platform/kernel/linux-stable.git
11 years agoMerge tag 'usb-3.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Tue, 2 Jul 2013 18:31:09 +0000 (11:31 -0700)]
Merge tag 'usb-3.11-rc1' of git://git./linux/kernel/git/gregkh/usb

Pull USB updates from Greg KH:
 "Here's the big USB 3.11-rc1 merge request.

  Lots of gadget and finally, chipidea driver updates (they were much
  needed), along with a new host controller driver, lots of little
  serial driver fixes, the removal of the 255 usb-serial device
  limitation, and a variety of other minor things.

  All of these have been in the linux-next releases for a while"

* tag 'usb-3.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (254 commits)
  usb: musb: omap2430: make it compile again
  usb: chipidea: ci_hdrc_imx: access phy via private data
  xhci: Add missing unlocks on error paths
  USB: option,qcserial: move Novatel Gobi1K IDs to qcserial
  ehci-atmel.c: prepare clk before calling enable
  USB: ohci-at91: prepare clk before calling enable
  USB: HWA: fix device probe failure
  wusbcore: add entries in Documentation/ABI for new wusbhc sysfs attributes
  wusbcore: add sysfs attribute for retry count
  wusbcore: add sysfs attribute for DNTS count and interval
  usb: chipidea: drop "13xxx" infix
  usb: phy: tegra: remove duplicated include from phy-tegra-usb.c
  usb: host: xhci-plat: release mem region while removing module
  usbmisc_imx: allow autoloading on according to dt ids
  usb: fix build error without CONFIG_USB_PHY
  usb: check usb_hub_to_struct_hub() return value
  xhci: check for failed dma pool allocation
  usb: gadget: f_subset: fix missing unlock on error in geth_alloc()
  usb: gadget: f_ncm: fix missing unlock on error in ncm_alloc()
  usb: gadget: f_ecm: fix missing unlock on error in ecm_alloc()
  ...

11 years agoMerge tag 'fscache-20130702' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowell...
Linus Torvalds [Tue, 2 Jul 2013 16:52:47 +0000 (09:52 -0700)]
Merge tag 'fscache-20130702' of git://git./linux/kernel/git/dhowells/linux-fs

Pull FS-Cache updates from David Howells:
 "This contains a number of fixes for various FS-Cache issues plus some
  cleanups.  The commits are, in order:

   1) Provide a system wait_on_atomic_t() and wake_up_atomic_t() sharing
      the bit-wait table (enhancement for #8).

   2) Don't put spin_lock() in a while-condition as spin_lock() may have
      a do {} while(0) wrapper (cleanup).

   3) Symbolically name i_mutex lock classes rather than using numbers
      in CacheFiles (cleanup).

   4) Don't sleep in page release if __GFP_FS is not set (deadlock vs
      ext4).

   5) Uninline fscache_object_init() (cleanup for #7).

   6) Wrap checks on object state (cleanup for #7).

   7) Simplify the object state machine by separating work states from
      wait states.

   8) Simplify cookie retention by objects (NULL pointer deref fix).

   9) Remove unused list_to_page() macro (cleanup).

  10) Make the remaining-pages counter in the retrieval op atomic
      (assertion failure fix).

  11) Don't use spin_is_locked() in assertions (assertion failure fix)"

* tag 'fscache-20130702' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  FS-Cache: Don't use spin_is_locked() in assertions
  FS-Cache: The retrieval remaining-pages counter needs to be atomic_t
  cachefiles: remove unused macro list_to_page()
  FS-Cache: Simplify cookie retention for fscache_objects, fixing oops
  FS-Cache: Fix object state machine to have separate work and wait states
  FS-Cache: Wrap checks on object state
  FS-Cache: Uninline fscache_object_init()
  FS-Cache: Don't sleep in page release if __GFP_FS is not set
  CacheFiles: name i_mutex lock class explicitly
  fs/fscache: remove spin_lock() from the condition in while()
  Add wait_on_atomic_t() and wake_up_atomic_t()

11 years agoMerge tag 'dlm-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm
Linus Torvalds [Tue, 2 Jul 2013 16:52:07 +0000 (09:52 -0700)]
Merge tag 'dlm-3.11' of git://git./linux/kernel/git/teigland/linux-dlm

Pull dlm updates from David Teigland:
 "This set includes a number of SCTP related fixes in the dlm, and a few
  other minor fixes and changes."

* tag 'dlm-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm:
  dlm: Avoid LVB truncation
  dlm: log an error for unmanaged lockspaces
  dlm: config: using strlcpy instead of strncpy
  dlm: remove duplicated include from lowcomms.c
  dlm: disable nagle for SCTP
  dlm: retry failed SCTP sends
  dlm: try other IPs when sctp init assoc fails
  dlm: clear correct bit during sctp init failure handling
  dlm: set sctp assoc id during setup
  dlm: clear correct init bit during sctp setup

11 years agoMerge tag 'for-f2fs-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk...
Linus Torvalds [Tue, 2 Jul 2013 16:42:38 +0000 (09:42 -0700)]
Merge tag 'for-f2fs-3.11' of git://git./linux/kernel/git/jaegeuk/f2fs

Pull f2fs updates from Jaegeuk Kim:
 "This patch-set includes the following major enhancement patches:
   - remount_fs callback function
   - restore parent inode number to enhance the fsync performance
   - xattr security labels
   - reduce the number of redundant lock/unlock data pages
   - avoid frequent write_inode calls

  The other minor bug fixes are as follows.
   - endian conversion bugs
   - various bugs in the roll-forward recovery routine"

* tag 'for-f2fs-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (56 commits)
  f2fs: fix to recover i_size from roll-forward
  f2fs: remove the unused argument "sbi" of func destroy_fsync_dnodes()
  f2fs: remove reusing any prefree segments
  f2fs: code cleanup and simplify in func {find/add}_gc_inode
  f2fs: optimize the init_dirty_segmap function
  f2fs: fix an endian conversion bug detected by sparse
  f2fs: fix crc endian conversion
  f2fs: add remount_fs callback support
  f2fs: recover wrong pino after checkpoint during fsync
  f2fs: optimize do_write_data_page()
  f2fs: make locate_dirty_segment() as static
  f2fs: remove unnecessary parameter "offset" from __add_sum_entry()
  f2fs: avoid freqeunt write_inode calls
  f2fs: optimise the truncate_data_blocks_range() range
  f2fs: use the F2FS specific flags in f2fs_ioctl()
  f2fs: sync dir->i_size with its block allocation
  f2fs: fix i_blocks translation on various types of files
  f2fs: set sb->s_fs_info before calling parse_options()
  f2fs: support xattr security labels
  f2fs: fix iget/iput of dir during recovery
  ...

11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw
Linus Torvalds [Tue, 2 Jul 2013 16:41:18 +0000 (09:41 -0700)]
Merge git://git./linux/kernel/git/steve/gfs2-3.0-nmw

Pull GFS2 updates from Steven Whitehouse:
 "There are a few bug fixes for various, mostly very minor corner cases,
  plus some interesting new features.

  The new features include atomic_open whose main benefit will be the
  reduction in locking overhead in case of combined lookup/create and
  open operations, sorting the log buffer lists by block number to
  improve the efficiency of AIL writeback, and aggressively issuing
  revokes in gfs2_log_flush to reduce overhead when dropping glocks."

