If we share the tag map between two or more queues, then we cannot
use __set_bit() to set the bit. In fact we need to make sure we
atomically acquire this tag, so loop using test_and_set_bit() to
protect from that.
Noticed by Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't need to disable irqs to clear current->io_context, it is protected
by ->alloc_lock. Even IF it was possible to submit I/O from IRQ on behalf of
current this irq_disable() can't help: current_io_context() will re-instantiate
->io_context after irq_enable().
We don't need task_lock() or local_irq_disable() to clear ioc->task. This can't
prevent other CPUs from playing with our io_context anyway.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We can use this information for making more intelligent priority
decisions, and it will also be useful for blktrace.
Signed-off-by: Jens Axboe <axboe@suse.de>
It can make sense to set read-ahead larger than a single request.
We should not be enforcing such policy on the user. Additionally,
using the BLKRASET ioctl doesn't impose such a restriction. So
additionally we now expose identical behaviour through the two.
Issue also reported by Anton <cbou@mail.ru>
Signed-off-by: Jens Axboe <axboe@suse.de>
CFQ implements this on its own now, but it's really block layer
knowledge. Tells a device queue to start dispatching requests to
the driver, taking care to unplug if needed. Also fixes the issue
where as/cfq will invoke a stopped queue, which we really don't
want.
Signed-off-by: Jens Axboe <axboe@suse.de>
Kill a few inlines that bring in too much code to more than one location
Shrinks kernel text by about 300 bytes on 32-bit x86.
Signed-off-by: Jens Axboe <axboe@suse.de>
cfq_exit_lock is protecting two things now:
- The per-ioc rbtree of cfq_io_contexts
- The per-cfqd linked list of cfq_io_contexts
The per-cfqd linked list can be protected by the queue lock, as it is (by
definition) per cfqd as the queue lock is.
The per-ioc rbtree is mainly used and updated by the process itself only.
The only outside use is the io priority changing. If we move the
priority changing to not browsing the rbtree, we can remove any locking
from the rbtree updates and lookup completely. Let the sys_ioprio syscall
just mark processes as having the iopriority changed and lazily update
the private cfq io contexts the next time io is queued, and we can
remove this locking as well.
Signed-off-by: Jens Axboe <axboe@suse.de>
- Don't assign variables that are only used once.
- Kill spin_lock() prefetching, it's opportunistic at best.
Signed-off-by: Jens Axboe <axboe@suse.de>
After Christophs SCSI change, the only usage left is RQ_ACTIVE
and RQ_INACTIVE. The block layer sets RQ_INACTIVE right before freeing
the request, so any check for RQ_INACTIVE in a driver is a bug and
indicates use-after-free.
So kill/clean the remaining users, straight forward.
Signed-off-by: Jens Axboe <axboe@suse.de>
It is always identical to &q->rq, and we only use it for detecting
whether this request came out of our mempool or not. So replace it
with an additional ->flags bit flag.
Signed-off-by: Jens Axboe <axboe@suse.de>
As the comments indicates in blkdev.h, we can fold it into ->end_io_data
usage as that is really what ->waiting is. Fixup the users of
blk_end_sync_rq().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The rbtree sort/lookup/reposition logic is mostly duplicated in
cfq/deadline/as, so move it to the elevator core. The io schedulers
still provide the actual rb root, as we don't want to impose any sort
of specific handling on the schedulers.
Introduce the helpers and rb_node in struct request to help migrate the
IO schedulers.
Signed-off-by: Jens Axboe <axboe@suse.de>
Right now, every IO scheduler implements its own backmerging (except for
noop, which does no merging). That results in duplicated code for
essentially the same operation, which is never a good thing. This patch
moves the backmerging out of the io schedulers and into the elevator
core. We save 1.6kb of text and as a bonus get backmerging for noop as
well. Win-win!
Signed-off-by: Jens Axboe <axboe@suse.de>
Right now ->flags is a bit of a mess: some are request types, and
others are just modifiers. Clean this up by splitting it into
->cmd_type and ->cmd_flags. This allows introduction of generic
Linux block message types, useful for sending generic Linux commands
to block devices.
Signed-off-by: Jens Axboe <axboe@suse.de>
The current block queue implementation already contains most of the
machinery for shared tag maps. The only remaining pieces are a way to
allocate and destroy a tag map independently of the queues (so that
the maps can be managed on the life cycle of the overseeing entity)
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
I know nothing about io scheduler, but I suspect set_task_ioprio() is not safe.
current_io_context() initializes "struct io_context", then sets ->io_context.
set_task_ioprio() running on another cpu may see the changes out of order, so
->set_ioprio(ioc) may use io_context which was not initialized properly.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Jens Axboe <axboe@suse.de>
lockdep needs to have the waitqueue lock initialized for on-stack waitqueues
implicitly initialized by DECLARE_COMPLETION(). Annotate on-stack completions
accordingly.
Has no effect on non-lockdep kernels.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The remaining counters in page_state after the zoned VM counter patches
have been applied are all just for show in /proc/vmstat. They have no
essential function for the VM.
We use a simple increment of per cpu variables. In order to avoid the most
severe races we disable preempt. Preempt does not prevent the race between
an increment and an interrupt handler incrementing the same statistics
counter. However, that race is exceedingly rare, we may only loose one
increment or so and there is no requirement (at least not in kernel) that
the vm event counters have to be accurate.
