[llvm] r273373 - llvm-ar: be more clever about default format
Saleem Abdulrasool via llvm-commits
llvm-commits at lists.llvm.org
Wed Jun 22 21:33:19 PDT 2016
On Wed, Jun 22, 2016 at 8:27 AM, Mike Aizatsky <aizatsky at google.com> wrote:
> Saleem,
>
> This breaks the build:
>
> http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/13963
>
> [5/5] Running the LLVM regression tests
> -- Testing: 17077 tests, 32 threads --
> Testing: 0 .. 10.. 20.. 30.. 40.. 50.. 60.. 70..
> FAIL: LLVM :: Object/mri1.test (13336 of 17077)
> ******************** TEST 'LLVM :: Object/mri1.test' FAILED ********************
> Script:
> --
> echo create /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.a > /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.mri
> echo save >> /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.mri
> echo end >> /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.mri
> /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/./bin/llvm-ar -M < /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.mri
> /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/./bin/llvm-ar t /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/test/Object/Output/mri1.test.tmp.a
> --
> Exit Code: 77
>
> Command Output (stderr):
> --
> ==21929==WARNING: MemorySanitizer: use-of-uninitialized-value
> #0 0xb694ba in llvm::writeArchive(llvm::StringRef, std::__1::vector<llvm::NewArchiveIterator, std::__1::allocator<llvm::NewArchiveIterator> >&, bool, llvm::object::Archive::Kind, bool, bool, std::__1::unique_ptr<llvm::MemoryBuffer, std::__1::default_delete<llvm::MemoryBuffer> >) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/lib/Object/ArchiveWriter.cpp:376:7
> #1 0x4dce49 in performWriteOperation(ArchiveOperation, llvm::object::Archive*, std::__1::unique_ptr<llvm::MemoryBuffer, std::__1::default_delete<llvm::MemoryBuffer> >, std::__1::vector<llvm::NewArchiveIterator, std::__1::allocator<llvm::NewArchiveIterator> >*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:670:7
> #2 0x4d6745 in performOperation(ArchiveOperation, llvm::object::Archive*, std::__1::unique_ptr<llvm::MemoryBuffer, std::__1::default_delete<llvm::MemoryBuffer> >, std::__1::vector<llvm::NewArchiveIterator, std::__1::allocator<llvm::NewArchiveIterator> >*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:703:5
> #3 0x4d2ecc in performOperation(ArchiveOperation, std::__1::vector<llvm::NewArchiveIterator, std::__1::allocator<llvm::NewArchiveIterator> >*) /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:726:5
> #4 0x4cc912 in runMRIScript /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:813:5
> #5 0x4cc912 in parseCommandLine /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:197
> #6 0x4cc912 in ar_main /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:820
> #7 0x4cc912 in main /mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/llvm-ar/llvm-ar.cpp:857
> #8 0x7f7bb7db0f44 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x21f44)
> #9 0x44ab5b in _start (/mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm_build_msan/bin/llvm-ar+0x44ab5b)
>
>
Thanks.
I tried to reproduce this locally, but all I seem to get indications of
issues in parseCommandLine and Triple construction. I even tried without
my changes locally and it still seems to trigger MSAN warnings. Also, it
seems that a build of MSAN enabled llvm-tblgen in debug mode also fails.
The positions that the symbolication indicates is also very weird:
lib/Object/ArchiveWriter.cpp:376
if (Kind != object::Archive::Kind)
The column indicated is actually Kind, which is an integral argument
(its an enumeration). The value is definitely initialised in all the
codepaths. Im not sure what to make of this warning.
>
> On Wed, Jun 22, 2016 at 6:45 AM Rafael Espíndola <
> llvm-commits at lists.llvm.org> wrote:
>
>> testcase? :-)
>>
>> On 22 June 2016 at 00:03, Saleem Abdulrasool via llvm-commits
>> <llvm-commits at lists.llvm.org> wrote:
>> > Author: compnerd
>> > Date: Tue Jun 21 23:03:28 2016
>> > New Revision: 273373
>> >
>> > URL: http://llvm.org/viewvc/llvm-project?rev=273373&view=rev
>> > Log:
>> > llvm-ar: be more clever about default format
>> >
>> > Try to be more clever about selecting the default format. When an
>> existing
>> > archive is used, use the type of the archive to determine the format.
>> When
>> > existing members are present, use the first member's format to
>> determine the
>> > format to use. If we are creating an empty archive (MRI mode) or are
>> adding
>> > non-object members, default to the current behaviour of using the host
>> type due
>> > to the lack of a better alternative. This aids in cross-compilation on
>> Darwin
>> > to non-Darwin platforms which rely on GNU format archives.
>> >
>> > Modified:
>> > llvm/trunk/tools/llvm-ar/llvm-ar.cpp
>> >
>> > Modified: llvm/trunk/tools/llvm-ar/llvm-ar.cpp
>> > URL:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/tools/llvm-ar/llvm-ar.cpp?rev=273373&r1=273372&r2=273373&view=diff
>> >
>> ==============================================================================
>> > --- llvm/trunk/tools/llvm-ar/llvm-ar.cpp (original)
>> > +++ llvm/trunk/tools/llvm-ar/llvm-ar.cpp Tue Jun 21 23:03:28 2016
>> > @@ -19,6 +19,7 @@
>> > #include "llvm/LibDriver/LibDriver.h"
>> > #include "llvm/Object/Archive.h"
>> > #include "llvm/Object/ArchiveWriter.h"
>> > +#include "llvm/Object/MachO.h"
>> > #include "llvm/Object/ObjectFile.h"
>> > #include "llvm/Support/CommandLine.h"
>> > #include "llvm/Support/Errc.h"
>> > @@ -575,21 +576,86 @@ computeNewArchiveMembers(ArchiveOperatio
>> > return Ret;
>> > }
>> >
>> > +static object::Archive::Kind getDefaultForHost() {
>> > + return Triple(sys::getProcessTriple()).isOSDarwin() ?
