[llvm-commits] [llvm] r81220 - in /llvm/trunk: lib/Target/ARM/ARMBaseInstrInfo.cpp lib/Target/ARM/ARMInstrNEON.td lib/Target/ARM/ARMRegisterInfo.td test/CodeGen/ARM/vmlal_lane.ll test/CodeGen/ARM/vmlsl_lane.ll test/CodeGen/ARM/vmul_lane.ll test/CodeGen/ARM/vmull_lane.ll test/CodeGen/ARM/vqRdmulh_lane.ll test/CodeGen/ARM/vqdmlal_lanes.ll test/CodeGen/ARM/vqdmlsl_lanes.ll test/CodeGen/ARM/vqdmulh_lane.ll test/CodeGen/ARM/vqdmull_lane.ll

Xerxes RĂ„nby xerxes at zafena.se
Tue Sep 8 12:00:26 PDT 2009


Den 2009-09-08 20:42, Anton Korobeynikov skrev:
> Hello Xerxes
>
>   
>> I belive this changeset caused about 230 CodeGen regression failures on
>> the llvm-arm-linux buildbot
>>     
> Could you please run some of these tests by hand and check how they
> failed? This really looks like miscompilation or something similar.
>
>   

looks like most errors are happens during ARM Assembly Printer pass like:

Running pass 'ARM Assembly Printer' on function '@foo'

the stdio logs are available on the buildbot server for about a week here:
http://google1.osuosl.org:8011/builders/llvm-arm-linux/builds/111/steps/test-llvm/logs/stdio
after that they seems to be removed.

i have attached the stdio of the failing ARM CodeGen tests in this mail.

I can get access to the machine first tomorrow during working hours GMT
to run some gdb sessions.

Cheers
Xerxes




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20090908/f392ff5f/attachment.html>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: llvm-arm-linux-111-make-test-CodeGen-stdio.txt
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20090908/f392ff5f/attachment.txt>


More information about the llvm-commits mailing list