NeilBrown [Fri, 9 May 2025 00:46:43 +0000 (10:46 +1000)]
nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer
Instead of calling xchg() and unrcu_pointer() before
nfsd_file_put_local(), we now pass pointer to the __rcu pointer and call
xchg() and unrcu_pointer() inside that function.
Where unrcu_pointer() is currently called the internals of "struct
nfsd_file" are not known and that causes older compilers such as gcc-8
to complain.
In some cases we have a __kernel (aka normal) pointer not an __rcu
pointer so we need to cast it to __rcu first. This is strictly a
weakening so no information is lost. Somewhat surprisingly, this cast
is accepted by gcc-8.
This has the pleasing result that the cmpxchg() which sets ro_file and
rw_file, and also the xchg() which clears them, are both now in the nfsd
code.
Reported-by: Pali Rohár <pali@kernel.org> Reported-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr> Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Fri, 9 May 2025 00:46:42 +0000 (10:46 +1000)]
nfs_localio: protect race between nfs_uuid_put() and nfs_close_local_fh()
nfs_uuid_put() and nfs_close_local_fh() can race if a "struct
nfs_file_localio" is released at the same time that nfsd calls
nfs_localio_invalidate_clients().
It is important that neither of these functions completes after the
other has started looking at a given nfs_file_localio and before it
finishes.
If nfs_uuid_put() exits while nfs_close_local_fh() is closing ro_file
and rw_file it could return to __nfd_file_cache_purge() while some files
are still referenced so the purge may not succeed.
If nfs_close_local_fh() exits while nfsd_uuid_put() is still closing the
files then the "struct nfs_file_localio" could be freed while
nfsd_uuid_put() is still looking at it. This side is currently handled
by copying the pointers out of ro_file and rw_file before deleting from
the list in nfsd_uuid. We need to preserve this while ensuring that
nfsd_uuid_put() does wait for nfs_close_local_fh().
This patch use nfl->uuid and nfl->list to provide the required
interlock.
nfs_uuid_put() removes the nfs_file_localio from the list, then drops
locks and puts the two files, then reclaims the spinlock and sets
->nfs_uuid to NULL.
nfs_close_local_fh() operates in the reverse order, setting ->nfs_uuid
to NULL, then closing the files, then unlinking from the list.
If nfs_uuid_put() finds that ->nfs_uuid is already NULL, it waits for
the nfs_file_localio to be removed from the list. If
nfs_close_local_fh() find that it has already been unlinked it waits for
->nfs_uuid to become NULL. This ensure that one of the two tries to
close the files, but they each waits for the other.
As nfs_uuid_put() is making the list empty, change from a
list_for_each_safe loop to a while that always takes the first entry.
This makes the intent more clear.
Also don't move the list to a temporary local list as this would defeat
the guarantees required for the interlock.
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Fri, 9 May 2025 00:46:41 +0000 (10:46 +1000)]
nfs_localio: duplicate nfs_close_local_fh()
nfs_close_local_fh() is called from two different places for quite
different use case.
It is called from nfs_uuid_put() when the nfs_uuid is being detached -
possibly because the nfs server is not longer serving that filesystem.
In this case there will always be an nfs_uuid and so rcu_read_lock() is
not needed.
It is also called when the nfs_file_localio is no longer needed. In
this case there may not be an active nfs_uuid.
These two can race, and handling the race properly while avoiding
excessive locking will require different handling on each side.
This patch prepares the way by opencoding nfs_close_local_fh() into
nfs_uuid_put(), then simplifying the code there as befits the context.
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Fri, 9 May 2025 00:46:40 +0000 (10:46 +1000)]
nfs_localio: simplify interface to nfsd for getting nfsd_file
The nfsd_localio_operations structure contains nfsd_file_get() to get a
reference to an nfsd_file. This is only used in one place, where
nfsd_open_local_fh() is also used.
This patch combines the two, calling nfsd_open_local_fh() passing a
pointer to where the nfsd_file pointer might be stored. If there is a
pointer there an nfsd_file_get() can get a reference, that reference is
returned. If not a new nfsd_file is acquired, stored at the pointer,
and returned. When we store a reference we also increase the refcount
on the net, as that refcount is decrements when we clear the stored
pointer.
We now get an extra reference *before* storing the new nfsd_file at the
given location. This avoids possible races with the nfsd_file being
freed before the final reference can be taken.
This patch moves the rcu_dereference() needed after fetching from
ro_file or rw_file into the nfsd code where the 'struct nfs_file' is
fully defined. This avoids an error reported by older versions of gcc
such as gcc-8 which complain about rcu_dereference() use in contexts
where the structure (which will supposedly be accessed) is not fully
defined.
Reported-by: Pali Rohár <pali@kernel.org> Reported-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr> Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Fri, 9 May 2025 00:46:39 +0000 (10:46 +1000)]
nfs_localio: always hold nfsd net ref with nfsd_file ref
Having separate nfsd_file_put and nfsd_file_put_local in struct
nfsd_localio_operations doesn't make much sense. The difference is that
nfsd_file_put doesn't drop a reference to the nfs_net which is what
keeps nfsd from shutting down.
Currently, if nfsd tries to shutdown it will invalidate the files stored
in the list from the nfs_uuid and this will drop all references to the
nfsd net that the client holds. But the client could still hold some
references to nfsd_files for active IO. So nfsd might think is has
completely shut down local IO, but hasn't and has no way to wait for
those active IO requests to complete.
So this patch changes nfsd_file_get to nfsd_file_get_local and has it
increase the ref count on the nfsd net and it replaces all calls to
->nfsd_put_file to ->nfsd_put_file_local.
It also changes ->nfsd_open_local_fh to return with the refcount on the
net elevated precisely when a valid nfsd_file is returned.
This means that whenever the client holds a valid nfsd_file, there will
be an associated count on the nfsd net, and so the count can only reach
zero when all nfsd_files have been returned.
nfs_local_file_put() is changed to call nfs_to_nfsd_file_put_local()
instead of replacing calls to one with calls to the other because this
will help a later patch which changes nfs_to_nfsd_file_put_local() to
take an __rcu pointer while nfs_local_file_put() doesn't.
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Fri, 9 May 2025 00:46:38 +0000 (10:46 +1000)]
nfs_localio: use cmpxchg() to install new nfs_file_localio
Rather than using nfs_uuid.lock to protect installing
a new ro_file or rw_file, change to use cmpxchg().
Removing the file already uses xchg() so this improves symmetry
and also makes the code a little simpler.
Also remove the optimisation of not taking the lock, and not removing
the nfs_file_localio from the linked list, when both ->ro_file and
->rw_file are already NULL. Given that ->nfs_uuid was not NULL, it is
extremely unlikely that neither ->ro_file or ->rw_file is NULL so
this optimisation can be of little value and it complicates
understanding of the code - why can the list_del_init() be skipped?
Finally, move the assignment of NULL to ->nfs_uuid until after
the last action on the nfs_file_localio (the list_del_init). As soon as
this is NULL a racing nfs_close_local_fh() can bypass all the locking
and go on to free the nfs_file_localio, so we must be certain to be
finished with it first.
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Chuck Lever [Wed, 21 May 2025 20:34:13 +0000 (16:34 -0400)]
SUNRPC: Prevent hang on NFS mount with xprtsec=[m]tls
Engineers at Hammerspace noticed that sometimes mounting with
"xprtsec=tls" hangs for a minute or so, and then times out, even
when the NFS server is reachable and responsive.
kTLS shuts off data_ready callbacks if strp->msg_ready is set to
mitigate data_ready callbacks when a full TLS record is not yet
ready to be read from the socket.
Normally msg_ready is clear when the first TLS record arrives on
a socket. However, I observed that sometimes tls_setsockopt() sets
strp->msg_ready, and that prevents forward progress because
tls_data_ready() becomes a no-op.
Moreover, Jakub says: "If there's a full record queued at the time
when [tlshd] passes the socket back to the kernel, it's up to the
reader to read the already queued data out." So SunRPC cannot
expect a data_ready call when ingress data is already waiting.
Add an explicit poll after SunRPC's upper transport is set up to
pick up any data that arrived after the TLS handshake but before
transport set-up is complete.
Reported-by: Steve Sears <sjs@hammerspace.com> Suggested-by: Jakub Kacinski <kuba@kernel.org> Fixes: 75eb6af7acdf ("SUNRPC: Add a TCP-with-TLS RPC transport class") Tested-by: Mike Snitzer <snitzer@kernel.org> Reviewed-by: Mike Snitzer <snitzer@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NeilBrown [Mon, 26 May 2025 04:08:25 +0000 (14:08 +1000)]
nfs: fix incorrect handling of large-number NFS errors in nfs4_do_mkdir()
A recent commit introduced nfs4_do_mkdir() which reports an error from
nfs4_call_sync() by returning it with ERR_PTR().
This is a problem as nfs4_call_sync() can return negative NFS-specific
errors with values larger than MAX_ERRNO (4095). One example is
NFS4ERR_DELAY which has value 10008.
This "pointer" gets to PTR_ERR_OR_ZERO() in nfs4_proc_mkdir() which
chooses ZERO because it isn't in the range of value errors. Ultimately
the pointer is dereferenced.
This patch changes nfs4_do_mkdir() to report the dentry pointer and
status separately - pointer as a return value, status in an "int *"
parameter.
The same separation is used for _nfs4_proc_mkdir() and the two are
combined only in nfs4_proc_mkdir() after the status has passed through
nfs4_handle_exception(), which ensures the error code does not exceed
MAX_ERRNO.
It also fixes a problem in the even when nfs4_handle_exception() updated
the error value, the original 'alias' was still returned.
Reported-by: Anna Schumaker <anna@kernel.org> Fixes: 8376583b84a1 ("nfs: change mkdir inode_operation to return alternate dentry if needed.") Signed-off-by: NeilBrown <neil@brown.name> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Li Lingfeng [Tue, 4 Mar 2025 13:05:33 +0000 (21:05 +0800)]
nfs: ignore SB_RDONLY when remounting nfs
In some scenarios, when mounting NFS, more than one superblock may be
created. The final superblock used is the last one created, but only the
first superblock carries the ro flag passed from user space. If a ro flag
is added to the superblock via remount, it will trigger the issue
described in Link[1].
