From owner-cvs-all Thu Jan 20 2:12:26 2000 Delivered-To: cvs-all@freebsd.org Received: from leia.creative.net.au (leia.creative.net.au [203.26.193.233]) by hub.freebsd.org (Postfix) with ESMTP id 970C5153C3 for ; Thu, 20 Jan 2000 02:12:21 -0800 (PST) (envelope-from adrian@leia.creative.net.au) Received: (from adrian@localhost) by leia.creative.net.au (8.9.3/8.9.3) id LAA01385 for cvs-all@freebsd.org; Thu, 20 Jan 2000 11:12:13 +0100 (CET) (envelope-from adrian) Date: Thu, 20 Jan 2000 11:12:13 +0100 From: Adrian Chadd To: cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/kern vfs_aio.c Message-ID: <20000120111213.A1357@leia.creative.net.au> References: <20000119210539.F20191@fw.wintelcom.net> <20000119225846.W27689@sturm.canonware.com> <20000119233702.E20751@fw.wintelcom.net> <20000119233311.X27689@sturm.canonware.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0pre2i In-Reply-To: <20000119233311.X27689@sturm.canonware.com> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk On Wed, Jan 19, 2000, Jason Evans wrote: > On Wed, Jan 19, 2000 at 11:37:02PM -0800, Alfred Perlstein wrote: > > Jason Evans [000119 23:24] wrote: > > > The person who sent me the patch was convinced that it was a necessary > > > change due to problems he had run into, and he convinced me at the time, > > > though my understanding of the code isn't great. > > > > > > Contrary to a previous statement, I'm going to leave the code as is for now > > > and actually write some test code to try to produce a problem with/without > > > the change. > > > > Can you post the issue he brought up that needed the spl lifted? > > No. It was in an xchat (IRC) buffer that crashed. =( I think he was seeing > permanently blocked processes. Twas me, and I was getting the system locking up in tsleep(). I wasn't entirely sure of the splx() fix, obviously I was looking at the wrong place, thanks for pointing it out. I'll spend some time today tidying up the other few places (eg lio_listio() ) where a similar race condition might occur, and I'll see if I can lock up the box with the splx() change reverted. Since I'm abusing the AIO code rather heavily now, I should be able to expose some bugs that might exist. The only other person I know of that is using the AIO code is Chris Sedore, but I don't know how heavy he's using the code (and whether he's doing disk IO + socket IO like I'm doing right now.) Adrian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message