Re: sata_mv port lockup on hotplug (kernel 2.6.38.2)

[Tejun Heo <htejun@gmail.com>]

Hello,

On Tue, Sep 06, 2011 at 01:19:44PM +0100, Bruce Stenning wrote:
> ata4: EH complete
> Waking error handler thread
> scsi_eh_wakeup: succeeded
> scsi_schedule_eh: succeeded
> scsi_restart_operations: waking up host to restart
> Error handler scsi_eh_3 sleeping

I think the following should fix the problem.  The code there is from
the time when libata shared scsi_host->host_lock.  libata no longer
does that so, in the current state, host_eh_scheduled can be cleared
with actual pending EH condition.

You should be able to reproduce the problem more easily by adding
delay (something like mdelay(5)) before host_eh_scheduled clearing
without the following patch applied.

Thanks.

diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index ed16fbe..e971784 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -771,6 +771,14 @@ void ata_scsi_port_error_handler(struct Scsi_Host *host, struct ata_port *ap)
 		/* process port suspend request */
 		ata_eh_handle_port_suspend(ap);
 
+		/*
+		 * Single iteration complete.  Clear SCSI EH scheduled
+		 * state and check whether another iteration is necessary.
+		 */
+		spin_lock_irqsave(host->host_lock, flags);
+		host->host_eh_scheduled = 0;
+		spin_unlock_irqrestore(host->host_lock, flags);
+
 		/* Exception might have happened after ->error_handler
 		 * recovered the port but before this point.  Repeat
 		 * EH in such case.
@@ -792,13 +800,6 @@ void ata_scsi_port_error_handler(struct Scsi_Host *host, struct ata_port *ap)
 		ata_for_each_link(link, ap, HOST_FIRST)
 			memset(&link->eh_info, 0, sizeof(link->eh_info));
 
-		/* Clear host_eh_scheduled while holding ap->lock such
-		 * that if exception occurs after this point but
-		 * before EH completion, SCSI midlayer will
-		 * re-initiate EH.
-		 */
-		host->host_eh_scheduled = 0;
-
 		spin_unlock_irqrestore(ap->lock, flags);
 		ata_eh_release(ap);
 	} else {