Using of two unconnected IDs req->in.h.unique and req->intr_unique does not
allow to link requests to a hash table. We need can't use none of them as a
key to calculate hash.
This patch changes the algorithm of allocation of IDs for a request. Plain
requests obtain even ID, while interrupt requests are encoded in the low
bit. So, in next patches we will be able to use the rest of ID bits to
calculate hash, and the hash will be the same for plain and interrupt
requests.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Currently, we take fc->lock there only to check for fc->connected.
But this flag is changed only on connection abort, which is very
rare operation.
So allow checking fc->connected under just fc->bg_lock and use this lock
(as well as fc->lock) when resetting fc->connected.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
To reduce contention of fc->lock, this patch introduces bg_lock for
protection of fields related to background queue. These are:
max_background, congestion_threshold, num_background, active_background,
bg_queue and blocked.
This allows next patch to make async reads not requiring fc->lock, so async
reads and writes will have better performance executed in parallel.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
This cleanup patch makes the function to use the primitive
instead of direct dereferencing.
Also, move fiq dereferencing out of cycle, since it's
always constant.
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Using waitqueue_active() is racy. Make sure we issue a wake_up()
unconditionally after storing into fc->blocked. After that it's okay to
optimize with waitqueue_active() since the first wake up provides the
necessary barrier for all waiters, not the just the woken one.
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Fixes: 3c18ef8117 ("fuse: optimize wake_up")
Cc: <stable@vger.kernel.org> # v3.10
Otherwise fuse_dev_do_write() could come in and finish off the request, and
the set_bit(FR_SENT, ...) could trigger the WARN_ON(test_bit(FR_SENT, ...))
in request_end().
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Reported-by: syzbot+ef054c4d3f64cd7f7cec@syzkaller.appspotmai
Fixes: 46c34a348b ("fuse: no fc->lock for pqueue parts")
Cc: <stable@vger.kernel.org> # v4.2
The 'bufs' array contains 'pipe->buffers' elements, but the
fuse_dev_splice_write() uses only 'pipe->nrbufs' elements.
So reduce the allocation size to 'pipe->nrbufs' elements.
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
The amount of pipe->buffers is basically controlled by userspace by
fcntl(... F_SETPIPE_SZ ...) so it could be large. High order allocations
could be slow (if memory is heavily fragmented) or may fail if the order
is larger than PAGE_ALLOC_COSTLY_ORDER.
Since the 'bufs' doesn't need to be physically contiguous, use
the kvmalloc_array() to allocate memory. If high order
page isn't available, the kvamalloc*() will fallback to 0-order.
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
fuse_dev_splice_write() reads pipe->buffers to determine the size of
'bufs' array before taking the pipe_lock(). This is not safe as
another thread might change the 'pipe->buffers' between the allocation
and taking the pipe_lock(). So we end up with too small 'bufs' array.
Move the bufs allocations inside pipe_lock()/pipe_unlock() to fix this.
Fixes: dd3bb14f44 ("fuse: support splice() writing to fuse device")
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: <stable@vger.kernel.org> # v2.6.35
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
fuse_abort_conn() does not guarantee that all async requests have actually
finished aborting (i.e. their ->end() function is called). This could
actually result in still used inodes after umount.
Add a helper to wait until all requests are fully done. This is done by
looking at the "num_waiting" counter. When this counter drops to zero, we
can be sure that no more requests are outstanding.
Fixes: 0d8e84b043 ("fuse: simplify request abort")
Cc: <stable@vger.kernel.org> # v4.2
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
fuse_dev_release() assumes that it's the only one referencing the
fpq->processing list, but that's not true, since fuse_abort_conn() can be
doing the same without any serialization between the two.
