From: malahal@us.ibm.com
To: linux-scsi@vger.kernel.org, jens.axboe@oracle.com
Subject: [RFC] [PATCH 2/2] blk request timeout handler patches
Date: Thu, 4 Oct 2007 11:20:03 -0700 [thread overview]
Message-ID: <20071004182003.GC16689@us.ibm.com> (raw)
In-Reply-To: <20071004181259.GA16689@us.ibm.com>
Fix scsi_dispatch_cmd() to stop timers.
Signed-off-by: Malahal Naineni <malahal@us.ibm.com>
diff -r 870bb598c977 drivers/scsi/scsi.c
--- a/drivers/scsi/scsi.c Thu Sep 27 00:25:38 2007 -0700
+++ b/drivers/scsi/scsi.c Thu Sep 27 01:04:10 2007 -0700
@@ -471,14 +471,19 @@ int scsi_dispatch_cmd(struct scsi_cmnd *
unsigned long timeout;
int rtn = 0;
+ /*
+ * We will use a queued command if possible, otherwise we will
+ * emulate the queuing and calling of completion function ourselves.
+ */
+ atomic_inc(&cmd->device->iorequest_cnt);
+
/* check if the device is still usable */
if (unlikely(cmd->device->sdev_state == SDEV_DEL)) {
/* in SDEV_DEL we error all commands. DID_NO_CONNECT
* returns an immediate error upwards, and signals
* that the device is no longer present */
cmd->result = DID_NO_CONNECT << 16;
- atomic_inc(&cmd->device->iorequest_cnt);
- __blk_complete_request(cmd->request);
+ scsi_done(cmd);
/* return 0 (because the command has been processed) */
goto out;
}
@@ -491,7 +496,8 @@ int scsi_dispatch_cmd(struct scsi_cmnd *
* future requests should not occur until the device
* transitions out of the suspend state.
*/
- scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY);
+
+ scsi_queue_retry(cmd, SCSI_MLQUEUE_DEVICE_BUSY);
SCSI_LOG_MLQUEUE(3, printk("queuecommand : device blocked \n"));
@@ -536,12 +542,6 @@ int scsi_dispatch_cmd(struct scsi_cmnd *
scsi_log_send(cmd);
/*
- * We will use a queued command if possible, otherwise we will
- * emulate the queuing and calling of completion function ourselves.
- */
- atomic_inc(&cmd->device->iorequest_cnt);
-
- /*
* Before we queue this command, check if the command
* length exceeds what the host adapter can handle.
*/
@@ -571,12 +571,8 @@ int scsi_dispatch_cmd(struct scsi_cmnd *
}
spin_unlock_irqrestore(host->host_lock, flags);
if (rtn) {
- if (blk_delete_timer(cmd->request)) {
- atomic_inc(&cmd->device->iodone_cnt);
- scsi_queue_insert(cmd,
- (rtn == SCSI_MLQUEUE_DEVICE_BUSY) ?
- rtn : SCSI_MLQUEUE_HOST_BUSY);
- }
+ scsi_queue_retry(cmd, (rtn == SCSI_MLQUEUE_DEVICE_BUSY) ?
+ rtn : SCSI_MLQUEUE_HOST_BUSY);
SCSI_LOG_MLQUEUE(3,
printk("queuecommand : request rejected\n"));
}
diff -r 870bb598c977 drivers/scsi/scsi_lib.c
--- a/drivers/scsi/scsi_lib.c Thu Sep 27 00:25:38 2007 -0700
+++ b/drivers/scsi/scsi_lib.c Thu Sep 27 01:16:28 2007 -0700
@@ -160,6 +160,36 @@ int scsi_queue_insert(struct scsi_cmnd *
return 0;
}
+
+/*
+ * Function: scsi_queue_retry()
+ *
+ * Purpose: Try inserting a command in the midlevel queue.
+ *
+ * Arguments: cmd - command that we are adding to queue.
+ * reason - why we are inserting command to queue.
+ *
+ * Lock status: Assumed that lock is not held upon entry.
+ *
+ * Returns: Nothing.
+ *
+ * Notes: This is very similar to scsi_queue_insert except that we
+ * call this function when we don't know if the blk layer timer
+ * is active or not. We could implement this either by calling
+ * blk_delete_timer and inserting in the midlevel queue if we
+ * successfully delete the timer OR setting appropriate result
+ * field in the cmd and letting it go through the normal done
+ * routines which will retry the command. For now, We call
+ * blk_delete_timer!
+ */
+void scsi_queue_retry(struct scsi_cmnd *cmd, int reason)
+{
+ if (blk_delete_timer(cmd->request)) {
+ atomic_inc(&cmd->device->iodone_cnt);
+ scsi_queue_insert(cmd, reason);
+ }
+}
+
/**
* scsi_execute - insert request and wait for the result
diff -r 870bb598c977 drivers/scsi/scsi_priv.h
--- a/drivers/scsi/scsi_priv.h Thu Sep 27 00:25:38 2007 -0700
+++ b/drivers/scsi/scsi_priv.h Thu Sep 27 01:03:39 2007 -0700
@@ -64,6 +64,7 @@ extern int scsi_maybe_unblock_host(struc
extern int scsi_maybe_unblock_host(struct scsi_device *sdev);
extern void scsi_device_unbusy(struct scsi_device *sdev);
extern int scsi_queue_insert(struct scsi_cmnd *cmd, int reason);
+extern void scsi_queue_retry(struct scsi_cmnd *cmd, int reason);
extern void scsi_next_command(struct scsi_cmnd *cmd);
extern void scsi_run_host_queues(struct Scsi_Host *shost);
extern struct request_queue *scsi_alloc_queue(struct scsi_device *sdev);
next prev parent reply other threads:[~2007-10-04 18:20 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-10-04 18:12 [RFC] [PATCH 0/2] blk request timeout handler patches malahal
2007-10-04 18:17 ` [RFC] [PATCH 1/2] " malahal
2007-10-04 18:52 ` Randy Dunlap
2007-10-04 20:40 ` Salyzyn, Mark
2007-10-05 12:49 ` Jens Axboe
2007-10-08 6:54 ` malahal
2007-10-08 7:04 ` Jens Axboe
2007-10-09 5:36 ` [RFC] [PATCH 1/1] " malahal
2007-10-09 9:14 ` Jens Axboe
2007-10-09 14:26 ` malahal
2007-10-09 12:00 ` Matthew Wilcox
2007-10-09 12:15 ` Jens Axboe
2007-10-09 15:56 ` James Bottomley
2007-10-09 17:23 ` malahal
2007-10-10 12:25 ` Jens Axboe
2007-10-10 16:58 ` malahal
2007-10-10 17:04 ` Jens Axboe
2007-10-11 18:01 ` malahal
2007-10-11 18:24 ` Jens Axboe
2007-10-11 18:33 ` Jens Axboe
2007-10-23 1:45 ` malahal
2007-10-23 6:30 ` malahal
2007-10-23 11:59 ` Jens Axboe
2007-10-05 12:50 ` [RFC] [PATCH 1/2] " Jens Axboe
2007-10-04 18:20 ` malahal [this message]
2007-10-04 18:32 ` [RFC] [PATCH 2/2] " Randy Dunlap
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20071004182003.GC16689@us.ibm.com \
--to=malahal@us.ibm.com \
--cc=jens.axboe@oracle.com \
--cc=linux-scsi@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).