From: Dan Williams <dan.j.williams@intel.com>
To: linux-scsi@vger.kernel.org
Cc: linux-ide@vger.kernel.org
Subject: [PATCH v4 09/10] libsas: pre-clean commands that won the eh vs completion race
Date: Mon, 16 Jan 2012 21:11:48 -0800 [thread overview]
Message-ID: <20120117051148.22344.90576.stgit@localhost6.localdomain6> (raw)
In-Reply-To: <20120117050847.22344.4961.stgit@localhost6.localdomain6>
When scrolling forward through the eh list (in a clear_q scenario) it is
possible to encounter commands that won the completion vs eh race. Rather
than sprinkle more "if (!task)" throughout the handler just make a pass
through the list and delete the race winners before handling the rest.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
drivers/scsi/libsas/sas_scsi_host.c | 25 ++++++++++++++++---------
1 files changed, 16 insertions(+), 9 deletions(-)
diff --git a/drivers/scsi/libsas/sas_scsi_host.c b/drivers/scsi/libsas/sas_scsi_host.c
index 94ef763..731c892 100644
--- a/drivers/scsi/libsas/sas_scsi_host.c
+++ b/drivers/scsi/libsas/sas_scsi_host.c
@@ -249,8 +249,8 @@ out_done:
static void sas_eh_finish_cmd(struct scsi_cmnd *cmd)
{
- struct sas_task *task = TO_SAS_TASK(cmd);
struct sas_ha_struct *sas_ha = SHOST_TO_SAS_HA(cmd->device->host);
+ struct sas_task *task = TO_SAS_TASK(cmd);
/* At this point, we only get called following an actual abort
* of the task, so we should be guaranteed not to be racing with
@@ -267,9 +267,9 @@ static void sas_eh_finish_cmd(struct scsi_cmnd *cmd)
static void sas_eh_defer_cmd(struct scsi_cmnd *cmd)
{
- struct sas_task *task = TO_SAS_TASK(cmd);
- struct domain_device *dev = task->dev;
+ struct domain_device *dev = cmd_to_domain_dev(cmd);
struct sas_ha_struct *ha = dev->port->ha;
+ struct sas_task *task = TO_SAS_TASK(cmd);
if (!dev_is_sata(dev)) {
sas_eh_finish_cmd(cmd);
@@ -530,8 +530,9 @@ static int sas_eh_handle_sas_errors(struct Scsi_Host *shost,
struct sas_internal *i = to_sas_internal(shost->transportt);
unsigned long flags;
struct sas_ha_struct *ha = SHOST_TO_SAS_HA(shost);
+ LIST_HEAD(done);
-Again:
+ /* clean out any commands that won the completion vs eh race */
list_for_each_entry_safe(cmd, n, work_q, eh_entry) {
struct domain_device *dev = cmd_to_domain_dev(cmd);
struct sas_task *task;
@@ -545,7 +546,12 @@ Again:
spin_unlock_irqrestore(&dev->done_lock, flags);
if (!task)
- continue;
+ list_move_tail(&cmd->eh_entry, &done);
+ }
+
+ Again:
+ list_for_each_entry_safe(cmd, n, work_q, eh_entry) {
+ struct sas_task *task = TO_SAS_TASK(cmd);
list_del_init(&cmd->eh_entry);
@@ -649,15 +655,16 @@ Again:
goto clear_q;
}
}
+ out:
+ list_splice_tail(&done, work_q);
list_splice_tail_init(&ha->eh_ata_q, work_q);
return list_empty(work_q);
-clear_q:
+
+ clear_q:
SAS_DPRINTK("--- Exit %s -- clear_q\n", __func__);
list_for_each_entry_safe(cmd, n, work_q, eh_entry)
sas_eh_finish_cmd(cmd);
-
- list_splice_tail_init(&ha->eh_ata_q, work_q);
- return list_empty(work_q);
+ goto out;
}
void sas_scsi_recover_host(struct Scsi_Host *shost)
next prev parent reply other threads:[~2012-01-17 5:11 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-01-17 5:11 [resend PATCH v4 00/10] libsas: eh reworks (ata-eh vs discovery, races, ...) Dan Williams
2012-01-17 5:11 ` [PATCH v4 01/10] libsas: convert dev->gone to flags Dan Williams
2012-01-17 5:11 ` [PATCH v4 02/10] libsas: prevent domain rediscovery competing with ata error handling Dan Williams
2012-01-17 5:11 ` [PATCH v4 03/10] libsas: fix timeout vs completion race Dan Williams
2012-01-17 5:11 ` [PATCH v4 04/10] libsas: don't mark expanders as gone when a child device is removed Dan Williams
2012-01-17 5:11 ` [PATCH v4 05/10] libsas: check for 'gone' expanders in smp_execute_task() Dan Williams
2012-01-17 5:11 ` [PATCH v4 06/10] libsas: fix sas_find_local_phy(), take phy references Dan Williams
2012-01-17 5:11 ` [PATCH v4 07/10] libsas: don't recover 'gone' devices in sas_ata_hard_reset() Dan Williams
2012-01-17 5:11 ` [PATCH v4 08/10] isci: ->lldd_ata_check_ready handler Dan Williams
2012-01-17 5:11 ` Dan Williams [this message]
2012-01-17 5:11 ` [PATCH v4 10/10] libsas: feed the scsi_block_when_processing_errors() meter Dan Williams
2012-02-02 8:04 ` Dan Williams
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=20120117051148.22344.90576.stgit@localhost6.localdomain6 \
--to=dan.j.williams@intel.com \
--cc=linux-ide@vger.kernel.org \
--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