linux.git/fs/tracefs, branch v6.18.21 Clone of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git tracefs: fix a leak in eventfs_create_events_dir() 2025-12-18T13:03:02+00:00 Al Viro viro@zeniv.linux.org.uk 2025-10-25T22:13:48+00:00 bc2b1bbc0329900b97034b6ad3728111fb4f4f01 [ Upstream commit 798a401660a151633cb171738a72a8f1efb9b0b4 ] If we have LOCKDOWN_TRACEFS, the function bails out - *after* having locked the parent directory and without bothering to undo that. Just check it before tracefs_start_creating()... Fixes: e24709454c45 "tracefs/eventfs: Add missing lockdown checks" Acked-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 798a401660a151633cb171738a72a8f1efb9b0b4 ]

If we have LOCKDOWN_TRACEFS, the function bails out - *after*
having locked the parent directory and without bothering to
undo that.  Just check it before tracefs_start_creating()...

Fixes: e24709454c45 "tracefs/eventfs: Add missing lockdown checks"
Acked-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Merge tag 'pull-rpc_pipefs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2025-07-28T16:56:09+00:00 Linus Torvalds torvalds@linux-foundation.org 2025-07-28T16:56:09+00:00 ddf52f12ef500d9f2a5e325e0c86449f594abb25 Pull rpc_pipefs updates from Al Viro: "Massage rpc_pipefs to use saner primitives and clean up the APIs provided to the rest of the kernel" * tag 'pull-rpc_pipefs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: rpc_create_client_dir(): return 0 or -E... rpc_create_client_dir(): don't bother with rpc_populate() rpc_new_dir(): the last argument is always NULL rpc_pipe: expand the calls of rpc_mkdir_populate() rpc_gssd_dummy_populate(): don't bother with rpc_populate() rpc_mkpipe_dentry(): switch to simple_start_creating() rpc_pipe: saner primitive for creating regular files rpc_pipe: saner primitive for creating subdirectories rpc_pipe: don't overdo directory locking rpc_mkpipe_dentry(): saner calling conventions rpc_unlink(): saner calling conventions rpc_populate(): lift cleanup into callers rpc_unlink(): use simple_recursive_removal() rpc_{rmdir_,}depopulate(): use simple_recursive_removal() instead rpc_pipe: clean failure exits in fill_super new helper: simple_start_creating() 
Pull rpc_pipefs updates from Al Viro:
 "Massage rpc_pipefs to use saner primitives and clean up the APIs
  provided to the rest of the kernel"

* tag 'pull-rpc_pipefs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  rpc_create_client_dir(): return 0 or -E...
  rpc_create_client_dir(): don't bother with rpc_populate()
  rpc_new_dir(): the last argument is always NULL
  rpc_pipe: expand the calls of rpc_mkdir_populate()
  rpc_gssd_dummy_populate(): don't bother with rpc_populate()
  rpc_mkpipe_dentry(): switch to simple_start_creating()
  rpc_pipe: saner primitive for creating regular files
  rpc_pipe: saner primitive for creating subdirectories
  rpc_pipe: don't overdo directory locking
  rpc_mkpipe_dentry(): saner calling conventions
  rpc_unlink(): saner calling conventions
  rpc_populate(): lift cleanup into callers
  rpc_unlink(): use simple_recursive_removal()
  rpc_{rmdir_,}depopulate(): use simple_recursive_removal() instead
  rpc_pipe: clean failure exits in fill_super
  new helper: simple_start_creating()
new helper: simple_start_creating() 2025-07-03T02:44:38+00:00 Al Viro viro@zeniv.linux.org.uk 2025-03-23T01:06:11+00:00 59200f45267481582f4e42334a510f01d0b89449 Set the things up for kernel-initiated creation of object in a tree-in-dcache filesystem. With respect to locking it's an equivalent of filename_create() - we either get a negative dentry with locked parent, or ERR_PTR() and no locks taken. tracefs and debugfs had that open-coded as part of their object creation machinery; switched to calling new helper. Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
Set the things up for kernel-initiated creation of object in
a tree-in-dcache filesystem.  With respect to locking it's
an equivalent of filename_create() - we either get a negative
dentry with locked parent, or ERR_PTR() and no locks taken.

tracefs and debugfs had that open-coded as part of their
object creation machinery; switched to calling new helper.

Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
tracefs: Add d_delete to remove negative dentries 2025-06-11T17:34:30+00:00 Steven Rostedt rostedt@goodmis.org 2025-06-11T16:18:15+00:00 d9b13cdad80dc11d74408cf201939a946e9303a6 If a lookup in tracefs is done on a file that does not exist, it leaves a dentry hanging around until memory pressure removes it. But eventfs dentries should hang around as when their ref count goes to zero, it requires more work to recreate it. For the rest of the tracefs dentries, they hang around as their dentry is used as a descriptor for the tracing system. But if a file lookup happens for a file in tracefs that does not exist, it should be deleted. Add a .d_delete callback that checks if dentry->fsdata is set or not. Only eventfs dentries set fsdata so if it has content it should not be deleted and should hang around in the cache. Reported-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
If a lookup in tracefs is done on a file that does not exist, it leaves a
dentry hanging around until memory pressure removes it. But eventfs
dentries should hang around as when their ref count goes to zero, it
requires more work to recreate it. For the rest of the tracefs dentries,
they hang around as their dentry is used as a descriptor for the tracing
system. But if a file lookup happens for a file in tracefs that does not
exist, it should be deleted.

Add a .d_delete callback that checks if dentry->fsdata is set or not. Only
eventfs dentries set fsdata so if it has content it should not be deleted
and should hang around in the cache.

Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
new helper: set_default_d_op() 2025-06-11T02:21:16+00:00 Al Viro viro@zeniv.linux.org.uk 2025-02-24T00:39:47+00:00 05fb0e666495cda068c068a681ecbbf8e57324d0 ... to be used instead of manually assigning to ->s_d_op. All in-tree filesystem converted (and field itself is renamed, so any out-of-tree ones in need of conversion will be caught by compiler). Reviewed-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
... to be used instead of manually assigning to ->s_d_op.
All in-tree filesystem converted (and field itself is renamed,
so any out-of-tree ones in need of conversion will be caught
by compiler).

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
VFS: rename lookup_one_len family to lookup_noperm and remove permission check 2025-04-08T09:24:36+00:00 NeilBrown neil@brown.name 2025-03-19T03:01:35+00:00 fa6fe07d1536361a227d655e69ca270faf28fdbe The lookup_one_len family of functions is (now) only used internally by a filesystem on itself either - in a context where permission checking is irrelevant such as by a virtual filesystem populating itself, or xfs accessing its ORPHANAGE or dquota accessing the quota file; or - in a context where a permission check (MAY_EXEC on the parent) has just been performed such as a network filesystem finding in "silly-rename" file in the same directory. This is also the context after the _parentat() functions where currently lookup_one_qstr_excl() is used. So the permission check is pointless. The name "one_len" is unhelpful in understanding the purpose of these functions and should be changed. Most of the callers pass the len as "strlen()" so using a qstr and QSTR() can simplify the code. This patch renames these functions (include lookup_positive_unlocked() which is part of the family despite the name) to have a name based on "lookup_noperm". They are changed to receive a 'struct qstr' instead of separate name and len. In a few cases the use of QSTR() results in a new call to strlen(). try_lookup_noperm() takes a pointer to a qstr instead of the whole qstr. This is consistent with d_hash_and_lookup() (which is nearly identical) and useful for lookup_noperm_unlocked(). The new lookup_noperm_common() doesn't take a qstr yet. That will be tidied up in a subsequent patch. Signed-off-by: NeilBrown <neil@brown.name> Link: https://lore.kernel.org/r/20250319031545.2999807-5-neil@brown.name Signed-off-by: Christian Brauner <brauner@kernel.org> 
The lookup_one_len family of functions is (now) only used internally by
a filesystem on itself either
- in a context where permission checking is irrelevant such as by a
  virtual filesystem populating itself, or xfs accessing its ORPHANAGE
  or dquota accessing the quota file; or
- in a context where a permission check (MAY_EXEC on the parent) has just
  been performed such as a network filesystem finding in "silly-rename"
  file in the same directory.  This is also the context after the
  _parentat() functions where currently lookup_one_qstr_excl() is used.

So the permission check is pointless.

The name "one_len" is unhelpful in understanding the purpose of these
functions and should be changed.  Most of the callers pass the len as
"strlen()" so using a qstr and QSTR() can simplify the code.

