Skip to contentSkip to conversation

Enable cached zone report

PATCH In review Damien Le Moal 3 participants 15 messages Dec 20, 2025 – Jan 7

lore
Share & tools
lore
#0 lore raw

v3 cover Enable cached zone report

Enable cached zone report to speed up mkfs and repair on a zoned block device (e.g. an SMR disk). Cached zone report support was introduced in the kernel with version 6.19-rc1.

Patch 1 and 2 are a couple of simple fixes/cleanups.
Patch 3 intorduces the new xfrog_report_zones() helper function, which is used in patch 4 and 5 for mkfs and repair respectively. Finally, patch 6 modifies xfrog_report_zones() to enable cached report zones.

Changes from v2:
- Complete rework of the series to make the zone reporting code common in libfrog
- Added patch 1 and 2 as small cleanups/improvements.

Changes from v1:
- Fix erroneous handling of ioctl(BLKREPORTZONEV2) error to correctly fallback to the regular ioctl(BLKREPORTZONE) if the kernel does not support BLKREPORTZONEV2.

Damien Le Moal (6):
libxfs: add missing forward declaration in xfs_zones.h
mkfs: remove unnecessary return value affectation
libfrog: introduce xfrog_report_zones
mkfs: use xfrog_report_zones()
repair: use xfrog_report_zones()
libfrog: enable cached report zones

 libfrog/Makefile   |  6 ++++--
 libfrog/zones.c    | 51 ++++++++++++++++++++++++++++++++++++++++++++++
 libfrog/zones.h    | 22 ++++++++++++++++++++
 libxfs/xfs_zones.c |  3 ++-
 libxfs/xfs_zones.h |  1 +
 mkfs/xfs_mkfs.c    | 39 +++++++++--------------------------
 repair/zoned.c     | 29 ++++++++------------------
 7 files changed, 98 insertions(+), 53 deletions(-)
 create mode 100644 libfrog/zones.c
 create mode 100644 libfrog/zones.h
Signature · 1 line
2.52.0
#1 lore raw

PATCH v3 1/6 libxfs: add missing forward declaration in xfs_zones.h

Add the missing forward declaration for struct blk_zone in xfs_zones.h. This avoids headaches with the order of header file inclusion to avoid compilation errors.

Fixes 48ccc2459039 ("xfs: parse and validate hardware zone information")
---
 libxfs/xfs_zones.h | 1 +
 1 file changed, 1 insertion(+)
libxfs/xfs_zones.h +1 -0
index c4f1367b2cca..6376bb0e6da6 100644
@@ -2,6 +2,7 @@
 #ifndef _LIBXFS_ZONES_H
 #define _LIBXFS_ZONES_H
 
+struct blk_zone;
 struct xfs_rtgroup;
 
 /*
Signature · 1 line
2.52.0
#2 lore raw

Re: [PATCH v3 1/6] libxfs: add missing forward declaration in xfs_zones.h

Quoted 8 lines · Damien Le Moal

Add the missing forward declaration for struct blk_zone in xfs_zones.h. This avoids headaches with the order of header file inclusion to avoid compilation errors.

Fixes 48ccc2459039 ("xfs: parse and validate hardware zone information")
---
 libxfs/xfs_zones.h | 1 +

We don't normally merge changes to xfsprogs' libxfs that haven't already gone into the kernel libxfs. If Andrey's ok with this then I am too.

(alternate suggestion: can the forward declaration go in platform_defs.h?)

--D

Quoted 18 lines
 1 file changed, 1 insertion(+)
libxfs/xfs_zones.h +1 -0
index c4f1367b2cca..6376bb0e6da6 100644
@@ -2,6 +2,7 @@
 #ifndef _LIBXFS_ZONES_H
 #define _LIBXFS_ZONES_H
 
+struct blk_zone;
 struct xfs_rtgroup;
 
 /*
Signature · 1 line
2.52.0
#3 lore raw

Re: [PATCH v3 1/6] libxfs: add missing forward declaration in xfs_zones.h

In reply to Damien Le Moal
Quoted 6 lines · Damien Le Moal

Add the missing forward declaration for struct blk_zone in xfs_zones.h. This avoids headaches with the order of header file inclusion to avoid compilation errors.

