From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757391AbXGSGNb (ORCPT ); Thu, 19 Jul 2007 02:13:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750866AbXGSGNX (ORCPT ); Thu, 19 Jul 2007 02:13:23 -0400 Received: from brick.kernel.dk ([80.160.20.94]:29442 "EHLO kernel.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751952AbXGSGNW (ORCPT ); Thu, 19 Jul 2007 02:13:22 -0400 Date: Thu, 19 Jul 2007 08:13:01 +0200 From: Jens Axboe To: Linus Torvalds Cc: Bartlomiej Zolnierkiewicz , Giacomo Catenazzi , Linux Kernel Mailing List , akpm@linux-foundation.org Subject: Re: regression: disk error loop (panic?) ide_do_rw_disk-bad: Message-ID: <20070719061300.GP11657@kernel.dk> References: <469D1D5E.8050609@cateee.net> <20070718202740.GN11657@kernel.dk> <200707190053.24578.bzolnier@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jul 18 2007, Linus Torvalds wrote: > > > On Thu, 19 Jul 2007, Bartlomiej Zolnierkiewicz wrote: > > > > Thanks for finding and fixing this. > > > > The latest patch (with additional cleanups) also looks good and should be > > safe enough (unchanged behavior for all non-pc requests) to merge it now. > > > > Acked-by: Bartlomiej Zolnierkiewicz > > Ok, Jens - mind signing off on the patch you sent out, and writing an > explanatory message? Feel free to just crib from my explanation of my > original patch, or whatever. Sure thing, it's below. > And it would be beautiful if people who saw the bad behaviour before > reverting the ide.c changes were to go back to that broken state, and > try the patch, and just verify that it acts like it should (ie you > should see just a few error messages, and it shouldn't cause the IDE > layer to go ballistic any more). --- [PATCH] IDE: fix termination of non-fs requests ide-disk calls ide_end_request(drive, 0, 0); to finish an unknown request, but this doesn't work so well for non-fs requests, since ide_end_request() internally looks at ->hard_cur_sectors to see how much data to end. Only file system requests store a transfer value in there, pc requests fill out ->data_len as a byte based transfer value instead. Since we ask to end 0 bytes of that request, it will never be terminated and ide-disk gets stuck in a loop "handling" that same request over and over. Switch __ide_end_request() to take a byte based transfer count, and adjust ide_end_request() to look at the right field to determine how much IO to end when it's being passed in 0. Acked-by: Bartlomiej Zolnierkiewicz Signed-off-by: Jens Axboe diff --git a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c index c5b5011..f9de798 100644 --- a/drivers/ide/ide-io.c +++ b/drivers/ide/ide-io.c @@ -55,7 +55,7 @@ #include static int __ide_end_request(ide_drive_t *drive, struct request *rq, - int uptodate, int nr_sectors) + int uptodate, unsigned int nr_bytes) { int ret = 1; @@ -64,7 +64,7 @@ static int __ide_end_request(ide_drive_t *drive, struct request *rq, * complete the whole request right now */ if (blk_noretry_request(rq) && end_io_error(uptodate)) - nr_sectors = rq->hard_nr_sectors; + nr_bytes = rq->hard_nr_sectors << 9; if (!blk_fs_request(rq) && end_io_error(uptodate) && !rq->errors) rq->errors = -EIO; @@ -78,7 +78,7 @@ static int __ide_end_request(ide_drive_t *drive, struct request *rq, HWGROUP(drive)->hwif->ide_dma_on(drive); } - if (!end_that_request_first(rq, uptodate, nr_sectors)) { + if (!end_that_request_chunk(rq, uptodate, nr_bytes)) { add_disk_randomness(rq->rq_disk); if (!list_empty(&rq->queuelist)) blkdev_dequeue_request(rq); @@ -103,6 +103,7 @@ static int __ide_end_request(ide_drive_t *drive, struct request *rq, int ide_end_request (ide_drive_t *drive, int uptodate, int nr_sectors) { + unsigned int nr_bytes = nr_sectors << 9; struct request *rq; unsigned long flags; int ret = 1; @@ -114,10 +115,14 @@ int ide_end_request (ide_drive_t *drive, int uptodate, int nr_sectors) spin_lock_irqsave(&ide_lock, flags); rq = HWGROUP(drive)->rq; - if (!nr_sectors) - nr_sectors = rq->hard_cur_sectors; + if (!nr_bytes) { + if (blk_pc_request(rq)) + nr_bytes = rq->data_len; + else + nr_bytes = rq->hard_cur_sectors << 9; + } - ret = __ide_end_request(drive, rq, uptodate, nr_sectors); + ret = __ide_end_request(drive, rq, uptodate, nr_bytes); spin_unlock_irqrestore(&ide_lock, flags); return ret; -- Jens Axboe