| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Pull NFS clent updates from Anna Schumaker:
"New Features:
- Implement the Sunrpc rfc2203 rpcsec_gss sequence number cache
- Add support for FALLOC_FL_ZERO_RANGE on NFS v4.2
- Add a localio sysfs attribute
Stable Fixes:
- Fix double-unlock bug in nfs_return_empty_folio()
- Don't check for OPEN feature support in v4.1
- Always probe for LOCALIO support asynchronously
- Prevent hang on NFS mounts with xprtsec=[m]tls
Other Bugfixes:
- xattr handlers should check for absent nfs filehandles
- Fix setattr caching of TIME_[MODIFY|ACCESS]_SET when timestamps are
delegated
- Fix listxattr to return selinux security labels
- Connect to NFSv3 DS using TLS if MDS connection uses TLS
- Clear SB_RDONLY before getting a superblock, and ignore when
remounting
- Fix incorrect handling of NFS error codes in nfs4_do_mkdir()
- Various nfs_localio fixes from Neil Brown that include fixing an
rcu compilation error found by older gcc versions.
- Update stats on flexfiles pNFS DSes when receiving NFS4ERR_DELAY
Cleanups:
- Add a refcount tracker for struct net in the nfs_client
- Allow FREE_STATEID to clean up delegations
- Always set NLINK even if the server doesn't support it
- Cleanups to the NFS folio writeback code
- Remove dead code from xs_tcp_tls_setup_socket()"
* tag 'nfs-for-6.16-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (30 commits)
flexfiles/pNFS: update stats on NFS4ERR_DELAY for v4.1 DSes
nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer
nfs_localio: protect race between nfs_uuid_put() and nfs_close_local_fh()
nfs_localio: duplicate nfs_close_local_fh()
nfs_localio: simplify interface to nfsd for getting nfsd_file
nfs_localio: always hold nfsd net ref with nfsd_file ref
nfs_localio: use cmpxchg() to install new nfs_file_localio
SUNRPC: Remove dead code from xs_tcp_tls_setup_socket()
SUNRPC: Prevent hang on NFS mount with xprtsec=[m]tls
nfs: fix incorrect handling of large-number NFS errors in nfs4_do_mkdir()
nfs: ignore SB_RDONLY when remounting nfs
nfs: clear SB_RDONLY before getting superblock
NFS: always probe for LOCALIO support asynchronously
pnfs/flexfiles: connect to NFSv3 DS using TLS if MDS connection uses TLS
NFS: add localio to sysfs
nfs: use writeback_iter directly
nfs: refactor nfs_do_writepage
nfs: don't return AOP_WRITEPAGE_ACTIVATE from nfs_do_writepage
nfs: fold nfs_page_async_flush into nfs_do_writepage
NFSv4: Always set NLINK even if the server doesn't support it
...
|
|
Pull netfs updates from Christian Brauner:
- The main API document has been extensively updated/rewritten
- Fix an oops in write-retry due to mis-resetting the I/O iterator
- Fix the recording of transferred bytes for short DIO reads
- Fix a request's work item to not require a reference, thereby
avoiding the need to get rid of it in BH/IRQ context
- Fix waiting and waking to be consistent about the waitqueue used
- Remove NETFS_SREQ_SEEK_DATA_READ, NETFS_INVALID_WRITE,
NETFS_ICTX_WRITETHROUGH, NETFS_READ_HOLE_CLEAR,
NETFS_RREQ_DONT_UNLOCK_FOLIOS, and NETFS_RREQ_BLOCKED
- Reorder structs to eliminate holes
- Remove netfs_io_request::ractl
- Only provide proc_link field if CONFIG_PROC_FS=y
- Remove folio_queue::marks3
- Fix undifferentiation of DIO reads from unbuffered reads
* tag 'vfs-6.16-rc1.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
netfs: Fix undifferentiation of DIO reads from unbuffered reads
netfs: Fix wait/wake to be consistent about the waitqueue used
netfs: Fix the request's work item to not require a ref
netfs: Fix setting of transferred bytes with short DIO reads
netfs: Fix oops in write-retry from mis-resetting the subreq iterator
fs/netfs: remove unused flag NETFS_RREQ_BLOCKED
fs/netfs: remove unused flag NETFS_RREQ_DONT_UNLOCK_FOLIOS
folio_queue: remove unused field `marks3`
fs/netfs: declare field `proc_link` only if CONFIG_PROC_FS=y
fs/netfs: remove `netfs_io_request.ractl`
fs/netfs: reorder struct fields to eliminate holes
fs/netfs: remove unused enum choice NETFS_READ_HOLE_CLEAR
fs/netfs: remove unused flag NETFS_ICTX_WRITETHROUGH
fs/netfs: remove unused source NETFS_INVALID_WRITE
fs/netfs: remove unused flag NETFS_SREQ_SEEK_DATA_READ
|
|
Pull automount updates from Al Viro:
"Automount wart removal
A bunch of odd boilerplate gone from instances - the reason for
those was the need to protect the yet-to-be-attched mount from
mark_mounts_for_expiry() deciding to take it out.