Fixes 48ccc2459039 ("xfs: parse and validate hardware zone information")

This needs to go into the kernel first. And yes, I can't blame new contributors for not knowing what files are shared..

#4 lore raw

PATCH v3 2/6 mkfs: remove unnecessary return value affectation

In reply to Damien Le Moal

The function report_zones() in mkfs/xfs_mkfs.c is a void function. So there is no need to set the variable ret to -EIO before returning if fstat() fails.

Fixes 2e5a737a61d3 ("xfs_mkfs: support creating file system with zoned RT devices")
---
 mkfs/xfs_mkfs.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
mkfs/xfs_mkfs.c +1 -3
index bc6a28b63c24..550fc011b614 100644
@@ -2589,10 +2589,8 @@ report_zones(
 		exit(1);
 	}
 
-	if (fstat(fd, &st) < 0) {
-		ret = -EIO;
+	if (fstat(fd, &st) < 0)
 		goto out_close;
-	}
 	if (!S_ISBLK(st.st_mode))
 		goto out_close;
 
Signature · 1 line
2.52.0
Reviewed-by on v3 2/6
#5 lore raw

Re: [PATCH v3 2/6] mkfs: remove unnecessary return value affectation

Quoted 6 lines · Damien Le Moal

The function report_zones() in mkfs/xfs_mkfs.c is a void function. So there is no need to set the variable ret to -EIO before returning if fstat() fails.

Fixes 2e5a737a61d3 ("xfs_mkfs: support creating file system with zoned RT devices")

LGTM,

--D

Quoted 24 lines
---
 mkfs/xfs_mkfs.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
mkfs/xfs_mkfs.c +1 -3
index bc6a28b63c24..550fc011b614 100644
@@ -2589,10 +2589,8 @@ report_zones(
 		exit(1);
 	}
 
-	if (fstat(fd, &st) < 0) {
-		ret = -EIO;
+	if (fstat(fd, &st) < 0)
 		goto out_close;
-	}
 	if (!S_ISBLK(st.st_mode))
 		goto out_close;
 
Signature · 1 line
2.52.0
#7 lore raw

PATCH v3 3/6 libfrog: introduce xfrog_report_zones

In reply to Damien Le Moal

Define the new helper function xfrog_report_zones() to report zones of a zoned block device. This function is implemented in the new file libfrog/zones.c and defined in the header file libfrog/zones.h.

xfrog_report_zones() allocates and returns a struct blk_zone_report structure. It is the responsibility of the caller to free this structure after use.

---
 libfrog/Makefile |  6 ++++--
 libfrog/zones.c  | 42 ++++++++++++++++++++++++++++++++++++++++++
 libfrog/zones.h  | 13 +++++++++++++
 3 files changed, 59 insertions(+), 2 deletions(-)
 create mode 100644 libfrog/zones.c
 create mode 100644 libfrog/zones.h
libfrog/Makefile +4 -2
index 268fa26638d7..9f405ffe3475 100644
@@ -35,7 +35,8 @@ radix-tree.c \
 randbytes.c \
 scrub.c \
 util.c \
-workqueue.c
+workqueue.c \
+zones.c
 
 HFILES = \
 avl64.h \
@@ -65,7 +66,8 @@ radix-tree.h \
 randbytes.h \
 scrub.h \
 statx.h \
-workqueue.h
+workqueue.h \
+zones.h
 
 GETTEXT_PY = \
 	gettext.py
libfrog/zones.c new file +42 -0

42 lines not on this page · raw message

libfrog/zones.h new file +13 -0
new file mode 100644
index 000000000000..66df7a426a27
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2025, Western Digital Corporation or its affiliates.
+ */
+#ifndef __LIBFROG_ZONE_H__
+#define __LIBFROG_ZONE_H__
+
+#include <stdint.h>
+#include <linux/blkzoned.h>
+
+struct blk_zone_report	*xfrog_report_zones(int	fd, uint64_t sector);
+
+#endif /* __LIBFROG_ZONE_H__ */
Signature · 1 line
2.52.0
#8 lore raw

