Skip to contentSkip to conversation

nvme_fc: fix double calls to nvme_cleanup_cmd()

PATCH New James Smart 4 participants 4 messages Jun 22, 2017 – Jun 27, 2017

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

PATCH 1/1 nvme_fc: fix double calls to nvme_cleanup_cmd()

Current fc transport code, on io termination, is calling nvme_cleanup_cmd() followed by the transport dma unmap routine which also calls nvme_cleanup_cmd(). Which means two kfrees occur on the same address, raising havoc. This resulted in odd data errors, effectively corruption..

Fix by removing the extraneous double calls. Call now occurs only in teardown paths and as part of dma unmap routine.

Signed-off-by James Smart <james.smart at broadcom.com>
---
 drivers/nvme/host/fc.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/nvme/host/fc.c +1 -4
index 5165007e86a6..02a6df84dc3c 100644
@@ -1957,10 +1957,8 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
 					queue->lldd_handle, &op->fcp_req);
 
 	if (ret) {
-		if (op->rq) {			/* normal request */
+		if (op->rq)			/* normal request */
 			nvme_fc_unmap_data(ctrl, op->rq, op);
-			nvme_cleanup_cmd(op->rq);
-		}
 		/* else - aen. no cleanup needed */
 
 		nvme_fc_ctrl_put(ctrl);
@@ -2078,7 +2076,6 @@ __nvme_fc_final_op_cleanup(struct request *rq)
 	op->flags &= ~(FCOP_FLAGS_TERMIO | FCOP_FLAGS_RELEASED |
 			FCOP_FLAGS_COMPLETE);
 
-	nvme_cleanup_cmd(rq);
 	nvme_fc_unmap_data(ctrl, rq, op);
 	nvme_complete_rq(rq);
 	nvme_fc_ctrl_put(ctrl);
Signature · 1 line
2.11.0
#1 lore raw

PATCH 1/1 nvme_fc: fix double calls to nvme_cleanup_cmd()

Quoted 32 lines · James Smart

Current fc transport code, on io termination, is calling nvme_cleanup_cmd() followed by the transport dma unmap routine which also calls nvme_cleanup_cmd(). Which means two kfrees occur on the same address, raising havoc. This resulted in odd data errors, effectively corruption..

Fix by removing the extraneous double calls. Call now occurs only in teardown paths and as part of dma unmap routine.

Signed-off-by James Smart <james.smart at broadcom.com>
---
 drivers/nvme/host/fc.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/nvme/host/fc.c +1 -3
index 5165007e86a6..02a6df84dc3c 100644
@@ -1957,10 +1957,8 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
 					queue->lldd_handle, &op->fcp_req);
 
 	if (ret) {
-		if (op->rq) {			/* normal request */
+		if (op->rq)			/* normal request */
 			nvme_fc_unmap_data(ctrl, op->rq, op);
-			nvme_cleanup_cmd(op->rq);
-		}
 		/* else - aen. no cleanup needed */
 
 		nvme_fc_ctrl_put(ctrl);

drivers/nvme/host/fc.c __nvme_fc_final_op_cleanup(struct request *rq) · line 2081

Quoted 7 lines
drivers/nvme/host/fc.c +0 -1
@@ -2078,7 +2076,6 @@ __nvme_fc_final_op_cleanup(struct request *rq)
 	op->flags &= ~(FCOP_FLAGS_TERMIO | FCOP_FLAGS_RELEASED |
 			FCOP_FLAGS_COMPLETE);
 
-	nvme_cleanup_cmd(rq);
 	nvme_fc_unmap_data(ctrl, rq, op);
 	nvme_complete_rq(rq);
 	nvme_fc_ctrl_put(ctrl);
Reviewed-by Hannes Reinecke <hare at suse.com>

Cheers,

Hannes

Signature · 5 lines
Dr. Hannes Reinecke		   Teamlead Storage & Networking
hare at suse.de			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 N?rnberg
GF: F. Imend?rffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG N?rnberg)
#2 lore raw

PATCH 1/1 nvme_fc: fix double calls to nvme_cleanup_cmd()

In reply to James Smart
Quoted 31 lines · James Smart

Current fc transport code, on io termination, is calling nvme_cleanup_cmd() followed by the transport dma unmap routine which also calls nvme_cleanup_cmd(). Which means two kfrees occur on the same address, raising havoc. This resulted in odd data errors, effectively corruption..

Fix by removing the extraneous double calls. Call now occurs only in teardown paths and as part of dma unmap routine.

Signed-off-by James Smart <james.smart at broadcom.com>
---
 drivers/nvme/host/fc.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/nvme/host/fc.c +1 -3
index 5165007e86a6..02a6df84dc3c 100644
@@ -1957,10 +1957,8 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
 					queue->lldd_handle, &op->fcp_req);
 
 	if (ret) {
-		if (op->rq) {			/* normal request */
+		if (op->rq)			/* normal request */
 			nvme_fc_unmap_data(ctrl, op->rq, op);
-			nvme_cleanup_cmd(op->rq);
-		}
 		/* else - aen. no cleanup needed */
 
 		nvme_fc_ctrl_put(ctrl);

drivers/nvme/host/fc.c __nvme_fc_final_op_cleanup(struct request *rq) · line 2081

Quoted 7 lines
drivers/nvme/host/fc.c +0 -1
@@ -2078,7 +2076,6 @@ __nvme_fc_final_op_cleanup(struct request *rq)
 	op->flags &= ~(FCOP_FLAGS_TERMIO | FCOP_FLAGS_RELEASED |
 			FCOP_FLAGS_COMPLETE);
 
-	nvme_cleanup_cmd(rq);
 	nvme_fc_unmap_data(ctrl, rq, op);
 	nvme_complete_rq(rq);
 	nvme_fc_ctrl_put(ctrl);

I had posted a patch for this back in April, but it doesn't look like it ever made it in. I had fixed it by removing the nvme_cleanup_cmd() call from nvme_fc_unmap_data(), but this looks OK too.

Reviewed-by Ewan D. Milne <emilne at redhat.com>