In the non preempt case this results in a simple increment for each
counter. For many architectures this will be reduced by the compiler to a
single instruction. This single instruction is atomic for i386 and x86_64.
And therefore even the rare race condition in an interrupt is avoided for
both architectures in most cases.
The patchset also adds an off switch for embedded systems that allows a
building of linux kernels without these counters.
The implementation of these counters is through inline code that hopefully
results in only a single instruction increment instruction being emitted
(i386, x86_64) or in the increment being hidden though instruction
concurrency (EPIC architectures such as ia64 can get that done).
Benefits:
- VM event counter operations usually reduce to a single inline instruction
on i386 and x86_64.
- No interrupt disable, only preempt disable for the preempt case.
Preempt disable can also be avoided by moving the counter into a spinlock.
- Handling is similar to zoned VM counters.
- Simple and easily extendable.
- Can be omitted to reduce memory use for embedded use.
References:
RFC http://marc.theaimsgroup.com/?l=linux-kernel&m=113512330605497&w=2
RFC http://marc.theaimsgroup.com/?l=linux-kernel&m=114988082814934&w=2
local_t http://marc.theaimsgroup.com/?l=linux-kernel&m=114991748606690&w=2
V2 http://marc.theaimsgroup.com/?t=115014808400007&r=1&w=2
V3 http://marc.theaimsgroup.com/?l=linux-kernel&m=115024767022346&w=2
V4 http://marc.theaimsgroup.com/?l=linux-kernel&m=115047968808926&w=2
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make use the of newly defined hotplug version of cpu_notifier functionality
wherever appropriate.
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Do a safer check for when to enable DMA. Currently we enable ISA DMA
for cases that do not need it, resulting in OOM conditions when ZONE_DMA
runs out of space.
Signed-off-by: Jens Axboe <axboe@suse.de>
A process flag to indicate whether we are doing sync io is incredibly
ugly. It also causes performance problems when one does a lot of async
io and then proceeds to sync it. Part of the io will go out as async,
and the other part as sync. This causes a disconnect between the
previously submitted io and the synced io. For io schedulers such as CFQ,
this will cause us lost merges and suboptimal behaviour in scheduling.
Remove PF_SYNCWRITE completely from the fsync/msync paths, and let
the O_DIRECT path just directly indicate that the writes are sync
by using WRITE_SYNC instead.
Signed-off-by: Jens Axboe <axboe@suse.de>
The queue lock can be taken from interrupts so it must always be taken with
irq disabling primitives. Some primitives already verify this.
blk_start_queue() is called under this lock, so interrupts must be
disabled.
Also document this requirement clearly in blk_init_queue(), where the queue
spinlock is set.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jens Axboe <axboe@suse.de>
list_splice_init(list, head) does unneeded job if it is known that
list_empty(head) == 1. We can use list_replace_init() instead.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
While executing barrrier sequence, the bar_rq which carries actual
write was accounted as normal IO on completion, while it wasn't on
queueing. This caused gendisk->in_flight to be decremented by 1 after
each barrier thus messed up statistics.
This patch makes bar_rq not accounted as normal IO. As the containing
barrier request as a whole is accounted, part of it shouldn't be.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't recurse back into the driver even if the unplug threshold is met,
when the driver asks for a requeue. This is both silly from a logical
point of view (requeues typically happen due to driver/hardware
shortage), and also dangerous since we could hit an endless request_fn
-> requeue -> unplug -> request_fn loop and crash on stack overrun.
Also limit blk_run_queue() to one level of recursion, similar to how
blk_start_queue() works.
This patch fixed a real problem with SLES10 and lpfc, and it could hit
any SCSI lld that returns non-zero from it's ->queuecommand() handler.
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Few of the notifier_chain_register() callers use __devinitdata in the
definition of notifier_block data structure. It is incorrect as the
data structure should be available after the initializations (they do
not unregister them during initializations).
This was leading to an oops when notifier_chain_register() call is
invoked for those callback chains after initialization.
This patch fixes all such usages to _not_ have the notifier_block data
structure in the init data section.
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch updates the comments to match the actual code.
Signed-off-by: Martin Waitz <tali@admingilde.org>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
On setups with many disks, we spend a considerable amount of time
looking up the process-disk mapping on each queue of io. Testing with
a NULL based block driver, this costs 40-50% reduction in throughput
for 1000 disks.
Signed-off-by: Jens Axboe <axboe@suse.de>
This flag should be set for a virtual device iff it is set for all underlying
devices.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Both elv_add_request() and generic_unplug_device() grab the queue lock
and disable interrupts, do that locally and use the __ variants.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jens Axboe <axboe@suse.de>
The low level PCI DMA mapping functions should handle it in most cases.
This should fix problems with depleting the DMA zone early. The old
code used precious GFP_DMA memory in many cases where it was not needed.
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
q->ordcolor must only be flipped on initial queueing of a hardbarrier
request.
Constructing ordered sequence and requeueing used to pass through
__elv_add_request() which flips q->ordcolor when it sees a barrier
request.
This patch separates out elv_insert() from __elv_add_request() and uses
elv_insert() when constructing ordered sequence and requeueing.
elv_insert() inserts the given request at the specified position and
does nothing else.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Acked-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Previously, if a fs request which was being drained failed and got
requeued, blk_do_ordered() didn't allow it to be reissued, which causes
queue stall. This patch makes blk_do_ordered() use the sequence of each
request to determine whether a request can be issued or not. This fixes
the bug and simplifies code.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Acked-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jens Axboe