This is one of the changes related to journal recovery I mentioned a
couple weeks ago. We can get into a situation where there are only
readonly nodes currently mounting the fs, but there are journals that need
to be recovered. Since the readonly nodes can't recover journals, the
next rw mounter needs to go through and check all journals and recover any
that are dirty (i.e. what the first node to mount the fs does). This rw
mounter needs to skip the journals held by the existing readonly nodes.
Skipping those journals amounts to using the TRY flag on the journal locks
so acquiring the lock of a journal held by a readonly node will fail
instead of blocking indefinately.
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
At some stage, a mutex was added to gfs2_glock_put() without
checking all its call sites. Two of them were called from
under a spinlock causing random delays at various points and
crashes.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
When allocating memory to sort directory entries, use vmalloc()
rather than kmalloc() since for larger directories, the required
size can easily be graeter than the 128k maximum of kmalloc().
Also adding the first steps towards getting the AOP_TRUNCATED_PAGE
return code get in the glock code by flagging all places where we
request a glock and we are holding a page lock.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
A typo in the directory code was causing postmark to fail
somewhere in the allocation code, since it was unable to
find newly allocated directory leaf blocks under certain
circumstances.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
A small update to the journaling code to change the way that
the "extra" blocks are accounted for in the journal. These are
used at a rate of one per 503 metadata blocks or one per 251
journaled data blocks (or just one if the total number of journaled
blocks in the transaction is smaller). Since we are using them at
two different rates the old method of accounting for them no longer
works and we count them up as required.
Since the "per transaction" accounting can't handle this (there is no
fixed number of header blocks per transaction) we have to account for
it in the general journal code. We now require that each transaction
reserves more blocks than it actually needs to take account of the
possible extra blocks.
Also a final fix to dir.c to ensure that all ref counts are handled
correctly.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The last patch missed some other instances of incorrect ref counting,
this fixes all of those too.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This fixes a ref count bug that sometimes showed up a umount time
(causing it to hang) but it otherwise mostly harmless. At the same
time there are some clean ups including making the log operations
structures const, moving a memory allocation so that its not done
in the fast path of checking to see if there is an outstanding
transaction related to a particular glock.
Removes the sd_log_wrap varaible which was updated, but never actually
used anywhere. Updates the gfs2 ioctl() to run without the kernel lock
(which it never needed anyway). Removes the "invalidate inodes" loop
from GFS2's put_super routine. This is done in kill super anyway so
we don't need to do it here. The loop was also bogus in that if there
are any inodes "stuck" at this point its a bug and we need to know
about it rather than hide it by hanging forever.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This puts the finishing touches to the ioctl support and also
removes a couple of unused fields from GFS2's private per file
structure.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Originally from Nick Piggin, just adapted to the newer branch.
You can't check PageLRU without holding zone->lru_lock. The page
release code can get away with it only because the page refcount is 0 at
that point. Also, you can't reliably remove pages from the LRU unless
the refcount is 0. Ever.
Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Jens Axboe <axboe@suse.de>
Thanks to Andrew for the good explanation of why this is so. akpm writes:
If a page is under writeback and we remove it from pagecache, it's still
going to get written to disk. But the VFS no longer knows about that page,
nor that this page is about to modify disk blocks.
So there might be scenarios in which those
blocks-which-are-about-to-be-written-to get reused for something else.
When writeback completes, it'll scribble on those blocks.
This won't happen in ext2/ext3-style filesystems in normal mode because the
page has buffers and try_to_release_page() will fail.
But ext2 in nobh mode doesn't attach buffers at all - it just sticks the
page in a BIO, finds some new blocks, points the BIO at those blocks and
lets it rip.
While that write IO's in flight, someone could truncate the file. Truncate
won't block on the writeout because the page isn't in pagecache any more.
So truncate will the free the blocks from the file under the page's feet.
Then something else can reallocate those blocks. Then write data to them.
Now, the original write completes, corrupting the filesystem.
Signed-off-by: Jens Axboe <axboe@suse.de>
By cleaning up the writeback logic (killing write_one_page() and the manual
set_page_dirty()), we can get rid of ->stolen inside the pipe_buffer and
just keep it local in pipe_to_file().
This also adds dirty page balancing logic and O_SYNC handling.
Signed-off-by: Jens Axboe <axboe@suse.de>
Clear the entire range, and don't increment pidx or we keep filling
the same position again and again.
Thanks to KAMEZAWA Hiroyuki.
Signed-off-by: Jens Axboe <axboe@suse.de>
* git://oss.sgi.com:8090/oss/git/xfs-2.6:
[XFS] Provide XFS support for the splice syscall.
