Linux currently has two models for block devices:
- The classic request_fn based approach, where drivers use struct
request units for IO. The block layer provides various helper
functionalities to let drivers share code, things like tag
management, timeout handling, queueing, etc.
- The "stacked" approach, where a driver squeezes in between the
block layer and IO submitter. Since this bypasses the IO stack,
driver generally have to manage everything themselves.
With drivers being written for new high IOPS devices, the classic
request_fn based driver doesn't work well enough. The design dates
back to when both SMP and high IOPS was rare. It has problems with
scaling to bigger machines, and runs into scaling issues even on
smaller machines when you have IOPS in the hundreds of thousands
per device.
The stacked approach is then most often selected as the model
for the driver. But this means that everybody has to re-invent
everything, and along with that we get all the problems again
that the shared approach solved.
This commit introduces blk-mq, block multi queue support. The
design is centered around per-cpu queues for queueing IO, which
then funnel down into x number of hardware submission queues.
We might have a 1:1 mapping between the two, or it might be
an N:M mapping. That all depends on what the hardware supports.
blk-mq provides various helper functions, which include:
- Scalable support for request tagging. Most devices need to
be able to uniquely identify a request both in the driver and
to the hardware. The tagging uses per-cpu caches for freed
tags, to enable cache hot reuse.
- Timeout handling without tracking request on a per-device
basis. Basically the driver should be able to get a notification,
if a request happens to fail.
- Optional support for non 1:1 mappings between issue and
submission queues. blk-mq can redirect IO completions to the
desired location.
- Support for per-request payloads. Drivers almost always need
to associate a request structure with some driver private
command structure. Drivers can tell blk-mq this at init time,
and then any request handed to the driver will have the
required size of memory associated with it.
- Support for merging of IO, and plugging. The stacked model
gets neither of these. Even for high IOPS devices, merging
sequential IO reduces per-command overhead and thus
increases bandwidth.
For now, this is provided as a potential 3rd queueing model, with
the hope being that, as it matures, it can replace both the classic
and stacked model. That would get us back to having just 1 real
model for block devices, leaving the stacked approach to dm/md
devices (as it was originally intended).
Contributions in this patch from the following people:
Shaohua Li <shli@fusionio.com>
Alexander Gordeev <agordeev@redhat.com>
Christoph Hellwig <hch@infradead.org>
Mike Christie <michaelc@cs.wisc.edu>
Matias Bjorling <m@bjorling.me>
Jeff Moyer <jmoyer@redhat.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This was the only real user of BIO_CLONED, which didn't have very clear
semantics. Convert to its own flag so we can get rid of BIO_CLONED.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Martin K. Petersen <martin.petersen@oracle.com>
This is for the new bio splitting code. When we split a bio, if the
split occured on a bvec boundry we reuse the bvec for the new bio. But
that means bio_free() can't free it, hence the explicit flag.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
Acked-by: Tejun Heo <tj@kernel.org>
__bio_for_each_segment() iterates bvecs from the specified index
instead of bio->bv_idx. Currently, the only usage is to walk all the
bvecs after the bio has been advanced by specifying 0 index.
For immutable bvecs, we need to split these apart;
bio_for_each_segment() is going to have a different implementation.
This will also help document the intent of code that's using it -
bio_for_each_segment_all() is only legal to use for code that owns the
bio.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Neil Brown <neilb@suse.de>
CC: Boaz Harrosh <bharrosh@panasas.com>
This gets open coded quite a bit and it's tricky to get right, so make a
generic version and convert some existing users over to it instead.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
Random cleanup - this code was duplicated and it's not really specific
to md.
Also added the ability to return the actual error code.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>
Acked-by: Tejun Heo <tj@kernel.org>
This is prep work for immutable bio vecs; we first want to centralize
where bvecs are modified.
Next two patches convert some existing code to use this function.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
This adds a pointer to the bvec array to struct bio_integrity_payload,
instead of the bvecs always being inline; then the bvecs are allocated
with bvec_alloc_bs().
Changed bvec_alloc_bs() and bvec_free_bs() to take a pointer to a
mempool instead of the bioset, so that bio integrity can use a different
mempool for its bvecs, and thus avoid a potential deadlock.
This is eventually for immutable bio vecs - immutable bvecs aren't
useful if we still have to copy them, hence the need for the pointer.
Less code is always nice too, though.
Also, bio_integrity_alloc() was using fs_bio_set if no bio_set was
specified. This was wrong - using the bio_set doesn't protect us from
memory allocation failures, because we just used kmalloc for the
bio_integrity_payload. But it does introduce the possibility of
deadlock, if for some reason we weren't supposed to be using fs_bio_set.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Martin K. Petersen <martin.petersen@oracle.com>
bio_integrity_split() seemed to be confusing pointers and arrays -
bip_vec in bio_integrity_payload was an array appended to the end of the
payload, so the bio_vecs in struct bio_pair should have come after the
bio_integrity_payload they're for.
