scsi: core: Allow passthrough to request midlayer retries

For passthrough we don't retry any error which we get a check condition
for. This results in a lot of callers driving their own retries for all
UAs, specific UAs, NOT_READY, specific sense values or any type of failure.

This adds the core code to allow passthrough users to specify what errors
they want the SCSI midlayer to retry for them. We can then convert users to
drop a lot of their sense parsing and retry handling.

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20240123002220.129141-2-michael.christie@oracle.com
Reviewed-by: John Garry <john.g.garry@oracle.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 95 insertions, 3 deletions

diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index cf3864f72093..16a33a558842 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -184,6 +184,92 @@ void scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
 	__scsi_queue_insert(cmd, reason, true);
 }
 
+void scsi_failures_reset_retries(struct scsi_failures *failures)
+{
+	struct scsi_failure *failure;
+
+	failures->total_retries = 0;
+
+	for (failure = failures->failure_definitions; failure->result;
+	     failure++)
+		failure->retries = 0;
+}
+EXPORT_SYMBOL_GPL(scsi_failures_reset_retries);
+
+/**
+ * scsi_check_passthrough - Determine if passthrough scsi_cmnd needs a retry.
+ * @scmd: scsi_cmnd to check.
+ * @failures: scsi_failures struct that lists failures to check for.
+ *
+ * Returns -EAGAIN if the caller should retry else 0.
+ */
+static int scsi_check_passthrough(struct scsi_cmnd *scmd,
+				  struct scsi_failures *failures)
+{
+	struct scsi_failure *failure;
+	struct scsi_sense_hdr sshdr;
+	enum sam_status status;
+
+	if (!failures)
+		return 0;
+
+	for (failure = failures->failure_definitions; failure->result;
+	     failure++) {
+		if (failure->result == SCMD_FAILURE_RESULT_ANY)
+			goto maybe_retry;
+
+		if (host_byte(scmd->result) &&
+		    host_byte(scmd->result) == host_byte(failure->result))
+			goto maybe_retry;
+
+		status = status_byte(scmd->result);
+		if (!status)
+			continue;
+
+		if (failure->result == SCMD_FAILURE_STAT_ANY &&
+		    !scsi_status_is_good(scmd->result))
+			goto maybe_retry;
+
+		if (status != status_byte(failure->result))
+			continue;
+
+		if (status_byte(failure->result) != SAM_STAT_CHECK_CONDITION ||
+		    failure->sense == SCMD_FAILURE_SENSE_ANY)
+			goto maybe_retry;
+
+		if (!scsi_command_normalize_sense(scmd, &sshdr))
+			return 0;
+
+		if (failure->sense != sshdr.sense_key)
+			continue;
+
+		if (failure->asc == SCMD_FAILURE_ASC_ANY)
+			goto maybe_retry;
+
+		if (failure->asc != sshdr.asc)
+			continue;
+
+		if (failure->ascq == SCMD_FAILURE_ASCQ_ANY ||
+		    failure->ascq == sshdr.ascq)
+			goto maybe_retry;
+	}
+
+	return 0;
+
+maybe_retry:
+	if (failure->allowed) {
+		if (failure->allowed == SCMD_FAILURE_NO_LIMIT ||
+		    ++failure->retries <= failure->allowed)
+			return -EAGAIN;
+	} else {
+		if (failures->total_allowed == SCMD_FAILURE_NO_LIMIT ||
+		    ++failures->total_retries <= failures->total_allowed)
+			return -EAGAIN;
+	}
+
+	return 0;
+}
+
 /**
  * scsi_execute_cmd - insert request and wait for the result
  * @sdev:	scsi_device
@@ -192,7 +278,7 @@ void scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
  * @buffer:	data buffer
  * @bufflen:	len of buffer
  * @timeout:	request timeout in HZ
- * @retries:	number of times to retry request
+ * @ml_retries:	number of times SCSI midlayer will retry request
  * @args:	Optional args. See struct definition for field descriptions
  *
  * Returns the scsi_cmnd result field if a command was executed, or a negative
@@ -200,7 +286,7 @@ void scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
  */
 int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd,
 		     blk_opf_t opf, void *buffer, unsigned int bufflen,
-		     int timeout, int retries,
+		     int timeout, int ml_retries,
 		     const struct scsi_exec_args *args)
 {
 	static const struct scsi_exec_args default_args;
@@ -214,6 +300,7 @@ int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd,
 			      args->sense_len != SCSI_SENSE_BUFFERSIZE))
 		return -EINVAL;
 
+retry:
 	req = scsi_alloc_request(sdev->request_queue, opf, args->req_flags);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
@@ -227,7 +314,7 @@ int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd,
 	scmd = blk_mq_rq_to_pdu(req);
 	scmd->cmd_len = COMMAND_SIZE(cmd[0]);
 	memcpy(scmd->cmnd, cmd, scmd->cmd_len);
-	scmd->allowed = retries;
+	scmd->allowed = ml_retries;
 	scmd->flags |= args->scmd_flags;
 	req->timeout = timeout;
 	req->rq_flags |= RQF_QUIET;
@@ -237,6 +324,11 @@ int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd,
 	 */
 	blk_execute_rq(req, true);
 
+	if (scsi_check_passthrough(scmd, args->failures) == -EAGAIN) {
+		blk_mq_free_request(req);
+		goto retry;
+	}
+
 	/*
 	 * Some devices (USB mass-storage in particular) may transfer
 	 * garbage data together with a residue indicating that the data