This patch renames these functions (include lookup_positive_unlocked()
which is part of the family despite the name) to have a name based on
"lookup_noperm".  They are changed to receive a 'struct qstr' instead
of separate name and len.  In a few cases the use of QSTR() results in a
new call to strlen().

try_lookup_noperm() takes a pointer to a qstr instead of the whole
qstr.  This is consistent with d_hash_and_lookup() (which is nearly
identical) and useful for lookup_noperm_unlocked().

The new lookup_noperm_common() doesn't take a qstr yet.  That will be
tidied up in a subsequent patch.

Signed-off-by: NeilBrown <neil@brown.name>
Link: https://lore.kernel.org/r/20250319031545.2999807-5-neil@brown.name
Signed-off-by: Christian Brauner <brauner@kernel.org>
Change inode_operations.mkdir to return struct dentry * 2025-02-27T19:00:17+00:00 NeilBrown neilb@suse.de 2025-02-27T01:32:53+00:00 88d5baf69082e5b410296435008329676b687549 Some filesystems, such as NFS, cifs, ceph, and fuse, do not have complete control of sequencing on the actual filesystem (e.g. on a different server) and may find that the inode created for a mkdir request already exists in the icache and dcache by the time the mkdir request returns. For example, if the filesystem is mounted twice the directory could be visible on the other mount before it is on the original mount, and a pair of name_to_handle_at(), open_by_handle_at() calls could instantiate the directory inode with an IS_ROOT() dentry before the first mkdir returns. This means that the dentry passed to ->mkdir() may not be the one that is associated with the inode after the ->mkdir() completes. Some callers need to interact with the inode after the ->mkdir completes and they currently need to perform a lookup in the (rare) case that the dentry is no longer hashed. This lookup-after-mkdir requires that the directory remains locked to avoid races. Planned future patches to lock the dentry rather than the directory will mean that this lookup cannot be performed atomically with the mkdir. To remove this barrier, this patch changes ->mkdir to return the resulting dentry if it is different from the one passed in. Possible returns are: NULL - the directory was created and no other dentry was used ERR_PTR() - an error occurred non-NULL - this other dentry was spliced in This patch only changes file-systems to return "ERR_PTR(err)" instead of "err" or equivalent transformations. Subsequent patches will make further changes to some file-systems to return a correct dentry. Not all filesystems reliably result in a positive hashed dentry: - NFS, cifs, hostfs will sometimes need to perform a lookup of the name to get inode information. Races could result in this returning something different. Note that this lookup is non-atomic which is what we are trying to avoid. Placing the lookup in filesystem code means it only happens when the filesystem has no other option. - kernfs and tracefs leave the dentry negative and the ->revalidate operation ensures that lookup will be called to correctly populate the dentry. This could be fixed but I don't think it is important to any of the users of vfs_mkdir() which look at the dentry. The recommendation to use d_drop();d_splice_alias() is ugly but fits with current practice. A planned future patch will change this. Reviewed-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: NeilBrown <neilb@suse.de> Link: https://lore.kernel.org/r/20250227013949.536172-2-neilb@suse.de Signed-off-by: Christian Brauner <brauner@kernel.org> 
Some filesystems, such as NFS, cifs, ceph, and fuse, do not have
complete control of sequencing on the actual filesystem (e.g.  on a
different server) and may find that the inode created for a mkdir
request already exists in the icache and dcache by the time the mkdir
request returns.  For example, if the filesystem is mounted twice the
directory could be visible on the other mount before it is on the
original mount, and a pair of name_to_handle_at(), open_by_handle_at()
calls could instantiate the directory inode with an IS_ROOT() dentry
before the first mkdir returns.

This means that the dentry passed to ->mkdir() may not be the one that
is associated with the inode after the ->mkdir() completes.  Some
callers need to interact with the inode after the ->mkdir completes and
they currently need to perform a lookup in the (rare) case that the
dentry is no longer hashed.

This lookup-after-mkdir requires that the directory remains locked to
avoid races.  Planned future patches to lock the dentry rather than the
directory will mean that this lookup cannot be performed atomically with
the mkdir.

To remove this barrier, this patch changes ->mkdir to return the
resulting dentry if it is different from the one passed in.
Possible returns are:
  NULL - the directory was created and no other dentry was used
  ERR_PTR() - an error occurred
  non-NULL - this other dentry was spliced in

This patch only changes file-systems to return "ERR_PTR(err)" instead of
"err" or equivalent transformations.  Subsequent patches will make
further changes to some file-systems to return a correct dentry.

Not all filesystems reliably result in a positive hashed dentry:

- NFS, cifs, hostfs will sometimes need to perform a lookup of
  the name to get inode information.  Races could result in this
  returning something different. Note that this lookup is
  non-atomic which is what we are trying to avoid.  Placing the
  lookup in filesystem code means it only happens when the filesystem
  has no other option.
- kernfs and tracefs leave the dentry negative and the ->revalidate
  operation ensures that lookup will be called to correctly populate
  the dentry.  This could be fixed but I don't think it is important
  to any of the users of vfs_mkdir() which look at the dentry.

The recommendation to use
    d_drop();d_splice_alias()
is ugly but fits with current practice.  A planned future patch will
change this.

Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: NeilBrown <neilb@suse.de>
Link: https://lore.kernel.org/r/20250227013949.536172-2-neilb@suse.de
Signed-off-by: Christian Brauner <brauner@kernel.org>
Pass parent directory inode and expected name to ->d_revalidate() 2025-01-28T00:25:23+00:00 Al Viro viro@zeniv.linux.org.uk 2024-12-08T05:28:51+00:00 5be1fa8abd7b049f51e6e98e75a37eef5ae2c296 ->d_revalidate() often needs to access dentry parent and name; that has to be done carefully, since the locking environment varies from caller to caller. We are not guaranteed that dentry in question will not be moved right under us - not unless the filesystem is such that nothing on it ever gets renamed. It can be dealt with, but that results in boilerplate code that isn't even needed - the callers normally have just found the dentry via dcache lookup and want to verify that it's in the right place; they already have the values of ->d_parent and ->d_name stable. There is a couple of exceptions (overlayfs and, to less extent, ecryptfs), but for the majority of calls that song and dance is not needed at all. It's easier to make ecryptfs and overlayfs find and pass those values if there's a ->d_revalidate() instance to be called, rather than doing that in the instances. This commit only changes the calling conventions; making use of supplied values is left to followups. NOTE: some instances need more than just the parent - things like CIFS may need to build an entire path from filesystem root, so they need more precautions than the usual boilerplate. This series doesn't do anything to that need - these filesystems have to keep their locking mechanisms (rename_lock loops, use of dentry_path_raw(), private rwsem a-la v9fs). One thing to keep in mind when using name is that name->name will normally point into the pathname being resolved; the filename in question occupies name->len bytes starting at name->name, and there is NUL somewhere after it, but it the next byte might very well be '/' rather than '\0'. Do not ignore name->len. Reviewed-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Gabriel Krisman Bertazi <gabriel@krisman.be> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
->d_revalidate() often needs to access dentry parent and name; that has
to be done carefully, since the locking environment varies from caller
to caller.  We are not guaranteed that dentry in question will not be
moved right under us - not unless the filesystem is such that nothing
on it ever gets renamed.

It can be dealt with, but that results in boilerplate code that isn't
even needed - the callers normally have just found the dentry via dcache
lookup and want to verify that it's in the right place; they already
have the values of ->d_parent and ->d_name stable.  There is a couple
of exceptions (overlayfs and, to less extent, ecryptfs), but for the
majority of calls that song and dance is not needed at all.

It's easier to make ecryptfs and overlayfs find and pass those values if
there's a ->d_revalidate() instance to be called, rather than doing that
in the instances.

This commit only changes the calling conventions; making use of supplied
values is left to followups.

NOTE: some instances need more than just the parent - things like CIFS
may need to build an entire path from filesystem root, so they need
more precautions than the usual boilerplate.  This series doesn't
do anything to that need - these filesystems have to keep their locking
mechanisms (rename_lock loops, use of dentry_path_raw(), private rwsem
a-la v9fs).

One thing to keep in mind when using name is that name->name will normally
point into the pathname being resolved; the filename in question occupies
name->len bytes starting at name->name, and there is NUL somewhere after it,
but it the next byte might very well be '/' rather than '\0'.  Do not
ignore name->len.

Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Gabriel Krisman Bertazi <gabriel@krisman.be>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
tracing: Fix tracefs mount options 2024-11-01T12:38:14+00:00 Kalesh Singh kaleshsingh@google.com 2024-10-30T17:17:48+00:00 e4d32142d1de8bcafd90ea5f4f557104f0969c41 Commit 78ff64081949 ("vfs: Convert tracefs to use the new mount API") converted tracefs to use the new mount APIs caused mount options (e.g. gid=<gid>) to not take effect. The tracefs superblock can be updated from multiple paths: - on fs_initcall() to init_trace_printk_function_export() - from a work queue to initialize eventfs tracer_init_tracefs_work_func() - fsconfig() syscall to mount or remount of tracefs The tracefs superblock root inode gets created early on in init_trace_printk_function_export(). With the new mount API, tracefs effectively uses get_tree_single() instead of the old API mount_single(). Previously, mount_single() ensured that the options are always applied to the superblock root inode: (1) If the root inode didn't exist, call fill_super() to create it and apply the options. (2) If the root inode exists, call reconfigure_single() which effectively calls tracefs_apply_options() to parse and apply options to the subperblock's fs_info and inode and remount eventfs (if necessary) On the other hand, get_tree_single() effectively calls vfs_get_super() which: (3) If the root inode doesn't exists, calls fill_super() to create it and apply the options. (4) If the root inode already exists, updates the fs_context root with the superblock's root inode. (4) above is always the case for tracefs mounts, since the super block's root inode will already be created by init_trace_printk_function_export(). This means that the mount options get ignored: - Since it isn't applied to the superblock's root inode, it doesn't get inherited by the children. - Since eventfs is initialized from a separate work queue and before call to mount with the options, and it doesn't get remounted for mount. Ensure that the mount options are applied to the super block and eventfs is remounted to respect the mount options. To understand this better, if fstab has the following: tracefs /sys/kernel/tracing tracefs nosuid,nodev,noexec,gid=tracing 0 0 On boot up, permissions look like: # ls -l /sys/kernel/tracing/trace -rw-r----- 1 root root 0 Nov 1 08:37 /sys/kernel/tracing/trace When it should look like: # ls -l /sys/kernel/tracing/trace -rw-r----- 1 root tracing 0 Nov 1 08:37 /sys/kernel/tracing/trace Link: https://lore.kernel.org/r/536e99d3-345c-448b-adee-a21389d7ab4b@redhat.com/ Cc: Eric Sandeen <sandeen@redhat.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Shuah Khan <shuah@kernel.org> Cc: Ali Zahraee <ahzahraee@gmail.com> Cc: Christian Brauner <brauner@kernel.org> Cc: David Howells <dhowells@redhat.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: stable@vger.kernel.org Fixes: 78ff64081949 ("vfs: Convert tracefs to use the new mount API") Link: https://lore.kernel.org/20241030171928.4168869-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> 
Commit 78ff64081949 ("vfs: Convert tracefs to use the new mount API")
converted tracefs to use the new mount APIs caused mount options
(e.g. gid=<gid>) to not take effect.

The tracefs superblock can be updated from multiple paths:
    - on fs_initcall() to init_trace_printk_function_export()
    - from a work queue to initialize eventfs
      tracer_init_tracefs_work_func()
    - fsconfig() syscall to mount or remount of tracefs

The tracefs superblock root inode gets created early on in
init_trace_printk_function_export().

With the new mount API, tracefs effectively uses get_tree_single() instead
of the old API mount_single().

Previously, mount_single() ensured that the options are always applied to
the superblock root inode:
    (1) If the root inode didn't exist, call fill_super() to create it
        and apply the options.
    (2) If the root inode exists, call reconfigure_single() which
        effectively calls tracefs_apply_options() to parse and apply
        options to the subperblock's fs_info and inode and remount
        eventfs (if necessary)

On the other hand, get_tree_single() effectively calls vfs_get_super()
which:
    (3) If the root inode doesn't exists, calls fill_super() to create it
        and apply the options.
    (4) If the root inode already exists, updates the fs_context root
        with the superblock's root inode.

(4) above is always the case for tracefs mounts, since the super block's
root inode will already be created by init_trace_printk_function_export().

This means that the mount options get ignored:
    - Since it isn't applied to the superblock's root inode, it doesn't
      get inherited by the children.
    - Since eventfs is initialized from a separate work queue and
      before call to mount with the options, and it doesn't get remounted
      for mount.

Ensure that the mount options are applied to the super block and eventfs
is remounted to respect the mount options.

