Due to the code that queues a subreq on the active subrequest list getting
moved to netfs_issue_read(), the NETFS_RREQ_ALL_QUEUED flag may now get set
before the list-add actually happens. This is not a problem if the
collection worker happens after the list-add, but it's a race - and, for
9P, where the read from the server is synchronous and done in the
submitting thread, this is a lot more likely.
The result is that, if the timing is wrong, a ref gets leaked because the
collector thinks that all the subreqs have completed (because it can't see
the last one yet) and clears NETFS_RREQ_IN_PROGRESS - at which point, the
collection worker no longer goes into the collector.
This can be provoked with AFS by injecting an msleep() right before the
final subreq is queued.
Fix this by splitting the queuing part out of netfs_issue_read() into a new
function, netfs_queue_read(), and calling it separately. The setting of
NETFS_RREQ_ALL_QUEUED is then done by netfs_queue_read() whilst it is
holding the spinlock (that's probably unnecessary, but shouldn't hurt).
It might be better to set a flag on the final subreq, but this could be a
problem if an error occurs and we can't queue it.
Fixes: e2d46f2ec3 ("netfs: Change the read result collector to only use one work item")
Reported-by: Ihor Solodrai <ihor.solodrai@pm.me>
Closes: https://lore.kernel.org/r/a7x33d4dnMdGTtRivptq6S1i8btK70SNBP2XyX_xwDAhLvgQoPox6FVBOkifq4eBinfFfbZlIkMZBe3QarlWTxoEtHZwJCZbNKtaqrR7PvI=@pm.me/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20250212222402.3618494-4-dhowells@redhat.com
Tested-by: Ihor Solodrai <ihor.solodrai@linux.dev>
cc: Eric Van Hensbergen <ericvh@kernel.org>
cc: Latchesar Ionkov <lucho@ionkov.net>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Christian Schoenebeck <linux_oss@crudebyte.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Steve French <stfrench@microsoft.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: v9fs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Fix a number of hangs in the netfslib read-retry code, including:
(1) netfs_reissue_read() doubles up the getting of references on
subrequests, thereby leaking the subrequest and causing inode eviction
to wait indefinitely. This can lead to the kernel reporting a hang in
the filesystem's evict_inode().
Fix this by removing the get from netfs_reissue_read() and adding one
to netfs_retry_read_subrequests() to deal with the one place that
didn't double up.
(2) The loop in netfs_retry_read_subrequests() that retries a sequence of
failed subrequests doesn't record whether or not it retried the one
that the "subreq" pointer points to when it leaves the loop. It may
not if renegotiation/repreparation of the subrequests means that fewer
subrequests are needed to span the cumulative range of the sequence.
Because it doesn't record this, the piece of code that discards
now-superfluous subrequests doesn't know whether it should discard the
one "subreq" points to - and so it doesn't.
Fix this by noting whether the last subreq it examines is superfluous
and if it is, then getting rid of it and all subsequent subrequests.
If that one one wasn't superfluous, then we would have tried to go
round the previous loop again and so there can be no further unretried
subrequests in the sequence.
(3) netfs_retry_read_subrequests() gets yet an extra ref on any additional
subrequests it has to get because it ran out of ones it could reuse to
to renegotiation/repreparation shrinking the subrequests.
Fix this by removing that extra ref.
(4) In netfs_retry_reads(), it was using wait_on_bit() to wait for
NETFS_SREQ_IN_PROGRESS to be cleared on all subrequests in the
sequence - but netfs_read_subreq_terminated() is now using a wait
queue on the request instead and so this wait will never finish.
Fix this by waiting on the wait queue instead. To make this work, a
new flag, NETFS_RREQ_RETRYING, is now set around the wait loop to tell
the wake-up code to wake up the wait queue rather than requeuing the
request's work item.
Note that this flag replaces the NETFS_RREQ_NEED_RETRY flag which is
no longer used.
(5) Whilst not strictly anything to do with the hang,
netfs_retry_read_subrequests() was also doubly incrementing the
subreq_counter and re-setting the debug index, leaving a gap in the
trace. This is also fixed.
One of these hangs was observed with 9p and with cifs. Others were forced
by manual code injection into fs/afs/file.c. Firstly, afs_prepare_read()
was created to provide an changing pattern of maximum subrequest sizes:
static int afs_prepare_read(struct netfs_io_subrequest *subreq)
{
struct netfs_io_request *rreq = subreq->rreq;
if (!S_ISREG(subreq->rreq->inode->i_mode))
return 0;
if (subreq->retry_count < 20)
rreq->io_streams[0].sreq_max_len =
umax(200, 2222 - subreq->retry_count * 40);
else
rreq->io_streams[0].sreq_max_len = 3333;
return 0;
}
and pointed to by afs_req_ops. Then the following:
struct netfs_io_subrequest *subreq = op->fetch.subreq;
if (subreq->error == 0 &&
S_ISREG(subreq->rreq->inode->i_mode) &&
subreq->retry_count < 20) {
subreq->transferred = subreq->already_done;
__clear_bit(NETFS_SREQ_HIT_EOF, &subreq->flags);
__set_bit(NETFS_SREQ_NEED_RETRY, &subreq->flags);
afs_fetch_data_notify(op);
return;
}
was inserted into afs_fetch_data_success() at the beginning and struct
netfs_io_subrequest given an extra field, "already_done" that was set to
the value in "subreq->transferred" by netfs_reissue_read().
When reading a 4K file, the subrequests would get gradually smaller, a new
subrequest would be allocated around the 3rd retry and then eventually be
rendered superfluous when the 20th retry was hit and the limit on the first
subrequest was eased.
Fixes: e2d46f2ec3 ("netfs: Change the read result collector to only use one work item")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20250212222402.3618494-2-dhowells@redhat.com
Tested-by: Marc Dionne <marc.dionne@auristor.com>
Tested-by: Steve French <stfrench@microsoft.com>
cc: Ihor Solodrai <ihor.solodrai@pm.me>
cc: Eric Van Hensbergen <ericvh@kernel.org>
cc: Latchesar Ionkov <lucho@ionkov.net>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Christian Schoenebeck <linux_oss@crudebyte.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: v9fs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZ4pRKQAKCRCRxhvAZXjc
ov2dAQCULWjTBWdF8Ro2bfNeXzWvUUnSPjoLJ9B4xlrOB9c2MAEAiwkKHkzAxUco
hCvaRJc3H2ze2wrgbIABPKB2noQVVwk=
=4ojv
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.14-rc1.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull vfs netfs updates from Christian Brauner:
"This contains read performance improvements and support for monolithic
single-blob objects that have to be read/written as such (e.g. AFS
directory contents). The implementation of the two parts is interwoven
as each makes the other possible.
- Read performance improvements
The read performance improvements are intended to speed up some
loss of performance detected in cifs and to a lesser extend in afs.
The problem is that we queue too many work items during the
collection of read results: each individual subrequest is collected
by its own work item, and then they have to interact with each
other when a series of subrequests don't exactly align with the
pattern of folios that are being read by the overall request.
Whilst the processing of the pages covered by individual
subrequests as they complete potentially allows folios to be woken
in parallel and with minimum delay, it can shuffle wakeups for
sequential reads out of order - and that is the most common I/O
pattern.
The final assessment and cleanup of an operation is then held up
until the last I/O completes - and for a synchronous sequential
operation, this means the bouncing around of work items just adds
latency.
Two changes have been made to make this work:
(1) All collection is now done in a single "work item" that works
progressively through the subrequests as they complete (and
also dispatches retries as necessary).
