[llvm-testresults] Broken buildbot: llvm-gcc-i386-linux-selfhost

Evan Cheng evan.cheng at apple.com
Fri Sep 24 10:42:44 PDT 2010


On Sep 24, 2010, at 10:36 AM, Jakob Stoklund Olesen wrote:

> 
> On Sep 24, 2010, at 10:29 AM, Rafael Espindola wrote:
> 
>> On 24 September 2010 13:19, Jakob Stoklund Olesen <stoklund at 2pi.dk> wrote:
>>> Hi,
>>> 
>>> The llvm-gcc-i386-linux-selfhost is currently broken in test.llvm.stage2, it looks like FileCheck is being miscompiled.
>>> 
>>> There has been some noise because of other failures, so the commit breaking this bot wasn't caught.
>>> 
>>> This is the last good build:
>>> 
>>> http://google1.osuosl.org:8011/builders/llvm-gcc-i386-linux-selfhost/builds/45 (r114628)
>>> 
>>> This is the first build to fail in the way it is currently failing:
>>> 
>>> http://google1.osuosl.org:8011/builders/llvm-gcc-i386-linux-selfhost/builds/55 (r114684)
>>> 
>>> 
>>> I suspect Rafael is the culprit since this is only happening on Linux, but Evan is a candidate as well.
>> 
>> So, my changes were to the ELF writer that is not used by default...
> 
> In that case, the failure almost certainly belongs to Evan.
> 
> Evan, could r114630 be triggering the same downstream bug that r114634 apparently did?

Anything is possible. But are we sure that's the case? The failures are restricted to a few llvm-gcc tests and on this tester only. I find that highly suspicious.

