Re: [PATCH] perf: introduce new two options for objdump.

From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: "Martin Liška" <mliska@suse.cz>
Cc: David Ahern <dsahern@gmail.com>,
	a.p.zijlstra@chello.nl, paulus@samba.org, mingo@redhat.com,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] perf: introduce new two options for objdump.
Date: Mon, 1 Jun 2015 09:54:32 -0300	[thread overview]
Message-ID: <20150601125432.GE624@kernel.org> (raw)
In-Reply-To: <556C3D8E.2060506@suse.cz>

Em Mon, Jun 01, 2015 at 01:10:06PM +0200, Martin Liška escreveu:
> On 05/20/2015 03:05 PM, Arnaldo Carvalho de Melo wrote:
> > Em Wed, May 20, 2015 at 03:02:01PM +0200, Martin Liška escreveu:
> >> On 05/19/2015 04:33 PM, Arnaldo Carvalho de Melo wrote:
> >>> Em Tue, May 19, 2015 at 08:29:58AM -0600, David Ahern escreveu:

<SNIP>

> >>>> Perhaps it should be generalized a bit? Right now we have symfs for the
> >>>> location of the tree for binaries. What about a srcfs for a source tree?
> >>>
> >>> That is a good idea indeed.
> >>
> >> I was motivated to be able to annotate source code, realized out-of-box. More precisely, on a machine A
> >> I run perf record, I moved the report file to machine B and I use these options to annotate source code.

<SNIP>

> > I guess so, and yeah, the feature implemented in your patch has value,
> > thanks for collaborating! It is just that it may fit better as a
> > complement to the --symfs stuff, that is also for off-box analysis:

> > --symfs=<directory>::
> >         Look for files with symbols relative to this directory.

> > Thanks, will wait for further discussion before applying,

> Are we going to wait any longer? It looks there's no discussion effort in this thread ;)

I was waiting for you to follow thru on David's suggestion.

- Arnaldo