lkml.org 
[lkml]   [2015]   [Jun]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 02/27] perf tools: Add comm string into struct thread_map
On Thu, Jun 25, 2015 at 11:39:27AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Tue, Jun 23, 2015 at 11:56:18PM +0200, Jiri Olsa escreveu:
> > On Tue, Jun 23, 2015 at 06:01:58PM -0300, Arnaldo Carvalho de Melo wrote:
> > > Em Tue, Jun 23, 2015 at 12:36:03AM +0200, Jiri Olsa escreveu:
> > > > Adding support to hold comm name together with pids in
> > > > 'struct thread_map'. It will be useful for --per-thread
> > > > option to display task pid together with task name.
>
> > > > Getting the task name from /proc/$pid/comm.
>
> > > Do we really need to read this all the time? Or just for tools that need
> > > this?
>
> > > Can't we have a thread_map__read_comms(map);
>
> > > That we can make this specific --per-thread option in 'stat' call?
>
> > ook
>
> Thanks, this made me think: thread_map is just an array that enumerates
> the pids ultimately for use with sys_perf_event_open(), bloating it with
> extra details used by just one oddball subfeature of just one tool seems
> overkill.
>
> Probably it would be better to rename it to pid_map, that is what is
> useful for all tools using it to create the event file descriptors and
> have a comm_map that would return another enumeration to obtain the
> comms, something like:
>
> struct thread_map *tids = thread__map_by_foo(parameters);
> struct comm_map *comms = comm_map__new(tids);
>
> That way the various ways of populating the thread_map->map[] (the
> _by_foo() variants of thread_map we have) would do that leg work and
> the comm_map__new(tids) would just read the comms.
>
> The thread_map users would not have to be touched at all, it would be
> kept as simple as it was and just tools that need those comms would go
> to the trouble of allocating the needed space for the array and to do
> the comm reading.
>
> But nah, I'll probably merge what you did and later, if I find the time,
> do it myself, better to have that new feature than to be picky, right?

right ;-) also the 'comm' string is small update after long time
of no change in thread_map and I guess it will take probably
another long time period before we will want to change it again..

thanks,
jirka


\
 
 \ /
  Last update: 2015-06-25 18:21    [W:0.100 / U:0.040 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site