redonkable/alistair23-linux
redonkable/alistair23-linux
1
0
Fork 0
Code Releases Activity

isci: Properly handle requests in the "aborting" state.

Browse source 
When a TMF times-out, the request is set back to "aborting".
Requests in the "aborting" state must be terminated when
LUN and device resets occur.

Signed-off-by: Jeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
This commit is contained in:
Jeff Skirvin 2011-03-31 13:10:34 -07:00 committed by Dan Williams
parent de728b7d72
commit f219f010a3
2 changed files with 12 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
drivers/scsi/isci/request.h
View file

@ -199,7 +199,7 @@ static inline enum isci_request_status isci_request_change_started_to_newstate(
old_state = isci_request->status;
if (old_state == started) {
if (old_state == started || old_state == aborting) {
BUG_ON(isci_request->io_request_completion != NULL);
isci_request->io_request_completion = completion_ptr;

14
drivers/scsi/isci/task.c
View file

@ -903,7 +903,9 @@ static void isci_terminate_request(
new_request_state
);
if ((old_state == started) || (old_state == completed)) {
if ((old_state == started) ||
(old_state == completed) ||
(old_state == aborting)) {
/* If the old_state is started:
* This request was not already being aborted. If it had been,
@ -920,6 +922,10 @@ static void isci_terminate_request(
* This request completed from the SCU hardware perspective
* and now just needs cleaning up in terms of freeing the
* request and potentially calling up to libsas.
*
* If old_state == aborting:
* This request has already gone through a TMF timeout, but may
* not have been terminated; needs cleaning up at least.
*/
isci_terminate_request_core(isci_host, isci_device,
isci_request);
@ -1297,14 +1303,16 @@ int isci_task_abort_task(struct sas_task *task)
spin_lock_irqsave(&isci_host->scic_lock, flags);
/* Check the request status and change to "aborting" if currently
/* Check the request status and change to "aborted" if currently
* "starting"; if true then set the I/O kernel completion
* struct that will be triggered when the request completes.
*/
old_state = isci_task_validate_request_to_abort(
old_request, isci_host, isci_device,
&aborted_io_completion);
if ((old_state != started) && (old_state != completed)) {
if ((old_state != started) &&
(old_state != completed) &&
(old_state != aborting)) {
spin_unlock_irqrestore(&isci_host->scic_lock, flags);