Fix it by making bip_vec a pointer to the inline vecs - a later patch is
going to make more use of this pointer.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Martin K. Petersen <martin.petersen@oracle.com>
Previously, if we ever try to allocate more than once from the same bio
set while running under generic_make_request() (i.e. a stacking block
driver), we risk deadlock.
This is because of the code in generic_make_request() that converts
recursion to iteration; any bios we submit won't actually be submitted
(so they can complete and eventually be freed) until after we return -
this means if we allocate a second bio, we're blocking the first one
from ever being freed.
Thus if enough threads call into a stacking block driver at the same
time with bios that need multiple splits, and the bio_set's reserve gets
used up, we deadlock.
This can be worked around in the driver code - we could check if we're
running under generic_make_request(), then mask out __GFP_WAIT when we
go to allocate a bio, and if the allocation fails punt to workqueue and
retry the allocation.
But this is tricky and not a generic solution. This patch solves it for
all users by inverting the previously described technique. We allocate a
rescuer workqueue for each bio_set, and then in the allocation code if
there are bios on current->bio_list we would be blocking, we punt them
to the rescuer workqueue to be submitted.
This guarantees forward progress for bio allocations under
generic_make_request() provided each bio is submitted before allocating
the next, and provided the bios are freed after they complete.
Note that this doesn't do anything for allocation from other mempools.
Instead of allocating per bio data structures from a mempool, code
should use bio_set's front_pad.
Tested it by forcing the rescue codepath to be taken (by disabling the
first GFP_NOWAIT) attempt, and then ran it with bcache (which does a lot
of arbitrary bio splitting) and verified that the rescuer was being
invoked.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Muthukumar Ratty <muthur@gmail.com>
This is prep work for the next patch, which embeds a struct bio_list in
struct bio_set.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
The WRITE SAME command supported on some SCSI devices allows the same
block to be efficiently replicated throughout a block range. Only a
single logical block is transferred from the host and the storage device
writes the same data to all blocks described by the I/O.
This patch implements support for WRITE SAME in the block layer. The
blkdev_issue_write_same() function can be used by filesystems and block
drivers to replicate a buffer across a block range. This can be used to
efficiently initialize software RAID devices, etc.
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove special-casing of non-rw fs style requests (discard). The nomerge
flags are consolidated in blk_types.h, and rq_mergeable() and
bio_mergeable() have been modified to use them.
bio_is_rw() is used in place of bio_has_data() a few places. This is
done to to distinguish true reads and writes from other fs type requests
that carry a payload (e.g. write same).
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Previously, there was bio_clone() but it only allocated from the fs bio
set; as a result various users were open coding it and using
__bio_clone().
This changes bio_clone() to become bio_clone_bioset(), and then we add
bio_clone() and bio_clone_kmalloc() as wrappers around it, making use of
the functionality the last patch adedd.
This will also help in a later patch changing how bio cloning works.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: NeilBrown <neilb@suse.de>
CC: Alasdair Kergon <agk@redhat.com>
CC: Boaz Harrosh <bharrosh@panasas.com>
CC: Jeff Garzik <jeff@garzik.org>
Acked-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Previously, bio_kmalloc() and bio_alloc_bioset() behaved slightly
different because there was some almost-duplicated code - this fixes
some of that.
The important change is that previously bio_kmalloc() always set
bi_io_vec = bi_inline_vecs, even if nr_iovecs == 0 - unlike
bio_alloc_bioset(). This would cause bio_has_data() to return true; I
don't know if this resulted in any actual bugs but it was certainly
wrong.
bio_kmalloc() and bio_alloc_bioset() also have different arbitrary
limits on nr_iovecs - 1024 (UIO_MAXIOV) for bio_kmalloc(), 256
(BIO_MAX_PAGES) for bio_alloc_bioset(). This patch doesn't fix that, but
at least they're enforced closer together and hopefully they will be
fixed in a later patch.
This'll also help with some future cleanups - there are a fair number of
functions that allocate bios (e.g. bio_clone()), and now they don't have
to be duplicated for bio_alloc(), bio_alloc_bioset(), and bio_kmalloc().
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
v7: Re-add dropped comments, improv patch description
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that we've got generic code for freeing bios allocated from bio
pools, this isn't needed anymore.
