Dmitry Vyukov managed to trigger a case where aio_migratepage can cause a
use-after-free during teardown of the aio ring buffer's mapping. This turns
out to be caused by access to the ioctx's ring_pages via the migratepage
operation which was not being protected by any locks during ioctx freeing.
Use the address_space's private_lock to protect use and updates of the mapping's
private_data, and make ioctx teardown unlink the ioctx from the address space.
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Tested-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Prior to the introduction of page migration support in "fs/aio: Add support
to aio ring pages migration" / 36bc08cc01,
mapping of the ring buffer pages was done via get_user_pages() while
retaining mmap_sem held for write. This avoided possible races with userland
racing an munmap() or mremap(). The page migration patch, however, switched
to using mm_populate() to prime the page mapping. mm_populate() cannot be
called with mmap_sem held.
Instead of dropping the mmap_sem, revert to the old behaviour and simply
drop the use of mm_populate() since get_user_pages() will cause the pages to
get mapped anyways. Thanks to Al Viro for spotting this issue.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Sseveral sparse warnings were caused by missing rcu_dereference() annotations
for dereferencing mm->ioctx_table. Thankfully, none of those were actual bugs
as the deref was protected by a spin lock in all instances.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
The commit 36bc08cc01 ("fs/aio: Add support to aio ring pages migration")
added some debugging code that is not required and resulted in a build error
when 98474236f7 ("vfs: make the dentry cache use the lockref infrastructure")
was added to the tree. The code is not required, so just delete it.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Another shortcoming of the table lookup patch was revealed where the pointer
was not being tested before being dereferenced. Verify this to avoid the
NULL pointer dereference.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
In the patch "aio: convert the ioctx list to table lookup v3", incorrect
handling in the ioctx_alloc() error path was introduced that lead to an
ioctx being added via ioctx_add_table() while freed when the ioctx_alloc()
call returned -EAGAIN due to hitting the aio_max_nr limit. Fix this by
only calling ioctx_add_table() as the last step in ioctx_alloc().
Also, several unnecessary rcu_dereference() calls were added that lead to
RCU warnings where the system was already protected by a spin lock for
accessing mm->ioctx_table.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
In the event that an overflow/underflow occurs while calculating req_batch,
clamp the minimum at 1 request instead of doing a BUG_ON().
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
On Wed, Jun 12, 2013 at 11:14:40AM -0700, Kent Overstreet wrote:
> On Mon, Apr 15, 2013 at 02:40:55PM +0300, Octavian Purdila wrote:
> > When using a large number of threads performing AIO operations the
> > IOCTX list may get a significant number of entries which will cause
> > significant overhead. For example, when running this fio script:
> >
> > rw=randrw; size=256k ;directory=/mnt/fio; ioengine=libaio; iodepth=1
> > blocksize=1024; numjobs=512; thread; loops=100
> >
> > on an EXT2 filesystem mounted on top of a ramdisk we can observe up to
> > 30% CPU time spent by lookup_ioctx:
> >
> > 32.51% [guest.kernel] [g] lookup_ioctx
> > 9.19% [guest.kernel] [g] __lock_acquire.isra.28
> > 4.40% [guest.kernel] [g] lock_release
> > 4.19% [guest.kernel] [g] sched_clock_local
> > 3.86% [guest.kernel] [g] local_clock
> > 3.68% [guest.kernel] [g] native_sched_clock
> > 3.08% [guest.kernel] [g] sched_clock_cpu
> > 2.64% [guest.kernel] [g] lock_release_holdtime.part.11
> > 2.60% [guest.kernel] [g] memcpy
> > 2.33% [guest.kernel] [g] lock_acquired
> > 2.25% [guest.kernel] [g] lock_acquire
> > 1.84% [guest.kernel] [g] do_io_submit
> >
> > This patchs converts the ioctx list to a radix tree. For a performance
> > comparison the above FIO script was run on a 2 sockets 8 core
> > machine. This are the results (average and %rsd of 10 runs) for the
> > original list based implementation and for the radix tree based
> > implementation:
> >
> > cores 1 2 4 8 16 32
> > list 109376 ms 69119 ms 35682 ms 22671 ms 19724 ms 16408 ms
> > %rsd 0.69% 1.15% 1.17% 1.21% 1.71% 1.43%
> > radix 73651 ms 41748 ms 23028 ms 16766 ms 15232 ms 13787 ms
> > %rsd 1.19% 0.98% 0.69% 1.13% 0.72% 0.75%
> > % of radix
> > relative 66.12% 65.59% 66.63% 72.31% 77.26% 83.66%
> > to list
> >
> > To consider the impact of the patch on the typical case of having
> > only one ctx per process the following FIO script was run:
> >
> > rw=randrw; size=100m ;directory=/mnt/fio; ioengine=libaio; iodepth=1
> > blocksize=1024; numjobs=1; thread; loops=100
> >
> > on the same system and the results are the following:
> >
> > list 58892 ms
> > %rsd 0.91%
> > radix 59404 ms
> > %rsd 0.81%
> > % of radix
> > relative 100.87%
> > to list
>
> So, I was just doing some benchmarking/profiling to get ready to send
> out the aio patches I've got for 3.11 - and it looks like your patch is
> causing a ~1.5% throughput regression in my testing :/
... <snip>
I've got an alternate approach for fixing this wart in lookup_ioctx()...
