buildbot failure in LLVM on clang-x64-ninja-win7

Fri Oct 23 14:13:19 PDT 2015

May be a bug unrelated to these specific revisions that manifests itself
randomally, I have seen such failures before, usually resulting from use
after free/invalidated pointers.
asan build may catch the issue but not always due to the use of BumpPtr
allocators that do not really free the memory so asan does not know about
it. Running the test under a debugger is another option.

2015-10-23 22:56 GMT+03:00 Sanjoy Das via llvm-commits <
llvm-commits at lists.llvm.org>:

> Okay, the bot is green now (since
> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6445) for
> some reason.  Any idea what happened?
>
> -- Sanjoy
>
>
> Sanjoy Das wrote:
>
>>
>>
>> Fedorova, Vera wrote:
>>
>>> Sanjoy,
>>>
>>> I have rebooted windows7-buildbot machine, deleted start directory and
>>> started full build.
>>> These fails are still present :
>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6429
>>>
>>
>> I have still not been unable to figure out which revision causes the
>> failure. For instance, in
>>
>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6444
>>
>> I built a rather old revision == 251040 which was supposed to be green;
>> but it still fails the two tests.
>>
>> -- Sanjoy
>>
>>
>>
>>> Vera Fedorova
>>> =============
>>> Software Engineer
>>> Intel Compiler Team
>>> Intel
>>>
>>>
>>> -----Original Message-----
>>> From: Sanjoy Das [mailto:sanjoy at playingwithpointers.com]
>>> Sent: Friday, October 23, 2015 05:58
>>> To: Evgenii Stepanov
>>> Cc: Galina Kistanova; Artur Pilipenko via llvm-commits;
>>> mstester.llvm at gmail.com
>>> Subject: Re: buildbot failure in LLVM on clang-x64-ninja-win7
>>>
>>> Galina,
>>>
>>> Can you please take a look to see if I'm setting the 'clobber' flag
>>> correctly in the two builds above? Are there other ways I can force
>>> the bot do a "full" build?
>>>
>>> Generally speaking, I don't think the second stage of a bootstrap bot
>>> should do an incremental build -- that is equivalent to running only a
>>> small percentage of the test suite. :)
>>>
>>> -- Sanjoy
>>>
>>>
>>> On Thu, Oct 22, 2015 at 6:29 PM, Sanjoy
>>> Das<sanjoy at playingwithpointers.com> wrote:
>>>
>>>> I see the same problem with "clobber" builds.
>>>>
>>>> 251048 with clobber fails:
>>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6418
>>>> 251045 with clobber fails:
>>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6420
>>>>
>>>> However, it does not look like that setting clobber to 1 actually
>>>> worked -- for instance, SmallVector.cpp isn't in the build logs for
>>>> 6420 or 6418.
>>>>
>>>> -- Sanjoy
>>>>
>>>> On Thu, Oct 22, 2015 at 5:10 PM, Evgenii Stepanov
>>>> <eugeni.stepanov at gmail.com> wrote:
>>>>
>>>>> You can request a full rebuild by adding "clobber" as "property 1
>>>>> name" and "1" as its value in the force build form, if you suspect
>>>>> that the problem lies there.
>>>>>
>>>>>
>>>>> On Thu, Oct 22, 2015 at 5:06 PM, Sanjoy Das via llvm-commits
>>>>> <llvm-commits at lists.llvm.org> wrote:
>>>>>
>>>>>> It looks like both 251045 and 251048 trigger the same failure:
>>>>>>
>>>>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6416
>>>>>> -- 251045
>>>>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6415
>>>>>> -- 251048
>>>>>>
>>>>>> 251048 is my first change, and 251045 is the change right before.
>>>>>> Also IIRC the buildbot *was* green on 251045. Any idea what's going
>>>>>> on?
>>>>>>
>>>>>> It is possible that the changes I made miscompiled a .o file that
>>>>>> was not re-built in the above two builds. However it was definitely