Re: [PATCH v3 3/6] libfrog: introduce xfrog_report_zones

Quoted 47 lines · Damien Le Moal

Define the new helper function xfrog_report_zones() to report zones of a zoned block device. This function is implemented in the new file libfrog/zones.c and defined in the header file libfrog/zones.h.

xfrog_report_zones() allocates and returns a struct blk_zone_report structure. It is the responsibility of the caller to free this structure after use.

---
 libfrog/Makefile |  6 ++++--
 libfrog/zones.c  | 42 ++++++++++++++++++++++++++++++++++++++++++
 libfrog/zones.h  | 13 +++++++++++++
 3 files changed, 59 insertions(+), 2 deletions(-)
 create mode 100644 libfrog/zones.c
 create mode 100644 libfrog/zones.h
libfrog/Makefile +4 -2
index 268fa26638d7..9f405ffe3475 100644
@@ -35,7 +35,8 @@ radix-tree.c \
 randbytes.c \
 scrub.c \
 util.c \
-workqueue.c
+workqueue.c \
+zones.c
 
 HFILES = \
 avl64.h \
@@ -65,7 +66,8 @@ radix-tree.h \
 randbytes.h \
 scrub.h \
 statx.h \
-workqueue.h
+workqueue.h \
+zones.h
 
 GETTEXT_PY = \
 	gettext.py
libfrog/zones.c new file +0 -0
new file mode 100644
index 000000000000..0187edce5fa4

libfrog/zones.c line 6

Quoted 6 lines
libfrog/zones.c new file +6 -0
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2025, Western Digital Corporation or its affiliates.
+ */
+#include "platform_defs.h"
+#include "zones.h"

Nit:

#include "libfrog/zones.h"

to make it more obvious that we're looking for the libfrog version (on the off chance /usr/include ever end up with a zones.h).

Quoted 20 lines
libfrog/zones.c new file +20 -0
+#include <sys/ioctl.h>
+
+/* random size that allows efficient processing */
+#define ZONES_PER_REPORT		16384
+
+struct blk_zone_report	*
+xfrog_report_zones(
+	int			fd,
+	uint64_t		sector)
+{
+	struct blk_zone_report	*rep;
+	size_t			rep_size;
+	int			ret;
+
+	rep_size = sizeof(struct blk_zone_report) +
+		   sizeof(struct blk_zone) * ZONES_PER_REPORT;
+	rep = calloc(1, rep_size);
+	if (!rep) {
+		fprintf(stderr,
+_("Failed to allocate memory for reporting zones.\n"));

libfrog/zones.c line 36

Quoted 10 lines
libfrog/zones.c new file +10 -0
+		return NULL;
+	}
+
+	rep->sector = sector;
+	rep->nr_zones = ZONES_PER_REPORT;
+
+	ret = ioctl(fd, BLKREPORTZONE, rep);
+	if (ret) {
+		fprintf(stderr,
+_("ioctl(BLKREPORTZONE) failed: %d!\n"), -errno);

Note that eventually gcc & friends will start warning about non-constant formatting strings because i18n catalogue attacks are a thing now.

<start rant>
aka the Rigellian translation pack maps that string to "Hork fubzzz %s %s %s %s" and now fprintf walks right off the varargs list and kaboom.

Annoyingly this makes output even worse because the canonical form of that becomes this monstrosity:

		fprintf(stderr, "%s: %s\n", _("BLKREPORTZONE failed"),
				strerror(-errno));

This also fails because the Rigellians use jack-o-lantern emoji to separate fields instead of colon-space so I don't know what to do :P <end rant>

That said, xfsprogs is full of potential i18n catalogue attacks so if Andrey's cool with it, then I'll go along with it.

