[PATCH v8 2/5] scsi: No retries on abort success
From: Muneendra <hidden>
Date: 2021-01-07 04:52:32
Subsystem:
scsi subsystem, the rest · Maintainers:
"James E.J. Bottomley", "Martin K. Petersen", Linus Torvalds
Added a new optional routine eh_should_retry_cmd in scsi_host_template that allows the transport to decide if a cmd is retryable.Return true if the transport is in a state the cmd should be retried on. Added a new interface scsi_eh_should_retry_cmd which checks and calls the new routine eh_should_retry_cmd. Made changes in scmd_eh_abort_handler and scsi_eh_flush_done_q which calls the scsi_eh_should_retry_cmd to check whether the command needs to be retried. The above changes were done based on a patch by Mike Christie. Reviewed-by: Himanshu Madhani <redacted> Reviewed-by: Ewan D. Milne <redacted> Reviewed-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Muneendra <redacted> --- v8: Rebased the patches on top of 5.11-rc2 v7: Added New routine in scsi_host_template to decide if a cmd is retryable instead of checking the same using SCMD_NORETRIES_ABORT bit as the cmd retry part can be checked by validating the port state. Moved the DID_TRANSPORT_MARGINAL changes to previous patch for reordering v6: Rearranged the patch by merging second hunk of the patch2 in v5 to this patch v5: added the DID_TRANSPORT_MARGINAL case to scsi_decide_disposition v4: Modified the comments in the code appropriately v3: Merged first part of the previous patch(v2 patch3) with this patch. v2: set the hostbyte as DID_TRANSPORT_MARGINAL instead of DID_TRANSPORT_FAILFAST. --- drivers/scsi/scsi_error.c | 17 +++++++++++++++-- include/scsi/scsi_host.h | 6 ++++++ 2 files changed, 21 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
index 28056ee498b3..1cdfa5a8ca09 100644
--- a/drivers/scsi/scsi_error.c
+++ b/drivers/scsi/scsi_error.c@@ -124,6 +124,17 @@ static bool scsi_cmd_retry_allowed(struct scsi_cmnd *cmd) return ++cmd->retries <= cmd->allowed; } +static bool scsi_eh_should_retry_cmd(struct scsi_cmnd *cmd) +{ + struct scsi_device *sdev = cmd->device; + struct Scsi_Host *host = sdev->host; + + if (host->hostt->eh_should_retry_cmd) + return host->hostt->eh_should_retry_cmd(cmd); + + return true; +} + /** * scmd_eh_abort_handler - Handle command aborts * @work: command to be aborted.
@@ -159,7 +170,8 @@ scmd_eh_abort_handler(struct work_struct *work) "eh timeout, not retrying " "aborted command\n")); } else if (!scsi_noretry_cmd(scmd) && - scsi_cmd_retry_allowed(scmd)) { + scsi_cmd_retry_allowed(scmd) && + scsi_eh_should_retry_cmd(scmd)) { SCSI_LOG_ERROR_RECOVERY(3, scmd_printk(KERN_WARNING, scmd, "retry aborted command\n"));
@@ -2111,7 +2123,8 @@ void scsi_eh_flush_done_q(struct list_head *done_q) list_for_each_entry_safe(scmd, next, done_q, eh_entry) { list_del_init(&scmd->eh_entry); if (scsi_device_online(scmd->device) && - !scsi_noretry_cmd(scmd) && scsi_cmd_retry_allowed(scmd)) { + !scsi_noretry_cmd(scmd) && scsi_cmd_retry_allowed(scmd) && + scsi_eh_should_retry_cmd(scmd)) { SCSI_LOG_ERROR_RECOVERY(3, scmd_printk(KERN_INFO, scmd, "%s: flush retry cmd\n",
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index 701f178b20ae..e30fd963b97d 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h@@ -314,6 +314,12 @@ struct scsi_host_template { * Status: OPTIONAL */ enum blk_eh_timer_return (*eh_timed_out)(struct scsi_cmnd *); + /* + * Optional routine that allows the transport to decide if a cmd + * is retryable. Return true if the transport is in a state the + * cmd should be retried on. + */ + bool (*eh_should_retry_cmd)(struct scsi_cmnd *scmd); /* This is an optional routine that allows transport to initiate * LLD adapter or firmware reset using sysfs attribute.
--
2.26.2