Flaky sanitizer test on ppc64 bot

Tue Mar 7 11:51:20 PST 2017

On Tue, Mar 7, 2017 at 6:41 AM, Bill Seurer <seurer at linux.vnet.ibm.com>
wrote:

> Actually there was a range of revision a bit further on where it worked,
> too.
>
> r279571 (and previous) - works
> r279572 - fails
> . . . fails
> r279642 - fails
> r279643 - works
> . . . works
> r279663 - works
> r279664 - fails
> (and fails from here on)
>
> IIRC this is the same range or revisions that affected some of of the
> other multithread test cases that are flaky on powerpcle and which are
> currently disabled.
>
> I wonder if some of these multithread tests should be run multiple times
> when the bots run them?

No. We simply need to figure out what's broken and fix that.
This will be one of the Aleksey's tasks for the nearest month or two.

>
>
> On 03/06/2017 06:42 PM, Kostya Serebryany wrote:
>
>> Good to know, thanks!
>> Aleksey is going to start digging around this code very soon.
>>
>> On Mon, Mar 6, 2017 at 9:10 AM, Bill Seurer <seurer at linux.vnet.ibm.com
>> <mailto:seurer at linux.vnet.ibm.com>> wrote:
>>
>>     I ran a script that dug back through LOTS of revisions and traced
>>     the start of this test cases failing to r279572.
>>
>>     ------------------------------------------------------------
>> ------------
>>     r279572 | kcc | 2016-08-23 16:19:47 -0500 (Tue, 23 Aug 2016) | 1 line
>>
>>     [sanitizer] change the 64-bit allocator to use a single array for
>>     free-d chunks instead of a lock-free linked list of tranfer batches.
>>     This change simplifies the code, makes th
>>     e allocator more 'hardened', and will allow simpler code to release
>>     RAM to OS. This may also slowdown malloc stress tests due to lock
>>     contension, but I did not observe noticeab
>>     le slowdown on various real multi-threaded benchmarks.
>>
>>
>>     BTW, I had to set the repeat count of
>>     projects/compiler-rt/lib/asan/tests/default/Asan-powerpc64le
>> -with-calls-Test
>>     up to 500 in order to get it to consistently fail.
>>
>>
>>     On 03/01/2017 06:24 PM, Kostya Serebryany wrote:
>>
>>         + more folks
>>
>>         On Wed, Mar 1, 2017 at 4:23 PM, Matthias Braun <matze at braunis.de
>>         <mailto:matze at braunis.de>
>>         <mailto:matze at braunis.de <mailto:matze at braunis.de>>> wrote:
>>
>>             We have another flaky sanitizer test on the ppc64 bot. Just
>>         look at
>>             the recent history and you see
>>             AddressSanitizer-Unit::AddressSanitizer.ManyThreadsTest
>>         failing a
>>             lot on unrelated commits:
>>
>>
>>         http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multis
>> tage?numbuilds=50
>>         <http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multi
>> stage?numbuilds=50>
>>
>>         <http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multi
>> stage?numbuilds=50
>>         <http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multi
>> stage?numbuilds=50>>
>>
>>             Please fix or disable this test.
>>
>>             - Matthias
>>
>>
>>
>>
>>     --
>>
>>     -Bill Seurer
>>
>>
>>
>
> --
>
> -Bill Seurer
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170307/2fedfa36/attachment.html>