(Though I'll at least try to remember not to type out the familiar old way even though I've been doing it for 30 years.)

libfrog/zones.c line 41

Quoted 5 lines
libfrog/zones.c new file +5 -0
+		free(rep);
+		return NULL;
+	}
+
+	return rep;

The logic in here looks good, though.

Quoted 7 lines
libfrog/zones.c new file +1 -0
+}
libfrog/zones.h new file +0 -0
new file mode 100644
index 000000000000..66df7a426a27

libfrog/zones.h line 9

Quoted 9 lines
libfrog/zones.h new file +9 -0
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2025, Western Digital Corporation or its affiliates.
+ */
+#ifndef __LIBFROG_ZONE_H__
+#define __LIBFROG_ZONE_H__
+
+#include <stdint.h>
+#include <linux/blkzoned.h>

Don't put includes in header files, please. uint64_t should be covered by platform_defs.h and you can just have a forward declaration of struct blk_zone_report.

--D

Quoted 8 lines
libfrog/zones.h new file +4 -0
+
+struct blk_zone_report	*xfrog_report_zones(int	fd, uint64_t sector);
+
+#endif /* __LIBFROG_ZONE_H__ */
Signature · 1 line
2.52.0
#9 lore raw

PATCH v3 4/6 mkfs: use xfrog_report_zones()

In reply to Damien Le Moal

Use the function xfrog_report_zones() to obtain zone information from a zoned device instead of directly issuing a BLKREPORTZONE ioctl.

---
 mkfs/xfs_mkfs.c | 35 +++++++++--------------------------
 1 file changed, 9 insertions(+), 26 deletions(-)
mkfs/xfs_mkfs.c +9 -26
index 550fc011b614..ac7ad0661805 100644
@@ -6,7 +6,6 @@
 #include "libfrog/util.h"
 #include "libxfs.h"
 #include <ctype.h>
-#include <linux/blkzoned.h>
 #include "libxfs/xfs_zones.h"
 #include "xfs_multidisk.h"
 #include "libxcmd.h"
@@ -15,6 +14,7 @@
 #include "libfrog/crc32cselftest.h"
 #include "libfrog/dahashselftest.h"
 #include "libfrog/fsproperties.h"
+#include "libfrog/zones.h"
 #include "proto.h"
 #include <ini.h>
 
@@ -2566,20 +2566,16 @@ struct zone_topology {
 	struct zone_info	log;
 };
 
-/* random size that allows efficient processing */
-#define ZONES_PER_IOCTL			16384
-
 static void
 report_zones(
 	const char		*name,
 	struct zone_info	*zi)
 {
-	struct blk_zone_report	*rep;
+	struct blk_zone_report	*rep = NULL;
 	bool			found_seq = false;
-	int			fd, ret = 0;
+	int			fd;
 	uint64_t		device_size;
 	uint64_t		sector = 0;
-	size_t			rep_size;
 	unsigned int		i, n = 0;
 	struct stat		st;
 
@@ -2606,31 +2602,18 @@ report_zones(
 	zi->nr_zones = device_size / zi->zone_size;
 	zi->nr_conv_zones = 0;
 
-	rep_size = sizeof(struct blk_zone_report) +
-		   sizeof(struct blk_zone) * ZONES_PER_IOCTL;
-	rep = malloc(rep_size);
-	if (!rep) {
-		fprintf(stderr,
-_("Failed to allocate memory for zone reporting.\n"));
-		exit(1);
-	}
-
 	while (n < zi->nr_zones) {
-		struct blk_zone *zones = (struct blk_zone *)(rep + 1);
+		struct blk_zone *zones;
 
-		memset(rep, 0, rep_size);
-		rep->sector = sector;
-		rep->nr_zones = ZONES_PER_IOCTL;
-
-		ret = ioctl(fd, BLKREPORTZONE, rep);
-		if (ret) {
-			fprintf(stderr,
-_("ioctl(BLKREPORTZONE) failed: %d!\n"), -errno);
+		free(rep);
+		rep = xfrog_report_zones(fd, sector);
+		if (!rep)
 			exit(1);
-		}
+
 		if (!rep->nr_zones)
 			break;
 
