Skip to contentSkip to conversation

Handle update hardware queues and queue freeze more carefully

PATCH Applied Daniel Wagner 5 participants 10 messages Sep 14, 2021 – Sep 21, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v7 cover Handle update hardware queues and queue freeze more carefully

Hi,

v7 is addressing the comments from Sagi (wrong commit id and add an explenation why the queues are empty).

Daniel

v1:
- https://lore.kernel.org/linux-nvme/[email protected]/
v2:
- https://lore.kernel.org/linux-nvme/[email protected]/
- reviewed tags collected
- added 'update hardware queues' for all transport
- added fix for fc hanger in nvme_wait_freeze_timeout
v3:
- https://lore.kernel.org/linux-nvme/[email protected]/
- dropped 'nvme-fc: Freeze queues before destroying them'
- added James' two patches
v4:
- https://lore.kernel.org/linux-nvme/[email protected]/
- added 'nvme-*: Unfreeze queues on reconnect'
- added Hannes' reviewed tags
v5:
- https://lore.kernel.org/linux-nvme/[email protected]/
- dropped non nvme-fc patches
- updated 'nvme-fc: fix controller reset hang during traffic'
v6:
- https://lore.kernel.org/linux-nvme/[email protected]/
- updated commit message 'nvme-fc: Update hardware queues before using them'
- dropped 'nvme-fc: fix controller reset hang during traffic'
- added 'nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues'
v7:
- updated commit id in last patch
- added explenation why the queues are empty in last patch

Daniel Wagner (1):
nvme-fc: Update hardware queues before using them

James Smart (2):
nvme-fc: avoid race between time out and tear down
nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues

 drivers/nvme/host/fc.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)
Signature · 7 lines
2.29.2


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
#1 lore raw

PATCH v7 1/3 nvme-fc: Update hardware queues before using them

In case the number of hardware queues changes, we need to update the tagset and the mapping of ctx to hctx first.

If we try to create and connect the I/O queues first, this operation will fail (target will reject the connect call due to the wrong number of queues) and hence we bail out of the recreate function. Then we will to try the very same operation again, thus we don't make any progress.

---
 drivers/nvme/host/fc.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
drivers/nvme/host/fc.c +8 -8
index b08a61ca283f..b5d9a5507de5 100644
@@ -2951,14 +2951,6 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 	if (ctrl->ctrl.queue_count == 1)
 		return 0;
 
-	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
-	if (ret)
-		goto out_free_io_queues;
-
-	ret = nvme_fc_connect_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
-	if (ret)
-		goto out_delete_hw_queues;
-
 	if (prior_ioq_cnt != nr_io_queues) {
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
@@ -2968,6 +2960,14 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		nvme_unfreeze(&ctrl->ctrl);
 	}
 
+	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
+	if (ret)
+		goto out_free_io_queues;
+
+	ret = nvme_fc_connect_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
+	if (ret)
+		goto out_delete_hw_queues;
+
 	return 0;
 
 out_delete_hw_queues:
Signature · 1 line
2.29.2
Reviewed-by on v7 1/3
#2 lore raw

Re: [PATCH v7 1/3] nvme-fc: Update hardware queues before using them

Quoted 54 lines · Daniel Wagner

In case the number of hardware queues changes, we need to update the tagset and the mapping of ctx to hctx first.

If we try to create and connect the I/O queues first, this operation will fail (target will reject the connect call due to the wrong number of queues) and hence we bail out of the recreate function. Then we will to try the very same operation again, thus we don't make any progress.

---
drivers/nvme/host/fc.c | 16 ++++++++--------

1 file changed, 8 insertions(+), 8 deletions(-)
drivers/nvme/host/fc.c +8 -8
index b08a61ca283f..b5d9a5507de5 100644
@@ -2951,14 +2951,6 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 	if (ctrl->ctrl.queue_count == 1)
 		return 0;
 
-	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
-	if (ret)
-		goto out_free_io_queues;
-
-	ret = nvme_fc_connect_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
-	if (ret)
-		goto out_delete_hw_queues;
-
 	if (prior_ioq_cnt != nr_io_queues) {
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
@@ -2968,6 +2960,14 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		nvme_unfreeze(&ctrl->ctrl);
 	}
 
+	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
+	if (ret)
+		goto out_free_io_queues;
+
+	ret = nvme_fc_connect_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
+	if (ret)
+		goto out_delete_hw_queues;
+
 	return 0;
 
Quoted 4 lines

out_delete_hw_queues:

Signature · 1 line
2.29.2

--
Himanshu Madhani Oracle Linux Engineering

#3 lore raw

PATCH v7 2/3 nvme-fc: avoid race between time out and tear down

In reply to Daniel Wagner

From: James Smart <[email protected]>

To avoid race between time out and tear down, in tear down process, first we quiesce the queue, and then delete the timer and cancel the time out work for the queue.