(2) For readahead and AIO, this work item be done on a workqueue
and can run in parallel with the ultimate consumer of the data;
for synchronous direct or unbuffered reads, the collection is
run in the application thread and not offloaded.
Functions such as smb2_readv_callback() then just tell netfslib
that the subrequest has terminated; netfslib does a minimal bit of
processing on the spot - stat counting and tracing mostly - and
then queues/wakes up the worker. This simplifies the logic as the
collector just walks sequentially through the subrequests as they
complete and walks through the folios, if buffered, unlocking them
as it goes. It also keeps to a minimum the amount of latency
injected into the filesystem's low-level I/O handling
The way netfs supports filesystems using the deprecated
PG_private_2 flag is changed: folios are flagged and added to a
write request as they complete and that takes care of scheduling
the writes to the cache. The originating read request can then just
unlock the pages whatever happens.
- Single-blob object support
Single-blob objects are files for which the content of the file
must be read from or written to the server in a single operation
because reading them in parts may yield inconsistent results. AFS
directories are an example of this as there exists the possibility
that the contents are generated on the fly and would differ between
reads or might change due to third party interference.
Such objects will be written to and retrieved from the cache if one
is present, though we allow/may need to propose multiple
subrequests to do so. The important part is that read from/write to
the *server* is monolithic.
Single blob reading is, for the moment, fully synchronous and does
result collection in the application thread and, also for the
moment, the API is supplied the buffer in the form of a folio_queue
chain rather than using the pagecache.
- Related afs changes
This series makes a number of changes to the kafs filesystem,
primarily in the area of directory handling:
- AFS's FetchData RPC reply processing is made partially
asynchronous which allows the netfs_io_request's outstanding
operation counter to be removed as part of reducing the
collection to a single work item.
- Directory and symlink reading are plumbed through netfslib using
the single-blob object API and are now cacheable with fscache.
This also allows the afs_read struct to be eliminated and
netfs_io_subrequest to be used directly instead.
- Directory and symlink content are now stored in a folio_queue
buffer rather than in the pagecache. This means we don't require
the RCU read lock and xarray iteration to access it, and folios
won't randomly disappear under us because the VM wants them
back.
- The vnode operation lock is changed from a mutex struct to a
private lock implementation. The problem is that the lock now
needs to be dropped in a separate thread and mutexes don't
permit that.
- When a new directory or symlink is created, we now initialise it
locally and mark it valid rather than downloading it (we know
what it's likely to look like).
- We now use the in-directory hashtable to reduce the number of
entries we need to scan when doing a lookup. The edit routines
have to maintain the hash chains.
- Cancellation (e.g. by signal) of an async call after the
rxrpc_call has been set up is now offloaded to the worker thread
as there will be a notification from rxrpc upon completion. This
avoids a double cleanup.
- A "rolling buffer" implementation is created to abstract out the
two separate folio_queue chaining implementations I had (one for
read and one for write).
- Functions are provided to create/extend a buffer in a folio_queue
chain and tear it down again.
This is used to handle AFS directories, but could also be used to
create bounce buffers for content crypto and transport crypto.
- The was_async argument is dropped from netfs_read_subreq_terminated()
Instead we wake the read collection work item by either queuing it
or waking up the app thread.
- We don't need to use BH-excluding locks when communicating between
the issuing thread and the collection thread as neither of them now
run in BH context.
- Also included are a number of new tracepoints; a split of the
netfslib write collection code to put retrying into its own file
(it gets more complicated with content encryption).
- There are also some minor fixes AFS included, including fixing the
AFS directory format struct layout, reducing some directory
over-invalidation and making afs_mkdir() translate EEXIST to
ENOTEMPY (which is not available on all systems the servers
support).
- Finally, there's a patch to try and detect entry into the folio
unlock function with no folio_queue structs in the buffer (which
isn't allowed in the cases that can get there).
This is a debugging patch, but should be minimal overhead"
* tag 'vfs-6.14-rc1.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (31 commits)
netfs: Report on NULL folioq in netfs_writeback_unlock_folios()
afs: Add a tracepoint for afs_read_receive()
afs: Locally initialise the contents of a new symlink on creation
afs: Use the contained hashtable to search a directory
afs: Make afs_mkdir() locally initialise a new directory's content
netfs: Change the read result collector to only use one work item
afs: Make {Y,}FS.FetchData an asynchronous operation
afs: Fix cleanup of immediately failed async calls
afs: Eliminate afs_read
afs: Use netfslib for symlinks, allowing them to be cached
afs: Use netfslib for directories
afs: Make afs_init_request() get a key if not given a file
netfs: Add support for caching single monolithic objects such as AFS dirs
netfs: Add functions to build/clean a buffer in a folio_queue
afs: Add more tracepoints to do with tracking validity
cachefiles: Add auxiliary data trace
cachefiles: Add some subrequest tracepoints
netfs: Remove some extraneous directory invalidations
afs: Fix directory format encoding struct
afs: Fix EEXIST error returned from afs_rmdir() to be ENOTEMPTY
...
Fix netfslib's read-retry to only call ->prepare_read() in the backing
filesystem such a function is provided. We can get to this point if a
there's an active cache as failed reads from the cache need negotiating
with the server instead.
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/529329.1736261010@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Netfslib needs to be able to handle kernel-initiated asynchronous DIO that
is supplied with a bio_vec[] array. Currently, because of the async flag,
this gets passed to netfs_extract_user_iter() which throws a warning and
fails because it only handles IOVEC and UBUF iterators. This can be
triggered through a combination of cifs and a loopback blockdev with
something like:
mount //my/cifs/share /foo
dd if=/dev/zero of=/foo/m0 bs=4K count=1K
losetup --sector-size 4096 --direct-io=on /dev/loop2046 /foo/m0
echo hello >/dev/loop2046
This causes the following to appear in syslog:
WARNING: CPU: 2 PID: 109 at fs/netfs/iterator.c:50 netfs_extract_user_iter+0x170/0x250 [netfs]
and the write to fail.
Fix this by removing the check in netfs_unbuffered_write_iter_locked() that
causes async kernel DIO writes to be handled as userspace writes. Note
that this change relies on the kernel caller maintaining the existence of
the bio_vec array (or kvec[] or folio_queue) until the op is complete.
Fixes: 153a9961b5 ("netfs: Implement unbuffered/DIO write support")
Reported-by: Nicolas Baranger <nicolas.baranger@3xo.fr>
Closes: https://lore.kernel.org/r/fedd8a40d54b2969097ffa4507979858@3xo.fr/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/608725.1736275167@warthog.procyon.org.uk
Tested-by: Nicolas Baranger <nicolas.baranger@3xo.fr>
Acked-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
cc: Steve French <smfrench@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
It seems that it's possible to get to netfs_writeback_unlock_folios() with
an empty rolling buffer during buffered writes. This should not be
possible as the rolling buffer is initialised as the write request is set
up and thereafter maintains at least one folio_queue struct therein until
it gets destroyed. This allows lockless addition and removal of
folio_queue structs in the buffer because, unlike with a ring buffer, the
producer and consumer each only need to look at and alter one pointer into
the buffer.
Now, the rolling buffer is only used for buffered I/O operations as
netfs_collect_write_results() should only call
netfs_writeback_unlock_folios() if the request is of origin type
NETFS_WRITEBACK, NETFS_WRITETHROUGH or NETFS_PGPRIV2_COPY_TO_CACHE.
