Message-ID: <1337103795.14207.343.camel@gandalf.stny.rr.com>
Subject: Re: [RFC][PATCH RT] rwsem_rt: Another (more sane) approach to mulit
 reader rt locks
From: Steven Rostedt <rostedt@goodmis.org>
To: Peter Zijlstra <peterz@infradead.org>
Cc: LKML <linux-kernel@vger.kernel.org>, RT <linux-rt-users@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Clark Williams <williams@redhat.com>
Date: Tue, 15 May 2012 13:43:15 -0400
In-Reply-To: <1337103096.27694.94.camel@twins>
References: <1337090625.14207.304.camel@gandalf.stny.rr.com>
	 <1337094383.27694.62.camel@twins>
	 <1337096542.14207.315.camel@gandalf.stny.rr.com>
	 <1337102732.14207.336.camel@gandalf.stny.rr.com>
	 <1337103096.27694.94.camel@twins>
Content-Type: text/plain; charset="ISO-8859-15"
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1719
Lines: 45

On Tue, 2012-05-15 at 19:31 +0200, Peter Zijlstra wrote:
> On Tue, 2012-05-15 at 13:25 -0400, Steven Rostedt wrote:
> > On Tue, 2012-05-15 at 11:42 -0400, Steven Rostedt wrote:
> > > On Tue, 2012-05-15 at 17:06 +0200, Peter Zijlstra wrote:
> > > > On Tue, 2012-05-15 at 10:03 -0400, Steven Rostedt wrote:
> > > > > 
> > > > > where readers may nest (the same task may grab the same rwsem for
> > > > > read multiple times), but only one task may hold the rwsem at any
> > > > > given
> > > > > time (for read or write).
> > > > 
> > > > Humm, that sounds iffy, rwsem isn't a recursive read lock only rwlock_t
> > > > is.
> > > 
> > > In that case, current -rt is broken. As it has it being a recursive lock
> > > (without my patch).
> 
> Nah not broken, just pointless. A recursive lock that's not used
> recursively is fine.

Heh, sure :-) But as -rt keeps it recursive, I didn't want to change
that.

> 
> > 
> > Why wouldn't it be recursive. If two different tasks are allowed to grab
> > a read lock at the same time, why can't the same task grab a read lock
> > twice? As long as it releases it the same amount of times.
> > 
> > Now you can't grab a read lock if you have the write lock.
> 
> rwsem is fifo-fair, if a writer comes in between the second read
> acquisition (even by the same task) would block and you'd be a deadlock
> since the write won't succeed since you're still holding a reader.

Yep agreed. And this patch didn't change that either.

-- Steve


--
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/