohnson, committed by ). Changed prior to commit: https://reviews.llvm.org/D50882?vs=161161&id=161223#toc Repository:

[PATCH] D46700: [ThinLTO] Add testing of new summary index format to a couple CFI tests

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: eraman, inglorion, mehdi_amini. Adds testing of combined index summary entries in disassembly format to CFI tests that were already testing the bitcode format. Depends on https://reviews.llvm.org/D46699.

[PATCH] D46700: [ThinLTO] Add testing of new summary index format to a couple CFI tests

tejohnson updated this revision to Diff 146836. tejohnson added a comment. Update tests for changes to https://reviews.llvm.org/D46699 Repository: rC Clang https://reviews.llvm.org/D46700 Files: test/CodeGen/thinlto-distributed-cfi-devirt.ll test/CodeGen/thinlto-distributed-cfi.ll Inde

[PATCH] D46700: [ThinLTO] Add testing of new summary index format to a couple CFI tests

tejohnson updated this revision to Diff 148740. tejohnson added a comment. Update tests so they won't get bot failures (don't try to match path, module hash). Ping - this can go in now that https://reviews.llvm.org/D46699 is in. Repository: rC Clang https://reviews.llvm.org/D46700 Files:

[PATCH] D38517: Enabling new pass manager in LTO (and thinLTO) link step via -fexperimental-new-pass-manager option

tejohnson added a comment. Thanks for adding this. Please add a test like the one for -ffunction-sections (tools/clang/test/Driver/gold-lto-sections.c). Repository: rL LLVM https://reviews.llvm.org/D38517 ___ cfe-commits mailing list cfe-commits

[PATCH] D38517: Enabling new pass manager in LTO (and thinLTO) link step via -fexperimental-new-pass-manager option

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM thanks! Repository: rL LLVM https://reviews.llvm.org/D38517 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.ll

[PATCH] D51198: [LTO] Fix -save-temps with LTO and unnamed globals.

tejohnson added a comment. Thanks. Can you fix the same code in EmitAssemblyWithNewPassManager? Repository: rC Clang https://reviews.llvm.org/D51198 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/lis

[PATCH] D51198: [LTO] Fix -save-temps with LTO and unnamed globals.

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM Repository: rC Clang https://reviews.llvm.org/D51198 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/

[PATCH] D42995: [ThinLTO] Ignore object files with empty ThinLTO index

tejohnson added a comment. > Empty ThinLTOIndexFile signals that we don't need this module during > linking. Not the only case actually. We now also pass an empty index file when we want to compile the bitcode down to object without applying any LTO optimization (there are a few cases where we

[PATCH] D42995: [ThinLTO] Ignore object files with empty ThinLTO index

tejohnson added a comment. In https://reviews.llvm.org/D42995#1000155, @vitalybuka wrote: > In https://reviews.llvm.org/D42995#125, @tejohnson wrote: > > > > Empty ThinLTOIndexFile signals that we don't need this module during > > > linking. > > > > Not the only case actually. We now also pa

[PATCH] D42995: [ThinLTO] Allow indexing to request backend to ignore the module

tejohnson added inline comments. Comment at: clang/lib/CodeGen/BackendUtil.cpp:1176 + // Output of this module is not needed, but build system may not know + // that. So we need to generated empty valid object file. + EmptyModule = llvm::make_unique("empty", M->get

[PATCH] D42995: [ThinLTO] Allow indexing to request backend to ignore the module

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM https://reviews.llvm.org/D42995 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo

[PATCH] D34055: Be more strict when checking the -flto option value

tejohnson added a comment. Make sure you add cfe-commits to the mailing list for clang changes (and llvm-commits for llvm changes, I accidentally added that one first and then fixed it), since all patches should go to the full mailing list. This fix looks correct to me. Please add a test case t

[PATCH] D34055: Be more strict when checking the -flto option value

tejohnson added a comment. In https://reviews.llvm.org/D34055#777005, @yamaguchi wrote: > I think we don't need additional testcase, because this is non-functional > change. Driver.cpp will emit error if value was not "thin" nor "full". This > testcase is at clang/test/CodeGen/thinlto-backend-o

[PATCH] D34055: Be more strict when checking the -flto option value

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM, thanks for the fix! https://reviews.llvm.org/D34055 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi

[PATCH] D34546: docs: Add documentation for the ThinLTO cache pruning policy string.

