diff mbox series

[for-rc,v3] RDMA/bnxt_re: Fix error recovery sequence

Message ID 20241231025008.2267162-1-kalesh-anakkur.purayil@broadcom.com (mailing list archive)
State Accepted
Headers show
Series [for-rc,v3] RDMA/bnxt_re: Fix error recovery sequence | expand

Commit Message

Kalesh Anakkur Purayil Dec. 31, 2024, 2:50 a.m. UTC 
Fixed to return ENXIO from __send_message_basic_sanity()
to indicate that device is in error state. In the case of
ERR_DEVICE_DETACHED state, the driver should not post the
commands to the firmware as it will time out eventually.

Removed bnxt_re_modify_qp() call from bnxt_re_dev_stop()
as it is a no-op.

Fixes: cc5b9b48d447 ("RDMA/bnxt_re: Recover the device when FW error is detected")
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Selvin Xavier <selvin.xavier@broadcom.com>
---
V3: As suggested by Leon, removed pci_channel_offline() check from this patch.
    This will be reviewed further and push a separate patch.
V2: No changes since v1 and is just a resend.
V1: https://patchwork.kernel.org/project/linux-rdma/patch/20241204075416.478431-5-kalesh-anakkur.purayil@broadcom.com/
---
 drivers/infiniband/hw/bnxt_re/main.c       | 8 +-------
 drivers/infiniband/hw/bnxt_re/qplib_rcfw.c | 5 +++--
 2 files changed, 4 insertions(+), 9 deletions(-)

Comments

Leon Romanovsky Dec. 31, 2024, 1:28 p.m. UTC | #1 
On Tue, 31 Dec 2024 08:20:08 +0530, Kalesh AP wrote:
> Fixed to return ENXIO from __send_message_basic_sanity()
> to indicate that device is in error state. In the case of
> ERR_DEVICE_DETACHED state, the driver should not post the
> commands to the firmware as it will time out eventually.
> 
> Removed bnxt_re_modify_qp() call from bnxt_re_dev_stop()
> as it is a no-op.
> 
> [...]

Applied, thanks!

[1/1] RDMA/bnxt_re: Fix error recovery sequence
      https://git.kernel.org/rdma/rdma/c/e6178bf78d0378

Best regards,
diff mbox series

Patch

diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index b7af0d5ff3b6..c143f273b759 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -1715,11 +1715,8 @@  static bool bnxt_re_is_qp1_or_shadow_qp(struct bnxt_re_dev *rdev,
 
 static void bnxt_re_dev_stop(struct bnxt_re_dev *rdev)
 {
-	int mask = IB_QP_STATE;
-	struct ib_qp_attr qp_attr;
 	struct bnxt_re_qp *qp;
 
-	qp_attr.qp_state = IB_QPS_ERR;
 	mutex_lock(&rdev->qp_lock);
 	list_for_each_entry(qp, &rdev->qp_list, list) {
 		/* Modify the state of all QPs except QP1/Shadow QP */
@@ -1727,12 +1724,9 @@  static void bnxt_re_dev_stop(struct bnxt_re_dev *rdev)
 			if (qp->qplib_qp.state !=
 			    CMDQ_MODIFY_QP_NEW_STATE_RESET &&
 			    qp->qplib_qp.state !=
-			    CMDQ_MODIFY_QP_NEW_STATE_ERR) {
+			    CMDQ_MODIFY_QP_NEW_STATE_ERR)
 				bnxt_re_dispatch_event(&rdev->ibdev, &qp->ib_qp,
 						       1, IB_EVENT_QP_FATAL);
-				bnxt_re_modify_qp(&qp->ib_qp, &qp_attr, mask,
-						  NULL);
-			}
 		}
 	}
 	mutex_unlock(&rdev->qp_lock);
diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
index 5e90ea232de8..17e62f22683b 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
+++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
@@ -424,7 +424,8 @@  static int __send_message_basic_sanity(struct bnxt_qplib_rcfw *rcfw,
 
 	/* Prevent posting if f/w is not in a state to process */
 	if (test_bit(ERR_DEVICE_DETACHED, &rcfw->cmdq.flags))
-		return bnxt_qplib_map_rc(opcode);
+		return -ENXIO;
+
 	if (test_bit(FIRMWARE_STALL_DETECTED, &cmdq->flags))
 		return -ETIMEDOUT;
 
@@ -493,7 +494,7 @@  static int __bnxt_qplib_rcfw_send_message(struct bnxt_qplib_rcfw *rcfw,
 
 	rc = __send_message_basic_sanity(rcfw, msg, opcode);
 	if (rc)
-		return rc;
+		return rc == -ENXIO ? bnxt_qplib_map_rc(opcode) : rc;
 
 	rc = __send_message(rcfw, msg, opcode);
 	if (rc)