So it would seem that one of the following occurred: (1) I/O started before
the request was fully initialised, (2) the origin got switched mid-flow or
(3) the request has already been freed and this is a UAF error. I think
the last is the most likely.
Make netfs_writeback_unlock_folios() report information about the request
and subrequests if folioq is seen to be NULL to try and help debug this,
throw a warning and return.
Note that this does not try to fix the problem.
Reported-by: syzbot+af5c06208fa71bf31b16@syzkaller.appspotmail.com
Link: https://syzkaller.appspot.com/bug?extid=af5c06208fa71bf31b16
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/ZxshMEW4U7MTgQYa@gmail.com/
Link: https://lore.kernel.org/r/20241216204124.3752367-33-dhowells@redhat.com
cc: Chang Yu <marcus.yu.56@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Since we know what the contents of a symlink will be when we create it on
the server, initialise its contents locally too to avoid the need to
download it.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-31-dhowells@redhat.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Change the way netfslib collects read results to do all the collection for
a particular read request using a single work item that walks along the
subrequest queue as subrequests make progress or complete, unlocking folios
progressively rather than doing the unlock in parallel as parallel requests
come in.
The code is remodelled to be more like the write-side code, though only
using a single stream. This makes it more directly comparable and thus
easier to duplicate fixes between the two sides.
This has a number of advantages:
(1) It's simpler. There doesn't need to be a complex donation mechanism
to handle mismatches between the size and alignment of subrequests and
folios. The collector unlocks folios as the subrequests covering each
complete.
(2) It should cause less scheduler overhead as there's a single work item
in play unlocking pages in parallel when a read gets split up into a
lot of subrequests instead of one per subrequest.
Whilst the parallellism is nice in theory, in practice, the vast
majority of loads are sequential reads of the whole file, so
committing a bunch of threads to unlocking folios out of order doesn't
help in those cases.
(3) It should make it easier to implement content decryption. A folio
cannot be decrypted until all the requests that contribute to it have
completed - and, again, most loads are sequential and so, most of the
time, we want to begin decryption sequentially (though it's great if
the decryption can happen in parallel).
There is a disadvantage in that we're losing the ability to decrypt and
unlock things on an as-things-arrive basis which may affect some
applications.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-28-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Now that directory and symlink reads go through netfslib, the afs_read
struct is mostly redundant with almost all data duplicated in the
netfs_io_request and netfs_io_subrequest structs that are also available
any time we're doing a fetch.
Eliminate afs_read by moving the one field we still need there to the
afs_call struct (we may be given a different amount of data than what we
asked for and have to track what remains of that) and using the
netfs_io_subrequest directly instead.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-24-dhowells@redhat.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Add two netfslib functions to build up or clean up a buffer in a
folio_queue. The first, netfs_alloc_folioq_buffer() will add folios to a
buffer, extending up at least to the given size. If it can, it will add
multipage folios. The folios are optionally have the mapping set and will
have the index set according to the distance from the front of the folio
queue.
The second function will free up a folio queue and put any folios in the
queue that have the first mark set.
The netfs_folio tracepoint is also altered to cope with folios that have a
NULL mapping, and the folios being added/put will have trace lines emitted
and will be accounted in the stats.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-19-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: netfs@lists.linux.dev
cc: linux-afs@lists.infradead.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Drop the was_async argument from netfs_read_subreq_terminated(). Almost
every caller is either in process context and passes false. Some
filesystems delegate the call to a workqueue to avoid doing the work in
their network message queue parsing thread.
The only exception is netfs_cache_read_terminated() which handles
completion in the cache - which is usually a callback from the backing
filesystem in softirq context, though it can be from process context if an
error occurred. In this case, delegate to a workqueue.
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/CAHk-=wiVC5Cgyz6QKXFu6fTaA6h4CjexDR-OV9kL6Vo5x9v8=A@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-10-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
A rolling buffer is a series of folios held in a list of folio_queues. New
folios and folio_queue structs may be inserted at the head simultaneously
with spent ones being removed from the tail without the need for locking.
The rolling buffer includes an iov_iter and it has to be careful managing
this as the list of folio_queues is extended such that an oops doesn't
incurred because the iterator was pointing to the end of a folio_queue
segment that got appended to and then removed.
We need to use the mechanism twice, once for read and once for write, and,
in future patches, we will use a second rolling buffer to handle bounce
buffering for content encryption.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-6-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Add a tracepoint to log the lifespan of folio_queue structs. For tracing
illustrative purposes, folio_queues are tagged with the debug ID of
whatever they're related to (typically a netfs_io_request) and a debug ID
of their own.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241216204124.3752367-5-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
netfs: Fix is-caching check in read-retry
The read-retry code checks the NETFS_RREQ_COPY_TO_CACHE flag to determine
if there might be failed reads from the cache that need turning into reads
from the server, with the intention of skipping the complicated part if it
can. The code that set the flag, however, got lost during the read-side
rewrite.
Fix the check to see if the cache_resources are valid instead. The flag
can then be removed.
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/3752048.1734381285@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
When the caching for a cookie is temporarily disabled (e.g. due to a DIO
write on that file), future copying to the cache for that file is disabled
until all fds open on that file are closed. However, if netfslib is using
the deprecated PG_private_2 method (such as is currently used by ceph), and
decides it wants to copy to the cache, netfs_advance_write() will just bail
at the first check seeing that the cache stream is unavailable, and
indicate that it dealt with all the content.
This means that we have no subrequests to provide notifications to drive
the state machine or even to pin the request and the request just gets
discarded, leaving the folios with PG_private_2 set.
Fix this by jumping directly to cancel the request if the cache is not
available. That way, we don't remove mark3 from the folio_queue list and
netfs_pgpriv2_cancel() will clean up the folios.
This was found by running the generic/013 xfstest against ceph with an
active cache and the "-o fsc" option passed to ceph. That would usually
hang
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Reported-by: Max Kellermann <max.kellermann@ionos.com>
Closes: https://lore.kernel.org/r/CAKPOu+_4m80thNy5_fvROoxBm689YtA0dZ-=gcmkzwYSY4syqw@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241213135013.2964079-11-dhowells@redhat.com
cc: Jeff Layton <jlayton@kernel.org>
cc: Ilya Dryomov <idryomov@gmail.com>
cc: Xiubo Li <xiubli@redhat.com>
cc: netfs@lists.linux.dev
cc: ceph-devel@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
At the end of netfs_unlock_read_folio() in which folios are marked
appropriately for copying to the cache (either with by being marked dirty
and having their private data set or by having PG_private_2 set) and then
unlocked, the folio_queue struct has the entry pointing to the folio
cleared. This presents a problem for netfs_pgpriv2_write_to_the_cache(),
which is used to write folios marked with PG_private_2 to the cache as it
expects to be able to trawl the folio_queue list thereafter to find the
relevant folios, leading to a hang.
Fix this by not clearing the folio_queue entry if we're going to do the
deprecated copy-to-cache. The clearance will be done instead as the folios
are written to the cache.
This can be reproduced by starting cachefiles, mounting a ceph filesystem
with "-o fsc" and writing to it.
Fixes: 796a404964 ("netfs: In readahead, put the folio refs as soon extracted")
Reported-by: Max Kellermann <max.kellermann@ionos.com>
Closes: https://lore.kernel.org/r/CAKPOu+_4m80thNy5_fvROoxBm689YtA0dZ-=gcmkzwYSY4syqw@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241213135013.2964079-10-dhowells@redhat.com
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
cc: Jeff Layton <jlayton@kernel.org>
cc: Ilya Dryomov <idryomov@gmail.com>
cc: Xiubo Li <xiubli@redhat.com>
cc: netfs@lists.linux.dev
cc: ceph-devel@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
syzkaller reported recursion with a loop of three calls (netfs_rreq_assess,
netfs_retry_reads and netfs_rreq_terminated) hitting the limit of the stack
during an unbuffered or direct I/O read.
