08 Feb, 2011
4 commits
-
* git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable: (33 commits)
Btrfs: Fix page count calculation
btrfs: Drop __exit attribute on btrfs_exit_compress
btrfs: cleanup error handling in btrfs_unlink_inode()
Btrfs: exclude super blocks when we read in block groups
Btrfs: make sure search_bitmap finds something in remove_from_bitmap
btrfs: fix return value check of btrfs_start_transaction()
btrfs: checking NULL or not in some functions
Btrfs: avoid uninit variable warnings in ordered-data.c
Btrfs: catch errors from btrfs_sync_log
Btrfs: make shrink_delalloc a little friendlier
Btrfs: handle no memory properly in prepare_pages
Btrfs: do error checking in btrfs_del_csums
Btrfs: use the global block reserve if we cannot reserve space
Btrfs: do not release more reserved bytes to the global_block_rsv than we need
Btrfs: fix check_path_shared so it returns the right value
btrfs: check return value of btrfs_start_ioctl_transaction() properly
btrfs: fix return value check of btrfs_join_transaction()
fs/btrfs/inode.c: Add missing IS_ERR test
btrfs: fix missing break in switch phrase
btrfs: fix several uncheck memory allocations
... -
* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
cifs: remove checks for ses->status == CifsExiting
cifs: add check for kmalloc in parse_dacl
cifs: don't send an echo request unless NegProt has been done
cifs: enable signing flag in SMB header when server has it on
cifs: Possible slab memory corruption while updating extended stats (repost)
CIFS: Fix variable types in cifs_iovec_read/write (try #2)
cifs: fix length vs. total_read confusion in cifs_demultiplex_thread -
take offset of start position into account when calculating page count.
Signed-off-by: Yan, Zheng
Signed-off-by: Chris Mason -
ses->status is never set to CifsExiting, so these checks are
always false.Tested-by: JG
Signed-off-by: Jeff Layton
Signed-off-by: Steve French
06 Feb, 2011
5 commits
-
As this function is called in some error paths while not
removing the module, the __exit attribute prevents the kernel
image from linking when btrfs is compiled in statically.Signed-off-by: Alexey Charkov
Signed-off-by: Chris Mason -
When btrfs_alloc_path() fails, btrfs_free_path() need not be called.
Therefore, it changes the branch ahead.Signed-off-by: Tsutomu Itoh
Signed-off-by: Chris Mason -
This has been resulting in a BUT_ON(ret) after btrfs_reserve_extent in
btrfs_cow_file_range. The reason is we don't actually calculate the bytes_super
for a block group until we go to cache it, which means that the space_info can
hand out reservations for space that it doesn't actually have, and we can run
out of data space. This is also a problem if you are using space caching since
we don't ever calculate bytes_super for the block groups. So instead everytime
we read a block group call exclude_super_stripes, which calculates the
bytes_super for the block group so it can be left out of the space_info. Then
whenever caching completes we just call free_excluded_extents so that the super
excluded extents are freed up. Also if we are unmounting and we hit any block
groups that haven't been cached we still need to call free_excluded_extents to
make sure things are cleaned up properly. Thanks,Reported-by: Arne Jansen
Signed-off-by: Josef Bacik
Signed-off-by: Chris Mason -
When we're cleaning up the tree log we need to be able to remove free space from
the block group. The problem is if that free space spans bitmaps we would not
find the space since we're looking for too many bytes. So make sure the amount
of bytes we search for is limited to either the number of bytes we want, or the
number of bytes left in the bitmap. This was tested by a user who was hitting
the BUG() after search_bitmap. With this patch he can now mount his fs.
Thanks,Signed-off-by: Josef Bacik
Signed-off-by: Chris Mason -
Exit from parse_dacl if no memory returned from the call to kmalloc.