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw:
  GFS2: Reserve journal space for quota change in do_grow
  GFS2: Fix fstrim boundary conditions
  GFS2: fix warning message
  GFS2: aggressively issue revokes in gfs2_log_flush
  GFS2: fix regression in dir_double_exhash
  GFS2: Add atomic_open support
  GFS2: Only do one directory search on create
  GFS2: fix error propagation in init_threads()
  GFS2: Remove no-op wrapper function
  GFS2: Cocci spatch "ptr_ret.spatch"
  GFS2: Eliminate gfs2_rg_lops
  GFS2: Sort buffer lists by inplace block number

11 years agoMerge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Tue, 2 Jul 2013 16:39:34 +0000 (09:39 -0700)]
Merge tag 'ext4_for_linus' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 update from Ted Ts'o:
 "Lots of bug fixes, cleanups and optimizations.  In the bug fixes
  category, of note is a fix for on-line resizing file systems where the
  block size is smaller than the page size (i.e., file systems 1k blocks
  on x86, or more interestingly file systems with 4k blocks on Power or
  ia64 systems.)

  In the cleanup category, the ext4's punch hole implementation was
  significantly improved by Lukas Czerner, and now supports bigalloc
  file systems.  In addition, Jan Kara significantly cleaned up the
  write submission code path.  We also improved error checking and added
  a few sanity checks.

  In the optimizations category, two major optimizations deserve
  mention.  The first is that ext4_writepages() is now used for
  nodelalloc and ext3 compatibility mode.  This allows writes to be
  submitted much more efficiently as a single bio request, instead of
  being sent as individual 4k writes into the block layer (which then
  relied on the elevator code to coalesce the requests in the block
  queue).  Secondly, the extent cache shrink mechanism, which was
  introduce in 3.9, no longer has a scalability bottleneck caused by the
  i_es_lru spinlock.  Other optimizations include some changes to reduce
  CPU usage and to avoid issuing empty commits unnecessarily."

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (86 commits)
  ext4: optimize starting extent in ext4_ext_rm_leaf()
  jbd2: invalidate handle if jbd2_journal_restart() fails
  ext4: translate flag bits to strings in tracepoints
  ext4: fix up error handling for mpage_map_and_submit_extent()
  jbd2: fix theoretical race in jbd2__journal_restart
  ext4: only zero partial blocks in ext4_zero_partial_blocks()
  ext4: check error return from ext4_write_inline_data_end()
  ext4: delete unnecessary C statements
  ext3,ext4: don't mess with dir_file->f_pos in htree_dirblock_to_tree()
  jbd2: move superblock checksum calculation to jbd2_write_superblock()
  ext4: pass inode pointer instead of file pointer to punch hole
  ext4: improve free space calculation for inline_data
  ext4: reduce object size when !CONFIG_PRINTK
  ext4: improve extent cache shrink mechanism to avoid to burn CPU time
  ext4: implement error handling of ext4_mb_new_preallocation()
  ext4: fix corruption when online resizing a fs with 1K block size
  ext4: delete unused variables
  ext4: return FIEMAP_EXTENT_UNKNOWN for delalloc extents
  jbd2: remove debug dependency on debug_fs and update Kconfig help text
  jbd2: use a single printk for jbd_debug()
  ...

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Tue, 2 Jul 2013 16:28:37 +0000 (09:28 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull VFS patches (part 1) from Al Viro:
 "The major change in this pile is ->readdir() replacement with
  ->iterate(), dealing with ->f_pos races in ->readdir() instances for
  good.

  There's a lot more, but I'd prefer to split the pull request into
  several stages and this is the first obvious cutoff point."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (67 commits)
  [readdir] constify ->actor
  [readdir] ->readdir() is gone
  [readdir] convert ecryptfs
  [readdir] convert coda
  [readdir] convert ocfs2
  [readdir] convert fatfs
  [readdir] convert xfs
  [readdir] convert btrfs
  [readdir] convert hostfs
  [readdir] convert afs
  [readdir] convert ncpfs
  [readdir] convert hfsplus
  [readdir] convert hfs
  [readdir] convert befs
  [readdir] convert cifs
  [readdir] convert freevxfs
  [readdir] convert fuse
  [readdir] convert hpfs
  reiserfs: switch reiserfs_readdir_dentry to inode
  reiserfs: is_privroot_deh() needs only directory inode, actually
  ...

11 years agosync: don't block the flusher thread waiting on IO
Dave Chinner [Tue, 2 Jul 2013 12:38:35 +0000 (22:38 +1000)]
sync: don't block the flusher thread waiting on IO

When sync does it's WB_SYNC_ALL writeback, it issues data Io and
then immediately waits for IO completion. This is done in the
context of the flusher thread, and hence completely ties up the
flusher thread for the backing device until all the dirty inodes
have been synced. On filesystems that are dirtying inodes constantly
and quickly, this means the flusher thread can be tied up for
minutes per sync call and hence badly affect system level write IO
performance as the page cache cannot be cleaned quickly.

We already have a wait loop for IO completion for sync(2), so cut
this out of the flusher thread and delegate it to wait_sb_inodes().
Hence we can do rapid IO submission, and then wait for it all to
complete.

Effect of sync on fsmark before the patch:

FSUse%        Count         Size    Files/sec     App Overhead
.....
     0       640000         4096      35154.6          1026984
     0       720000         4096      36740.3          1023844
     0       800000         4096      36184.6           916599
     0       880000         4096       1282.7          1054367
     0       960000         4096       3951.3           918773
     0      1040000         4096      40646.2           996448
     0      1120000         4096      43610.1           895647
     0      1200000         4096      40333.1           921048

And a single sync pass took:

  real    0m52.407s
  user    0m0.000s
  sys     0m0.090s

After the patch, there is no impact on fsmark results, and each
individual sync(2) operation run concurrently with the same fsmark
workload takes roughly 7s:

  real    0m6.930s
  user    0m0.000s
  sys     0m0.039s

IOWs, sync is 7-8x faster on a busy filesystem and does not have an
adverse impact on ongoing async data write operations.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agof2fs: fix to recover i_size from roll-forward
Jaegeuk Kim [Thu, 27 Jun 2013 04:04:08 +0000 (13:04 +0900)]
f2fs: fix to recover i_size from roll-forward

If user requests many data writes and fsync together, the last updated i_size
should be stored to the inode block consistently.

But, previous write_end just marks the inode as dirty and doesn't update its
metadata into its inode block.
After that, fsync just writes the inode block with newly updated data index
excluding inode metadata updates.

So, this patch introduces write_end in which updates inode block too when the
i_size is changed.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: remove the unused argument "sbi" of func destroy_fsync_dnodes()
Gu Zheng [Thu, 27 Jun 2013 01:28:54 +0000 (09:28 +0800)]
f2fs: remove the unused argument "sbi" of func destroy_fsync_dnodes()

As destroy_fsync_dnodes() is a simple list-cleanup func, so delete the unused
and unrelated f2fs_sb_info argument of it.

Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: remove reusing any prefree segments
Jaegeuk Kim [Thu, 27 Jun 2013 00:59:40 +0000 (09:59 +0900)]
f2fs: remove reusing any prefree segments

This patch removes check_prefree_segments initially designed to enhance the
performance by narrowing the range of LBA usage across the whole block device.

