Skip to contentSkip to conversation

ceph: add debugfs entries signifying new mount syntax support

RFC In review Venky Shankar 4 participants 12 messages Aug 18, 2021 – Aug 23, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover ceph: add debugfs entries signifying new mount syntax support

[This is based on top of new mount syntax series]

Patrick proposed the idea of having debugfs entries to signify if kernel supports the new (v2) mount syntax. The primary use of this information is to catch any bugs in the new syntax implementation.

This would be done as follows::

The userspace mount helper tries to mount using the new mount syntax and fallsback to using old syntax if the mount using new syntax fails. However, a bug in the new mount syntax implementation can silently result in the mount helper switching to old syntax.

So, the debugfs entries can be relied upon by the mount helper to check if the kernel supports the new mount syntax. Cases when the mount using the new syntax fails, but the kernel does support the new mount syntax, the mount helper could probably log before switching to the old syntax (or fail the mount altogether when run in test mode).

Debugfs entries are as follows::

    /sys/kernel/debug/ceph/
    ....
    ....
    /sys/kernel/debug/ceph/dev_support
    /sys/kernel/debug/ceph/dev_support/v2
    ....
    ....

Note that there is no entry signifying v1 mount syntax. That's because the kernel still supports mounting with old syntax and older kernels do not have debug entries for the same.

Venky Shankar (2):
ceph: add helpers to create/cleanup debugfs sub-directories under
"ceph" directory
ceph: add debugfs entries for v2 (new) mount syntax support

 fs/ceph/debugfs.c            | 28 ++++++++++++++++++++++++++++
 fs/ceph/super.c              |  3 +++
 fs/ceph/super.h              |  2 ++
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 5 files changed, 60 insertions(+), 2 deletions(-)
Signature · 1 line
2.27.0
#1 lore raw

RFC 1/2 ceph: add helpers to create/cleanup debugfs sub-directories under "ceph" directory

Callers can use this helper to create a subdirectory under "ceph" directory in debugfs to place custom files for exporting information to userspace.

---
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 2 files changed, 27 insertions(+), 2 deletions(-)
include/linux/ceph/debugfs.h +3 -0
index 8b3a1a7a953a..c372e6cb8aae 100644
@@ -10,5 +10,8 @@ extern void ceph_debugfs_cleanup(void);
 extern void ceph_debugfs_client_init(struct ceph_client *client);
 extern void ceph_debugfs_client_cleanup(struct ceph_client *client);
 
+extern struct dentry *ceph_debugfs_create_subdir(const char *subdir);
+extern void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry);
+
 #endif
 
net/ceph/debugfs.c +24 -2
index 2110439f8a24..eabac20b3ff4 100644
@@ -404,6 +404,18 @@ void ceph_debugfs_cleanup(void)
 	debugfs_remove(ceph_debugfs_dir);
 }
 
+struct dentry *ceph_debugfs_create_subdir(const char *subdir)
+{
+	return debugfs_create_dir(subdir, ceph_debugfs_dir);
+}
+EXPORT_SYMBOL(ceph_debugfs_create_subdir);
+
+void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry)
+{
+	debugfs_remove(subdir_dentry);
+}
+EXPORT_SYMBOL(ceph_debugfs_cleanup_subdir);
+
 void ceph_debugfs_client_init(struct ceph_client *client)
 {
 	char name[80];
@@ -413,7 +425,7 @@ void ceph_debugfs_client_init(struct ceph_client *client)
 
 	dout("ceph_debugfs_client_init %p %s\n", client, name);
 
-	client->debugfs_dir = debugfs_create_dir(name, ceph_debugfs_dir);
+	client->debugfs_dir = ceph_debugfs_create_subdir(name);
 
 	client->monc.debugfs_file = debugfs_create_file("monc",
 						      0400,
@@ -454,7 +466,7 @@ void ceph_debugfs_client_cleanup(struct ceph_client *client)
 	debugfs_remove(client->debugfs_monmap);
 	debugfs_remove(client->osdc.debugfs_file);
 	debugfs_remove(client->monc.debugfs_file);
-	debugfs_remove(client->debugfs_dir);
+	ceph_debugfs_cleanup_subdir(client->debugfs_dir);
 }
 