Link[2] attempted to address this by marking the superblock as ro during
the initial mount. However, this introduced a new problem in scenarios
where multiple mount points share the same superblock:
[root@a ~]# mount /dev/sdb /mnt/sdb
[root@a ~]# echo "/mnt/sdb *(rw,no_root_squash)" > /etc/exports
[root@a ~]# echo "/mnt/sdb/test_dir2 *(ro,no_root_squash)" >> /etc/exports
[root@a ~]# systemctl restart nfs-server
[root@a ~]# mount -t nfs -o rw 127.0.0.1:/mnt/sdb/test_dir1 /mnt/test_mp1
[root@a ~]# mount | grep nfs4
127.0.0.1:/mnt/sdb/test_dir1 on /mnt/test_mp1 type nfs4 (rw,relatime,...
[root@a ~]# mount -t nfs -o ro 127.0.0.1:/mnt/sdb/test_dir2 /mnt/test_mp2
[root@a ~]# mount | grep nfs4
127.0.0.1:/mnt/sdb/test_dir1 on /mnt/test_mp1 type nfs4 (ro,relatime,...
127.0.0.1:/mnt/sdb/test_dir2 on /mnt/test_mp2 type nfs4 (ro,relatime,...
[root@a ~]#
When mounting the second NFS, the shared superblock is marked as ro,
causing the previous NFS mount to become read-only.
To resolve both issues, the ro flag is no longer applied to the superblock
during remount. Instead, the ro flag on the mount is used to control
whether the mount point is read-only.
Fixes: 281cad46b34d ("NFS: Create a submount rpc_op")
Link[1]: https://lore.kernel.org/all/20240604112636.236517-3-lilingfeng@huaweicloud.com/
Link[2]: https://lore.kernel.org/all/20241130035818.1459775-1-lilingfeng3@huawei.com/ Signed-off-by: Li Lingfeng <lilingfeng3@huawei.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Li Lingfeng [Tue, 4 Mar 2025 13:05:32 +0000 (21:05 +0800)]
nfs: clear SB_RDONLY before getting superblock
As described in the link, commit 52cb7f8f1778 ("nfs: ignore SB_RDONLY when
mounting nfs") removed the check for the ro flag when determining whether
to share the superblock, which caused issues when mounting different
subdirectories under the same export directory via NFSv3. However, this
change did not affect NFSv4.
For NFSv3:
1) A single superblock is created for the initial mount.
2) When mounted read-only, this superblock carries the SB_RDONLY flag.
3) Before commit 52cb7f8f1778 ("nfs: ignore SB_RDONLY when mounting nfs"):
Subsequent rw mounts would not share the existing ro superblock due to
flag mismatch, creating a new superblock without SB_RDONLY.
After the commit:
The SB_RDONLY flag is ignored during superblock comparison, and this leads
to sharing the existing superblock even for rw mounts.
Ultimately results in write operations being rejected at the VFS layer.
For NFSv4:
1) Multiple superblocks are created and the last one will be kept.
2) The actually used superblock for ro mounts doesn't carry SB_RDONLY flag.
Therefore, commit 52cb7f8f1778 doesn't affect NFSv4 mounts.
Clear SB_RDONLY before getting superblock when NFS_MOUNT_UNSHARED is not
set to fix it.
Fixes: 52cb7f8f1778 ("nfs: ignore SB_RDONLY when mounting nfs") Closes: https://lore.kernel.org/all/12d7ea53-1202-4e21-a7ef-431c94758ce5@app.fastmail.com/T/ Signed-off-by: Li Lingfeng <lilingfeng3@huawei.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Mike Snitzer [Tue, 13 May 2025 16:08:31 +0000 (12:08 -0400)]
NFS: always probe for LOCALIO support asynchronously
It was reported that NFS client mounts of AWS Elastic File System
(EFS) volumes is slow, this is because the AWS firewall disallows
LOCALIO (because it doesn't consider the use of NFS_LOCALIO_PROGRAM
valid), see: https://bugzilla.redhat.com/show_bug.cgi?id=2335129
Switch to performing the LOCALIO probe asynchronously to address the
potential for the NFS LOCALIO protocol being disallowed and/or slowed
by the remote server's response.
While at it, fix nfs_local_probe_async() to always take/put a
reference on the nfs_client that is using the LOCALIO protocol.
Also, unexport the nfs_local_probe() symbol and make it private to
fs/nfs/localio.c
This change has the side-effect of initially issuing reads, writes and
commits over the wire via SUNRPC until the LOCALIO probe completes.
Suggested-by: Jeff Layton <jlayton@kernel.org> # to always probe async Fixes: 76d4cb6345da ("nfs: probe for LOCALIO when v4 client reconnects to server") Cc: stable@vger.kernel.org # 6.14+ Signed-off-by: Mike Snitzer <snitzer@kernel.org> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Mike Snitzer [Tue, 13 May 2025 19:50:44 +0000 (15:50 -0400)]
NFS: add localio to sysfs
The Linux NFS client and server added support for LOCALIO in Linux
v6.12. It is useful to know if a client and server negotiated LOCALIO
be used, so expose it through the 'localio' attribute.
Suggested-by: Anna Schumaker <anna.schumaker@oracle.com> Signed-off-by: Mike Snitzer <snitzer@kernel.org> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
nfs: don't return AOP_WRITEPAGE_ACTIVATE from nfs_do_writepage
nfs_do_writepage is a successful return that requires the caller to
unlock the folio. Using it here requires special casing both in
nfs_do_writepage and nfs_writepages_callback and leaves a land mine in
nfs_wb_folio in case it ever set the flag. Remove it and just
unconditionally unlock in nfs_writepages_callback.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Han Young [Sun, 4 May 2025 12:57:04 +0000 (20:57 +0800)]
NFSv4: Always set NLINK even if the server doesn't support it
fattr4_numlinks is a recommended attribute, so the client should emulate
it even if the server doesn't support it. In decode_attr_nlink function
in nfs4xdr.c, nlink is initialized to 1. However, this default value
isn't set to the inode due to the check in nfs_fhget.
So if the server doesn't support numlinks, inode's nlink will be zero,
the mount will fail with error "Stale file handle". Set the nlink to 1
if the server doesn't support it.
Signed-off-by: Han Young <hanyang.tony@bytedance.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
The NFS client's list of delegations can grow quite large (well beyond the
delegation watermark) if the server is revoking or there are repeated
events that expire state. Once this happens, the revoked delegations can
cause a performance problem for subsequent walks of the
servers->delegations list when the client tries to test and free state.
If we can determine that the FREE_STATEID operation has completed without
error, we can prune the delegation from the list.
Since the NFS client combines TEST_STATEID with FREE_STATEID in its minor
version operations, there isn't an easy way to communicate success of
FREE_STATEID. Rather than re-arrange quite a number of calling paths to
break out the separate procedures, let's signal the success of FREE_STATEID
by setting the stateid's type.
Set NFS4_FREED_STATEID_TYPE for stateids that have been successfully
discarded from the server, and use that type to signal that the delegation
can be cleaned up.
Signed-off-by: Benjamin Coddington <bcodding@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Scott Mayhew [Wed, 30 Apr 2025 11:12:29 +0000 (07:12 -0400)]
NFSv4: Don't check for OPEN feature support in v4.1
fattr4_open_arguments is a v4.2 recommended attribute, so we shouldn't
be sending it to v4.1 servers.
Fixes: cb78f9b7d0c0 ("nfs: fix the fetch of FATTR4_OPEN_ARGUMENTS") Signed-off-by: Scott Mayhew <smayhew@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Benjamin Coddington <bcodding@redhat.com> Cc: stable@vger.kernel.org # 6.11+ Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NFSv4.2: fix listxattr to return selinux security label
Currently, when NFS is queried for all the labels present on the
file via a command example "getfattr -d -m . /mnt/testfile", it
does not return the security label. Yet when asked specifically for
the label (getfattr -n security.selinux) it will be returned.
Include the security label when all attributes are queried.
Signed-off-by: Olga Kornievskaia <okorniev@redhat.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
NFSv4.2: fix setattr caching of TIME_[MODIFY|ACCESS]_SET when timestamps are delegated
nfs_setattr will flush all pending writes before updating a file time
attributes. However when the client holds delegated timestamps, it can
update its timestamps locally as it is the authority for the file
times attributes. The client will later set the file attributes by
adding a setattr to the delegreturn compound updating the server time
attributes.
Fix nfs_setattr to avoid flushing pending writes when the file time
attributes are delegated and the mtime/atime are set to a fixed
timestamp (ATTR_[MODIFY|ACCESS]_SET. Also, when sending the setattr
procedure over the wire, we need to clear the correct attribute bits
from the bitmask.
I was able to measure a noticable speedup when measuring untar performance.
Test: $ time tar xzf ~/dir.tgz
Baseline: 1m13.072s
Patched: 0m49.038s
Which is more than 30% latency improvement.
Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Anna Schumaker [Fri, 11 Apr 2025 19:47:32 +0000 (15:47 -0400)]
NFS: Add support for fallocate(FALLOC_FL_ZERO_RANGE)
This implements a suggestion from Trond that we can mimic
FALLOC_FL_ZERO_RANGE by sending a compound that first does a DEALLOCATE
to punch a hole in a file, and then an ALLOCATE to fill the hole with
zeroes. There might technically be a race here, but once the DEALLOCATE
finishes any reads from the region would return zeroes anyway, so I
don't expect it to cause problems.
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Max Kellermann [Wed, 23 Apr 2025 13:22:50 +0000 (15:22 +0200)]
fs/nfs/read: fix double-unlock bug in nfs_return_empty_folio()
Sometimes, when a file was read while it was being truncated by
another NFS client, the kernel could deadlock because folio_unlock()
was called twice, and the second call would XOR back the `PG_locked`
flag.
Most of the time (depending on the timing of the truncation), nobody
notices the problem because folio_unlock() gets called three times,
which flips `PG_locked` back off:
The problem is that nfs_read_add_folio() is not supposed to unlock the
folio if fscache is enabled, and a nfs_netfs_folio_unlock() check is
missing in nfs_return_empty_folio().
Rarely this leads to a warning in netfs_read_collection():
------------[ cut here ]------------
R=0000031c: folio 10 is not locked
WARNING: CPU: 0 PID: 29 at fs/netfs/read_collect.c:133 netfs_read_collection+0x7c0/0xf00
[...]
