linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Robert Richter <rric@kernel.org>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>, Ingo Molnar <mingo@kernel.org>,
	Arnaldo Carvalho de Melo <acme@infradead.org>,
	Jiri Olsa <jolsa@redhat.com>,
	linux-kernel@vger.kernel.org, Borislav Petkov <bp@suse.de>
Subject: Re: [PATCH v2 03/14] perf: Add persistent event facilities
Date: Tue, 25 Jun 2013 10:41:35 +0200	[thread overview]
Message-ID: <20130625084135.GA21579@rric.localhost> (raw)
In-Reply-To: <20130624094458.GO28407@twins.programming.kicks-ass.net>

On 24.06.13 11:44:58, Peter Zijlstra wrote:
> > +static void del_persistent_event(int cpu, struct perf_event_attr *attr)
> > +{
> > +	struct pers_event_desc *desc, *tmp;
> > +	struct perf_event *event = NULL;
> > +
> > +	list_for_each_entry_safe(desc, tmp, &per_cpu(pers_events, cpu), plist) {
> > +		if (desc->attr->config == attr->config) {
> 
> attr->config might not be enough; you might also want to compare type,
> config1, config2, and possible bp_type as well; although I think we want
> to avoid having hwbp style persistent events, but then what do I know.

[ I would call this what it is: perf_add_persistent_tracepoint(), or
so :-) ]

We chosed a tracepoint-only implementation as a first try and since
this is our use case. The goal is to have support for any event type.

I also was thinking of comparing the whole attr structure. But since
attr are similar (or almost identical) only you can't compare each
member of attr and need exceptions (e.g. cpu number). These exceptions
are ugly and depending on the specific use case.

I rather tend to assign unique identifiers to each persistent event
and then use attr->config to select the event, e.g.:

		if (desc->unique_id == attr->config) ...

Tracepoints have this unique id in it already, thus tp only first.

The change can be made without breaking ABI, since the event setup is
described in sysfs.

-Robert

> 
> > +			event = desc->event;
> > +			break;
> > +		}
> > +	}

  reply	other threads:[~2013-06-25  8:41 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-06-11 16:42 [PATCH v2 00/14] perf, persistent: Kernel updates for perf tool integration Robert Richter
2013-06-11 16:42 ` [PATCH v2 01/14] perf, ring_buffer: Use same prefix Robert Richter
2013-06-11 16:42 ` [PATCH v2 02/14] perf: Add persistent events Robert Richter
2013-06-24  9:28   ` Peter Zijlstra
2013-06-24 19:24     ` Borislav Petkov
2013-06-25  8:46       ` Robert Richter
2013-06-11 16:42 ` [PATCH v2 03/14] perf: Add persistent event facilities Robert Richter
2013-06-14  2:15   ` Namhyung Kim
2013-06-14  7:20     ` Robert Richter
2013-06-24  9:32   ` Peter Zijlstra
2013-06-25  8:47     ` Robert Richter
2013-06-24  9:44   ` Peter Zijlstra
2013-06-25  8:41     ` Robert Richter [this message]
2013-06-24  9:48   ` Peter Zijlstra
2013-06-24 19:26     ` Borislav Petkov
2013-06-25  7:44       ` Peter Zijlstra
2013-06-25  9:24         ` Robert Richter
2013-06-25  9:37           ` Borislav Petkov
2013-06-25 10:51             ` Robert Richter
2013-06-25 15:29               ` Borislav Petkov
2013-06-25 16:14                 ` Robert Richter
2013-06-11 16:42 ` [PATCH v2 04/14] MCE: Enable persistent event Robert Richter
2013-06-11 16:42 ` [PATCH v2 05/14] perf, persistent: Rework struct pers_event_desc Robert Richter
2013-06-11 16:42 ` [PATCH v2 06/14] perf, persistent: Remove rb_put() Robert Richter
2013-06-11 16:42 ` [PATCH v2 07/14] perf, persistent: Introduce get_persistent_event() Robert Richter
2013-06-11 16:42 ` [PATCH v2 08/14] perf, persistent: Reworking perf_get_persistent_event_fd() Robert Richter
2013-06-11 16:42 ` [PATCH v2 09/14] perf, persistent: Protect event lists with mutex Robert Richter
2013-06-11 16:42 ` [PATCH v2 10/14] perf, persistent: Avoid adding identical events Robert Richter
2013-06-11 16:42 ` [PATCH v2 11/14] perf, persistent: Implementing a persistent pmu Robert Richter
2013-06-11 16:42 ` [PATCH v2 12/14] perf, persistent: Name each persistent event Robert Richter
2013-06-11 16:42 ` [PATCH v2 13/14] perf, persistent: Exposing persistent events using sysfs Robert Richter
2013-06-14  2:36   ` Namhyung Kim
2013-06-14  8:57     ` Robert Richter
2013-06-11 16:42 ` [PATCH v2 14/14] perf, persistent: Allow multiple users for an event Robert Richter
2013-06-24 10:08 ` [PATCH v2 00/14] perf, persistent: Kernel updates for perf tool integration Peter Zijlstra
2013-06-25 10:46   ` Robert Richter
2013-06-24 10:22 ` Peter Zijlstra
2013-06-25 16:56   ` Robert Richter
2013-06-24 10:24 ` Peter Zijlstra
2013-06-24 15:25 ` Peter Zijlstra
2013-06-24 19:45   ` Ingo Molnar
2013-06-25 17:57     ` Robert Richter
2013-06-25 19:16       ` Borislav Petkov
2013-06-26  8:12         ` Robert Richter
2013-06-26  8:24           ` Borislav Petkov
2013-06-26  9:46             ` Ingo Molnar
2013-06-26  9:56               ` Borislav Petkov
2013-06-26 10:11             ` Robert Richter
2013-06-26 11:45               ` Ingo Molnar
2013-06-26 12:25                 ` Ingo Molnar
2013-06-26 12:44                 ` Robert Richter
2013-06-27  5:46                   ` Namhyung Kim
2013-06-27  8:35                     ` Borislav Petkov
2013-06-27  8:50                       ` Ingo Molnar

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20130625084135.GA21579@rric.localhost \
    --to=rric@kernel.org \
    --cc=acme@infradead.org \
    --cc=bp@alien8.de \
    --cc=bp@suse.de \
    --cc=jolsa@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).