Date: Mon, 9 May 2005 18:11:43 -0700
From: "Paul E. McKenney" <paulmck@us.ibm.com>
To: Joe Seigh <jseigh_02@xemaps.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: RCU + SMR for preemptive kernel/user threads.
Message-ID: <20050510011143.GC1337@us.ibm.com>
Reply-To: paulmck@us.ibm.com
References: <opsqivh7agehbc72@grunion>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <opsqivh7agehbc72@grunion>
User-Agent: Mutt/1.4.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 986
Lines: 25

On Mon, May 09, 2005 at 03:09:33PM -0400, Joe Seigh wrote:

[ . . . ]

> *** The per cpu times are in jiffies rounded off from some higher  
> resolution value.
> This means that for n time counters, you may have to wait up to n jiffies  
> for
> at least one round up to occur.  It would be nicer to have the higher  
> resolution
> value or even better a per processor number of interrupts to allow more  
> flexibility
> and finer granularity in polling intervals.

You can use smp_call_function() to invoke a function on each of the
other CPUs.  The implementations of smp_call_function() use inter-processor
interrupts, which should suffice, but you can always make the called
function just execute a memory barrier.

							Thanx, Paul
-
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/