Workqueue: events_unbound netfs_read_collection_worker
RIP: 0010:netfs_read_collection+0x7c0/0xf00
[...]
Call Trace:
<TASK>
netfs_read_collection_worker+0x67/0x80
process_one_work+0x12e/0x2c0
worker_thread+0x295/0x3a0
Most of the time, however, processes just get stuck forever in
folio_wait_bit_common(), waiting for `PG_locked` to disappear, which
never happens because nobody is really holding the folio lock.
Fixes: 000dbe0bec05 ("NFS: Convert buffered read paths to use netfs when fscache is enabled") Cc: stable@vger.kernel.org Signed-off-by: Max Kellermann <max.kellermann@ionos.com> Reviewed-by: Dave Wysochanski <dwysocha@redhat.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Scott Mayhew [Wed, 16 Apr 2025 15:23:38 +0000 (11:23 -0400)]
NFSv4: xattr handlers should check for absent nfs filehandles
The nfs inodes for referral anchors that have not yet been followed have
their filehandles zeroed out.
Attempting to call getxattr() on one of these will cause the nfs client
to send a GETATTR to the nfs server with the preceding PUTFH sans
filehandle. The server will reply NFS4ERR_NOFILEHANDLE, leading to -EIO
being returned to the application.
Nikhil Jha [Wed, 19 Mar 2025 17:02:40 +0000 (13:02 -0400)]
sunrpc: don't immediately retransmit on seqno miss
RFC2203 requires that retransmitted messages use a new gss sequence
number, but the same XID. This means that if the server is just slow
(e.x. overloaded), the client might receive a response using an older
seqno than the one it has recorded.
Currently, Linux's client immediately retransmits in this case. However,
this leads to a lot of wasted retransmits until the server eventually
responds faster than the client can resend.
Client -> SEQ 1 -> Server
Client -> SEQ 2 -> Server
Client <- SEQ 1 <- Server (misses, expecting seqno = 2)
Client -> SEQ 3 -> Server (immediate retransmission on miss)
Client <- SEQ 2 <- Server (misses, expecting seqno = 3)
Client -> SEQ 4 -> Server (immediate retransmission on miss)
... and so on ...
This commit makes it so that we ignore messages with bad checksums
due to seqnum mismatch, and rely on the usual timeout behavior for
retransmission instead of doing so immediately.
Signed-off-by: Nikhil Jha <njha@janestreet.com> Acked-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Nikhil Jha [Wed, 19 Mar 2025 17:02:39 +0000 (13:02 -0400)]
sunrpc: implement rfc2203 rpcsec_gss seqnum cache
This implements a sequence number cache of the last three (right now
hardcoded) sent sequence numbers for a given XID, as suggested by the
RFC.
From RFC2203 5.3.3.1:
"Note that the sequence number algorithm requires that the client
increment the sequence number even if it is retrying a request with
the same RPC transaction identifier. It is not infrequent for
clients to get into a situation where they send two or more attempts
and a slow server sends the reply for the first attempt. With
RPCSEC_GSS, each request and reply will have a unique sequence
number. If the client wishes to improve turn around time on the RPC
call, it can cache the RPCSEC_GSS sequence number of each request it
sends. Then when it receives a response with a matching RPC
transaction identifier, it can compute the checksum of each sequence
number in the cache to try to match the checksum in the reply's
verifier."
Signed-off-by: Nikhil Jha <njha@janestreet.com> Acked-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
Linus Torvalds [Sat, 17 May 2025 19:53:39 +0000 (12:53 -0700)]
Merge tag 'dmaengine-fix-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine
Pull dmaengine fixes from Vinod Koul:
"This has a bunch of idxd driver fixes, dmatest revert and bunch of
smaller driver fixes:
- a bunch of idxd potential mem leak fixes
- dmatest revert for waiting for interrupt fix as that causes issue
- a couple of ti k3 udma fixes for locking and cap_mask
- mediatek deadlock fix and unused variable cleanup fix"
* tag 'dmaengine-fix-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine:
dmaengine: mediatek: drop unused variable
dmaengine: fsl-edma: Fix return code for unhandled interrupts
dmaengine: mediatek: Fix a possible deadlock error in mtk_cqdma_tx_status()
dmaengine: idxd: Fix ->poll() return value
dmaengine: idxd: Refactor remove call with idxd_cleanup() helper
dmaengine: idxd: Add missing idxd cleanup to fix memory leak in remove call
dmaengine: idxd: fix memory leak in error handling path of idxd_pci_probe
dmaengine: idxd: fix memory leak in error handling path of idxd_alloc
dmaengine: idxd: Add missing cleanups in cleanup internals
dmaengine: idxd: Add missing cleanup for early error out in idxd_setup_internals
dmaengine: idxd: fix memory leak in error handling path of idxd_setup_groups
dmaengine: idxd: fix memory leak in error handling path of idxd_setup_engines
dmaengine: idxd: fix memory leak in error handling path of idxd_setup_wqs
dmaengine: ptdma: Move variable condition check to the first place and remove redundancy
dmaengine: idxd: Fix allowing write() from different address spaces
dmaengine: ti: k3-udma: Add missing locking
dmaengine: ti: k3-udma: Use cap_mask directly from dma_device structure instead of a local copy
dmaengine: Revert "dmaengine: dmatest: Fix dmatest waiting less when interrupted"
dmaengine: idxd: cdev: Fix uninitialized use of sva in idxd_cdev_open
Linus Torvalds [Sat, 17 May 2025 19:49:47 +0000 (12:49 -0700)]
Merge tag 'phy-fixes-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy
Pull phy fixes from Vinod Koul:
"A bunch of renesas fixes and few smaller fixes in other drivers:
- Rensas fixes for unbind ole detection, irq, locking etc
- tegra fixes for error handling at init and UTMI power states and
stray unlock fix
- rockchip missing assignment and pll output fixes
- startfive usb host detection fixes"
* tag 'phy-fixes-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy:
phy: Fix error handling in tegra_xusb_port_init
phy: renesas: rcar-gen3-usb2: Set timing registers only once
phy: renesas: rcar-gen3-usb2: Assert PLL reset on PHY power off
phy: renesas: rcar-gen3-usb2: Lock around hardware registers and driver data
phy: renesas: rcar-gen3-usb2: Move IRQ request in probe
phy: renesas: rcar-gen3-usb2: Fix role detection on unbind/bind
phy: tegra: xusb: remove a stray unlock
phy: phy-rockchip-samsung-hdptx: Fix PHY PLL output 50.25MHz error
phy: starfive: jh7110-usb: Fix USB 2.0 host occasional detection failure
phy: rockchip-samsung-dcphy: Add missing assignment
phy: can-transceiver: Re-instate "mux-states" property presence check
phy: qcom-qmp-ufs: check for mode type for phy setting
phy: tegra: xusb: Use a bitmask for UTMI pad power state tracking
Linus Torvalds [Sat, 17 May 2025 19:04:19 +0000 (12:04 -0700)]
Merge tag 'soundwire-6.15-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire
Pull soundwire fix from Vinod Koul:
- Fix for irq domain creation race in the core
* tag 'soundwire-6.15-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire:
soundwire: bus: Fix race on the creation of the IRQ domain
Linus Torvalds [Sat, 17 May 2025 17:56:52 +0000 (10:56 -0700)]
Merge tag 'mm-hotfixes-stable-2025-05-17-09-41' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Pull hotfixes from Andrew Morton:
"Nine singleton hotfixes, all MM. Four are cc:stable"
* tag 'mm-hotfixes-stable-2025-05-17-09-41' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
mm: userfaultfd: correct dirty flags set for both present and swap pte
zsmalloc: don't underflow size calculation in zs_obj_write()
mm/page_alloc: fix race condition in unaccepted memory handling
mm/page_alloc: ensure try_alloc_pages() plays well with unaccepted memory
MAINTAINERS: add mm GUP section
mm/codetag: move tag retrieval back upfront in __free_pages()
mm/memory: fix mapcount / refcount sanity check for mTHP reuse
kernel/fork: only call untrack_pfn_clear() on VMAs duplicated for fork()
mm: hugetlb: fix incorrect fallback for subpool
Linus Torvalds [Sat, 17 May 2025 16:02:11 +0000 (09:02 -0700)]
Merge tag 'irq-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull misc irqchip driver fixes from Ingo Molnar:
- Remove the MSI_CHIP_FLAG_SET_ACK flag from 5 irqchip drivers
that did not require it
- Fix IRQ handling delays in the riscv-imsic irqchip driver
* tag 'irq-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/riscv-imsic: Start local sync timer on correct CPU
irqchip: Drop MSI_CHIP_FLAG_SET_ACK from unsuspecting MSI drivers
Linus Torvalds [Sat, 17 May 2025 15:43:51 +0000 (08:43 -0700)]
Merge tag 'x86-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull misc x86 fixes from Ingo Molnar:
- Fix SEV-SNP kdump bugs
- Update the email address of Alexey Makhalov in MAINTAINERS
- Add the CPU feature flag for the Zen6 microarchitecture
- Fix typo in system message
* tag 'x86-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/mm: Remove duplicated word in warning message
x86/CPU/AMD: Add X86_FEATURE_ZEN6
x86/sev: Make sure pages are not skipped during kdump
x86/sev: Do not touch VMSA pages during SNP guest memory kdump
MAINTAINERS: Update Alexey Makhalov's email address
x86/sev: Fix operator precedence in GHCB_MSR_VMPL_REQ_LEVEL macro
Linus Torvalds [Sat, 17 May 2025 15:38:11 +0000 (08:38 -0700)]
Merge tag 'perf-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 perf event fix from Ingo Molnar:
"Fix PEBS-via-PT crash"
* tag 'perf-urgent-2025-05-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Fix segfault with PEBS-via-PT with sample_freq
Linus Torvalds [Sat, 17 May 2025 14:48:25 +0000 (07:48 -0700)]
Merge tag 'loongarch-fixes-6.15-2' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson
Pull LoongArch fixes from Huacai Chen:
"Fix some bugs in kernel-fpu, cpu idle function, hibernation and
uprobes"
* tag 'loongarch-fixes-6.15-2' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson:
LoongArch: uprobes: Remove redundant code about resume_era
LoongArch: uprobes: Remove user_{en,dis}able_single_step()
LoongArch: Save and restore CSR.CNTC for hibernation
LoongArch: Move __arch_cpu_idle() to .cpuidle.text section
LoongArch: Fix MAX_REG_OFFSET calculation
LoongArch: Prevent cond_resched() occurring within kernel-fpu
Linus Torvalds [Sat, 17 May 2025 01:02:41 +0000 (18:02 -0700)]
Merge tag '6.15-rc6-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
- Fix memory leak in mkdir error path
- Fix max rsize miscalculation after channel reconnect
* tag '6.15-rc6-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: fix zero rsize error messages
smb: client: fix memory leak during error handling for POSIX mkdir
Linus Torvalds [Fri, 16 May 2025 23:07:39 +0000 (16:07 -0700)]
Merge tag 'drm-fixes-2025-05-17' of https://gitlab.freedesktop.org/drm/kernel
Pull drm fixes from Dave Airlie:
"Weekly drm fixes, I'll be honest and say I think this is larger than
I'd prefer at this point, the main blow out point is that xe has two
larger fixes.