This patch also makes bio_free() static, since without bi_destructor
there should be no need for it to be called anywhere else.
bio_free() is now only called from bio_put, so we can refactor those a
bit - move some code from bio_put() to bio_free() and kill the redundant
bio->bi_next = NULL.
v5: Switch to BIO_KMALLOC_POOL ((void *)~0), per Boaz
v6: BIO_KMALLOC_POOL now NULL, drop bio_free's EXPORT_SYMBOL
v7: No #define BIO_KMALLOC_POOL anymore
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Reusing bios is something that's been highly frowned upon in the past,
but driver code keeps doing it anyways. If it's going to happen anyways,
we should provide a generic method.
This'll help with getting rid of bi_destructor - drivers/block/pktcdvd.c
was open coding it, by doing a bio_init() and resetting bi_destructor.
This required reordering struct bio, but the block layer is not yet
nearly fast enough for any cacheline effects to matter here.
v5: Add a define BIO_RESET_BITS, to be very explicit about what parts of
bio->bi_flags are saved.
v6: Further commenting verbosity, per Tejun
v9: Add a function comment
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that bios keep track of where they were allocated from,
bio_integrity_alloc_bioset() becomes redundant.
Remove bio_integrity_alloc_bioset() and drop bio_set argument from the
related functions and make them use bio->bi_pool.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
CC: Jens Axboe <axboe@kernel.dk>
CC: Martin K. Petersen <martin.petersen@oracle.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
cgroup/for-3.5 contains the following changes which blk-cgroup needs
to proceed with the on-going cleanup.
* Dynamic addition and removal of cftypes to make config/stat file
handling modular for policies.
* cgroup removal update to not wait for css references to drain to fix
blkcg removal hang caused by cfq caching cfqgs.
Pull in cgroup/for-3.5 into block/for-3.5/core. This causes the
following conflicts in block/blk-cgroup.c.
* 761b3ef50e "cgroup: remove cgroup_subsys argument from callbacks"
conflicts with blkiocg_pre_destroy() addition and blkiocg_attach()
removal. Resolved by removing @subsys from all subsys methods.
* 676f7c8f84 "cgroup: relocate cftype and cgroup_subsys definitions in
controllers" conflicts with ->pre_destroy() and ->attach() updates
and removal of modular config. Resolved by dropping forward
declarations of the methods and applying updates to the relocated
blkio_subsys.
* 4baf6e3325 "cgroup: convert all non-memcg controllers to the new
cftype interface" builds upon the previous item. Resolved by adding
->base_cftypes to the relocated blkio_subsys.
Signed-off-by: Tejun Heo <tj@kernel.org>
Pull <linux/bug.h> cleanup from Paul Gortmaker:
"The changes shown here are to unify linux's BUG support under the one
<linux/bug.h> file. Due to historical reasons, we have some BUG code
in bug.h and some in kernel.h -- i.e. the support for BUILD_BUG in
linux/kernel.h predates the addition of linux/bug.h, but old code in
kernel.h wasn't moved to bug.h at that time. As a band-aid, kernel.h
was including <asm/bug.h> to pseudo link them.
This has caused confusion[1] and general yuck/WTF[2] reactions. Here
is an example that violates the principle of least surprise:
CC lib/string.o
lib/string.c: In function 'strlcat':
lib/string.c:225:2: error: implicit declaration of function 'BUILD_BUG_ON'
make[2]: *** [lib/string.o] Error 1
$
$ grep linux/bug.h lib/string.c
#include <linux/bug.h>
$
We've included <linux/bug.h> for the BUG infrastructure and yet we
still get a compile fail! [We've not kernel.h for BUILD_BUG_ON.] Ugh -
very confusing for someone who is new to kernel development.
With the above in mind, the goals of this changeset are:
1) find and fix any include/*.h files that were relying on the
implicit presence of BUG code.
2) find and fix any C files that were consuming kernel.h and hence
relying on implicitly getting some/all BUG code.
3) Move the BUG related code living in kernel.h to <linux/bug.h>
4) remove the asm/bug.h from kernel.h to finally break the chain.
During development, the order was more like 3-4, build-test, 1-2. But
to ensure that git history for bisect doesn't get needless build
failures introduced, the commits have been reorderd to fix the problem
areas in advance.
[1] https://lkml.org/lkml/2012/1/3/90
[2] https://lkml.org/lkml/2012/1/17/414"
Fix up conflicts (new radeon file, reiserfs header cleanups) as per Paul
and linux-next.
* tag 'bug-for-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux:
kernel.h: doesn't explicitly use bug.h, so don't include it.
bug: consolidate BUILD_BUG_ON with other bug code
BUG: headers with BUG/BUG_ON etc. need linux/bug.h
bug.h: add include of it to various implicit C users
lib: fix implicit users of kernel.h for TAINT_WARN
spinlock: macroize assert_spin_locked to avoid bug.h dependency
x86: relocate get/set debugreg fcns to include/asm/debugreg.