But that's easy to detect and take care of in mark_mounts_for_expiry()
itself; no need to have every instance simulate mount being busy by
grabbing an extra reference to it, with finish_automount() undoing
that once it attaches that mount.
Should've done it that way from the very beginning... This is a
flagday change, thankfully there are very few instances.
vfs_submount() is gone - its sole remaining user (trace_automount)
had been switched to saner primitives"
* tag 'pull-automount' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
kill vfs_submount()
saner calling conventions for ->d_automount()
|
|
On NFS4ERR_DELAY nfs slient updates its stats, but misses for
flexfiles v4.1 DSes.
Signed-off-by: Tigran Mkrtchyan <tigran.mkrtchyan@desy.de>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
Implementation follows bones of the pattern that was established in
commit a35518cae4b325 ("NFSv4.1/pnfs: fix NFS with TLS in pnfs").
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
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>
|
|
Stop using write_cache_pages and use writeback_iter directly. This
removes an indirect call per written folio and makes the code easier
to follow.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
Use early returns wherever possible to simplify the code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
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>
|
|
Fold nfs_page_async_flush into its only caller to clean up the code a
bit.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
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:
1. vfs_read, nfs_read_folio, ... nfs_read_add_folio,
nfs_return_empty_folio
2. vfs_read, nfs_read_folio, ... netfs_read_collection,
netfs_unlock_abandoned_read_pages
3. vfs_read, ... nfs_do_read_folio, nfs_read_add_folio,
nfs_return_empty_folio
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>
|
|
Pull misc vfs updates from Christian Brauner:
"This contains the usual selections of misc updates for this cycle.
Features:
- Use folios for symlinks in the page cache
FUSE already uses folios for its symlinks. Mirror that conversion
in the generic code and the NFS code. That lets us get rid of a few
folio->page->folio conversions in this path, and some of the few
remaining users of read_cache_page() / read_mapping_page()
- Try and make a few filesystem operations killable on the VFS
inode->i_mutex level
- Add sysctl vfs_cache_pressure_denom for bulk file operations
Some workloads need to preserve more dentries than we currently
allow through out sysctl interface
A HDFS servers with 12 HDDs per server, on a HDFS datanode startup
involves scanning all files and caching their metadata (including
dentries and inodes) in memory. Each HDD contains approximately 2
million files, resulting in a total of ~20 million cached dentries
after initialization
To minimize dentry reclamation, they set vfs_cache_pressure to 1.
