Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 26 Jun 2006 09:55:25 -0400
From:      John Baldwin <jhb@freebsd.org>
To:        John Birrell <jb@freebsd.org>
Cc:        Perforce Change Reviews <perforce@freebsd.org>
Subject:   Re: PERFORCE change 100025 for review
Message-ID:  <200606260955.25591.jhb@freebsd.org>
In-Reply-To: <200606260537.k5Q5bOgw020000@repoman.freebsd.org>
References:  <200606260537.k5Q5bOgw020000@repoman.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 26 June 2006 01:37, John Birrell wrote:
> http://perforce.freebsd.org/chv.cgi?CH=100025
> 
> Change 100025 by jb@jb_freebsd2 on 2006/06/26 05:37:23
> 
> 	The use of FreeBSD's semaphores here creates scheduler instability.
> 	This device needs a clean path from the timer interrupt.
> 	
> 	For the time being, since I'm only working on single processor
> 	systems, comment out the semaphore calls. When I get access to a 
> 	multi-cpu machine, I'll need to implement a simple semaphore system
> 	for use here.

The only thing you can use from the timer interrupt is a spin mutex as far
as synchronization primitives.  Note that you can provide sleep/wakeup
via msleep_spin() and wakeup() + wakeup_one().

-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200606260955.25591.jhb>