Mark functions as coroutine_fn when they are only called by other coroutine_fns
and they can suspend. Change calls to co_wrappers to use the non-wrapped
functions, which in turn requires adding GRAPH_RDLOCK annotations.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20230601115145.196465-3-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
raw_co_getlength is called by handle_aiocb_write_zeroes, which is not a coroutine
function. This is harmless because raw_co_getlength does not actually suspend,
but in the interest of clarity make it a non-coroutine_fn that is just wrapped
by the coroutine_fn raw_co_getlength. Likewise, check_cache_dropped was only
a coroutine_fn because it called raw_co_getlength, so it can be made non-coroutine
as well.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20230601115145.196465-2-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Now that bdrv_graph_wrlock() temporarily drops the AioContext lock that
its caller holds, it can poll without causing deadlocks. We can now
re-enable graph locking.
This reverts commit ad128dff0bf4b6f971d05eb4335a627883a19c1d.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-12-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If the caller keeps the AioContext lock for a block node in an iothread,
polling in bdrv_graph_wrlock() deadlocks if the condition isn't
fulfilled immediately.
Now that all callers make sure to actually have the AioContext locked
when they call bdrv_replace_child_noperm() like they should, we can
change bdrv_graph_wrlock() to take a BlockDriverState whose AioContext
lock the caller holds (NULL if it doesn't) and unlock it temporarily
while polling.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-11-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We use the user_ss[] array to hold the user emulation sources,
and the softmmu_ss[] array to hold the system emulation ones.
Hold the latter in the 'system_ss[]' array for parity with user
emulation.
Mechanical change doing:
$ sed -i -e s/softmmu_ss/system_ss/g $(git grep -l softmmu_ss)
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230613133347.82210-10-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Mechanical change running Coccinelle spatch with content
generated from the qom-cast-macro-clean-cocci-gen.py added
in the previous commit.
Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230601093452.38972-3-philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
When we for example have a sparse qcow2 image and discard: unmap is enabled,
there can be a lot of fragmentation in the image after some time. Especially on VM's
that do a lot of writes/deletes.
This causes the qcow2 image to grow even over 110% of its virtual size,
because the free gaps in the image get too small to allocate new
continuous clusters. So it allocates new space at the end of the image.
Disabling discard is not an option, as discard is needed to keep the
incremental backup size as low as possible. Without discard, the
incremental backups would become large, as qemu thinks it's just dirty
blocks but it doesn't know the blocks are unneeded.
So we need to avoid fragmentation but also 'empty' the unneeded blocks in
the image to have a small incremental backup.
In addition, we also want to send the discards further down the stack, so
the underlying blocks are still discarded.
Therefor we introduce a new qcow2 option "discard-no-unref".
When setting this option to true, discards will no longer have the qcow2
driver relinquish cluster allocations. Other than that, the request is
handled as normal: All clusters in range are marked as zero, and, if
pass-discard-request is true, it is passed further down the stack.
The only difference is that the now-zero clusters are preallocated
instead of being unallocated.
This will avoid fragmentation on the qcow2 image.
Fixes: https://gitlab.com/qemu-project/qemu/-/issues/1621
Signed-off-by: Jean-Louis Dupond <jean-louis@dupond.be>
Message-Id: <20230605084523.34134-2-jean-louis@dupond.be>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
All the offsets in the BAT must be lower than the file size.