Fixes: c3696046be ("fuse: separate pqueue for clones")
Cc: <stable@vger.kernel.org> # v4.2
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Refcounting of request is broken when fuse_abort_conn() is called and
request is on the fpq->io list:
- ref is taken too late
- then it is not dropped
Fixes: 0d8e84b043 ("fuse: simplify request abort")
Cc: <stable@vger.kernel.org> # v4.2
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
If a connection gets aborted while congested, FUSE can leave
nr_wb_congested[] stuck until reboot causing wait_iff_congested() to
wait spuriously which can lead to severe performance degradation.
The leak is caused by gating congestion state clearing with
fc->connected test in request_end(). This was added way back in 2009
by 26c3679101 ("fuse: destroy bdi on umount"). While the commit
description doesn't explain why the test was added, it most likely was
to avoid dereferencing bdi after it got destroyed.
Since then, bdi lifetime rules have changed many times and now we're
always guaranteed to have access to the bdi while the superblock is
alive (fc->sb).
Drop fc->connected conditional to avoid leaking congestion states.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Joshua Miller <joshmiller@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: stable@vger.kernel.org # v2.6.29+
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
In order to support mounts from namespaces other than init_user_ns, fuse
must translate uids and gids to/from the userns of the process servicing
requests on /dev/fuse. This patch does that, with a couple of restrictions
on the namespace:
- The userns for the fuse connection is fixed to the namespace
from which /dev/fuse is opened.
- The namespace must be the same as s_user_ns.
These restrictions simplify the implementation by avoiding the need to pass
around userns references and by allowing fuse to rely on the checks in
setattr_prepare for ownership changes. Either restriction could be relaxed
in the future if needed.
For cuse the userns used is the opener of /dev/cuse. Semantically the cuse
support does not appear safe for unprivileged users. Practically the
permissions on /dev/cuse only make it accessible to the global root user.
If something slips through the cracks in a user namespace the only users
who will be able to use the cuse device are those users mapped into the
user namespace.
Translation in the posix acl is updated to use the uuser namespace of the
filesystem. Avoiding cases which might bypass this translation is handled
in a following change.
This change is stronlgy based on a similar change from Seth Forshee and
Dongsu Park.
Cc: Seth Forshee <seth.forshee@canonical.com>
Cc: Dongsu Park <dongsu@kinvolk.io>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Upon a cursory examinination the uid and gid of a fuse request are
necessary for correct operation. Failing a fuse request where those
values are not reliable seems a straight forward and reliable means of
ensuring that fuse requests with bad data are not sent or processed.
In most cases the vfs will avoid actions it suspects will cause
an inode write back of an inode with an invalid uid or gid. But that does
not map precisely to what fuse is doing, so test for this and solve
this at the fuse level as well.
Performing this work in fuse_req_init_context is cheap as the code is
already performing the translation here and only needs to check the
result of the translation to see if things are not representable in
a form the fuse server can handle.
[SzM] Don't zero the context for the nofail case, just keep using the
munging version (makes sense for debugging and doesn't hurt).
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
At the point of fuse_dev_do_read the user space process that initiated the
action on the fuse filesystem may no longer exist. The process have been
killed or may have fired an asynchronous request and exited.
If the initial process has exited, the code "pid_vnr(find_pid_ns(in->h.pid,
fc->pid_ns)" will either return a pid of 0, or in the unlikely event that
the pid has been reallocated it can return practically any pid. Any pid is
possible as the pid allocator allocates pid numbers in different pid
namespaces independently.
The only way to make translation in fuse_dev_do_read reliable is to call
get_pid in fuse_req_init_context, and pid_vnr followed by put_pid in
fuse_dev_do_read. That reference counting in other contexts has been shown
to bounce cache lines between processors and in general be slow. So that
is not desirable.
The only known user of running the fuse server in a different pid namespace
from the filesystem does not care what the pids are in the fuse messages so
removing this code should not matter.
Getting the translation to a server running outside of the pid namespace of
a container can still be achieved by playing setns games at mount time. It
is also possible to add an option to pass a pid namespace into the fuse
filesystem at mount time.