>>>>>> re-built when going *from* 251045 to 251053 (since otherwise the
>>>>>> miscompile would not have manifested), so the source file in 251045
>>>>>> vs. in 251053 should be different; and that file have been re-built
>>>>>> when coming back to 251045 as well.
>>>>>>
>>>>>> -- Sanjoy
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Thu, Oct 22, 2015 at 4:24 PM, Sanjoy Das
>>>>>> <sanjoy at playingwithpointers.com> wrote:
>>>>>>
>>>>>>> Thanks Galina, this should help me discover the first breaking
>>>>>>> revision quickly.
>>>>>>>
>>>>>>> -- Sanjoy
>>>>>>>
>>>>>>> On Thu, Oct 22, 2015 at 4:17 PM, Galina
>>>>>>> Kistanova<gkistanova at gmail.com> wrote:
>>>>>>>
>>>>>>>> PS: I noticed the `Force build` form, but wasn't sure if I was
>>>>>>>>> okay
>>>>>>>>>
>>>>>>>> using it
>>>>>>>>
>>>>>>>> It's ok. Please fill "Branch to build" and "Revision to build:"
>>>>>>>> fields.
>>>>>>>>
>>>>>>>> Thanks
>>>>>>>>
>>>>>>>> Galina
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Thu, Oct 22, 2015 at 4:10 PM, Sanjoy Das
>>>>>>>> <sanjoy at playingwithpointers.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> I have not been able to reproduce this issue on a Release without
>>>>>>>>> asserts bootstrap on OSX or Linux. I'm trying a Release+Asserts
>>>>>>>>> bootstrap now.
>>>>>>>>>
>>>>>>>>> Since I don't have access to a Windows box, some help on
>>>>>>>>> narrowing down the issue will be much appreciated.
>>>>>>>>>
>>>>>>>>> For instance, is it possible to build and test each of the
>>>>>>>>> revisions between r251045 and r251053? Narrowing down the
>>>>>>>>> failure to one revision will be very helpful.
>>>>>>>>>
>>>>>>>>> PS: I noticed the `Force build` form, but wasn't sure if I was
>>>>>>>>> okay using it.
>>>>>>>>>
>>>>>>>>> -- Sanjoy
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> llvm.buildmaster at lab.llvm.org wrote:
>>>>>>>>>
>>>>>>>>>> The Buildbot has detected a new failure on builder
>>>>>>>>>> clang-x64-ninja-win7 while building llvm.
>>>>>>>>>> Full details are available at:
>>>>>>>>>>
>>>>>>>>>> http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/6
>>>>>>>>>> 404
>>>>>>>>>>
>>>>>>>>>> Buildbot URL: http://lab.llvm.org:8011/
>>>>>>>>>>
>>>>>>>>>> Buildslave for this Build: windows7-buildbot
>>>>>>>>>>
>>>>>>>>>> Build Reason: scheduler
>>>>>>>>>> Build Source Stamp: [branch trunk] 251053
>>>>>>>>>> Blamelist: sanjoy
>>>>>>>>>>
>>>>>>>>>> BUILD FAILED: failed ninja check 2
>>>>>>>>>>
>>>>>>>>>> sincerely,
>>>>>>>>>> -The Buildbot
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Sanjoy Das
>>>>>>> http://playingwithpointers.com
>>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Sanjoy Das
>>>>>> http://playingwithpointers.com
>>>>>> _______________________________________________
>>>>>> llvm-commits mailing list
>>>>>> llvm-commits at lists.llvm.org
>>>>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>>>>>
>>>>>
>>>>
>>>> --
>>>> Sanjoy Das
>>>> http://playingwithpointers.com
>>>>
>>>
>>>
>>>
>>> --
>>> Sanjoy Das
>>> http://playingwithpointers.com
>>>
>>> --------------------------------------------------------------------
>>> Joint Stock Company Intel A/O
>>> Registered legal address: Krylatsky Hills Business Park,
>>> 17 Krylatskaya Str., Bldg 4, Moscow 121614,
>>> Russian Federation
>>>
>>> This e-mail and any attachments may contain confidential material for
>>> the sole use of the intended recipient(s). Any review or distribution
>>> by others is strictly prohibited. If you are not the intended
>>> recipient, please contact the sender and delete all copies.
>>>
>> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151024/f33856e0/attachment.html>