diff mbox series

mpi3mr: Return IOs to an unrecoverable HBA with DID_ERROR

Message ID 20220505184808.24049-1-sreekanth.reddy@broadcom.com
State New
Headers show
Series mpi3mr: Return IOs to an unrecoverable HBA with DID_ERROR | expand

Commit Message

Sreekanth Reddy May 5, 2022, 6:48 p.m. UTC
Complete all new I/O requests issued to an unrecoverable
controller with DID_ERROR status instead of returning the
I/O requests with SCSI_MLQUEUE_HOST_BUSY. This will prevent
the infinite retries of the new I/Os when a controller
is in the unrecoverable state.

Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
---
 drivers/scsi/mpi3mr/mpi3mr_os.c | 6 ++++++
 1 file changed, 6 insertions(+)

Comments

Martin K. Petersen May 17, 2022, 2:16 a.m. UTC | #1
On Fri, 6 May 2022 00:18:08 +0530, Sreekanth Reddy wrote:

> Set each SCSI device's default IO timeout and
> default error handling IO timeout to 60s.
> 
> 

Applied to 5.19/scsi-queue, thanks!

[1/1] mpi3mr: Increase IO timeout value to 60s
      https://git.kernel.org/mkp/scsi/c/1aa529d40025
diff mbox series

Patch

diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index 89a4918..5c57519 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -3996,6 +3996,12 @@  static int mpi3mr_qcmd(struct Scsi_Host *shost,
 	int iprio_class;
 	u8 is_pcie_dev = 0;
 
+	if (mrioc->unrecoverable) {
+		scmd->result = DID_ERROR << 16;
+		scsi_done(scmd);
+		goto out;
+	}
+
 	sdev_priv_data = scmd->device->hostdata;
 	if (!sdev_priv_data || !sdev_priv_data->tgt_priv_data) {
 		scmd->result = DID_NO_CONNECT << 16;