 #else  /* CONFIG_DEBUG_FS */
@@ -475,4 +487,14 @@ void ceph_debugfs_client_cleanup(struct ceph_client *client)
 {
 }
 
+struct dentry *ceph_debugfs_create_subdir(const char *subdir)
+{
+}
+EXPORT_SYMBOL(ceph_debugfs_create_subdir);
+
+void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry)
+{
+}
+EXPORT_SYMBOL(ceph_debugfs_cleanup_subdir);
+
 #endif  /* CONFIG_DEBUG_FS */
Signature · 1 line
2.27.0
#2 lore raw

Re: [RFC 1/2] ceph: add helpers to create/cleanup debugfs sub-directories under "ceph" directory

[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch]

Hi Venky,

[FYI, it's a private test report for your RFC patch.]
[auto build test ERROR on ceph-client/for-linus]
[also build test ERROR on v5.14-rc6 next-20210818]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Venky-Shankar/ceph-add-debugfs-entries-signifying-new-mount-syntax-support/20210818-140321
base:   https://github.com/ceph/ceph-client.git for-linus
config: hexagon-randconfig-r025-20210817 (attached as .config)
compiler: clang version 12.0.0

reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# https://github.com/0day-ci/linux/commit/93ace7e15eedcf62ee9a96c7396482133470e651
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Venky-Shankar/ceph-add-debugfs-entries-signifying-new-mount-syntax-support/20210818-140321 git checkout 93ace7e15eedcf62ee9a96c7396482133470e651
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=hexagon

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>):

>> net/ceph/debugfs.c:492:1: error: non-void function does not return a value [-Werror,-Wreturn-type]

}
^
1 error generated.

vim +492 net/ceph/debugfs.c

489
490 struct dentry *ceph_debugfs_create_subdir(const char *subdir)
491 {

 > 492	}

493 EXPORT_SYMBOL(ceph_debugfs_create_subdir);
494

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#3 lore raw

Re: [RFC 1/2] ceph: add helpers to create/cleanup debugfs sub-directories under "ceph" directory

In reply to Venky Shankar

There's a build issue with non-debugfs builds -- I'll fix it up and resend.

Quoted 83 lines · Venky Shankar

Callers can use this helper to create a subdirectory under "ceph" directory in debugfs to place custom files for exporting information to userspace.

---
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 2 files changed, 27 insertions(+), 2 deletions(-)
include/linux/ceph/debugfs.h +3 -0
index 8b3a1a7a953a..c372e6cb8aae 100644
@@ -10,5 +10,8 @@ extern void ceph_debugfs_cleanup(void);
 extern void ceph_debugfs_client_init(struct ceph_client *client);
 extern void ceph_debugfs_client_cleanup(struct ceph_client *client);
 
+extern struct dentry *ceph_debugfs_create_subdir(const char *subdir);
+extern void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry);
+
 #endif
 
net/ceph/debugfs.c +24 -3
index 2110439f8a24..eabac20b3ff4 100644
@@ -404,6 +404,18 @@ void ceph_debugfs_cleanup(void)
        debugfs_remove(ceph_debugfs_dir);
 }
 
+struct dentry *ceph_debugfs_create_subdir(const char *subdir)
+{
+       return debugfs_create_dir(subdir, ceph_debugfs_dir);
+}
+EXPORT_SYMBOL(ceph_debugfs_create_subdir);
+
+void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry)
+{
+       debugfs_remove(subdir_dentry);
+}
+EXPORT_SYMBOL(ceph_debugfs_cleanup_subdir);
+
 void ceph_debugfs_client_init(struct ceph_client *client)
 {
        char name[80];
@@ -413,7 +425,7 @@ void ceph_debugfs_client_init(struct ceph_client *client)
 
        dout("ceph_debugfs_client_init %p %s\n", client, name);
 
