This patch is based on Dan's original patch. His original description is
below:
Smatch complained about a couple checking for NULL after dereferencing
bugs. I'm not super familiar with the code so I did the conservative
thing and move the dereferences after the checks.
The dereferences in cifs_lock() and cifs_fsync() were added in
ba00ba64cf "cifs: make various routines use the cifsFileInfo->tcon
pointer". The dereference in find_writable_file() was added in
6508d904e6 "cifs: have find_readable/writable_file filter by fsuid".
The comments there say it's possible to trigger the NULL dereference
under stress.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Simplify many places when we need to set oplock level on an inode.
Signed-off-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Stanse found that pSMBFile in cifs_ioctl and file->f_path.dentry in
cifs_user_write are dereferenced prior their test to NULL.
The alternative is not to dereference them before the tests. The patch is
to point out the problem, you have to decide.
While at it we cache the inode in cifs_user_write to a local variable
and use all over the function.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Cc: Steve French <sfrench@samba.org>
Cc: linux-cifs@vger.kernel.org
Cc: Jeff Layton <jlayton@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Steve French <sfrench@us.ibm.com>
* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
cifs: Cleanup and thus reduce smb session structure and fields used during authentication
NTLM auth and sign - Use appropriate server challenge
cifs: add kfree() on error path
NTLM auth and sign - minor error corrections and cleanup
NTLM auth and sign - Use kernel crypto apis to calculate hashes and smb signatures
NTLM auth and sign - Define crypto hash functions and create and send keys needed for key exchange
cifs: cifs_convert_address() returns zero on error
NTLM auth and sign - Allocate session key/client response dynamically
cifs: update comments - [s/GlobalSMBSesLock/cifs_file_list_lock/g]
cifs: eliminate cifsInodeInfo->write_behind_rc (try #6)
[CIFS] Fix checkpatch warnings and bump cifs version number
cifs: wait for writeback to complete in cifs_flush
cifs: convert cifsFileInfo->count to non-atomic counter
We leak 256 bytes here on this error path.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
This removes more dead code that was somehow missed by commit 0d99519efe
(writeback: remove unused nonblocking and congestion checks). There are
no behavior change except for the removal of two entries from one of the
ext4 tracing interface.
The nonblocking checks in ->writepages are no longer used because the
flusher now prefer to block on get_request_wait() than to skip inodes on
IO congestion. The latter will lead to more seeky IO.
The nonblocking checks in ->writepage are no longer used because it's
redundant with the WB_SYNC_NONE check.
We no long set ->nonblocking in VM page out and page migration, because
a) it's effectively redundant with WB_SYNC_NONE in current code
b) it's old semantic of "Don't get stuck on request queues" is mis-behavior:
that would skip some dirty inodes on congestion and page out others, which
is unfair in terms of LRU age.
Inspired by Christoph Hellwig. Thanks!
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: David Howells <dhowells@redhat.com>
Cc: Sage Weil <sage@newdream.net>
Cc: Steve French <sfrench@samba.org>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
GlobalSMBSesLock is now cifs_file_list_lock. Update comments to reflect this.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
write_behind_rc is redundant and just adds complexity to the code. What
we really want to do instead is to use mapping_set_error to reset the
flags on the mapping when we find a writeback error and can't report it
to userspace yet.
For cifs_flush and cifs_fsync, we shouldn't reset the flags since errors
returned there do get reported to userspace.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Reviewed-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
The f_op->flush operation is the last chance to return a writeback
related error when closing a file. Ensure that we don't miss reporting
any errors by waiting for writeback to complete in cifs_flush before
proceeding.
There's no reason to do this when the file isn't open for write
however.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Reviewed-by: David Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
The count for cifsFileInfo is currently an atomic, but that just adds
complexity for little value. We generally need to hold cifs_file_list_lock
to traverse the lists anyway so we might as well make this counter
non-atomic and simply use the cifs_file_list_lock to protect it.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Now that it's feasible for a cifsFileInfo to outlive the filp under
which it was created, move the close processing into cifsFileInfo_put.