[XFS] Reenable write barriers by default.
[XFS] Make project quota enforcement return an error code consistent with
[XFS] Implement the silent parameter to fill_super, previously ignored.
[XFS] Cleanup comment to remove reference to obsoleted function
* master.kernel.org:/pub/scm/linux/kernel/git/perex/alsa: (28 commits)
[ALSA] Kconfig SND_SEQUENCER_OSS help text fix
[ALSA] Add Aux input switch control for Aureon Universe
[ALSA] pcxhr - Fix the crash with REV01 board
[ALSA] sound/pci/hda: use create_singlethread_workqueue()
[ALSA] hda-intel - Add support of ATI SB600
[ALSA] cs4281 - Fix the check of timeout in probe
[ALSA] cs4281 - Fix the check of right channel
[ALSA] Test volume resolution of usb audio at initialization
[ALSA] maestro3.c: fix BUG, optimization
[ALSA] HDA/Realtek: multiple input mux definitions and pin mode additions
[ALSA] AdLib FM card driver
[ALSA] Fix / clean up PCM-OSS setup hooks
[ALSA] Clean up PCM codes (take 2)
[ALSA] Tiny clean up of PCM codes
[ALSA] ISA drivers bailing on first !enable[i]
[ALSA] Remove obsolete kfree_nocheck call
[ALSA] Remove obsolete kfree_nocheck call
[ALSA] Add snd-als300 driver for Avance Logic ALS300/ALS300+ soundcards
[ALSA] Add snd-riptide driver for Conexant Riptide chip
[ALSA] hda-codec - Fix noisy output wtih AD1986A 3stack model
...
No one should be writing a PAGE_SIZE worth of data to a normal sysfs
file, so properly terminate the buffer.
Thanks to Al Viro for pointing out my supidity here.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This reverts the mutex conversion that was recently done to the hdaps
driver; this coversion was buggy because the hdaps driver started using
this semaphore in IRQ context, which mutexes do not allow. Easiest
solution for now is to just revert the patch (the patch was part of a
bigger GIT commit, 9a61bf6300 but this
only reverts this one file)
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* git://git.kernel.org/pub/scm/linux/kernel/git/bunk/trivial: (48 commits)
Documentation: fix minor kernel-doc warnings
BUG_ON() Conversion in drivers/net/
BUG_ON() Conversion in drivers/s390/net/lcs.c
BUG_ON() Conversion in mm/slab.c
BUG_ON() Conversion in mm/highmem.c
BUG_ON() Conversion in kernel/signal.c
BUG_ON() Conversion in kernel/signal.c
BUG_ON() Conversion in kernel/ptrace.c
BUG_ON() Conversion in ipc/shm.c
BUG_ON() Conversion in fs/freevxfs/
BUG_ON() Conversion in fs/udf/
BUG_ON() Conversion in fs/sysv/
BUG_ON() Conversion in fs/inode.c
BUG_ON() Conversion in fs/fcntl.c
BUG_ON() Conversion in fs/dquot.c
BUG_ON() Conversion in md/raid10.c
BUG_ON() Conversion in md/raid6main.c
BUG_ON() Conversion in md/raid5.c
Fix minor documentation typo
BFP->BPF in Documentation/networking/tuntap.txt
...
sbp2util_mark_command_completed takes a lock which was already taken by
sbp2scsi_complete_all_commands. This is a regression in Linux 2.6.15.
Reported by Kristian Harms at
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=187394
[ More complete commentary, as response to questions by Andrew: ]
> This changes the call environment for all implementations of
> ->Current_done(). Are they all safe to call under this lock?
Short answer: Yes, trust me. ;-) Long answer:
The done() callbacks are passed on to sbp2 from the SCSI stack along
with each SCSI command via the queuecommand hook. The done() callback
is safe to call in atomic context. So does
Documentation/scsi/scsi_mid_low_api.txt say, and many if not all SCSI
low-level handlers rely on this fact. So whatever this callback does,
it is "self-contained" and it won't conflict with sbp2's internal ORB
list handling. In particular, it won't race with the
sbp2_command_orb_lock.
Moreover, sbp2 already calls the done() handler with
sbp2_command_orb_lock taken in sbp2scsi_complete_all_commands(). I
admit this is ultimately no proof of correctness, especially since this
portion of code introduced the spinlock recursion in the first place and
we didn't realize it since this code's submission before 2.6.15 until
now. (I have learned a lesson from this.)