-       client->debugfs_dir = debugfs_create_dir(name, ceph_debugfs_dir);
+       client->debugfs_dir = ceph_debugfs_create_subdir(name);
 
        client->monc.debugfs_file = debugfs_create_file("monc",
                                                      0400,
@@ -454,7 +466,7 @@ void ceph_debugfs_client_cleanup(struct ceph_client *client)
        debugfs_remove(client->debugfs_monmap);
        debugfs_remove(client->osdc.debugfs_file);
        debugfs_remove(client->monc.debugfs_file);
-       debugfs_remove(client->debugfs_dir);
+       ceph_debugfs_cleanup_subdir(client->debugfs_dir);
 }
 
 #else  /* CONFIG_DEBUG_FS */
@@ -475,4 +487,14 @@ void ceph_debugfs_client_cleanup(struct ceph_client *client)
 {
 }
 
+struct dentry *ceph_debugfs_create_subdir(const char *subdir)
+{
+}
+EXPORT_SYMBOL(ceph_debugfs_create_subdir);
+
+void ceph_debugfs_cleanup_subdir(struct dentry *subdir_dentry)
+{
+}
+EXPORT_SYMBOL(ceph_debugfs_cleanup_subdir);
+
 #endif  /* CONFIG_DEBUG_FS */
--

2.27.0

Signature · 2 lines
Cheers,
Venky
#4 lore raw

RFC 2/2 ceph: add debugfs entries for v2 (new) mount syntax support

In reply to Venky Shankar
---
 fs/ceph/debugfs.c | 28 ++++++++++++++++++++++++++++
 fs/ceph/super.c   |  3 +++
 fs/ceph/super.h   |  2 ++
 3 files changed, 33 insertions(+)
fs/ceph/debugfs.c +28 -0
index 66989c880adb..d19f15ace781 100644
@@ -22,6 +22,12 @@
 #include "mds_client.h"
 #include "metric.h"
 
+#define MNT_DEV_SUPPORT_DIR   "dev_support"
+#define MNT_DEV_V2_FILE  "v2"
+
+static struct dentry *ceph_mnt_dev_support_dir;
+static struct dentry *ceph_mnt_dev_v2_file;
+
 static int mdsmap_show(struct seq_file *s, void *p)
 {
 	int i;
@@ -416,6 +422,20 @@ void ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 						  &status_fops);
 }
 
+void ceph_fs_debugfs_mnt_dev_init(void)
+{
+	ceph_mnt_dev_support_dir = ceph_debugfs_create_subdir(MNT_DEV_SUPPORT_DIR);
+	ceph_mnt_dev_v2_file = debugfs_create_file(MNT_DEV_V2_FILE,
+						   0400,
+						   ceph_mnt_dev_support_dir,
+						   NULL, NULL);
+}
+
+void ceph_fs_debugfs_mnt_dev_cleanup(void)
+{
+	debugfs_remove(ceph_mnt_dev_v2_file);
+	ceph_debugfs_cleanup_subdir(ceph_mnt_dev_support_dir);
+}
 
 #else  /* CONFIG_DEBUG_FS */
 
@@ -427,4 +447,12 @@ void ceph_fs_debugfs_cleanup(struct ceph_fs_client *fsc)
 {
 }
 
+void ceph_fs_debugfs_mnt_dev_init(void)
+{
+}
+
+void ceph_fs_debugfs_mnt_dev_cleanup(void)
+{
+}
+
 #endif  /* CONFIG_DEBUG_FS */
fs/ceph/super.c +3 -0
index 609ffc8c2d78..21e4a8249afd 100644
@@ -1404,6 +1404,8 @@ static int __init init_ceph(void)
 	if (ret)
 		goto out_caches;
 