Signed-off-by: Stanislav Fomichev
Signed-off-by: Steve French
05 Feb, 2011
4 commits
-
When the socket to the server is disconnected, the client more or less
immediately calls cifs_reconnect to reconnect the socket. The NegProt
and SessSetup however are not done until an actual call needs to be
made.With the addition of the SMB echo code, it's possible that the server
will initiate a disconnect on an idle socket. The client will then
reconnect the socket but no NegotiateProtocol request is done. The
SMBEcho workqueue job will then eventually pop, and an SMBEcho will be
sent on the socket. The server will then reject it since no NegProt was
done.The ideal fix would be to either have the socket not be reconnected
until we plan to use it, or to immediately do a NegProt when the
reconnect occurs. The code is not structured for this however. For now
we must just settle for not sending any echoes until the NegProt is
done.Reported-by: JG
Signed-off-by: Jeff Layton
Signed-off-by: Steve French -
cifs_sign_smb only generates a signature if the correct Flags2 bit is
set. Make sure that it gets set correctly if we're sending an async
call.This patch fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=28142
Reported-and-Tested-by: JG
Signed-off-by: Jeff Layton
Signed-off-by: Steve French -
Updating extended statistics here can cause slab memory corruption
if a callback function frees slab memory (mid_entry).Signed-off-by: Shirish Pargaonkar
Reviewed-by: Jeff Layton
Signed-off-by: Steve French -
In get_empty_filp() since 2.6.29, file_free(f) is called with f->f_cred == NULL
when security_file_alloc() returned an error. As a result, kernel will panic()
due to put_cred(NULL) call within RCU callback.Fix this bug by assigning f->f_cred before calling security_file_alloc().
Signed-off-by: Tetsuo Handa
Signed-off-by: David Howells
Signed-off-by: Linus Torvalds
04 Feb, 2011
6 commits
-
Variable 'i' should be unsigned long as it's used in circle with num_pages,
and bytes_read/total_written should be ssize_t according to return value.Signed-off-by: Pavel Shilovsky
Reviewed-by: Shirish Pargaonkar
Signed-off-by: Steve French -
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hch/hfsplus:
hfsplus: fix up a comparism in hfsplus_file_extend
hfsplus: fix two memory leaks in wrapper.c
hfsplus: do not leak buffer on error
hfsplus: fix failed mount handling -
Revert an incorrect hunk from commit b2837fcf4994e699a4def002e26f274d95b387c1,
"hfsplus: %L-to-%ll, macro correction, and remove unneeded braces"
revert a pointless change of comparism operation argument order, which turned
out to not even be equivalent.Reported-by: Joe Perches
Signed-off-by: Christoph Hellwig -
Signed-Off-By: Chuck Ebbert
Signed-off-by: Christoph Hellwig -
Signed-Off-By: Chuck Ebbert
Signed-off-by: Christoph Hellwig -
Currently the error handling in hfsplus_fill_super is a mess, and can
lead to accessing fields in the superblock that haven't been even set
up yet. Fix this by making sure we do not set up sb->s_root until we
have the mount fully set up, and before that do proper step by step
unwinding instead of using hfsplus_put_super as a big hammer.Reported-by: Dan Williams
Signed-off-by: Christoph Hellwig
03 Feb, 2011
4 commits
-
This reverts commit 115e19c53501edc11f730191f7f047736815ae3d.
