Instead of taking the writer lock internally, require callers to already
hold it when calling bdrv_attach_child_common(). These callers will
typically already hold the graph lock once the locking work is
completed, which means that they can't call functions that take it
internally.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-13-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In previous patches, we changed some transactionable functions to be
marked as GRAPH_WRLOCK, but required that tran_finalize() is still
called without the lock. This was because all callbacks that can be in
the same transaction need to follow the same convention.
Now that we don't have conflicting requirements any more, we can switch
all of the transaction callbacks to be declared GRAPH_WRLOCK, too, and
call tran_finalize() with the lock held.
Document for each of these transactionable functions that the lock needs
to be held when completing the transaction, and make sure that all
callers down to the place where the transaction is finalised actually
have the writer lock.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-12-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Instead of taking the writer lock internally, require callers to already
hold it when calling bdrv_attach_child_common(). These callers will
typically already hold the graph lock once the locking work is
completed, which means that they can't call functions that take it
internally.
Note that the transaction callbacks still take the lock internally, so
tran_finalize() must be called without the lock held. This is because
bdrv_append() also calls bdrv_replace_node_noperm(), which currently
requires the transaction callbacks to be called unlocked. In the next
step, both of them can be switched to locked tran_finalize() calls
together.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-11-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Instead of taking the writer lock internally, require callers to already
hold it when calling bdrv_replace_child_tran(). These callers will
typically already hold the graph lock once the locking work is
completed, which means that they can't call functions that take it
internally.
While a graph lock is held, polling is not allowed. Therefore draining
the necessary nodes can no longer be done in bdrv_remove_child() and
bdrv_replace_node_noperm(), but the callers must already make sure that
they are drained.
Note that the transaction callbacks still take the lock internally, so
tran_finalize() must be called without the lock held. This is because
bdrv_append() also calls bdrv_attach_child_noperm(), which currently
requires to be called unlocked. Once it changes, the transaction
callbacks can be changed, too.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-10-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Instead of taking the writer lock internally, require callers to already
hold it when calling bdrv_replace_child_noperm(). These callers will
typically already hold the graph lock once the locking work is
completed, which means that they can't call functions that take it
internally.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-9-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_unref() is called by a lot of places that need to hold the graph
lock (it naturally happens in the context of operations that change the
graph). However, bdrv_unref() takes the graph writer lock internally, so
it can't actually be called while already holding a graph lock without
causing a deadlock.
bdrv_unref() also can't just become GRAPH_WRLOCK because it drains the
node before closing it, and draining requires that the graph is
unlocked.
The solution is to defer deleting the node until we don't hold the lock
any more and draining is possible again.
Note that keeping images open for longer than necessary can create
problems, too: You can't open an image again before it is really closed
(if image locking didn't prevent it, it would cause corruption).
Reopening an image immediately happens at least during bdrv_open() and
bdrv_co_create().
In order to solve this problem, make sure to run the deferred unref in
bdrv_graph_wrunlock(), i.e. the first possible place where we can drain
again. This is also why bdrv_schedule_unref() is marked GRAPH_WRLOCK.
The output of iotest 051 is updated because the additional polling
changes the order of HMP output, resulting in a new "(qemu)" prompt in
the test output that was previously on a separate line and filtered out.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230911094620.45040-6-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This field has been unused since commit 72373e40fb ('block:
bdrv_reopen_multiple: refresh permissions on updated graph').
Remove it.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230911094620.45040-2-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For image creation code, we have central fallback code for protocols
that do not support creating new images (like NBD or iscsi). So for
them, you can only specify existing paths/exports that are overwritten
to make clean new images. In such a case, if the given path cannot be
opened (assuming a pre-existing image there), we print an error message
that tries to describe what is going on: That with this protocol, you
cannot create new images, but only overwrite existing ones; and the
given path could not be opened as a pre-existing image.
However, the current message is confusing, because it does not say that
the protocol in question does not support creating new images, but
instead that "image creation" is unsupported. This can be interpreted
to mean that `qemu-img create` will not work in principle, which is not
true. Be more verbose for clarity.