+		zones = (struct blk_zone *)(rep + 1);
 		for (i = 0; i < rep->nr_zones; i++) {
 			if (n >= zi->nr_zones)
 				break;
Signature · 1 line
2.52.0
#10 lore raw

Re: [PATCH v3 4/6] mkfs: use xfrog_report_zones()

Quoted 68 lines · Damien Le Moal

Use the function xfrog_report_zones() to obtain zone information from a zoned device instead of directly issuing a BLKREPORTZONE ioctl.

---
 mkfs/xfs_mkfs.c | 35 +++++++++--------------------------
 1 file changed, 9 insertions(+), 26 deletions(-)
mkfs/xfs_mkfs.c +4 -17
index 550fc011b614..ac7ad0661805 100644
@@ -6,7 +6,6 @@
 #include "libfrog/util.h"
 #include "libxfs.h"
 #include <ctype.h>
-#include <linux/blkzoned.h>
 #include "libxfs/xfs_zones.h"
 #include "xfs_multidisk.h"
 #include "libxcmd.h"
@@ -15,6 +14,7 @@
 #include "libfrog/crc32cselftest.h"
 #include "libfrog/dahashselftest.h"
 #include "libfrog/fsproperties.h"
+#include "libfrog/zones.h"
 #include "proto.h"
 #include <ini.h>
 
@@ -2566,20 +2566,16 @@ struct zone_topology {
 	struct zone_info	log;
 };
 
-/* random size that allows efficient processing */
-#define ZONES_PER_IOCTL			16384
-
 static void
 report_zones(
 	const char		*name,
 	struct zone_info	*zi)
 {
-	struct blk_zone_report	*rep;
+	struct blk_zone_report	*rep = NULL;
 	bool			found_seq = false;
-	int			fd, ret = 0;
+	int			fd;
 	uint64_t		device_size;
 	uint64_t		sector = 0;
-	size_t			rep_size;
 	unsigned int		i, n = 0;
 	struct stat		st;
 
@@ -2606,31 +2602,18 @@ report_zones(
 	zi->nr_zones = device_size / zi->zone_size;
 	zi->nr_conv_zones = 0;
 
-	rep_size = sizeof(struct blk_zone_report) +
-		   sizeof(struct blk_zone) * ZONES_PER_IOCTL;
-	rep = malloc(rep_size);
-	if (!rep) {
-		fprintf(stderr,
-_("Failed to allocate memory for zone reporting.\n"));
-		exit(1);
-	}
-
 	while (n < zi->nr_zones) {
-		struct blk_zone *zones = (struct blk_zone *)(rep + 1);
+		struct blk_zone *zones;
 

mkfs/xfs_mkfs.c report_zones( · line 2609

Quoted 10 lines
mkfs/xfs_mkfs.c +2 -8
-		memset(rep, 0, rep_size);
-		rep->sector = sector;
-		rep->nr_zones = ZONES_PER_IOCTL;
-
-		ret = ioctl(fd, BLKREPORTZONE, rep);
-		if (ret) {
-			fprintf(stderr,
-_("ioctl(BLKREPORTZONE) failed: %d!\n"), -errno);
+		free(rep);
+		rep = xfrog_report_zones(fd, sector);

Now that I see the actual usage patterns I wonder if this would be more efficient if you had a separate allocator function to avoid repeatedly freeing and re-allocating the buffer?

Also this could be combined with the previous patch since you're really just hoisting code from mkfs to libfrog.