There are a number of issues:
(1) There is no limit on the number of retries.
(2) A subrequest is supposed to be abandoned if it does not transfer
anything (NETFS_SREQ_NO_PROGRESS), but that isn't checked under all
circumstances.
(3) The actual root cause, which is this:
if (atomic_dec_and_test(&rreq->nr_outstanding))
netfs_rreq_terminated(rreq, ...);
When we do a retry, we bump the rreq->nr_outstanding counter to
prevent the final cleanup phase running before we've finished
dispatching the retries. The problem is if we hit 0, we have to do
the cleanup phase - but we're in the cleanup phase and end up
repeating the retry cycle, hence the recursion.
Work around the problem by limiting the number of retries. This is based
on Lizhi Xu's patch[1], and makes the following changes:
(1) Replace NETFS_SREQ_NO_PROGRESS with NETFS_SREQ_MADE_PROGRESS and make
the filesystem set it if it managed to read or write at least one byte
of data. Clear this bit before issuing a subrequest.
(2) Add a ->retry_count member to the subrequest and increment it any time
we do a retry.
(3) Remove the NETFS_SREQ_RETRYING flag as it is superfluous with
->retry_count. If the latter is non-zero, we're doing a retry.
(4) Abandon a subrequest if retry_count is non-zero and we made no
progress.
(5) Use ->retry_count in both the write-side and the read-size.
[?] Question: Should I set a hard limit on retry_count in both read and
write? Say it hits 50, we always abandon it. The problem is that
these changes only mitigate the issue. As long as it made at least one
byte of progress, the recursion is still an issue. This patch
mitigates the problem, but does not fix the underlying cause. I have
patches that will do that, but it's an intrusive fix that's currently
pending for the next merge window.
The oops generated by KASAN looks something like:
BUG: TASK stack guard page was hit at ffffc9000482ff48 (stack is ffffc90004830000..ffffc90004838000)
Oops: stack guard page: 0000 [#1] PREEMPT SMP KASAN NOPTI
...
RIP: 0010:mark_lock+0x25/0xc60 kernel/locking/lockdep.c:4686
...
mark_usage kernel/locking/lockdep.c:4646 [inline]
__lock_acquire+0x906/0x3ce0 kernel/locking/lockdep.c:5156
lock_acquire.part.0+0x11b/0x380 kernel/locking/lockdep.c:5825
local_lock_acquire include/linux/local_lock_internal.h:29 [inline]
___slab_alloc+0x123/0x1880 mm/slub.c:3695
__slab_alloc.constprop.0+0x56/0xb0 mm/slub.c:3908
__slab_alloc_node mm/slub.c:3961 [inline]
slab_alloc_node mm/slub.c:4122 [inline]
kmem_cache_alloc_noprof+0x2a7/0x2f0 mm/slub.c:4141
radix_tree_node_alloc.constprop.0+0x1e8/0x350 lib/radix-tree.c:253
idr_get_free+0x528/0xa40 lib/radix-tree.c:1506
idr_alloc_u32+0x191/0x2f0 lib/idr.c:46
idr_alloc+0xc1/0x130 lib/idr.c:87
p9_tag_alloc+0x394/0x870 net/9p/client.c:321
p9_client_prepare_req+0x19f/0x4d0 net/9p/client.c:644
p9_client_zc_rpc.constprop.0+0x105/0x880 net/9p/client.c:793
p9_client_read_once+0x443/0x820 net/9p/client.c:1570
p9_client_read+0x13f/0x1b0 net/9p/client.c:1534
v9fs_issue_read+0x115/0x310 fs/9p/vfs_addr.c:74
netfs_retry_read_subrequests fs/netfs/read_retry.c:60 [inline]
netfs_retry_reads+0x153a/0x1d00 fs/netfs/read_retry.c:232
netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371
netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407
netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235
netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371
netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407
netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235
netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371
...
netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407
netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235
netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371
netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407
netfs_retry_reads+0x155e/0x1d00 fs/netfs/read_retry.c:235
netfs_rreq_assess+0x5d3/0x870 fs/netfs/read_collect.c:371
netfs_rreq_terminated+0xe5/0x110 fs/netfs/read_collect.c:407
netfs_dispatch_unbuffered_reads fs/netfs/direct_read.c:103 [inline]
netfs_unbuffered_read fs/netfs/direct_read.c:127 [inline]
netfs_unbuffered_read_iter_locked+0x12f6/0x19b0 fs/netfs/direct_read.c:221
netfs_unbuffered_read_iter+0xc5/0x100 fs/netfs/direct_read.c:256
v9fs_file_read_iter+0xbf/0x100 fs/9p/vfs_file.c:361
do_iter_readv_writev+0x614/0x7f0 fs/read_write.c:832
vfs_readv+0x4cf/0x890 fs/read_write.c:1025
do_preadv fs/read_write.c:1142 [inline]
__do_sys_preadv fs/read_write.c:1192 [inline]
__se_sys_preadv fs/read_write.c:1187 [inline]
__x64_sys_preadv+0x22d/0x310 fs/read_write.c:1187
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Closes: https://syzkaller.appspot.com/bug?extid=1fc6f64c40a9d143cfb6
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241108034020.3695718-1-lizhi.xu@windriver.com/ [1]
Link: https://lore.kernel.org/r/20241213135013.2964079-9-dhowells@redhat.com
Tested-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com
Suggested-by: Lizhi Xu <lizhi.xu@windriver.com>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Jeff Layton <jlayton@kernel.org>
cc: v9fs@lists.linux.dev
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Reported-by: syzbot+885c03ad650731743489@syzkaller.appspotmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Use clear_and_wake_up_bit() rather than something like:
clear_bit_unlock(NETFS_RREQ_IN_PROGRESS, &rreq->flags);
wake_up_bit(&rreq->flags, NETFS_RREQ_IN_PROGRESS);
as there needs to be a barrier inserted between which is present in
clear_and_wake_up_bit().
Fixes: 288ace2f57 ("netfs: New writeback implementation")
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241213135013.2964079-8-dhowells@redhat.com
Reviewed-by: Akira Yokosawa <akiyks@gmail.com>
cc: Zilin Guan <zilin@seu.edu.cn>
cc: Akira Yokosawa <akiyks@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
The function netfs_unbuffered_write_iter_locked() in
fs/netfs/direct_write.c contains an unnecessary smp_rmb() call after
wait_on_bit(). Since wait_on_bit() already incorporates a memory barrier
that ensures the flag update is visible before the function returns, the
smp_rmb() provides no additional benefit and incurs unnecessary overhead.
This patch removes the redundant barrier to simplify and optimize the code.
Signed-off-by: Zilin Guan <zilin@seu.edu.cn>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20241207021952.2978530-1-zilin@seu.edu.cn/
Link: https://lore.kernel.org/r/20241213135013.2964079-7-dhowells@redhat.com
Reviewed-by: Akira Yokosawa <akiyks@gmail.com>
cc: Akira Yokosawa <akiyks@gmail.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
When a read subrequest finishes, if it doesn't have sufficient coverage to
complete the folio(s) covering either side of it, it will donate the excess
coverage to the adjacent subrequests on either side, offloading
responsibility for unlocking the folio(s) covered to them.