One is a fix for active context utilisation reporting, it's for a
reported regression and will end up in stable anyways, so I don't see
any point in holding it up.
The second is a fix for mixed cpu/gpu atomics, which are currently
broken, but are also not something your average desktop/laptop user is
going to hit in normal operation, and having them fixed now is better
than threading them through stable later.
Other than those, it's mostly the usual, a bunch of amdgpu randoms and
a few other minor fixes.
dma-buf:
- Avoid memory reordering in fence handling
meson:
- Avoid integer overflow in mode-clock calculations
panel-mipi-dbi:
- Fix output with drm_client_setup_with_fourcc()
gpuvm:
- Add timeslicing and allocation restriction for SVM
xe:
- Fix shrinker debugfs name
- Add HW workaround to Xe2
- Fix SVM when mixing GPU and CPU atomics
- Fix per client engine utilization due to active contexts not saving
timestamp with lite restore enabled"
* tag 'drm-fixes-2025-05-17' of https://gitlab.freedesktop.org/drm/kernel: (24 commits)
drm/xe: Add WA BB to capture active context utilization
drm/xe: Save the gt pointer in lrc and drop the tile
drm/xe: Save CTX_TIMESTAMP mmio value instead of LRC value
drm/xe: Timeslice GPU on atomic SVM fault
drm/gpusvm: Add timeslicing support to GPU SVM
drm/xe: Strict migration policy for atomic SVM faults
drm/gpusvm: Introduce devmem_only flag for allocation
drm/xe/xe2hpg: Add Wa_22021007897
drm/amdgpu: read back register after written for VCN v4.0.5
Revert "drm/amd/display: Hardware cursor changes color when switched to software cursor"
dma-buf: insert memory barrier before updating num_fences
drm/xe: Fix the gem shrinker name
drm/amd/display: Avoid flooding unnecessary info messages
drm/amd/display: Fix null check of pipe_ctx->plane_state for update_dchubp_dpp
drm/amd/display: check stream id dml21 wrapper to get plane_id
drm/amd/display: fix link_set_dpms_off multi-display MST corner case
drm/amd/display: Defer BW-optimization-blocked DRR adjustments
Revert: "drm/amd/display: Enable urgent latency adjustment on DCN35"
drm/amd/display: Correct the reply value when AUX write incomplete
drm/amdgpu: fix incorrect MALL size for GFX1151
...
Linus Torvalds [Fri, 16 May 2025 21:29:12 +0000 (14:29 -0700)]
Merge tag 'nfs-for-6.15-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Pull NFS client bugfixes from Trond Myklebust:
- NFS: Fix a couple of missed handlers for the ENETDOWN and ENETUNREACH
transport errors
- NFS: Handle Oopsable failure of nfs_get_lock_context in the unlock
path
- NFSv4: Fix a race in nfs_local_open_fh()
- NFSv4/pNFS: Fix a couple of layout segment leaks in layoutreturn
- NFSv4/pNFS Avoid sharing pNFS DS connections between net namespaces
since IP addresses are not guaranteed to refer to the same nodes
- NFS: Don't flush file data while holding multiple directory locks in
nfs_rename()
* tag 'nfs-for-6.15-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
NFS: Avoid flushing data while holding directory locks in nfs_rename()
NFS/pnfs: Fix the error path in pnfs_layoutreturn_retry_later_locked()
NFSv4/pnfs: Reset the layout state after a layoutreturn
NFS/localio: Fix a race in nfs_local_open_fh()
nfs: nfs3acl: drop useless assignment in nfs3_get_acl()
nfs: direct: drop useless initializer in nfs_direct_write_completion()
nfs: move the nfs4_data_server_cache into struct nfs_net
nfs: don't share pNFS DS connections between net namespaces
nfs: handle failure of nfs_get_lock_context in unlock path
pNFS/flexfiles: Record the RPC errors in the I/O tracepoints
NFSv4/pnfs: Layoutreturn on close must handle fatal networking errors
NFSv4: Handle fatal ENETDOWN and ENETUNREACH errors
NFS: Avoid flushing data while holding directory locks in nfs_rename()
The Linux client assumes that all filehandles are non-volatile for
renames within the same directory (otherwise sillyrename cannot work).
However, the existence of the Linux 'subtree_check' export option has
meant that nfs_rename() has always assumed it needs to flush writes
before attempting to rename.
Since NFSv4 does allow the client to query whether or not the server
exhibits this behaviour, and since knfsd does actually set the
appropriate flag when 'subtree_check' is enabled on an export, it
should be OK to optimise away the write flushing behaviour in the cases
where it is clearly not needed.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com> Reviewed-by: Jeff Layton <jlayton@kernel.org>
Trond Myklebust [Sat, 10 May 2025 14:50:13 +0000 (10:50 -0400)]
NFSv4/pnfs: Reset the layout state after a layoutreturn
If there are still layout segments in the layout plh_return_lsegs list
after a layout return, we should be resetting the state to ensure they
eventually get returned as well.
Fixes: 68f744797edd ("pNFS: Do not free layout segments that are marked for return") Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Linus Torvalds [Fri, 16 May 2025 17:21:25 +0000 (10:21 -0700)]
Merge tag 'block-6.15-20250515' of git://git.kernel.dk/linux
Pull block fixes from Jens Axboe:
- NVMe pull request via Christoph:
- fixes for atomic writes (Alan Adamson)
- fixes for polled CQs in nvmet-epf (Damien Le Moal)
- fix for polled CQs in nvme-pci (Keith Busch)
- fix compile on odd configs that need to be forced to inline
(Kees Cook)
- one more quirk (Ilya Guterman)
- Fix for missing allocation of an integrity buffer for some cases
- Fix for a regression with ublk command cancelation
* tag 'block-6.15-20250515' of git://git.kernel.dk/linux:
ublk: fix dead loop when canceling io command
nvme-pci: add NVME_QUIRK_NO_DEEPEST_PS quirk for SOLIDIGM P44 Pro
nvme: all namespaces in a subsystem must adhere to a common atomic write size
nvme: multipath: enable BLK_FEAT_ATOMIC_WRITES for multipathing
nvmet: pci-epf: remove NVMET_PCI_EPF_Q_IS_SQ
nvmet: pci-epf: improve debug message
nvmet: pci-epf: cleanup nvmet_pci_epf_raise_irq()
nvmet: pci-epf: do not fall back to using INTX if not supported
nvmet: pci-epf: clear completion queue IRQ flag on delete
nvme-pci: acquire cq_poll_lock in nvme_poll_irqdisable
nvme-pci: make nvme_pci_npages_prp() __always_inline
block: always allocate integrity buffer when required
Linus Torvalds [Fri, 16 May 2025 16:59:24 +0000 (09:59 -0700)]
Merge tag 'io_uring-6.15-20250515' of git://git.kernel.dk/linux
Pull io_uring fixes from Jens Axboe:
- Fix a regression with highmem and mapping of regions, where
the coalescing code assumes any page is directly mapped
- Fix an issue with HYBRID_IOPOLL and passthrough commands,
where the timer wasn't always setup correctly
- Fix an issue with fdinfo not correctly locking around reading
the rings, which can be an issue if the ring is being resized
at the same time
* tag 'io_uring-6.15-20250515' of git://git.kernel.dk/linux:
io_uring/fdinfo: grab ctx->uring_lock around io_uring_show_fdinfo()
io_uring/memmap: don't use page_address() on a highmem page
io_uring/uring_cmd: fix hybrid polling initialization issue
Linus Torvalds [Fri, 16 May 2025 16:51:49 +0000 (09:51 -0700)]
Merge tag 'xfs-fixes-6.15-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull xfs fixes from Carlos Maiolino:
"This includes a bug fix for a possible data corruption vector on the
zoned allocator garbage collector"
* tag 'xfs-fixes-6.15-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
xfs: Fix comment on xfs_trans_ail_update_bulk()
xfs: Fix a comment on xfs_ail_delete
xfs: Fail remount with noattr2 on a v5 with v4 enabled
xfs: fix zoned GC data corruption due to wrong bv_offset
xfs: free up mp->m_free[0].count in error case
Linus Torvalds [Fri, 16 May 2025 16:40:07 +0000 (09:40 -0700)]
Merge tag 'acpi-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull ACPI fix from Rafael Wysocki:
"Fix ACPI PPTT parsing code to address a regression introduced recently
and add more sanity checking of data supplied by the platform firmware
to avoid using invalid data (Jeremy Linton)"
* tag 'acpi-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI: PPTT: Fix processor subtable walk
Linus Torvalds [Fri, 16 May 2025 16:24:54 +0000 (09:24 -0700)]
Merge tag 'spi-fix-v6.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fixes from Mark Brown:
"A few small driver specific fixes, the most substantial one being the
Tegra one which fixes spurious errors with default delays for chip
select hold times"
* tag 'spi-fix-v6.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: spi-sun4i: fix early activation
spi: tegra114: Use value to check for invalid delays
spi: loopback-test: Do not split 1024-byte hexdumps
Linus Torvalds [Fri, 16 May 2025 16:19:37 +0000 (09:19 -0700)]
Merge tag 'regulator-fix-v6.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator
Pull regulator fix from Mark Brown:
"This fixes an invalid memory access in the MAX20086 driver which could
occur during error handling for failed probe due to a hidden use of
devres in the core DT parsing code"
* tag 'regulator-fix-v6.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
regulator: max20086: fix invalid memory access
Linus Torvalds [Fri, 16 May 2025 16:13:51 +0000 (09:13 -0700)]
Merge tag 'gpio-fixes-for-v6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux
Pull gpio fixes from Bartosz Golaszewski:
- fix an interrupt storm on system wake-up in gpio-pca953x
- fix an out-of-bounds write in gpio-virtuser
- update MAINTAINERS with an entry for the sloppy logic analyzer
* tag 'gpio-fixes-for-v6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
gpio: virtuser: fix potential out-of-bound write
gpio: pca953x: fix IRQ storm on system wake up
MAINTAINERS: add me as maintainer for the gpio sloppy logic analyzer
Linus Torvalds [Fri, 16 May 2025 16:06:12 +0000 (09:06 -0700)]
Merge tag 'sound-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
"A handful small fixes. The only significant change is the fix for MIDI
2.0 UMP handling in ALSA sequencer, but as MIDI 2.0 stuff is still new
and rarely used, the impact should be pretty limited.