Despite this configuration, memory pressure conditions can still
trigger reclamation of up to 50% of cached dentries, reducing the
cache from 20 million to approximately 10 million entries. During
the subsequent cache rebuild period, any HDFS datanode restart
operation incurs substantial latency penalties until full cache
recovery completes
To maintain service stability, more dentries need to be preserved
during memory reclamation. The current minimum reclaim ratio (1/100
of total dentries) remains too aggressive for such workload. This
patch introduces vfs_cache_pressure_denom for more granular cache
pressure control
The configuration [vfs_cache_pressure=1,
vfs_cache_pressure_denom=10000] effectively maintains the full 20
million dentry cache under memory pressure, preventing datanode
restart performance degradation
- Avoid some jumps in inode_permission() using likely()/unlikely()
- Avid a memory access which is most likely a cache miss when
descending into devcgroup_inode_permission()
- Add fastpath predicts for stat() and fdput()
- Anonymous inodes currently don't come with a proper mode causing
issues in the kernel when we want to add useful VFS debug assert.
Fix that by giving them a proper mode and masking it off when we
report it to userspace which relies on them not having any mode
- Anonymous inodes currently allow to change inode attributes because
the VFS falls back to simple_setattr() if i_op->setattr isn't
implemented. This means the ownership and mode for every single
user of anon_inode_inode can be changed. Block that as it's either
useless or actively harmful. If specific ownership is needed the
respective subsystem should allocate anonymous inodes from their
own private superblock
- Raise SB_I_NODEV and SB_I_NOEXEC on the anonymous inode superblock
- Add proper tests for anonymous inode behavior
- Make it easy to detect proper anonymous inodes and to ensure that
we can detect them in codepaths such as readahead()
Cleanups:
- Port pidfs to the new anon_inode_{g,s}etattr() helpers
- Try to remove the uselib() system call
- Add unlikely branch hint return path for poll
- Add unlikely branch hint on return path for core_sys_select
- Don't allow signals to interrupt getdents copying for fuse
- Provide a size hint to dir_context for during readdir()
- Use writeback_iter directly in mpage_writepages
- Update compression and mtime descriptions in initramfs
documentation
- Update main netfs API document
- Remove useless plus one in super_cache_scan()
- Remove unnecessary NULL-check guards during setns()
- Add separate separate {get,put}_cgroup_ns no-op cases
Fixes:
- Fix typo in root= kernel parameter description
- Use KERN_INFO for infof()|info_plog()|infofc()
- Correct comments of fs_validate_description()
- Mark an unlikely if condition with unlikely() in
vfs_parse_monolithic_sep()
- Delete macro fsparam_u32hex()
- Remove unused and problematic validate_constant_table()
- Fix potential unsigned integer underflow in fs_name()
- Make file-nr output the total allocated file handles"
* tag 'vfs-6.16-rc1.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (43 commits)
fs: Pass a folio to page_put_link()
nfs: Use a folio in nfs_get_link()
fs: Convert __page_get_link() to use a folio
fs/read_write: make default_llseek() killable
fs/open: make do_truncate() killable
fs/open: make chmod_common() and chown_common() killable
include/linux/fs.h: add inode_lock_killable()
readdir: supply dir_context.count as readdir buffer size hint
vfs: Add sysctl vfs_cache_pressure_denom for bulk file operations
fuse: don't allow signals to interrupt getdents copying
Documentation: fix typo in root= kernel parameter description
include/cgroup: separate {get,put}_cgroup_ns no-op case
kernel/nsproxy: remove unnecessary guards
fs: use writeback_iter directly in mpage_writepages
fs: remove useless plus one in super_cache_scan()
fs: add S_ANON_INODE
fs: remove uselib() system call
device_cgroup: avoid access to ->i_rdev in the common case in devcgroup_inode_permission()
fs/fs_parse: Remove unused and problematic validate_constant_table()
fs: touch up predicts in inode_permission()
...
|
|
Pull vfs directory lookup updates from Christian Brauner:
"This contains cleanups for the lookup_one*() family of helpers.
We expose a set of functions with names containing "lookup_one_len"
and others without the "_len". This difference has nothing to do with
"len". It's rater a historical accident that can be confusing.