Fixes: 5d6d3a301c ("fuse: allow server to run in different pid_ns")
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Currently the userspace has no way of knowing whether the fuse
connection ended because of umount or abort via sysfs. It makes it hard
for filesystems to free the mountpoint after abort without worrying
about removing some new mount.
The patch fixes it by returning different errors when userspace reads
from /dev/fuse (-ENODEV for umount and -ECONNABORTED for abort).
Add a new capability flag FUSE_ABORT_ERROR. If set and the connection is
gone because of sysfs abort, reading from the device will return
-ECONNABORTED.
Signed-off-by: Szymon Lukasz <noh4hss@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
This is the mindless scripted replacement of kernel use of POLL*
variables as described by Al, done by this script:
for V in IN OUT PRI ERR RDNORM RDBAND WRNORM WRBAND HUP RDHUP NVAL MSG; do
L=`git grep -l -w POLL$V | grep -v '^t' | grep -v /um/ | grep -v '^sa' | grep -v '/poll.h$'|grep -v '^D'`
for f in $L; do sed -i "-es/^\([^\"]*\)\(\<POLL$V\>\)/\\1E\\2/" $f; done
done
with de-mangling cleanups yet to come.
NOTE! On almost all architectures, the EPOLL* constants have the same
values as the POLL* constants do. But they keyword here is "almost".
For various bad reasons they aren't the same, and epoll() doesn't
actually work quite correctly in some cases due to this on Sparc et al.
The next patch from Al will sort out the final differences, and we
should be all done.
Scripted-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
All callers of release_pages claim the pages being released are cache
hot. As no one cares about the hotness of pages being released to the
allocator, just ditch the parameter.
No performance impact is expected as the overhead is marginal. The
parameter is removed simply because it is a bit stupid to have a useless
parameter copied everywhere.
Link: http://lkml.kernel.org/r/20171018075952.10627-7-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Please do not apply this to mainline directly, instead please re-run the
coccinelle script shown below and apply its output.
For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't harmful, and changing them results in
churn.
However, for some features, the read/write distinction is critical to
correct operation. To distinguish these cases, separate read/write
accessors must be used. This patch migrates (most) remaining
ACCESS_ONCE() instances to {READ,WRITE}_ONCE(), using the following
coccinelle script:
----
// Convert trivial ACCESS_ONCE() uses to equivalent READ_ONCE() and
// WRITE_ONCE()
// $ make coccicheck COCCI=/home/mark/once.cocci SPFLAGS="--include-headers" MODE=patch
virtual patch
@ depends on patch @
expression E1, E2;
@@
- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)
@ depends on patch @
expression E;
@@
- ACCESS_ONCE(E)
+ READ_ONCE(E)
----
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: davem@davemloft.net
Cc: linux-arch@vger.kernel.org
Cc: mpe@ellerman.id.au
Cc: shuah@kernel.org
Cc: snitzer@redhat.com
Cc: thor.thayer@linux.intel.com
Cc: tj@kernel.org
Cc: viro@zeniv.linux.org.uk
Cc: will.deacon@arm.com
Link: http://lkml.kernel.org/r/1508792849-3115-19-git-send-email-paulmck@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit 0b6e9ea041 ("fuse: Add support for pid namespaces") broke
Sandstorm.io development tools, which have been sending FUSE file
descriptors across PID namespace boundaries since early 2014.
The above patch added a check that prevented I/O on the fuse device file
descriptor if the pid namespace of the reader/writer was different from the
pid namespace of the mounter. With this change passing the device file
descriptor to a different pid namespace simply doesn't work. The check was
added because pids are transferred to/from the fuse userspace server in the
namespace registered at mount time.
To fix this regression, remove the checks and do the following:
1) the pid in the request header (the pid of the task that initiated the
filesystem operation) is translated to the reader's pid namespace. If a
mapping doesn't exist for this pid, then a zero pid is used. Note: even if
a mapping would exist between the initiator task's pid namespace and the
reader's pid namespace the pid will be zero if either mapping from
initator's to mounter's namespace or mapping from mounter's to reader's
namespace doesn't exist.