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM https://reviews.llvm.org/D34546 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo

[PATCH] D47597: IRGen: Write .dwo files when -split-dwarf-file is used together with -fthinlto-index.

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM thanks! https://reviews.llvm.org/D47597 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/

[PATCH] D46700: [ThinLTO] Add testing of new summary index format to a couple CFI tests

tejohnson added a comment. Herald added a subscriber: steven_wu. ping Repository: rC Clang https://reviews.llvm.org/D46700 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D46700: [ThinLTO] Add testing of new summary index format to a couple CFI tests

This revision was automatically updated to reflect the committed changes. Closed by commit rL333966: [ThinLTO] Add testing of new summary index format to a couple CFI tests (authored by tejohnson, committed by ). Herald added a subscriber: llvm-commits. Repository: rL LLVM https://reviews.llvm

[PATCH] D47906: [ThinLTO] Add testing of summary index parsing to a couple CFI tests

tejohnson created this revision. tejohnson added reviewers: pcc, dexonsmith, mehdi_amini. Herald added subscribers: steven_wu, eraman, inglorion. Changes to some clang side tests to go with the summary parsing patch. Depends on https://reviews.llvm.org/D47905. Repository: rC Clang https://re

[PATCH] D34156: [LTO] Enable module summary emission by default for regular LTO

tejohnson accepted this revision. tejohnson added a comment. LGTM with following suggestions. Comment at: clang/lib/CodeGen/BackendUtil.cpp:776 + bool EmitLTOSummary = + (CodeGenOpts.PrepareForLTO && !CodeGenOpts.PrepareForThinLTO && + llvm::Triple(TheModule->getTar

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: dexonsmith, steven_wu, inglorion, mehdi_amini. Currently, -flto-unit is specified whenever LTO options are used (unless using the old LTO API). This causes vtable defs to be processed using regular LTO, whi

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson added a comment. In https://reviews.llvm.org/D53524#1271357, @pcc wrote: > The reason why LTO unit is always enabled is so that you can link translation > units compiled with `-fsanitize=cfi` and/or `-fwhole-program-vtables` against > translation units compiled without CFI/WPD. With t

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson added a comment. In https://reviews.llvm.org/D53524#1271387, @tejohnson wrote: > In https://reviews.llvm.org/D53524#1271357, @pcc wrote: > > > The reason why LTO unit is always enabled is so that you can link > > translation units compiled with `-fsanitize=cfi` and/or > > `-fwhole-pro

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson added a comment. In https://reviews.llvm.org/D53524#1276038, @pcc wrote: > In https://reviews.llvm.org/D53524#1274505, @tejohnson wrote: > > > In https://reviews.llvm.org/D53524#1271387, @tejohnson wrote: > > > > > In https://reviews.llvm.org/D53524#1271357, @pcc wrote: > > > > > > > Th

[PATCH] D53891: [LTO] Pass down LTOUnit codegen flag to bitcode writer

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: cfe-commits, dexonsmith, steven_wu, eraman, inglorion, mehdi_amini. Clang side patch for recording the LTOUnit flag in the index. Depends on https://reviews.llvm.org/D53890. Repository: rC Clang http

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson added a comment. In https://reviews.llvm.org/D53524#1279288, @tejohnson wrote: > In https://reviews.llvm.org/D53524#1276038, @pcc wrote: > > > In https://reviews.llvm.org/D53524#1274505, @tejohnson wrote: > > > > > In https://reviews.llvm.org/D53524#1271387, @tejohnson wrote: > > > > >

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

tejohnson updated this revision to Diff 181317. tejohnson marked 4 inline comments as done. tejohnson added a comment. Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D53891/new/ https://reviews.llvm.org/D53891 Files: include/clang/Basic/CodeGen

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

This revision was automatically updated to reflect the committed changes. Closed by commit rL350949: [LTO] Add option to enable LTOUnit splitting, and disable unless needed (authored by tejohnson, committed by ). Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D53891/n

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson updated this revision to Diff 172181. tejohnson added a comment. Address comments: Promote -flto-unit to clang driver option (and test it) Adjust LTOVisibility.rst to reflect change of default and new option. Repository: rC Clang https://reviews.llvm.org/D53524 Files: docs/LTOVis