The functions without "_len" take a "mnt_idmap" pointer. This is found
in the "vfsmount" and that is an important question when choosing
which to use: do you have a vfsmount, or are you "inside" the
filesystem. A related question is "is permission checking relevant
here?".
nfsd and cachefiles *do* have a vfsmount but *don't* use the non-_len
functions. They pass nop_mnt_idmap and refuse to work on filesystems
which have any other idmap.
This work changes nfsd and cachefile to use the lookup_one family of
functions and to explictily pass &nop_mnt_idmap which is consistent
with all other vfs interfaces used where &nop_mnt_idmap is explicitly
passed.
The remaining uses of the "_one" functions do not require permission
checks so these are renamed to be "_noperm" and the permission
checking is removed.
This series also changes these lookup function to take a qstr instead
of separate name and len. In many cases this simplifies the call"
* tag 'vfs-6.16-rc1.async.dir' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
VFS: change lookup_one_common and lookup_noperm_common to take a qstr
Use try_lookup_noperm() instead of d_hash_and_lookup() outside of VFS
VFS: rename lookup_one_len family to lookup_noperm and remove permission check
cachefiles: Use lookup_one() rather than lookup_one_len()
nfsd: Use lookup_one() rather than lookup_one_len()
VFS: improve interface for lookup_one functions
|
|
On cifs, "DIO reads" (specified by O_DIRECT) need to be differentiated from
"unbuffered reads" (specified by cache=none in the mount parameters). The
difference is flagged in the protocol and the server may behave
differently: Windows Server will, for example, mandate that DIO reads are
block aligned.
Fix this by adding a NETFS_UNBUFFERED_READ to differentiate this from
NETFS_DIO_READ, parallelling the write differentiation that already exists.
cifs will then do the right thing.
Fixes: 016dc8516aec ("netfs: Implement unbuffered/DIO read support")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/3444961.1747987072@warthog.procyon.org.uk
Reviewed-by: "Paulo Alcantara (Red Hat)" <pc@manguebit.com>
Reviewed-by: Viacheslav Dubeyko <Slava.Dubeyko@ibm.com>
cc: Steve French <sfrench@samba.org>
cc: netfs@lists.linux.dev
cc: v9fs@lists.linux.dev
cc: linux-afs@lists.infradead.org
cc: linux-cifs@vger.kernel.org
cc: ceph-devel@vger.kernel.org
cc: linux-nfs@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
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.
For example:
$ strace -e trace=getxattr getfattr -n system.nfs4_acl /mnt/t/ref
getxattr("/mnt/t/ref", "system.nfs4_acl", NULL, 0) = -1 EIO (Input/output error)
/mnt/t/ref: system.nfs4_acl: Input/output error
+++ exited with 1 +++
Have the xattr handlers return -ENODATA instead.
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
These are long-held references to the netns, so make sure the refcount
tracker is aware of them.
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Anna Schumaker <anna.schumaker@oracle.com>
|
|
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>
|
|
If there isn't a valid layout, or the layout stateid has changed, the
cleanup after a layout return should clear out the old data.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
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>
|
|
All callers now have a folio. Pass it to page_put_link(), saving a
hidden call to compound_head(). Also add kernel-doc for page_get_link()
and page_put_link().
Signed-off-by: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Link: https://lore.kernel.org/20250514171316.3002934-4-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
Mirror the changes to __page_get_link() by retrieving a folio from
the page cache instead of a page. Removes two hidden calls to
compound_head().
Signed-off-by: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Link: https://lore.kernel.org/20250514171316.3002934-3-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
Currently the calling conventions for ->d_automount() instances have
an odd wart - returned new mount to be attached is expected to have
refcount 2.
That kludge is intended to make sure that mark_mounts_for_expiry() called
before we get around to attaching that new mount to the tree won't decide
to take it out. finish_automount() drops the extra reference after it's
done with attaching mount to the tree - or drops the reference twice in
case of error. ->d_automount() instances have rather counterintuitive
boilerplate in them.
There's a much simpler approach: have mark_mounts_for_expiry() skip the
mounts that are yet to be mounted. And to hell with grabbing/dropping
those extra references. Makes for simpler correctness analysis, at that...
