Quoting Linus (3 years ago, FUSE inclusion discussions):
"User-space filesystems are hard to get right. I'd claim that they
are almost impossible, unless you limit them somehow (shared
writable mappings are the nastiest part - if you don't have those,
you can reasonably limit your problems by limiting the number of
dirty pages you accept through normal "write()" calls)."
Instead of attempting the impossible, I've just waited for the dirty page
accounting infrastructure to materialize (thanks to Peter Zijlstra and
others). This nicely solved the biggest problem: limiting the number of pages
used for write caching.
Some small details remained, however, which this largish patch attempts to
address. It provides a page writeback implementation for fuse, which is
completely safe against VM related deadlocks. Performance may not be very
good for certain usage patterns, but generally it should be acceptable.
It has been tested extensively with fsx-linux and bash-shared-mapping.
Fuse page writeback design
--------------------------
fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM.
It copies the contents of the original page, and queues a WRITE request to the
userspace filesystem using this temp page.
The writeback is finished instantly from the MM's point of view: the page is
removed from the radix trees, and the PageDirty and PageWriteback flags are
cleared.
For the duration of the actual write, the NR_WRITEBACK_TEMP counter is
incremented. The per-bdi writeback count is not decremented until the actual
write completes.
On dirtying the page, fuse waits for a previous write to finish before
proceeding. This makes sure, there can only be one temporary page used at a
time for one cached page.
This approach is wasteful in both memory and CPU bandwidth, so why is this
complication needed?
The basic problem is that there can be no guarantee about the time in which
the userspace filesystem will complete a write. It may be buggy or even
malicious, and fail to complete WRITE requests. We don't want unrelated parts
of the system to grind to a halt in such cases.
Also a filesystem may need additional resources (particularly memory) to
complete a WRITE request. There's a great danger of a deadlock if that
allocation may wait for the writepage to finish.
Currently there are several cases where the kernel can block on page
writeback:
- allocation order is larger than PAGE_ALLOC_COSTLY_ORDER
- page migration
- throttle_vm_writeout (through NR_WRITEBACK)
- sync(2)
Of course in some cases (fsync, msync) we explicitly want to allow blocking.
So for these cases new code has to be added to fuse, since the VM is not
tracking writeback pages for us any more.
As an extra safetly measure, the maximum dirty ratio allocated to a single
fuse filesystem is set to 1% by default. This way one (or several) buggy or
malicious fuse filesystems cannot slow down the rest of the system by hogging
dirty memory.
With appropriate privileges, this limit can be raised through
'/sys/class/bdi/<bdi>/max_ratio'.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move the fields 'dentry' and 'vfsmount' into the request specific union, since
these are only used for the RELEASE request.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some open flags (O_APPEND, O_DIRECT) can be changed with fcntl(F_SETFL, ...)
after open, but fuse currently only sends the flags to userspace in open.
To make it possible to correcly handle changing flags, send the
current value to userspace in each read and write.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently reading a fuse file will stop at cached i_size and return
EOF, even though the file might have grown since the attributes were
last updated.
So detect if trying to read past EOF, and refresh the attributes
before continuing with the read.
Thanks to mpb for the report.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For mandatory locking the userspace filesystem needs to know the lock
ownership for read, write and truncate operations.
This patch adds the necessary fields to the protocol.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch adds a new helper function fuse_write_fill() which makes it
possible to send WRITE requests asynchronously.
A new flag for WRITE requests is also added which indicates that this a write
from the page cache, and not a "normal" file write.
This patch is in preparation for writable mmap support.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Each WRITE request must carry a valid file descriptor. When a page is written
back from a memory mapping, the file through which the page was dirtied is not
available, so a new mechananism is needed to find a suitable file in
->writepage(s).
A list of fuse_files is added to fuse_inode. The file is removed from the
list in fuse_release().
This patch is in preparation for writable mmap support.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It is trivial to add support for flock(2) semantics to the existing protocol,
by setting the lock owner field to the file pointer, and passing a new
FUSE_LK_FLOCK flag with the locking request.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch allows fuse filesystems to implement open(..., O_TRUNC) as a single
request, instead of separate truncate and open requests.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Getattr and lookup operations can be running in parallel to attribute changing
operations, such as write and setattr.
This means, that if for example getattr was slower than a write, the cached
size attribute could be set to a stale value.
To prevent this race, introduce a per-filesystem attribute version counter.
This counter is incremented whenever cached attributes are modified, and the
incremented value stored in the inode.
Before storing new attributes in the cache, getattr and lookup check, using
the version number, whether the attributes have been modified during the
request's lifetime. If so, the returned attributes are not cached, because
they might be stale.
Thanks to Jakub Bogusz for the bug report and test program.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: Jakub Bogusz <jakub.bogusz@gemius.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move dput/mntput pair from request_end() to fuse_release_end(), because
there's no other place they are used.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fuse_permission() didn't refresh inode attributes before using them, even if
the validity has already expired.