[PATCH] D53891: [LTO] Pass down LTOUnit codegen flag to bitcode writer

tejohnson updated this revision to Diff 174329. tejohnson added a comment. - As discussed off-patch, will use a new, separate option to control this splitting, here I am using -f[no]split-lto-unit. - Switch the default of splitting lto units to off by default, unless compiled with CFI or -fwhol

[PATCH] D53891: [LTO] Pass down LTOUnit codegen flag to bitcode writer

tejohnson updated this revision to Diff 174332. tejohnson added a comment. Update a couple tests due to new default of splitting off and new flag. Repository: rC Clang https://reviews.llvm.org/D53891 Files: include/clang/Driver/Options.td include/clang/Driver/SanitizerArgs.h include/cl

[PATCH] D53891: [LTO] Pass down LTOUnit codegen flag to bitcode writer

tejohnson added a comment. In https://reviews.llvm.org/D53891#1300789, @tejohnson wrote: > - Switch the default of splitting lto units to off by default, unless > compiled with CFI or -fwhole-program-vtables. Thinking ahead to when I add the index based WPD implementation, we are going to wan

[PATCH] D53891: [LTO] Pass down LTOUnit codegen flag to bitcode writer

tejohnson updated this revision to Diff 174835. tejohnson added a comment. Update to use new module flag Repository: rC Clang https://reviews.llvm.org/D53891 Files: include/clang/Driver/Options.td include/clang/Driver/SanitizerArgs.h include/clang/Frontend/CodeGenOptions.def lib/Code

[PATCH] D53524: [ThinLTO] Enable LTOUnit only when it is needed

tejohnson abandoned this revision. tejohnson added a comment. Abandoned in favor of new approach in https://reviews.llvm.org/D53890/https://reviews.llvm.org/D53891. Repository: rC Clang https://reviews.llvm.org/D53524 ___ cfe-commits mailing lis

[PATCH] D55048: [ThinLTO] Allow importing of multiple symbols with same GUID

tejohnson created this revision. tejohnson added a reviewer: evgeny777. Herald added subscribers: dexonsmith, steven_wu, eraman, inglorion, mehdi_amini. The is the clang side of the fix in D55047 , to handle the case where two different modules have local variable

[PATCH] D55048: [ThinLTO] Allow importing of multiple symbols with same GUID

This revision was automatically updated to reflect the committed changes. Closed by commit rC347887: [ThinLTO] Allow importing of multiple symbols with same GUID (authored by tejohnson, committed by ). Changed prior to commit: https://reviews.llvm.org/D55048?vs=175814&id=175881#toc Repository:

[PATCH] D59304: Fix invocation of Gold plugin with LTO after r355331

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. lgtm Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59304/new/ https://reviews.llvm.org/D59304 ___ cfe-comm

[PATCH] D60163: [ThinLTO] Handle -fno-builtin* options for distributed backends

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: kristina, dexonsmith, steven_wu, eraman, inglorion, mehdi_amini. Herald added a project: clang. Without this patch, we were ignoring the -fno-builtin* options for clang invocations performing ThinLTO distr

[PATCH] D60516: [LTO] Add plumbing to save stats during LTO on Darwin.

tejohnson added a comment. LGTM with a minor fix needed below. Darwin still uses the old LTO API, which is why the lto Config based handling in the new LTO API (LTO.h/LTO.cpp) are not being used on this path. Comment at: llvm/lib/LTO/LTOCodeGenerator.cpp:101 +"lto-stats-f

[PATCH] D60163: [ThinLTO] Handle -fno-builtin* options for distributed backends

tejohnson added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60163/new/ https://reviews.llvm.org/D60163 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/

[PATCH] D61022: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM

tejohnson created this revision. tejohnson added a reviewer: xur. Herald added subscribers: dexonsmith, steven_wu, hiraditya, eraman, inglorion, mehdi_amini. Herald added projects: clang, LLVM. The opt level was not being passed down to the ThinLTO backend when invoked via clang (for distributed

[PATCH] D61022: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM

tejohnson marked an inline comment as done. tejohnson added inline comments. Comment at: llvm/test/tools/gold/X86/opt-level.ll:53 + ; CHECK-O1-OLDPM: select + ; The new PM does not do as many optimizations at O1 + ; CHECK-O1-NEWPM: phi mehdi_amini wrote: > Thi