Reviewed-by: Christian Brauner <brauner@kernel.org>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Acked-by: David Howells <dhowells@redhat.com>
Tested-by: David Howells <dhowells@redhat.com>
Acked-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
Once the clp->cl_uuid.lock has been dropped, another CPU could come in
and free the struct nfsd_file that was just added. To prevent that from
happening, take the RCU read lock before dropping the spin lock.
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Reviewed-by: Mike Snitzer <snitzer@kernel.org>
|
|
In nfs3_get_acl(), the local variable status is assigned the result of
nfs_refresh_inode() inside the *switch* statement, but that value gets
overwritten in the next *if* statement's true branch and is completely
ignored if that branch isn't taken...
Found by Linux Verification Center (linuxtesting.org) with the Svace static
analysis tool.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Reviewed-by: Benjamin Coddington <bcodding@redhat.com>
Link: https://lore.kernel.org/r/c32dced7-a4fa-43c0-aafe-ef6c819c2f91@omp.ru
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
In nfs_direct_write_completion(), the local variable req isn't used outside
the *while* loop and is assigned to right at the start of that loop's body,
so its initializer appears useless -- drop it; then move the declaration to
the loop body (which happens to have a pointless empty line anyway)...
Found by Linux Verification Center (linuxtesting.org) with the Svace static
analysis tool.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Reviewed-by: Benjamin Coddington <bcodding@redhat.com>
Link: https://lore.kernel.org/r/416219f5-7983-484b-b5a7-5fb7da9561f7@omp.ru
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
Since struct nfs4_pnfs_ds should not be shared between net namespaces,
move from a global list of objects to a per-netns list and spinlock.
Tested-by: Sargun Dillon <sargun@sargun.me>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Benjamin Coddington <bcodding@redhat.com>
Link: https://lore.kernel.org/r/20250410-nfs-ds-netns-v2-2-f80b7979ba80@kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
Currently, different NFS clients can share the same DS connections, even
when they are in different net namespaces. If a containerized client
creates a DS connection, another container can find and use it. When the
first client exits, the connection will close which can lead to stalls
in other clients.
Add a net namespace pointer to struct nfs4_pnfs_ds, and compare those
value to the caller's netns in _data_server_lookup_locked() when
searching for a nfs4_pnfs_ds to match.
Reported-by: Omar Sandoval <osandov@osandov.com>
Reported-by: Sargun Dillon <sargun@sargun.me>
Closes: https://lore.kernel.org/linux-nfs/Z_ArpQC_vREh_hEA@telecaster/
Tested-by: Sargun Dillon <sargun@sargun.me>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Benjamin Coddington <bcodding@redhat.com>
Link: https://lore.kernel.org/r/20250410-nfs-ds-netns-v2-1-f80b7979ba80@kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
When memory is insufficient, the allocation of nfs_lock_context in
nfs_get_lock_context() fails and returns -ENOMEM. If we mistakenly treat
an nfs4_unlockdata structure (whose l_ctx member has been set to -ENOMEM)
as valid and proceed to execute rpc_run_task(), this will trigger a NULL
pointer dereference in nfs4_locku_prepare. For example:
BUG: kernel NULL pointer dereference, address: 000000000000000c
PGD 0 P4D 0
Oops: Oops: 0000 [#1] SMP PTI
CPU: 15 UID: 0 PID: 12 Comm: kworker/u64:0 Not tainted 6.15.0-rc2-dirty #60
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-2.fc40
Workqueue: rpciod rpc_async_schedule
RIP: 0010:nfs4_locku_prepare+0x35/0xc2
Code: 89 f2 48 89 fd 48 c7 c7 68 69 ef b5 53 48 8b 8e 90 00 00 00 48 89 f3
RSP: 0018:ffffbbafc006bdb8 EFLAGS: 00010246
RAX: 000000000000004b RBX: ffff9b964fc1fa00 RCX: 0000000000000000
