linux-stable.git/block, branch linux-2.6.37.y

block: fix mis-synchronisation in blkdev_issue_zeroout()

2011-03-23T19:49:58+00:00

commit 0aeea18964173715a1037034ef6838198f319319 upstream.

BZ29402
https://bugzilla.kernel.org/show_bug.cgi?id=29402

We can hit serious mis-synchronization in bio completion path of
blkdev_issue_zeroout() leading to a panic.

The problem is that when we are going to wait_for_completion() in
blkdev_issue_zeroout() we check if the bb.done equals issued (number of
submitted bios). If it does, we can skip the wait_for_completition()
and just out of the function since there is nothing to wait for.
However, there is a ordering problem because bio_batch_end_io() is
calling atomic_inc(&bb->done) before complete(), hence it might seem to
blkdev_issue_zeroout() that all bios has been completed and exit. At
this point when bio_batch_end_io() is going to call complete(bb->wait),
bb and wait does not longer exist since it was allocated on stack in
blkdev_issue_zeroout() ==> panic!

(thread 1)                      (thread 2)
bio_batch_end_io()              blkdev_issue_zeroout()
  if(bb) {                      ...
    if (bb->end_io)             ...
      bb->end_io(bio, err);     ...
    atomic_inc(&bb->done);      ...
    ...                         while (issued != atomic_read(&bb.done))
    ...                         (let issued == bb.done)
    ...                         (do the rest of the function)
    ...                         return ret;
    complete(bb->wait);
    ^^^^^^^^
    panic

We can fix this easily by simplifying bio_batch and completion counting.

Also remove bio_end_io_t *end_io since it is not used.

Signed-off-by: Lukas Czerner 
Reported-by: Eric Whitney 
Tested-by: Eric Whitney 
Reviewed-by: Jeff Moyer 
CC: Dmitry Monakhov 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: blk-flush shouldn't call directly into q->request_fn() __blk_run_queue()

2011-03-07T23:05:24+00:00

commit 255bb490c8c27eed484d538efe6ef6a7473bd3f6 upstream.

blk-flush decomposes a flush into sequence of multiple requests.  On
completion of a request, the next one is queued; however, block layer
must not implicitly call into q->request_fn() directly from completion
path.  This makes the queue behave unexpectedly when seen from the
drivers and violates the assumption that q->request_fn() is called
with process context + queue_lock.

This patch makes blk-flush the following two changes to make sure
q->request_fn() is not called directly from request completion path.

- blk_flush_complete_seq_end_io() now asks __blk_run_queue() to always
  use kblockd instead of calling directly into q->request_fn().

- queue_next_fseq() uses ELEVATOR_INSERT_REQUEUE instead of
  ELEVATOR_INSERT_FRONT so that elv_insert() doesn't try to unplug the
  request queue directly.

Reported by Jan in the following threads.

 http://thread.gmane.org/gmane.linux.ide/48778
 http://thread.gmane.org/gmane.linux.ide/48786

stable: applicable to v2.6.37.

Signed-off-by: Tejun Heo 
Reported-by: Jan Beulich 
Cc: "David S. Miller" 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: add @force_kblockd to __blk_run_queue()

2011-03-07T23:05:23+00:00

commit 1654e7411a1ad4999fe7890ef51d2a2bbb1fcf76 upstream.

__blk_run_queue() automatically either calls q->request_fn() directly
or schedules kblockd depending on whether the function is recursed.
blk-flush implementation needs to be able to explicitly choose
kblockd.  Add @force_kblockd.

All the current users are converted to specify %false for the
parameter and this patch doesn't introduce any behavior change.

stable: This is prerequisite for fixing ide oops caused by the new
        blk-flush implementation.

