Re: [PATCH perf/core v9 00/16] perf-probe --cache and SDT support

From: Masami Hiramatsu
Date: Tue May 31 2016 - 05:01:48 EST


On Mon, 30 May 2016 13:10:20 -0300
Arnaldo Carvalho de Melo <acme@xxxxxxxxxx> wrote:

> Em Mon, May 30, 2016 at 12:54:58PM +0900, Namhyung Kim escreveu:
> > Hi Masami,
> >
> > On Sun, May 29, 2016 at 12:15:01AM +0900, Masami Hiramatsu wrote:
> > > Hi,
> > >
> > > Here is the 9th version of the patchset for probe-cache and
> > > initial SDT support.
> > >
> > > The previous version is here; https://lkml.org/lkml/2016/5/14/107
> > >
> > > This version just fixes 2 points according to Hemant's comment,
> > > and split out the filename__readable() patch.
> > >
> > > Changes in v9:
> > > - [1/16,2/16] split out the filename__readable()
> > > - [7/16] Fix to show which event is deleted.
> > > - [12/16] Update list_usage to show sdt option.
> > >
> > > Thank you,
> >
> > Thanks for doing this, I hope this gets merged soon. The patch 4 and
> > 14 look a bit larger and would be better to be splitted IMHO. Other
> > than that, all look good to me.
> >
> > Acked-by: Namhyung Kim <namhyung@xxxxxxxxxx>
>
> Agreed, please split both patches, I applied 1-3, please resend from 4
> onwards.

Hmm, for patch#4, it includes
A. Rename and Export build_id_cache__cachedir() [14 LOC]
B. Add perf_probe_event__copy() [73 LOC]
C. Add (or fix and reenable) synthesize_perf_probe_point() [38 LOC]
D. Introduce perf_cache interfaces (new/add/commit/delete) [320 LOC]
E. Add --cache option for perf-probe [15 LOC]

And E depends on D, D depends on A,B and C. And of course without E,
other parts are just a deadweight.(no other one call it, just exported)
Should I split them all?

Thank you,

>
> - Arnaldo


--
Masami Hiramatsu <mhiramat@xxxxxxxxxx>