From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750736AbWF2O1F (ORCPT ); Thu, 29 Jun 2006 10:27:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750750AbWF2O1F (ORCPT ); Thu, 29 Jun 2006 10:27:05 -0400 Received: from e31.co.us.ibm.com ([32.97.110.149]:57748 "EHLO e31.co.us.ibm.com") by vger.kernel.org with ESMTP id S1750736AbWF2O1D (ORCPT ); Thu, 29 Jun 2006 10:27:03 -0400 Date: Thu, 29 Jun 2006 07:27:41 -0700 From: "Paul E. McKenney" To: Josh Triplett Cc: linux-kernel@vger.kernel.org, Dipkanar Sarma , Andrew Morton , Samuel Ortiz Subject: Re: [PATCH] irda: Fix RCU lock pairing on error path Message-ID: <20060629142741.GA1294@us.ibm.com> Reply-To: paulmck@us.ibm.com References: <1151542602.18723.19.camel@josh-work.beaverton.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1151542602.18723.19.camel@josh-work.beaverton.ibm.com> User-Agent: Mutt/1.4.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jun 28, 2006 at 05:56:41PM -0700, Josh Triplett wrote: > irlan_client_discovery_indication calls rcu_read_lock and rcu_read_unlock, but > returns without unlocking in an error case. Fix that by replacing the return > with a goto so that the rcu_read_unlock always gets executed. Good catch!!! Looks good from an RCU viewpoint, in fact, looks absolutely necessary to permit IRDA to work in debug mode, particularly in either a CONFIG_PREEMPT or a -rt kernel. One question below, but up to the maintainer. ;-) Thanx, Paul Acked-by: Paul E. McKenney > Signed-off-by: Josh Triplett > > --- > > net/irda/irlan/irlan_client.c | 3 ++- > 1 files changed, 2 insertions(+), 1 deletions(-) > > e20ab96944814489277c3bfd4e69133854ab01e9 > diff --git a/net/irda/irlan/irlan_client.c b/net/irda/irlan/irlan_client.c > index f8e6cb0..5ce7d2e 100644 > --- a/net/irda/irlan/irlan_client.c > +++ b/net/irda/irlan/irlan_client.c > @@ -173,13 +173,14 @@ void irlan_client_discovery_indication(d > rcu_read_lock(); > self = irlan_get_any(); > if (self) { > - IRDA_ASSERT(self->magic == IRLAN_MAGIC, return;); > + IRDA_ASSERT(self->magic == IRLAN_MAGIC, goto out;); > > IRDA_DEBUG(1, "%s(), Found instance (%08x)!\n", __FUNCTION__ , > daddr); > > irlan_client_wakeup(self, saddr, daddr); > } > +out: Should the above label instead be as follows? IRDA_ASSERT_LABEL(out:) Just in case some variant of gcc, sparse, or whatever complains about labels that don't have a corresponding goto (in CONFIG_IRDA_DEBUG=n builds). > rcu_read_unlock(); > } > > >