The perf tool segmentation fault which we were observing and this
patch series intended to fix it, has been fixed due to change in
default behaviour of perf stat[1][2].
[1]
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v4.12&id=0d79f8b93187c771b6971acfaba67f4e2f1e07
ping?
On Tue, May 23, 2017 at 8:44 PM, Ganapatrao Kulkarni
wrote:
> On Tue, May 23, 2017 at 3:34 PM, Mark Rutland wrote:
>> Hi,
>>
>> Does anyone have any comments on these?
>>
>> I'm happy to rebase/resend if necessary.
>>
>> I'd very much like to see this fixed.
>
> arm64 platforms have depend
On Tue, May 23, 2017 at 3:34 PM, Mark Rutland wrote:
> Hi,
>
> Does anyone have any comments on these?
>
> I'm happy to rebase/resend if necessary.
>
> I'd very much like to see this fixed.
arm64 platforms have dependency on this patch set to use latest perf tool.
Earliest merge/review is much ap
Hi,
Does anyone have any comments on these?
I'm happy to rebase/resend if necessary.
I'd very much like to see this fixed.
Thanks,
Mark.
On Fri, Apr 28, 2017 at 12:47:08PM +0100, Mark Rutland wrote:
> Hi,
>
> These patches fix a segfault seen in some cases when perf stat is exiting.
>
> We d
Hi,
These patches fix a segfault seen in some cases when perf stat is exiting.
We don't balance opening/closing of events in all cases, and go out-of-bounds
when we close events. Full details in the patch 2 commit message.
Thanks,
Mark.
Mark Rutland (2):
perf evsel: add per{cpu,thread} close
5 matches
Mail list logo