Fix the check condition for correct check.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Message-Id: <20230424093147.197643-13-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Replace the way we use mutex in parallels_co_check() for simplier
and less error prone code.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Message-Id: <20230424093147.197643-12-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
We will add more and more checks so we need a better code structure
in parallels_co_check. Let each check performs in a separate loop
in a separate helper.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20230424093147.197643-11-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
We will add more and more checks so we need a better code structure
in parallels_co_check. Let each check performs in a separate loop
in a separate helper.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Message-Id: <20230424093147.197643-10-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Exclude out-of-image clusters from allocated and fragmented clusters
calculation.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Message-Id: <20230424093147.197643-9-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
We will add more and more checks so we need a better code structure in
parallels_co_check. Let each check performs in a separate loop in a
separate helper.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Message-Id: <20230424093147.197643-8-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
We will add more and more checks so we need a better code structure
in parallels_co_check. Let each check performs in a separate loop
in a separate helper.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20230424093147.197643-7-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
BAT is written in the context of conventional operations over the image
inside bdrv_co_flush() when it calls parallels_co_flush_to_os() callback.
Thus we should not modify BAT array directly, but call
parallels_set_bat_entry() helper and bdrv_co_flush() further on. After
that there is no need to manually write BAT and track its modification.
This makes code more generic and allows to split parallels_set_bat_entry()
for independent pieces.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Message-Id: <20230424093147.197643-6-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
This helper will be reused in next patches during parallels_co_check
rework to simplify its code.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20230424093147.197643-5-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Set data_end to the end of the last cluster inside the image. In such a
way we can be sure that corrupted offsets in the BAT can't affect on the
image size. If there are no allocated clusters set image_end_offset by
data_end.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Message-Id: <20230424093147.197643-4-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Don't let high_off be more than the file size even if we don't fix the
image.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20230424093147.197643-3-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
data_end field in BDRVParallelsState is set to the biggest offset present
in BAT. If this offset is outside of the image, any further write will
create the cluster at this offset and/or the image will be truncated to
this offset on close. This is definitely not correct.
Raise an error in parallels_open() if data_end points outside the image
and it is not a check (let the check to repaire the image). Set data_end
to the end of the cluster with the last correct offset.
Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Message-Id: <20230424093147.197643-2-alexander.ivanov@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
When processing vectored guest requests that are not aligned to the
storage request alignment, we pad them by adding head and/or tail
buffers for a read-modify-write cycle.
The guest can submit I/O vectors up to IOV_MAX (1024) in length, but
with this padding, the vector can exceed that limit. As of
4c002cef0e ("util/iov: make
qemu_iovec_init_extended() honest"), we refuse to pad vectors beyond the
limit, instead returning an error to the guest.
To the guest, this appears as a random I/O error. We should not return
an I/O error to the guest when it issued a perfectly valid request.
Before 4c002cef0e, we just made the vector
longer than IOV_MAX, which generally seems to work (because the guest
assumes a smaller alignment than we really have, file-posix's
raw_co_prw() will generally see bdrv_qiov_is_aligned() return false, and
so emulate the request, so that the IOV_MAX does not matter). However,
that does not seem exactly great.
I see two ways to fix this problem:
1. We split such long requests into two requests.
2. We join some elements of the vector into new buffers to make it
shorter.
I am wary of (1), because it seems like it may have unintended side
effects.
(2) on the other hand seems relatively simple to implement, with
hopefully few side effects, so this patch does that.
To do this, the use of qemu_iovec_init_extended() in bdrv_pad_request()
is effectively replaced by the new function bdrv_create_padded_qiov(),
which not only wraps the request IOV with padding head/tail, but also
ensures that the resulting vector will not have more than IOV_MAX
elements. Putting that functionality into qemu_iovec_init_extended() is
infeasible because it requires allocating a bounce buffer; doing so
would require many more parameters (buffer alignment, how to initialize
the buffer, and out parameters like the buffer, its length, and the
original elements), which is not reasonable.
Conversely, it is not difficult to move qemu_iovec_init_extended()'s
functionality into bdrv_create_padded_qiov() by using public
qemu_iovec_* functions, so that is what this patch does.
Because bdrv_pad_request() was the only "serious" user of
qemu_iovec_init_extended(), the next patch will remove the latter
function, so the functionality is not implemented twice.