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2217204
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230720140024.46836-1-hreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
CoMutex has poor performance when lock contention is high. The tracked
requests list is accessed frequently and performance suffers in QEMU
multi-queue block layer scenarios.
It is not necessary to use CoMutex for the requests lock. The lock is
always released across coroutine yield operations. It is held for
relatively short periods of time and it is not beneficial to yield when
the lock is held by another coroutine.
Change the lock type from CoMutex to QemuMutex to improve multi-queue
block layer performance. fio randread bs=4k iodepth=64 with 4 IOThreads
handling a virtio-blk device with 8 virtqueues improves from 254k to
517k IOPS (+203%). Full benchmark results and configuration details are
available here:
980c40845d
In the future we may wish to introduce thread-local tracked requests
lists to avoid lock contention completely. That would be much more
involved though.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230808155852.2745350-3-stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This is going to be used in the subsequent commit as requests alignment
(in particular, during copy-on-read). This value only makes sense for
the formats which support subclusters (currently QCOW2 only). If this
field isn't set by driver's own bdrv_get_info() implementation, we
simply set it equal to the cluster size thus treating each cluster as
having a single subcluster.
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230711172553.234055-2-andrey.drobyshev@virtuozzo.com>
Mark functions as coroutine_fn when they are only called by other coroutine_fns
and they can suspend. Because this function operates on a BlockBackend, mark it
GRAPH_UNLOCKED.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20230601115145.196465-6-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@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>
bdrv_set_backing() requires the caller to hold the AioContext lock for
@backing_hd. Take it in bdrv_open_backing_file() before calling the
function.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-9-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_open_inherit() calls several functions for which it needs to hold
the AioContext lock, but currently doesn't. This includes calls in
bdrv_append_temp_snapshot(), for which bdrv_open_inherit() is the only
caller. Fix the locking in these places.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-8-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_set_file_or_backing_noperm() requires the caller to hold the
AioContext lock for the child node, but we hold the one for the parent
node in bdrv_reopen_parse_file_or_backing(). Take the other one
temporarily.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-7-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The function can move the child node to a different AioContext. In this
case, it also must take the AioContext lock for the new context before
calling functions that require the caller to hold the AioContext for the
child node.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-6-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_attach_child() requires that the caller holds the AioContext lock
for the new child node. Take it in bdrv_open_child() and document that
the caller must not hold any AioContext apart from the main AioContext.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230605085711.21261-5-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@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>
While calling bdrv_new_open_driver_opts(), the main AioContext lock must
be held, not the lock of the AioContext of the block subtree it will be
added to afterwards.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-11-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_refresh_total_sectors() and bdrv_refresh_limits() expect to be
called under the AioContext lock of the node. Take the lock.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-10-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>
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>
These functions specify that the caller must hold the "@filename
AioContext lock". This doesn't make sense, file names don't have an
AioContext. New BlockDriverStates always start in the main AioContext,
so this is what we really need here.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-3-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
These are functions that modify the graph, so they must be able to take
a writer lock. This is impossible if they already hold the reader lock.
If they need a reader lock for some of their operations, they should
take it internally.
Many of them go through blk_*(), which will always take the lock itself.
Direct calls of bdrv_*() need to take the reader lock. Note that while
locking for bdrv_co_*() calls is checked by TSA, this is not the case
for the mixed_coroutine_fns bdrv_*(). Holding the lock is still required
when they are called from coroutine context like here!
This effectively reverts 4ec8df0183, but adds some internal locking
instead.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230510203601.418015-2-kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Putting zoned/non-zoned BlockDrivers on top of each other is not
allowed.
Signed-off-by: Sam Li <faithilikerun@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20230508045533.175575-6-faithilikerun@gmail.com
Message-id: 20230324090605.28361-6-faithilikerun@gmail.com
[Adjust commit message prefix as suggested by Philippe Mathieu-Daudé
<philmd@linaro.org> and clarify that the check is about zoned
BlockDrivers.