Now, preference is given to donating down to a lower file offset over
donating up because that check is done first - but there's no check that
the lower subreq is actually contiguous, and so we can end up donating
incorrectly.
The scenario seen[1] is that an 8MiB readahead request spanning four 2MiB
folios is split into eight 1MiB subreqs (numbered 1 through 8). These
terminate in the order 1,6,2,5,3,7,4,8. What happens is:
- 1 donates to 2
- 6 donates to 5
- 2 completes, unlocking the first folio (with 1).
- 5 completes, unlocking the third folio (with 6).
- 3 donates to 4
- 7 donates to 4 incorrectly
- 4 completes, unlocking the second folio (with 3), but can't use
the excess from 7.
- 8 donates to 4, also incorrectly.
Fix this by preventing downward donation if the subreqs are not contiguous
(in the example above, 7 donates to 4 across the gap left by 5 and 6).
Reported-by: Shyam Prasad N <nspmangalore@gmail.com>
Closes: https://lore.kernel.org/r/CANT5p=qBwjBm-D8soFVVtswGEfmMtQXVW83=TNfUtvyHeFQZBA@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/526707.1733224486@warthog.procyon.org.uk/ [1]
Link: https://lore.kernel.org/r/20241213135013.2964079-3-dhowells@redhat.com
cc: Steve French <sfrench@samba.org>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZzcUoQAKCRCRxhvAZXjc
omxAAP9WE8zSxeu7Foa6+OmLO62mYdB8rRrQ4OjzX+zunL0UnAD9FAHPsB4amWm4
/zK3Nf7ipijop5+RgSJTgURffASKOgI=
=ifkk
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.13.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull netfs updates from Christian Brauner:
"Various fixes for the netfs library and related infrastructure:
cachefiles:
- Fix a dentry leak in cachefiles_open_file()
- Fix incorrect length return value in
cachefiles_ondemand_fd_write_iter()
- Fix missing pos updates in cachefiles_ondemand_fd_write_iter()
- Clean up in cachefiles_commit_tmpfile()
- Fix NULL pointer dereference in object->file
- Add a memory barrier for FSCACHE_VOLUME_CREATING
netfs:
- Remove call to folio_index()
- Fix a few minor bugs in netfs_page_mkwrite()
- Remove unnecessary references to pages"
* tag 'vfs-6.13.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
netfs/fscache: Add a memory barrier for FSCACHE_VOLUME_CREATING
cachefiles: Fix NULL pointer dereference in object->file
cachefiles: Clean up in cachefiles_commit_tmpfile()
cachefiles: Fix missing pos updates in cachefiles_ondemand_fd_write_iter()
cachefiles: Fix incorrect length return value in cachefiles_ondemand_fd_write_iter()
netfs: Remove unnecessary references to pages
netfs: Fix a few minor bugs in netfs_page_mkwrite()
netfs: Remove call to folio_index()
In fscache_create_volume(), there is a missing memory barrier between the
bit-clearing operation and the wake-up operation. This may cause a
situation where, after a wake-up, the bit-clearing operation hasn't been
detected yet, leading to an indefinite wait. The triggering process is as
follows:
[cookie1] [cookie2] [volume_work]
fscache_perform_lookup
fscache_create_volume
fscache_perform_lookup
fscache_create_volume
fscache_create_volume_work
cachefiles_acquire_volume
clear_and_wake_up_bit
test_and_set_bit
test_and_set_bit
goto maybe_wait
goto no_wait
In the above process, cookie1 and cookie2 has the same volume. When cookie1
enters the -no_wait- process, it will clear the bit and wake up the waiting
process. If a barrier is missing, it may cause cookie2 to remain in the
-wait- process indefinitely.
In commit 3288666c72 ("fscache: Use clear_and_wake_up_bit() in
fscache_create_volume_work()"), barriers were added to similar operations
in fscache_create_volume_work(), but fscache_create_volume() was missed.
By combining the clear and wake operations into clear_and_wake_up_bit() to
fix this issue.
Fixes: bfa22da3ed ("fscache: Provide and use cache methods to lookup/create/free a volume")
Signed-off-by: Zizhi Wo <wozizhi@huawei.com>
Link: https://lore.kernel.org/r/20241107110649.3980193-6-wozizhi@huawei.com
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
In the I/O locking code borrowed from NFS into netfslib, i_rwsem is held
locked across a buffered write - but this causes a performance regression
in cifs as it excludes buffered reads for the duration (cifs didn't use any
locking for buffered reads).
Mitigate this somewhat by downgrading the i_rwsem to a read lock across the
buffered write. This at least allows parallel reads to occur whilst
excluding other writes, DIO, truncate and setattr.
Note that this shouldn't be a problem for a buffered write as a read
through an mmap can circumvent i_rwsem anyway.
Also note that we might want to make this change in NFS also.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/1317958.1729096113@warthog.procyon.org.uk
cc: Steve French <sfrench@samba.org>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Trond Myklebust <trondmy@kernel.org>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: linux-nfs@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
netfslib currently defers dropping the ref on the folios it obtains during
readahead to after it has started I/O on the basis that we can do it whilst
we wait for the I/O to complete, but this runs the risk of the I/O
collection racing with this in future.
Furthermore, Matthew Wilcox strongly suggests that the refs should be
dropped immediately, as readahead_folio() does (netfslib is using
__readahead_batch() which doesn't drop the refs).
Fixes: ee4cdf7ba8 ("netfs: Speed up buffered reading")
Suggested-by: Matthew Wilcox <willy@infradead.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/3771538.1728052438@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Matthew Wilcox (Oracle) <willy@infradead.org> says:
A few minor fixes; nothing earth-shattering.
Matthew Wilcox (Oracle) (3):
netfs: Remove call to folio_index()
netfs: Fix a few minor bugs in netfs_page_mkwrite()
netfs: Remove unnecessary references to pages
Link: https://lore.kernel.org/r/20241005182307.3190401-1-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
We can't return with VM_FAULT_SIGBUS | VM_FAULT_LOCKED; the core
code will not unlock the folio in this instance. Introduce a new
"unlock" error exit to handle this case. Use it to handle
the "folio is truncated" check, and change the "writeback interrupted
by a fatal signal" to do a NOPAGE exit instead of letting the core
code install the folio currently under writeback before killing the
process.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Link: https://lore.kernel.org/r/20241005182307.3190401-3-willy@infradead.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
After dividing up a proposed write into subrequests, netfslib sets
NETFS_RREQ_ALL_QUEUED to indicate to the collector that it can move on to
the final cleanup once it has emptied the subrequest queues.
Now, whilst the collector will normally end up running at least once after
this bit is set just because it takes a while to process all the write
subrequests before the collector runs out of subrequests, there exists the
possibility that the issuing thread will be forced to sleep and the
collector thread will clean up all the subrequests before ALL_QUEUED gets
set.
In such a case, the collector thread will not get triggered again and will
never clear NETFS_RREQ_IN_PROGRESS thus leaving a request uncompleted and
causing a potential futute hang.
Fix this by scheduling the write collector if all the subrequest queues are
empty (and thus no writes pending issuance).
Note that we'd do this ideally before queuing the subrequest, but in the
case of buffered writeback, at least, we can't find out that we've run out
of folios until after we've called writeback_iter() and it has returned
NULL - at which point we might not actually have any subrequests still
under construction.
