Skip to contentSkip to conversation

file-posix: fix cache.direct=on check for zoned block devices

PATCH Applied Niklas Cassel 3 participants 5 messages Jun 17 – Jul 29 Block layer

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 file-posix: fix cache.direct=on check for zoned block devices

For zoned block devices with SWR zones, the order of the writes must be maintained to avoid write not at write pointer errors.
Thus, cache.direct=on must be used to bypass the page cache, since the page cache uses write-back caching, where the physical order of the writes is not guaranteed.

There is already a check for cache.direct=on in raw_open_common(), however, this check is done before raw_refresh_zoned_limits() has been called (which initializes bs->bl.zoned), so it is currently dead code.

Fix this by moving the check to raw_refresh_zoned_limits(), such that the check is done after bs->bl.zoned has been initialized.

Fixes 774c726ceb2a ("block: add zoned BlockDriver check to block layer")

---
Changes since v1:
-Picked up tags from Damien and Dmitry
-Modified existing code comment to use plural instead of singular

 block/file-posix.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)
block/file-posix.c +10 -12
index 5019ddade9..32750ac509 100644
@@ -800,18 +800,6 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,
             goto fail;
         }
     }
-#ifdef CONFIG_BLKZONED
-    /*
-     * The kernel page cache does not reliably work for writes to SWR zones
-     * of zoned block device because it can not guarantee the order of writes.
-     */
-    if ((bs->bl.zoned != BLK_Z_NONE) &&
-        (!(s->open_flags & O_DIRECT))) {
-        error_setg(errp, "The driver supports zoned devices, and it requires "
-                         "cache.direct=on, which was not specified.");
-        return -EINVAL; /* No host kernel page cache */
-    }
-#endif
 
 #ifdef __FreeBSD__
     if (S_ISCHR(st.st_mode)) {
@@ -1462,6 +1450,16 @@ static void raw_refresh_zoned_limits(BlockDriverState *bs, struct stat *st,
     }
     bs->bl.zoned = zoned;
 
+    /*
+     * The kernel page cache does not reliably work for writes to SWR zones of
+     * zoned block devices because it can not guarantee the order of writes.
+     */
+    if (!(s->open_flags & O_DIRECT)) {
+        error_setg(errp, "The driver supports zoned devices, and it requires "
+                         "cache.direct=on, which was not specified.");
+        goto no_zoned;
+    }
+
     ret = get_sysfs_long_val(st, "max_open_zones");
     if (ret >= 0) {
         bs->bl.max_open_zones = ret;
Signature · 1 line
2.54.0
Reviewed-by on v2 1/1
#2 lore raw
In reply to Niklas Cassel
Quoted 19 lines · Niklas Cassel

For zoned block devices with SWR zones, the order of the writes must be maintained to avoid write not at write pointer errors.
Thus, cache.direct=on must be used to bypass the page cache, since the page cache uses write-back caching, where the physical order of the writes is not guaranteed.

There is already a check for cache.direct=on in raw_open_common(), however, this check is done before raw_refresh_zoned_limits() has been called (which initializes bs->bl.zoned), so it is currently dead code.

Fix this by moving the check to raw_refresh_zoned_limits(), such that the check is done after bs->bl.zoned has been initialized.

Fixes 774c726ceb2a ("block: add zoned BlockDriver check to block layer")

---
Changes since v1:

Review

Quoted 5 lines
(quoted diff) +0 -2
-Picked up tags from Damien and Dmitry
-Modified existing code comment to use plural instead of singular
 
 block/file-posix.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)
#3 lore raw
Quoted 23 lines · Stefan Hajnoczi
Quoted 22 lines · Niklas Cassel

For zoned block devices with SWR zones, the order of the writes must be maintained to avoid write not at write pointer errors.
Thus, cache.direct=on must be used to bypass the page cache, since the page cache uses write-back caching, where the physical order of the writes is not guaranteed.

There is already a check for cache.direct=on in raw_open_common(), however, this check is done before raw_refresh_zoned_limits() has been called (which initializes bs->bl.zoned), so it is currently dead code.

Fix this by moving the check to raw_refresh_zoned_limits(), such that the check is done after bs->bl.zoned has been initialized.

Fixes 774c726ceb2a ("block: add zoned BlockDriver check to block layer")

---
Changes since v1:

(quoted diff) +0 -2
-Picked up tags from Damien and Dmitry
-Modified existing code comment to use plural instead of singular
Quoted 4 lines
Quoted 2 lines
 block/file-posix.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

Gentle ping on this patch.

Would be nice if anyone could pick it up :)

Kind regards,
Niklas

#4 lore raw
In reply to Niklas Cassel
Quoted 19 lines · Niklas Cassel

For zoned block devices with SWR zones, the order of the writes must be maintained to avoid write not at write pointer errors.
Thus, cache.direct=on must be used to bypass the page cache, since the page cache uses write-back caching, where the physical order of the writes is not guaranteed.

There is already a check for cache.direct=on in raw_open_common(), however, this check is done before raw_refresh_zoned_limits() has been called (which initializes bs->bl.zoned), so it is currently dead code.

Fix this by moving the check to raw_refresh_zoned_limits(), such that the check is done after bs->bl.zoned has been initialized.

Fixes 774c726ceb2a ("block: add zoned BlockDriver check to block layer")

---
Changes since v1:

Review

Quoted 5 lines
(quoted diff) +0 -2
-Picked up tags from Damien and Dmitry
-Modified existing code comment to use plural instead of singular
 
 block/file-posix.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

No comments from Kevin and this is a zoned storage fix, so I'm picking it up.

QEMU is currently frozen for the 11.1 release and only serious regressions, security fixes, or build fixes should go into the next release candidate (-rc3).

I'm applying this to my block-next tree for the next release (11.2). It will hit the master branch in about 2 weeks when 11.1 has been released.

Thanks, applied to my block-next tree:
https://gitlab.com/stefanha/qemu/commits/block-next

Stefan