[PATCH] D61022: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM

tejohnson marked an inline comment as done. tejohnson added inline comments. Comment at: llvm/test/tools/gold/X86/opt-level.ll:53 + ; CHECK-O1-OLDPM: select + ; The new PM does not do as many optimizations at O1 + ; CHECK-O1-NEWPM: phi tejohnson wrote: > mehdi

[PATCH] D61022: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM

tejohnson marked an inline comment as done. tejohnson added inline comments. Comment at: llvm/test/tools/gold/X86/opt-level.ll:53 + ; CHECK-O1-OLDPM: select + ; The new PM does not do as many optimizations at O1 + ; CHECK-O1-NEWPM: phi chandlerc wrote: > tejoh

[PATCH] D61022: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM

This revision was automatically updated to reflect the committed changes. Closed by commit rL359025: [ThinLTO] Pass down opt level to LTO backend and handle -O0 LTO in new PM (authored by tejohnson, committed by ). Changed prior to commit: https://reviews.llvm.org/D61022?vs=196263&id=196296#toc

[PATCH] D52323: Add necessary support for storing code-model to module IR.

tejohnson added a comment. Note that if you add a line like: "Depends on https://reviews.llvm.org/D52322"; in the summary that Phabricator will automatically link the two in the right way. Comment at: lib/CodeGen/CodeGenModule.cpp:569 + .Default(~0u); +if

[PATCH] D52323: Add necessary support for storing code-model to module IR.

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM https://reviews.llvm.org/D52323 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo

[PATCH] D57265: [PM/CC1] Add -f[no-]split-cold-code CC1 options to toggle splitting

tejohnson added a comment. Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Overall it looks ok to me, but I'd like Chandler to comment regarding the preferred way to do this with the new PM, since we don't tend to use booleans there in the PassBuilder to control passes. Is

[PATCH] D54176: [PGO] clang part of change for context-sensitive PGO.

tejohnson added a comment. Herald added a subscriber: jdoerfert. LGTM except for place noted below where I disagree with a change made earlier. Will let @davidxl chime in if he disagrees with me or has any other comments. Comment at: lib/CodeGen/BackendUtil.cpp:967 + if (Code

[PATCH] D37995: [Docs] Document cache pruning support for gold

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. LGTM with one fix noted. Comment at: docs/ThinLTO.rst:144 To help keep the size of the cache under control, ThinLTO supports cache pruning. Cache pruning is supported

[PATCH] D59673: [Clang] Harmonize Split DWARF options with llc

tejohnson added inline comments. Comment at: lib/CodeGen/BackendUtil.cpp:1345 Conf.RemarksPasses = CGOpts.OptRecordPasses; - Conf.DwoPath = CGOpts.SplitDwarfFile; + Conf.DwoPath = CGOpts.SplitDwarfOutput; switch (Action) { aaronpuchert wrote: > aaronpuche

[PATCH] D59673: [Clang] Harmonize Split DWARF options with llc

tejohnson added inline comments. Comment at: lib/CodeGen/BackendUtil.cpp:1345 Conf.RemarksPasses = CGOpts.OptRecordPasses; - Conf.DwoPath = CGOpts.SplitDwarfFile; + Conf.DwoPath = CGOpts.SplitDwarfOutput; switch (Action) { aaronpuchert wrote: > tejohnson

[PATCH] D59673: [Clang] Harmonize Split DWARF options with llc

tejohnson added a comment. lgtm for the LTO bits. Suggestion below for comment. Comment at: llvm/include/llvm/LTO/Config.h:92 + /// The name for the split debug info file used for the DW_AT_[GNU_]dwo_name + /// attribute in the skeleton CU. + std::string SplitDwarfFile;

[PATCH] D59673: [Clang] Harmonize Split DWARF options with llc

tejohnson added inline comments. Comment at: llvm/include/llvm/LTO/Config.h:92 + /// The name for the split debug info file used for the DW_AT_[GNU_]dwo_name + /// attribute in the skeleton CU. + std::string SplitDwarfFile; aaronpuchert wrote: > tejohnson wrot

[PATCH] D64458: add -fthinlto-index= option to clang-cl

tejohnson added a comment. LGTM, but please wait for @pcc to make sure he is satisfied. I think it would be good to port the file type deduction to clang (possibly as a follow on), to make these consistent. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.

