diff mbox series

[for-next,2/2] RDMA/bnxt_re: refactor code around bnxt_qplib_map_rc

Message ID 20230616061700.741769-2-kashyap.desai@broadcom.com (mailing list archive)
State Accepted
Delegated to: Jason Gunthorpe
Headers show
Series [for-next,1/2] RDMA/bnxt_re: Remove incorrect return check from slow path | expand

Commit Message

Kashyap Desai June 16, 2023, 6:17 a.m. UTC 
Updated function comment of bnxt_qplib_map_rc
Removed intermediate return value ENXIO and directly called bnxt_qplib_map_rc
from __send_message_basic_sanity.

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
---
 drivers/infiniband/hw/bnxt_re/qplib_rcfw.c | 23 ++++++++++++++--------
 1 file changed, 15 insertions(+), 8 deletions(-)

Comments

Dan Carpenter June 16, 2023, 7:04 a.m. UTC | #1 
On Fri, Jun 16, 2023 at 11:47:00AM +0530, Kashyap Desai wrote:
> Updated function comment of bnxt_qplib_map_rc
> Removed intermediate return value ENXIO and directly called bnxt_qplib_map_rc
> from __send_message_basic_sanity.
> 
> Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
> Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
> ---

Probably Reported-by is giving me to much credit but I do like this
change.  It's more clear now.  Thanks, a lot!

regards,
dan carpenter
diff mbox series

Patch

diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
index 30c6e865d691..a8323054cfee 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
+++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c
@@ -57,13 +57,20 @@  static void bnxt_qplib_service_creq(struct tasklet_struct *t);
  * bnxt_qplib_map_rc  -  map return type based on opcode
  * @opcode    -  roce slow path opcode
  *
- * In some cases like firmware halt is detected, the driver is supposed to
- * remap the error code of the timed out command.
+ * case #1
+ * Firmware initiated error recovery is a safe state machine and
+ * driver can consider all the underlying rdma resources are free.
+ * In this state, it is safe to return success for opcodes related to
+ * destroying rdma resources (like destroy qp, destroy cq etc.).
  *
- * It is not safe to assume hardware is really inactive so certain opcodes
- * like destroy qp etc are not safe to be returned success, but this function
- * will be called when FW already reports a timeout. This would be possible
- * only when FW crashes and resets. This will clear all the HW resources.
+ * case #2
+ * If driver detect potential firmware stall, it is not safe state machine
+ * and the driver can not consider all the underlying rdma resources are
+ * freed.
+ * In this state, it is not safe to return success for opcodes related to
+ * destroying rdma resources (like destroy qp, destroy cq etc.).
+ *
+ * Scope of this helper function is only for case #1.
  *
  * Returns:
  * 0 to communicate success to caller.
@@ -418,7 +425,7 @@  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 -ENXIO;
+		return bnxt_qplib_map_rc(opcode);
 	if (test_bit(FIRMWARE_STALL_DETECTED, &cmdq->flags))
 		return -ETIMEDOUT;
 
@@ -488,7 +495,7 @@  static int __bnxt_qplib_rcfw_send_message(struct bnxt_qplib_rcfw *rcfw,
 
 	rc = __send_message_basic_sanity(rcfw, msg, opcode);
 	if (rc)
-		return rc == -ENXIO ? bnxt_qplib_map_rc(opcode) : rc;
+		return rc;
 
 	rc = __send_message(rcfw, msg);
 	if (rc)