This patch merges the admin and io sync ops into the queue teardown logic as shown in the RDMA patch 3017013dcc "nvme-rdma: avoid race between time out and tear down". There is no teardown_lock in nvme-fc.

---
 drivers/nvme/host/fc.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/nvme/host/fc.c +2 -0
index b5d9a5507de5..6ebe68396712 100644
@@ -2487,6 +2487,7 @@ __nvme_fc_abort_outstanding_ios(struct nvme_fc_ctrl *ctrl, bool start_queues)
 	 */
 	if (ctrl->ctrl.queue_count > 1) {
 		nvme_stop_queues(&ctrl->ctrl);
+		nvme_sync_io_queues(&ctrl->ctrl);
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
 		blk_mq_tagset_wait_completed_request(&ctrl->tag_set);
@@ -2510,6 +2511,7 @@ __nvme_fc_abort_outstanding_ios(struct nvme_fc_ctrl *ctrl, bool start_queues)
 	 * clean up the admin queue. Same thing as above.
 	 */
 	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
+	blk_sync_queue(ctrl->ctrl.admin_q);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
 	blk_mq_tagset_wait_completed_request(&ctrl->admin_tag_set);
Signature · 1 line
2.29.2
Reviewed-by on v7 2/3
#4 lore raw

Re: [PATCH v7 2/3] nvme-fc: avoid race between time out and tear down

Quoted 46 lines · Daniel Wagner

From: James Smart <[email protected]>

To avoid race between time out and tear down, in tear down process, first we quiesce the queue, and then delete the timer and cancel the time out work for the queue.

This patch merges the admin and io sync ops into the queue teardown logic as shown in the RDMA patch 3017013dcc "nvme-rdma: avoid race between time out and tear down". There is no teardown_lock in nvme-fc.

---
drivers/nvme/host/fc.c | 2 ++

1 file changed, 2 insertions(+)
drivers/nvme/host/fc.c +2 -0
index b5d9a5507de5..6ebe68396712 100644
@@ -2487,6 +2487,7 @@ __nvme_fc_abort_outstanding_ios(struct nvme_fc_ctrl *ctrl, bool start_queues)
 	 */
 	if (ctrl->ctrl.queue_count > 1) {
 		nvme_stop_queues(&ctrl->ctrl);
+		nvme_sync_io_queues(&ctrl->ctrl);
 		blk_mq_tagset_busy_iter(&ctrl->tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
 		blk_mq_tagset_wait_completed_request(&ctrl->tag_set);
@@ -2510,6 +2511,7 @@ __nvme_fc_abort_outstanding_ios(struct nvme_fc_ctrl *ctrl, bool start_queues)
 	 * clean up the admin queue. Same thing as above.
 	 */
 	blk_mq_quiesce_queue(ctrl->ctrl.admin_q);
+	blk_sync_queue(ctrl->ctrl.admin_q);
 	blk_mq_tagset_busy_iter(&ctrl->admin_tag_set,
 				nvme_fc_terminate_exchange, &ctrl->ctrl);
 	blk_mq_tagset_wait_completed_request(&ctrl->admin_tag_set);
Signature · 1 line
2.29.2

--
Himanshu Madhani Oracle Linux Engineering

#5 lore raw

PATCH v7 3/3 nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues

In reply to Daniel Wagner

From: James Smart <[email protected]>

Remove the freeze/unfreeze around changes to the number of hardware queues. Study and retest has indicated there are no ios that can be active at this point so there is nothing to freeze.

nvme-fc is draining the queues in the shutdown and error recovery path in __nvme_fc_abort_outstanding_ios.

This patch primarily reverts 88e837ed0f1f "nvme-fc: wait for queues to freeze before calling update_hr_hw_queues". It's not an exact revert as it leaves the adjusting of hw queues only if the count changes.

[dwagner: added explanation why no IO is pending]
---
 drivers/nvme/host/fc.c | 2 --
 1 file changed, 2 deletions(-)
drivers/nvme/host/fc.c +0 -2
index 6ebe68396712..aa14ad963d91 100644
@@ -2957,9 +2957,7 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
 			prior_ioq_cnt, nr_io_queues);
-		nvme_wait_freeze(&ctrl->ctrl);
 		blk_mq_update_nr_hw_queues(&ctrl->tag_set, nr_io_queues);
-		nvme_unfreeze(&ctrl->ctrl);
 	}
 
 	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
Signature · 1 line
2.29.2
Reviewed-by on v7 3/3
#6 lore raw

Re: [PATCH v7 3/3] nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues

Quoted 26 lines · Daniel Wagner

From: James Smart <[email protected]>

Remove the freeze/unfreeze around changes to the number of hardware queues. Study and retest has indicated there are no ios that can be active at this point so there is nothing to freeze.

nvme-fc is draining the queues in the shutdown and error recovery path in __nvme_fc_abort_outstanding_ios.

