Skip to contentSkip to conversation

nvme: fix identify error status silent ignore

PATCH Applied Sagi Grimberg 3 participants 5 messages Jun 26, 2020 – Jun 29, 2020 NVMe

lore
Share & tools
lore

Sagi Grimberg

author maintainer
#0 lore raw

PATCH v3 1/1 nvme: fix identify error status silent ignore

Patch 59c7c3caaaf8 intended to only silently ignore non retry-able errors (DNR bit set) such that we can still identify misbehaving controllers, and in the other hand propagate retry-able errors (DNR bit cleared) so we don't wrongly abandon a namespace just because it happens to be temporarily inaccessible.

The goal remains the same as the original commit where this was introduced but unfortunately had the logic backwards.

Fixes 59c7c3caaaf8 ("nvme: fix possible hang when ns

scanning fails during error recovery")

---
Changes from v2:
- added comment on non-trivial code

Changes from v1:
- remove paranthesis

 drivers/nvme/host/core.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/nvme/host/core.c +8 -2
index 2afed32d3892..92dc2327bf3a 100644
@@ -1128,9 +1128,15 @@ static int nvme_identify_ns_descs(struct nvme_ctrl *ctrl, unsigned nsid,
 			"Identify Descriptors failed (%d)\n", status);
 		 /*
 		  * Don't treat an error as fatal, as we potentially already
-		  * have a NGUID or EUI-64.
+		  * have a NGUID or EUI-64. If we failed with DNR set, we want
+		  * to silently ignore the error as we can still identify
+		  * the device, but if the status has DNR set, we want
+		  * to propogate the error back specifically for the disk
+		  * revalidation flow to make sure we don't abandon the
+		  * device just because of a temporal retry-able error (such
+		  * as path of transport errors).
 		  */
-		if (status > 0 && !(status & NVME_SC_DNR))
+		if (status > 0 && status & NVME_SC_DNR)
 			status = 0;
 		goto free_data;
 	}
Signature · 7 lines
2.25.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
#1 lore raw
Quoted 9 lines · Sagi Grimberg

Patch 59c7c3caaaf8 intended to only silently ignore non retry-able errors (DNR bit set) such that we can still identify misbehaving controllers, and in the other hand propagate retry-able errors (DNR bit cleared) so we don't wrongly abandon a namespace just because it happens to be temporarily inaccessible.

The goal remains the same as the original commit where this was introduced but unfortunately had the logic backwards.

So how did this work all the time? I'm really worried that someone this is actually going to break things. What do you think about only queueing it up for 5.9 for now?

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author maintainer
#2 lore raw
Quoted 8 lines
Quoted 8 lines

Patch 59c7c3caaaf8 intended to only silently ignore non retry-able errors (DNR bit set) such that we can still identify misbehaving controllers, and in the other hand propagate retry-able errors (DNR bit cleared) so we don't wrongly abandon a namespace just because it happens to be temporarily inaccessible.

Quoted 6 lines
Quoted 2 lines

The goal remains the same as the original commit where this was introduced but unfortunately had the logic backwards.

So how did this work all the time? I'm really worried that someone this is actually going to break things. What do you think about only queueing it up for 5.9 for now?

The code is broken now, the original patch was a mistake. It didn't fix the scan/error-recovery race, but also broke the ignore we had for pci. So I think this should go to 5.8-rc and also be picked up for stable 5.7.

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

#4 lore raw
In reply to Sagi Grimberg
Quoted 30 lines · Sagi Grimberg

Patch 59c7c3caaaf8 intended to only silently ignore non retry-able errors (DNR bit set) such that we can still identify misbehaving controllers, and in the other hand propagate retry-able errors (DNR bit cleared) so we don't wrongly abandon a namespace just because it happens to be temporarily inaccessible.

The goal remains the same as the original commit where this was introduced but unfortunately had the logic backwards.

Fixes 59c7c3caaaf8 ("nvme: fix possible hang when ns

scanning fails during error recovery")

---
Changes from v2:
- added comment on non-trivial code

Changes from v1:
- remove paranthesis

 drivers/nvme/host/core.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/nvme/host/core.c +0 -0
index 2afed32d3892..92dc2327bf3a 100644

drivers/nvme/host/core.c static int nvme_identify_ns_descs(struct nvme_ctrl *ctrl, unsigned nsid, · line 1134

Quoted 8 lines
drivers/nvme/host/core.c +4 -1
@@ -1128,9 +1128,15 @@ static int nvme_identify_ns_descs(struct nvme_ctrl *ctrl, unsigned nsid,
 			"Identify Descriptors failed (%d)\n", status);
 		 /*
 		  * Don't treat an error as fatal, as we potentially already
-		  * have a NGUID or EUI-64.
+		  * have a NGUID or EUI-64. If we failed with DNR set, we want
+		  * to silently ignore the error as we can still identify
+		  * the device, but if the status has DNR set, we want
+		  * to propogate the error back specifically for the disk

s/propogate/propagate/

Perhaps this minor nit could be fixed up while applying?

Kind regards,
Niklas

Quoted 17 lines
drivers/nvme/host/core.c +4 -1
+		  * revalidation flow to make sure we don't abandon the
+		  * device just because of a temporal retry-able error (such
+		  * as path of transport errors).
 		  */
-		if (status > 0 && !(status & NVME_SC_DNR))
+		if (status > 0 && status & NVME_SC_DNR)
 			status = 0;
 		goto free_data;
 	}
Signature · 7 lines
2.25.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme