Pull more smb client updates from Steve French:
- fix i_size in fallocate
- two truncate fixes
- utime fix
- minor cleanups
- SMB1 fixes
- improve error check in read
- improve perf of copy file_range (copy_chunk)
* tag 'v6.18-rc-part2-smb-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
cifs: update internal version number
cifs: Add comments for DeletePending assignments in open functions
cifs: Add fallback code path for cifs_mkdir_setinfo()
cifs: Allow fallback code in smb_set_file_info() also for directories
cifs: Query EA $LXMOD in cifs_query_path_info() for WSL reparse points
smb: client: remove cfids_invalidation_worker
smb: client: remove redudant assignment in cifs_strict_fsync()
smb: client: fix race with fallocate(2) and AIO+DIO
smb: client: fix missing timestamp updates after utime(2)
smb: client: fix missing timestamp updates after ftruncate(2)
smb: client: fix missing timestamp updates with O_TRUNC
cifs: Fix copy_to_iter return value check
smb: client: batch SRV_COPYCHUNK entries to cut round trips
smb: client: Omit an if branch in smb2_find_smb_tcon()
smb: client: Return directly after a failed genlmsg_new() in cifs_swn_send_register_message()
smb: client: Use common code in cifs_do_create()
smb: client: Improve unlocking of a mutex in cifs_get_swn_reg()
smb: client: Return a status code only as a constant in cifs_spnego_key_instantiate()
smb: client: Use common code in cifs_lookup()
smb: client: Reduce the scopes for a few variables in two functions
On more places is set DeletePending member to 0. Add comments why is 0 the
correct value. Paths in DELETE_PENDING state cannot be opened by new calls.
So if the newly issued open for that path succeed then it means that the
path cannot be in DELETE_PENDING state.
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Use SMBSetInformation() as a fallback function (when CIFSSMBSetPathInfo()
fails) which can set attribudes on the directory, including changing
read-only attribute.
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
On NT systems, it is possible to do SMB open call also for directories.
Open argument CREATE_NOT_DIR disallows opening directories. So in fallback
code path in smb_set_file_info() remove CREATE_NOT_DIR restriction to allow
it also for directories.
Similar fallback is implemented also in CIFSSMBSetPathInfoFB() function and
this function already allows to call operation for directories.
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
EA $LXMOD is required for WSL non-symlink reparse points.
Fixes: ef86ab131d ("cifs: Fix querying of WSL CHR and BLK reparse points over SMB1")
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
We can do the same cleanup on laundromat.
On invalidate_all_cached_dirs(), run laundromat worker with 0 timeout
and flush it for immediate + sync cleanup.
Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
Remove redudant assignment of @rc as it will be overwritten by the
following cifs_file_flush() call.
Reported-by: Steve French <stfrench@microsoft.com>
Addresses-Coverity: 1665925
Fixes: 210627b0aca9 ("smb: client: fix missing timestamp updates with O_TRUNC")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
AIO+DIO may extend the file size, hence we need to make sure ->i_size
is stable across the entire fallocate(2) operation, otherwise it would
become a truncate and then inode size reduced back down when it
finishes.
Fix this by calling netfs_wait_for_outstanding_io() right after
acquiring ->i_rwsem exclusively in cifs_fallocate() and then guarantee
a stable ->i_size across fallocate(2).
Also call netfs_wait_for_outstanding_io() after truncating pagecache
to avoid any potential races with writeback.
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Fixes: 210627b0aca9 ("smb: client: fix missing timestamp updates with O_TRUNC")
Cc: Frank Sorenson <sorenson@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
Don't reuse open handle when changing timestamps to prevent the server
from disabling automatic timestamp updates as per MS-FSA 2.1.4.17.
---8<---
import os
import time
filename = '/mnt/foo'
def print_stat(prefix):
st = os.stat(filename)
print(prefix, ': ', time.ctime(st.st_atime), time.ctime(st.st_ctime))
fd = os.open(filename, os.O_CREAT|os.O_TRUNC|os.O_WRONLY, 0o644)
print_stat('old')
os.utime(fd, None)
time.sleep(2)
os.write(fd, b'foo')
os.close(fd)
time.sleep(2)
print_stat('new')
---8<---
Before patch:
$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old : Fri Oct 3 14:01:21 2025 Fri Oct 3 14:01:21 2025
new : Fri Oct 3 14:01:21 2025 Fri Oct 3 14:01:21 2025
After patch:
$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old : Fri Oct 3 17:03:34 2025 Fri Oct 3 17:03:34 2025
new : Fri Oct 3 17:03:36 2025 Fri Oct 3 17:03:36 2025
Fixes: b6f2a0f89d ("cifs: for compound requests, use open handle if possible")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: Frank Sorenson <sorenson@redhat.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
Mask off ATTR_MTIME|ATTR_CTIME bits on ATTR_SIZE (e.g. ftruncate(2))
to prevent the client from sending set info calls and then disabling
automatic timestamp updates on server side as per MS-FSA 2.1.4.17.
---8<---
import os
import time
filename = '/mnt/foo'
def print_stat(prefix):
st = os.stat(filename)
print(prefix, ': ', time.ctime(st.st_atime), time.ctime(st.st_ctime))
fd = os.open(filename, os.O_CREAT|os.O_TRUNC|os.O_WRONLY, 0o644)
print_stat('old')
os.ftruncate(fd, 10)
time.sleep(2)
os.write(fd, b'foo')
os.close(fd)
time.sleep(2)
print_stat('new')
---8<---
Before patch:
$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old : Fri Oct 3 13:47:03 2025 Fri Oct 3 13:47:03 2025
new : Fri Oct 3 13:47:00 2025 Fri Oct 3 13:47:03 2025
After patch:
$ mount.cifs //srv/share /mnt -o ...
$ python3 run.py
old : Fri Oct 3 13:48:39 2025 Fri Oct 3 13:48:39 2025
new : Fri Oct 3 13:48:41 2025 Fri Oct 3 13:48:41 2025
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Cc: Frank Sorenson <sorenson@redhat.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
Don't call ->set_file_info() on open handle to prevent the server from
stopping [cm]time updates automatically as per MS-FSA 2.1.4.17.
Fix this by checking for ATTR_OPEN bit earlier in cifs_setattr() to
prevent ->set_file_info() from being called when opening a file with
O_TRUNC. Do the truncation in ->open() instead.
This also saves two roundtrips when opening a file with O_TRUNC and
there are currently no open handles to be reused.
Before patch:
$ mount.cifs //srv/share /mnt -o ...
$ cd /mnt
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
old: 2025-10-03 13:26:23.151030500 -0300 2025-10-03 13:26:23.151030500 -0300
new: 2025-10-03 13:26:23.151030500 -0300 2025-10-03 13:26:23.151030500 -0300
After patch:
$ mount.cifs //srv/share /mnt -o ...
$ cd /mnt
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
$ exec 3>foo; stat -c 'old: %z %y' foo; sleep 2; echo test >&3; exec 3>&-; sleep 2; stat -c 'new: %z %y' foo
old: 2025-10-03 13:28:13.911933800 -0300 2025-10-03 13:28:13.911933800 -0300
new: 2025-10-03 13:28:26.647492700 -0300 2025-10-03 13:28:26.647492700 -0300
Reported-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
The return value of copy_to_iter() function will never be negative,
it is the number of bytes copied, or zero if nothing was copied.
Update the check to treat 0 as an error, and return -1 in that case.
Fixes: d08089f649 ("cifs: Change the I/O paths to use an iterator rather than a page list")
Acked-by: Tom Talpey <tom@talpey.com>
Reviewed-by: David Howells <dhowells@redhat.com>
Signed-off-by: Fushuai Wang <wangfushuai@baidu.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
smb2_copychunk_range() used to send a single SRV_COPYCHUNK per
SRV_COPYCHUNK_COPY IOCTL.
Implement variable Chunks[] array in struct copychunk_ioctl and fill it
with struct copychunk (MS-SMB2 2.2.31.1.1), bounded by server-advertised
limits.
This reduces the number of IOCTL requests for large copies.
While we are at it, rename a couple variables to follow the terminology
used in the specification.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Statements from an if branch and the end of this function implementation
were equivalent.
Thus delete duplicate source code.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
* Return directly after a call of the function “genlmsg_new” failed
at the beginning.
* Delete the label “fail” which became unnecessary
with this refactoring.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
Use two additional labels so that another bit of common code can be better
reused at the end of this function implementation.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
* Return a status code without storing it in an intermediate variable.
* Delete the local variable “ret” and the label “error”
which became unnecessary with this refactoring.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
Use three additional labels so that another bit of common code can be
better reused at the end of this function implementation.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Steve French <stfrench@microsoft.com>
* cifs_lookup():
Move the definition for the local variable “cfid” into an else branch
so that the corresponding setting will only be performed if a NULL inode
was detected during lookup by this function.
* cifs_d_revalidate():
Move the definition for the local variables “inode” and “rc” into
an if branch so that the corresponding setting will only be performed
after a d_really_is_positive() call.
Move the definition for the local variable “cfid” into an else branch
so that the corresponding setting will only be performed if further data
processing will be needed for an open_cached_dir_by_dentry() call.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Pull file->f_path constification from Al Viro:
"Only one thing was modifying ->f_path of an opened file - acct(2).
Massaging that away and constifying a bunch of struct path * arguments
in functions that might be given &file->f_path ends up with the
situation where we can turn ->f_path into an anon union of const
struct path f_path and struct path __f_path, the latter modified only
in a few places in fs/{file_table,open,namei}.c, all for struct file
instances that are yet to be opened"
* tag 'pull-f_path' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (23 commits)
Have cc(1) catch attempts to modify ->f_path
kernel/acct.c: saner struct file treatment
configfs:get_target() - release path as soon as we grab configfs_item reference
apparmor/af_unix: constify struct path * arguments
ovl_is_real_file: constify realpath argument
ovl_sync_file(): constify path argument
ovl_lower_dir(): constify path argument
ovl_get_verity_digest(): constify path argument
ovl_validate_verity(): constify {meta,data}path arguments
ovl_ensure_verity_loaded(): constify datapath argument
ksmbd_vfs_set_init_posix_acl(): constify path argument
ksmbd_vfs_inherit_posix_acl(): constify path argument
ksmbd_vfs_kern_path_unlock(): constify path argument
ksmbd_vfs_path_lookup_locked(): root_share_path can be const struct path *
check_export(): constify path argument
export_operations->open(): constify path argument
rqst_exp_get_by_name(): constify path argument
nfs: constify path argument of __vfs_getattr()
bpf...d_path(): constify path argument
done_path_create(): constify path argument
...
Pull smb client updates from Steve French:
- Fix oops in crypt message
- Remove duplicate arc4 code
- Fix potential io_uring reconnect
- Two important directory leases fixes and three perf improvements
- Three minor cleanups
- Four debug improvements (e.g. for showing more information on leases,
and one for adding more helpful information on reconnect)
* tag '6.18-rc-part1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
cifs: client: force multichannel=off when max_channels=1
smb client: fix bug with newly created file in cached dir
smb: client: short-circuit negative lookups when parent dir is fully cached
smb: client: short-circuit in open_cached_dir_by_dentry() if !dentry
smb: client: remove pointless cfid->has_lease check
smb: client: transport: minor indentation style fix
smb: client: transport: avoid reconnects triggered by pending task work
smb: client: remove unused fid_lock
smb: client: update cfid->last_access_time in open_cached_dir_by_dentry()
smb: client: ensure open_cached_dir_by_dentry() only returns valid cfid
smb: client: account smb directory cache usage and per-tcon totals
smb: client: add drop_dir_cache module parameter to invalidate cached dirents
smb: client: show lease state as R/H/W (or NONE) in open_files
smb: client: fix crypto buffers in non-linear memory
smb: Use arc4 library instead of duplicate arc4 code
smb: client: add tcon information to smb2_reconnect() debug messages
Pull smb server fixes from Steve French:
- Fix potential UAFs and corruptions in rpc open and close
- Fix copy_file_range when ranges overlap
- Improve session, share, connection lookup performance
- Fix potential hash collisions in share and session lists
- Debugging improvement - making per-connection threads easier to
identify
- Improve socket creation
- Fix return code mapping for posix query fs info
- Add support for limiting the maximum number of connections per IP
address, extending the existing connection limiting mechanism to
enforce per-IP connection limits alongside the global connection
limit
* tag 'v6.18rc1-part1-ksmbd-server-fixes' of git://git.samba.org/ksmbd:
ksmbd: increase session and share hash table bits
ksmbd: replace connection list with hash table
ksmbd: add an error print when maximum IP connections limit is reached
ksmbd: add max ip connections parameter
ksmbd: fix error code overwriting in smb2_get_info_filesystem()
ksmbd: copy overlapped range within the same file
ksmbd: use sock_create_kern interface to create kernel socket
ksmbd: make ksmbd thread names distinct by client IP
ksmbd: Fix race condition in RPC handle list access
Pull fs_context updates from Al Viro:
"Change vfs_parse_fs_string() calling conventions
Get rid of the length argument (almost all callers pass strlen() of
the string argument there), add vfs_parse_fs_qstr() for the cases that
do want separate length"
* tag 'pull-fs_context' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
do_nfs4_mount(): switch to vfs_parse_fs_string()
change the calling conventions for vfs_parse_fs_string()
Previously, specifying both multichannel and max_channels=1 as mount
options would leave multichannel enabled, even though it is not
meaningful when only one channel is allowed. This led to confusion and
inconsistent behavior, as the client would advertise multichannel
capability but never establish secondary channels.