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2141964
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-Id: <20230411173418.19549-3-hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
It's already confusing that we have two very similar functions for
wrapping the parse of a 64-bit unsigned value, differing mainly on
whether they permit leading '-'. Adjust the signature of parse_uint()
and parse_uint_full() to be like all of qemu_strto*(): put the result
parameter last, use the same types (uint64_t and unsigned long long
have the same width, but are not always the same type), and mark
endptr const (this latter change only affects the rare caller of
parse_uint). Adjust all callers in the tree.
While at it, note that since cutils.c already includes:
QEMU_BUILD_BUG_ON(sizeof(int64_t) != sizeof(long long));
we are guaranteed that the result of parse_uint* cannot exceed
UINT64_MAX (or the build would have failed), so we can drop
pre-existing dead comparisons in opts-visitor.c that were never false.
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-Id: <20230522190441.64278-8-eblake@redhat.com>
[eblake: Drop dead code spotted by Markus]
Signed-off-by: Eric Blake <eblake@redhat.com>
Some virtio-blk drivers (e.g. virtio-blk-vhost-vdpa) supports the fd
passing. Let's expose this to the user, so the management layer
can pass the file descriptor of an already opened path.
If the libblkio virtio-blk driver supports fd passing, let's always
use qemu_open() to open the `path`, so we can handle fd passing
from the management layer through the "/dev/fdset/N" special path.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20230530071941.8954-2-sgarzare@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
No block driver implements .bdrv_co_io_plug() anymore. Get rid of the
function pointers.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20230530180959.1108766-7-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stop using the .bdrv_co_io_plug() API because it is not multi-queue
block layer friendly. Use the new blk_io_plug_call() API to batch I/O
submission instead.
Note that a dev_max_batch check is dropped in laio_io_unplug() because
the semantics of unplug_fn() are different from .bdrv_co_unplug():
1. unplug_fn() is only called when the last blk_io_unplug() call occurs,
not every time blk_io_unplug() is called.
2. unplug_fn() is per-thread, not per-BlockDriverState, so there is no
way to get per-BlockDriverState fields like dev_max_batch.
Therefore this condition cannot be moved to laio_unplug_fn(). It is not
obvious that this condition affects performance in practice, so I am
removing it instead of trying to come up with a more complex mechanism
to preserve the condition.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20230530180959.1108766-6-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stop using the .bdrv_co_io_plug() API because it is not multi-queue
block layer friendly. Use the new blk_io_plug_call() API to batch I/O
submission instead.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20230530180959.1108766-5-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stop using the .bdrv_co_io_plug() API because it is not multi-queue
block layer friendly. Use the new blk_io_plug_call() API to batch I/O
submission instead.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20230530180959.1108766-4-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stop using the .bdrv_co_io_plug() API because it is not multi-queue
block layer friendly. Use the new blk_io_plug_call() API to batch I/O
submission instead.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20230530180959.1108766-3-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Introduce a new API for thread-local blk_io_plug() that does not
traverse the block graph. The goal is to make blk_io_plug() multi-queue
friendly.
Instead of having block drivers track whether or not we're in a plugged
section, provide an API that allows them to defer a function call until
we're unplugged: blk_io_plug_call(fn, opaque). If blk_io_plug_call() is
called multiple times with the same fn/opaque pair, then fn() is only
called once at the end of the function - resulting in batching.
This patch introduces the API and changes blk_io_plug()/blk_io_unplug().
blk_io_plug()/blk_io_unplug() no longer require a BlockBackend argument
because the plug state is now thread-local.
Later patches convert block drivers to blk_io_plug_call() and then we
can finally remove .bdrv_co_io_plug() once all block drivers have been
converted.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20230530180959.1108766-2-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
All callers now pass is_external=false to aio_set_fd_handler() and
aio_set_event_notifier(). The aio_disable_external() API that
temporarily disables fd handlers that were registered is_external=true
is therefore dead code.
Remove aio_disable_external(), aio_enable_external(), and the
is_external arguments to aio_set_fd_handler() and
aio_set_event_notifier().