Other than that, quirks for USB-audio and a few cosmetic fixes and
changes in drivers that should be safe to apply"
* tag 'sound-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: usb-audio: Add sample rate quirk for Microdia JP001 USB Camera
ALSA: es1968: Add error handling for snd_pcm_hw_constraint_pow2()
ALSA: sh: SND_AICA should depend on SH_DMA_API
ALSA: usb-audio: Add sample rate quirk for Audioengine D1
ALSA: ump: Fix a typo of snd_ump_stream_msg_device_info
ALSA/hda: intel-sdw-acpi: Correct sdw_intel_acpi_scan() function parameter
ALSA: seq: Fix delivery of UMP events to group ports
Lukas Bulwahn [Fri, 16 May 2025 09:08:10 +0000 (11:08 +0200)]
x86/mm: Remove duplicated word in warning message
Commit bbeb69ce3013 ("x86/mm: Remove CONFIG_HIGHMEM64G support") introduces
a new warning message MSG_HIGHMEM_TRIMMED, which accidentally introduces a
duplicated 'for for' in the warning message.
Remove this duplicated word.
This was noticed while reviewing for references to obsolete kernel build
config options.
Fixes: bbeb69ce3013 ("x86/mm: Remove CONFIG_HIGHMEM64G support") Signed-off-by: Lukas Bulwahn <lukas.bulwahn@redhat.com> Signed-off-by: Ingo Molnar <mingo@kernel.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: H. Peter Anvin <hpa@zytor.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: kernel-janitors@vger.kernel.org Link: https://lore.kernel.org/r/20250516090810.556623-1-lukas.bulwahn@redhat.com
Dave Airlie [Fri, 16 May 2025 01:21:29 +0000 (11:21 +1000)]
Merge tag 'drm-xe-fixes-2025-05-15-1' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes
Core Changes:
- Add timeslicing and allocation restriction for SVM
Driver Changes:
- Fix shrinker debugfs name
- Add HW workaround to Xe2
- Fix SVM when mixing GPU and CPU atomics
- Fix per client engine utilization due to active contexts
not saving timestamp with lite restore enabled.
Linus Torvalds [Thu, 15 May 2025 21:20:48 +0000 (14:20 -0700)]
Merge tag 'bcachefs-2025-05-15' of git://evilpiepirate.org/bcachefs
Pull bcachefs fixes from Kent Overstreet:
"The main user reported ones are:
- Fix a btree iterator locking inconsistency that's been causing us
to go emergency read-only in evacuate: "Fix broken btree_path lock
invariants in next_node()"
- Minor btree node cache reclaim tweak that should help with OOMs:
don't set btree nodes as accessed on fill
- Fix a bch2_bkey_clear_rebalance() issue that was causing rebalance
to do needless work"
* tag 'bcachefs-2025-05-15' of git://evilpiepirate.org/bcachefs:
bcachefs: fix wrong arg to fsck_err()
bcachefs: Fix missing commit in backpointer to missing target
bcachefs: Fix accidental O(n^2) in fiemap
bcachefs: Fix set_should_be_locked() call in peek_slot()
bcachefs: Fix self deadlock
bcachefs: Don't set btree nodes as accessed on fill
bcachefs: Fix livelock in journal_entry_open()
bcachefs: Fix broken btree_path lock invariants in next_node()
bcachefs: Don't strip rebalance_opts from indirect extents
Linus Torvalds [Thu, 15 May 2025 19:26:09 +0000 (12:26 -0700)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
"Four small fixes for crashes:
- Double free in rxe
- UAF in irdma from early freeing the rf
- Off by one undoing the IRQ allocations during error unwind in irdma
- Another race with device rename and uevent generation. uevents
accesses the struct device name and UAF when it is changed"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/core: Fix "KASAN: slab-use-after-free Read in ib_register_device" problem
ice, irdma: fix an off by one in error handling code
irdma: free iwdev->rf after removing MSI-X
RDMA/rxe: Fix slab-use-after-free Read in rxe_queue_cleanup bug
Linus Torvalds [Thu, 15 May 2025 18:04:36 +0000 (11:04 -0700)]
Merge tag 'landlock-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux
Pull landlock fixes from Mickaël Salaün:
"This fixes a KUnit issue, simplifies code, and adds new tests"
* tag 'landlock-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux:
landlock: Improve bit operations in audit code
landlock: Remove KUnit test that triggers a warning
Linus Torvalds [Thu, 15 May 2025 17:57:24 +0000 (10:57 -0700)]
Merge tag 'hid-for-linus-2025051501' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Pull HID fixes from Benjamin Tissoires:
- fix a few potential memory leaks in the wacom driver (Qasim Ijaz)
- AMD SFH fixes when there is only one SRA sensor (Mario Limonciello)
- HID-BPF dispatch UAF fix that happens on removal of the Logitech DJ
receiver (Rong Zhang)
- various minor fixes and usual device ID additions
* tag 'hid-for-linus-2025051501' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
HID: bpf: abort dispatch if device destroyed
HID: quirks: Add ADATA XPG alpha wireless mouse support
HID: hid-steam: Remove the unused variable connected
HID: amd_sfh: Avoid clearing reports for SRA sensor
HID: amd_sfh: Fix SRA sensor when it's the only sensor
HID: wacom: fix shift OOB in kfifo allocation for zero pktlen
HID: uclogic: Add NULL check in uclogic_input_configured()
HID: wacom: fix memory leak on size mismatch in wacom_wac_queue_flush()
HID: wacom: handle kzalloc() allocation failure in wacom_wac_queue_flush()
HID: thrustmaster: fix memory leak in thrustmaster_interrupts()
HID: hid-appletb-kbd: Fix wrong date and kernel version in sysfs interface docs
HID: bpf: fix BTN_STYLUS for the XP Pen ACK05 remote
Linus Torvalds [Thu, 15 May 2025 17:40:49 +0000 (10:40 -0700)]
Merge tag 'net-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from Bluetooth and wireless.
A few more fixes for the locking changes trickling in. Nothing too
alarming, I suspect those will continue for another release. Other
than that things are slowing down nicely.
Current release - fix to a fix:
- Bluetooth: hci_event: use key encryption size when its known
- tools: ynl-gen: allow multi-attr without nested-attributes again
Current release - regressions:
- locking fixes:
- lock lower level devices when updating features
- eth: bnxt_en: bring back rtnl_lock() in the bnxt_open() path
- devmem: fix panic when Netlink socket closes after module unload
Current release - new code bugs:
- eth: txgbe: fixes for FW communication on new AML devices
Previous releases - always broken:
- sched: flush gso_skb list too during ->change(), avoid potential
null-deref on reconfig
Ming Lei [Thu, 15 May 2025 16:26:01 +0000 (00:26 +0800)]
ublk: fix dead loop when canceling io command
Commit:
f40139fde527 ("ublk: fix race between io_uring_cmd_complete_in_task and
ublk_cancel_cmd")
adds a request state check in ublk_cancel_cmd(), and if the request is
started, skips canceling this uring_cmd.
However, the current uring_cmd may be in ACTIVE state, without block
request coming to the uring command. Meantime, if the cached request in
tag_set.tags[tag] has been delivered to ublk server and reycycled, then
this uring_cmd can't be canceled.
ublk requests are aborted in ublk char device release handler, which
depends on canceling all ACTIVE uring_cmd. So it causes a dead loop.
Fix this issue by not taking a stale request into account when canceling
uring_cmd in ublk_cancel_cmd().
Reported-by: Shinichiro Kawasaki <shinichiro.kawasaki@wdc.com> Closes: https://lore.kernel.org/linux-block/mruqwpf4tqenkbtgezv5oxwq7ngyq24jzeyqy4ixzvivatbbxv@4oh2wzz4e6qn/ Fixes: f40139fde527 ("ublk: fix race between io_uring_cmd_complete_in_task and ublk_cancel_cmd") Signed-off-by: Ming Lei <ming.lei@redhat.com> Link: https://lore.kernel.org/r/20250515162601.77346-1-ming.lei@redhat.com
[axboe: rewording of commit message] Signed-off-by: Jens Axboe <axboe@kernel.dk>
That happens because intel_pmu_pebs_event_update_no_drain() assumes all the
pebs_enabled bits represent counter indexes, which is not always the case.
In this particular case, bits 60 and 61 are set for PEBS-via-PT purposes.
The behaviour of PEBS-via-PT with sample frequency is questionable because
although a PMI is generated (PEBS_PMI_AFTER_EACH_RECORD), the period is not
adjusted anyway.
Putting that aside, fix intel_pmu_pebs_event_update_no_drain() by passing
the mask of counter bits instead of 'size'. Note, prior to the Fixes
commit, 'size' would be limited to the maximum counter index, so the issue
was not hit.
Fixes: 722e42e45c2f1 ("perf/x86: Support counter mask") Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ingo Molnar <mingo@kernel.org> Reviewed-by: Kan Liang <kan.liang@linux.intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ian Rogers <irogers@google.com> Cc: linux-perf-users@vger.kernel.org Link: https://lore.kernel.org/r/20250508134452.73960-1-adrian.hunter@intel.com
Taehee Yoo [Wed, 14 May 2025 15:40:28 +0000 (15:40 +0000)]
net: devmem: fix kernel panic when netlink socket close after module unload
Kernel panic occurs when a devmem TCP socket is closed after NIC module
is unloaded.