--Stefan]
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_refresh_limits() need to hold a reader lock for the graph because
it accesses the children list of a node.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230504115750.54437-21-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_debug_event() need to hold a reader lock for the graph.
Unfortunately we cannot use a co_wrapper_bdrv_rdlock (i.e. make the
coroutine wrapper a no_coroutine_fn), because the function is called
(using the BLKDBG_EVENT macro) by mixed functions that run both in
coroutine and non-coroutine context (for example many of the functions
in qcow2-cluster.c and qcow2-refcount.c).
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230504115750.54437-16-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_get_info() need to hold a reader lock for the graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230504115750.54437-15-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_get_allocated_file_size() need to hold a reader lock for the
graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20230504115750.54437-14-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Drivers were a bit confused about whether .bdrv_open can run in a
coroutine and whether or not it holds a graph lock.
It cannot keep a graph lock from the caller across the whole function
because it both changes the graph (requires a writer lock) and does I/O
(requires a reader lock). Therefore, it should take these locks
internally as needed.
The functions used to be called in coroutine context during image
creation. This was buggy for other reasons, and as of commit 32192301,
all block drivers go through no_co_wrappers. So it is not called in
coroutine context any more.
Fix qcow2 and qed to work with the correct assumptions: The graph lock
needs to be taken internally instead of just assuming it's already
there, and the coroutine path is dead code that can be removed.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230504115750.54437-9-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
These functions must not be called in coroutine context, because they
need write access to the graph.
Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230504115750.54437-4-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Not a coroutine_fn, you say?
static int64_t bdrv_sum_allocated_file_size(BlockDriverState *bs)
{
BdrvChild *child;
int64_t child_size, sum = 0;
QLIST_FOREACH(child, &bs->children, next) {
if (child->role & (BDRV_CHILD_DATA | BDRV_CHILD_METADATA |
BDRV_CHILD_FILTERED))
{
child_size = bdrv_co_get_allocated_file_size(child->bs);
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Well what do we have here?!
I rest my case, your honor.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230308211435.346375-1-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The introduction of the graph lock is causing blk_get_geometry, a hot function
used in the I/O path, to create a coroutine. However, the only part that really
needs to run in coroutine context is the call to bdrv_co_refresh_total_sectors,
which in turn only happens in the rare case of host CD-ROM devices.
So, write by hand the three wrappers on the path from blk_co_get_geometry to
bdrv_co_refresh_total_sectors, so that the coroutine wrapper is only created
if bdrv_nb_sectors actually calls bdrv_refresh_total_sectors.
Reported-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230407153303.391121-9-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_co_get_geometry is only used in blk_co_get_geometry. Inline it in
there, to reduce the number of wrappers for bs->total_sectors.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230407153303.391121-7-pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
After reopening a BlockDriverState, it's possible that the size of the
underlying file has changed. This for example is covered by test 171.
Right now, this is handled by the raw driver's has_variable_length = true
setting. Since this will be removed by the next patch, handle it on
reopen instead, together with the existing bdrv_refresh_limits.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230407153303.391121-4-pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
At the protocol level, has_variable_length only needs to be true in the
very special case of host CD-ROM drives, so that they do not need an
explicit monitor command to read the new size when a disc is loaded
in the tray.
However, at the format level has_variable_length has to be true for all
raw blockdevs and for all filters, even though in practice the length
depends on the underlying file and thus will not change except in the
case of host CD-ROM drives.
As a first step towards computing an accurate value of has_variable_length,
add the value into the BlockLimits structure and initialize the field
from the BlockDriver.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230407153303.391121-2-pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_get_geometry() eventually calls bdrv_nb_sectors(), which is a
co_wrapper_mixed_bdrv_rdlock. This means that when it is called from
coroutine context, it already assume to have the graph locked.
However, virtio_blk_sect_range_ok() in block/export/virtio-blk-handler.c
(used by vhost-user-blk and VDUSE exports) runs in a coroutine, but
doesn't take the graph lock - blk_*() functions are generally expected
to do that internally. This causes an assertion failure when accessing
an export for the first time if it runs in an iothread.