Fix this by forcing multichannel to false whenever max_channels=1,
ensuring the mount configuration is consistent and matches user intent.
This prevents the client from advertising or attempting multichannel
support when it is not possible.
Signed-off-by: Rajasi Mandal <rajasimandal@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Test generic/637 spotted a problem with create of a new file in a
cached directory (by the same client) could cause cases where the
new file does not show up properly in ls on that client until the
lease times out.
Fixes: 037e1bae58 ("smb: client: use ParentLeaseKey in cifs_do_create")
Cc: stable@vger.kernel.org
Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
When the parent directory has a valid and complete cached enumeration we
can assume that negative dentries are not present in the directory, thus
we can return without issuing a request.
This reduces traffic for common ENOENT when the directory entries are
cached.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
When dentry is NULL, the current code acquires the spinlock and traverses
the entire list, but the condition (dentry && cfid->dentry == dentry)
ensures no match will ever be found.
Return -ENOENT early in this case, avoiding unnecessary lock acquisition
and list traversal.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
open_cached_dir() will only return a valid cfid, which has both
has_lease = true and time != 0.
Remove the pointless check of cfid->has_lease right after
open_cached_dir() returns no error.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
When io_uring is used in the same task as CIFS, there might be
unnecessary reconnects, causing issues in user-space applications
like QEMU with a log like:
> CIFS: VFS: \\10.10.100.81 Error -512 sending data on socket to server
Certain io_uring completions might be added to task_work with
notify_method being TWA_SIGNAL and thus TIF_NOTIFY_SIGNAL is set for
the task.
In __smb_send_rqst(), signals are masked before calling
smb_send_kvec(), but the masking does not apply to TIF_NOTIFY_SIGNAL.
If sk_stream_wait_memory() is reached via sock_sendmsg() while
TIF_NOTIFY_SIGNAL is set, signal_pending(current) will evaluate to
true there, and -EINTR will be propagated all the way from
sk_stream_wait_memory() to sock_sendmsg() in smb_send_kvec().
Afterwards, __smb_send_rqst() will see that not everything was written
and reconnect.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
The fid_lock in struct cached_fid does not currently provide any real
synchronization. Previously, it had the intention to prevent a double
release of the dentry, but every change to cfid->dentry is already
protected either by cfid_list_lock (while the entry is in the list) or
happens after the cfid has been removed (so no other thread should find
it).
Since there is no scenario in which fid_lock prevents any race, it is
vestigial and can be removed along with its associated
spin_lock()/spin_unlock() calls.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
open_cached_dir_by_dentry() was missing an update of
cfid->last_access_time to jiffies, similar to what open_cached_dir()
has.
Add it to the function.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
open_cached_dir_by_dentry() was exposing an invalid cached directory to
callers. The validity check outside the function was exclusively based
on cfid->time.
Add validity check before returning success and introduce
is_valid_cached_dir() helper for consistent checks across the code.
Signed-off-by: Henrique Carvalho <henrique.carvalho@suse.com>
Reviwed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
Add lightweight accounting for directory lease cache usage
to aid debugging and limiting cache size in future. Track
per-directory entry/byte counts and maintain per-tcon
aggregates. Also expose the totals in /proc/fs/cifs/open_dirs.
Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Add write-only /sys/module/cifs/parameters/drop_dir_cache. Writing a
non-zero value iterates all tcons and calls invalidate_all_cached_dirs()
to drop cached directory entries. This is useful to force a dirent cache
drop across mounts for debugging and testing purpose.
Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Print the lease/oplock caching state for each open file as a
compact string of letters: R (read), H (handle), W (write).
Signed-off-by: Bharath SM <bharathsm@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Increases the number of bits for the hash table from 3 to 12.
The thousands of sessions and shares can be connected.
So the current 3-bit size can lead to frequent hash collisions.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Replace connection list with hash table to improve lookup performance.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
This change introduces an error print using pr_info_ratelimited()
to prevent excessive logging. This message will inform the user that
the limit for maximum IP connections has been hit and what that
current count is, which can be useful for debugging and monitoring
connection limits.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
This parameter set the maximum number of connections per ip address.
The default is 8.
Cc: stable@vger.kernel.org
Fixes: c0d41112f1 ("ksmbd: extend the connection limiting mechanism to support IPv6")
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
If client doesn't negotiate with SMB3.1.1 POSIX Extensions,
then proper error code won't be returned due to overwriting.
Return error immediately.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: e2f34481b2 ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Signed-off-by: Matvey Kovalev <matvey.kovalev@ispras.ru>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
cifs.ko request to copy overlapped range within the same file.
ksmbd is using vfs_copy_file_range for this, vfs_copy_file_range() does not
allow overlapped copying within the same file.
This patch use do_splice_direct() if offset and length are overlapped.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
we should use sock_create_kern() if the socket resides in kernel space.
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
This patch makes ksmbd thread names distinct by client IP address.
100943 ? S 0:00 [ksmbd:::ffff:10.177.110.57]
or
101752 ? S 0:00 [ksmbd:10.177.110.57]
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
The 'sess->rpc_handle_list' XArray manages RPC handles within a ksmbd
session. Access to this list is intended to be protected by
'sess->rpc_lock' (an rw_semaphore). However, the locking implementation was
flawed, leading to potential race conditions.
In ksmbd_session_rpc_open(), the code incorrectly acquired only a read lock
before calling xa_store() and xa_erase(). Since these operations modify
the XArray structure, a write lock is required to ensure exclusive access
and prevent data corruption from concurrent modifications.
Furthermore, ksmbd_session_rpc_method() accessed the list using xa_load()
without holding any lock at all. This could lead to reading inconsistent
data or a potential use-after-free if an entry is concurrently removed and
the pointer is dereferenced.
Fix these issues by:
1. Using down_write() and up_write() in ksmbd_session_rpc_open()
to ensure exclusive access during XArray modification, and ensuring
the lock is correctly released on error paths.
2. Adding down_read() and up_read() in ksmbd_session_rpc_method()
to safely protect the lookup.
Fixes: a1f46c99d9 ("ksmbd: fix use-after-free in ksmbd_session_rpc_open")
Fixes: b685757c7b ("ksmbd: Implements sess->rpc_handle_list as xarray")
Cc: stable@vger.kernel.org
Signed-off-by: Yunseong Kim <ysk@kzalloc.com>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
The crypto API, through the scatterlist API, expects input buffers to be
in linear memory. We handle this with the cifs_sg_set_buf() helper
that converts vmalloc'd memory to their corresponding pages.
However, when we allocate our aead_request buffer (@creq in
smb2ops.c::crypt_message()), we do so with kvzalloc(), which possibly
puts aead_request->__ctx in vmalloc area.
AEAD algorithm then uses ->__ctx for its private/internal data and
operations, and uses sg_set_buf() for such data on a few places.
This works fine as long as @creq falls into kmalloc zone (small
requests) or vmalloc'd memory is still within linear range.
Tasks' stacks are vmalloc'd by default (CONFIG_VMAP_STACK=y), so too
many tasks will increment the base stacks' addresses to a point where
virt_addr_valid(buf) will fail (BUG() in sg_set_buf()) when that
happens.
In practice: too many parallel reads and writes on an encrypted mount
will trigger this bug.
To fix this, always alloc @creq with kmalloc() instead.
Also drop the @sensitive_size variable/arguments since
kfree_sensitive() doesn't need it.
Backtrace:
[ 945.272081] ------------[ cut here ]------------
[ 945.272774] kernel BUG at include/linux/scatterlist.h:209!
