Re: [PATCH v2] perf report: Add wall-clock and parallelism profiling

[Arnaldo Carvalho de Melo <acme@kernel.org>]

On Tue, Jan 14, 2025 at 02:53:03PM -0300, Arnaldo Carvalho de Melo wrote:
> On Tue, Jan 14, 2025 at 05:07:14PM +0100, Dmitry Vyukov wrote:
> > I agree it would be nice to have these usability improvements.
 
> Right, and I'm not talking about those as a requirement for your patch
> to get accepted.
 
> > But this is also a large change, so I would prefer to merge the basic
> > functionality, and then we can do some improvements on top. This is
> > also the first time I am touching perf code, there are too many new
> > things to learn for me.
 
> I'm happy you made the effort and will try to help you as much as I can.
 
> I agree that, with the patch below, that I'll try to split and merge
> before travelling, we'll have a good starting point.
> 
> Here is the patch, I'll split the perf_config__set_variable into a
> separate patch and merge, so that you can continue from there. The
> "<LONGER EXPLANATION>" part should tell the user that 'perf config
> report.workload=yes|no' can be done later, to switch the mode, if one
> doesn't want to continue with it as default.

Its now in the tmp.perf-tools-next branch at:

https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git

I'm testing it together with a batch of patches one of which I had to
bisect and remove, soon will be in the main perf-tools-next branch.

I mean this patch:

"perf config: Add a function to set one variable in .perfconfig"
https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/commit/?h=tmp.perf-tools-next

Thanks,

- Arnaldo