RDX: 0000000000000000 RSI: fffffffffffffff4 RDI: ffff9ba53fddbf40
RBP: ffff9ba539934000 R08: 0000000000000000 R09: ffffbbafc006bc38
R10: ffffffffb6b689c8 R11: 0000000000000003 R12: ffff9ba539934030
R13: 0000000000000001 R14: 0000000004248060 R15: ffffffffb56d1c30
FS: 0000000000000000(0000) GS:ffff9ba5881f0000(0000) knlGS:00000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000000000c CR3: 000000093f244000 CR4: 00000000000006f0
Call Trace:
<TASK>
__rpc_execute+0xbc/0x480
rpc_async_schedule+0x2f/0x40
process_one_work+0x232/0x5d0
worker_thread+0x1da/0x3d0
? __pfx_worker_thread+0x10/0x10
kthread+0x10d/0x240
? __pfx_kthread+0x10/0x10
ret_from_fork+0x34/0x50
? __pfx_kthread+0x10/0x10
ret_from_fork_asm+0x1a/0x30
</TASK>
Modules linked in:
CR2: 000000000000000c
---[ end trace 0000000000000000 ]---
Free the allocated nfs4_unlockdata when nfs_get_lock_context() fails and
return NULL to terminate subsequent rpc_run_task, preventing NULL pointer
dereference.
Fixes: f30cb757f680 ("NFS: Always wait for I/O completion before unlock")
Signed-off-by: Li Lingfeng <lilingfeng3@huawei.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Link: https://lore.kernel.org/r/20250417072508.3850532-1-lilingfeng3@huawei.com
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
When debugging I/O issues, we want to see not just the NFS level errors,
but also the RPC level problems, so record both in the tracepoints.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
|
If we have a fatal ENETDOWN or ENETUNREACH error, then the layoutreturn
on close code should also handle that as fatal, and free the layouts.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
|
|
Ensure that the NFSv4 error handling code recognises the
RPC_TASK_NETUNREACH_FATAL flag, and handles the ENETDOWN and ENETUNREACH
errors accordingly.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
|
|
Pull nfsd fixes from Chuck Lever:
- v6.15 libcrc clean-up makes invalid configurations possible
- Fix a potential deadlock introduced during the v6.15 merge window
* tag 'nfsd-6.15-1' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
nfsd: decrease sc_count directly if fail to queue dl_recall
nfs: add missing selections of CONFIG_CRC32
|
|
nfs.ko, nfsd.ko, and lockd.ko all use crc32_le(), which is available
only when CONFIG_CRC32 is enabled. But the only NFS kconfig option that
selected CONFIG_CRC32 was CONFIG_NFS_DEBUG, which is client-specific and
did not actually guard the use of crc32_le() even on the client.
The code worked around this bug by only actually calling crc32_le() when
CONFIG_CRC32 is built-in, instead hard-coding '0' in other cases. This
avoided randconfig build errors, and in real kernels the fallback code
was unlikely to be reached since CONFIG_CRC32 is 'default y'. But, this
really needs to just be done properly, especially now that I'm planning
to update CONFIG_CRC32 to not be 'default y'.
Therefore, make CONFIG_NFS_FS, CONFIG_NFSD, and CONFIG_LOCKD select
CONFIG_CRC32. Then remove the fallback code that becomes unnecessary,
as well as the selection of CONFIG_CRC32 from CONFIG_NFS_DEBUG.
Fixes: 1264a2f053a3 ("NFS: refactor code for calculating the crc32 hash of a filehandle")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Anna Schumaker <anna.schumaker@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
|
|
The lookup_one_len family of functions is (now) only used internally by
a filesystem on itself either
- in a context where permission checking is irrelevant such as by a
virtual filesystem populating itself, or xfs accessing its ORPHANAGE
or dquota accessing the quota file; or
- in a context where a permission check (MAY_EXEC on the parent) has just
been performed such as a network filesystem finding in "silly-rename"
file in the same directory. This is also the context after the
_parentat() functions where currently lookup_one_qstr_excl() is used.
