From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luben Tuikov Subject: Re: [PATCH] 2.5.x use list_head to handle scsi starved request queues Date: Mon, 24 Mar 2003 16:25:42 -0500 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <3E7F77D6.9040300@splentec.com> References: <20030319182755.A9535@beaverton.ibm.com> <3E7A1EF5.3050501@splentec.com> <20030320203912.A18471@beaverton.ibm.com> <3E7B7A9A.6030007@splentec.com> <20030321165050.B9578@beaverton.ibm.com> <3E7F3C67.5010704@splentec.com> <20030324112940.A11000@beaverton.ibm.com> <3E7F688C.3020009@splentec.com> <20030324202509.GF2371@suse.de> <20030324123813.A11614@beaverton.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: List-Id: linux-scsi@vger.kernel.org To: Patrick Mansfield Cc: Jens Axboe , James Bottomley , linux-scsi@vger.kernel.org Patrick Mansfield wrote: > > That is another patch. Then you should've sent it before this one. But nevertheless, the code of the patch you posted, assumed that q->queue_lock is the same for all devices' request queues. This was my beef since the very beginning of this thread. > >>Irk no, that's quite a bad idea. >> >>I completely agree with you that making assumptions about q->queue_lock >>== host lock is really really bad. Don't do that. > > Yes, it's bad (with or without the patch in question), I am working on No! With the patch in question. The rest of SCSI Core assumes that q->queue_lock has nothing to do with any other q's lock. > patches to make queue_lock per scsi_device (queue_lock != host_lock), > where I'm forced to make changes in this same area. -- Luben