When allocating a new segment, previous f2fs tries to find proper prefree
segments, and then, if finds a segment, it reuses the segment for further
data or node block allocation.

However, I found that this was totally wrong approach since the prefree segments
have several data or node blocks that will be used by the roll-forward mechanism
operated after sudden-power-off.

Let's assume the following scenario.

/* write 8MB with fsync */
for (i = 0; i < 2048; i++) {
offset = i * 4096;
write(fd, offset, 4KB);
fsync(fd);
}

In this case, naive segment allocation sequence will be like:
 data segment: x, x+1, x+2, x+3
 node segment: y, y+1, y+2, y+3.

But, if we can reuse prefree segments, the sequence can be like:
 data segment: x, x+1, y, y+1
 node segment: y, y+1, y+2, y+3.
Because, y, y+1, and y+2 became prefree segments one by one, and those are
reused by data allocation.

After conducting this workload, we should consider how to recover the latest
inode with its data.
If we reuse the prefree segments such as y or y+1, we lost the old node blocks
so that f2fs even cannot start roll-forward recovery.

Therefore, I suggest that we should remove reusing prefree segments.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: code cleanup and simplify in func {find/add}_gc_inode
Gu Zheng [Thu, 20 Jun 2013 09:52:39 +0000 (17:52 +0800)]
f2fs: code cleanup and simplify in func {find/add}_gc_inode

This patch simplifies list operations in find_gc_inode and add_gc_inode.
Just simple code cleanup.

Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
[Jaegeuk Kim: add description]
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: optimize the init_dirty_segmap function
Namjae Jeon [Sun, 16 Jun 2013 00:49:11 +0000 (09:49 +0900)]
f2fs: optimize the init_dirty_segmap function

Optimize the while loop condition

Since this condition will always be true and while loop will
be terminated by the following condition in code:

if (segno >= TOTAL_SEGS(sbi))
    break;
Hence we can replace the while loop condition with while(1)
instead of always checking for segno to be less than Total segs.

Also we do not need to use TOTAL_SEGS() everytime. We can store
this value in a local variable since this value is constant.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Pankaj Kumar <pankaj.km@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: fix an endian conversion bug detected by sparse
Jaegeuk Kim [Sun, 23 Jun 2013 22:47:23 +0000 (07:47 +0900)]
f2fs: fix an endian conversion bug detected by sparse

This patch should fix the following bug reported by kbuild test robot.

fs/f2fs/recovery.c:233:33: sparse: incorrect type in assignment
(different base types)

parse warnings: (new ones prefixed by >>)

>> recovery.c:233: sparse: incorrect type in assignment (different base types)
   recovery.c:233:    expected unsigned int [unsigned] [assigned] ofs_in_node
   recovery.c:233:    got restricted __le16 [assigned] [usertype] ofs_in_node
>> recovery.c:238: sparse: incorrect type in assignment (different base types)
   recovery.c:238:    expected unsigned int [unsigned] ofs_in_node
   recovery.c:238:    got restricted __le16 [assigned] [usertype] ofs_in_node

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agof2fs: fix crc endian conversion
Jaegeuk Kim [Wed, 19 Jun 2013 11:47:19 +0000 (20:47 +0900)]
f2fs: fix crc endian conversion

While calculating CRC for the checkpoint block, we use __u32, but when storing
the crc value to the disk, we use __le32.

Let's fix the inconsistency.

Reported-and-Tested-by: Oded Gabbay <ogabbay@advaoptical.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
11 years agoext4: optimize starting extent in ext4_ext_rm_leaf()
Ashish Sangwan [Mon, 1 Jul 2013 12:12:41 +0000 (08:12 -0400)]
ext4: optimize starting extent in ext4_ext_rm_leaf()

Both hole punch and truncate use ext4_ext_rm_leaf() for removing
blocks.  Currently we choose the last extent as the starting
point for removing blocks:

ex = EXT_LAST_EXTENT(eh);

This is OK for truncate but for hole punch we can optimize the extent
selection as the path is already initialized.  We could use this
information to select proper starting extent.  The code change in this
patch will not affect truncate as for truncate path[depth].p_ext will
always be NULL.

Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agojbd2: invalidate handle if jbd2_journal_restart() fails
Theodore Ts'o [Mon, 1 Jul 2013 12:12:41 +0000 (08:12 -0400)]
jbd2: invalidate handle if jbd2_journal_restart() fails

If jbd2_journal_restart() fails the handle will have been disconnected
from the current transaction.  In this situation, the handle must not
be used for for any jbd2 function other than jbd2_journal_stop().
Enforce this with by treating a handle which has a NULL transaction
pointer as an aborted handle, and issue a kernel warning if
jbd2_journal_extent(), jbd2_journal_get_write_access(),
jbd2_journal_dirty_metadata(), etc. is called with an invalid handle.

This commit also fixes a bug where jbd2_journal_stop() would trip over
a kernel jbd2 assertion check when trying to free an invalid handle.

Also move the responsibility of setting current->journal_info to
start_this_handle(), simplifying the three users of this function.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reported-by: Younger Liu <younger.liu@huawei.com>
Cc: Jan Kara <jack@suse.cz>
11 years agoext4: translate flag bits to strings in tracepoints
Theodore Ts'o [Mon, 1 Jul 2013 12:12:40 +0000 (08:12 -0400)]
ext4: translate flag bits to strings in tracepoints

Translate the bitfields used in various flags argument to strings to
make the tracepoint output more human-readable.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: fix up error handling for mpage_map_and_submit_extent()
Theodore Ts'o [Mon, 1 Jul 2013 12:12:40 +0000 (08:12 -0400)]
ext4: fix up error handling for mpage_map_and_submit_extent()

The function mpage_released_unused_page() must only be called once;
otherwise the kernel will BUG() when the second call to
mpage_released_unused_page() tries to unlock the pages which had been
unlocked by the first call.

Also restructure the error handling so that we only give up on writing
the dirty pages in the case of ENOSPC where retrying the allocation
won't help.  Otherwise, a transient failure, such as a kmalloc()
failure in calling ext4_map_blocks() might cause us to give up on
those pages, leading to a scary message in /var/log/messages plus data
loss.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Jan Kara <jack@suse.cz>
11 years agojbd2: fix theoretical race in jbd2__journal_restart
Theodore Ts'o [Mon, 1 Jul 2013 12:12:40 +0000 (08:12 -0400)]
jbd2: fix theoretical race in jbd2__journal_restart

Once we decrement transaction->t_updates, if this is the last handle
holding the transaction from closing, and once we release the
t_handle_lock spinlock, it's possible for the transaction to commit
and be released.  In practice with normal kernels, this probably won't
happen, since the commit happens in a separate kernel thread and it's
unlikely this could all happen within the space of a few CPU cycles.

On the other hand, with a real-time kernel, this could potentially
happen, so save the tid found in transaction->t_tid before we release
t_handle_lock.  It would require an insane configuration, such as one
where the jbd2 thread was set to a very high real-time priority,
perhaps because a high priority real-time thread is trying to read or
write to a file system.  But some people who use real-time kernels
have been known to do insane things, including controlling
laser-wielding industrial robots.  :-)

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org
11 years agoext4: only zero partial blocks in ext4_zero_partial_blocks()
Lukas Czerner [Mon, 1 Jul 2013 12:12:39 +0000 (08:12 -0400)]
ext4: only zero partial blocks in ext4_zero_partial_blocks()

Currently if we pass range into ext4_zero_partial_blocks() which covers
entire block we would attempt to zero it even though we should only zero
unaligned part of the block.