Signed-off-by: Tejun Heo 
Cc: Jan Beulich 
Cc: James Bottomley 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

blk-throttle: Do not use kblockd workqueue for throtl work

2011-03-07T23:05:23+00:00

commit 450adcbe518ab3a3953d8475309525d22de77cba upstream.

o Dominik Klein reported a system hang issue while doing some blkio
  throttling testing.

  https://lkml.org/lkml/2011/2/24/173

o Some tracing revealed that CFQ was not dispatching any more jobs as
  queue unplug was not happening. And queue unplug was not happening
  because unplug work was not being called as there was one throttling
  work on same cpu which as not finished yet. And throttling work had not
  finished as it was tyring to dispatch a bio to CFQ but all the request
  descriptors were consume to it was put to sleep.

o So basically it is a cyclic dependecny between CFQ unplug work and
  throtl dispatch work. Tejun suggested that use separate workqueue for
  such cases.

o This patch uses a separate workqueue for throttle related work and
  does not rely on kblockd workqueue anymore.

Reported-by: Dominik Klein 
Signed-off-by: Vivek Goyal 
Acked-by: Tejun Heo 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

Fix over-zealous flush_disk when changing device size.

2011-03-07T23:05:12+00:00

commit 93b270f76e7ef3b81001576860c2701931cdc78b upstream.

There are two cases when we call flush_disk.
In one, the device has disappeared (check_disk_change) so any
data will hold becomes irrelevant.
In the oter, the device has changed size (check_disk_size_change)
so data we hold may be irrelevant.

In both cases it makes sense to discard any 'clean' buffers,
so they will be read back from the device if needed.

In the former case it makes sense to discard 'dirty' buffers
as there will never be anywhere safe to write the data.  In the
second case it *does*not* make sense to discard dirty buffers
as that will lead to file system corruption when you simply enlarge
the containing devices.

flush_disk calls __invalidate_devices.
__invalidate_device calls both invalidate_inodes and invalidate_bdev.

invalidate_inodes *does* discard I_DIRTY inodes and this does lead
to fs corruption.

invalidate_bev *does*not* discard dirty pages, but I don't really care
about that at present.

So this patch adds a flag to __invalidate_device (calling it
__invalidate_device2) to indicate whether dirty buffers should be
killed, and this is passed to invalidate_inodes which can choose to
skip dirty inodes.

flusk_disk then passes true from check_disk_change and false from
check_disk_size_change.

dm avoids tripping over this problem by calling i_size_write directly
rathher than using check_disk_size_change.

md does use check_disk_size_change and so is affected.

This regression was introduced by commit 608aeef17a which causes
check_disk_size_change to call flush_disk, so it is suitable for any
kernel since 2.6.27.

Acked-by: Jeff Moyer 
Cc: Andrew Patterson 
Cc: Jens Axboe 
Signed-off-by: NeilBrown 
Signed-off-by: Greg Kroah-Hartman

cfq-iosched: Don't wait if queue already has requests.

2011-02-17T23:14:40+00:00

commit 02a8f01b5a9f396d0327977af4c232d0f94c45fd upstream.

Commit 7667aa0630407bc07dc38dcc79d29cc0a65553c1 added logic to wait for
the last queue of the group to become busy (have at least one request),
so that the group does not lose out for not being continuously
backlogged. The commit did not check for the condition that the last
queue already has some requests. As a result, if the queue already has
requests, wait_busy is set. Later on, cfq_select_queue() checks the
flag, and decides that since the queue has a request now and wait_busy
is set, the queue is expired.  This results in early expiration of the
queue.

This patch fixes the problem by adding a check to see if queue already
has requests. If it does, wait_busy is not set. As a result, time slices
do not expire early.

The queues with more than one request are usually buffered writers.
Testing shows improvement in isolation between buffered writers.

Signed-off-by: Justin TerAvest 
Reviewed-by: Gui Jianfeng 
Acked-by: Vivek Goyal 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

block: fix accounting bug on cross partition merges

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

commit 09e099d4bafea3b15be003d548bdf94b4b6e0e17 upstream.

/proc/diskstats would display a strange output as follows.