[PATCH] D64610: [clang] allow -fthinlto-index= without -x ir

tejohnson accepted this revision. tejohnson added a comment. LGTM with the comment fix from @MaskRay Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D64610/new/ https://reviews.llvm.org/D64610 ___ cfe-com

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1515176 , @tejohnson wrote: > In D61634#1512020 , @gchatelet wrote: > > > AFAIU here is a coarse plan of what needs to happen > > > > 1. Add a `no-builtin` clang function attribu

[PATCH] D65009: [LTO] Don't mark regular LTO units with EnableSplitLTOUnit

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: dexonsmith, steven_wu, inglorion, Prazek, mehdi_amini. Herald added a project: clang. LTO Unit splitting does not apply to regular LTO modules, only ThinLTO (which must be consistently split into regular an

[PATCH] D65009: [LTO] Don't mark regular LTO units with EnableSplitLTOUnit

tejohnson added a comment. In D65009#1594009 , @pcc wrote: > Sorry, just realized this. If I do > > clang++ -c -flto a.cpp # "split" > clang++ -c -flto=thin b.cpp -fwhole-program-vtables # non-split > clang++ a.o b.o > > > this should fail, right? If

[PATCH] D65009: [LTO] Always mark regular LTO units with EnableSplitLTOUnit=1

tejohnson updated this revision to Diff 210918. tejohnson added a comment. Address comments Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D65009/new/ https://reviews.llvm.org/D65009 Files: lib/CodeGen/BackendUtil.cpp test/CodeGen/split-lto-unit.c Index: tes

[PATCH] D65009: [LTO] Always mark regular LTO units with EnableSplitLTOUnit=1

This revision was automatically updated to reflect the committed changes. Closed by commit rL366623: [LTO] Always mark regular LTO units with EnableSplitLTOUnit=1 (authored by tejohnson, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Repository: rL LLVM

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. Herald added a reviewer: jdoerfert. I had some time to work on this finally late last week. I decided the most straightforward thing was to implement the necessary interface changes to the TLI analysis to make it require a Function (without any changes yet to how that

[PATCH] D63626: [clang][NewPM] Remove exception handling before loading pgo sample profile data

tejohnson added a comment. Adding Wei who works on SamplePGO on our end currently. Comment at: llvm/lib/Passes/PassBuilder.cpp:665-668 +// We must also remove exception handling before attaching sample profile +// data. +MPM.addPass( +createModuleToPostOrder

[PATCH] D63976: Allow clang -Os and -Oz to work with -flto and lld

tejohnson added inline comments. Comment at: llvm-9.0.0-20190629/clang/lib/Driver/ToolChains/CommonArgs.cpp:395 + if(OOpt == "s" || OOpt == "z") +OOpt = "3"; +} Os/Oz are closer to O2 than O3 (which allows much more aggressive code size increasi

[PATCH] D60162: [ThinLTO] Support TargetLibraryInfoImpl in the backend

tejohnson updated this revision to Diff 199320. tejohnson added a comment. Herald added subscribers: cfe-commits, hiraditya, eraman. Herald added a project: clang. Rework using module flags. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60162/new/

[PATCH] D60162: [ThinLTO] Support TargetLibraryInfoImpl in the backend

tejohnson added a comment. In D60162#1498392 , @tejohnson wrote: > In D60162#1498288 , @tejohnson wrote: > > > Working on this one again as it previously wasn't causing any issues but > > just now got exposed in mu

[PATCH] D60163: [ThinLTO] Handle -fno-builtin* options for distributed backends

tejohnson abandoned this revision. tejohnson added a comment. The parent revision now includes both clang and llvm side changes. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60163/new/ https://reviews.llvm.org/D60163 __

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1501066 , @gchatelet wrote: > In D61634#1500453 , @efriedma wrote: > > > My main blocker is that I want to make sure we're moving in the right > > direction: towards LLVM IR wit

[PATCH] D60162: [ThinLTO] Add module flags for TargetLibraryInfoImpl and use in LTO backends

tejohnson marked 2 inline comments as done. tejohnson added inline comments. Comment at: clang/test/CodeGen/svml-calls.ll:16 + +define void @sin_f64(double* nocapture %varray) { +; CHECK-LABEL: @sin_f64( steven_wu wrote: > Personally, I think codegen tests like t