Evan

> 
> /jakob
> 
> 
>> 
>>> 
>>> These are the possible offending commits:
>>> 
>>> commit de4ed4eaaf8bfae847502dffb3d72f2578353d75
>>> Author: Benjamin Kramer <benny.kra at googlemail.com>
>>> Date:   Thu Sep 23 11:57:26 2010
>>> 
>>>   Unbreak build. Jim, please review.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114684 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Target/ARM/ARMAsmPrinter.cpp |    8 ++++----
>>> 1 files changed, 4 insertions(+), 4 deletions(-)
>>> 
>>> commit 5e291de8ea80c858d2ee0285bd9cdfc422432e05
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Thu Sep 23 11:32:19 2010
>>> 
>>>   Revert 114634 for now since buildbot claim it broke Clang self-hosting. I doubt it but it's possible it's exposing another bug somewhere.
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114681 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp |   10 ++++++----
>>> 1 files changed, 6 insertions(+), 4 deletions(-)
>>> 
>>> commit 75c8d1e486f4b99a3a27bebbaf2f5af2b2b6f6e2
>>> Author: Jim Grosbach <grosbach at apple.com>
>>> Date:   Thu Sep 23 11:07:04 2010
>>> 
>>>   trailing whitespace
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114680 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> include/llvm/MC/MCAsmInfo.h |   56 +++++++++++++++++++++---------------------
>>> 1 files changed, 28 insertions(+), 28 deletions(-)
>>> 
>>> commit 73492f1b048c89fba56b712a902108ed81929c4a
>>> Author: Jim Grosbach <grosbach at apple.com>
>>> Date:   Thu Sep 23 11:05:37 2010
>>> 
>>>   Clean up the 'trap' instruction printing a bit. Non-Darwin assemblers don't
>>>   (yet) recognize the 'trap' mnemonic, so we use .short/.long to emit the
>>>   opcode directly. On Darwin, however, we do want the mnemonic for more
>>>   readable assembly code and better disassembly.
>>> 
>>>   Adjust the .td file to use the 'trap' mnemonic and handle using the binutils
>>>   workaround in the assembly printer. Also tweak the formatting of the opcode
>>>   values to make them consistent between the MC printer and the old printer.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114679 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Target/ARM/ARMAsmPrinter.cpp |   34 ++++++++++++++++++++++++++++++++++
>>> lib/Target/ARM/ARMInstrInfo.td   |    4 +---
>>> lib/Target/ARM/ARMInstrThumb.td  |    4 +---
>>> 3 files changed, 36 insertions(+), 6 deletions(-)
>>> 
>>> commit 2d882eee1e9077fecad73401fbf553e3f31569ba
>>> Author: Rafael Espindola <rafael.espindola at gmail.com>
>>> Date:   Thu Sep 23 11:01:31 2010
>>> 
>>>   Correctly compute the offset of the symbol. Forgot these bits from the
>>>   last commit.
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114678 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/MC/ELFObjectWriter.cpp |    5 ++++-
>>> 1 files changed, 4 insertions(+), 1 deletions(-)
>>> 
>>> commit 0d8df808cdf2be57d31824177a15da37f3e6b10a
>>> Author: Jim Grosbach <grosbach at apple.com>
>>> Date:   Thu Sep 23 10:58:00 2010
>>> 
>>>   nuke unused var
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114676 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Target/ARM/AsmPrinter/ARMInstPrinter.h |    1 -
>>> 1 files changed, 0 insertions(+), 1 deletions(-)
>>> 
>>> commit 2742d6a114cc1054726d792475b24875602045d0
>>> Author: Chris Lattner <sabre at nondot.org>
>>> Date:   Thu Sep 23 10:37:11 2010
>>> 
>>>   rewrite the copyright section to match reality: llvm does't have a
>>>   copyright assignment process.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114673 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> docs/DeveloperPolicy.html |   32 ++++++++++++++++++--------------
>>> 1 files changed, 18 insertions(+), 14 deletions(-)
>>> 
>>> commit dc8482873b01c2ebdddb0d6dc291b76699dd6b82
>>> Author: Chris Lattner <sabre at nondot.org>
>>> Date:   Thu Sep 23 10:27:54 2010
>>> 
>>>   add lldb and libc++ code owners.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114669 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> docs/DeveloperPolicy.html |    6 +++++-
>>> 1 files changed, 5 insertions(+), 1 deletions(-)
>>> 
>>> commit 45e8cea79f2dc14db8745a97a23e320f9479ecc8
>>> Author: Rafael Espindola <rafael.espindola at gmail.com>
>>> Date:   Thu Sep 23 10:25:18 2010
>>> 
>>>   Represent relocations against local symbols as relocations against the section
>>>   they are in. Both ways should be equivalent, but gas produces relocations
>>>   against the section.
>>> 
>>>   Roman wrote the patch, I added the test.
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114667 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/MC/ELFObjectWriter.cpp |    2 +-
>>> test/MC/ELF/local-reloc.s  |   32 ++++++++++++++++++++++++++++++++
>>> 2 files changed, 33 insertions(+), 1 deletions(-)
>>> 
>>> commit 4c12f7c23aa503cc597fab4b29902d4cfbc84b23
>>> Author: Bob Wilson <bob.wilson at apple.com>
>>> Date:   Thu Sep 23 10:25:06 2010
>>> 
>>>   Fix llvm-extract so that it changes the linkage of all GlobalValues to
>>>   "external" even when doing lazy bitcode loading.  This was broken because
>>>   a function that is not materialized fails the !isDeclaration() test.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114666 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Transforms/IPO/ExtractGV.cpp |   30 ++++++++++++++----------------
>>> test/Other/extract.ll            |    7 ++++++-
>>> 2 files changed, 20 insertions(+), 17 deletions(-)
>>> 
>>> commit f5c3617ad7cf284b7c9d65f8a350f91a51f05593
>>> Author: Oscar Fuentes <ofv at wanadoo.es>