--D

Quoted 15 lines
mkfs/xfs_mkfs.c +3 -1
+		if (!rep)
 			exit(1);
-		}
+
 		if (!rep->nr_zones)
 			break;
 
+		zones = (struct blk_zone *)(rep + 1);
 		for (i = 0; i < rep->nr_zones; i++) {
 			if (n >= zi->nr_zones)
 				break;
Signature · 1 line
2.52.0
#11 lore raw

PATCH v3 5/6 repair: use xfrog_report_zones()

In reply to Damien Le Moal

Use the function xfrog_report_zones() to obtain zones information from a zoned device instead of directly issuing a BLKREPORTZONE ioctl.

---
 repair/zoned.c | 29 ++++++++---------------------
 1 file changed, 8 insertions(+), 21 deletions(-)
repair/zoned.c +8 -21
index 206b0158f95f..c211e02448da 100644
@@ -3,7 +3,7 @@
  * Copyright (c) 2024 Christoph Hellwig.
  */
 #include <ctype.h>
-#include <linux/blkzoned.h>
+#include "libfrog/zones.h"
 #include "libxfs_priv.h"
 #include "libxfs.h"
 #include "xfs_zones.h"
@@ -51,8 +51,7 @@ check_zones(
 	uint64_t		sector = XFS_FSB_TO_BB(mp, mp->m_sb.sb_rtstart);
 	unsigned int		zone_size, zone_capacity;
 	uint64_t		device_size;
-	size_t			rep_size;
-	struct blk_zone_report	*rep;
+	struct blk_zone_report	*rep = NULL;
 	unsigned int		i, n = 0;
 
 	if (ioctl(fd, BLKGETSIZE64, &device_size))
@@ -67,30 +66,18 @@ check_zones(
 		return;
 	}
 
-	rep_size = sizeof(struct blk_zone_report) +
-		   sizeof(struct blk_zone) * ZONES_PER_IOCTL;
-	rep = malloc(rep_size);
-	if (!rep) {
-		do_warn(_("malloc failed for zone report\n"));
-		return;
-	}
-
 	while (n < mp->m_sb.sb_rgcount) {
-		struct blk_zone *zones = (struct blk_zone *)(rep + 1);
-		int ret;
+		struct blk_zone *zones;
 
-		memset(rep, 0, rep_size);
-		rep->sector = sector;
-		rep->nr_zones = ZONES_PER_IOCTL;
+		free(rep);
+		rep = xfrog_report_zones(fd, sector);
+		if (!rep)
+			return;
 
-		ret = ioctl(fd, BLKREPORTZONE, rep);
-		if (ret) {
-			do_error(_("ioctl(BLKREPORTZONE) failed: %d!\n"), ret);
-			goto out_free;
-		}
 		if (!rep->nr_zones)
 			break;
 
+		zones = (struct blk_zone *)(rep + 1);
 		for (i = 0; i < rep->nr_zones; i++) {
 			if (n >= mp->m_sb.sb_rgcount)
 				break;
Signature · 1 line
2.52.0
#12 lore raw

PATCH v3 6/6 libfrog: enable cached report zones

In reply to Damien Le Moal

Modify the function xfrog_report_zones() to default to always trying first a cached report zones using the BLKREPORTZONEV2 ioctl. If the kernel does not support BLKREPORTZONEV2, fall back to the (slower) regular report zones BLKREPORTZONE ioctl.

TO enable this feature even if xfsprogs is compiled on a system where linux/blkzoned.h does not define BLKREPORTZONEV2, this ioctl is defined in libfrog/zones.h, together with the BLK_ZONE_REP_CACHED flag and the BLK_ZONE_COND_ACTIVE zone condition.

Since a cached report zone always return the condition BLK_ZONE_COND_ACTIVE for any zone that is implicitly open, explicitly open or closed, the function xfs_zone_validate_seq() is modified to handle this new condition as being equivalent to the implicit open, explicit open or closed conditions.