+	ceph_fs_debugfs_mnt_dev_init();
+
 	pr_info("loaded (mds proto %d)\n", CEPH_MDSC_PROTOCOL);
 
 	return 0;
@@ -1417,6 +1419,7 @@ static int __init init_ceph(void)
 static void __exit exit_ceph(void)
 {
 	dout("exit_ceph\n");
+	ceph_fs_debugfs_mnt_dev_cleanup();
 	unregister_filesystem(&ceph_fs_type);
 	destroy_caches();
 }
fs/ceph/super.h +2 -0
index 8f71184b7c85..3c63c1adcfaa 100644
@@ -1231,6 +1231,8 @@ extern int ceph_locks_to_pagelist(struct ceph_filelock *flocks,
 /* debugfs.c */
 extern void ceph_fs_debugfs_init(struct ceph_fs_client *client);
 extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client *client);
+extern void ceph_fs_debugfs_mnt_dev_init(void);
+extern void ceph_fs_debugfs_mnt_dev_cleanup(void);
 
 /* quota.c */
 static inline bool __ceph_has_any_quota(struct ceph_inode_info *ci)
Signature · 1 line
2.27.0
#5 lore raw

Re: [RFC 2/2] ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 2 lines · Venky Shankar

[...]

Is "debugfs" the right place for this? I do wonder if that can be dropped/disabled via some obscure kernel config?

Also "debugX" doesn't sound like the proper place for a feature flag of the kernel. I just did a quick check on my system and I do see:

$ ls /sys/fs/ext4/features
batched_discard  casefold  encryption  fast_commit  lazy_itable_init
meta_bg_resize  metadata_csum_seed  test_dummy_encryption_v2  verity

Perhaps we need something similar for fs/ceph?

Signature · 5 lines
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
#6 lore raw

Re: [RFC 2/2] ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 7 lines · Patrick Donnelly
Quoted 2 lines · Venky Shankar

[...]

Is "debugfs" the right place for this? I do wonder if that can be dropped/disabled via some obscure kernel config?

The primary use for this (v2 syntax entry) is for catching bugs in v2 mount syntax implementation which sounds more like a form of "debugging". Sysfs represents the whole device model as seen from the kernel.

And, sysfs is optional too (CONFIG_SYSFS).

Quoted 17 lines

Also "debugX" doesn't sound like the proper place for a feature flag of the kernel. I just did a quick check on my system and I do see:

$ ls /sys/fs/ext4/features
batched_discard  casefold  encryption  fast_commit  lazy_itable_init
meta_bg_resize  metadata_csum_seed  test_dummy_encryption_v2  verity

Perhaps we need something similar for fs/ceph?

--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

Signature · 2 lines
Cheers,
Venky
#7 lore raw

Re: [RFC 2/2] ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 7 lines · Venky Shankar
Quoted 5 lines · Patrick Donnelly
Quoted 2 lines · Venky Shankar

[...]

Quoted 7 lines
Quoted 2 lines

Is "debugfs" the right place for this? I do wonder if that can be dropped/disabled via some obscure kernel config?

The primary use for this (v2 syntax entry) is for catching bugs in v2 mount syntax implementation which sounds more like a form of "debugging". Sysfs represents the whole device model as seen from the kernel.

So, Jeff in another thread suggested that we make these debugfs entries generic (client_features). In that case, I'm ok with having these in sysfs.

Quoted 25 lines

And, sysfs is optional too (CONFIG_SYSFS).

Quoted 8 lines

Also "debugX" doesn't sound like the proper place for a feature flag of the kernel. I just did a quick check on my system and I do see:

$ ls /sys/fs/ext4/features
batched_discard  casefold  encryption  fast_commit  lazy_itable_init
meta_bg_resize  metadata_csum_seed  test_dummy_encryption_v2  verity
Quoted 9 lines

Perhaps we need something similar for fs/ceph?

--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

--
Cheers,
Venky

Signature · 2 lines
Cheers,
Venky
#8 lore raw

Re: [RFC 2/2] ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 11 lines · Venky Shankar
Quoted 10 lines · Venky Shankar
Quoted 7 lines · Patrick Donnelly
Quoted 2 lines · Venky Shankar

[...]

Is "debugfs" the right place for this? I do wonder if that can be dropped/disabled via some obscure kernel config?

Quoted 9 lines
Quoted 4 lines

The primary use for this (v2 syntax entry) is for catching bugs in v2 mount syntax implementation which sounds more like a form of "debugging". Sysfs represents the whole device model as seen from the kernel.