[ 945.273520] Oops: invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC NOPTI
[ 945.274412] CPU: 7 UID: 0 PID: 56 Comm: kworker/u33:0 Kdump: loaded Not tainted 6.15.0-lku-11779-g8e9d6efccdd7-dirty #1 PREEMPT(voluntary)
[ 945.275736] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-2-gc13ff2cd-prebuilt.qemu.org 04/01/2014
[ 945.276877] Workqueue: writeback wb_workfn (flush-cifs-2)
[ 945.277457] RIP: 0010:crypto_gcm_init_common+0x1f9/0x220
[ 945.278018] Code: b0 00 00 00 48 83 c4 08 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc cc cc cc 48 c7 c0 00 00 00 80 48 2b 05 5c 58 e5 00 e9 58 ff ff ff <0f> 0b 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b 48 c7 04 24 01 00 00 00 48 8b
[ 945.279992] RSP: 0018:ffffc90000a27360 EFLAGS: 00010246
[ 945.280578] RAX: 0000000000000000 RBX: ffffc90001d85060 RCX: 0000000000000030
[ 945.281376] RDX: 0000000000080000 RSI: 0000000000000000 RDI: ffffc90081d85070
[ 945.282145] RBP: ffffc90001d85010 R08: ffffc90001d85000 R09: 0000000000000000
[ 945.282898] R10: ffffc90001d85090 R11: 0000000000001000 R12: ffffc90001d85070
[ 945.283656] R13: ffff888113522948 R14: ffffc90001d85060 R15: ffffc90001d85010
[ 945.284407] FS: 0000000000000000(0000) GS:ffff8882e66cf000(0000) knlGS:0000000000000000
[ 945.285262] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 945.285884] CR2: 00007fa7ffdd31f4 CR3: 000000010540d000 CR4: 0000000000350ef0
[ 945.286683] Call Trace:
[ 945.286952] <TASK>
[ 945.287184] ? crypt_message+0x33f/0xad0 [cifs]
[ 945.287719] crypto_gcm_encrypt+0x36/0xe0
[ 945.288152] crypt_message+0x54a/0xad0 [cifs]
[ 945.288724] smb3_init_transform_rq+0x277/0x300 [cifs]
[ 945.289300] smb_send_rqst+0xa3/0x160 [cifs]
[ 945.289944] cifs_call_async+0x178/0x340 [cifs]
[ 945.290514] ? __pfx_smb2_writev_callback+0x10/0x10 [cifs]
[ 945.291177] smb2_async_writev+0x3e3/0x670 [cifs]
[ 945.291759] ? find_held_lock+0x32/0x90
[ 945.292212] ? netfs_advance_write+0xf2/0x310
[ 945.292723] netfs_advance_write+0xf2/0x310
[ 945.293210] netfs_write_folio+0x346/0xcc0
[ 945.293689] ? __pfx__raw_spin_unlock_irq+0x10/0x10
[ 945.294250] netfs_writepages+0x117/0x460
[ 945.294724] do_writepages+0xbe/0x170
[ 945.295152] ? find_held_lock+0x32/0x90
[ 945.295600] ? kvm_sched_clock_read+0x11/0x20
[ 945.296103] __writeback_single_inode+0x56/0x4b0
[ 945.296643] writeback_sb_inodes+0x229/0x550
[ 945.297140] __writeback_inodes_wb+0x4c/0xe0
[ 945.297642] wb_writeback+0x2f1/0x3f0
[ 945.298069] wb_workfn+0x300/0x490
[ 945.298472] process_one_work+0x1fe/0x590
[ 945.298949] worker_thread+0x1ce/0x3c0
[ 945.299397] ? __pfx_worker_thread+0x10/0x10
[ 945.299900] kthread+0x119/0x210
[ 945.300285] ? __pfx_kthread+0x10/0x10
[ 945.300729] ret_from_fork+0x119/0x1b0
[ 945.301163] ? __pfx_kthread+0x10/0x10
[ 945.301601] ret_from_fork_asm+0x1a/0x30
[ 945.302055] </TASK>
Fixes: d08089f649 ("cifs: Change the I/O paths to use an iterator rather than a page list")
Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
fs/smb/common/cifs_arc4.c has an implementation of ARC4, but a copy of
this same code is also present in lib/crypto/arc4.c to serve the other
users of this legacy algorithm in the kernel. Remove the duplicate
implementation in fs/smb/, which seems to have been added because of a
misunderstanding, and just use the lib/crypto/ one.
Signed-off-by: Eric Biggers <ebiggers@kernel.org>
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Linus Torvalds