[PATCH 0/2] ceph: use entity name from new device string

[PATCH 0/2] ceph: use entity name from new device string

[Patrick Donnelly]

From: Patrick Donnelly <pdonnell@redhat.com>

Respinning this because the last series accidentally included patches from
another set.

Patrick Donnelly (2):
  ceph: requalify some char pointers as const
  ceph: extract entity name from device id

 fs/ceph/super.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)


base-commit: 7234e2ea0edd00bfb6bb2159e55878c19885ce68
-- 
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

[PATCH 1/2] ceph: requalify some char pointers as const

[Patrick Donnelly]

From: Patrick Donnelly <pdonnell@redhat.com>

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
---
 fs/ceph/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 73f321b52895..42bdbe5b7ef9 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -286,7 +286,7 @@ static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end,
 	struct ceph_fsid fsid;
 	struct ceph_parse_opts_ctx *pctx = fc->fs_private;
 	struct ceph_mount_options *fsopt = pctx->opts;
-	char *fsid_start, *fs_name_start;
+	const char *fsid_start, *fs_name_start;
 
 	if (*dev_name_end != '=') {
 		dout("separator '=' missing in source");
-- 
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

[PATCH 2/2] ceph: extract entity name from device id

[Patrick Donnelly]

From: Patrick Donnelly <pdonnell@redhat.com>

Previously, the "name" in the new device syntax "<name>@<fsid>.<fsname>" was
ignored because (presumably) tests were done using mount.ceph which also passed
the entity name using "-o name=foo". If mounting is done without the mount.ceph
helper, the new device id syntax fails to set the name properly.

Cc: stable@vger.kernel.org
Resolves: https://tracker.ceph.com/issues/68516
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
---
 fs/ceph/super.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 42bdbe5b7ef9..de03cd6eb86e 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -285,7 +285,9 @@ static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end,
 	size_t len;
 	struct ceph_fsid fsid;
 	struct ceph_parse_opts_ctx *pctx = fc->fs_private;
+	struct ceph_options *opts = pctx->copts;
 	struct ceph_mount_options *fsopt = pctx->opts;
+	const char *name_start = dev_name;
 	const char *fsid_start, *fs_name_start;
 
 	if (*dev_name_end != '=') {
@@ -296,8 +298,14 @@ static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end,
 	fsid_start = strchr(dev_name, '@');
 	if (!fsid_start)
 		return invalfc(fc, "missing cluster fsid");
-	++fsid_start; /* start of cluster fsid */
+	len = fsid_start - name_start;
+	kfree(opts->name);
+	opts->name = kstrndup(name_start, len, GFP_KERNEL);
+	if (!opts->name)
+		return -ENOMEM;
+	dout("using %s entity name", opts->name);
 
+	++fsid_start; /* start of cluster fsid */
 	fs_name_start = strchr(fsid_start, '.');
 	if (!fs_name_start)
 		return invalfc(fc, "missing file system name");
-- 
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

Re: [PATCH 0/2] ceph: use entity name from new device string

[Ilya Dryomov]

On Mon, Oct 14, 2024 at 3:46 PM Patrick Donnelly <batrick@batbytes.com> wrote:
>
> From: Patrick Donnelly <pdonnell@redhat.com>
>
> Respinning this because the last series accidentally included patches from
> another set.
>
> Patrick Donnelly (2):
>   ceph: requalify some char pointers as const
>   ceph: extract entity name from device id

Hi Patrick,

Since the fix (the second patch) is marked for stable, I reordered
these so that the first patch which just tacks on const doesn't get in
the way when backporting the fix.

Thanks,

                Ilya

[PATCH 1/3] ceph: correct ceph_mds_cap_item field name

[Patrick Donnelly]

The issue_seq is sent with bulk cap releases, not the current sequence number.

See also ceph.git commit: "include/ceph_fs: correct ceph_mds_cap_item field name".

See-also: https://tracker.ceph.com/issues/66704
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
---
 fs/ceph/mds_client.c         | 2 +-
 include/linux/ceph/ceph_fs.h | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index c4a5fd94bbbb..0be82de8a6da 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -2362,7 +2362,7 @@ static void ceph_send_cap_releases(struct ceph_mds_client *mdsc,
 		item->ino = cpu_to_le64(cap->cap_ino);
 		item->cap_id = cpu_to_le64(cap->cap_id);
 		item->migrate_seq = cpu_to_le32(cap->mseq);
-		item->seq = cpu_to_le32(cap->issue_seq);
+		item->issue_seq = cpu_to_le32(cap->issue_seq);
 		msg->front.iov_len += sizeof(*item);
 
 		ceph_put_cap(mdsc, cap);
diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h
index ee1d0e5f9789..4ff3ad5e9210 100644
--- a/include/linux/ceph/ceph_fs.h
+++ b/include/linux/ceph/ceph_fs.h
@@ -822,7 +822,7 @@ struct ceph_mds_cap_release {
 struct ceph_mds_cap_item {
 	__le64 ino;
 	__le64 cap_id;
-	__le32 migrate_seq, seq;
+	__le32 migrate_seq, issue_seq;
 } __attribute__ ((packed));
 
 #define CEPH_MDS_LEASE_REVOKE           1  /*    mds  -> client */

base-commit: 75b607fab38d149f232f01eae5e6392b394dd659
-- 
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

[PATCH 2/2] ceph: extract entity name from device id

[Patrick Donnelly]

From: Patrick Donnelly <pdonnell@redhat.com>

Previously, the "name" in the new device syntax "<name>@<fsid>.<fsname>" was
ignored because (presumably) tests were done using mount.ceph which also passed
the entity name using "-o name=foo". If mounting is done without the mount.ceph
helper, the new device id syntax fails to set the name properly.

Resolves: https://tracker.ceph.com/issues/68516
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
---
 fs/ceph/super.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 42bdbe5b7ef9..de03cd6eb86e 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -285,7 +285,9 @@ static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end,
 	size_t len;
 	struct ceph_fsid fsid;
 	struct ceph_parse_opts_ctx *pctx = fc->fs_private;
+	struct ceph_options *opts = pctx->copts;
 	struct ceph_mount_options *fsopt = pctx->opts;
+	const char *name_start = dev_name;
 	const char *fsid_start, *fs_name_start;
 
 	if (*dev_name_end != '=') {
@@ -296,8 +298,14 @@ static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end,
 	fsid_start = strchr(dev_name, '@');
 	if (!fsid_start)
 		return invalfc(fc, "missing cluster fsid");
-	++fsid_start; /* start of cluster fsid */
+	len = fsid_start - name_start;
+	kfree(opts->name);
+	opts->name = kstrndup(name_start, len, GFP_KERNEL);
+	if (!opts->name)
+		return -ENOMEM;
+	dout("using %s entity name", opts->name);
 
+	++fsid_start; /* start of cluster fsid */
 	fs_name_start = strchr(fsid_start, '.');
 	if (!fs_name_start)
 		return invalfc(fc, "missing file system name");
-- 
Patrick Donnelly, Ph.D.
He / Him / His
Red Hat Partner Engineer
IBM, Inc.
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D