Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754375AbYJASxU (ORCPT ); Wed, 1 Oct 2008 14:53:20 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753416AbYJASxJ (ORCPT ); Wed, 1 Oct 2008 14:53:09 -0400 Received: from mx2.redhat.com ([66.187.237.31]:48647 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752570AbYJASxG (ORCPT ); Wed, 1 Oct 2008 14:53:06 -0400 Date: Wed, 01 Oct 2008 14:50:39 -0400 (EDT) Message-Id: <20081001.145039.39152159.k-ueda@ct.jp.nec.com> To: James.Bottomley@HansenPartnership.com Cc: linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org, dm-devel@redhat.com, akpm@linux-foundation.org, michaelc@cs.wisc.edu, j-nomura@ce.jp.nec.com, k-ueda@ct.jp.nec.com Subject: [PATCH 1/1] scsi: export busy state via q->lld_busy_fn() From: Kiyoshi Ueda Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6284 Lines: 179 Hi James, I hope the previous RFC patch(*) would be no problem, since I haven't gotten any negative comment. (*) http://lkml.org/lkml/2008/9/25/262 So could you take this patch for 2.6.28 additionally? This patch implements a new interface of the block layer for request stacking drivers. There should be no effect on existing drivers' behavior. This patch was created on top of the commit below of scsi-post-merge-2.6. --------------------------------------------------------------------- commit e49f03f37104c0e7cae7c455480069bada00932f Author: James Bottomley Date: Fri Sep 12 16:46:51 2008 -0500 [SCSI] scsi_error: fix target reset handling --------------------------------------------------------------------- And this patch depends on the following block layer patch, which is in Jens' for-2.6.28 of linux-2.6-block. http://lkml.org/lkml/2008/9/29/142 Thanks, Kiyoshi Ueda Subject: [PATCH 1/1] scsi: export busy state via q->lld_busy_fn() This patch implements q->lld_busy_fn() for scsi mid layer to export its busy state. Please note that it checks the cached information (sdev->lld_busy) for efficiency, instead of checking the actual state of sdev/starget/shost everytime. So the care must be taken not to leave sdev->lld_busy = 1 for the following cases: - when there is no request in the sdev queue - when scsi can't dispatch I/Os anymore and needs to kill I/Os Otherwise, request stacking drivers may not submit any request, and then, nobody sets back lld_busy = 0 and that causes deadlock. OTOH, it has no major problem in setting sdev->lld_busy = 0 even when the starget/shost is actually busy, because newly submitted request will soon turn it to 1 in scsi_request_fn(). Signed-off-by: Kiyoshi Ueda Signed-off-by: Jun'ichi Nomura Cc: Andrew Morton Cc: Mike Christie Cc: James Bottomley --- drivers/scsi/scsi.c | 4 ++-- drivers/scsi/scsi_lib.c | 20 +++++++++++++++++++- include/scsi/scsi_device.h | 13 +++++++++++++ 3 files changed, 34 insertions(+), 3 deletions(-) Index: scsi-post-merge-2.6/drivers/scsi/scsi_lib.c =================================================================== --- scsi-post-merge-2.6.orig/drivers/scsi/scsi_lib.c +++ scsi-post-merge-2.6/drivers/scsi/scsi_lib.c @@ -480,6 +480,8 @@ void scsi_device_unbusy(struct scsi_devi spin_unlock(shost->host_lock); spin_lock(sdev->request_queue->queue_lock); sdev->device_busy--; + if (sdev->device_busy < sdev->queue_depth && !sdev->device_blocked) + sdev->lld_busy = 0; spin_unlock_irqrestore(sdev->request_queue->queue_lock, flags); } @@ -1535,6 +1537,13 @@ static void scsi_softirq_done(struct req } } +static int scsi_lld_busy(struct request_queue *q) +{ + struct scsi_device *sdev = q->queuedata; + + return sdev ? sdev->lld_busy : 0; +} + /* * Function: scsi_request_fn() * @@ -1577,9 +1586,14 @@ static void scsi_request_fn(struct reque * accept it. */ req = elv_next_request(q); - if (!req || !scsi_dev_queue_ready(q, sdev)) + if (!req) break; + if (!scsi_dev_queue_ready(q, sdev)) { + sdev->lld_busy = 1; + break; + } + if (unlikely(!scsi_device_online(sdev))) { sdev_printk(KERN_ERR, sdev, "rejecting I/O to offline device\n"); @@ -1649,6 +1663,8 @@ static void scsi_request_fn(struct reque rtn = scsi_dispatch_cmd(cmd); spin_lock_irq(q->queue_lock); if(rtn) { + sdev->lld_busy = 1; + /* we're refusing the command; because of * the way locks get dropped, we need to * check here if plugging is required */ @@ -1673,6 +1689,7 @@ static void scsi_request_fn(struct reque * later time. */ spin_lock_irq(q->queue_lock); + sdev->lld_busy = 1; blk_requeue_request(q, req); sdev->device_busy--; if(sdev->device_busy == 0) @@ -1756,6 +1773,7 @@ struct request_queue *scsi_alloc_queue(s blk_queue_prep_rq(q, scsi_prep_fn); blk_queue_softirq_done(q, scsi_softirq_done); blk_queue_rq_timed_out(q, scsi_times_out); + blk_queue_lld_busy(q, scsi_lld_busy); return q; } Index: scsi-post-merge-2.6/drivers/scsi/scsi.c =================================================================== --- scsi-post-merge-2.6.orig/drivers/scsi/scsi.c +++ scsi-post-merge-2.6/drivers/scsi/scsi.c @@ -809,8 +809,6 @@ void scsi_finish_command(struct scsi_cmn struct scsi_driver *drv; unsigned int good_bytes; - scsi_device_unbusy(sdev); - /* * Clear the flags which say that the device/host is no longer * capable of accepting new commands. These are set in scsi_queue.c @@ -823,6 +821,8 @@ void scsi_finish_command(struct scsi_cmn starget->target_blocked = 0; sdev->device_blocked = 0; + scsi_device_unbusy(sdev); + /* * If we have valid sense information, then some kind of recovery * must have taken place. Make a note of this. Index: scsi-post-merge-2.6/include/scsi/scsi_device.h =================================================================== --- scsi-post-merge-2.6.orig/include/scsi/scsi_device.h +++ scsi-post-merge-2.6/include/scsi/scsi_device.h @@ -145,6 +145,19 @@ struct scsi_device { unsigned retry_hwerror:1; /* Retry HARDWARE_ERROR */ unsigned last_sector_bug:1; /* do not use multisector accesses on SD_LAST_BUGGY_SECTORS */ + unsigned lld_busy:1; /* Exporting busy state for stacking + * drivers. + * + * 1 - There is one or more requests + * in the device's queue, + * and the device (or target/host) + * can not dispatch any request + * immediately. + * 0 - The device can dispatch + * (or kill) requests immediately, + * or there is no request in + * the device's queue even though + * target/host may busy. */ DECLARE_BITMAP(supported_events, SDEV_EVT_MAXBITS); /* supported events */ struct list_head event_list; /* asserted events */ -- 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/