This is an example of the crash:
$ ./storage-daemon/qemu-storage-daemon --object iothread,id=th0 --blockdev file,filename=/home/kwolf/images/hd.img,node-name=disk --export vhost-user-blk,addr.type=unix,addr.path=/tmp/vhost.sock,node-name=disk,id=exp0,iothread=th0
qemu-storage-daemon: ../block/graph-lock.c:268: void assert_bdrv_graph_readable(void): Assertion `qemu_in_main_thread() || reader_count()' failed.
(gdb) bt
#0 0x00007ffff6eafe5c in __pthread_kill_implementation () from /lib64/libc.so.6
#1 0x00007ffff6e5fa76 in raise () from /lib64/libc.so.6
#2 0x00007ffff6e497fc in abort () from /lib64/libc.so.6
#3 0x00007ffff6e4971b in __assert_fail_base.cold () from /lib64/libc.so.6
#4 0x00007ffff6e58656 in __assert_fail () from /lib64/libc.so.6
#5 0x00005555556337a3 in assert_bdrv_graph_readable () at ../block/graph-lock.c:268
#6 0x00005555555fd5a2 in bdrv_co_nb_sectors (bs=0x5555564c5ef0) at ../block.c:5847
#7 0x00005555555ee949 in bdrv_nb_sectors (bs=0x5555564c5ef0) at block/block-gen.c:256
#8 0x00005555555fd6b9 in bdrv_get_geometry (bs=0x5555564c5ef0, nb_sectors_ptr=0x7fffef7fedd0) at ../block.c:5884
#9 0x000055555562ad6d in blk_get_geometry (blk=0x5555564cb200, nb_sectors_ptr=0x7fffef7fedd0) at ../block/block-backend.c:1624
#10 0x00005555555ddb74 in virtio_blk_sect_range_ok (blk=0x5555564cb200, block_size=512, sector=0, size=512) at ../block/export/virtio-blk-handler.c:44
#11 0x00005555555dd80d in virtio_blk_process_req (handler=0x5555564cbb98, in_iov=0x7fffe8003830, out_iov=0x7fffe8003860, in_num=1, out_num=0) at ../block/export/virtio-blk-handler.c:189
#12 0x00005555555dd546 in vu_blk_virtio_process_req (opaque=0x7fffe8003800) at ../block/export/vhost-user-blk-server.c:66
#13 0x00005555557bf4a1 in coroutine_trampoline (i0=-402635264, i1=32767) at ../util/coroutine-ucontext.c:177
#14 0x00007ffff6e75c20 in ?? () from /lib64/libc.so.6
#15 0x00007fffefffa870 in ?? ()
#16 0x0000000000000000 in ?? ()
Fix this by creating a new blk_co_get_geometry() that takes the lock,
and changing blk_get_geometry() to be a co_wrapper_mixed around it.
To make the resulting code cleaner, virtio-blk-handler.c can directly
call the coroutine version now (though that wouldn't be necessary for
fixing the bug, taking the lock in blk_co_get_geometry() is what fixes
it).
Fixes: 8ab8140a04
Reported-by: Lukáš Doktor <ldoktor@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230327113959.60071-1-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_refresh_total_sectors() need to hold a reader lock for the
graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-24-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_delete_file() need to hold a reader lock for the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-22-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_eject() and bdrv_co_lock_medium() need to hold a reader lock for
the graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-20-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_is_inserted() need to hold a reader lock for the graph.
blk_is_inserted() is done as a co_wrapper_mixed_bdrv_rdlock (unlike most
other blk_* functions) because it is called a lot from other blk_co_*()
functions that already hold the lock. These calls go through
blk_is_available(), which becomes a co_wrapper_mixed_bdrv_rdlock, too,
for the same reason.