The entire test-fdmon-epoll test is removed because its sole purpose was
testing aio_disable_external().
Parts of this patch were generated using the following coccinelle
(https://coccinelle.lip6.fr/) semantic patch:
@@
expression ctx, fd, is_external, io_read, io_write, io_poll, io_poll_ready, opaque;
@@
- aio_set_fd_handler(ctx, fd, is_external, io_read, io_write, io_poll, io_poll_ready, opaque)
+ aio_set_fd_handler(ctx, fd, io_read, io_write, io_poll, io_poll_ready, opaque)
@@
expression ctx, notifier, is_external, io_read, io_poll, io_poll_ready;
@@
- aio_set_event_notifier(ctx, notifier, is_external, io_read, io_poll, io_poll_ready)
+ aio_set_event_notifier(ctx, notifier, io_read, io_poll, io_poll_ready)
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-21-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This is part of ongoing work to remove the aio_disable_external() API.
Use BlockDevOps .drained_begin/end/poll() instead of
aio_set_fd_handler(is_external=true).
As a side-effect the FUSE export now follows AioContext changes like the
other export types.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-16-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The FUSE export calls blk_exp_ref/unref() without the AioContext lock.
Instead of fixing the FUSE export, adjust blk_exp_ref/unref() so they
work without the AioContext lock. This way it's less error-prone.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-15-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
vduse_blk_detach_ctx() waits for in-flight requests using
AIO_WAIT_WHILE(). This is not allowed according to a comment in
bdrv_set_aio_context_commit():
/*
* Take the old AioContex when detaching it from bs.
* At this point, new_context lock is already acquired, and we are now
* also taking old_context. This is safe as long as bdrv_detach_aio_context
* does not call AIO_POLL_WHILE().
*/
Use this opportunity to rewrite the drain code in vduse-blk:
- Use the BlockExport refcount so that vduse_blk_exp_delete() is only
called when there are no more requests in flight.
- Implement .drained_poll() so in-flight request coroutines are stopped
by the time .bdrv_detach_aio_context() is called.
- Remove AIO_WAIT_WHILE() from vduse_blk_detach_ctx() to solve the
.bdrv_detach_aio_context() constraint violation. It's no longer
needed due to the previous changes.
- Always handle the VDUSE file descriptor, even in drained sections. The
VDUSE file descriptor doesn't submit I/O, so it's safe to handle it in
drained sections. This ensures that the VDUSE kernel code gets a fast
response.
- Suspend virtqueue fd handlers in .drained_begin() and resume them in
.drained_end(). This eliminates the need for the
aio_set_fd_handler(is_external=true) flag, which is being removed from
QEMU.
This is a long list but splitting it into individual commits would
probably lead to git bisect failures - the changes are all related.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-14-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For simplicity, always run BlockDevOps .drained_begin/end/poll()
callbacks in the main loop thread. This makes it easier to implement the
callbacks and avoids extra locks.
Move the function pointer declarations from the I/O Code section to the
Global State section for BlockDevOps, BdrvChildClass, and BlockDriver.
Narrow IO_OR_GS_CODE() to GLOBAL_STATE_CODE() where appropriate.
The test-bdrv-drain test case calls bdrv_drain() from an IOThread. This
is now only allowed from coroutine context, so update the test case to
run in a coroutine.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-11-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The BlockBackend quiesce_counter is greater than zero during drained
sections. Add an API to check whether the BlockBackend is in a drained
section.
The next patch will use this API.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-10-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
vhost-user activity must be suspended during bdrv_drained_begin/end().
This prevents new requests from interfering with whatever is happening
in the drained section.
Previously this was done using aio_set_fd_handler()'s is_external
argument. In a multi-queue block layer world the aio_disable_external()
API cannot be used since multiple AioContext may be processing I/O, not
just one.