Apparently setting inode->bdi to one's own sb->s_bdi stops VFS from
sending *read-aheads*. This problem was bisected to this commit. A
revert fixes it. I'll investigate farther why is this happening for the
next Kernel, but for now a revert.I'm sending to stable@kernel.org as well, since it exists also in
2.6.37. 2.6.36 is good and does not have this patch.CC: Stable Tree
Signed-off-by: Boaz Harrosh
Signed-off-by: Linus Torvalds -
Some filesystems don't deal well with being asked to map less than
blocksize blocks (GFS2 for example). Since we are always mapping at least
blocksize sections anyway, just make sure len is at least as big as a
blocksize so we don't trip up any filesystems. Thanks,Signed-off-by: Josef Bacik
Cc: Steven Whitehouse
Cc: Christoph Hellwig
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
FMODE_EXEC is a constant type of fmode_t but was used with normal integer
constants. This results in following warnings from sparse. Fix it using
new macro __FMODE_EXEC.fs/exec.c:116:58: warning: restricted fmode_t degrades to integer
fs/exec.c:689:58: warning: restricted fmode_t degrades to integer
fs/fcntl.c:777:9: warning: restricted fmode_t degrades to integerSigned-off-by: Namhyung Kim
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
commit 95aac7b1cd224f ("epoll: make epoll_wait() use the hrtimer range
feature") added a performance regression because it uses timespec_add_ns()
with potential very large 'ns' values.[akpm@linux-foundation.org: s/epoll_set_mstimeout/ep_set_mstimeout/, per Davide]
Reported-by: Simon Kirby
Signed-off-by: Eric Dumazet
Cc: Shawn Bohrer
Acked-by: Davide Libenzi
Cc: [2.6.37.x]
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
02 Feb, 2011
2 commits
-
length at this point is the length returned by the last kernel_recvmsg
call. total_read is the length of all of the data read so far. length
is more or less meaningless at this point, so use total_read for
everything.Signed-off-by: Jeff Layton
Reviewed-by: Pavel Shilovsky
Signed-off-by: Steve French -
* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
cifs: fix length checks in checkSMB
[CIFS] Update cifs minor version
cifs: No need to check crypto blockcipher allocation
cifs: clean up some compiler warnings
cifs: make CIFS depend on CRYPTO_MD4
cifs: force a reconnect if there are too many MIDs in flight
cifs: don't pop a printk when sending on a socket is interrupted
cifs: simplify SMB header check routine
cifs: send an NT_CANCEL request when a process is signalled
cifs: handle cancelled requests better
cifs: fix two compiler warning about uninitialized vars
01 Feb, 2011
10 commits
-
The error check of btrfs_start_transaction() is added, and the mistake
of the error check on several places is corrected.Signed-off-by: Tsutomu Itoh
Signed-off-by: Chris Mason -
Because NULL is returned when the memory allocation fails,
it is checked whether it is NULL.Signed-off-by: Tsutomu Itoh
Signed-off-by: Chris Mason -
This one isn't really an uninit variable, but for pretty
obscure reasons. Let's make it clearly correct.Signed-off-by: Chris Mason
-
* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
NFS: NFSv4 readdir loses entries
NFS: Micro-optimize nfs4_decode_dirent()
NFS: Fix an NFS client lockdep issue
NFS construct consistent co_ownerid for v4.1
NFS: nfs_wcc_update_inode() should set nfsi->attr_gencount
NFS improve pnfs_put_deviceid_cache debug print
NFS fix cb_sequence error processing
NFS do not find client in NFSv4 pg_authenticate
NLM: Fix "kernel BUG at fs/lockd/host.c:417!" or ".../host.c:283!"
NFS: Prevent memory allocation failure in nfsacl_encode()
NFS: nfsacl_{encode,decode} should return signed integer
NFS: Fix "kernel BUG at fs/nfs/nfs3xdr.c:1338!"
NFS: Fix "kernel BUG at fs/aio.c:554!"
NFS4: Avoid potential NULL pointer dereference in decode_and_add_ds().
NFS: fix handling of malloc failure during nfs_flush_multi() -
The cERROR message in checkSMB when the calculated length doesn't match
the RFC1001 length is incorrect in many cases. It always says that the
RFC1001 length is bigger than the SMB, even when it's actually the
reverse.Fix the error message to say the reverse of what it does now when the
SMB length goes beyond the end of the received data. Also, clarify the
error message when the RFC length is too big. Finally, clarify the
comments to show that the 512 byte limit on extra data at the end of
the packet is arbitrary.Signed-off-by: Jeff Layton
Signed-off-by: Steve French -
* 'for-linus' of git://oss.sgi.com/xfs/xfs:
xfs: xfs_bmap_add_extent_delay_real should init br_startblock
xfs: fix dquot shaker deadlock
xfs: handle CIl transaction commit failures correctly
xfs: limit extsize to size of AGs and/or MAXEXTLEN
xfs: prevent extsize alignment from exceeding maximum extent size
xfs: limit extent length for allocation to AG size
xfs: speculative delayed allocation uses rounddown_power_of_2 badly
xfs: fix efi item leak on forced shutdown
xfs: fix log ticket leak on forced shutdown. -
Signed-off-by: Steve French
-
btrfs_sync_log returns -EAGAIN when we need full transaction commits
instead of small log commits, but sometimes we were dropping the return
value.In practice, we check for this a few different ways, but this is still a
bug that can leave off full log commits when we really need them.Signed-off-by: Chris Mason
-
Xfstests 224 will just sit there and spin for ever until eventually we give up
flushing delalloc and exit. On my box this took several hours. I could not
interrupt this process either, even though we use INTERRUPTIBLE. So do 2 things1) Keep us from looping over and over again without reclaiming anything
2) If we get interrupted exit the loopI tested this and the test now exits in a reasonable amount of time, and can be
interrupted with ctrl+c. Thanks,Signed-off-by: Josef Bacik
Signed-off-by: Chris Mason -
Missed one change as per earlier suggestion.