[PATCH] D60162: [ThinLTO] Add module flags for TargetLibraryInfoImpl and use in LTO backends

tejohnson marked an inline comment as done. tejohnson added inline comments. Comment at: llvm/lib/LTO/LTOBackend.cpp:221 +static TargetLibraryInfoImpl *createTLII(Module &Mod, TargetMachine *TM) { + TargetLibraryInfoImpl *TLII = tejohnson wrote: > steven_wu wr

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1502138 , @hfinkel wrote: > In D61634#1502043 , @efriedma wrote: > > > > I have a related patch that turns -fno-builtin* options into module flags > > > > Do you have any opinion

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1502685 , @gchatelet wrote: > In D61634#1502201 , @tejohnson wrote: > > > Using function level attributes instead of module flags does provide finer > > grained control and avoi

[PATCH] D60162: [ThinLTO] Add module flags for TargetLibraryInfoImpl and use in LTO backends

tejohnson added a comment. There is discussion of using function attributes to control this instead, see https://reviews.llvm.org/D61634. I'll hold off on making the planned changes here until the overall approach is decided. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION h

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1512020 , @gchatelet wrote: > AFAIU here is a coarse plan of what needs to happen > > 1. Add a `no-builtin` clang function attribute that has the same semantic as > the `no-builtin` cmd line argument >

[PATCH] D60162: [ThinLTO] Add module flags for TargetLibraryInfoImpl and use in LTO backends

tejohnson abandoned this revision. tejohnson added a comment. As per D61634 , we are going to use function attributes instead. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60162/new/ https://reviews.llvm.org/D6016

[PATCH] D55620: [ThinLTO] Clang changes to utilize new pass to handle chains of aliases

tejohnson created this revision. tejohnson added reviewers: pcc, davidxl. Herald added subscribers: dexonsmith, steven_wu, inglorion, mehdi_amini. As with NameAnonGlobals, invoke the new CanonicalizeAliases via clang when using the new PM. Depends on D54507 . Re

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

tejohnson marked 2 inline comments as done. tejohnson added inline comments. Comment at: lib/CodeGen/BackendUtil.cpp:831 + *OS, CodeGenOpts.EmitLLVMUseLists, EmitLTOSummary, + /*EmitModuleHash=*/false)); } pcc wrote: > Why add this argument

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

tejohnson updated this revision to Diff 179305. tejohnson added a comment. Implement comment suggestions. Added test for new error. Also test changes for LLVM side changes in D53890 . Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

tejohnson added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D53891/new/ https://reviews.llvm.org/D53891 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/l

[PATCH] D55620: [ThinLTO] Clang changes to utilize new pass to handle chains of aliases

tejohnson added a comment. ping on the clang side change as well Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D55620/new/ https://reviews.llvm.org/D55620 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http:/

[PATCH] D55620: [ThinLTO] Clang changes to utilize new pass to handle chains of aliases

tejohnson updated this revision to Diff 180272. tejohnson added a comment. Update comment Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D55620/new/ https://reviews.llvm.org/D55620 Files: lib/CodeGen/BackendUtil.cpp test/CodeGen/lto-newpm-pipeline.c Index: t

[PATCH] D55620: [ThinLTO] Clang changes to utilize new pass to handle chains of aliases

This revision was automatically updated to reflect the committed changes. Closed by commit rL350424: [ThinLTO] Clang changes to utilize new pass to handle chains of aliases (authored by tejohnson, committed by ). Herald added a subscriber: llvm-commits. Repository: rL LLVM CHANGES SINCE LAST A

[PATCH] D53891: [LTO] Add option to enable LTOUnit splitting, and disable unless needed

tejohnson added a comment. ping - @pcc do you have any more comments or can this be committed along with D53890 ? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D53891/new/ https://reviews.llvm.org/D53891 ___

[PATCH] D67385: Pass -mcmodel to LTO plugin

tejohnson added a comment. Typically -mcmodel is passed to the clang compile invocation which sets a module flag in the IR, which is then used by LTO (see calls to Module::setCodeModel() and Module::getCodeModel()). Why is it necessary to pass through the mcmodel passed to the link invocation?

