Date: Fri, 24 Jun 2011 08:49:22 -0400
From: Eric B Munson <emunson@mgebm.net>
To: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: mingo@elte.hu, borislav.petkov@amd.com, bblum@andrew.cmu.edu,
        linux-kernel@vger.kernel.org, mhack@us.ibm.com, eranian@google.com
Subject: Re: [PATCH 3/3] events: Ensure that timers are updated without
 requiring read() call
Message-ID: <20110624124921.GB3727@mgebm.net>
References: <1308861279-15216-1-git-send-email-emunson@mgebm.net>
 <1308861279-15216-3-git-send-email-emunson@mgebm.net>
 <1308908656.27849.21.camel@twins>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="98e8jtXdkpgskNou"
Content-Disposition: inline
In-Reply-To: <1308908656.27849.21.camel@twins>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3075
Lines: 89


--98e8jtXdkpgskNou
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, 24 Jun 2011, Peter Zijlstra wrote:

> On Thu, 2011-06-23 at 16:34 -0400, Eric B Munson wrote:
> > The event tracing infrastructure exposes two timers which should be upd=
ated
> > each time the value of the counter is updated.  Currently, these counte=
rs are
> > only updated when userspace calls read() on the fd associated with an e=
vent.
> > This means that counters which are read via the mmap'd page exclusively=
 never
> > have their timers updated.  This patch adds ensures that the timers are=
 updated
> > each time the values in the mmap'd page are updated.
> >=20
> > Signed-off-by: Eric B Munson <emunson@mgebm.net>
> > ---
> >  kernel/events/core.c |   12 ++++++++++++
> >  1 files changed, 12 insertions(+), 0 deletions(-)
> >=20
> > diff --git a/kernel/events/core.c b/kernel/events/core.c
> > index 9e9a7fa..e3be175 100644
> > --- a/kernel/events/core.c
> > +++ b/kernel/events/core.c
> > @@ -3382,6 +3382,18 @@ void perf_event_update_userpage(struct perf_even=
t *event)
> >  	struct perf_buffer *buffer;
> > =20
> >  	rcu_read_lock();
> > +	/*
> > +	 * compute total_time_enabled, total_time_running
> > +	 * based on snapshot values taken when the event
> > +	 * was last scheduled in.
> > +	 *
> > +	 * we cannot simply called update_context_time()
> > +	 * because of locking issue as we are called in
>=20
> s/are/can be/
>=20
> > +	 * NMI context
> > +	 */
> > +	calc_timer_values(event,
> > +				&event->total_time_enabled,
> > +				&event->total_time_running);
>=20
> I'm not sure writing those from NMI context is a sane thing to do, best
> is to compute the values into a local variable and use that variable
> below.
>=20
> Took the first two patches.
>=20

Now that I think about it, this will just mask the problem.  I have a test
program uses the mmap'd user space page to access event counters (it never
calls read()).  In this case, the timer values in the event will never be
updated.  It will display "properly" but the structure won't ever be correc=
t.
Given that, how can we keep the event values current?

Eric

--98e8jtXdkpgskNou
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iQEcBAEBAgAGBQJOBIfRAAoJEH65iIruGRnNpP8IAK+MdHBptF/irAhndCAEq5LK
Q3N1D08tjGHMl5OntHR4/eKJOh7rOvr3oMxn6HaYn58MfA4Z1Q7+fX2wt0BdC+7B
kUnylNuxArRFiqEiZnIqY8yyz+NHnRIKVdXzwyl2pOwgwkT9OePiSCQLsvnwTc3g
GSx7GTqRUXK2Ozk6+iOgMd8v77aCajkIUn5B/JQlQD9XF4ktvc1e68O5DOEtBY1i
aW1mkOfsxOkHjty7SzNeP2QEtHsU5l3fauYNCEACoK/UEAwH/Dn29Z0TNXjM3/nx
Xx18MHyWhyDAjLytfeKwZ5guaiphG0KRvdD6qXZ/iRjtLHMoyeYW1qlUaAkkPws=
=Nv7q
-----END PGP SIGNATURE-----

--98e8jtXdkpgskNou--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/