So, Jeff in another thread suggested that we make these debugfs entries generic (client_features). In that case, I'm ok with having these in sysfs.

I think debugfs is fine for this. This is mainly for teuthology, and in that case we'll have debugfs available.

One of the reasons to use debugfs here is that stuff in there is specifically _not_ considered part of the kernel's ABI, so we can more easily make changes to it in the future w/o worrying as much about backward compatibility.

Quoted 28 lines
Quoted 25 lines

And, sysfs is optional too (CONFIG_SYSFS).

Quoted 8 lines

Also "debugX" doesn't sound like the proper place for a feature flag of the kernel. I just did a quick check on my system and I do see:

$ ls /sys/fs/ext4/features
batched_discard  casefold  encryption  fast_commit  lazy_itable_init
meta_bg_resize  metadata_csum_seed  test_dummy_encryption_v2  verity
Quoted 9 lines

Perhaps we need something similar for fs/ceph?

--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D

--
Cheers,
Venky

Signature · 1 line
Jeff Layton <[email protected]>
#9 lore raw
In reply to Venky Shankar
Quoted 6 lines · Venky Shankar

[This is based on top of new mount syntax series]

Patrick proposed the idea of having debugfs entries to signify if kernel supports the new (v2) mount syntax. The primary use of this information is to catch any bugs in the new syntax implementation.

Quoted 7 lines

This would be done as follows::

The userspace mount helper tries to mount using the new mount syntax and fallsback to using old syntax if the mount using new syntax fails. However, a bug in the new mount syntax implementation can silently result in the mount helper switching to old syntax.

Is this a known bug you're talking about or are you just speculating about the potential for bugs there?

Quoted 25 lines

So, the debugfs entries can be relied upon by the mount helper to check if the kernel supports the new mount syntax. Cases when the mount using the new syntax fails, but the kernel does support the new mount syntax, the mount helper could probably log before switching to the old syntax (or fail the mount altogether when run in test mode).

Debugfs entries are as follows::

    /sys/kernel/debug/ceph/
    ....
    ....
    /sys/kernel/debug/ceph/dev_support
    /sys/kernel/debug/ceph/dev_support/v2
    ....
    ....

Note that there is no entry signifying v1 mount syntax. That's because the kernel still supports mounting with old syntax and older kernels do not have debug entries for the same.

Venky Shankar (2):
ceph: add helpers to create/cleanup debugfs sub-directories under
"ceph" directory
ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 7 lines
 fs/ceph/debugfs.c            | 28 ++++++++++++++++++++++++++++
 fs/ceph/super.c              |  3 +++
 fs/ceph/super.h              |  2 ++
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 5 files changed, 60 insertions(+), 2 deletions(-)

I'm not a huge fan of this approach overall as it requires that you have access to debugfs, and that's not guaranteed to be available everywhere. If you want to add this for debugging purposes, that's fine, but I don't think you want the mount helper to rely on this infrastructure.

Signature · 1 line
Jeff Layton <[email protected]>
#10 lore raw
Quoted 10 lines · Jeff Layton
Quoted 8 lines · Venky Shankar

[This is based on top of new mount syntax series]

Patrick proposed the idea of having debugfs entries to signify if kernel supports the new (v2) mount syntax. The primary use of this information is to catch any bugs in the new syntax implementation.

This would be done as follows::

Quoted 8 lines
Quoted 5 lines

The userspace mount helper tries to mount using the new mount syntax and fallsback to using old syntax if the mount using new syntax fails. However, a bug in the new mount syntax implementation can silently result in the mount helper switching to old syntax.

Is this a known bug you're talking about or are you just speculating about the potential for bugs there?

Potential bugs.

Quoted 34 lines
Quoted 26 lines

So, the debugfs entries can be relied upon by the mount helper to check if the kernel supports the new mount syntax. Cases when the mount using the new syntax fails, but the kernel does support the new mount syntax, the mount helper could probably log before switching to the old syntax (or fail the mount altogether when run in test mode).