Fix this by checking whether the range covers the whole block skip
zeroing if so.

Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: check error return from ext4_write_inline_data_end()
Theodore Ts'o [Mon, 1 Jul 2013 12:12:39 +0000 (08:12 -0400)]
ext4: check error return from ext4_write_inline_data_end()

The function ext4_write_inline_data_end() can return an error.  So we
need to assign it to a signed integer variable to check for an error
return (since copied is an unsigned int).

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Zheng Liu <wenqing.lz@taobao.com>
Cc: stable@vger.kernel.org
11 years agoext4: delete unnecessary C statements
jon ernst [Mon, 1 Jul 2013 12:12:39 +0000 (08:12 -0400)]
ext4: delete unnecessary C statements

Comparing unsigned variable with 0 always returns false.
err = 0 is duplicated and unnecessary.

[ tytso: Also cleaned up error handling in ext4_block_zero_page_range() ]

Signed-off-by: "Jon Ernst" <jonernst07@gmx.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext3,ext4: don't mess with dir_file->f_pos in htree_dirblock_to_tree()
Al Viro [Mon, 1 Jul 2013 12:12:38 +0000 (08:12 -0400)]
ext3,ext4: don't mess with dir_file->f_pos in htree_dirblock_to_tree()

Both ext3 and ext4 htree_dirblock_to_tree() is just filling the
in-core rbtree for use by call_filldir().  All updates of ->f_pos are
done by the latter; bumping it here (on error) is obviously wrong - we
might very well have it nowhere near the block we'd found an error in.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@vger.kernel.org
11 years agojbd2: move superblock checksum calculation to jbd2_write_superblock()
Theodore Ts'o [Mon, 1 Jul 2013 12:12:38 +0000 (08:12 -0400)]
jbd2: move superblock checksum calculation to jbd2_write_superblock()

Some of the functions which modify the jbd2 superblock were not
updating the checksum before calling jbd2_write_superblock().  Move
the call to jbd2_superblock_csum_set() to jbd2_write_superblock(), so
that the checksum is calculated consistently.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Darrick J. Wong <darrick.wong@oracle.com>
Cc: stable@vger.kernel.org
11 years agoext4: pass inode pointer instead of file pointer to punch hole
Ashish Sangwan [Mon, 1 Jul 2013 12:12:38 +0000 (08:12 -0400)]
ext4: pass inode pointer instead of file pointer to punch hole

No need to pass file pointer when we can directly pass inode pointer.

Signed-off-by: Ashish Sangwan <a.sangwan@samsung.com>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: improve free space calculation for inline_data
boxi liu [Mon, 1 Jul 2013 12:12:37 +0000 (08:12 -0400)]
ext4: improve free space calculation for inline_data

In ext4 feature inline_data,it use the xattr's space to store the
inline data in inode.When we calculate the inline data as the xattr,we
add the pad.But in get_max_inline_xattr_value_size() function we count
the free space without pad.It cause some contents are moved to a block
even if it can be
stored in the inode.

Signed-off-by: liulei <lewis.liulei@huawei.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Tao Ma <boyu.mt@taobao.com>
11 years agoext4: reduce object size when !CONFIG_PRINTK
Joe Perches [Mon, 1 Jul 2013 12:12:37 +0000 (08:12 -0400)]
ext4: reduce object size when !CONFIG_PRINTK

Reduce the object size ~10% could be useful for embedded systems.

Add #ifdef CONFIG_PRINTK #else #endif blocks to hold formats and
arguments, passing " " to functions when !CONFIG_PRINTK and still
verifying format and arguments with no_printk.

$ size fs/ext4/built-in.o*
   text    data     bss     dec     hex filename
 239375     610     888  240873   3ace9 fs/ext4/built-in.o.new
 264167     738     888  265793   40e41 fs/ext4/built-in.o.old

    $ grep -E "CONFIG_EXT4|CONFIG_PRINTK" .config
    # CONFIG_PRINTK is not set
    CONFIG_EXT4_FS=y
    CONFIG_EXT4_USE_FOR_EXT23=y
    CONFIG_EXT4_FS_POSIX_ACL=y
    # CONFIG_EXT4_FS_SECURITY is not set
    # CONFIG_EXT4_DEBUG is not set

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: improve extent cache shrink mechanism to avoid to burn CPU time
Zheng Liu [Mon, 1 Jul 2013 12:12:37 +0000 (08:12 -0400)]
ext4: improve extent cache shrink mechanism to avoid to burn CPU time

Now we maintain an proper in-order LRU list in ext4 to reclaim entries
from extent status tree when we are under heavy memory pressure.  For
keeping this order, a spin lock is used to protect this list.  But this
lock burns a lot of CPU time.  We can use the following steps to trigger
it.

  % cd /dev/shm
  % dd if=/dev/zero of=ext4-img bs=1M count=2k
  % mkfs.ext4 ext4-img
  % mount -t ext4 -o loop ext4-img /mnt
  % cd /mnt
  % for ((i=0;i<160;i++)); do truncate -s 64g $i; done
  % for ((i=0;i<160;i++)); do cp $i /dev/null &; done
  % perf record -a -g
  % perf report

This commit tries to fix this problem.  Now a new member called
i_touch_when is added into ext4_inode_info to record the last access
time for an inode.  Meanwhile we never need to keep a proper in-order
LRU list.  So this can avoid to burns some CPU time.  When we try to
reclaim some entries from extent status tree, we use list_sort() to get
a proper in-order list.  Then we traverse this list to discard some
entries.  In ext4_sb_info, we use s_es_last_sorted to record the last
time of sorting this list.  When we traverse the list, we skip the inode
that is newer than this time, and move this inode to the tail of LRU
list.  When the head of the list is newer than s_es_last_sorted, we will
sort the LRU list again.

In this commit, we break the loop if s_extent_cache_cnt == 0 because
that means that all extents in extent status tree have been reclaimed.

Meanwhile in this commit, ext4_es_{un}register_shrinker()'s prototype is
changed to save a local variable in these functions.

Reported-by: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: implement error handling of ext4_mb_new_preallocation()
Alexey Khoroshilov [Mon, 1 Jul 2013 12:12:36 +0000 (08:12 -0400)]
ext4: implement error handling of ext4_mb_new_preallocation()

If memory allocation in ext4_mb_new_group_pa() is failed,
it returns error code, ext4_mb_new_preallocation() propages it,
but ext4_mb_new_blocks() ignores it.

An observed result was:

- allocation fail means ext4_mb_new_group_pa() does not update
  ext4_allocation_context;

- ext4_mb_new_blocks() sets ext4_allocation_request->len (ar->len =
  ac->ac_b_ex.fe_len;) to number of blocks preallocated (512) instead
  of number of blocks requested (1);

- that activates update cycle in ext4_splice_branch():
    for (i = 1; i < blks; i++) <-- blks is 512 instead of 1 here
      *(where->p + i) = cpu_to_le32(current_block++);

- it iterates 511 times and corrupts a chunk of memory including inode
  structure;

- page fault happens at EXT4_SB(inode->i_sb) in ext4_mark_inode_dirty();

- system hangs with 'scheduling while atomic' BUG.