2) The lk.pid value in setlk/setlkw requests and getlk reply is left alone.
Userspace should not interpret this value anyway. Also allow the
setlk/setlkw operations if the pid of the task cannot be represented in the
mounter's namespace (pid being zero in that case).
Reported-by: Kenton Varda <kenton@sandstorm.io>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Fixes: 0b6e9ea041 ("fuse: Add support for pid namespaces")
Cc: <stable@vger.kernel.org> # v4.12+
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Seth Forshee <seth.forshee@canonical.com>
Pull fuse updates from Miklos Szeredi:
"Support for pid namespaces from Seth and refcount_t work from Elena"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: Add support for pid namespaces
fuse: convert fuse_conn.count from atomic_t to refcount_t
fuse: convert fuse_req.count from atomic_t to refcount_t
fuse: convert fuse_file.count from atomic_t to refcount_t
It is not needed anymore since bdi is initialized whenever superblock
exists.
CC: Miklos Szeredi <miklos@szeredi.hu>
CC: linux-fsdevel@vger.kernel.org
Suggested-by: Miklos Szeredi <mszeredi@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@fb.com>
Allocate struct backing_dev_info separately instead of embedding it
inside the superblock. This unifies handling of bdi among users.
CC: Miklos Szeredi <miklos@szeredi.hu>
CC: linux-fsdevel@vger.kernel.org
Acked-by: Miklos Szeredi <mszeredi@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@fb.com>
When the userspace process servicing fuse requests is running in
a pid namespace then pids passed via the fuse fd are not being
translated into that process' namespace. Translation is necessary
for the pid to be useful to that process.
Since no use case currently exists for changing namespaces all
translations can be done relative to the pid namespace in use
when fuse_conn_init() is called. For fuse this translates to
mount time, and for cuse this is when /dev/cuse is opened. IO for
this connection from another namespace will return errors.
Requests from processes whose pid cannot be translated into the
target namespace will have a value of 0 for in.h.pid.
File locking changes based on previous work done by Eric
Biederman.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
refcount_t type and corresponding API should be
used instead of atomic_t when the variable is used as
a reference counter. This allows to avoid accidental
refcounter overflows that might lead to use-after-free
situations.
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David Windsor <dwindsor@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Fix up affected files that include this signal functionality via sched.h.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is a potential race between fuse_dev_do_write()
and request_wait_answer() contexts as shown below:
TASK 1:
__fuse_request_send():
|--spin_lock(&fiq->waitq.lock);
|--queue_request();
|--spin_unlock(&fiq->waitq.lock);
|--request_wait_answer():
|--if (test_bit(FR_SENT, &req->flags))
<gets pre-empted after it is validated true>
TASK 2:
fuse_dev_do_write():
|--clears bit FR_SENT,
|--request_end():
|--sets bit FR_FINISHED
|--spin_lock(&fiq->waitq.lock);
|--list_del_init(&req->intr_entry);
|--spin_unlock(&fiq->waitq.lock);
|--fuse_put_request();
|--queue_interrupt();
<request gets queued to interrupts list>
|--wake_up_locked(&fiq->waitq);
|--wait_event_freezable();
<as FR_FINISHED is set, it returns and then
the caller frees this request>
Now, the next fuse_dev_do_read(), see interrupts list is not empty
and then calls fuse_read_interrupt() which tries to access the request
which is already free'd and gets the below crash:
[11432.401266] Unable to handle kernel paging request at virtual address
6b6b6b6b6b6b6b6b
...
[11432.418518] Kernel BUG at ffffff80083720e0
[11432.456168] PC is at __list_del_entry+0x6c/0xc4
[11432.463573] LR is at fuse_dev_do_read+0x1ac/0x474
...