>> object::Archive::K_BSD
>> > + :
>> object::Archive::K_GNU;
>> > +}
>> > +
>> > +static object::Archive::Kind
>> > +getKindFromMember(const NewArchiveIterator &Member) {
>> > + if (Member.isNewMember()) {
>> > + object::Archive::Kind Kind = getDefaultForHost();
>> > +
>> > + sys::fs::file_status Status;
>> > + if (auto OptionalFD = Member.getFD(Status)) {
>> > + auto OptionalMB = MemoryBuffer::getOpenFile(*OptionalFD,
>> Member.getName(),
>> > + Status.getSize(),
>> false);
>> > + if (OptionalMB) {
>> > + MemoryBufferRef MemoryBuffer =
>> (*OptionalMB)->getMemBufferRef();
>> > +
>> > + Expected<std::unique_ptr<object::ObjectFile>> OptionalObject =
>> > + object::ObjectFile::createObjectFile(MemoryBuffer);
>> > +
>> > + if (OptionalObject)
>> > + Kind = isa<object::MachOObjectFile>(**OptionalObject)
>> > + ? object::Archive::K_BSD
>> > + : object::Archive::K_GNU;
>> > +
>> > + // squelch the error in case we had a non-object file
>> > + consumeError(OptionalObject.takeError());
>> > + }
>> > +
>> > + if (close(*OptionalFD) != 0)
>> > + failIfError(std::error_code(errno, std::generic_category()),
>> > + "failed to close file");
>> > + }
>> > +
>> > + return Kind;
>> > + } else {
>> > + const object::Archive::Child &OldMember = Member.getOld();
>> > + if (OldMember.getParent()->isThin())
>> > + return object::Archive::Kind::K_GNU;
>> > +
>> > + auto OptionalMB = OldMember.getMemoryBufferRef();
>> > + failIfError(OptionalMB.getError());
>> > +
>> > + Expected<std::unique_ptr<object::ObjectFile>> OptionalObject =
>> > + object::ObjectFile::createObjectFile(*OptionalMB);
>> > +
>> > + if (OptionalObject)
>> > + return isa<object::MachOObjectFile>(*OptionalObject->get())
>> > + ? object::Archive::K_BSD
>> > + : object::Archive::K_GNU;
>> > +
>> > + // squelch the error in case we had a non-object file
>> > + consumeError(OptionalObject.takeError());
>> > + return getDefaultForHost();
>> > + }
>> > +}
>> > +
>> > static void
>> > performWriteOperation(ArchiveOperation Operation,
>> > object::Archive *OldArchive,
>> > std::unique_ptr<MemoryBuffer> OldArchiveBuf,
>> > std::vector<NewArchiveIterator> *NewMembersP) {
>> > + std::vector<NewArchiveIterator> NewMembers;
>> > + if (!NewMembersP)
>> > + NewMembers = computeNewArchiveMembers(Operation, OldArchive);
>> > +
>> > object::Archive::Kind Kind;
>> > switch (FormatOpt) {
>> > - case Default: {
>> > - Triple T(sys::getProcessTriple());
>> > - if (T.isOSDarwin() && !Thin)
>> > - Kind = object::Archive::K_BSD;
>> > - else
>> > + case Default:
>> > + if (Thin)
>> > Kind = object::Archive::K_GNU;
>> > + else if (OldArchive)
>> > + Kind = OldArchive->kind();
>> > + else if (NewMembersP)
>> > + Kind = NewMembersP->size() ?
>> getKindFromMember(NewMembersP->front())
>> > + : getDefaultForHost();
>> > + else
>> > + Kind = NewMembers.size() ? getKindFromMember(NewMembers.front())
>> > + : getDefaultForHost();
>> > break;
>> > - }
>> > case GNU:
>> > Kind = object::Archive::K_GNU;
>> > break;
>> > @@ -599,18 +665,10 @@ performWriteOperation(ArchiveOperation O
>> > Kind = object::Archive::K_BSD;
>> > break;
>> > }
>> > - if (NewMembersP) {
>> > - std::pair<StringRef, std::error_code> Result = writeArchive(
>> > - ArchiveName, *NewMembersP, Symtab, Kind, Deterministic, Thin,
>> > - std::move(OldArchiveBuf));
>> > - failIfError(Result.second, Result.first);
>> > - return;
>> > - }
>> > - std::vector<NewArchiveIterator> NewMembers =
>> > - computeNewArchiveMembers(Operation, OldArchive);
>> > - auto Result =
>> > - writeArchive(ArchiveName, NewMembers, Symtab, Kind,
>> Deterministic, Thin,
>> > - std::move(OldArchiveBuf));
>> > +
>> > + std::pair<StringRef, std::error_code> Result =
>> > + writeArchive(ArchiveName, NewMembersP ? *NewMembersP :
>> NewMembers, Symtab,
>> > + Kind, Deterministic, Thin,
>> std::move(OldArchiveBuf));
>> > failIfError(Result.second, Result.first);
>> > }
>> >
>> >
>> >
>> > _______________________________________________
>> > llvm-commits mailing list
>> > llvm-commits at lists.llvm.org
>> > http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>
> --
> Mike
> Sent from phone
>
--
Saleem Abdulrasool
compnerd (at) compnerd (dot) org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160622/a9192875/attachment.html>
More information about the llvm-commits
mailing list