Re: [PATCH 0/4] perf: add support for profiling jitted code

[Stephane Eranian]

Peter,

On Wed, Feb 11, 2015 at 6:39 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> On Wed, Feb 11, 2015 at 12:42:41AM +0100, Stephane Eranian wrote:
>> To enable synchronization of the runtime MMAPs with those recorded by
>> the kernel on behalf of the perf tool, the runtime needs to timestamp
>> any record in the dump file using the same time source. The current
>> patch series is relying on Sonny Rao's posix-clock patch series
>> posted on LKML in 2013. The link to the patches is:
>>       https://lkml.org/lkml/2013/12/6/1044
>>
>
> At least for x86 you could use something like this:
>
>   lkml.kernel.org/r/aa2dd2e4f1e9f2225758be5ba00f14d6909a8ce1.1423180257.git.shli@xxxxxx
>
> We can re-create the perf_clock() from the tsc with the mult, shift and
> offset provided in the perf userpage.

I had forgotten that I had modified Sonny's patch to use sched_clock(). I will
post V2 using David Ahern's driver instead.

But, we need a portable solution, there are jitted environment on other
architectures.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/