This patch primarily reverts 88e837ed0f1f "nvme-fc: wait for queues to freeze before calling update_hr_hw_queues". It's not an exact revert as it leaves the adjusting of hw queues only if the count changes.

[dwagner: added explanation why no IO is pending]
---
 drivers/nvme/host/fc.c | 2 --
 1 file changed, 2 deletions(-)
drivers/nvme/host/fc.c +0 -0
index 6ebe68396712..aa14ad963d91 100644

drivers/nvme/host/fc.c nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl) · line 2963

Quoted 9 lines
drivers/nvme/host/fc.c +0 -2
@@ -2957,9 +2957,7 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
 			prior_ioq_cnt, nr_io_queues);
-		nvme_wait_freeze(&ctrl->ctrl);
 		blk_mq_update_nr_hw_queues(&ctrl->tag_set, nr_io_queues);
-		nvme_unfreeze(&ctrl->ctrl);
 	}
 
 	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);

Cheers,

Hannes

Signature · 4 lines
Dr. Hannes Reinecke		           Kernel Storage Architect
[email protected]			                  +49 911 74053 688
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), GF: Felix Imendörffer
Reviewed-by on v7 3/3
#7 lore raw

Re: [PATCH v7 3/3] nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues

In reply to Daniel Wagner
Quoted 41 lines · Daniel Wagner

From: James Smart <[email protected]>

Remove the freeze/unfreeze around changes to the number of hardware queues. Study and retest has indicated there are no ios that can be active at this point so there is nothing to freeze.

nvme-fc is draining the queues in the shutdown and error recovery path in __nvme_fc_abort_outstanding_ios.

This patch primarily reverts 88e837ed0f1f "nvme-fc: wait for queues to freeze before calling update_hr_hw_queues". It's not an exact revert as it leaves the adjusting of hw queues only if the count changes.

[dwagner: added explanation why no IO is pending]

---
drivers/nvme/host/fc.c | 2 --

1 file changed, 2 deletions(-)
drivers/nvme/host/fc.c +0 -2
index 6ebe68396712..aa14ad963d91 100644
@@ -2957,9 +2957,7 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
 			prior_ioq_cnt, nr_io_queues);
-		nvme_wait_freeze(&ctrl->ctrl);
 		blk_mq_update_nr_hw_queues(&ctrl->tag_set, nr_io_queues);
-		nvme_unfreeze(&ctrl->ctrl);
 	}
 
 	ret = nvme_fc_create_hw_io_queues(ctrl, ctrl->ctrl.sqsize + 1);
Signature · 1 line
2.29.2

Looks Good.

--
Himanshu Madhani Oracle Linux Engineering

Reviewed-by on v7 3/3
#8 lore raw

Re: [PATCH v7 3/3] nvme-fc: Remove freeze/unfreeze around update_nr_hw_queues

In reply to Daniel Wagner
Quoted 25 lines · Daniel Wagner

From: James Smart <[email protected]>

Remove the freeze/unfreeze around changes to the number of hardware queues. Study and retest has indicated there are no ios that can be active at this point so there is nothing to freeze.

nvme-fc is draining the queues in the shutdown and error recovery path in __nvme_fc_abort_outstanding_ios.

This patch primarily reverts 88e837ed0f1f "nvme-fc: wait for queues to freeze before calling update_hr_hw_queues". It's not an exact revert as it leaves the adjusting of hw queues only if the count changes.

[dwagner: added explanation why no IO is pending]
---
 drivers/nvme/host/fc.c | 2 --
 1 file changed, 2 deletions(-)
drivers/nvme/host/fc.c +0 -0
index 6ebe68396712..aa14ad963d91 100644

drivers/nvme/host/fc.c nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl) · line 2961

Quoted 7 lines
drivers/nvme/host/fc.c +0 -2
@@ -2957,9 +2957,7 @@ nvme_fc_recreate_io_queues(struct nvme_fc_ctrl *ctrl)
 		dev_info(ctrl->ctrl.device,
 			"reconnect: revising io queue count from %d to %d\n",
 			prior_ioq_cnt, nr_io_queues);
-		nvme_wait_freeze(&ctrl->ctrl);
 		blk_mq_update_nr_hw_queues(&ctrl->tag_set, nr_io_queues);
-		nvme_unfreeze(&ctrl->ctrl);
 	}

Both connect queue and ns queues are added to this tag_set, so blk_mq_update_nr_hw_queues() will freeze them all before updating nr_hw_queues.

Signature · 1 line
Ming