---
 libfrog/zones.c    | 11 ++++++++++-
 libfrog/zones.h    |  9 +++++++++
 libxfs/xfs_zones.c |  3 ++-
 3 files changed, 21 insertions(+), 2 deletions(-)
libfrog/zones.c +10 -1
index 0187edce5fa4..8b45066de176 100644
@@ -27,10 +27,19 @@ _("Failed to allocate memory for reporting zones.\n"));
 		return NULL;
 	}
 
+	/*
+	 * Try cached report zones first. If this fails, fallback to the regular
+	 * (slower) report zones.
+	 */
 	rep->sector = sector;
 	rep->nr_zones = ZONES_PER_REPORT;
+	rep->flags = BLK_ZONE_REP_CACHED;
 
-	ret = ioctl(fd, BLKREPORTZONE, rep);
+	ret = ioctl(fd, BLKREPORTZONEV2, rep);
+	if (ret < 0 && errno == ENOTTY) {
+		rep->flags = 0;
+		ret = ioctl(fd, BLKREPORTZONE, rep);
+	}
 	if (ret) {
 		fprintf(stderr,
 _("ioctl(BLKREPORTZONE) failed: %d!\n"), -errno);
libfrog/zones.h +9 -0
index 66df7a426a27..4605aea93114 100644
@@ -8,6 +8,15 @@
 #include <stdint.h>
 #include <linux/blkzoned.h>
 
+/*
+ * Cached report ioctl (/usr/include/linux/blkzoned.h)
+ */
+#ifndef BLKREPORTZONEV2
+#define BLKREPORTZONEV2		_IOWR(0x12, 142, struct blk_zone_report)
+#define BLK_ZONE_REP_CACHED	(1U << 31)
+#define BLK_ZONE_COND_ACTIVE	0xff
+#endif
+
 struct blk_zone_report	*xfrog_report_zones(int	fd, uint64_t sector);
 
 #endif /* __LIBFROG_ZONE_H__ */
libxfs/xfs_zones.c +2 -1
index 7a81d83f5b3e..3c89a89ca21e 100644
@@ -3,7 +3,7 @@
  * Copyright (c) 2023-2025 Christoph Hellwig.
  * Copyright (c) 2024-2025, Western Digital Corporation or its affiliates.
  */
-#include <linux/blkzoned.h>
+#include <libfrog/zones.h>
 #include "libxfs_priv.h"
 #include "xfs.h"
 #include "xfs_fs.h"
@@ -97,6 +97,7 @@ xfs_zone_validate_seq(
 	case BLK_ZONE_COND_IMP_OPEN:
 	case BLK_ZONE_COND_EXP_OPEN:
 	case BLK_ZONE_COND_CLOSED:
+	case BLK_ZONE_COND_ACTIVE:
 		return xfs_zone_validate_wp(zone, rtg, write_pointer);
 	case BLK_ZONE_COND_FULL:
 		return xfs_zone_validate_full(zone, rtg, write_pointer);
Signature · 1 line
2.52.0
#13 lore raw

Re: [PATCH v3 6/6] libfrog: enable cached report zones

Quoted 53 lines · Damien Le Moal

Modify the function xfrog_report_zones() to default to always trying first a cached report zones using the BLKREPORTZONEV2 ioctl. If the kernel does not support BLKREPORTZONEV2, fall back to the (slower) regular report zones BLKREPORTZONE ioctl.

TO enable this feature even if xfsprogs is compiled on a system where linux/blkzoned.h does not define BLKREPORTZONEV2, this ioctl is defined in libfrog/zones.h, together with the BLK_ZONE_REP_CACHED flag and the BLK_ZONE_COND_ACTIVE zone condition.

Since a cached report zone always return the condition BLK_ZONE_COND_ACTIVE for any zone that is implicitly open, explicitly open or closed, the function xfs_zone_validate_seq() is modified to handle this new condition as being equivalent to the implicit open, explicit open or closed conditions.