Debugfs entries are as follows::

    /sys/kernel/debug/ceph/
    ....
    ....
    /sys/kernel/debug/ceph/dev_support
    /sys/kernel/debug/ceph/dev_support/v2
    ....
    ....

Note that there is no entry signifying v1 mount syntax. That's because the kernel still supports mounting with old syntax and older kernels do not have debug entries for the same.

Venky Shankar (2):
ceph: add helpers to create/cleanup debugfs sub-directories under
"ceph" directory
ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 7 lines
 fs/ceph/debugfs.c            | 28 ++++++++++++++++++++++++++++
 fs/ceph/super.c              |  3 +++
 fs/ceph/super.h              |  2 ++
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 5 files changed, 60 insertions(+), 2 deletions(-)
Quoted 4 lines

I'm not a huge fan of this approach overall as it requires that you have access to debugfs, and that's not guaranteed to be available everywhere. If you want to add this for debugging purposes, that's fine, but I don't think you want the mount helper to rely on this infrastructure.

Right. The use-case here is probably to rely on it during teuthology tests where the mount fails (and the tests) when using the new syntax but the kernel has v2 syntax support.

I recall the discussion on having some sort of --no-fallback option to not fall-back to old syntax, but since we have the debugfs entries, we may as well rely on those (at least for testing).

Quoted 4 lines

--
Jeff Layton <[email protected]>

Signature · 2 lines
Cheers,
Venky
#11 lore raw
Quoted 61 lines · Venky Shankar
Quoted 10 lines · Jeff Layton
Quoted 8 lines · Venky Shankar

[This is based on top of new mount syntax series]

Patrick proposed the idea of having debugfs entries to signify if kernel supports the new (v2) mount syntax. The primary use of this information is to catch any bugs in the new syntax implementation.

This would be done as follows::

Quoted 8 lines
Quoted 5 lines

The userspace mount helper tries to mount using the new mount syntax and fallsback to using old syntax if the mount using new syntax fails. However, a bug in the new mount syntax implementation can silently result in the mount helper switching to old syntax.

Is this a known bug you're talking about or are you just speculating about the potential for bugs there?

Potential bugs.

Quoted 34 lines
Quoted 26 lines

So, the debugfs entries can be relied upon by the mount helper to check if the kernel supports the new mount syntax. Cases when the mount using the new syntax fails, but the kernel does support the new mount syntax, the mount helper could probably log before switching to the old syntax (or fail the mount altogether when run in test mode).

Debugfs entries are as follows::

    /sys/kernel/debug/ceph/
    ....
    ....
    /sys/kernel/debug/ceph/dev_support
    /sys/kernel/debug/ceph/dev_support/v2
    ....
    ....

Note that there is no entry signifying v1 mount syntax. That's because the kernel still supports mounting with old syntax and older kernels do not have debug entries for the same.

Venky Shankar (2):
ceph: add helpers to create/cleanup debugfs sub-directories under
"ceph" directory
ceph: add debugfs entries for v2 (new) mount syntax support

Quoted 7 lines
 fs/ceph/debugfs.c            | 28 ++++++++++++++++++++++++++++
 fs/ceph/super.c              |  3 +++
 fs/ceph/super.h              |  2 ++
 include/linux/ceph/debugfs.h |  3 +++
 net/ceph/debugfs.c           | 26 ++++++++++++++++++++++++--
 5 files changed, 60 insertions(+), 2 deletions(-)
Quoted 4 lines

I'm not a huge fan of this approach overall as it requires that you have access to debugfs, and that's not guaranteed to be available everywhere. If you want to add this for debugging purposes, that's fine, but I don't think you want the mount helper to rely on this infrastructure.

Quoted 8 lines

Right. The use-case here is probably to rely on it during teuthology tests where the mount fails (and the tests) when using the new syntax but the kernel has v2 syntax support.

I recall the discussion on having some sort of --no-fallback option to not fall-back to old syntax, but since we have the debugfs entries, we may as well rely on those (at least for testing).

Ok, I think that sounds reasonable.

Signature · 1 line
Jeff Layton <[email protected]>