>>> Date:   Thu Sep 23 09:59:44 2010
>>> 
>>>   Avoid warnings about conversions to `bool' in MS compilers.
>>> 
>>>   Patch by Nathan Jeffords!
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114662 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> include/llvm/Attributes.h |    2 +-
>>> include/llvm/Operator.h   |    2 +-
>>> 2 files changed, 2 insertions(+), 2 deletions(-)
>>> 
>>> commit f0e9075c4e46dde7fedeaf68e60f719cd43c04d0
>>> Author: Oscar Fuentes <ofv at wanadoo.es>
>>> Date:   Thu Sep 23 09:59:36 2010
>>> 
>>>   Fix VS 2010 build.
>>> 
>>>   Patch by Nathan Jeffords!
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114661 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp |    8 ++++----
>>> 1 files changed, 4 insertions(+), 4 deletions(-)
>>> 
>>> commit 51e88552f74c337c6203e431fcdcb794347465ef
>>> Author: Nate Begeman <natebegeman at mac.com>
>>> Date:   Thu Sep 23 09:49:17 2010
>>> 
>>>   Revert r114596, it's breaking a few tests.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114659 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> utils/TableGen/NeonEmitter.cpp |   27 +++++----------------------
>>> 1 files changed, 5 insertions(+), 22 deletions(-)
>>> 
>>> commit 5e7cbf2cb28956a666b6d425f6178607ca7e7740
>>> Author: Douglas Gregor <dgregor at apple.com>
>>> Date:   Thu Sep 23 07:19:21 2010
>>> 
>>>   Bump LLVM in CMake makefiles to 2.9
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114640 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> CMakeLists.txt |    2 +-
>>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>> 
>>> commit 2da8af24fb7d1881384c14534ef2ffa2ba0ec902
>>> Author: Rafael Espindola <rafael.espindola at gmail.com>
>>> Date:   Thu Sep 23 07:14:56 2010
>>> 
>>>   Fix the FIXME.
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114639 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/MC/ELFObjectWriter.cpp |    2 +-
>>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>> 
>>> commit 45d67090cf9a397bcfef88e3bf70273ff0b764fd
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Thu Sep 23 00:18:35 2010
>>> 
>>>   Follow up to r114630. Do not optimize away unconditional branch following a conditional one.
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114634 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp |   10 ++++------
>>> 1 files changed, 4 insertions(+), 6 deletions(-)
>>> 
>>> commit dc90cc9c94ab4197d7e56fa31808c9a3fcf2c59b
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Wed Sep 22 23:55:34 2010
>>> 
>>>   Disable codegen prepare critical edge splitting. Machine instruction passes now
>>>   break critical edges on demand.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114633 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Transforms/Scalar/CodeGenPrepare.cpp       |    6 +++++-
>>> test/CodeGen/ARM/lsr-on-unrolled-loops.ll      |    6 +++---
>>> test/CodeGen/Thumb/2010-07-15-debugOrdering.ll |    2 +-
>>> test/CodeGen/X86/2008-02-18-TailMergingBug.ll  |    2 +-
>>> test/CodeGen/X86/lsr-reuse.ll                  |    5 ++---
>>> test/CodeGen/X86/phi-immediate-factoring.ll    |    2 +-
>>> test/CodeGen/X86/phys_subreg_coalesce-2.ll     |    2 +-
>>> test/CodeGen/X86/postra-licm.ll                |    2 +-
>>> test/CodeGen/X86/pr2659.ll                     |    7 ++++---
>>> test/CodeGen/X86/tail-opts.ll                  |   15 ++++++---------
>>> test/CodeGen/X86/tailcallfp2.ll                |    4 +++-
>>> 11 files changed, 28 insertions(+), 25 deletions(-)
>>> 
>>> commit a892ef1e568ae5c2cb35b06f7f4ca59d4d86cdc4
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Wed Sep 22 23:54:40 2010
>>> 
>>>   If there are multiple unconditional branches terminating a block, eliminate all
>>>   but the first one. Those will never be executed. There was logic to do this
>>>   but it was faulty.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114632 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/Target/ARM/ARMBaseInstrInfo.cpp |   18 +++++++++++++++++-
>>> 1 files changed, 17 insertions(+), 1 deletions(-)
>>> 
>>> commit 0454f4e2a638d6718b8fbb6a3182a5ff46fad33d
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Wed Sep 22 23:53:00 2010
>>> 
>>>   Don't sink insert_subreg, subreg_to_reg, reg_sequence. They are meant to be
>>>   close to their sources to facilitate coalescing.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114631 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/CodeGen/MachineSink.cpp |   11 ++++++++++-
>>> 1 files changed, 10 insertions(+), 1 deletions(-)
>>> 
>>> commit 9c8dc77c7ecf3b458bae08502564256cd91ea5e8
>>> Author: Evan Cheng <evan.cheng at apple.com>
>>> Date:   Wed Sep 22 23:51:55 2010
>>> 
>>>   SDISel should not optimize a unconditional branch following a conditional branch
>>>   when the unconditional branch destination is the fallthrough block. The
>>>   canonicalization makes it easier to allow optimizations on DAGs to invert
>>>   conditional branches. The branch folding pass (and AnalyzeBranch) will clean up
>>>   the unnecessary unconditional branches later.
>>> 
>>>   This is one of the patches leading up to disabling codegen prepare critical edge
>>>   splitting.
>>> 
>>> 
>>>   git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114630 91177308-0d34-0410-b5e6-96231b3b80d8
>>> 
>>> lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp |    9 +++++----
>>> 1 files changed, 5 insertions(+), 4 deletions(-)
>>> 
>>> 
>> 
>> 
>> 
>> -- 
>> Rafael Ávila de Espíndola
> 





More information about the llvm-testresults mailing list