To understand this better, if fstab has the following:

 tracefs  /sys/kernel/tracing  tracefs   nosuid,nodev,noexec,gid=tracing 0  0

On boot up, permissions look like:

 # ls -l /sys/kernel/tracing/trace
 -rw-r----- 1 root root 0 Nov  1 08:37 /sys/kernel/tracing/trace

When it should look like:

 # ls -l /sys/kernel/tracing/trace
 -rw-r----- 1 root tracing 0 Nov  1 08:37 /sys/kernel/tracing/trace

Link: https://lore.kernel.org/r/536e99d3-345c-448b-adee-a21389d7ab4b@redhat.com/

Cc: Eric Sandeen <sandeen@redhat.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Ali Zahraee <ahzahraee@gmail.com>
Cc: Christian Brauner <brauner@kernel.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: stable@vger.kernel.org
Fixes: 78ff64081949 ("vfs: Convert tracefs to use the new mount API")
Link: https://lore.kernel.org/20241030171928.4168869-2-kaleshsingh@google.com
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
eventfs: Use list_del_rcu() for SRCU protected list variable 2024-09-05T14:18:48+00:00 Steven Rostedt rostedt@goodmis.org 2024-09-04T17:16:05+00:00 d2603279c7d645bf0d11fa253b23f1ab48fc8d3c Chi Zhiling reported: We found a null pointer accessing in tracefs[1], the reason is that the variable 'ei_child' is set to LIST_POISON1, that means the list was removed in eventfs_remove_rec. so when access the ei_child->is_freed, the panic triggered. by the way, the following script can reproduce this panic loop1 (){ while true do echo "p:kp submit_bio" > /sys/kernel/debug/tracing/kprobe_events echo "" > /sys/kernel/debug/tracing/kprobe_events done } loop2 (){ while true do tree /sys/kernel/debug/tracing/events/kprobes/ done } loop1 & loop2 [1]: [ 1147.959632][T17331] Unable to handle kernel paging request at virtual address dead000000000150 [ 1147.968239][T17331] Mem abort info: [ 1147.971739][T17331] ESR = 0x0000000096000004 [ 1147.976172][T17331] EC = 0x25: DABT (current EL), IL = 32 bits [ 1147.982171][T17331] SET = 0, FnV = 0 [ 1147.985906][T17331] EA = 0, S1PTW = 0 [ 1147.989734][T17331] FSC = 0x04: level 0 translation fault [ 1147.995292][T17331] Data abort info: [ 1147.998858][T17331] ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000 [ 1148.005023][T17331] CM = 0, WnR = 0, TnD = 0, TagAccess = 0 [ 1148.010759][T17331] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 [ 1148.016752][T17331] [dead000000000150] address between user and kernel address ranges [ 1148.024571][T17331] Internal error: Oops: 0000000096000004 [#1] SMP [ 1148.030825][T17331] Modules linked in: team_mode_loadbalance team nlmon act_gact cls_flower sch_ingress bonding tls macvlan dummy ib_core bridge stp llc veth amdgpu amdxcp mfd_core gpu_sched drm_exec drm_buddy radeon crct10dif_ce video drm_suballoc_helper ghash_ce drm_ttm_helper sha2_ce ttm sha256_arm64 i2c_algo_bit sha1_ce sbsa_gwdt cp210x drm_display_helper cec sr_mod cdrom drm_kms_helper