The patch implements a check for ext4_mb_new_preallocation() error
code and handles its failure as if ext4_mb_regular_allocator() fails.

Found by Linux File System Verification project (linuxtesting.org).

[ Patch restructed by tytso to make the flow of control easier to follow. ]

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoext4: fix corruption when online resizing a fs with 1K block size
Maarten ter Huurne [Mon, 1 Jul 2013 12:12:08 +0000 (08:12 -0400)]
ext4: fix corruption when online resizing a fs with 1K block size

Subtracting the number of the first data block places the superblock
backups one block too early, corrupting the file system. When the block
size is larger than 1K, the first data block is 0, so the subtraction
has no effect and no corruption occurs.

Signed-off-by: Maarten ter Huurne <maarten@treewalker.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Jan Kara <jack@suse.cz>
CC: stable@vger.kernel.org
11 years agoMerge branch 'for-next/hugepages' of git://git.linaro.org/people/stevecapper/linux...
Catalin Marinas [Mon, 1 Jul 2013 10:20:58 +0000 (11:20 +0100)]
Merge branch 'for-next/hugepages' of git://git.linaro.org/people/stevecapper/linux into upstream-hugepages

* 'for-next/hugepages' of git://git.linaro.org/people/stevecapper/linux:
  ARM64: mm: THP support.
  ARM64: mm: Raise MAX_ORDER for 64KB pages and THP.
  ARM64: mm: HugeTLB support.
  ARM64: mm: Move PTE_PROT_NONE bit.
  ARM64: mm: Make PAGE_NONE pages read only and no-execute.
  ARM64: mm: Restore memblock limit when map_mem finished.
  mm: thp: Correct the HPAGE_PMD_ORDER check.
  x86: mm: Remove general hugetlb code from x86.
  mm: hugetlb: Copy general hugetlb code from x86 to mm.
  x86: mm: Remove x86 version of huge_pmd_share.
  mm: hugetlb: Copy huge_pmd_share from x86 to mm.

Conflicts:
arch/arm64/Kconfig
arch/arm64/include/asm/pgtable-hwdef.h
arch/arm64/include/asm/pgtable.h

11 years agoMerge tag 'v3.10' into sched/core
Ingo Molnar [Mon, 1 Jul 2013 09:16:54 +0000 (11:16 +0200)]
Merge tag 'v3.10' into sched/core

Merge in a recent upstream commit:

  c2853c8df57f include/linux/math64.h: add div64_ul()

because:

  72a4cf20cb71 sched: Change cfs_rq load avg to unsigned long

relies on it.

[ We don't rebase sched/core for this, because the handful of
  followup commits after the broken commit are not behavioral
  changes so are unlikely to be needed during bisection. ]

Signed-off-by: Ingo Molnar <mingo@kernel.org>
11 years agoLinux 3.10 v3.10
Linus Torvalds [Sun, 30 Jun 2013 22:13:29 +0000 (15:13 -0700)]
Linux 3.10

11 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Sun, 30 Jun 2013 22:08:15 +0000 (15:08 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull another powerpc fix from Benjamin Herrenschmidt:
 "I mentioned that while we had fixed the kernel crashes, EEH error
  recovery didn't always recover...  It appears that I had a fix for
  that already in powerpc-next (with a stable CC).

  I cherry-picked it today and did a few tests and it seems that things
  now work quite well.  The patch is also pretty simple, so I see no
  reason to wait before merging it."

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/eeh: Fix fetching bus for single-dev-PE

11 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sun, 30 Jun 2013 22:06:25 +0000 (15:06 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is a set of seven bug fixes.  Several fcoe fixes for locking
  problems, initiator issues and a VLAN API change, all of which could
  eventually lead to data corruption, one fix for a qla2xxx locking
  problem which could lead to multiple completions of the same request
  (and subsequent data corruption) and a use after free in the ipr
  driver.  Plus one minor MAINTAINERS file update"

(only six bugfixes in this pull, since I had already pulled the fcoe API
fix directly from Robert Love)

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] ipr: Avoid target_destroy accessing memory after it was freed
  [SCSI] qla2xxx: Fix for locking issue between driver ISR and mailbox routines
  MAINTAINERS: Fix fcoe mailing list
  libfc: extend ex_lock to protect all of fc_seq_send
  libfc: Correct check for initiator role
  libfcoe: Fix Conflicting FCFs issue in the fabric

11 years agoRevert "serial: 8250_pci: add support for another kind of NetMos Technology PCI 9835...
Greg Kroah-Hartman [Sun, 30 Jun 2013 16:03:06 +0000 (09:03 -0700)]
Revert "serial: 8250_pci: add support for another kind of NetMos Technology PCI 9835 Multi-I/O Controller"

This reverts commit 8d2f8cd424ca0b99001f3ff4f5db87c4e525f366.

As reported by Stefan, this device already works with the parport_serial
driver, so the 8250_pci driver should not also try to grab it as well.

Reported-by: Stefan Seyfried <stefan.seyfried@googlemail.com>
Cc: Wang YanQing <udknight@gmail.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agopowerpc/eeh: Fix fetching bus for single-dev-PE
Gavin Shan [Wed, 5 Jun 2013 07:34:02 +0000 (15:34 +0800)]
powerpc/eeh: Fix fetching bus for single-dev-PE

While running Linux as guest on top of phyp, we possiblly have
PE that includes single PCI device. However, we didn't return
its PCI bus correctly and it leads to failure on recovery from
EEH errors for single-dev-PE. The patch fixes the issue.

Cc: <stable@vger.kernel.org> # v3.7+
Cc: Steve Best <sbest@us.ibm.com>
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Sun, 30 Jun 2013 00:02:48 +0000 (17:02 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull powerpc fixes from Ben Herrenschmidt:
 "We discovered some breakage in our "EEH" (PCI Error Handling) code
  while doing error injection, due to a couple of regressions.  One of
  them is due to a patch (37f02195bee9 "powerpc/pci: fix PCI-e devices
  rescan issue on powerpc platform") that, in hindsight, I shouldn't
  have merged considering that it caused more problems than it solved.

  Please pull those two fixes.  One for a simple EEH address cache
  initialization issue.  The other one is a patch from Guenter that I
  had originally planned to put in 3.11 but which happens to also fix
  that other regression (a kernel oops during EEH error handling and
  possibly hotplug).

  With those two, the couple of test machines I've hammered with error
  injection are remaining up now.  EEH appears to still fail to recover
  on some devices, so there is another problem that Gavin is looking
  into but at least it's no longer crashing the kernel."

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/pci: Improve device hotplug initialization
  powerpc/eeh: Add eeh_dev to the cache during boot

11 years agoARM: dt: Only print warning, not WARN() on bad cpu map in device tree
Olof Johansson [Sat, 29 Jun 2013 23:25:14 +0000 (16:25 -0700)]
ARM: dt: Only print warning, not WARN() on bad cpu map in device tree

Due to recent changes and expecations of proper cpu bindings, there are
now cases for many of the in-tree devicetrees where a WARN() will hit
on boot due to badly formatted /cpus nodes.