Fixes: 288ace2f57 ("netfs: New writeback implementation")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/3317784.1727880350@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZvqoUgAKCRCRxhvAZXjc
oveLAQC952BB+giATJvX5lPz400g+MoDVlWiPdqAVbF0PCGRRwEA+/7TedOfOkQx
Df3iAzDXVbX/WvWYdZ/DyLp/r44WHAw=
=tGkE
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.12-rc2.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull vfs fixes from Christian Brauner:
"afs:
- Fix setting of the server responding flag
- Remove unused struct afs_address_list and afs_put_address_list()
function
- Fix infinite loop because of unresponsive servers
- Ensure that afs_retry_request() function is correctly added to the
afs_req_ops netfs operations table
netfs:
- Fix netfs_folio tracepoint handling to handle NULL mappings
- Add a missing folio_queue API documentation
- Ensure that netfs_write_folio() correctly advances the iterator via
iov_iter_advance()
- Fix a dentry leak during concurrent cull and cookie lookup
operations in cachefiles
pidfs:
- Correctly handle accessing another task's pid namespace"
* tag 'vfs-6.12-rc2.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
netfs: Fix the netfs_folio tracepoint to handle NULL mapping
netfs: Add folio_queue API documentation
netfs: Advance iterator correctly rather than jumping it
afs: Fix the setting of the server responding flag
afs: Remove unused struct and function prototype
afs: Fix possible infinite loop with unresponsive servers
pidfs: check for valid pid namespace
afs: Fix missing wire-up of afs_retry_request()
cachefiles: fix dentry leak in cachefiles_open_file()
In netfs_write_folio(), use iov_iter_advance() to advance the folio as we
split bits of it off to subrequests rather than manually jumping the
->iov_offset value around. This becomes more problematic when we use a
bounce buffer made out of single-page folios to cover a multipage pagecache
folio.
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/2238548.1727424522@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
In netfslib, a buffered writeback operation has a 'write queue' of folios
that are being written, held in a linear sequence of folio_queue structs.
The 'issuer' adds new folio_queues on the leading edge of the queue and
populates each one progressively; the 'collector' pops them off the
trailing edge and discards them and the folios they point to as they are
consumed.
The queue is required to always retain at least one folio_queue structure.
This allows the queue to be accessed without locking and with just a bit of
barriering.
When a new subrequest is prepared, its ->io_iter iterator is pointed at the
current end of the write queue and then the iterator is extended as more
data is added to the queue until the subrequest is committed.
Now, the problem is that the folio_queue at the leading edge of the write
queue when a subrequest is prepared might have been entirely consumed - but
not yet removed from the queue as it is the only remaining one and is
preventing the queue from collapsing.
So, what happens is that subreq->io_iter is pointed at the spent
folio_queue, then a new folio_queue is added, and, at that point, the
collector is at entirely at liberty to immediately delete the spent
folio_queue.
This leaves the subreq->io_iter pointing at a freed object. If the system
is lucky, iterate_folioq() sees ->io_iter, sees the as-yet uncorrupted
freed object and advances to the next folio_queue in the queue.
In the case seen, however, the freed object gets recycled and put back onto
the queue at the tail and filled to the end. This confuses
iterate_folioq() and it tries to step ->next, which may be NULL - resulting
in an oops.
Fix this by the following means:
(1) When preparing a write subrequest, make sure there's a folio_queue
struct with space in it at the leading edge of the queue. A function
to make space is split out of the function to append a folio so that
it can be called for this purpose.
(2) If the request struct iterator is pointing to a completely spent
folio_queue when we make space, then advance the iterator to the newly
allocated folio_queue. The subrequest's iterator will then be set
from this.
The oops could be triggered using the generic/346 xfstest with a filesystem
on9P over TCP with cache=loose. The oops looked something like:
BUG: kernel NULL pointer dereference, address: 0000000000000008
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
...
RIP: 0010:_copy_from_iter+0x2db/0x530
...
Call Trace:
<TASK>
...
p9pdu_vwritef+0x3d8/0x5d0
p9_client_prepare_req+0xa8/0x140
p9_client_rpc+0x81/0x280
p9_client_write+0xcf/0x1c0
v9fs_issue_write+0x87/0xc0
netfs_advance_write+0xa0/0xb0
netfs_write_folio.isra.0+0x42d/0x500
netfs_writepages+0x15a/0x1f0
do_writepages+0xd1/0x220
filemap_fdatawrite_wbc+0x5c/0x80
v9fs_mmap_vm_close+0x7d/0xb0
remove_vma+0x35/0x70
vms_complete_munmap_vmas+0x11a/0x170
do_vmi_align_munmap+0x17d/0x1c0
do_vmi_munmap+0x13e/0x150
__vm_munmap+0x92/0xd0
__x64_sys_munmap+0x17/0x20
do_syscall_64+0x80/0xe0
entry_SYSCALL_64_after_hwframe+0x71/0x79
This also fixed a similar-looking issue with cifs and generic/074.
Fixes: cd0277ed0c ("netfs: Use new folio_queue data type and iterator instead of xarray iter")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202409180928.f20b5a08-oliver.sang@intel.com
Closes: https://lore.kernel.org/oe-lkp/202409131438.3f225fbf-oliver.sang@intel.com
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: kernel test robot <oliver.sang@intel.com>
cc: Eric Van Hensbergen <ericvh@kernel.org>
cc: Latchesar Ionkov <lucho@ionkov.net>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Christian Schoenebeck <linux_oss@crudebyte.com>
cc: Paulo Alcantara <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: v9fs@lists.linux.dev
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
The cifs flag CIFS_INO_MODIFIED_ATTR, which indicates that the mtime and
ctime need to be written back on close, got taken over by netfs as
NETFS_ICTX_MODIFIED_ATTR to avoid the need to call a function pointer to
set it.
The flag gets set correctly on buffered writes, but doesn't get set by
netfs_page_mkwrite(), leading to occasional failures in generic/080 and
generic/215.
Fix this by setting the flag in netfs_page_mkwrite().
Fixes: 73425800ac ("netfs, cifs: Move CIFS_INO_MODIFIED_ATTR to netfs_inode")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202409161629.98887b2-oliver.sang@intel.com
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cifs@vger.kernel.org
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZuQEvgAKCRCRxhvAZXjc
onQWAQD6IxAKPU0zom2FoWNilvSzPs7WglTtvddX9pu/lT1RNAD/YC/wOLW8mvAv
9oTAmigQDQQhEWdJA9RgLZBiw7k+DAw=
=zWFb
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.12.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull netfs updates from Christian Brauner:
"This contains the work to improve read/write performance for the new
netfs library.
The main performance enhancing changes are:
- Define a structure, struct folio_queue, and a new iterator type,
ITER_FOLIOQ, to hold a buffer as a replacement for ITER_XARRAY. See
that patch for questions about naming and form.
ITER_FOLIOQ is provided as a replacement for ITER_XARRAY. The
problem with an xarray is that accessing it requires the use of a
lock (typically the RCU read lock) - and this means that we can't
supply iterate_and_advance() with a step function that might sleep
(crypto for example) without having to drop the lock between pages.
ITER_FOLIOQ is the iterator for a chain of folio_queue structs,
where each folio_queue holds a small list of folios. A folio_queue
struct is a simpler structure than xarray and is not subject to
concurrent manipulation by the VM. folio_queue is used rather than
a bvec[] as it can form lists of indefinite size, adding to one end
and removing from the other on the fly.
- Provide a copy_folio_from_iter() wrapper.
- Make cifs RDMA support ITER_FOLIOQ.
