Skip to contentSkip to conversation

Backport of bae818ee1577

PATCH New Ilya Dryomov 2 participants 4 messages Nov 1, 2015 – Nov 11, 2015 Stable branch

lore
Share & tools
lore
#0 lore raw

Hello,

Please apply the following backport of bae818ee1577 ("rbd: require stable pages if message data CRCs are enabled") to 3.9+.

I have two patches: one for 3.9-3.17 and another one for 3.18+. It really is a single backport (git cherry-pick / git am -3 can handle the remaining context just fine), but I know some of you use quilt for queueing up patches, so I'm sending both just in case.

Thanks,

Ilya

#1 lore raw

PATCH 1/1 rbd: require stable pages if message data CRCs are enabled

From: Ronny Hegewald <[email protected]>

rbd requires stable pages, as it performs a crc of the page data before they are send to the OSDs.

But since kernel 3.9 (patch 1d1d1a767206fbe5d4c69493b7e6d2a8d08cc0a0 "mm: only enforce stable page writes if the backing device requires it") it is not assumed anymore that block devices require stable pages.

This patch sets the necessary flag to get stable pages back for rbd.

In a ceph installation that provides multiple ext4 formatted rbd devices "bad crc" messages appeared regularly (ca 1 message every 1-2 minutes on every OSD that provided the data for the rbd) in the OSD-logs before this patch. After this patch this messages are pretty much gone (only ca 1-2 / month / OSD).

Cc [email protected] # 3.9+, needs backporting
[[email protected]: require stable pages only in crc case, changelog]
[[email protected]: backport to 3.9-3.17: context]
---
 drivers/block/rbd.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/block/rbd.c +3 -0
index 2ea515509ca6..ff0dd53554cf 100644
@@ -3443,6 +3443,9 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
 	blk_queue_io_opt(q, segment_size);
 
 	blk_queue_merge_bvec(q, rbd_merge_bvec);
+	if (!ceph_test_opt(rbd_dev->rbd_client->client, NOCRC))
+		q->backing_dev_info.capabilities |= BDI_CAP_STABLE_WRITES;
+
 	disk->queue = q;
 
 	q->queuedata = rbd_dev;
Signature · 1 line
2.4.3
#2 lore raw

PATCH 1/1 rbd: require stable pages if message data CRCs are enabled

In reply to Ilya Dryomov

From: Ronny Hegewald <[email protected]>

rbd requires stable pages, as it performs a crc of the page data before they are send to the OSDs.

But since kernel 3.9 (patch 1d1d1a767206fbe5d4c69493b7e6d2a8d08cc0a0 "mm: only enforce stable page writes if the backing device requires it") it is not assumed anymore that block devices require stable pages.

This patch sets the necessary flag to get stable pages back for rbd.

In a ceph installation that provides multiple ext4 formatted rbd devices "bad crc" messages appeared regularly (ca 1 message every 1-2 minutes on every OSD that provided the data for the rbd) in the OSD-logs before this patch. After this patch this messages are pretty much gone (only ca 1-2 / month / OSD).

Cc [email protected] # 3.9+, needs backporting
[[email protected]: require stable pages only in crc case, changelog]
[[email protected]: backport to 3.18-4.2: context]
---
 drivers/block/rbd.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/block/rbd.c +3 -0
index 324bf35ec4dd..77324e9a9221 100644
@@ -3819,6 +3819,9 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
 	q->limits.discard_zeroes_data = 1;
 
 	blk_queue_merge_bvec(q, rbd_merge_bvec);
+	if (!ceph_test_opt(rbd_dev->rbd_client->client, NOCRC))
+		q->backing_dev_info.capabilities |= BDI_CAP_STABLE_WRITES;
+
 	disk->queue = q;
 
 	q->queuedata = rbd_dev;
Signature · 1 line
2.4.3
#3 lore raw
In reply to Ilya Dryomov
Quoted 10 lines · Ilya Dryomov

Hello,

Please apply the following backport of bae818ee1577 ("rbd: require stable pages if message data CRCs are enabled") to 3.9+.

I have two patches: one for 3.9-3.17 and another one for 3.18+. It really is a single backport (git cherry-pick / git am -3 can handle the remaining context just fine), but I know some of you use quilt for queueing up patches, so I'm sending both just in case.

Quoted 7 lines

Thanks,

Ilya
--
To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Thanks for providing these backports. I'll queue it for the 3.16 kernel.

Cheers,
--
Lu�s