This means that the last user of the filehandle always does the actual
on the wire close call. This also allows us to get rid of the closePend
flag from cifsFileInfo. If we have an active reference to the file
then it's never going to have a close pending.
cifs_close is converted to simply put the filehandle.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
...and make it non-inlined in preparation for the move of most of
cifs_close to it.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Convert this lock to a regular spinlock
A rwlock_t offers little value here. It's more expensive than a regular
spinlock unless you have a fairly large section of code that runs under
the read lock and can benefit from the concurrency.
Additionally, we need to ensure that the refcounting for files isn't
racy and to do that we need to lock areas that can increment it for
write. That means that the areas that can actually use a read_lock are
very few and relatively infrequently used.
While we're at it, change the name to something easier to type, and fix
a bug in find_writable_file. cifsFileInfo_put can sleep and shouldn't be
called while holding the lock.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
It's currently in dir.c which makes little sense...
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
All the remaining users of cifsFileInfo->pfile just use it to get
at the f_flags/f_mode. Now that we store that separately in the
cifsFileInfo, there's no need to consult the pfile at all from
a cifsFileInfo pointer.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Have cifs_write take a cifsFileInfo pointer instead of a filp. Since
cifsFileInfo holds references on the dentry, and that holds one to
the inode, we can eliminate some unneeded NULL pointer checks.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Add a f_flags field that holds the f_flags field from the filp. We'll
need this info in case the filp ever goes away before the cifsFileInfo
does. Have cifs_reopen_file use that value instead of filp->f_flags
too and have it take a cifsFileInfo arg instead of a filp.
While we're at it, get rid of some bogus cargo-cult NULL pointer
checks in that function and reduce the level of indentation.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
It already takes a file pointer. The inode associated with that had damn
well better be the same one we're passing in anyway. Thus, there's no
need for a separate argument here.
Also, get rid of the bogus check for a null pCifsInode pointer. The
CIFS_I macro uses container_of(), and that will virtually never return a
NULL pointer anyway.
Finally, move the setting of the canCache* flags outside of the lock.
Other places in the code don't hold that lock when setting it, so I
assume it's not really needed here either.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Eliminate the poor, misunderstood "oflags" option from cifs_new_fileinfo.
The callers mostly pass in the filp->f_flags here.
That's not correct however since we're checking that value for
the presence of FMODE_READ. Luckily that only affects how the f_list is
ordered. What it really wants here is the file->f_mode. Just use that
field from the filp to determine it.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
The way flags are passed and converted for cifs_posix_open is rather
non-sensical. Some callers call cifs_posix_convert_flags on the flags
before they pass them to cifs_posix_open, whereas some don't. Two flag
conversion steps is just confusing though.
Change the function instead to clearly expect input in f_flags format,
and fix the callers to pass that in. Then, have cifs_posix_open call
cifs_convert_posix_flags to do the conversion. Move cifs_posix_open to
file.c as well so we can keep cifs_convert_posix_flags as a static
function.
Fix it also to not ignore O_CREAT, O_EXCL and O_TRUNC, and instead have
cifs_reopen_file mask those bits off before calling cifs_posix_open.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
cifs: eliminate cifs_posix_open_inode_helper
This function is redundant. The only thing it does is set the canCache
flags, but those get set in cifs_new_fileinfo anyway.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Filesystems aren't really supposed to do anything with a vfsmount. It's
considered a layering violation since vfsmounts are entirely managed at
the VFS layer.
CIFS currently keeps an active reference to a vfsmount in order to
prevent the superblock vanishing before an oplock break has completed.
What we really want to do instead is to keep sb->s_active high until the
oplock break has completed. This patch borrows the scheme that NFS uses
for handling sillyrenames.
An atomic_t is added to the cifs_sb_info. When it transitions from 0 to
1, an extra reference to the superblock is taken (by bumping the
s_active value). When it transitions from 1 to 0, that reference is
dropped and a the superblock teardown may proceed if there are no more
references to it.