- Use folio queues in the write-side helpers instead of xarrays.
- Add a function to reset the iterator in a subrequest.
- Simplify the write-side helpers to use sheaves to skip gaps rather
than trying to work out where gaps are.
- In afs, make the read subrequests asynchronous, putting them into
work items to allow the next patch to do progressive
unlocking/reading.
- Overhaul the read-side helpers to improve performance.
- Fix the caching of a partial block at the end of a file.
- Allow a store to be cancelled.
Then some changes for cifs to make it use folio queues instead of
xarrays for crypto bufferage:
- Use raw iteration functions rather than manually coding iteration
when hashing data.
- Switch to using folio_queue for crypto buffers.
- Remove the xarray bits.
Make some adjustments to the /proc/fs/netfs/stats file such that:
- All the netfs stats lines begin 'Netfs:' but change this to
something a bit more useful.
- Add a couple of stats counters to track the numbers of skips and
waits on the per-inode writeback serialisation lock to make it
easier to check for this as a source of performance loss.
Miscellaneous work:
- Ensure that the sb_writers lock is taken around
vfs_{set,remove}xattr() in the cachefiles code.
- Reduce the number of conditional branches in netfs_perform_write().
- Move the CIFS_INO_MODIFIED_ATTR flag to the netfs_inode struct and
remove cifs_post_modify().
- Move the max_len/max_nr_segs members from netfs_io_subrequest to
netfs_io_request as they're only needed for one subreq at a time.
- Add an 'unknown' source value for tracing purposes.
- Remove NETFS_COPY_TO_CACHE as it's no longer used.
- Set the request work function up front at allocation time.
- Use bh-disabling spinlocks for rreq->lock as cachefiles completion
may be run from block-filesystem DIO completion in softirq context.
- Remove fs/netfs/io.c"
* tag 'vfs-6.12.netfs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (25 commits)
docs: filesystems: corrected grammar of netfs page
cifs: Don't support ITER_XARRAY
cifs: Switch crypto buffer to use a folio_queue rather than an xarray
cifs: Use iterate_and_advance*() routines directly for hashing
netfs: Cancel dirty folios that have no storage destination
cachefiles, netfs: Fix write to partial block at EOF
netfs: Remove fs/netfs/io.c
netfs: Speed up buffered reading
afs: Make read subreqs async
netfs: Simplify the writeback code
netfs: Provide an iterator-reset function
netfs: Use new folio_queue data type and iterator instead of xarray iter
cifs: Provide the capability to extract from ITER_FOLIOQ to RDMA SGEs
iov_iter: Provide copy_folio_from_iter()
mm: Define struct folio_queue and ITER_FOLIOQ to handle a sequence of folios
netfs: Use bh-disabling spinlocks for rreq->lock
netfs: Set the request work function upon allocation
netfs: Remove NETFS_COPY_TO_CACHE
netfs: Reserve netfs_sreq_source 0 as unset/unknown
netfs: Move max_len/max_nr_segs from netfs_io_subrequest to netfs_io_stream
...
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZuQEGwAKCRCRxhvAZXjc
ojIuAQC433+hBkvjvmQ7H0r5rgZSjUuCTG3bSmdU7RJmPHUHhwEA85v/NGq53f+W
IhandK6t+Cf0JYpFZ3N0bT88hDYVhQQ=
=9zGL
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.12.misc' of gitolite.kernel.org:pub/scm/linux/kernel/git/vfs/vfs
Pull misc vfs updates from Christian Brauner:
"This contains the usual pile of misc updates:
Features:
- Add F_CREATED_QUERY fcntl() that allows userspace to query whether
a file was actually created. Often userspace wants to know whether
an O_CREATE request did actually create a file without using
O_EXCL. The current logic is that to first attempts to open the
file without O_CREAT | O_EXCL and if ENOENT is returned userspace
tries again with both flags. If that succeeds all is well. If it
now reports EEXIST it retries.
That works fairly well but some corner cases make this more
involved. If this operates on a dangling symlink the first openat()
without O_CREAT | O_EXCL will return ENOENT but the second openat()
with O_CREAT | O_EXCL will fail with EEXIST.
The reason is that openat() without O_CREAT | O_EXCL follows the
symlink while O_CREAT | O_EXCL doesn't for security reasons. So
it's not something we can really change unless we add an explicit
opt-in via O_FOLLOW which seems really ugly.
All available workarounds are really nasty (fanotify, bpf lsm etc)
so add a simple fcntl().
- Try an opportunistic lookup for O_CREAT. Today, when opening a file
we'll typically do a fast lookup, but if O_CREAT is set, the kernel
always takes the exclusive inode lock. This was likely done with
the expectation that O_CREAT means that we always expect to do the
create, but that's often not the case. Many programs set O_CREAT
even in scenarios where the file already exists (see related
F_CREATED_QUERY patch motivation above).
The series contained in the pr rearranges the pathwalk-for-open
code to also attempt a fast_lookup in certain O_CREAT cases. If a
positive dentry is found, the inode_lock can be avoided altogether
and it can stay in rcuwalk mode for the last step_into.
- Expose the 64 bit mount id via name_to_handle_at()
Now that we provide a unique 64-bit mount ID interface in statx(2),
we can now provide a race-free way for name_to_handle_at(2) to
provide a file handle and corresponding mount without needing to
worry about racing with /proc/mountinfo parsing or having to open a
file just to do statx(2).
While this is not necessary if you are using AT_EMPTY_PATH and
don't care about an extra statx(2) call, users that pass full paths
into name_to_handle_at(2) need to know which mount the file handle
comes from (to make sure they don't try to open_by_handle_at a file
handle from a different filesystem) and switching to AT_EMPTY_PATH
would require allocating a file for every name_to_handle_at(2) call
- Add a per dentry expire timeout to autofs
There are two fairly well known automounter map formats, the autofs
format and the amd format (more or less System V and Berkley).
Some time ago Linux autofs added an amd map format parser that
implemented a fair amount of the amd functionality. This was done
within the autofs infrastructure and some functionality wasn't
implemented because it either didn't make sense or required extra
kernel changes. The idea was to restrict changes to be within the
existing autofs functionality as much as possible and leave changes
with a wider scope to be considered later.
One of these changes is implementing the amd options:
1) "unmount", expire this mount according to a timeout (same as
the current autofs default).
2) "nounmount", don't expire this mount (same as setting the
autofs timeout to 0 except only for this specific mount) .
3) "utimeout=<seconds>", expire this mount using the specified
timeout (again same as setting the autofs timeout but only for
this mount)
To implement these options per-dentry expire timeouts need to be
implemented for autofs indirect mounts. This is because all map
keys (mounts) for autofs indirect mounts use an expire timeout
stored in the autofs mount super block info. structure and all
indirect mounts use the same expire timeout.