Signed-off-by: Shirish Pargaonkar
Reviewed-by: Jeff Layton
Signed-off-by: Steve French
31 Jan, 2011
5 commits
-
New compiler warnings that I noticed when building a patchset based
on recent Fedora kernel:fs/cifs/cifssmb.c: In function 'CIFSSMBSetFileSize':
fs/cifs/cifssmb.c:4813:8: warning: variable 'data_offset' set but not used
[-Wunused-but-set-variable]fs/cifs/file.c: In function 'cifs_open':
fs/cifs/file.c:349:24: warning: variable 'pCifsInode' set but not used
[-Wunused-but-set-variable]
fs/cifs/file.c: In function 'cifs_partialpagewrite':
fs/cifs/file.c:1149:23: warning: variable 'cifs_sb' set but not used
[-Wunused-but-set-variable]
fs/cifs/file.c: In function 'cifs_iovec_write':
fs/cifs/file.c:1740:9: warning: passing argument 6 of 'CIFSSMBWrite2' from
incompatible pointer type [enabled by default]
fs/cifs/cifsproto.h:337:12: note: expected 'unsigned int *' but argument is
of type 'size_t *'fs/cifs/readdir.c: In function 'cifs_readdir':
fs/cifs/readdir.c:767:23: warning: variable 'cifs_sb' set but not used
[-Wunused-but-set-variable]fs/cifs/cifs_dfs_ref.c: In function 'cifs_dfs_d_automount':
fs/cifs/cifs_dfs_ref.c:342:2: warning: 'rc' may be used uninitialized in
this function [-Wuninitialized]
fs/cifs/cifs_dfs_ref.c:278:6: note: 'rc' was declared hereSigned-off-by: Jeff Layton
Reviewed-by: Pavel Shilovsky
Signed-off-by: Steve French -
Recently CIFS was changed to use the kernel crypto API for MD4 hashes,
but the Kconfig dependencies were not changed to reflect this.Signed-off-by: Jeff Layton
Reported-and-Tested-by: Suresh Jayaraman
Reviewed-by: Shirish Pargaonkar
Signed-off-by: Steve French -
Currently, we allow the pending_mid_q to grow without bound with
SIGKILL'ed processes. This could eventually be a DoS'able problem. An
unprivileged user could a process that does a long-running call and then
SIGKILL it.If he can also intercept the NT_CANCEL calls or the replies from the
server, then the pending_mid_q could grow very large, possibly even to
2^16 entries which might leave GetNextMid in an infinite loop. Fix this
by imposing a hard limit of 32k calls per server. If we cross that
limit, set the tcpStatus to CifsNeedReconnect to force cifsd to
eventually reconnect the socket and clean out the pending_mid_q.While we're at it, clean up the function a bit and eliminate an
unnecessary NULL pointer check.Signed-off-by: Jeff Layton
Reviewed-by: Shirish Pargaonkar
Signed-off-by: Steve French -
If we kill the process while it's sending on a socket then the
kernel_sendmsg will return -EINTR. This is normal. No need to spam the
ring buffer with this info.Signed-off-by: Jeff Layton
Signed-off-by: Steve French -
...just cleanup. There should be no behavior change.
Signed-off-by: Jeff Layton
Reviewed-by: Pavel Shilovsky
Signed-off-by: Steve French