This is Devmem TCP unregistration scenarios. number is an order.
(a)netlink socket close (b)pp destroy (c)uninstall result
1 2 3 OK
1 3 2 (d)Impossible
2 1 3 OK
3 1 2 (e)Kernel panic
2 3 1 (d)Impossible
3 2 1 (d)Impossible
(a) netdev_nl_sock_priv_destroy() is called when devmem TCP socket is
closed.
(b) page_pool_destroy() is called when the interface is down.
(c) mp_ops->uninstall() is called when an interface is unregistered.
(d) There is no scenario in mp_ops->uninstall() is called before
page_pool_destroy().
Because unregister_netdevice_many_notify() closes interfaces first
and then calls mp_ops->uninstall().
(e) netdev_nl_sock_priv_destroy() accesses struct net_device to acquire
netdev_lock().
But if the interface module has already been removed, net_device
pointer is invalid, so it causes kernel panic.
In summary, there are only 3 possible scenarios.
A. sk close -> pp destroy -> uninstall.
B. pp destroy -> sk close -> uninstall.
C. pp destroy -> uninstall -> sk close.
Case C is a kernel panic scenario.
In order to fix this problem, It makes mp_dmabuf_devmem_uninstall() set
binding->dev to NULL.
It indicates an bound net_device was unregistered.
It makes netdev_nl_sock_priv_destroy() do not acquire netdev_lock()
if binding->dev is NULL.
A new binding->lock is added to protect a dev of a binding.
So, lock ordering is like below.
priv->lock
netdev_lock(dev)
binding->lock
Tests:
Scenario A:
./ncdevmem -s 192.168.1.4 -c 192.168.1.2 -f $interface -l -p 8000 \
-v 7 -t 1 -q 1 &
pid=$!
sleep 10
kill $pid
ip link set $interface down
modprobe -rv $module
Scenario B:
./ncdevmem -s 192.168.1.4 -c 192.168.1.2 -f $interface -l -p 8000 \
-v 7 -t 1 -q 1 &
pid=$!
sleep 10
ip link set $interface down
kill $pid
modprobe -rv $module
Gerhard Engleder [Wed, 14 May 2025 19:56:57 +0000 (21:56 +0200)]
tsnep: fix timestamping with a stacked DSA driver
This driver is susceptible to a form of the bug explained in commit c26a2c2ddc01 ("gianfar: Fix TX timestamping with a stacked DSA driver")
and in Documentation/networking/timestamping.rst section "Other caveats
for MAC drivers", specifically it timestamps any skb which has
SKBTX_HW_TSTAMP, and does not consider if timestamping has been enabled
in adapter->hwtstamp_config.tx_type.
Evaluate the proper TX timestamping condition only once on the TX
path (in tsnep_xmit_frame_ring()) and store the result in an additional
TX entry flag. Evaluate the new TX entry flag in the TX confirmation path
(in tsnep_tx_poll()).
This way SKBTX_IN_PROGRESS is set by the driver as required, but never
evaluated. SKBTX_IN_PROGRESS shall not be evaluated as it can be set
by a stacked DSA driver and evaluating it would lead to unwanted
timestamps.
Fixes: 403f69bbdbad ("tsnep: Add TSN endpoint Ethernet MAC driver") Suggested-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: Gerhard Engleder <gerhard@engleder-embedded.com> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com> Link: https://patch.msgid.link/20250514195657.25874-1-gerhard@engleder-embedded.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pengtao He [Wed, 14 May 2025 13:20:13 +0000 (21:20 +0800)]
net/tls: fix kernel panic when alloc_page failed
We cannot set frag_list to NULL pointer when alloc_page failed.
It will be used in tls_strp_check_queue_ok when the next time
tls_strp_read_sock is called.
This is because we don't reset full_len in tls_strp_flush_anchor_copy()
so the recv path will try to continue handling the partial record
on the next call but we dettached the rcvq from the frag list.
Alternative fix would be to reset full_len.
Unable to handle kernel NULL pointer dereference
at virtual address 0000000000000028
Call trace:
tls_strp_check_rcv+0x128/0x27c
tls_strp_data_ready+0x34/0x44
tls_data_ready+0x3c/0x1f0
tcp_data_ready+0x9c/0xe4
tcp_data_queue+0xf6c/0x12d0
tcp_rcv_established+0x52c/0x798
Jakub Kicinski [Thu, 15 May 2025 14:19:49 +0000 (07:19 -0700)]
Merge tag 'wireless-2025-05-15' of https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless
Johannes Berg says:
====================
Couple of stragglers:
- mac80211: fix syzbot/ubsan in scan counted-by
- mt76: fix NAPI handling on driver remove
- mt67: fix multicast/ipv6 receive
* tag 'wireless-2025-05-15' of https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
wifi: mac80211: Set n_channels after allocating struct cfg80211_scan_request
wifi: mt76: mt7925: fix missing hdr_trans_tlv command for broadcast wtbl
wifi: mt76: disable napi on driver removal
====================
Michael Chan [Wed, 14 May 2025 06:29:08 +0000 (23:29 -0700)]
bnxt_en: bring back rtnl_lock() in the bnxt_open() path
Error recovery, PCIe AER, resume, and TX timeout will invoke bnxt_open()
with netdev_lock only. This will cause RTNL assert failure in
netif_set_real_num_tx_queues(), netif_set_real_num_tx_queues(),
and netif_set_real_num_tx_queues().
Example error recovery assert:
RTNL: assertion failed at net/core/dev.c (3178)
WARNING: CPU: 3 PID: 3392 at net/core/dev.c:3178 netif_set_real_num_tx_queues+0x1fd/0x210
For now, bring back rtnl_lock() in all these code paths that can invoke
bnxt_open(). In the bnxt_queue_start() error path, we don't have
rtnl_lock held so we just change it to call netif_close() instead of
bnxt_reset_task() for simplicity. This error path is unlikely so it
should be fine.
Fixes: 004b5008016a ("eth: bnxt: remove most dependencies on RTNL") Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20250514062908.2766677-1-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ido Schimmel [Wed, 14 May 2025 12:48:05 +0000 (14:48 +0200)]
mlxsw: spectrum_router: Fix use-after-free when deleting GRE net devices
The driver only offloads neighbors that are constructed on top of net
devices registered by it or their uppers (which are all Ethernet). The
device supports GRE encapsulation and decapsulation of forwarded
traffic, but the driver will not offload dummy neighbors constructed on
top of GRE net devices as they are not uppers of its net devices:
# ip link add name gre1 up type gre tos inherit local 192.0.2.1 remote 198.51.100.1
# ip neigh add 0.0.0.0 lladdr 0.0.0.0 nud noarp dev gre1
$ ip neigh show dev gre1 nud noarp
0.0.0.0 lladdr 0.0.0.0 NOARP
(Note that the neighbor is not marked with 'offload')
When the driver is reloaded and the existing configuration is replayed,
the driver does not perform the same check regarding existing neighbors
and offloads the previously added one:
# devlink dev reload pci/0000:01:00.0
$ ip neigh show dev gre1 nud noarp
0.0.0.0 lladdr 0.0.0.0 offload NOARP
If the neighbor is later deleted, the driver will ignore the
notification (given the GRE net device is not its upper) and will
therefore keep referencing freed memory, resulting in a use-after-free
[1] when the net device is deleted:
# ip neigh del 0.0.0.0 lladdr 0.0.0.0 dev gre1
# ip link del dev gre1
Fix by skipping neighbor replay if the net device for which the replay
is performed is not our upper.
irqchip/riscv-imsic: Start local sync timer on correct CPU
When starting the local sync timer to synchronize the state of a remote
CPU it should be added on the CPU to be synchronized, not the initiating
CPU. This results in interrupt delivery being delayed until the timer
eventually runs (due to another mask/unmask/migrate operation) on the
target CPU.
Fixes: 0f67911e821c ("irqchip/riscv-imsic: Separate next and previous pointers in IMSIC vector") Signed-off-by: Andrew Bresticker <abrestic@rivosinc.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Anup Patel <anup@brainfault.org> Link: https://lore.kernel.org/all/20250514171320.3494917-1-abrestic@rivosinc.com
Jens Axboe [Thu, 15 May 2025 12:20:56 +0000 (06:20 -0600)]
Merge tag 'nvme-6.15-2025-05-15' of git://git.infradead.org/nvme into block-6.15
Pull NVMe fixes from Christoph:
"nvme fixes for linux 6.15
- fixes for atomic writes (Alan Adamson)
- fixes for polled CQs in nvmet-epf (Damien Le Moal)
- fix for polled CQs in nvme-pci (Keith Busch)
- fix compile on odd configs that need to be forced to inline
(Kees Cook)
- one more quirk (Ilya Guterman)"
* tag 'nvme-6.15-2025-05-15' of git://git.infradead.org/nvme:
nvme-pci: add NVME_QUIRK_NO_DEEPEST_PS quirk for SOLIDIGM P44 Pro
nvme: all namespaces in a subsystem must adhere to a common atomic write size
nvme: multipath: enable BLK_FEAT_ATOMIC_WRITES for multipathing
nvmet: pci-epf: remove NVMET_PCI_EPF_Q_IS_SQ
nvmet: pci-epf: improve debug message
nvmet: pci-epf: cleanup nvmet_pci_epf_raise_irq()
nvmet: pci-epf: do not fall back to using INTX if not supported
nvmet: pci-epf: clear completion queue IRQ flag on delete
nvme-pci: acquire cq_poll_lock in nvme_poll_irqdisable
nvme-pci: make nvme_pci_npages_prp() __always_inline
Kees Cook [Fri, 9 May 2025 18:46:45 +0000 (11:46 -0700)]
wifi: mac80211: Set n_channels after allocating struct cfg80211_scan_request
Make sure that n_channels is set after allocating the
struct cfg80211_registered_device::int_scan_req member. Seen with
syzkaller:
UBSAN: array-index-out-of-bounds in net/mac80211/scan.c:1208:5
index 0 is out of range for type 'struct ieee80211_channel *[] __counted_by(n_channels)' (aka 'struct ieee80211_channel *[]')
This was missed in the initial conversions because I failed to locate
the allocation likely due to the "sizeof(void *)" not matching the
"channels" array type.