Fixes:
- Fix missing fput for FSCONFIG_SET_FD in autofs
- Use param->file for FSCONFIG_SET_FD in coda
- Delete the 'fs/netfs' proc subtreee when netfs module exits
- Make sure that struct uid_gid_map fits into a single cacheline
- Don't flush in-flight wb switches for superblocks without cgroup
writeback
- Correcting the idmapping mount example in the idmapping
documentation
- Fix a race between evice_inodes() and find_inode() and iput()
- Refine the show_inode_state() macro definition in writeback code
- Prevent dump_mapping() from accessing invalid dentry.d_name.name
- Show actual source for debugfs in /proc/mounts
- Annotate data-race of busy_poll_usecs in eventpoll
- Don't WARN for racy path_noexec check in exec code
- Handle OOM on mnt_warn_timestamp_expiry()
- Fix some spelling in the iomap design documentation
- Fix typo in procfs comment
- Fix typo in fs/namespace.c comment
Cleanups:
- Add the VFS git tree to the MAINTAINERS file
- Move FMODE_UNSIGNED_OFFSET to fop_flags freeing up another f_mode
bit in struct file bringing us to 5 free f_mode bits
- Remove the __I_DIO_WAKEUP bit from i_state flags as we can simplify
the wait mechanism
- Remove the unused path_put_init() helper
- Replace a __u32 with u32 for s_fsnotify_mask as __u32 is uapi
specific
- Replace the unsigned long i_state member with a u32 i_state member
in struct inode freeing up 4 bytes in struct inode. Instead of
using the bit based wait apis we're now using the var event apis
and using the individual bytes of the i_state member to wait on
state changes
- Explain how per-syscall AT_* flags should be allocated
- Use in_group_or_capable() helper to simplify the posix acl mode
update code
- Switch to LIST_HEAD() in fsync_buffers_list() to simplify the code
- Removed comment about d_rcu_to_refcount() as that function doesn't
exist anymore
- Add kernel documentation for lookup_fast()
- Don't re-zero evenpoll fields
- Remove outdated comment after close_fd()
- Fix imprecise wording in comment about the pipe filesystem
- Drop GFP_NOFAIL mode from alloc_page_buffers
- Missing blank line warnings and struct declaration improved in
file_table
- Annotate struct poll_list with __counted_by()
- Remove the unused read parameter in percpu-rwsem
- Remove linux/prefetch.h include from direct-io code
- Use kmemdup_array instead of kmemdup for multiple allocation in
mnt_idmapping code
- Remove unused mnt_cursor_del() declaration
Performance tweaks:
- Dodge smp_mb in break_lease and break_deleg in the common case
- Only read fops once in fops_{get,put}()
- Use RCU in ilookup()
- Elide smp_mb in iversion handling in the common case
- Drop one lock trip in evict()"
* tag 'vfs-6.12.misc' of gitolite.kernel.org:pub/scm/linux/kernel/git/vfs/vfs: (58 commits)
uidgid: make sure we fit into one cacheline
proc: Fix typo in the comment
fs/pipe: Correct imprecise wording in comment
fhandle: expose u64 mount id to name_to_handle_at(2)
uapi: explain how per-syscall AT_* flags should be allocated
fs: drop GFP_NOFAIL mode from alloc_page_buffers
writeback: Refine the show_inode_state() macro definition
fs/inode: Prevent dump_mapping() accessing invalid dentry.d_name.name
mnt_idmapping: Use kmemdup_array instead of kmemdup for multiple allocation
netfs: Delete subtree of 'fs/netfs' when netfs module exits
fs: use LIST_HEAD() to simplify code
inode: make i_state a u32
inode: port __I_LRU_ISOLATING to var event
vfs: fix race between evice_inodes() and find_inode()&iput()
inode: port __I_NEW to var event
inode: port __I_SYNC to var event
fs: reorder i_state bits
fs: add i_state helpers
MAINTAINERS: add the VFS git tree
fs: s/__u32/u32/ for s_fsnotify_mask
...
Kafs wants to be able to cache the contents of directories (and symlinks),
but whilst these are downloaded from the server with the FS.FetchData RPC
op and similar, the same as for regular files, they can't be updated by
FS.StoreData, but rather have special operations (FS.MakeDir, etc.).
Now, rather than redownloading a directory's content after each change made
to that directory, kafs modifies the local blob. This blob can be saved
out to the cache, and since it's using netfslib, kafs just marks the folios
dirty and lets ->writepages() on the directory take care of it, as for an
regular file.
This is fine as long as there's a cache as although the upload stream is
disabled, there's a cache stream to drive the procedure. But if the cache
goes away in the meantime, suddenly there's no way do any writes and the
code gets confused, complains "R=%x: No submit" to dmesg and leaves the
dirty folio hanging.
Fix this by just cancelling the store of the folio if neither stream is
active. (If there's no cache at the time of dirtying, we should just not
mark the folio dirty).
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20240814203850.2240469-23-dhowells@redhat.com/ # v2
Signed-off-by: Christian Brauner <brauner@kernel.org>
Because it uses DIO writes, cachefiles is unable to make a write to the
backing file if that write is not aligned to and sized according to the
backing file's DIO block alignment. This makes it tricky to handle a write
to the cache where the EOF on the network file is not correctly aligned.
To get around this, netfslib attempts to tell the driver it is calling how
much more data there is available beyond the EOF that it can use to pad the
write (netfslib preclears the part of the folio above the EOF). However,
it tries to tell the cache what the maximum length is, but doesn't
calculate this correctly; and, in any case, cachefiles actually ignores the
value and just skips the block.
Fix this by:
(1) Change the value passed to indicate the amount of extra data that can
be added to the operation (now ->submit_extendable_to). This is much
simpler to calculate as it's just the end of the folio minus the top
of the data within the folio - rather than having to account for data
spread over multiple folios.
(2) Make cachefiles add some of this data if the subrequest it is given
ends at the network file's i_size if the extra data is sufficient to
pad out to a whole block.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20240814203850.2240469-22-dhowells@redhat.com/ # v2
Signed-off-by: Christian Brauner <brauner@kernel.org>
Improve the efficiency of buffered reads in a number of ways:
(1) Overhaul the algorithm in general so that it's a lot more compact and
split the read submission code between buffered and unbuffered
versions. The unbuffered version can be vastly simplified.
(2) Read-result collection is handed off to a work queue rather than being
done in the I/O thread. Multiple subrequests can be processes
simultaneously.
(3) When a subrequest is collected, any folios it fully spans are
collected and "spare" data on either side is donated to either the
previous or the next subrequest in the sequence.
Notes:
(*) Readahead expansion is massively slows down fio, presumably because it
causes a load of extra allocations, both folio and xarray, up front
before RPC requests can be transmitted.
(*) RDMA with cifs does appear to work, both with SIW and RXE.
(*) PG_private_2-based reading and copy-to-cache is split out into its own
file and altered to use folio_queue. Note that the copy to the cache
now creates a new write transaction against the cache and adds the
folios to be copied into it. This allows it to use part of the
writeback I/O code.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20240814203850.2240469-20-dhowells@redhat.com/ # v2
Signed-off-by: Christian Brauner <brauner@kernel.org>
Use the new folio_queue structures to simplify the writeback code. The
problem with referring to the i_pages xarray directly is that we may have
gaps in the sequence of folios we're writing from that we need to skip when
we're removing the writeback mark from the folios we're writing back from.
At the moment the code tries to deal with this by carefully tracking the
gaps in each writeback stream (eg. write to server and write to cache) and
divining when there's a gap that spans folios (something that's not helped
by folios not being a consistent size).
Instead, the folio_queue buffer contains pointers only the folios we're
dealing with, has them in ascending order and indicates a gap by placing
non-consequitive folios next to each other. This makes it possible to
track where we need to clean up to by just keeping track of where we've
processed to on each stream and taking the minimum.
Note that the I/O iterator is always rounded up to the end of the folio,
even if that is beyond the EOF position, so that the cache can do DIO from
the page. The excess space is cleared, though mmapped writes clobber it.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: netfs@lists.linux.dev
cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20240814203850.2240469-18-dhowells@redhat.com/ # v2
Signed-off-by: Christian Brauner <brauner@kernel.org>