[PATCH] D67385: Pass -mcmodel to LTO plugin

tejohnson added a comment. In D67385#1679127 , @khchen wrote: > @tejohnson when I run the `clang -flto -Wl,-plugin-opt=-help` command, it > shows the > > --code-model= - Choose code model > =tiny

[PATCH] D61634: [clang/llvm] Allow efficient implementation of libc's memory functions in C/C++

tejohnson added a comment. In D61634#1635595 , @tejohnson wrote: > I had some time to work on this finally late last week. I decided the most > straightforward thing was to implement the necessary interface changes to the > TLI analysis to make it requir

[PATCH] D67385: Pass -mcmodel to LTO plugin

tejohnson added a comment. In D67385#1680942 , @khchen wrote: > @tejohnson for example: > > $ clang -flto a.c -c -o a.o > $ llvm-ar q a.a a.o // archive libraries > $ clang -flto a.a b.c -O2 -o main -mcmodel=small > > > In above case

[PATCH] D68029: [ThinLTO] Enable index-only WPD from clang

tejohnson created this revision. tejohnson added a reviewer: pcc. Herald added subscribers: arphaman, dexonsmith, steven_wu, inglorion, mehdi_amini. Herald added a project: clang. To trigger the index-only Whole Program Devirt support added to LLVM, we need to be able to specify -fno-split-lto-un

[PATCH] D68028: [clang] Add no_builtin attribute

tejohnson added a comment. In D68028#1685912 , @gchatelet wrote: > @tejohnson I believe this is the missing part for D67923 > . Thanks, yep I will take a closer look at the patch today. > I'm unsure if we still need the

[PATCH] D67592: [Clang] Use -main-file-name for source filename if not set

tejohnson added a comment. Please upload patches with context (https://llvm.org/docs/Phabricator.html#requesting-a-review-via-the-web-interface) I'm not terribly familiar with this part of clang, but my concern would be that overriding the PresumedInputFile here could have unintended side effec

[PATCH] D67592: [Clang] Use -main-file-name for source filename if not set

tejohnson accepted this revision. tejohnson added a comment. This revision is now accepted and ready to land. lgtm CHANGES SINCE LAST ACTION https://reviews.llvm.org/D67592/new/ https://reviews.llvm.org/D67592 ___ cfe-commits mailing list cfe-com

[PATCH] D67592: [Clang] Use -main-file-name for source filename if not set

tejohnson added a comment. In D67592#1686898 , @the_jk wrote: > I don't have commit access so I'd need someone to push this, thanks. I can commit for you today. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D67592/new/ https://reviews.llvm.org/

[PATCH] D67592: [Clang] Use -main-file-name for source filename if not set

This revision was automatically updated to reflect the committed changes. Closed by commit rL373217: [Clang] Use -main-file-name for source filename if not set (authored by tejohnson, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to commit:

[PATCH] D67592: [Clang] Use -main-file-name for source filename if not set

tejohnson added a comment. In D67592#1688413 , @thakis wrote: > This is failing on my mac like so: This should be fixed by r373237. I didn't get any email with buildbot failures, but someone mailed one to me manually. Not sure why it only fails sometime

[PATCH] D68029: [ThinLTO] Enable index-only WPD from clang

tejohnson added a comment. Herald added a subscriber: hiraditya. Ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D68029/new/ https://reviews.llvm.org/D68029 ___ cfe-commits mailing list cfe-commits@lists.llvm.org h

[PATCH] D68029: [ThinLTO] Enable index-only WPD from clang

This revision was automatically updated to reflect the committed changes. Closed by commit rL373370: [ThinLTO] Enable index-only WPD from clang (authored by tejohnson, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Repository: rL LLVM CHANGES SINCE LAST

[PATCH] D45217: [ThinLTO] Pass -save-temps to LTO backend for distributed ThinLTO builds

tejohnson added a comment. ping Repository: rC Clang https://reviews.llvm.org/D45217 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D45217: [ThinLTO] Pass -save-temps to LTO backend for distributed ThinLTO builds

tejohnson marked 2 inline comments as done. tejohnson added inline comments. Comment at: include/clang/Driver/Options.td:2255 HelpText<"Save intermediate compilation results.">; -def save_temps : Flag<["-", "--"], "save-temps">, Flags<[DriverOption]>, +def save_temps : Flag<["

  1   2   3   4   5   6   7   >