Switch to BlockDevOps->drained_begin/end() callbacks.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-8-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Each vhost-user-blk request runs in a coroutine. When the BlockBackend
enters a drained section we need to enter a quiescent state. Currently
any in-flight requests race with bdrv_drained_begin() because it is
unaware of vhost-user-blk requests.
When blk_co_preadv/pwritev()/etc returns it wakes the
bdrv_drained_begin() thread but vhost-user-blk request processing has
not yet finished. The request coroutine continues executing while the
main loop thread thinks it is in a drained section.
One example where this is unsafe is for blk_set_aio_context() where
bdrv_drained_begin() is called before .aio_context_detached() and
.aio_context_attach(). If request coroutines are still running after
bdrv_drained_begin(), then the AioContext could change underneath them
and they race with new requests processed in the new AioContext. This
could lead to virtqueue corruption, for example.
(This example is theoretical, I came across this while reading the
code and have not tried to reproduce it.)
It's easy to make bdrv_drained_begin() wait for in-flight requests: add
a .drained_poll() callback that checks the VuServer's in-flight counter.
VuServer just needs an API that returns true when there are requests in
flight. The in-flight counter needs to be atomic.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-7-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The VuServer object has a refcount field and ref/unref APIs. The name is
confusing because it's actually an in-flight request counter instead of
a refcount.
Normally a refcount destroys the object upon reaching zero. The VuServer
counter is used to wake up the vhost-user coroutine when there are no
more requests.
Avoid confusing by renaming refcount and ref/unref to in_flight and
inc/dec.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-6-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_set_aio_context() is not fully transactional because
blk_do_set_aio_context() updates blk->ctx outside the transaction. Most
of the time this goes unnoticed but a BlockDevOps.drained_end() callback
that invokes blk_get_aio_context() fails assert(ctx == blk->ctx). This
happens because blk->ctx is only assigned after
BlockDevOps.drained_end() is called and we're in an intermediate state
where BlockDrvierState nodes already have the new context and the
BlockBackend still has the old context.
Making blk_set_aio_context() fully transactional solves this assertion
failure because the BlockBackend's context is updated as part of the
transaction (before BlockDevOps.drained_end() is called).
Split blk_do_set_aio_context() in order to solve this assertion failure.
This helper function actually serves two different purposes:
1. It drives blk_set_aio_context().
2. It responds to BdrvChildClass->change_aio_ctx().
Get rid of the helper function. Do #1 inside blk_set_aio_context() and
do #2 inside blk_root_set_aio_ctx_commit(). This simplifies the code.
The only drawback of the fully transactional approach is that
blk_set_aio_context() must contend with blk_root_set_aio_ctx_commit()
being invoked as part of the AioContext change propagation. This can be
solved by temporarily setting blk->allow_aio_context_change to true.
Future patches call blk_get_aio_context() from
BlockDevOps->drained_end(), so this patch will become necessary.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-2-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The AioContext lock must not be held for bdrv_open_child(), but it is
necessary for the following operations, in particular those using nested
event loops in coroutine wrappers.
Temporarily dropping the main AioContext lock is not necessary because
we know we run in the main thread.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-9-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When opening the 'file' child moves bs to an iothread, we need to hold
the AioContext lock of it before we can call raw_apply_options() (and
more specifically, bdrv_getlength() inside of it).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-8-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
qcow2_open() doesn't work correctly when opening the 'file' child moves
bs to an iothread, for several reasons:
- It uses BDRV_POLL_WHILE() to wait for the qcow2_open_entry()
coroutine, which involves dropping the AioContext lock for bs when it
is not in the main context - but we don't hold it, so this crashes.
- It runs the qcow2_open_entry() coroutine in the current thread instead
of the new AioContext of bs.
- qcow2_open_entry() doesn't notify the main loop when it's done.
This patches fixes these issues around delegating work to a coroutine.