[11432.679999] [<ffffff80083720e0>] __list_del_entry+0x6c/0xc4
[11432.687794] [<ffffff80082c65e0>] fuse_dev_do_read+0x1ac/0x474
[11432.693180] [<ffffff80082c6b14>] fuse_dev_read+0x6c/0x78
[11432.699082] [<ffffff80081d5638>] __vfs_read+0xc0/0xe8
[11432.704459] [<ffffff80081d5efc>] vfs_read+0x90/0x108
[11432.709406] [<ffffff80081d67f0>] SyS_read+0x58/0x94
As FR_FINISHED bit is set before deleting the intr_entry with input
queue lock in request completion path, do the testing of this flag and
queueing atomically with the same lock in queue_interrupt().
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Fixes: fd22d62ed0 ("fuse: no fc->lock for iqueue parts")
Cc: <stable@vger.kernel.org> # 4.2+
fuse_abort_conn() moves requests from pending list to a temporary list
before canceling them. This operation races with request_wait_answer()
which also tries to remove the request after it gets a fatal signal. It
checks FR_PENDING flag to determine whether the request is still in the
pending list.
Make fuse_abort_conn() clear FR_PENDING flag so that request_wait_answer()
does not remove the request from temporary list.
This bug causes an Oops when trying to delete an already deleted list entry
in end_requests().
Fixes: ee314a870e ("fuse: abort: no fc->lock needed for request ending")
Signed-off-by: Tahsin Erdogan <tahsin@google.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Cc: <stable@vger.kernel.org> # 4.2+
Pull VFS splice updates from Al Viro:
"There's a bunch of branches this cycle, both mine and from other folks
and I'd rather send pull requests separately.
This one is the conversion of ->splice_read() to ITER_PIPE iov_iter
(and introduction of such). Gets rid of a lot of code in fs/splice.c
and elsewhere; there will be followups, but these are for the next
cycle... Some pipe/splice-related cleanups from Miklos in the same
branch as well"
* 'work.splice_read' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
pipe: fix comment in pipe_buf_operations
pipe: add pipe_buf_steal() helper
pipe: add pipe_buf_confirm() helper
pipe: add pipe_buf_release() helper
pipe: add pipe_buf_get() helper
relay: simplify relay_file_read()
switch default_file_splice_read() to use of pipe-backed iov_iter
switch generic_file_splice_read() to use of ->read_iter()
new iov_iter flavour: pipe-backed
fuse_dev_splice_read(): switch to add_to_pipe()
skb_splice_bits(): get rid of callback
new helper: add_to_pipe()
splice: lift pipe_lock out of splice_to_pipe()
splice: switch get_iovec_page_array() to iov_iter
splice_to_pipe(): don't open-code wakeup_pipe_readers()
consistent treatment of EFAULT on O_DIRECT read/write
* splice_to_pipe() stops at pipe overflow and does *not* take pipe_lock
* ->splice_read() instances do the same
* vmsplice_to_pipe() and do_splice() (ultimate callers of splice_to_pipe())
arrange for waiting, looping, etc. themselves.
That should make pipe_lock the outermost one.
Unfortunately, existing rules for the amount passed by vmsplice_to_pipe()
and do_splice() are quite ugly _and_ userland code can be easily broken
by changing those. It's not even "no more than the maximal capacity of
this pipe" - it's "once we'd fed pipe->nr_buffers pages into the pipe,
leave instead of waiting".
Considering how poorly these rules are documented, let's try "wait for some
space to appear, unless given SPLICE_F_NONBLOCK, then push into pipe
and if we run into overflow, we are done".
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Pull fuse updates from Miklos Szeredi:
"This fixes error propagation from writeback to fsync/close for
writeback cache mode as well as adding a missing capability flag to
the INIT message. The rest are cleanups.
(The commits are recent but all the code actually sat in -next for a
while now. The recommits are due to conflict avoidance and the
addition of Cc: stable@...)"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: use filemap_check_errors()
mm: export filemap_check_errors() to modules
fuse: fix wrong assignment of ->flags in fuse_send_init()
fuse: fuse_flush must check mapping->flags for errors
fuse: fsync() did not return IO errors
fuse: don't mess with blocking signals
new helper: wait_event_killable_exclusive()
fuse: improve aio directIO write performance for size extending writes
We always mixed in the parent pointer into the dentry name hash, but we
did it late at lookup time. It turns out that we can simplify that
lookup-time action by salting the hash with the parent pointer early
instead of late.