Instead of using an rbtree, just use the reserved id in the ring buffer
header to index an array pointing the ioctx. It's not finished yet, and
it needs to be tidied up, but is most of the way there.
-ben
--
"Thought is the essence of where you are now."
--
kmo> And, a rework of Ben's code, but this was entirely his idea
kmo> -Kent
bcrl> And fix the code to use the right mm_struct in kill_ioctx(), actually
free memory.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
With the changes to use percpu counters for aio event ring size calculation,
existing increases to aio_max_nr are now insufficient to allow for the
allocation of enough events. Double the value used for aio_max_nr to account
for the doubling introduced by the percpu slack.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Originally, io_event() was documented to return the io_event if
cancellation succeeded - the io_event wouldn't be delivered via the ring
buffer like it normally would.
But this isn't what the implementation was actually doing; the only
driver implementing cancellation, the usb gadget code, never returned an
io_event in its cancel function. And aio_complete() was recently changed
to no longer suppress event delivery if the kiocb had been cancelled.
This gets rid of the unused io_event argument to kiocb_cancel() and
kiocb->ki_cancel(), and changes io_cancel() to return -EINPROGRESS if
kiocb->ki_cancel() returned success.
Also tweak the refcounting in kiocb_cancel() to make more sense.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
Cc: Zach Brown <zab@redhat.com>
Cc: Felipe Balbi <balbi@ti.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: Selvan Mani <smani@micron.com>
Cc: Sam Bradshaw <sbradshaw@micron.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
See the previous patch ("aio: reqs_active -> reqs_available") for why we
want to do this - this basically implements a per cpu allocator for
reqs_available that doesn't actually allocate anything.
Note that we need to increase the size of the ringbuffer we allocate,
since a single thread won't necessarily be able to use all the
reqs_available slots - some (up to about half) might be on other per cpu
lists, unavailable for the current thread.
We size the ringbuffer based on the nr_events userspace passed to
io_setup(), so this is a slight behaviour change - but nr_events wasn't
being used as a hard limit before, it was being rounded up to the next
page before so this doesn't change the actual semantics.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
Cc: Zach Brown <zab@redhat.com>
Cc: Felipe Balbi <balbi@ti.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: Selvan Mani <smani@micron.com>
Cc: Sam Bradshaw <sbradshaw@micron.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Reviewed-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
When "fs/aio: Add support to aio ring pages migration" was applied, it
broke the build when CONFIG_MIGRATION was disabled. Wrap the migration
code with a test for CONFIG_MIGRATION to fix this and save a few bytes
when migration is disabled.
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
As the aio job will pin the ring pages, that will lead to mem migrated
failed. In order to fix this problem we use an anon inode to manage the aio ring
pages, and setup the migratepage callback in the anon inode's address space, so
that when mem migrating the aio ring pages will be moved to other mem node safely.
Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
There was a regression introduced by 36f5588905 ("aio: refcounting
cleanup"), reported by Jens Axboe - the refcounting cleanup switched to
using RCU in the shutdown path, but the synchronize_rcu() was done in
the context of the io_destroy() syscall greatly increasing the time it
could block.
This patch switches it to call_rcu() and makes shutdown asynchronous
(more asynchronous than it was originally; before the refcount changes
io_destroy() would still wait on pending kiocbs).
Note that there's a global quota on the max outstanding kiocbs, and that
quota must be manipulated synchronously; otherwise io_setup() could
return -EAGAIN when there isn't quota available, and userspace won't
have any way of waiting until shutdown of the old kioctxs has finished
(besides busy looping).
So we release our quota before kioctx shutdown has finished, which
should be fine since the quota never corresponded to anything real
anyways.
Signed-off-by: Kent Overstreet <koverstreet@google.com>
Cc: Zach Brown <zab@redhat.com>
Cc: Felipe Balbi <balbi@ti.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Reported-by: Jens Axboe <axboe@kernel.dk>
Tested-by: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: Selvan Mani <smani@micron.com>
Cc: Sam Bradshaw <sbradshaw@micron.com>
Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Tested-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The recent changes overhauling fs/aio.c introduced a bug that results in
the kioctx not being freed when outstanding kiocbs are cancelled at
exit_aio() time. Specifically, a kiocb that is cancelled has its
completion events discarded by batch_complete_aio(), which then fails to
wake up the process stuck in free_ioctx(). Fix this by modifying the
wait_event() condition in free_ioctx() appropriately.
This patch was tested with the cancel operation in the thread based code
posted yesterday.
[akpm@linux-foundation.org: fix build]
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Kent Overstreet <koverstreet@google.com>
Cc: Kent Overstreet <koverstreet@google.com>
Cc: Josh Boyer <jwboyer@redhat.com>
Cc: Zach Brown <zab@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Benjamin LaHaise
Artem Savkov
Kent Overstreet
Gu Zheng
Tang Chen
Al Viro