So the permission check is pointless.
The name "one_len" is unhelpful in understanding the purpose of these
functions and should be changed. Most of the callers pass the len as
"strlen()" so using a qstr and QSTR() can simplify the code.
This patch renames these functions (include lookup_positive_unlocked()
which is part of the family despite the name) to have a name based on
"lookup_noperm". They are changed to receive a 'struct qstr' instead
of separate name and len. In a few cases the use of QSTR() results in a
new call to strlen().
try_lookup_noperm() takes a pointer to a qstr instead of the whole
qstr. This is consistent with d_hash_and_lookup() (which is nearly
identical) and useful for lookup_noperm_unlocked().
The new lookup_noperm_common() doesn't take a qstr yet. That will be
tidied up in a subsequent patch.
Signed-off-by: NeilBrown <neil@brown.name>
Link: https://lore.kernel.org/r/20250319031545.2999807-5-neil@brown.name
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
Pull NFS client updates from Trond Myklebust:
"Bugfixes:
- Three fixes for looping in the NFSv4 state manager delegation code
- Fix for the NFSv4 state XDR code (Neil Brown)
- Fix a leaked reference in nfs_lock_and_join_requests()
- Fix a use-after-free in the delegation return code
Features:
- Implement the NFSv4.2 copy offload OFFLOAD_STATUS operation to
allow monitoring of an in-progress copy
- Add a mount option to force NFSv3/NFSv4 to use READDIRPLUS in a
getdents() call
- SUNRPC now allows some basic management of an existing RPC client's
connections using sysfs
- Improvements to the automated teardown of a NFS client when the
container it was initiated from gets killed
- Improvements to prevent tasks from getting stuck in a killable wait
state after calling exit_signals()"
* tag 'nfs-for-6.15-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (29 commits)
nfs: Add missing release on error in nfs_lock_and_join_requests()
NFSv4: Check for delegation validity in nfs_start_delegation_return_locked()
NFS: Don't allow waiting for exiting tasks
SUNRPC: Don't allow waiting for exiting tasks
NFSv4: Treat ENETUNREACH errors as fatal for state recovery
NFSv4: clp->cl_cons_state < 0 signifies an invalid nfs_client
NFSv4: Further cleanups to shutdown loops
NFS: Shut down the nfs_client only after all the superblocks
SUNRPC: rpc_clnt_set_transport() must not change the autobind setting
SUNRPC: rpcbind should never reset the port to the value '0'
pNFS/flexfiles: Report ENETDOWN as a connection error
pNFS/flexfiles: Treat ENETUNREACH errors as fatal in containers
NFS: Treat ENETUNREACH errors as fatal in containers
NFS: Add a mount option to make ENETUNREACH errors fatal
sunrpc: Add a sysfs file for one-step xprt deletion
sunrpc: Add a sysfs file for adding a new xprt
sunrpc: Add a sysfs files for rpc_clnt information
sunrpc: Add a sysfs attr for xprtsec
NFS: Add implid to sysfs
NFS: Extend rdirplus mount option with "force|none"
...
|
|
Call nfs_release_request() on this error path before returning.
Fixes: c3f2235782c3 ("nfs: fold nfs_folio_find_and_lock_request into nfs_lock_and_join_requests")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://lore.kernel.org/r/3aaaa3d5-1c8a-41e4-98c7-717801ddd171@stanley.mountain
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
|
Linus Torvalds
Tigran Mkrtchyan
NeilBrown
Li Lingfeng
Mike Snitzer
Christoph Hellwig
Han Young
Benjamin Coddington
Scott Mayhew
Olga Kornievskaia
Sagi Grimberg
Anna Schumaker
Max Kellermann
David Howells
Jeff Layton
Trond Myklebust
Matthew Wilcox (Oracle)
Al Viro
Sergey Shtylyov
Eric Biggers
Dan Carpenter