A few other users of our string hashes also wanted to mix in their own
pointers into the hash, and those are updated to use the same mechanism.
Hash users that don't have any particular initial salt can just use the
NULL pointer as a no-salt.
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: George Spelvin <linux@sciencehorizons.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time
ago with promise that one day it will be possible to implement page
cache with bigger chunks than PAGE_SIZE.
This promise never materialized. And unlikely will.
We have many places where PAGE_CACHE_SIZE assumed to be equal to
PAGE_SIZE. And it's constant source of confusion on whether
PAGE_CACHE_* or PAGE_* constant should be used in a particular case,
especially on the border between fs and mm.
Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much
breakage to be doable.
Let's stop pretending that pages in page cache are special. They are
not.
The changes are pretty straight-forward:
- <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;
- <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;
- PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN};
- page_cache_get() -> get_page();
- page_cache_release() -> put_page();
This patch contains automated changes generated with coccinelle using
script below. For some reason, coccinelle doesn't patch header files.
I've called spatch for them manually.
The only adjustment after coccinelle is revert of changes to
PAGE_CAHCE_ALIGN definition: we are going to drop it later.
There are few places in the code where coccinelle didn't reach. I'll
fix them manually in a separate patch. Comments and documentation also
will be addressed with the separate patch.
virtual patch
@@
expression E;
@@
- E << (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E
@@
expression E;
@@
- E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E
@@
@@
- PAGE_CACHE_SHIFT
+ PAGE_SHIFT
@@
@@
- PAGE_CACHE_SIZE
+ PAGE_SIZE
@@
@@
- PAGE_CACHE_MASK
+ PAGE_MASK
@@
expression E;
@@
- PAGE_CACHE_ALIGN(E)
+ PAGE_ALIGN(E)
@@
expression E;
@@
- page_cache_get(E)
+ get_page(E)
@@
expression E;
@@
- page_cache_release(E)
+ put_page(E)
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fuse_dev_ioctl() performed fuse_get_dev() on a user-supplied fd,
leading to a type confusion issue. Fix it by checking file->f_op.
Signed-off-by: Jann Horn <jann@thejh.net>
Acked-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Make each fuse device clone refer to a separate processing queue. The only
constraint on userspace code is that the request answer must be written to
the same device clone as it was read off.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Allow fuse device clones to refer to be distinguished. This patch just
adds the infrastructure by associating a separate "struct fuse_dev" with
each clone.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Allow an open fuse device to be "cloned". Userspace can create a clone by:
newfd = open("/dev/fuse", O_RDWR)
ioctl(newfd, FUSE_DEV_IOC_CLONE, &oldfd);
At this point newfd will refer to the same fuse connection as oldfd.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
In fuse_abort_conn() when all requests are on private lists we no longer
need fc->lock protection.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
No longer need to call request_end() with the connection lock held. We
still protect the background counters and queue with fc->lock, so acquire
it if necessary.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Now that we atomically test having already done everything we no longer
need other protection.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
When the connection is aborted it is possible that request_end() will be
called twice. Use atomic test and set to do the actual ending only once.
test_and_set_bit() also provides the necessary barrier semantics so no
explicit smp_wmb() is necessary.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
When an unlocked request is aborted, it is moved from fpq->io to a private
list. Then, after unlocking fpq->lock, the private list is processed and
the requests are finished off.
To protect the private list, we need to mark the request with a flag, so if
in the meantime the request is unlocked the list is not corrupted.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Add a fpq->lock for protecting members of struct fuse_pqueue and FR_LOCKED
request flag.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
- locked list_add() + list_del_init() cancel out
- common handling of case when request is ended here in the read phase
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
This will allow checking ->connected just with the processing queue lock.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
This is just two fields: fc->io and fc->processing.
