Skip to content

L
linux
Switch branch/tag
  1. 06 Nov, 2008 7 commits
    • Alan Stern's avatar
      Add round_jiffies_up and related routines · 9c133c46
      Alan Stern authored 
      This patch (as1158b) adds round_jiffies_up() and friends.  These
routines work like the analogous round_jiffies() functions, except
that they will never round down.

The new routines will be useful for timeouts where we don't care
exactly when the timer expires, provided it doesn't expire too soon.
Signed-off-by: default avatarAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      9c133c46
    • Tejun Heo's avatar
      block: fix __blkdev_get() for removable devices · 89f97496
      Tejun Heo authored 
      Commit 0762b8bd moved disk_get_part()
in front of recursive get on the whole disk, which caused removable
devices to try disk_get_part() before rescanning after a new media is
inserted, which might fail legit open attempts or give the old
partition.

This patch fixes the problem by moving disk_get_part() after
__blkdev_get() on the whole disk.

This problem was spotted by Borislav Petkov.
Signed-off-by: default avatarTejun Heo <tj@kernel.org>
Tested-by: default avatarBorislav Petkov <petkovbb@gmail.com>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      89f97496
    • Suresh Siddha's avatar
      generic-ipi: fix the smp_mb() placement · 561920a0
      Suresh Siddha authored 
      smp_mb() is needed (to make the memory operations visible globally) before
sending the ipi on the sender and the receiver (on Alpha atleast) needs
smp_read_barrier_depends() in the handler before reading the call_single_queue
list in a lock-free fashion.

On x86, x2apic mode register accesses for sending IPI's don't have serializing
semantics. So the need for smp_mb() before sending the IPI becomes more
critical in x2apic mode.

Remove the unnecessary smp_mb() in csd_flag_wait(), as the presence of that
smp_mb() doesn't mean anything on the sender, when the ipi receiver is not
doing any thing special (like memory fence) after clearing the CSD_FLAG_WAIT.
Signed-off-by: default avatarSuresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      561920a0
    • Mike Anderson's avatar
      blk: move blk_delete_timer call in end_that_request_last · e78042e5
      Mike Anderson authored 
      Move the calling  blk_delete_timer to later in end_that_request_last to
address an issue where blkdev_dequeue_request may have add a timer for the
request.
Signed-off-by: default avatarMike Anderson <andmike@linux.vnet.ibm.com>
Acked-by: default avatarTejun Heo <tj@kernel.org>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      e78042e5
    • Tejun Heo's avatar
      block: add timer on blkdev_dequeue_request() not elv_next_request() · 2920ebbd
      Tejun Heo authored 
      Block queue supports two usage models - one where block driver peeks
at the front of queue using elv_next_request(), processes it and
finishes it and the other where block driver peeks at the front of
queue, dequeue the request using blkdev_dequeue_request() and finishes
it.  The latter is more flexible as it allows the driver to process
multiple commands concurrently.

These two inconsistent usage models affect the block layer
implementation confusing.  For some, elv_next_request() is considered
the issue point while others consider blkdev_dequeue_request() the
issue point.

Till now the inconsistency mostly affect only accounting, so it didn't
really break anything seriously; however, with block layer timeout,
this inconsistency hits hard.  Block layer considers
elv_next_request() the issue point and adds timer but SCSI layer
thinks it was just peeking and when the request can't process the
command right away, it's just left there without further processing.
This makes the request dangling on the timer list and, when the timer
goes off, the request which the SCSI layer and below think is still on
the block queue ends up in the EH queue, causing various problems - EH
hang (failed count goes over busy count and EH never wakes up),
WARN_ON() and oopses as low level driver trying to handle the unknown
command, etc. depending on the timing.

As SCSI midlayer is the only user of block layer timer at the moment,
moving blk_add_timer() to elv_dequeue_request() fixes the problem;
however, this two usage models definitely need to be cleaned up in the
future.
Signed-off-by: default avatarTejun Heo <tj@kernel.org>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      2920ebbd
    • Jeremy Fitzhardinge's avatar
      bio: define __BIOVEC_PHYS_MERGEABLE · f92131c3
      Jeremy Fitzhardinge authored 
      Define __BIOVEC_PHYS_MERGEABLE as the default implementation of
BIOVEC_PHYS_MERGEABLE, so that its available for reuse within an
arch-specific definition of BIOVEC_PHYS_MERGEABLE.
Signed-off-by: default avatarJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      f92131c3
    • FUJITA Tomonori's avatar
      block: remove unused ll_new_mergeable() · 43381785
      FUJITA Tomonori authored 
      Signed-off-by: default avatarFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: default avatarJens Axboe <jens.axboe@oracle.com>
      43381785
  3. 04 Nov, 2008 15 commits
  5. 03 Nov, 2008 18 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7