scsi: smartpqi: Reporting unhandled SCSI errors

When a HARDWARE_ERROR is triggered for ASC=0x3e, the existing code is only
considering the case where ASCQ=0x1.

According to the http://www.t10.org/lists/asc-num.htm#ASC_3E specification,
other values may occur like a timeout (ASCQ=0x2).

This patch prints an error message when a non-handled message is received.
This can help diagnose a possible misbehavior of the controller or a
missing implementation in the Linux kernel.

This patch keeps the exact same error handling but prints a message if an
ASCQ != 1 is reported.

[mkp: clarified commit message]

Signed-off-by: Erwan Velu <e.velu@criteo.com>
Acked-by: Don Brace <don.brace@microsemi.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c
index c26cac819f9e7a2ceaf9638fdebd600a7639f21f..b17761eafca91eeb8ff68126567f46f4942184f6 100644 (file)
--- a/drivers/scsi/smartpqi/smartpqi_init.c
+++ b/drivers/scsi/smartpqi/smartpqi_init.c
@@ -2755,16 +2755,25 @@ static void pqi_process_raid_io_error(struct pqi_io_request *io_request)
                        scsi_normalize_sense(error_info->data,
                                sense_data_length, &sshdr) &&
                                sshdr.sense_key == HARDWARE_ERROR &&
-                               sshdr.asc == 0x3e &&
-                               sshdr.ascq == 0x1) {
+                               sshdr.asc == 0x3e) {
                        struct pqi_ctrl_info *ctrl_info = shost_to_hba(scmd->device->host);
                        struct pqi_scsi_dev *device = scmd->device->hostdata;
 
-                       if (printk_ratelimit())
-                               scmd_printk(KERN_ERR, scmd, "received 'logical unit failure' from controller for scsi %d:%d:%d:%d\n",
-                                       ctrl_info->scsi_host->host_no, device->bus, device->target, device->lun);
-                       pqi_take_device_offline(scmd->device, "RAID");
-                       host_byte = DID_NO_CONNECT;
+                       switch (sshdr.ascq) {
+                       case 0x1: /* LOGICAL UNIT FAILURE */
+                               if (printk_ratelimit())
+                                       scmd_printk(KERN_ERR, scmd, "received 'logical unit failure' from controller for scsi %d:%d:%d:%d\n",
+                                               ctrl_info->scsi_host->host_no, device->bus, device->target, device->lun);
+                               pqi_take_device_offline(scmd->device, "RAID");
+                               host_byte = DID_NO_CONNECT;
+                               break;
+
+                       default: /* See http://www.t10.org/lists/asc-num.htm#ASC_3E */
+                               if (printk_ratelimit())
+                                       scmd_printk(KERN_ERR, scmd, "received unhandled error %d from controller for scsi %d:%d:%d:%d\n",
+                                               sshdr.ascq, ctrl_info->scsi_host->host_no, device->bus, device->target, device->lun);
+                               break;
+                       }
                }
 
                if (sense_data_length > SCSI_SENSE_BUFFERSIZE)