Skip to contentSkip to conversation

scsi: retry MODE SENSE on unit attention

PATCH In review Hannes Reinecke 2 participants 4 messages Jun 11, 2015 – Jun 12, 2015 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi: retry MODE SENSE on unit attention

The 'sd' driver is calling scsi_mode_sense() to figure out internal details. But scsi_mode_sense() never checks for any pending unit attentions, so we're getting annoying error messages like:

MODE SENSE: unimplemented page/subpage: 0x00/0x00

and a possible wrong decision for device cache handling.

---
 drivers/scsi/scsi_lib.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/scsi/scsi_lib.c +6 -1
index 2428d96..d7915c8 100644
@@ -2423,7 +2423,7 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 	unsigned char cmd[12];
 	int use_10_for_ms;
 	int header_length;
-	int result;
+	int result, retry_count = retries;
 	struct scsi_sense_hdr my_sshdr;
 
 	memset(data, 0, sizeof(*data));
@@ -2502,6 +2502,11 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 			data->block_descriptor_length = buffer[3];
 		}
 		data->header_length = header_length;
+	} else if ((status_byte(result) == CHECK_CONDITION) &&
+		   scsi_sense_valid(sshdr) &&
+		   sshdr->sense_key == UNIT_ATTENTION && retry_count) {
+		retry_count--;
+		goto retry;
 	}
 
 	return result;
Signature · 1 line
1.8.5.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 29 lines · Hannes Reinecke

The 'sd' driver is calling scsi_mode_sense() to figure out internal details. But scsi_mode_sense() never checks for any pending unit attentions, so we're getting annoying error messages like:

MODE SENSE: unimplemented page/subpage: 0x00/0x00

and a possible wrong decision for device cache handling.

---
 drivers/scsi/scsi_lib.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/scsi/scsi_lib.c +1 -1
index 2428d96..d7915c8 100644
@@ -2423,7 +2423,7 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 	unsigned char cmd[12];
 	int use_10_for_ms;
 	int header_length;
-	int result;
+	int result, retry_count = retries;
 	struct scsi_sense_hdr my_sshdr;
 
 	memset(data, 0, sizeof(*data));
@@ -2502,6 +2502,11 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 			data->block_descriptor_length = buffer[3];

drivers/scsi/scsi_lib.c scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, · line 2512

Quoted 10 lines
drivers/scsi/scsi_lib.c +5 -0
 		}
 		data->header_length = header_length;
+	} else if ((status_byte(result) == CHECK_CONDITION) &&
+		   scsi_sense_valid(sshdr) &&
+		   sshdr->sense_key == UNIT_ATTENTION && retry_count) {
+		retry_count--;
+		goto retry;
 	}
 
 	return result;

Great, but shouldn't we be doing this more generally? What about scsi_mode_select()?

(And, with the number of status changes that can get reported by
UAs, we might want to think about increasing the retry count on these commands up from 3 at some point.)

#2 lore raw
Quoted 39 lines · Ewan Milne
Quoted 38 lines · Hannes Reinecke

The 'sd' driver is calling scsi_mode_sense() to figure out internal details. But scsi_mode_sense() never checks for any pending unit attentions, so we're getting annoying error messages like:

MODE SENSE: unimplemented page/subpage: 0x00/0x00

and a possible wrong decision for device cache handling.

---
 drivers/scsi/scsi_lib.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/scsi/scsi_lib.c +6 -1
index 2428d96..d7915c8 100644
@@ -2423,7 +2423,7 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 	unsigned char cmd[12];
 	int use_10_for_ms;
 	int header_length;
-	int result;
+	int result, retry_count = retries;
 	struct scsi_sense_hdr my_sshdr;
 
 	memset(data, 0, sizeof(*data));
@@ -2502,6 +2502,11 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 			data->block_descriptor_length = buffer[3];
 		}
 		data->header_length = header_length;
+	} else if ((status_byte(result) == CHECK_CONDITION) &&
+		   scsi_sense_valid(sshdr) &&
+		   sshdr->sense_key == UNIT_ATTENTION && retry_count) {
+		retry_count--;
+		goto retry;
 	}
 
Quoted 5 lines

drivers/scsi/scsi_lib.c scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, · line 2512

Quoted 1 line
drivers/scsi/scsi_lib.c +0 -0
 	return result;

Great, but shouldn't we be doing this more generally? What about scsi_mode_select()?

I haven't seen any issues with scsi_mode_select() as of now, so I didn't do anything about this :-)

Quoted 4 lines

(And, with the number of status changes that can get reported by
UAs, we might want to think about increasing the retry count on these commands up from 3 at some point.)

Hmm. _Actually_, we're not getting _more_ UAs (neither the number nor the situation at which UAs are being send has changed). It's just that we're trying to _use_ UAs so these things pop up. But yeah, raising the number or retries to eg 5 is probably a good idea.

Quoted 2 lines

Cheers,

Hannes

Signature · 9 lines
Dr. Hannes Reinecke		               zSeries & Storage
[email protected]			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#3 lore raw
Quoted 48 lines · Hannes Reinecke
Quoted 39 lines · Ewan Milne
Quoted 38 lines · Hannes Reinecke

The 'sd' driver is calling scsi_mode_sense() to figure out internal details. But scsi_mode_sense() never checks for any pending unit attentions, so we're getting annoying error messages like:

MODE SENSE: unimplemented page/subpage: 0x00/0x00

and a possible wrong decision for device cache handling.

---
 drivers/scsi/scsi_lib.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/scsi/scsi_lib.c +6 -1
index 2428d96..d7915c8 100644
@@ -2423,7 +2423,7 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 	unsigned char cmd[12];
 	int use_10_for_ms;
 	int header_length;
-	int result;
+	int result, retry_count = retries;
 	struct scsi_sense_hdr my_sshdr;
 
 	memset(data, 0, sizeof(*data));
@@ -2502,6 +2502,11 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
 			data->block_descriptor_length = buffer[3];
 		}
 		data->header_length = header_length;
+	} else if ((status_byte(result) == CHECK_CONDITION) &&
+		   scsi_sense_valid(sshdr) &&
+		   sshdr->sense_key == UNIT_ATTENTION && retry_count) {
+		retry_count--;
+		goto retry;
 	}
 
Quoted 5 lines

drivers/scsi/scsi_lib.c scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, · line 2512

Quoted 1 line
drivers/scsi/scsi_lib.c +0 -0
 	return result;

Great, but shouldn't we be doing this more generally? What about scsi_mode_select()?

I haven't seen any issues with scsi_mode_select() as of now, so I didn't do anything about this :-)

Quoted 8 lines
Quoted 4 lines

(And, with the number of status changes that can get reported by
UAs, we might want to think about increasing the retry count on these commands up from 3 at some point.)

Hmm. _Actually_, we're not getting _more_ UAs (neither the number nor the situation at which UAs are being send has changed). It's just that we're trying to _use_ UAs so these things pop up. But yeah, raising the number or retries to eg 5 is probably a good idea.

Yeah, or maybe add a time limit to we don't drag out the boot time. We can do this later, though, I'm fine with your patch as it is.

Quoted 6 lines
Quoted 3 lines

Cheers,

Hannes