Also, the vfsmount pointer is removed from cifsFileInfo and from
cifs_new_fileinfo, and some bogus forward declarations are removed from
cifsfs.h.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
cifsFileInfo is a bit problematic. It contains a reference back to the
struct file itself. This makes it difficult for a cifsFileInfo to exist
without a corresponding struct file.
It would be better instead of the cifsFileInfo just held info pertaining
to the open file on the server instead without any back refrences to the
struct file. This would allow it to exist after the filp to which it was
originally attached was closed.
Much of the use of the file pointer in this struct is to get at the
dentry. Begin divorcing the cifsFileInfo from the struct file by
keeping a reference to the dentry. Since the dentry will have a
reference to the inode, we can eliminate the "pInode" field too and
convert the igrab/iput to dget/dput.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
When we implement multiuser mounts, we'll need to filter filehandles
by fsuid. Add a flag for multiuser mounts and code to filter by
fsuid when it's set.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
cifsFileInfo needs a pointer to a tcon, but it doesn't currently hold a
reference to it. Change it to keep a pointer to a tcon_link instead and
hold a reference to it.
That will keep the tcon from being freed until the file is closed.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Eventually, we'll need to track the use of tcons on a per-sb basis, so that
we know when it's ok to tear them down. Begin this conversion by adding a
new "tcon_link" struct and accessors that get it. For now, the core data
structures are untouched -- cifs_sb still just points to a single tcon and
the pointers are just cast to deal with the accessor functions. A later
patch will flesh this out.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Get a reference to the file early so we can eventually base the decision
about signing on the correct tcon. If that doesn't work for some reason,
then fall back to generic_writepages. That's just as likely to fail, but
it simplifies the error handling.
In truth, I'm not sure how that could occur anyway, so maybe a NULL
open_file here ought to be a BUG()?
After that, we drop the reference to the open_file and then we re-get
one prior to each WriteAndX call. This helps ensure that the filehandle
isn't held open any longer than necessary and that open files are
reclaimed prior to each write call.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
To minimize calls to cifs_sb_tcon and to allow for a clear error path if
a tcon can't be acquired.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
When we convert cifs to do multiple sessions per mount, we'll need more
than one tcon per superblock. At that point "cifs_sb->tcon" will make
no sense. Add a new accessor function that gets a tcon given a cifs_sb.
For now, it just returns cifs_sb->tcon. Later it'll do more.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
cifs_new_fileinfo() does not use the 'oplock' value from the callers. Instead,
it sets it to REQ_OPLOCK which seems wrong. We should be using the oplock value
obtained from the Server to set the inode's clientCanCacheAll or
clientCanCacheRead flags. Fix this by passing oplock from the callers to
cifs_new_fileinfo().
This change dates back to commit a6ce4932 (2.6.30-rc3). So, all the affected
versions will need this fix. Please Cc stable once reviewed and accepted.
Cc: Stable <stable@kernel.org>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
... and avoid implicit casting from a signed type. Also, pass oplock by value
instead by reference as we don't intend to change the value in
cifs_open_inode_helper().
Thanks to Jeff Layton for spotting this.
Reviewed-by: Jeff Layton <jlayton@samba.org>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
cifs has a lot of complicated functions that have to clean up things on
error, but some of them don't have all of the cleanup code
well-consolidated. Clean up and consolidate error handling in several
functions.
This is in preparation of later patches that will need to put references
to the tcon link container.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: (55 commits)
workqueue: mark init_workqueues() as early_initcall()
workqueue: explain for_each_*cwq_cpu() iterators
fscache: fix build on !CONFIG_SYSCTL
slow-work: kill it
gfs2: use workqueue instead of slow-work
drm: use workqueue instead of slow-work
cifs: use workqueue instead of slow-work
fscache: drop references to slow-work
fscache: convert operation to use workqueue instead of slow-work
fscache: convert object to use workqueue instead of slow-work
workqueue: fix how cpu number is stored in work->data
workqueue: fix mayday_mask handling on UP
workqueue: fix build problem on !CONFIG_SMP
workqueue: fix locking in retry path of maybe_create_worker()
async: use workqueue for worker pool
workqueue: remove WQ_SINGLE_CPU and use WQ_UNBOUND instead
workqueue: implement unbound workqueue
workqueue: prepare for WQ_UNBOUND implementation
libata: take advantage of cmwq and remove concurrency limitations
workqueue: fix worker management invocation without pending works
...
