linux-toradex.git/drivers/block, branch v2.6.37.2

nbd: remove module-level ioctl mutex

2011-02-17T23:14:50+00:00

commit de1f016f882e52facc3c8609599f827bcdd14af9 upstream.

Commit 2a48fc0ab242417 ("block: autoconvert trivial BKL users to private
mutex") replaced uses of the BKL in the nbd driver with mutex
operations.  Since then, I've been been seeing these lock ups:

 INFO: task qemu-nbd:16115 blocked for more than 120 seconds.
 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
 qemu-nbd      D 0000000000000001     0 16115  16114 0x00000004
  ffff88007d775d98 0000000000000082 ffff88007d775fd8 ffff88007d774000
  0000000000013a80 ffff8800020347e0 ffff88007d775fd8 0000000000013a80
  ffff880133730000 ffff880002034440 ffffea0004333db8 ffffffffa071c020
 Call Trace:
  [] __mutex_lock_slowpath+0xf7/0x180
  [] mutex_lock+0x2b/0x50
  [] nbd_ioctl+0x6c/0x1c0 [nbd]
  [] blkdev_ioctl+0x230/0x730
  [] block_ioctl+0x41/0x50
  [] do_vfs_ioctl+0x93/0x370
  [] sys_ioctl+0x81/0xa0
  [] system_call_fastpath+0x16/0x1b

Instrumenting the nbd module's ioctl handler with some extra logging
clearly shows the NBD_DO_IT ioctl being invoked which is a long-lived
ioctl in the sense that it doesn't return until another ioctl asks the
driver to disconnect.  However, that other ioctl blocks, waiting for the
module-level mutex that replaced the BKL, and then we're stuck.

This patch removes the module-level mutex altogether.  It's clearly
wrong, and as far as I can see, it's entirely unnecessary, since the nbd
driver maintains per-device mutexes, and I don't see anything that would
require a module-level (or kernel-level, for that matter) mutex.

Signed-off-by: Soren Hansen 
Acked-by: Serge Hallyn 
Acked-by: Paul Clements 
Cc: Arnd Bergmann 
Cc: Jens Axboe 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

Fix build error in drivers/block/cciss.c

2010-12-21T05:21:49+00:00

.. caused by a missing semi-colon, introduced in commit 0fc13c8995cd
("cciss: fix cciss_revalidate panic").

Reported-by: Stephen Rothwell 
Reported-by: Thiago Farina 
Cc: Jens Axboe 
Signed-off-by: Linus Torvalds

Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

2010-12-20T17:19:46+00:00

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  cciss: fix cciss_revalidate panic
  block: max hardware sectors limit wrapper
  block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead
  blk-throttle: Correct the placement of smp_rmb()
  blk-throttle: Trim/adjust slice_end once a bio has been dispatched
  block: check for proper length of iov entries earlier in blk_rq_map_user_iov()
  drbd: fix for spin_lock_irqsave in endio callback
  drbd: don't recvmsg with zero length

cciss: fix cciss_revalidate panic

2010-12-17T08:01:37+00:00

If you delete a logical drive, and then run BLKRRPART (e.g. via fdisk)
on a logical drive which is "after" the deleted logical drive in the h->drv[]
array, then cciss_revalidate panics because it will access the null pointer
h->drv[x] when x hits the deleted drive.

Signed-off-by: Stephen M. Cameron 
Cc: stable@kernel.org
Signed-off-by: Jens Axboe

xen: Provide a variant of __RING_SIZE() that is an integer constant expression

2010-12-15T20:34:28+00:00

Without this, gcc 4.5 won't compile xen-netfront and xen-blkfront, where
this is being used to specify array sizes.

Signed-off-by: Jan Beulich 
Signed-off-by: Jeremy Fitzhardinge 
Cc: Jens Axboe 
Cc: David Miller 
Cc: Stable Kernel 
Signed-off-by: Linus Torvalds

Merge branch 'rbd-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client

2010-12-02T16:05:22+00:00

* 'rbd-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: replace the rbd sysfs interface

rbd: replace the rbd sysfs interface

2010-12-01T23:53:22+00:00

The new interface creates directories per mapped image
and under each it creates a subdir per available snapshot.
This allows keeping a cleaner interface within the sysfs
guidelines. The ABI documentation was updated too.

Acked-by: Greg Kroah-Hartman 
Signed-off-by: Yehuda Sadeh 
Signed-off-by: Sage Weil

drbd: fix for spin_lock_irqsave in endio callback

2010-11-27T18:50:43+00:00

In commit 9b7f76dc37919ea36caa9680a3f765e5b19b25fb,
 Author: Lars Ellenberg 
 Date:   Wed Aug 11 23:40:24 2010 +0200

    drbd: new configuration parameter c-min-rate

a bad chunk slipped through, which is now reverted as well,
restoring the correct irqsave for the endio callback.

This patch also add comments at both req_mod()
and in the endio callback so it should not happen again.

Signed-off-by: Philipp Reisner 
Signed-off-by: Lars Ellenberg

drbd: don't recvmsg with zero length

2010-11-27T18:50:43+00:00

This should fix a performance degradation we observed recently.

If we don't expect any subheader, we should not call into the tcp stack,
as that may add considerable latency if there is no data available at
this point.

For a synthetic synchronous write load with single outstanding writes,
this additional latency when processing the "unplug remote" packet
added up to a performance degradation factor >= 10.

Signed-off-by: Philipp Reisner 
Signed-off-by: Lars Ellenberg

Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

2010-11-26T22:17:50+00:00

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  cciss: fix build for PROC_FS disabled
  block: fix amiga and atari floppy driver compile warning
  blk-throttle: Fix calculation of max number of WRITES to be dispatched
  ioprio: grab rcu_read_lock in sys_ioprio_{set,get}()
  xen/blkfront: cope with backend that fail empty BLKIF_OP_WRITE_BARRIER requests
  xen/blkfront: Implement FUA with BLKIF_OP_WRITE_BARRIER
  xen/blkfront: change blk_shadow.request to proper pointer
  xen/blkfront: map REQ_FLUSH into a full barrier