Downgrade this to a pr_warn() to be less alarmist, since it's not a
new problem.

Tested on Arndale, Cubox, Seaboard and Panda ES. Panda hits the WARN
without this, the others do not.

Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agopowerpc/pci: Improve device hotplug initialization
Guenter Roeck [Mon, 10 Jun 2013 17:18:08 +0000 (10:18 -0700)]
powerpc/pci: Improve device hotplug initialization

Commit 37f02195b (powerpc/pci: fix PCI-e devices rescan issue on powerpc
platform) fixes a problem with interrupt and DMA initialization on hot
plugged devices. With this commit, interrupt and DMA initialization for
hot plugged devices is handled in the pci device enable function.

This approach has a couple of drawbacks. First, it creates two code paths
for device initialization, one for hot plugged devices and another for devices
known during the initial PCI scan. Second, the initialization code for hot
plugged devices is only called when the device is enabled, ie typically
in the probe function. Also, the platform specific setup code is called each
time pci_enable_device() is called, not only once during device discovery,
meaning it is actually called multiple times, once for devices discovered
during the initial scan and again each time a driver is re-loaded.

The visible result is that interrupt pins are only assigned to hot plugged
devices when the device driver is loaded. Effectively this changes the PCI
probe API, since pci_dev->irq and the device's dma configuration will now
only be valid after pci_enable() was called at least once. A more subtle
change is that platform specific PCI device setup is moved from device
discovery into the driver's probe function, more specifically into the
pci_enable_device() call.

To fix the inconsistencies, add new function pcibios_add_device.
Call pcibios_setup_device from pcibios_setup_bus_devices if device setup
is not complete, and from pcibios_add_device if bus setup is complete.

With this change, device setup code is moved back into device initialization,
and called exactly once for both static and hot plugged devices.

[ This also fixes a regression introduced by the above patch which
  causes dev->irq to be overwritten under some cirumstances after
  MSIs have been enabled for the device which leads to crashes due
  to the MSI core "hijacking" dev->irq to store the base MSI number
  and not the LSI. --BenH
]

Cc: Yuanquan Chen <Yuanquan.Chen@freescale.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Hiroo Matsumoto <matsumoto.hiroo@jp.fujitsu.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agocgroup: CGRP_ROOT_SUBSYS_BOUND should also be ignored when mounting an existing hierarchy
Tejun Heo [Sat, 29 Jun 2013 21:06:10 +0000 (14:06 -0700)]
cgroup: CGRP_ROOT_SUBSYS_BOUND should also be ignored when mounting an existing hierarchy

0ce6cba357 ("cgroup: CGRP_ROOT_SUBSYS_BOUND should be ignored when
comparing mount options") only updated the remount path but
CGRP_ROOT_SUBSYS_BOUND should also be ignored when comparing options
while mounting an existing hierarchy.  As option mismatch triggers a
warning but doesn't fail the mount without sane_behavior, this only
triggers a spurious warning message.

Fix it by only comparing CGRP_ROOT_OPTION_MASK bits when comparing new
and existing root options.

Signed-off-by: Tejun Heo <tj@kernel.org>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Sat, 29 Jun 2013 18:34:18 +0000 (11:34 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fix from Herbert Xu:
 "This fixes a crash in the crypto layer exposed by an SCTP test tool"

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: algboss - Hold ref count on larval

11 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Sat, 29 Jun 2013 18:32:05 +0000 (11:32 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm/qxl fix from Dave Airlie:
 "Bad me forgot an access check, possible security issue, but since this
  is the first kernel with it, should be fine to just put it in now"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/qxl: add missing access check for execbuffer ioctl

11 years agoFix: kernel/ptrace.c: ptrace_peek_siginfo() missing __put_user() validation
Mathieu Desnoyers [Fri, 28 Jun 2013 13:49:46 +0000 (09:49 -0400)]
Fix: kernel/ptrace.c: ptrace_peek_siginfo() missing __put_user() validation

This __put_user() could be used by unprivileged processes to write into
kernel memory.  The issue here is that even if copy_siginfo_to_user()
fails, the error code is not checked before __put_user() is executed.

Luckily, ptrace_peek_siginfo() has been added within the 3.10-rc cycle,
so it has not hit a stable release yet.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andrey Vagin <avagin@openvz.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Pedro Alves <palves@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph...
Linus Torvalds [Sat, 29 Jun 2013 17:31:15 +0000 (10:31 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/sage/ceph-client

Pull Ceph fix from Sage Weil:
 "This is a recently spotted regression in the snapshot behavior...

  It turns out several tests weren't being run in the nightlies so this
  took a while to spot"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: send snapshot context with writes

11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sat, 29 Jun 2013 17:30:31 +0000 (10:30 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull ubifs fixes from Al Viro:
 "A couple of ubifs readdir/lseek race fixes.  Stable fodder, really
  nasty..."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  UBIFS: fix a horrid bug
  UBIFS: prepare to fix a horrid bug

11 years agoMerge tag 'for-linus-20130628' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowe...
Linus Torvalds [Sat, 29 Jun 2013 17:28:52 +0000 (10:28 -0700)]
Merge tag 'for-linus-20130628' of git://git./linux/kernel/git/dhowells/linux-mn10300

Pull two MN10300 fixes from David Howells:
 "The first fixes a problem with passing arrays rather than pointers to
  get_user() where __typeof__ then wants to declare and initialise an
  array variable which gcc doesn't like.

  The second fixes a problem whereby putting mem=xxx into the kernel
  command line causes init=xxx to get an incorrect value."

* tag 'for-linus-20130628' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-mn10300:
  mn10300: Use early_param() to parse "mem=" parameter
  mn10300: Allow to pass array name to get_user()

11 years agoMerge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 29 Jun 2013 17:27:19 +0000 (10:27 -0700)]
Merge branch 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull timer fix from Thomas Gleixner:
 "Correct an ordering issue in the tick broadcast code.  I really wish
  we'd get compensation for pain and suffering for each line of code we
  write to work around dysfunctional timer hardware."

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tick: Fix tick_broadcast_pending_mask not cleared

11 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 29 Jun 2013 17:26:50 +0000 (10:26 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fix from Ingo Molnar:
 "One more fix for a recently discovered bug"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf: Disable monitoring on setuid processes for regular users

11 years agoMerge branch 'devel-stable' into for-next
Russell King [Sat, 29 Jun 2013 10:44:43 +0000 (11:44 +0100)]
Merge branch 'devel-stable' into for-next

Conflicts:
arch/arm/Makefile
arch/arm/include/asm/glue-proc.h

11 years agoMerge branches 'fixes', 'mcpm', 'misc' and 'mmci' into for-next
Russell King [Sat, 29 Jun 2013 10:43:28 +0000 (11:43 +0100)]
Merge branches 'fixes', 'mcpm', 'misc' and 'mmci' into for-next

11 years agoARM: 7775/1: mm: Remove do_sect_fault from LPAE code
Steven Capper [Tue, 25 Jun 2013 07:45:51 +0000 (08:45 +0100)]
ARM: 7775/1: mm: Remove do_sect_fault from LPAE code

For LPAE, do_sect_fault used to be invoked as the second level access
flag handler. When transparent huge pages were introduced for LPAE,
do_page_fault was used instead.

