Re: [PATCH v6 3/5] tools, perf, script: Add --call-trace and --call-ret-trace
From: leo . yan
Date: Sat Sep 29 2018 - 03:39:29 EST
On Fri, Sep 28, 2018 at 10:19:44AM -0700, Andi Kleen wrote:
> > Seems to me, these two features are _NOT_ only benefit for intel_pt,
> > other hardware tracing (e.g. Arm CoreSight) can enable these features
> > as well. This patch is to document only for intel_pt, later if we
> > enable this feature on Arm platform we need to change the doc;
> > alternatively we can use more general description for these two options
> > at the first place. How about you think for this?
>
> Likely it already works for CoreSight
I think Kim played with this patch series and he also pointed me for
this series.
> I specified intel_pt, because if we just say traces the users won't
> know what PMU to specify for record. Being too abstract is
> often not helpful.
>
> If someone successfully tests it on CoreSight they could submit
> a patch to the documentation to add "or <coresightpmu>" to these
> two cases. That would make it then clear for those users too.
Okay, agree.
Actually I applied your patch series v6 on the perf latest core branch
and tested on Arm Juno board, I observed there have couple issues, one
is CoreSight trace data doesn't support timestamp so I need to use
'-F,-time' to workaround the command failure; another issue is now
CoreSight is absent to set sample flags so perf fails to resolve
symbols [1]; these two issues are only related with CoreSight decoder
and it's no matter with this patch, so I didn't mention in my previous
replying.
I need a bit more time to work out more formal CoreSight fixing patches
and will send for reviewing (also will include one patch to clarifying
Arm Coresight support in doc as suggested).
Thanks,
Leo Yan
[1] https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-script.c?h=perf/core#n1128