[RFC] [PATCH] scsi_eh_abort_cmds on failed devices

[Hannes Reinecke <hare@xxxxxxx>]

Hi all,

I keep on stumbling across this in scsi_eh_abort_cmds():

		rtn = scsi_try_to_abort_cmd(scmd);
		if (rtn == SUCCESS) {
			scmd->eh_eflags &= ~SCSI_EH_CANCEL_CMD;
			if (!scsi_device_online(scmd->device) ||
			    !scsi_eh_tur(scmd)) {
				scsi_eh_finish_cmd(scmd, done_q);
			}

Now scsi_device_online() translates to:

static inline int scsi_device_online(struct scsi_device *sdev)
{
	return sdev->sdev_state != SDEV_OFFLINE;
}

Assume that the device has already vanished at this point
(ie sdev_state would be SDEV_DEL) we would be happily sending
a TUR command to a deleted device, right?
And as scsi_eh_tur command is using the ->queuecommand callback
directly, any normal checks for the sdev_state do not apply here.

Isn't something like this in order here?

diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
index f394649..6e2305d 100644
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c
@@ -778,6 +778,9 @@ static int scsi_send_eh_cmnd(struct scsi_cmnd *scmd, unsigne
d char *cmnd,
       struct scsi_eh_save ses;
       int rtn;

+       if (sdev->sdev_state == SDEV_CANCEL || sdev->sdev_state == SDEV_DEL)
+               return FAILED;
+
       scsi_eh_prep_cmnd(scmd, &ses, cmnd, cmnd_size, sense_bytes);
       shost->eh_action = &done;

Thanks for any feedback.

Cheers,

Hannes
--
Dr. Hannes Reinecke		      zSeries & Storage
hare@xxxxxxx			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: Markus Rex, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html