Fixed up conflicts in fs/cifs/* as per Tejun. Other trivial conflicts in
include/linux/workqueue.h, kernel/trace/Kconfig and kernel/workqueue.c
Read pages from a FS-Cache data storage object into a CIFS inode.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Store pages from an CIFS inode into the data storage object associated with
that inode.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Takes care of invalidation and release of FS-Cache marked pages and also
invalidation of the FsCache page flag when the inode is removed.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Define inode-level data storage objects (managed by cifsInodeInfo structs).
Each inode-level object is created in a super-block level object and is itself
a data storage object in to which pages from the inode are stored.
The inode object is keyed by UniqueId. The coherency data being used is
LastWriteTime, LastChangeTime and end of file reported by the server.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Workqueue can now handle high concurrency. Use system_nrt_wq
instead of slow-work.
* Updated is_valid_oplock_break() to not call cifs_oplock_break_put()
as advised by Steve French. It might cause deadlock. Instead,
reference is increased after queueing succeeded and
cifs_oplock_break() briefly grabs GlobalSMBSeslock before putting
the cfile to make sure it doesn't put before the matching get is
finished.
* Anton Blanchard reported that cifs conversion was using now gone
system_single_wq. Use system_nrt_wq which provides non-reentrance
guarantee which is enough and much better.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Steve French <sfrench@samba.org>
Cc: Anton Blanchard <anton@samba.org>
It's currently possible for cifs_open to fail after it has already
called cifs_new_fileinfo. In that situation, the new fileinfo will be
leaked as the caller doesn't call fput. That in turn leads to a busy
inodes after umount problem since the fileinfo holds an extra inode
reference now. Shuffle cifs_open around a bit so that it only calls
cifs_new_fileinfo if it's going to succeed.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-and-Tested-by: Suresh Jayaraman <sjayaraman@suse.de>
...and ensure that we propagate the error back to avoid any surprises.
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Reviewed-and-Tested-by: Jeff Layton <jlayton@redhat.com>
...which takes a ton of unneeded arguments and does a lot more pointer
dereferencing than is really needed.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-and-Tested-by: Suresh Jayaraman <sjayaraman@suse.de>
The current scheme of sticking open files on a list and assuming that
cifs_open will scoop them off of it is broken and leads to "Busy
inodes after umount..." errors at unmount time.
The problem is that there is no guarantee that cifs_open will always
be called after a ->lookup or ->create operation. If there are
permissions or other problems, then it's quite likely that it *won't*
be called.
Fix this by fully instantiating the filp whenever the file is created
and pass that filp back to the VFS. If there is a problem, the VFS
can clean up the references.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-and-Tested-by: Suresh Jayaraman <sjayaraman@suse.de>
Having cifs_posix_open call cifs_new_fileinfo is problematic and
inconsistent with how "regular" opens work. It's also buggy as
cifs_reopen_file calls this function on a reconnect, which creates a new
struct cifsFileInfo that just gets leaked.
Push it out into the callers. This also allows us to get rid of the
"mnt" arg to cifs_posix_open.
Finally, in the event that a cifsFileInfo isn't or can't be created, we
always want to close the filehandle out on the server as the client
won't have a record of the filehandle and can't actually use it. Make
sure that CIFSSMBClose is called in those cases.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-and-Tested-by: Suresh Jayaraman <sjayaraman@suse.de>
Commit 315e995c63 is causing OOM kills
when stress-testing a CIFS filesystem. The VFS readpages operation takes
a page reference. The older code just handed this reference off to the
page cache, but the new code takes an extra one. The simplest fix is to
put the new reference after add_to_page_cache_lru.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
..a left over from the commit 3321b791b2.
Cc: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <sfrench@us.ibm.com>
On lease break we were breaking to readonly leases always
even if write requested. Also removed experimental
ifdef around setlease code
Signed-off-by: Steve French <sfrench@us.ibm.com>
