[SCSI] libsas: fix timeout vs completion race

Until we have told the lldd to forget a task a timed out operation can
return from the hardware at any time.  Since completion frees the task
we need to make sure that no tasks run their normal completion handler
once eh has decided to manage the task.  Similar to
ata_scsi_cmd_error_handler() freeze completions to let eh judge the
outcome of the race.

Task collector mode is problematic because it presents a situation where
a task can be timed out and aborted before the lldd has even seen it.
For this case we need to guarantee that a task that an lldd has been
told to forget does not get queued after the lldd says "never seen it".
With sas_scsi_timed_out we achieve this with the ->task_queue_flush
mutex, rather than adding more time.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>

1 files changed, 13 insertions, 22 deletions

diff --git a/drivers/scsi/libsas/sas_ata.c b/drivers/scsi/libsas/sas_ata.c
index 903bb441b9f9..4c2a1402373c 100644
--- a/drivers/scsi/libsas/sas_ata.c
+++ b/drivers/scsi/libsas/sas_ata.c
@@ -93,21 +93,30 @@ static enum ata_completion_errors sas_to_ata_err(struct task_status_struct *ts)
 static void sas_ata_task_done(struct sas_task *task)
 {
         struct ata_queued_cmd *qc = task->uldd_task;
-        struct domain_device *dev;
+        struct domain_device *dev = task->dev;
         struct task_status_struct *stat = &task->task_status;
         struct ata_task_resp *resp = (struct ata_task_resp *)stat->buf;
-        struct sas_ha_struct *sas_ha;
+        struct sas_ha_struct *sas_ha = dev->port->ha;
         enum ata_completion_errors ac;
         unsigned long flags;
         struct ata_link *link;
         struct ata_port *ap;
 
+        spin_lock_irqsave(&dev->done_lock, flags);
+        if (test_bit(SAS_HA_FROZEN, &sas_ha->state))
+                task = NULL;
+        else if (qc && qc->scsicmd)
+                ASSIGN_SAS_TASK(qc->scsicmd, NULL);
+        spin_unlock_irqrestore(&dev->done_lock, flags);
+
+        /* check if libsas-eh got to the task before us */
+        if (unlikely(!task))
+                return;
+
         if (!qc)
                 goto qc_already_gone;
 
         ap = qc->ap;
-        dev = ap->private_data;
-        sas_ha = dev->port->ha;
         link = &ap->link;
 
         spin_lock_irqsave(ap->lock, flags);
@@ -156,8 +165,6 @@ static void sas_ata_task_done(struct sas_task *task)
         }
 
         qc->lldd_task = NULL;
-        if (qc->scsicmd)
-                ASSIGN_SAS_TASK(qc->scsicmd, NULL);
         ata_qc_complete(qc);
         spin_unlock_irqrestore(ap->lock, flags);
 
@@ -633,22 +640,6 @@ void sas_ata_strategy_handler(struct Scsi_Host *shost)
         sas_enable_revalidation(sas_ha);
 }
 
-int sas_ata_timed_out(struct scsi_cmnd *cmd, struct sas_task *task,
-                      enum blk_eh_timer_return *rtn)
-{
-        struct domain_device *ddev = cmd_to_domain_dev(cmd);
-
-        if (!dev_is_sata(ddev) || task)
-                return 0;
-
-        /* we're a sata device with no task, so this must be a libata
-         * eh timeout.  Ideally should hook into libata timeout
-         * handling, but there's no point, it just wants to activate
-         * the eh thread */
-        *rtn = BLK_EH_NOT_HANDLED;
-        return 1;
-}
-
 int sas_ata_eh(struct Scsi_Host *shost, struct list_head *work_q,
                struct list_head *done_q)
 {