binfmt_misc sg loop fuse drm dm_mod nfnetlink ip_tables autofs4 [last unloaded: tls] [ 1148.072808][T17331] CPU: 3 PID: 17331 Comm: ls Tainted: G W ------- ---- 6.6.43 #2 [ 1148.081751][T17331] Source Version: 21b3b386e948bedd29369af66f3e98ab01b1c650 [ 1148.088783][T17331] Hardware name: Greatwall GW-001M1A-FTF/GW-001M1A-FTF, BIOS KunLun BIOS V4.0 07/16/2020 [ 1148.098419][T17331] pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 1148.106060][T17331] pc : eventfs_iterate+0x2c0/0x398 [ 1148.111017][T17331] lr : eventfs_iterate+0x2fc/0x398 [ 1148.115969][T17331] sp : ffff80008d56bbd0 [ 1148.119964][T17331] x29: ffff80008d56bbf0 x28: ffff001ff5be2600 x27: 0000000000000000 [ 1148.127781][T17331] x26: ffff001ff52ca4e0 x25: 0000000000009977 x24: dead000000000100 [ 1148.135598][T17331] x23: 0000000000000000 x22: 000000000000000b x21: ffff800082645f10 [ 1148.143415][T17331] x20: ffff001fddf87c70 x19: ffff80008d56bc90 x18: 0000000000000000 [ 1148.151231][T17331] x17: 0000000000000000 x16: 0000000000000000 x15: ffff001ff52ca4e0 [ 1148.159048][T17331] x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000 [ 1148.166864][T17331] x11: 0000000000000000 x10: 0000000000000000 x9 : ffff8000804391d0 [ 1148.174680][T17331] x8 : 0000000180000000 x7 : 0000000000000018 x6 : 0000aaab04b92862 [ 1148.182498][T17331] x5 : 0000aaab04b92862 x4 : 0000000080000000 x3 : 0000000000000068 [ 1148.190314][T17331] x2 : 000000000000000f x1 : 0000000000007ea8 x0 : 0000000000000001 [ 1148.198131][T17331] Call trace: [ 1148.201259][T17331] eventfs_iterate+0x2c0/0x398 [ 1148.205864][T17331] iterate_dir+0x98/0x188 [ 1148.210036][T17331] __arm64_sys_getdents64+0x78/0x160 [ 1148.215161][T17331] invoke_syscall+0x78/0x108 [ 1148.219593][T17331] el0_svc_common.constprop.0+0x48/0xf0 [ 1148.224977][T17331] do_el0_svc+0x24/0x38 [ 1148.228974][T17331] el0_svc+0x40/0x168 [ 1148.232798][T17331] el0t_64_sync_handler+0x120/0x130 [ 1148.237836][T17331] el0t_64_sync+0x1a4/0x1a8 [ 1148.242182][T17331] Code: 54ffff6c f9400676 910006d6 f9000676 (b9405300) [ 1148.248955][T17331] ---[ end trace 0000000000000000 ]--- The issue is that list_del() is used on an SRCU protected list variable before the synchronization occurs. This can poison the list pointers while there is a reader iterating the list. This is simply fixed by using list_del_rcu() that is specifically made for this purpose. Link: https://lore.kernel.org/linux-trace-kernel/20240829085025.3600021-1-chizhiling@163.com/ Cc: stable@vger.kernel.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Link: https://lore.kernel.org/20240904131605.640d42b1@gandalf.local.home Fixes: 43aa6f97c2d03 ("eventfs: Get rid of dentry pointers without refcounts") Reported-by: Chi Zhiling <chizhiling@kylinos.cn> Tested-by: Chi Zhiling <chizhiling@kylinos.cn> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> 
Chi Zhiling reported:

  We found a null pointer accessing in tracefs[1], the reason is that the
  variable 'ei_child' is set to LIST_POISON1, that means the list was
  removed in eventfs_remove_rec. so when access the ei_child->is_freed, the
  panic triggered.

  by the way, the following script can reproduce this panic

  loop1 (){
      while true
      do
          echo "p:kp submit_bio" > /sys/kernel/debug/tracing/kprobe_events
          echo "" > /sys/kernel/debug/tracing/kprobe_events
      done
  }
  loop2 (){
      while true
      do
          tree /sys/kernel/debug/tracing/events/kprobes/
      done
  }
  loop1 &
  loop2

  [1]:
  [ 1147.959632][T17331] Unable to handle kernel paging request at virtual address dead000000000150
  [ 1147.968239][T17331] Mem abort info:
  [ 1147.971739][T17331]   ESR = 0x0000000096000004
  [ 1147.976172][T17331]   EC = 0x25: DABT (current EL), IL = 32 bits
  [ 1147.982171][T17331]   SET = 0, FnV = 0
  [ 1147.985906][T17331]   EA = 0, S1PTW = 0
  [ 1147.989734][T17331]   FSC = 0x04: level 0 translation fault
  [ 1147.995292][T17331] Data abort info:
  [ 1147.998858][T17331]   ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
  [ 1148.005023][T17331]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
  [ 1148.010759][T17331]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
  [ 1148.016752][T17331] [dead000000000150] address between user and kernel address ranges
  [ 1148.024571][T17331] Internal error: Oops: 0000000096000004 [#1] SMP
  [ 1148.030825][T17331] Modules linked in: team_mode_loadbalance team nlmon act_gact cls_flower sch_ingress bonding tls macvlan dummy ib_core bridge stp llc veth amdgpu amdxcp mfd_core gpu_sched drm_exec drm_buddy radeon crct10dif_ce video drm_suballoc_helper ghash_ce drm_ttm_helper sha2_ce ttm sha256_arm64 i2c_algo_bit sha1_ce sbsa_gwdt cp210x drm_display_helper cec sr_mod cdrom drm_kms_helper binfmt_misc sg loop fuse drm dm_mod nfnetlink ip_tables autofs4 [last unloaded: tls]
  [ 1148.072808][T17331] CPU: 3 PID: 17331 Comm: ls Tainted: G        W         ------- ----  6.6.43 #2
  [ 1148.081751][T17331] Source Version: 21b3b386e948bedd29369af66f3e98ab01b1c650
  [ 1148.088783][T17331] Hardware name: Greatwall GW-001M1A-FTF/GW-001M1A-FTF, BIOS KunLun BIOS V4.0 07/16/2020
  [ 1148.098419][T17331] pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
  [ 1148.106060][T17331] pc : eventfs_iterate+0x2c0/0x398
  [ 1148.111017][T17331] lr : eventfs_iterate+0x2fc/0x398
  [ 1148.115969][T17331] sp : ffff80008d56bbd0
  [ 1148.119964][T17331] x29: ffff80008d56bbf0 x28: ffff001ff5be2600 x27: 0000000000000000
  [ 1148.127781][T17331] x26: ffff001ff52ca4e0 x25: 0000000000009977 x24: dead000000000100
  [ 1148.135598][T17331] x23: 0000000000000000 x22: 000000000000000b x21: ffff800082645f10
  [ 1148.143415][T17331] x20: ffff001fddf87c70 x19: ffff80008d56bc90 x18: 0000000000000000
  [ 1148.151231][T17331] x17: 0000000000000000 x16: 0000000000000000 x15: ffff001ff52ca4e0
  [ 1148.159048][T17331] x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
  [ 1148.166864][T17331] x11: 0000000000000000 x10: 0000000000000000 x9 : ffff8000804391d0
  [ 1148.174680][T17331] x8 : 0000000180000000 x7 : 0000000000000018 x6 : 0000aaab04b92862
  [ 1148.182498][T17331] x5 : 0000aaab04b92862 x4 : 0000000080000000 x3 : 0000000000000068
  [ 1148.190314][T17331] x2 : 000000000000000f x1 : 0000000000007ea8 x0 : 0000000000000001
  [ 1148.198131][T17331] Call trace:
  [ 1148.201259][T17331]  eventfs_iterate+0x2c0/0x398
  [ 1148.205864][T17331]  iterate_dir+0x98/0x188
  [ 1148.210036][T17331]  __arm64_sys_getdents64+0x78/0x160
  [ 1148.215161][T17331]  invoke_syscall+0x78/0x108
  [ 1148.219593][T17331]  el0_svc_common.constprop.0+0x48/0xf0
  [ 1148.224977][T17331]  do_el0_svc+0x24/0x38
  [ 1148.228974][T17331]  el0_svc+0x40/0x168
  [ 1148.232798][T17331]  el0t_64_sync_handler+0x120/0x130
  [ 1148.237836][T17331]  el0t_64_sync+0x1a4/0x1a8
  [ 1148.242182][T17331] Code: 54ffff6c f9400676 910006d6 f9000676 (b9405300)
  [ 1148.248955][T17331] ---[ end trace 0000000000000000 ]---

The issue is that list_del() is used on an SRCU protected list variable
before the synchronization occurs. This can poison the list pointers while
there is a reader iterating the list.

This is simply fixed by using list_del_rcu() that is specifically made for
this purpose.

Link: https://lore.kernel.org/linux-trace-kernel/20240829085025.3600021-1-chizhiling@163.com/

Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/20240904131605.640d42b1@gandalf.local.home
Fixes: 43aa6f97c2d03 ("eventfs: Get rid of dentry pointers without refcounts")
Reported-by: Chi Zhiling <chizhiling@kylinos.cn>
Tested-by: Chi Zhiling <chizhiling@kylinos.cn>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>