Thanks to Junjiro Okajima for spotting this.
Also remove some old code to unconditionally refresh the attributes on the
root inode.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Other than truncate, there are two cases, when fuse tries to get rid
of cached pages:
a) in open, if KEEP_CACHE flag is not set
b) in getattr, if file size changed spontaneously
Until now invalidate_mapping_pages() were used, which didn't get rid
of mapped pages. This is wrong, and becomes more wrong as dirty pages
are introduced. So instead properly invalidate all pages with
invalidate_inode_pages2().
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Make lifetime of 'struct fuse_file' independent from 'struct file' by adding a
reference counter and destructor.
This will enable asynchronous page writeback, where it cannot be guaranteed,
that the file is not released while a request with this file handle is being
served.
The actual RELEASE request is only sent when there are no more references to
the fuse_file.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[mszeredi]
- don't send zero length write requests
- it is not legal for the filesystem to return with zero written bytes
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
They can use generic_file_splice_read() instead. Since sys_sendfile() now
prefers that, there should be no change in behaviour.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
This fixes O_APPEND in direct IO mode. Also checks writes against file size
limits, notably rlimits.
Reported by Greg Bruno.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
First thing mm.h does is including sched.h solely for can_do_mlock() inline
function which has "current" dereference inside. By dealing with can_do_mlock()
mm.h can be detached from sched.h which is good. See below, why.
This patch
a) removes unconditional inclusion of sched.h from mm.h
b) makes can_do_mlock() normal function in mm/mlock.c
c) exports can_do_mlock() to not break compilation
d) adds sched.h inclusions back to files that were getting it indirectly.
e) adds less bloated headers to some files (asm/signal.h, jiffies.h) that were
getting them indirectly
Net result is:
a) mm.h users would get less code to open, read, preprocess, parse, ... if
they don't need sched.h
b) sched.h stops being dependency for significant number of files:
on x86_64 allmodconfig touching sched.h results in recompile of 4083 files,
after patch it's only 3744 (-8.3%).
Cross-compile tested on
all arm defconfigs, all mips defconfigs, all powerpc defconfigs,
alpha alpha-up
arm
i386 i386-up i386-defconfig i386-allnoconfig
ia64 ia64-up
m68k
mips
parisc parisc-up
powerpc powerpc-up
s390 s390-up
sparc sparc-up
sparc64 sparc64-up
um-x86_64
x86_64 x86_64-up x86_64-defconfig x86_64-allnoconfig
as well as my two usual configs.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
posix_test_lock() and ->lock() do the same job but have gratuitously
different interfaces. Modify posix_test_lock() so the two agree,
simplifying some code in the process.
Signed-off-by: Marc Eshel <eshel@almaden.ibm.com>
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Convert all calls to invalidate_inode_pages() into open-coded calls to
invalidate_mapping_pages().
Leave the invalidate_inode_pages() wrapper in place for now, marked as
deprecated.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The use by FUSE was just a remnant of an optimization from the time
when writable mappings were supported.
Now FUSE never actually allows the creation of dirty pages, so this
invocation of clear_page_dirty() is effectively a no-op.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add support for the BMAP operation for block device based filesystems. This
is needed to support swap-files and lilo.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This just ignore the remaining pages.
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Steven French <sfrench@us.ibm.com>
Cc: Miklos Szeredi <miklos@szeredi.hu>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fuse didn't always call i_size_write() with i_mutex held which caused rare
hangs on SMP/32bit. This bug has been present since fuse-2.2, well before
being merged into mainline.
The simplest solution is to protect i_size_write() with the per-connection
spinlock. Using i_mutex for this purpose would require some restructuring of
the code and I'm not even sure it's always safe to acquire i_mutex in all
places i_size needs to be set.
Since most of vmtruncate is already duplicated for other reasons, duplicate
the remaining part as well, making all i_size_write() calls internal to fuse.
Using i_size_write() was unnecessary in fuse_init_inode(), since this function
is only called on a newly created locked inode.
Reported by a few people over the years, but special thanks to Dana Henriksen
who was persistent enough in helping me debug it.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch cleans up generic_file_*_read/write() interfaces. Christoph
Hellwig gave me the idea for this clean ups.
In a nutshell, all filesystems should set .aio_read/.aio_write methods and use
do_sync_read/ do_sync_write() as their .read/.write methods. This allows us
to cleanup all variants of generic_file_* routines.
Final available interfaces:
generic_file_aio_read() - read handler
generic_file_aio_write() - write handler
generic_file_aio_write_nolock() - no lock write handler
__generic_file_aio_write_nolock() - internal worker routine
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't let fuse_readpages leave the @pages list not empty when exiting
on error.
[akpm@osdl.org: kernel-doc fixes]
Signed-off-by: Alexander Zarochentsev <zam@namesys.com>
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Same as with already do with the file operations: keep them in .rodata and
prevents people from doing runtime patching.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Steven French <sfrench@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
VFS uses current->files pointer as lock owner ID, and it wouldn't be
prudent to expose this value to userspace. So scramble it with XTEA using
a per connection random key, known only to the kernel. Only one direction
needs to be implemented, since the ID is never sent in the reverse
direction.