Reported-by: syzbot+4bcdddd48bb6f0be0da1@syzkaller.appspotmail.com Closes: https://lore.kernel.org/lkml/680fd171.050a0220.2b69d1.045e.GAE@google.com/ Fixes: e3eac9f32ec0 ("wifi: cfg80211: Annotate struct cfg80211_scan_request with __counted_by") Signed-off-by: Kees Cook <kees@kernel.org> Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org> Link: https://patch.msgid.link/20250509184641.work.542-kees@kernel.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Nicolas Chauvet [Thu, 15 May 2025 10:21:32 +0000 (12:21 +0200)]
ALSA: usb-audio: Add sample rate quirk for Microdia JP001 USB Camera
Microdia JP001 does not support reading the sample rate which leads to
many lines of "cannot get freq at ep 0x84".
This patch adds the USB ID to quirks.c and avoids those error messages.
usb 7-4: New USB device found, idVendor=0c45, idProduct=636b, bcdDevice= 1.00
usb 7-4: New USB device strings: Mfr=2, Product=1, SerialNumber=3
usb 7-4: Product: JP001
usb 7-4: Manufacturer: JP001
usb 7-4: SerialNumber: JP001
usb 7-4: 3:1: cannot get freq at ep 0x84
Vinod Koul [Thu, 15 May 2025 10:42:13 +0000 (11:42 +0100)]
dmaengine: mediatek: drop unused variable
Commit 157ae5ffd76a dmaengine: mediatek: Fix a possible deadlock error
in mtk_cqdma_tx_status() fixed locks but kept unused varibale leading to
warning and build failure (due to warning treated as errors)
drivers/dma/mediatek/mtk-cqdma.c: In function 'mtk_cqdma_find_active_desc':
drivers/dma/mediatek/mtk-cqdma.c:423:23: error: unused variable 'flags' [-Werror=unused-variable]
423 | unsigned long flags;
| ^~~~~
Fix by dropping this unused flag
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Fixes: 157ae5ffd76a ("dmaengine: mediatek: Fix a possible deadlock error in mtk_cqdma_tx_status()") Signed-off-by: Vinod Koul <vkoul@kernel.org>
octeontx2-pf: Do not reallocate all ntuple filters
If ntuple filters count is modified followed by
unicast filters count using devlink then the ntuple count
set by user is ignored and all the ntuple filters are
being reallocated. Fix this by storing the ntuple count
set by user. Without this patch, say if user tries
to modify ntuple count as 8 followed by ucast filter count as 4
using devlink commands then ntuple count is being reverted to
default value 16 i.e, not retaining user set value 8.
Fixes: 39c469188b6d ("octeontx2-pf: Add ucast filter count configurability via devlink.") Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/1747054357-5850-1-git-send-email-sbhatta@marvell.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Fedor Pchelkin [Tue, 6 May 2025 11:55:39 +0000 (14:55 +0300)]
wifi: mt76: disable napi on driver removal
A warning on driver removal started occurring after commit 9dd05df8403b
("net: warn if NAPI instance wasn't shut down"). Disable tx napi before
deleting it in mt76_dma_cleanup().
Tested with mt7921e but the same pattern can be actually applied to other
mt76 drivers calling mt76_dma_cleanup() during removal. Tx napi is enabled
in their *_dma_init() functions and only toggled off and on again inside
their suspend/resume/reset paths. So it should be okay to disable tx
napi in such a generic way.
Found by Linux Verification Center (linuxtesting.org).
Fixes: 2ac515a5d74f ("mt76: mt76x02: use napi polling for tx cleanup") Cc: stable@vger.kernel.org Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru> Tested-by: Ming Yen Hsieh <mingyen.hsieh@mediatek.com> Link: https://patch.msgid.link/20250506115540.19045-1-pchelkin@ispras.ru Signed-off-by: Felix Fietkau <nbd@nbd.name>
Linus Torvalds [Thu, 15 May 2025 05:24:17 +0000 (22:24 -0700)]
Merge tag 'kbuild-fixes-v6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild
Pull Kbuild fixes from Masahiro Yamada:
- Add proper pahole version dependency to CONFIG_GENDWARFKSYMS to avoid
module loading errors
- Fix UAPI header tests for the OpenRISC architecture
- Add dependency on the libdw package in Debian and RPM packages
- Disable -Wdefault-const-init-unsafe warnings on Clang
- Make "make clean ARCH=um" also clean the arch/x86/ directory
- Revert the use of -fmacro-prefix-map=, which causes issues with
debugger usability
* tag 'kbuild-fixes-v6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
kbuild: fix typos "module.builtin" to "modules.builtin"
Revert "kbuild, rust: use -fremap-path-prefix to make paths relative"
Revert "kbuild: make all file references relative to source root"
kbuild: fix dependency on sorttable
init: remove unused CONFIG_CC_CAN_LINK_STATIC
um: let 'make clean' properly clean underlying SUBARCH as well
kbuild: Disable -Wdefault-const-init-unsafe
kbuild: rpm-pkg: Add (elfutils-devel or libdw-devel) to BuildRequires
kbuild: deb-pkg: Add libdw-dev:native to Build-Depends-Arch
usr/include: openrisc: don't HDRTEST bpf_perf_event.h
kbuild: Require pahole <v1.28 or >v1.29 with GENDWARFKSYMS on X86
Starting with commit dca5161f9bd0 in the 6.3 kernel, the Linux driver
for Hyper-V synthetic networking (netvsc) occasionally reports
"nvsp_rndis_pkt_complete error status: 2".[1] This error indicates
that Hyper-V has rejected a network packet transmit request from the
guest, and the outgoing network packet is dropped. Higher level
network protocols presumably recover and resend the packet so there is
no functional error, but performance is slightly impacted. Commit dca5161f9bd0 is not the cause of the error -- it only added reporting
of an error that was already happening without any notice. The error
has presumably been present since the netvsc driver was originally
introduced into Linux.
This patch set fixes the root cause of the problem, which is that the
netvsc driver in Linux may send an incorrectly formatted VMBus message
to Hyper-V when transmitting the network packet. The incorrect
formatting occurs when the rndis header of the VMBus message crosses a
page boundary due to how the Linux skb head memory is aligned. In such
a case, two PFNs are required to describe the location of the rndis
header, even though they are contiguous in guest physical address
(GPA) space. Hyper-V requires that two PFNs be in a single "GPA range"
data struture, but current netvsc code puts each PFN in its own GPA
range, which Hyper-V rejects as an error in the case of the rndis
header.
The incorrect formatting occurs only for larger packets that netvsc
must transmit via a VMBus "GPA Direct" message. There's no problem
when netvsc transmits a smaller packet by copying it into a pre-
allocated send buffer slot because the pre-allocated slots don't have
page crossing issues.
After commit 14ad6ed30a10 in the 6.14 kernel, the error occurs much
more frequently in VMs with 16 or more vCPUs. It may occur every few
seconds, or even more frequently, in a ssh session that outputs a lot
of text. Commit 14ad6ed30a10 subtly changes how skb head memory is
allocated, making it much more likely that the rndis header will cross
a page boundary when the vCPU count is 16 or more. The changes in
commit 14ad6ed30a10 are perfectly valid -- they just had the side
effect of making the netvsc bug more prominent.
One fix is to check for adjacent PFNs in vmbus_sendpacket_pagebuffer()
and just combine them into a single GPA range. Such a fix is very
contained. But conceptually it is fixing the problem at the wrong
level. So this patch set takes the broader approach of maintaining
the already known grouping of contiguous PFNs at a higher level in
the netvsc driver code, and propagating that grouping down to the
creation of the VMBus message to send to Hyper-V. Maintaining the
grouping fixes this problem, and has the added benefit of allowing
netvsc_dma_map() to make fewer calls to dma_map_single() to do bounce
buffering in CoCo VMs.
Patch 1 is a preparatory change to allow vmbus_sendpacket_mpb_desc()
to specify multiple GPA ranges. In current code
vmbus_sendpacket_mpb_desc() is used only by the storvsc synthetic SCSI
driver, and it always creates a single GPA range.
Patch 2 updates the netvsc driver to use vmbus_sendpacket_mpb_desc()
instead of vmbus_sendpacket_pagebuffer(). Because the higher levels of
netvsc still don't group contiguous PFNs, this patch is functionally
neutral. The VMBus message to Hyper-V still has many GPA ranges, each
with a single PFN. But it lays the groundwork for the next patch.
Patch 3 changes the higher levels of netvsc to preserve the already
known grouping of contiguous PFNs. When the contiguous groupings are
passed to vmbus_sendpacket_mpb_desc(), GPA ranges containing multiple
PFNs are produced, as expected by Hyper-V. This is point at which the
core problem is fixed.
Patches 4 and 5 remove code that is no longer necessary after the
previous patches.
These changes provide a net reduction of about 65 lines of code, which
is an added benefit.
These changes have been tested in normal VMs, in SEV-SNP and TDX CoCo
VMs, and in Dv6-series VMs where the netvsp implementation is in the
OpenHCL paravisor instead of the Hyper-V host.
These changes are built against kernel version 6.15-rc6.
With the netvsc driver changed to use vmbus_sendpacket_mpb_desc()
instead of vmbus_sendpacket_pagebuffer(), the latter has no remaining
callers. Remove it.
Michael Kelley [Tue, 13 May 2025 00:06:03 +0000 (17:06 -0700)]
hv_netvsc: Remove rmsg_pgcnt
init_page_array() now always creates a single page buffer array entry
for the rndis message, even if the rndis message crosses a page
boundary. As such, the number of page buffer array entries used for
the rndis message must no longer be tracked -- it is always just 1.
Remove the rmsg_pgcnt field and use "1" where the value is needed.