Unfortunately, do_sect_fault remains defined but not used for LPAE code
resulting in a compile warning.

This patch surrounds do_sect_fault with #ifndef CONFIG_ARM_LPAE to fix
this warning.

Signed-off-by: Steve Capper <steve.capper@linaro.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7777/1: Avoid extra calls to the C compiler
Douglas Anderson [Thu, 27 Jun 2013 03:32:06 +0000 (04:32 +0100)]
ARM: 7777/1: Avoid extra calls to the C compiler

Starting up the C compiler can be a slow operation on some systems.
Though these calls don't individually take a lot of time, they add up.
Rearrange the ARM Makefile a bit to avoid extra calls to the compiler
when they can be easily avoided.

When running with the Chrome OS ARM cross compiler
"armv7a-cros-linux-gnueabi-", this shaved .55 seconds (from 5.31
seconds to 4.76 seconds) off an incremental build of the kernel:
  time make -j32 ARCH=arm CROSS_COMPILE=armv7a-cros-linux-gnueabi-

Thanks to Mike Frysinger for the clean trick to make this work.

Signed-off-by: Doug Anderson <dianders@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7774/1: Fix dtb dependency to use order-only prerequisites
Douglas Anderson [Mon, 24 Jun 2013 21:03:50 +0000 (22:03 +0100)]
ARM: 7774/1: Fix dtb dependency to use order-only prerequisites

The %.dtb dependency is specified to depend on the PHONY "scripts".
That means that it'll build every time even if the underlying dtb file
hasn't been touched.  Use an order-only prerequisites to fix this.
Also mark "dtbs" as PHONY for correctness.

This was broken in (70b0476 ARM: 7513/1: Make sure dtc is built before
running it).

Reported-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Doug Anderson <dianders@chromium.org>
Acked-by: Olof Johansson <olof@lixom.net>
Reviewed-by: David Brown <davidb@codeaurora.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agolseek_execute() doesn't need an inode passed to it
Al Viro [Sun, 23 Jun 2013 18:45:30 +0000 (22:45 +0400)]
lseek_execute() doesn't need an inode passed to it

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoblock_dev: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 17:34:45 +0000 (21:34 +0400)]
block_dev: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agocpqphp_sysfs: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:46:49 +0000 (12:46 +0400)]
cpqphp_sysfs: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agotile-srom: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:43:48 +0000 (12:43 +0400)]
tile-srom: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoproc_powerpc: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:35:26 +0000 (12:35 +0400)]
proc_powerpc: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoubi/cdev: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:30:24 +0000 (12:30 +0400)]
ubi/cdev: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agopci/proc: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:29:00 +0000 (12:29 +0400)]
pci/proc: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoisapnp: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:09:11 +0000 (12:09 +0400)]
isapnp: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolpfc: switch to fixed_size_llseek()
Al Viro [Sun, 23 Jun 2013 08:08:05 +0000 (12:08 +0400)]
lpfc: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: give the blocked_hash its own spinlock
Jeff Layton [Fri, 21 Jun 2013 12:58:20 +0000 (08:58 -0400)]
locks: give the blocked_hash its own spinlock

There's no reason we have to protect the blocked_hash and file_lock_list
with the same spinlock. With the tests I have, breaking it in two gives
a barely measurable performance benefit, but it seems reasonable to make
this locking as granular as possible.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: add a new "lm_owner_key" lock operation
Jeff Layton [Fri, 21 Jun 2013 12:58:19 +0000 (08:58 -0400)]
locks: add a new "lm_owner_key" lock operation

Currently, the hashing that the locking code uses to add these values
to the blocked_hash is simply calculated using fl_owner field. That's
valid in most cases except for server-side lockd, which validates the
owner of a lock based on fl_owner and fl_pid.

In the case where you have a small number of NFS clients doing a lot
of locking between different processes, you could end up with all
the blocked requests sitting in a very small number of hash buckets.

Add a new lm_owner_key operation to the lock_manager_operations that
will generate an unsigned long to use as the key in the hashtable.
That function is only implemented for server-side lockd, and simply
XORs the fl_owner and fl_pid.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: turn the blocked_list into a hashtable
Jeff Layton [Fri, 21 Jun 2013 12:58:18 +0000 (08:58 -0400)]
locks: turn the blocked_list into a hashtable

Break up the blocked_list into a hashtable, using the fl_owner as a key.
This speeds up searching the hash chains, which is especially significant
for deadlock detection.

Note that the initial implementation assumes that hashing on fl_owner is
sufficient. In most cases it should be, with the notable exception being
server-side lockd, which compares ownership using a tuple of the
nlm_host and the pid sent in the lock request. So, this may degrade to a
single hash bucket when you only have a single NFS client. That will be
addressed in a later patch.

The careful observer may note that this patch leaves the file_lock_list
alone. There's much less of a case for turning the file_lock_list into a
hashtable. The only user of that list is the code that generates
/proc/locks, and it always walks the entire list.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: convert fl_link to a hlist_node
Jeff Layton [Fri, 21 Jun 2013 12:58:17 +0000 (08:58 -0400)]
locks: convert fl_link to a hlist_node

Testing has shown that iterating over the blocked_list for deadlock
detection turns out to be a bottleneck. In order to alleviate that,
begin the process of turning it into a hashtable. We start by turning
the fl_link into a hlist_node and the global lists into hlists. A later
patch will do the conversion of the blocked_list to a hashtable.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: avoid taking global lock if possible when waking up blocked waiters
Jeff Layton [Fri, 21 Jun 2013 12:58:16 +0000 (08:58 -0400)]
locks: avoid taking global lock if possible when waking up blocked waiters

Since we always hold the i_lock when inserting a new waiter onto the
fl_block list, we can avoid taking the global lock at all if we find
that it's empty when we go to wake up blocked waiters.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: protect most of the file_lock handling with i_lock
Jeff Layton [Fri, 21 Jun 2013 12:58:15 +0000 (08:58 -0400)]
locks: protect most of the file_lock handling with i_lock

Having a global lock that protects all of this code is a clear
scalability problem. Instead of doing that, move most of the code to be
protected by the i_lock instead. The exceptions are the global lists
that the ->fl_link sits on, and the ->fl_block list.

->fl_link is what connects these structures to the
global lists, so we must ensure that we hold those locks when iterating
over or updating these lists.

Furthermore, sound deadlock detection requires that we hold the
blocked_list state steady while checking for loops. We also must ensure
that the search and update to the list are atomic.

For the checking and insertion side of the blocked_list, push the
acquisition of the global lock into __posix_lock_file and ensure that
checking and update of the  blocked_list is done without dropping the
lock in between.

On the removal side, when waking up blocked lock waiters, take the
global lock before walking the blocked list and dequeue the waiters from
the global list prior to removal from the fl_block list.

With this, deadlock detection should be race free while we minimize
excessive file_lock_lock thrashing.