---
 libfrog/zones.c    | 11 ++++++++++-
 libfrog/zones.h    |  9 +++++++++
 libxfs/xfs_zones.c |  3 ++-
 3 files changed, 21 insertions(+), 2 deletions(-)
libfrog/zones.c +10 -1
index 0187edce5fa4..8b45066de176 100644
@@ -27,10 +27,19 @@ _("Failed to allocate memory for reporting zones.\n"));
 		return NULL;
 	}
 
+	/*
+	 * Try cached report zones first. If this fails, fallback to the regular
+	 * (slower) report zones.
+	 */
 	rep->sector = sector;
 	rep->nr_zones = ZONES_PER_REPORT;
+	rep->flags = BLK_ZONE_REP_CACHED;
 
-	ret = ioctl(fd, BLKREPORTZONE, rep);
+	ret = ioctl(fd, BLKREPORTZONEV2, rep);
+	if (ret < 0 && errno == ENOTTY) {
+		rep->flags = 0;
+		ret = ioctl(fd, BLKREPORTZONE, rep);
+	}
 	if (ret) {
 		fprintf(stderr,
 _("ioctl(BLKREPORTZONE) failed: %d!\n"), -errno);
libfrog/zones.h +0 -0
index 66df7a426a27..4605aea93114 100644

libfrog/zones.h line 17

Quoted 10 lines
libfrog/zones.h +7 -0
@@ -8,6 +8,15 @@
 #include <stdint.h>
 #include <linux/blkzoned.h>
 
+/*
+ * Cached report ioctl (/usr/include/linux/blkzoned.h)
+ */
+#ifndef BLKREPORTZONEV2
+#define BLKREPORTZONEV2		_IOWR(0x12, 142, struct blk_zone_report)
+#define BLK_ZONE_REP_CACHED	(1U << 31)
+#define BLK_ZONE_COND_ACTIVE	0xff

Of these three definitions, only BLK_ZONE_COND_ACTIVE is actually used outside of libfrog/zones.c, right? I suggest defining BLKREPORTZONEV2 and BLK_ZONE_REP_CACHED there instead of zones.h to reduce the number of symbols floating around.

(The rest of the code changes look ok to me)

--D

Quoted 30 lines
libfrog/zones.h +2 -0
+#endif
+
 struct blk_zone_report	*xfrog_report_zones(int	fd, uint64_t sector);
 
 #endif /* __LIBFROG_ZONE_H__ */
libxfs/xfs_zones.c +2 -1
index 7a81d83f5b3e..3c89a89ca21e 100644
@@ -3,7 +3,7 @@
  * Copyright (c) 2023-2025 Christoph Hellwig.
  * Copyright (c) 2024-2025, Western Digital Corporation or its affiliates.
  */
-#include <linux/blkzoned.h>
+#include <libfrog/zones.h>
 #include "libxfs_priv.h"
 #include "xfs.h"
 #include "xfs_fs.h"
@@ -97,6 +97,7 @@ xfs_zone_validate_seq(
 	case BLK_ZONE_COND_IMP_OPEN:
 	case BLK_ZONE_COND_EXP_OPEN:
 	case BLK_ZONE_COND_CLOSED:
+	case BLK_ZONE_COND_ACTIVE:
 		return xfs_zone_validate_wp(zone, rtg, write_pointer);
 	case BLK_ZONE_COND_FULL:
 		return xfs_zone_validate_full(zone, rtg, write_pointer);
Signature · 1 line
2.52.0
#14 lore raw

Re: [PATCH v3 6/6] libfrog: enable cached report zones

In reply to Damien Le Moal

libxfs/xfs_zones.c

Quoted 4 lines · Damien Le Moal
libxfs/xfs_zones.c +0 -0
index 7a81d83f5b3e..3c89a89ca21e 100644

This file is shared with the kernel, so we'll need to get the changes from a libxfs sync as well.

libxfs/xfs_zones.c line 6

Quoted 6 lines
libxfs/xfs_zones.c +1 -1
@@ -3,7 +3,7 @@
  * Copyright (c) 2023-2025 Christoph Hellwig.
  * Copyright (c) 2024-2025, Western Digital Corporation or its affiliates.
  */
-#include <linux/blkzoned.h>
+#include <libfrog/zones.h>

.. and we'll need this include to be hidden in the xfsprogs-wide headers.