The XTEA algorithm is implemented inline since it's simple enough to do so,
and this adds less complexity than if the crypto API were used.
Thanks to Jesper Juhl for the idea.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add synchronous request interruption. This is needed for file locking
operations which have to be interruptible. However filesystem may implement
interruptibility of other operations (e.g. like NFS 'intr' mount option).
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
All POSIX locks owned by the current task are removed on close(). If the
FLUSH request resulting initiated by close() fails to reach userspace, there
might be locks remaining, which cannot be removed.
The only reason it could fail, is if allocating the request fails. In this
case use the request reserved for RELEASE, or if that is currently used by
another FLUSH, wait for it to become available.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch adds POSIX file locking support to the fuse interface.
This implementation doesn't keep any locking state in kernel. Unlocking on
close() is handled by the FLUSH message, which now contains the lock owner id.
Mandatory locking is not supported. The filesystem may enfoce mandatory
locking in userspace if needed.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't put requests into the background when a fatal interrupt occurs while the
request is in userspace. This removes a major wart from the implementation.
Backgrounding of requests was introduced to allow breaking of deadlocks.
However now the same can be achieved by aborting the filesystem through the
'abort' sysfs attribute.
This is a change in the interface, but should not cause problems, since these
kinds of deadlocks never happen during normal operation.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Pass the POSIX lock owner ID to the flush operation.
This is useful for filesystems which don't want to store any locking state
in inode->i_flock but want to handle locking/unlocking POSIX locks
internally. FUSE is one such filesystem but I think it possible that some
network filesystems would need this also.
Also add a flag to indicate that a POSIX locking request was generated by
close(), so filesystems using the above feature won't send an extra locking
request in this case.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It's cleaner to allocate a new request, otherwise the uid/gid/pid
fields of the request won't be filled in.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
FUSE allocated most requests from a fixed size pool filled at mount time.
However in some cases (release/forget) non-pool requests were used. File
locking operations aren't well served by the request pool, since they may
block indefinetly thus exhausting the pool.
This patch removes the request pool and always allocates requests on demand.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
During heavy parallel filesystem activity it was possible to Oops the kernel.
The reason is that read_cache_pages() could skip pages which have already been
inserted into the cache by another task. Occasionally this may result in zero
pages actually being sent, while fuse_send_readpages() relies on at least one
page being in the request.
So check this corner case and just free the request instead of trying to send
it.
Reported and tested by Konstantin Isakov.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a conversion to make the various file_operations structs in fs/
const. Basically a regexp job, with a few manual fixups
The goal is both to increase correctness (harder to accidentally write to
shared datastructures) and reducing the false sharing of cachelines with
things that get dirty in .data (while .rodata is nicely read only and thus
cache clean)
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There's a rather theoretical case of the BUG triggering in
fuse_reset_request():
- iget() fails because of OOM after a successful CREATE_OPEN request
- during IO on the resulting RELEASE request the connection is aborted
Fix and add warning to fuse_reset_request().
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
While asynchronous reads mean a performance improvement in most cases, if
the filesystem assumed that reads are synchronous, then async reads may
degrade performance (filesystem may receive reads out of order, which can
confuse it's own readahead logic).
With sshfs a 1.5 to 4 times slowdown can be measured.
There's also a need for userspace filesystems to know whether asynchronous
reads are supported by the kernel or not.
To achive these, negotiate in the INIT request whether async reads will be
used and the maximum readahead value. Update interface version to 7.6
If userspace uses a version earlier than 7.6, then disable async reads, and
set maximum readahead value to the maximum read size, as done in previous
versions.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch changes fuse_readpages() to send READ requests asynchronously.
This makes it possible for userspace filesystems to utilize the kernel
readahead logic instead of having to implement their own (resulting in double
caching).
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add a separate function for filling in the READ request. This will make it
possible to send asynchronous READ requests as well as synchronous ones.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add possibility for requests to run asynchronously and call an 'end' callback
when finished.
With this, the special handling of the INIT and RELEASE requests can be
cleaned up too.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Inline keyword is unnecessary in most cases. Clean them up.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch converts the inode semaphore to a mutex. I have tested it on
XFS and compiled as much as one can consider on an ia64. Anyway your
luck with it might be different.
Modified-by: Ingo Molnar <mingo@elte.hu>
(finished the conversion)
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
In direct_io mode, send at least one page per reqest. Previously it was
possible that reqests with zero data were sent, and hence the read/write
didn't make any progress, resulting in an infinite (though interruptible)
loop.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make file operations on a bad inode fail. This just makes things a
bit more consistent.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch adds an atomic create+open operation. This does not yet work if
the file type changes between lookup and create+open, but solves the
permission checking problems for the separte create and open methods.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
