Re: perf regression testing. was Re: [PATCH 1/2] perf symbols: Fixinverted logic for showing kallsyms as the source of symbols
From: Arnaldo Carvalho de Melo
Date: Sat Jan 23 2010 - 10:53:19 EST
Em Sat, Jan 23, 2010 at 08:04:51AM +0100, Ingo Molnar escreveu:
> * Arnaldo Carvalho de Melo <acme@xxxxxxxxxxxxx> wrote:
> > Em Fri, Jan 22, 2010 at 05:43:28PM +0100, Mike Galbraith escreveu:
> > > On Fri, 2010-01-22 at 14:35 -0200, Arnaldo Carvalho de Melo wrote:
> > > > From: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> > > >
> > > > Only if we parsed /proc/kallsyms (or a copy found in the buildid cache)
> > > > we should set the dso long name to "[kernel.kallsyms]".
> > >
> > > Yup, works just as well as my busted one :)
> > yeah :-)
> > I'll soon start working on 'perf regtest', to start accumulating regression
> > tests a la 'perf bench', as these silly bugs are worrying me...
> Excellent, it's much needed!
> The best example i know about is KVM: they have built an impressive regression
> test suite and add testcases for new regressions all the time.
> The key is to have something simple to run, which developers (and
> maintainers) can run after every new perf commit or so.
Yes, that is the idea, I'm finishing the first test (/proc/kallsyms
versus what we get from a auto-located vmlinux symtab) and will submit
[acme@doppio linux-2.6-tip]$ perf regtest
1: vmlinux symtab matches kallsyms...
Not quite working yet, tho :o)
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/