Finally, in order to avoid a lock inversion problem when handling
/proc/locks output we must ensure that manipulations of the fl_block
list are also protected by the file_lock_lock.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: encapsulate the fl_link list handling
Jeff Layton [Fri, 21 Jun 2013 12:58:14 +0000 (08:58 -0400)]
locks: encapsulate the fl_link list handling

Move the fl_link list handling routines into a separate set of helpers.
Also ensure that locks and requests are always put on global lists
last (after fully initializing them) and are taken off before unintializing
them.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: make "added" in __posix_lock_file a bool
Jeff Layton [Fri, 21 Jun 2013 12:58:13 +0000 (08:58 -0400)]
locks: make "added" in __posix_lock_file a bool

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: comment cleanups and clarifications
Jeff Layton [Fri, 21 Jun 2013 12:58:12 +0000 (08:58 -0400)]
locks: comment cleanups and clarifications

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: make generic_add_lease and generic_delete_lease static
Jeff Layton [Fri, 21 Jun 2013 12:58:11 +0000 (08:58 -0400)]
locks: make generic_add_lease and generic_delete_lease static

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agocifs: use posix_unblock_lock instead of locks_delete_block
Jeff Layton [Fri, 21 Jun 2013 12:58:10 +0000 (08:58 -0400)]
cifs: use posix_unblock_lock instead of locks_delete_block

commit 66189be74 (CIFS: Fix VFS lock usage for oplocked files) exported
the locks_delete_block symbol. There's already an exported helper
function that provides this capability however, so make cifs use that
instead and turn locks_delete_block back into a static function.

Note that if fl->fl_next == NULL then this lock has already been through
locks_delete_block(), so we should be OK to ignore an ENOENT error here
and simply not retry the lock.

Cc: Pavel Shilovsky <piastryyy@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agolocks: drop the unused filp argument to posix_unblock_lock
Jeff Layton [Fri, 21 Jun 2013 12:58:09 +0000 (08:58 -0400)]
locks: drop the unused filp argument to posix_unblock_lock

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoDon't pass inode to ->d_hash() and ->d_compare()
Linus Torvalds [Tue, 21 May 2013 22:22:44 +0000 (15:22 -0700)]
Don't pass inode to ->d_hash() and ->d_compare()

Instances either don't look at it at all (the majority of cases) or
only want it to find the superblock (which can be had as dentry->d_sb).
A few cases that want more are actually safe with dentry->d_inode -
the only precaution needed is the check that it hadn't been replaced with
NULL by rmdir() or by overwriting rename(), which case should be simply
treated as cache miss.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agominix: bug widening a binary "not" operation
Dan Carpenter [Wed, 19 Jun 2013 00:08:10 +0000 (10:08 +1000)]
minix: bug widening a binary "not" operation

"chunk_size" is an unsigned int and "pos" is an unsigned long.  The
"& ~(chunk_size-1)" operation clears the high 32 bits unintentionally.

The ALIGN() macro does the correct thing.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
11 years agosplice: lift checks from do_splice_from() into callers
Al Viro [Wed, 19 Jun 2013 11:41:54 +0000 (15:41 +0400)]
splice: lift checks from do_splice_from() into callers

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoconstify rw_verify_area()
Al Viro [Wed, 19 Jun 2013 11:26:04 +0000 (15:26 +0400)]
constify rw_verify_area()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agops3flash: switch to generic_file_llseek_size()
Al Viro [Mon, 17 Jun 2013 06:05:35 +0000 (10:05 +0400)]
ps3flash: switch to generic_file_llseek_size()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agowlcore: use *ppos, not file->f_pos
Al Viro [Mon, 17 Jun 2013 11:26:19 +0000 (15:26 +0400)]
wlcore: use *ppos, not file->f_pos

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agobfa: switch to fixed_size_llseek()
Al Viro [Mon, 17 Jun 2013 13:45:46 +0000 (17:45 +0400)]
bfa: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agofnic: switch to fixed_size_llseek()
Al Viro [Mon, 17 Jun 2013 13:44:23 +0000 (17:44 +0400)]
fnic: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agovc: switch to fixed_size_llseek()
Al Viro [Mon, 17 Jun 2013 11:31:22 +0000 (15:31 +0400)]
vc: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoeisa_eeprom: switch to fixed_size_llseek()
Al Viro [Mon, 17 Jun 2013 11:27:47 +0000 (15:27 +0400)]
eisa_eeprom: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agobna: switch to fixed_size_llseek()
Al Viro [Mon, 17 Jun 2013 11:23:54 +0000 (15:23 +0400)]
bna: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agozorro: switch to fixed_size_llseek()
Al Viro [Sat, 22 Jun 2013 08:10:22 +0000 (12:10 +0400)]
zorro: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agomtdchar: switch to fixed_size_llseek()
Al Viro [Sun, 16 Jun 2013 16:27:42 +0000 (20:27 +0400)]
mtdchar: switch to fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agonew helper: fixed_size_llseek()
Al Viro [Sun, 16 Jun 2013 16:27:42 +0000 (20:27 +0400)]
new helper: fixed_size_llseek()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoecryptfs: switch ecryptfs_decode_and_decrypt_filename() from dentry to sb
Al Viro [Sun, 16 Jun 2013 16:05:38 +0000 (20:05 +0400)]
ecryptfs: switch ecryptfs_decode_and_decrypt_filename() from dentry to sb

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agofuse: another open-coded file_inode()
Al Viro [Sun, 16 Jun 2013 16:05:23 +0000 (20:05 +0400)]
fuse: another open-coded file_inode()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agobtrfs: more open-coded file_inode()
Al Viro [Sun, 16 Jun 2013 15:32:35 +0000 (19:32 +0400)]
btrfs: more open-coded file_inode()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agofanotify: quit wanking with FASYNC in ->release()
Al Viro [Sun, 16 Jun 2013 15:08:36 +0000 (19:08 +0400)]
fanotify: quit wanking with FASYNC in ->release()

... especially since there's no way to get that sucker
on the list fsnotify_fasync() works with - the only thing
adding to it is fsnotify_fasync() itself and it's never
called for fanotify files while they are opened.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agocomedi: quit wanking with FASYNC in ->release()
Al Viro [Sun, 16 Jun 2013 15:05:07 +0000 (19:05 +0400)]
comedi: quit wanking with FASYNC in ->release()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agomore open-coded file_inode() calls
Al Viro [Sun, 16 Jun 2013 13:25:12 +0000 (17:25 +0400)]
more open-coded file_inode() calls

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agokill find_inode_number()
Al Viro [Sat, 15 Jun 2013 07:37:47 +0000 (11:37 +0400)]
kill find_inode_number()

the only remaining caller (in ncpfs) is guaranteed to return 0 -
we only hit it if we'd just checked that there's no dentry with
such name.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agocoda: don't bother with find_inode_number()
Al Viro [Wed, 19 Jun 2013 09:21:03 +0000 (13:21 +0400)]
coda: don't bother with find_inode_number()

the fallback it's using for dcache misses is actually the
same value we would've used for inumber anyway.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoproc_fill_cache(): clean up, get rid of pointless find_inode_number() use
Al Viro [Sat, 15 Jun 2013 07:33:10 +0000 (11:33 +0400)]
proc_fill_cache(): clean up, get rid of pointless find_inode_number() use

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
11 years agoproc_fill_cache(): just make instantiate_t return int
Al Viro [Sat, 15 Jun 2013 07:15:20 +0000 (11:15 +0400)]
proc_fill_cache(): just make instantiate_t return int

all instances always return ERR_PTR(-E...) or NULL, anyway

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>