This patch just rearranges the fields, no functional change.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
wait_event_interruptible_exclusive_locked() will do everything
request_wait() does, so replace it.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Interrupt is only queued after the request has been sent to userspace.
This is either done in request_wait_answer() or fuse_dev_do_read()
depending on which state the request is in at the time of the interrupt.
If it's not yet sent, then queuing the interrupt is postponed until the
request is read. Otherwise (the request has already been read and is
waiting for an answer) the interrupt is queued immedidately.
We want to call queue_interrupt() without fc->lock protection, in which
case there can be a race between the two functions:
- neither of them queue the interrupt (thinking the other one has already
done it).
- both of them queue the interrupt
The first one is prevented by adding memory barriers, the second is
prevented by checking (under fiq->waitq.lock) if the interrupt has already
been queued.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Use fiq->waitq.lock for protecting members of struct fuse_iqueue and
FR_PENDING request flag, previously protected by fc->lock.
Following patches will remove fc->lock protection from these members.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
This will allow checking ->connected just with the input queue lock.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
The input queue contains normal requests (fc->pending), forgets
(fc->forget_*) and interrupts (fc->interrupts). There's also fc->waitq and
fc->fasync for waking up the readers of the fuse device when a request is
available.
The fc->reqctr is also moved to the input queue (assigned to the request
when the request is added to the input queue.
This patch just rearranges the fields, no functional change.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Use flags for representing the state in fuse_req. This is needed since
req->list will be protected by different locks in different states, hence
we'll want the state itself to be split into distinct bits, each protected
with the relevant lock in that state.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
FUSE_REQ_INIT is actually the same state as FUSE_REQ_PENDING and
FUSE_REQ_READING and FUSE_REQ_WRITING can be merged into a common
FUSE_REQ_IO state.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Only hold fc->lock over sections of request_wait_answer() that actually
need it. If wait_event_interruptible() returns zero, it means that the
request finished. Need to add memory barriers, though, to make sure that
all relevant data in the request is synchronized.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Since it's a 64bit counter, it's never gonna wrap around. Remove code
dealing with that possibility.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Splice fc->pending and fc->processing lists into a common kill list while
holding fc->lock.
By the time we release fc->lock, pending and processing lists are empty and
the io list contains only locked requests.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Finer grained locking will mean there's no single lock to protect
modification of bitfileds in fuse_req.
So move to using bitops. Can use the non-atomic variants for those which
happen while the request definitely has only one reference.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
- don't end the request while req->locked is true
- make unlock_request() return an error if the connection was aborted
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
fuse_abort_conn() does all the work done by fuse_dev_release() and more.
"More" consists of:
end_io_requests(fc);
wake_up_all(&fc->waitq);
kill_fasync(&fc->fasync, SIGIO, POLL_IN);
All of which should be no-op (WARN_ON's added).
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
And the same with fuse_request_send_nowait_locked().
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
fc->conn_error is set once in FUSE_INIT reply and never cleared. Check it
in request allocation, there's no sense in doing all the preparation if
sending will surely fail.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Move accounting of fc->num_waiting to the point where the request actually
starts waiting. This is earlier than the current queue_request() for
background requests, since they might be waiting on the fc->bg_queue before
being queued on fc->pending.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
Reset req->waiting in fuse_put_request(). This is needed for correct
accounting in fc->num_waiting for reserved requests.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
request_end() expects fc->num_background and fc->active_background to have
been incremented, which is not the case in fuse_request_send_nowait()
failure path. So instead just call the ->end() callback (which is actually
set by all callers).
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
struct kiocb now is a generic I/O container, so move it to fs.h.
Also do a #include diet for aio.h while we're at it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The misc subsystem (which is used for /dev/fuse) initializes private_data to
point to the misc device when a driver has registered a custom open file
operation, and initializes it to NULL when a custom open file operation has
*not* been provided.
