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

[David Ahern <dsahern@gmail.com>]

On 2/11/15 9:15 AM, Peter Zijlstra wrote:
> On Wed, Feb 11, 2015 at 08:18:37AM -0500, Stephane Eranian wrote:
>
>> But, we need a portable solution, there are jitted environment on other
>> architectures.
>
> Yeah, I'm aware of that. But the time people are not liking any of those
> patches.
>
> The thing that seems to have most traction is something like
>
>    lkml.kernel.org/r/20150211161256.GH2896@worktop.programming.kicks-ass.net
>
> Which would basically make the trace clock be CLOCK_MONOTONIC.
>

Sure, but IF that solution is ever adopted it will apply to top of tree 
kernels. In the meantime users of long term kernels (enterprise, LTS, 
other) need to be able to use this feature.

I understand that Ingo (and probably you as well) have a disdain for the 
kernel module approach mentioned above and it is not the preferred 
approach. But given the resistance for over *4 years* to solutions to 
this problem the module does provide a solution for the existing user base.

David