$ cat /proc/diskstats |grep sda
   8       0 sda 90524 7579 102154 20464 0 0 0 0 0 14096 20089
   8       1 sda1 19085 1352 21841 4209 0 0 0 0 4294967064 15689 4293424691
                                                ~~~~~~~~~~
   8       2 sda2 71252 3624 74891 15950 0 0 0 0 232 23995 1562390
   8       3 sda3 54 487 2188 92 0 0 0 0 0 88 92
   8       4 sda4 4 0 8 0 0 0 0 0 0 0 0
   8       5 sda5 81 2027 2130 138 0 0 0 0 0 87 137

Its reason is the wrong way of accounting hd_struct->in_flight. When a bio is
merged into a request belongs to different partition by ELEVATOR_FRONT_MERGE.

The detailed root cause is as follows.

Assuming that there are two partition, sda1 and sda2.

1. A request for sda2 is in request_queue. Hence sda1's hd_struct->in_flight
   is 0 and sda2's one is 1.

        | hd_struct->in_flight
   ---------------------------
   sda1 |          0
   sda2 |          1
   ---------------------------

2. A bio belongs to sda1 is issued and is merged into the request mentioned on
   step1 by ELEVATOR_BACK_MERGE. The first sector of the request is changed
   from sda2 region to sda1 region. However the two partition's
   hd_struct->in_flight are not changed.

        | hd_struct->in_flight
   ---------------------------
   sda1 |          0
   sda2 |          1
   ---------------------------

3. The request is finished and blk_account_io_done() is called. In this case,
   sda2's hd_struct->in_flight, not a sda1's one, is decremented.

        | hd_struct->in_flight
   ---------------------------
   sda1 |         -1
   sda2 |          1
   ---------------------------

The patch fixes the problem by caching the partition lookup
inside the request structure, hence making sure that the increment
and decrement will always happen on the same partition struct. This
also speeds up IO with accounting enabled, since it cuts down on
the number of lookups we have to do.

Also add a refcount to struct hd_struct to keep the partition in
memory as long as users exist. We use kref_test_and_get() to ensure
we don't add a reference to a partition which is going away.

Signed-off-by: Jerome Marchand 
Signed-off-by: Yasuaki Ishimatsu 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

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

block: max hardware sectors limit wrapper

2010-12-17T07:36:01+00:00

Implement blk_limits_max_hw_sectors() and make
blk_queue_max_hw_sectors() a wrapper around it.

DM needs this to avoid setting queue_limits' max_hw_sectors and
max_sectors directly.  dm_set_device_limits() now leverages
blk_limits_max_hw_sectors() logic to establish the appropriate
max_hw_sectors minimum (PAGE_SIZE).  Fixes issue where DM was
incorrectly setting max_sectors rather than max_hw_sectors (which
caused dm_merge_bvec()'s max_hw_sectors check to be ineffective).

Signed-off-by: Mike Snitzer 
Cc: stable@kernel.org
Acked-by: Martin K. Petersen 
Signed-off-by: Jens Axboe

block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead

2010-12-17T07:35:53+00:00

When stacking devices, a request_queue is not always available. This
forced us to have a no_cluster flag in the queue_limits that could be
used as a carrier until the request_queue had been set up for a
metadevice.

There were several problems with that approach. First of all it was up
to the stacking device to remember to set queue flag after stacking had
completed. Also, the queue flag and the queue limits had to be kept in
sync at all times. We got that wrong, which could lead to us issuing
commands that went beyond the max scatterlist limit set by the driver.

The proper fix is to avoid having two flags for tracking the same thing.
We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the
block layer merging functions. The queue_limit 'no_cluster' is turned
into 'cluster' to avoid double negatives and to ease stacking.
Clustering defaults to being enabled as before. The queue flag logic is
removed from the stacking function, and explicitly setting the cluster
flag is no longer necessary in DM and MD.

Reported-by: Ed Lin 
Signed-off-by: Martin K. Petersen 
Acked-by: Mike Snitzer 
Cc: stable@kernel.org
Signed-off-by: Jens Axboe