Message-ID: <3DB5D67A.26137CFC@splentec.com>
Date: Tue, 22 Oct 2002 18:51:38 -0400
From: Luben Tuikov <luben@splentec.com>
Organization: Splentec Ltd.
MIME-Version: 1.0
To: Christoph Hellwig <hch@infradead.org>
CC: Stephen Lord <lord@sgi.com>, linux-xfs <linux-xfs@oss.sgi.com>,
   Eric Sandeen <sandeen@sgi.com>
Subject: Re: PATCH: sleeping while holding a lock in	
 _pagebuf_free_bh()::page_buf.c
References: <3DB49424.9E4CAC0F@splentec.com> <1035289272.9684.13.camel@laptop.americas.sgi.com> <3DB5AD3F.80402@splentec.com> <20021022210658.A10581@infradead.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: linux-xfs-bounce@oss.sgi.com
Errors-to: linux-xfs-bounce@oss.sgi.com
Precedence: bulk

Christoph Hellwig wrote:
> 
> On Tue, Oct 22, 2002 at 03:55:43PM -0400, Luben Tuikov wrote:
> > This is what wake_up_sync() does. wake_up() may reschedule
> > on an SMP system, since the other CPUs may be able to get
> > another task to run. This is why there are two different
> > implementations wake_up() and wake_up_sync().
> >
> > If you have 4 CPUs, 3 idle and call wake_up() why should
> > the 3 idle CPUs _STAY_ idle, _until_ you (later) call schedule()
> > and _then_ to schedule another task on the idle CPUs, and
> > you to continue to go on, say if you were a SCHED_FIFO, or
> > SCHED_OTHER...
> >
> > BTW, that the _whole_ point of SMP -- _symmetric_ MP.
> 
> Could you please give me some of thev crack you're smoking?

Yep, real professional!
 
-- 
Luben