linux-kernel - Re: [PATCH 2/3 -tip] perf_counter tools: Add support to set of multiple events in one shot

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20090629035708.GD8059@elte.hu>
Date:	Mon, 29 Jun 2009 05:57:08 +0200
From:	Ingo Molnar <mingo@...e.hu>
To:	Jaswinder Singh Rajput <jaswinder@...nel.org>
Cc:	Thomas Gleixner <tglx@...utronix.de>,
	Peter Zijlstra <peterz@...radead.org>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/3 -tip] perf_counter tools: Add support to set of
	multiple events in one shot


* Jaswinder Singh Rajput <jaswinder@...nel.org> wrote:

> On Sun, 2009-06-28 at 15:29 +0200, Ingo Molnar wrote:
> > * Jaswinder Singh Rajput <jaswinder@...nel.org> wrote:
> > 
> > > On Sat, 2009-06-27 at 18:38 +0200, Ingo Molnar wrote:
> > > > * Jaswinder Singh Rajput <jaswinder@...nel.org> wrote:
> > > > 
> > > > > Add support for HARDWARE and SOFTWARE events :
> > > > >  perf stat -e all-sw-events
> > > > >  perf stat -e sw-events
> > > > >  perf stat -e all-hw-events
> > > > >  perf stat -e hw-events
> > > > 
> > > > > +static struct event_type_symbol event_type_symbols[] = {
> > > > > + [PERF_TYPE_HARDWARE]	= { "hw-events",	"all-hw-events",	},
> > > > > + [PERF_TYPE_SOFTWARE]	= { "sw-events",	"all-sw-events",	},
> > > > 
> > > > Hm, this really just special-cases and open-codes these. The better 
> > > > solution is what i suggested in my review of your prior patches: 
> > > > regex pattern matching.
> > > > 
> > > 
> > > parse_events() is responsible to parse events for 'perf stat -e' 
> > > and it is parsing by parse_event_symbol()
> > > 
> > > If you want to use regex pattern matching then either we should 
> > > make some another option or if we need to rewrite 
> > > parse_event_symbol to use regex pattern matching which will be 
> > > applicable to all the events.
> > 
> > As i mentioned it before, i think the most intuitive solution is to 
> > extend the --event syntax with regex patterns. No new option - just 
> > richer -e syntax.
> > 
> > We could have this syntax:
> > 
> >   hw-cpu-cycles
> >   hw-instructions
> >   hw-cache-references
> >   hw-cache-misses
> >   hw-branch-instructions
> >   hw-branch-misses
> >   hw-bus-cycles
> > 
> >   sw-cpu-clock
> >   sw-task-clock
> >   sw-page-faults
> >   sw-minor-faults
> >   sw-major-faults
> >   sw-context-switches
> >   sw-cpu-migrations
> > 
> > regex patterns like:
> > 
> >   hw-*
> >   sw-*
> >   *
> > 
> > the first one would select all hardware events - the second all 
> > software events - the third all events in general. But other regex 
> > patterns make sense too, like:
> > 
> >   *branch*
> >   *cache*
> >   *fault*
> > 
> > And as the number of generic events increases, so will regex 
> > patterns become more and more useful.
> > 
> 
> This is a good approach, but this should be done separately 
> because these will effect all the events.
> 
> Currently I send patches to support multiple events in one shot 
> which is adding new feature.
> 
> And you telling to use regex wildcard for all the events in 
> parse_event_symbol() which is enhancement and should be done in 
> series of patches.

The above patterns i suggested _already cover_ 'multiple events'.

We might define further aliases like:

	all := "*"
        all-sw := "sw-*"

but it should all be in terms of patterns and regular expressions, 
not via some hardcoded special-case thing as your posted patches 
did.

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/