Temporarily dropping the main AioContext lock is not necessary because
we know we run in the main thread.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-7-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_open_backing_file() calls bdrv_open_inherit(), so all callers must
hold the main AioContext lock.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-6-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This fixes blk_new_open() to not assume that bs is in the main context.
In particular, the BlockBackend must be created with the right
AioContext because it will refuse to move to a different context
afterwards. (blk->allow_aio_context_change is false.)
Use this opportunity to use blk_insert_bs() instead of duplicating the
bdrv_root_attach_child() call. This is consistent with what
blk_new_with_bs() does. Add comments to document the locking rules.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-5-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The function documentation already says that all callers must hold the
main AioContext lock, but not all of them do. This can cause assertion
failures when functions called by bdrv_open() try to drop the lock. Fix
a few more callers to take the lock before calling bdrv_open().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-4-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
All of the functions that currently take a BlockDriverState, BdrvChild
or BlockBackend as their first parameter expect the associated
AioContext to be locked when they are called. In the case of
no_co_wrappers, they are called from bottom halves directly in the main
loop, so no other caller can be expected to take the lock for them. This
can result in assertion failures because a lock that isn't taken is
released in nested event loops.
Looking at the first parameter is already done by co_wrappers to decide
where the coroutine should run, so doing the same in no_co_wrappers is
only consistent. Take the lock in the generated bottom halves to fix the
problem.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-2-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In QEMU 8.0, we've been seeing deadlocks in bdrv_graph_wrlock(). They
come from callers that hold an AioContext lock, which is not allowed
during polling. In theory, we could temporarily release the lock, but
callers are inconsistent about whether they hold a lock, and if they do,
some are also confused about which one they hold. While all of this is
fixable, it's not trivial, and the best course of action for 8.0.1 is
probably just disabling the graph locking code temporarily.
We don't currently rely on graph locking yet. It is supposed to replace
the AioContext lock eventually to enable multiqueue support, but as long
as we still have the AioContext lock, it is sufficient without the graph
lock. Once the AioContext lock goes away, the deadlock doesn't exist any
more either and this commit can be reverted. (Of course, it can also be
reverted while the AioContext lock still exists if the callers have been
fixed.)
Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230517152834.277483-2-kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
There are some conditions under which we don't actually need to do
anything for taking a reader lock: Writing the graph is only possible
from the main context while holding the BQL. So if a reader is running
in the main context under the BQL and knows that it won't be interrupted
until the next writer runs, we don't actually need to do anything.
This is the case if the reader code neither has a nested event loop
(this is forbidden anyway while you hold the lock) nor is a coroutine
(because a writer could run when the coroutine has yielded).
These conditions are exactly what bdrv_graph_rdlock_main_loop() asserts.
They are not fulfilled in bdrv_graph_co_rdlock(), which always runs in a
coroutine.
This deletes the shortcuts in bdrv_graph_co_rdlock() that skip taking
the reader lock in the main thread.
Reported-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230510203601.418015-9-kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When jobs are sleeping, for example to enforce a given rate limit, they
can be reentered early, in particular in order to get paused, to update
the rate limit or to get cancelled.
Before this patch, they behave in this case as if they had fully
completed their rate limiting delay. This means that requests are sped
up beyond their limit, violating the constraints that the user gave us.
Change the block jobs to sleep in a loop until the necessary delay is
completed, while still allowing cancelling them immediately as well
pausing (handled by the pause point in job_sleep_ns()) and updating the
rate limit.
This change is also motivated by iotests cases being prone to fail
because drain operations pause and unpause them so often that block jobs
complete earlier than they are supposed to. In particular, the next
commit would fail iotests 030 without this change.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230510203601.418015-8-kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
qcow2_do_open() calls a few no_co_wrappers that wrap functions taking
the graph lock internally as a writer. Therefore, it can't hold the
reader lock across these calls, it causes deadlocks. Drop the lock
temporarily around the calls.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230510203601.418015-4-kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