This subtle quirk is confusing, to the point where kernel code registers
*empty* file open operations to have private_data point to the misc device
structure. And it leads to bugs, where the addition or removal of a custom open
file operation surprisingly changes the initial contents of a file's
private_data structure.
So to simplify things in the misc subsystem, a patch [1] has been proposed to
*always* set the private_data to point to the misc device, instead of only
doing this when a custom open file operation has been registered.
But before this patch can be applied we need to modify drivers that make the
assumption that a misc device file's private_data is initialized to NULL
because they didn't register a custom open file operation, so they don't rely
on this assumption anymore. FUSE uses private_data to store the fuse_conn and
errors out if this is not initialized to NULL at mount time.
Hence, we now set a file's private_data to NULL explicitly, to be independent
of whatever value the misc subsystem initializes it to by default.
[1] https://lkml.org/lkml/2014/12/4/939
Reported-by: Giedrius Statkevicius <giedriuswork@gmail.com>
Reported-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Tom Van Braeckel <tomvanbraeckel@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Regular pipe buffers' ->steal method (generic_pipe_buf_steal()) doesn't set
PG_uptodate.
Don't warn on this condition, just set the uptodate flag.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: stable@vger.kernel.org
fuse_try_move_page() is not prepared for replacing pages that have already
been read.
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: stable@vger.kernel.org
Theoretically we need to order setting of various fields in fc with
fc->initialized.
No known bug reports related to this yet.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Analysis from Marc:
"Commit 7078187a79 ("fuse: introduce fuse_simple_request() helper")
from the above pull request triggers some EIO errors for me in some tests
that rely on fuse
Looking at the code changes and a bit of debugging info I think there's a
general problem here that fuse_get_req checks and possibly waits for
fc->initialized, and this was always called first. But this commit
changes the ordering and in many places fc->minor is now possibly used
before fuse_get_req, and we can't be sure that fc has been initialized.
In my case fuse_lookup_init sets req->out.args[0].size to the wrong size
because fc->minor at that point is still 0, leading to the EIO error."
Fix by moving the compat adjustments into fuse_simple_request() to after
fuse_get_req().
This is also more readable than the original, since now compatibility is
handled in a single function instead of cluttering each operation.
Reported-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: Marc Dionne <marc.c.dionne@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Fixes: 7078187a79 ("fuse: introduce fuse_simple_request() helper")
The following pattern is repeated many times:
req = fuse_get_req_nopages(fc);
/* Initialize req->(in|out).args */
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(req);
Create a new replacement helper:
/* Initialize args */
err = fuse_simple_request(fc, &args);
In addition to reducing the code size, this will ease moving from the
complex arg-based to a simpler page-based I/O on the fuse device.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Pull fuse fixes from Miklos Szeredi:
"This contains miscellaneous fixes"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: replace count*size kzalloc by kcalloc
fuse: release temporary page if fuse_writepage_locked() failed
fuse: restructure ->rename2()
fuse: avoid scheduling while atomic
fuse: handle large user and group ID
fuse: inode: drop cast
fuse: ignore entry-timeout on LOOKUP_REVAL
fuse: timeout comparison fix
As reported by Richard Sharpe, an attempt to use fuse_notify_inval_entry()
triggers complains about scheduling while atomic:
BUG: scheduling while atomic: fuse.hf/13976/0x10000001
This happens because fuse_notify_inval_entry() attempts to allocate memory
with GFP_KERNEL, holding "struct fuse_copy_state" mapped by kmap_atomic().
Introduced by commit 58bda1da4b "fuse/dev: use atomic maps"
Fix by moving the map/unmap to just cover the actual memcpy operation.
Original patch from Maxim Patlasov <mpatlasov@parallels.com>
Reported-by: Richard Sharpe <realrichardsharpe@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: <stable@vger.kernel.org> # v3.15+
cold is a bool, make it one. Make the likely case the "if" part of the
block instead of the else as according to the optimisation manual this is
preferred.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jan Kara <jack@suse.cz>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