I stress-tested my patch on x86 uniprocessor with a preemptible SMP
kernel (alas I have no SMP machine yet) and made sure that all code
paths which involve the sbp2_command_orb_lock were gone through multiple
times.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
IB/ipath: kbuild infrastructure
IB/ipath: infiniband verbs support
IB/ipath: misc infiniband code, part 2
IB/ipath: misc infiniband code, part 1
IB/ipath: infiniband RC protocol support
IB/ipath: infiniband UC and UD protocol support
IB/ipath: infiniband header files
IB/ipath: layering interfaces used by higher-level driver code
IB/ipath: support for userspace apps using core driver
IB/ipath: sysfs and ipathfs support for core driver
IB/ipath: misc driver support code
IB/ipath: chip initialisation code, and diag support
IB/ipath: support for PCI Express devices
IB/ipath: support for HyperTransport devices
IB/ipath: core driver header files
IB/ipath: core device driver
* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
[SPARC]: Wire up sys_sync_file_range() into syscall tables.
[SPARC]: Wire up sys_splice() into the syscall tables.
[SPARC64]: Update defconfig.
[SPARC64]: Align address in huge_pte_alloc().
[SPARC64]: Document the instruction checks we do in do_sparc64_fault().
[SPARC64]: Make tsb_sync() mm comparison more precise.
It doesn't make the splice itself necessarily nonblocking (because the
actual file descriptors that are spliced from/to may block unless they
have the O_NONBLOCK flag set), but it makes the splice pipe operations
nonblocking.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Patch from Pavel Pisa
This patch adds support of i.MX/MX1 SD/MMC controller.
It has been significantly redesigned from the original Sascha Hauer's
version to support scatter-gather DMA, to conform to latest Pierre Ossman's
and Russell King's MMC-SD Linux 2.6.x infrastructure.
The handling of all events has been moved to the softirq context
and is designed with no busy-looping in mind. Unfortunately
some controller bugs has to be overcome by limited looping
about 2-20 usec but these are observed only for initial card
recognition phase.
There are still some missing/missed IRQs problems under heavy load.
Help of somebody with access to the full SDHC design information
is probably necessary.
Regenerated against 2.6.16-git-060402 to solve clash with other patches.
Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Acked-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Patch from Andrew Victor
This patch adds support for the MMC/SD card interface on the Atmel
AT91RM9200 processor.
Original driver was by Nick Randell, but a number of people have
subsequently worked on it. It's currently maintained by Malcolm Noyes.
Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
(We'd only added the number, which meant that actually trying
to use splice just went off into la-la-land)
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Watchdog driver for the Atmel AT91RM9200 processor.
Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Andrew Morton <akpm@osdl.org>
The Coverity checker noted that this resulted in a NULL pointer
reference if we were coming from
if (usb_pcwd == NULL) {
printk(KERN_ERR PFX "Out of memory\n");
goto error;
}
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Andrew Morton <akpm@osdl.org>
change sprintf(pcwd_private.fw_ver_str, "ERROR");
to strcpy... as pointed out by Andrew Morton.
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Andrew Morton <akpm@osdl.org>
removal of includes (since we don't use kmalloc and
TASK_INTERRUPTABLE anymore).
Addition of missing commands.
Printk that lets the user know when the module was
unloaded.
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Patch from Tony Lindgren
This patch syncs OMAP board support with linux-omap tree.
The highlights of the patch are:
- Add support for Nokia 770 by Juha Yrjola
- Add support for Samsung Apollon by Kyungmin Park
- Add support for Amstrad E3 videophone by Jonathan McDowell
- Remove board-netstar.c board support as requested by Ladislav Michl
- Do platform_device registration in board files by Komal Shah et al.
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Patch from Tony Lindgren
Update misc OMAP core code from linux-omap tree:
- McBSP updates by Samuel Ortiz, Andrzej Zaborowski
- Whitespace cleanups by Ladislav Michl
- Other fixes by various linux-omap developers
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Patch from Tony Lindgren
Update OMAP framebuffer low-level init code from linux-omap tree
by Imre Deak.
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Patch from Tony Lindgren
Update OMAP PM code from linux-omap tree:
- Move PM code from plat-omap to mach-omap1 and mach-omap2
by Tony Lindgren
- Add minimal PM support for omap24xx by Tony Lindgren and
Richard Woodruff
- Misc updates to omap1 PM code by Tuukka Tikkanen et al
- Updates to the SRAM code needed for PM and FB by Imre Deak
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Patch from Tony Lindgren
Update OMAP GPIO code from linux-omap tree:
- Fix omap16xx edge control by Juha Yrjola
- Support for additional omap16xx trigger modes by Dirk Behme
- Fix edge detection by Tony Lindgren et al.
- Better support for omap15xx and omap310 by Andrej Zaborowski
- Fix omap15xx interrupt bug by Petukhov Nikolay
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