Michael Kelley [Tue, 13 May 2025 00:06:02 +0000 (17:06 -0700)]
hv_netvsc: Preserve contiguous PFN grouping in the page buffer array
Starting with commit dca5161f9bd0 ("hv_netvsc: Check status in
SEND_RNDIS_PKT completion message") in the 6.3 kernel, the Linux
driver for Hyper-V synthetic networking (netvsc) occasionally reports
"nvsp_rndis_pkt_complete error status: 2".[1] This error indicates
that Hyper-V has rejected a network packet transmit request from the
guest, and the outgoing network packet is dropped. Higher level
network protocols presumably recover and resend the packet so there is
no functional error, but performance is slightly impacted. Commit dca5161f9bd0 is not the cause of the error -- it only added reporting
of an error that was already happening without any notice. The error
has presumably been present since the netvsc driver was originally
introduced into Linux.
The root cause of the problem is that the netvsc driver in Linux may
send an incorrectly formatted VMBus message to Hyper-V when
transmitting the network packet. The incorrect formatting occurs when
the rndis header of the VMBus message crosses a page boundary due to
how the Linux skb head memory is aligned. In such a case, two PFNs are
required to describe the location of the rndis header, even though
they are contiguous in guest physical address (GPA) space. Hyper-V
requires that two rndis header PFNs be in a single "GPA range" data
struture, but current netvsc code puts each PFN in its own GPA range,
which Hyper-V rejects as an error.
The incorrect formatting occurs only for larger packets that netvsc
must transmit via a VMBus "GPA Direct" message. There's no problem
when netvsc transmits a smaller packet by copying it into a pre-
allocated send buffer slot because the pre-allocated slots don't have
page crossing issues.
After commit 14ad6ed30a10 ("net: allow small head cache usage with
large MAX_SKB_FRAGS values") in the 6.14-rc4 kernel, the error occurs
much more frequently in VMs with 16 or more vCPUs. It may occur every
few seconds, or even more frequently, in an ssh session that outputs a
lot of text. Commit 14ad6ed30a10 subtly changes how skb head memory is
allocated, making it much more likely that the rndis header will cross
a page boundary when the vCPU count is 16 or more. The changes in
commit 14ad6ed30a10 are perfectly valid -- they just had the side
effect of making the netvsc bug more prominent.
Current code in init_page_array() creates a separate page buffer array
entry for each PFN required to identify the data to be transmitted.
Contiguous PFNs get separate entries in the page buffer array, and any
information about contiguity is lost.
Fix the core issue by having init_page_array() construct the page
buffer array to represent contiguous ranges rather than individual
pages. When these ranges are subsequently passed to
netvsc_build_mpb_array(), it can build GPA ranges that contain
multiple PFNs, as required to avoid the error "nvsp_rndis_pkt_complete
error status: 2". If instead the network packet is sent by copying
into a pre-allocated send buffer slot, the copy proceeds using the
contiguous ranges rather than individual pages, but the result of the
copying is the same. Also fix rndis_filter_send_request() to construct
a contiguous range, since it has its own page buffer array.
This change has a side benefit in CoCo VMs in that netvsc_dma_map()
calls dma_map_single() on each contiguous range instead of on each
page. This results in fewer calls to dma_map_single() but on larger
chunks of memory, which should reduce contention on the swiotlb.
Since the page buffer array now contains one entry for each contiguous
range instead of for each individual page, the number of entries in
the array can be reduced, saving 208 bytes of stack space in
netvsc_xmit() when MAX_SKG_FRAGS has the default value of 17.
Michael Kelley [Tue, 13 May 2025 00:06:01 +0000 (17:06 -0700)]
hv_netvsc: Use vmbus_sendpacket_mpb_desc() to send VMBus messages
netvsc currently uses vmbus_sendpacket_pagebuffer() to send VMBus
messages. This function creates a series of GPA ranges, each of which
contains a single PFN. However, if the rndis header in the VMBus
message crosses a page boundary, the netvsc protocol with the host
requires that both PFNs for the rndis header must be in a single "GPA
range" data structure, which isn't possible with
vmbus_sendpacket_pagebuffer(). As the first step in fixing this, add a
new function netvsc_build_mpb_array() to build a VMBus message with
multiple GPA ranges, each of which may contain multiple PFNs. Use
vmbus_sendpacket_mpb_desc() to send this VMBus message to the host.
There's no functional change since higher levels of netvsc don't
maintain or propagate knowledge of contiguous PFNs. Based on its
input, netvsc_build_mpb_array() still produces a separate GPA range
for each PFN and the behavior is the same as with
vmbus_sendpacket_pagebuffer(). But the groundwork is laid for a
subsequent patch to provide the necessary grouping.
Michael Kelley [Tue, 13 May 2025 00:06:00 +0000 (17:06 -0700)]
Drivers: hv: Allow vmbus_sendpacket_mpb_desc() to create multiple ranges
vmbus_sendpacket_mpb_desc() is currently used only by the storvsc driver
and is hardcoded to create a single GPA range. To allow it to also be
used by the netvsc driver to create multiple GPA ranges, no longer
hardcode as having a single GPA range. Allow the calling driver to
specify the rangecount in the supplied descriptor.
Update the storvsc driver to reflect this new approach.
Linus Torvalds [Thu, 15 May 2025 02:33:18 +0000 (19:33 -0700)]
Merge tag 'tpmdd-next-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd
Pull tpm fixes from Jarkko Sakkinen:
"A few last minute fixes for v6.15"
* tag 'tpmdd-next-6.15-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd:
tpm: tis: Double the timeout B to 4s
char: tpm: tpm-buf: Add sanity check fallback in read helpers
tpm: Mask TPM RC in tpm2_start_auth_session()
Bo-Cun Chen [Tue, 13 May 2025 04:27:30 +0000 (05:27 +0100)]
net: ethernet: mtk_eth_soc: fix typo for declaration MT7988 ESW capability
Since MTK_ESW_BIT is a bit number rather than a bitmap, it causes
MTK_HAS_CAPS to produce incorrect results. This leads to the ETH
driver not declaring MAC capabilities correctly for the MT7988 ESW.
Fixes: 445eb6448ed3 ("net: ethernet: mtk_eth_soc: add basic support for MT7988 SoC") Signed-off-by: Bo-Cun Chen <bc-bocun.chen@mediatek.com> Signed-off-by: Daniel Golle <daniel@makrotopia.org> Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Link: https://patch.msgid.link/b8b37f409d1280fad9c4d32521e6207f63cd3213.1747110258.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jiawen Wu [Tue, 13 May 2025 02:10:09 +0000 (10:10 +0800)]
net: libwx: Fix FW mailbox unknown command
For the new SW-FW interaction, missing the error return if there is an
unknown command. It causes the driver to mistakenly believe that the
interaction is complete. This problem occurs when new driver is paired
with old firmware, which does not support the new mailbox commands.
Jiawen Wu [Tue, 13 May 2025 02:10:08 +0000 (10:10 +0800)]
net: libwx: Fix FW mailbox reply timeout
For the new SW-FW interaction, the timeout waiting for the firmware to
return is too short. So that some mailbox commands cannot be completed.
Use the 'timeout' parameter instead of fixed timeout value for flexible
configuration.
Jiawen Wu [Tue, 13 May 2025 02:10:07 +0000 (10:10 +0800)]
net: txgbe: Fix to calculate EEPROM checksum for AML devices
In the new firmware version, the shadow ram reserves some space to store
I2C information, so the checksum calculation needs to skip this section.
Otherwise, the driver will fail to probe because the invalid EEPROM
checksum.
octeontx2-pf: macsec: Fix incorrect max transmit size in TX secy
MASCEC hardware block has a field called maximum transmit size for
TX secy. Max packet size going out of MCS block has be programmed
taking into account full packet size which has L2 header,SecTag
and ICV. MACSEC offload driver is configuring max transmit size as
macsec interface MTU which is incorrect. Say with 1500 MTU of real
device, macsec interface created on top of real device will have MTU of
1468(1500 - (SecTag + ICV)). This is causing packets from macsec
interface of size greater than or equal to 1468 are not getting
transmitted out because driver programmed max transmit size as 1468
instead of 1514(1500 + ETH_HDR_LEN).
Jakub Kicinski [Tue, 13 May 2025 22:16:38 +0000 (15:16 -0700)]
netlink: specs: tc: all actions are indexed arrays
Some TC filters have actions listed as indexed arrays of nests
and some as just nests. They are all indexed arrays, the handling
is common across filters.
Jakub Kicinski [Tue, 13 May 2025 22:13:16 +0000 (15:13 -0700)]
netlink: specs: tc: fix a couple of attribute names
Fix up spelling of two attribute names. These are clearly typoes
and will prevent C codegen from working. Let's treat this as
a fix to get the correction into users' hands ASAP, and prevent
anyone depending on the wrong names.
Michal Suchanek [Fri, 4 Apr 2025 08:23:14 +0000 (10:23 +0200)]
tpm: tis: Double the timeout B to 4s
With some Infineon chips the timeouts in tpm_tis_send_data (both B and
C) can reach up to about 2250 ms.
Timeout C is retried since
commit de9e33df7762 ("tpm, tpm_tis: Workaround failed command reception on Infineon devices")
Timeout B still needs to be extended.
The problem is most commonly encountered with context related operation
such as load context/save context. These are issued directly by the
kernel, and there is no retry logic for them.
When a filesystem is set up to use the TPM for unlocking the boot fails,
and restarting the userspace service is ineffective. This is likely
because ignoring a load context/save context result puts the real TPM
state and the TPM state expected by the kernel out of sync.
Chips known to be affected:
tpm_tis IFX1522:00: 2.0 TPM (device-id 0x1D, rev-id 54)
Description: SLB9672
Firmware Revision: 15.22
char: tpm: tpm-buf: Add sanity check fallback in read helpers
Fix Smatch-detected issue:
drivers/char/tpm/tpm-buf.c:208 tpm_buf_read_u8() error:
uninitialized symbol 'value'.
drivers/char/tpm/tpm-buf.c:225 tpm_buf_read_u16() error:
uninitialized symbol 'value'.
drivers/char/tpm/tpm-buf.c:242 tpm_buf_read_u32() error:
uninitialized symbol 'value'.
Zero-initialize the return values in tpm_buf_read_u8(), tpm_buf_read_u16(),
and tpm_buf_read_u32() to guard against uninitialized data in case of a
boundary overflow.
Add defensive initialization ensures the return values are always defined,
preventing undefined behavior if the unexpected happens.
Signed-off-by: Purva Yeshi <purvayeshi550@gmail.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
projects / thirdparty / kernel / linux.git / log