Functions that run in a coroutine and can call bdrv_co_is_available()
directly are changed to do so, which results in better TSA coverage.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-19-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
bdrv_co_create() need to hold a reader lock for the graph.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-17-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
It is never called outside of block.c.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230203152202.49054-2-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_append() is called with bs_top AioContext held, but
bdrv_attach_child_noperm() could change the AioContext of bs_top.
bdrv_replace_node_noperm() calls bdrv_drained_begin() starting from
commit 2398747128 ("block: Don't poll in bdrv_replace_child_noperm()").
bdrv_drained_begin() can call BDRV_POLL_WHILE that assumes the new lock
is taken, so let's temporarily hold the new AioContext to prevent QEMU
from failing in BDRV_POLL_WHILE when it tries to release the wrong
AioContext.
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2168209
Reported-by: Aihua Liang <aliang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20230214171621.11574-1-sgarzare@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230126172432.436111-14-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_co_create_opts_simple() runs in a coroutine. Therefore it is not
allowed to open images directly. Fix the call to use the corresponding
no_co_wrapper instead.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230126172432.436111-13-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_debug_event() is categorized as an I/O function, and it currently
doesn't run in a coroutine. We should let it take a graph rdlock since
it traverses the block nodes graph, which however is only possible in a
coroutine.
Therefore turn it into a co_wrapper_mixed to move the actual function
into a coroutine where the lock can be taken.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-14-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_lock_medium() is categorized as an I/O function, and it currently
doesn't run in a coroutine. We should let it take a graph rdlock since
it traverses the block nodes graph, which however is only possible in a
coroutine.
The only caller of this function is blk_lock_medium(). Therefore make
blk_lock_medium() a co_wrapper, so that it always creates a new
coroutine, and then make bdrv_lock_medium() a coroutine_fn where the
lock can be taken.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-13-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_eject() is categorized as an I/O function, and it currently
doesn't run in a coroutine. We should let it take a graph rdlock since
it traverses the block nodes graph, which however is only possible in a
coroutine.
The only caller of this function is blk_eject(). Therefore make
blk_eject() a co_wrapper, so that it always creates a new coroutine, and
then make bdrv_eject() coroutine_fn where the lock can be taken.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-12-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_get_info() is categorized as an I/O function, and it currently
doesn't run in a coroutine. We should let it take a graph rdlock since
it traverses the block nodes graph, which however is only possible in a
coroutine.
Therefore turn it into a co_wrapper to move the actual function into a
coroutine where the lock can be taken.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-11-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_get_allocated_file_size() is categorized as an I/O function, and it
currently doesn't run in a coroutine. We should let it take a graph
rdlock since it traverses the block nodes graph, which however is only
possible in a coroutine.
Therefore turn it into a co_wrapper to move the actual function into a
coroutine where the lock can be taken.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-10-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
BlockDriver->bdrv_getlength is categorized as IO callback, and it
currently doesn't run in a coroutine. We should let it take a graph
rdlock since the callback traverses the block nodes graph, which however
is only possible in a coroutine.
Therefore turn it into a co_wrapper to move the actual function into a
coroutine where the lock can be taken.
Because now this function creates a new coroutine and polls, we need to
take the AioContext lock where it is missing, for the only reason that
internally co_wrapper calls AIO_WAIT_WHILE and it expects to release the
AioContext lock.
This is especially messy when a co_wrapper creates a coroutine and polls
in bdrv_open_driver, because this function has so many callers in so
many context that it can easily lead to deadlocks. Therefore the new
rule for bdrv_open_driver is that the caller must always hold the
AioContext lock of the given bs (except if it is a coroutine), because
the function calls bdrv_refresh_total_sectors() which is now a
co_wrapper.
Once the rwlock is ultimated and placed in every place it needs to be,
we will poll using AIO_WAIT_WHILE_UNLOCKED and remove the AioContext
lock.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-7-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The name is not good, not the least because we are going to convert this
to a generated co_wrapper, which adds a _co infix after the first part
of the name.
No functional change intended.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-6-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_is_inserted() is categorized as an I/O function, and it currently
doesn't run in a coroutine. We should let it take a graph rdlock since
it traverses the block nodes graph, which however is only possible in a
coroutine.
Therefore turn it into a co_wrapper to move the actual function into a
coroutine where the lock can be taken.
At the same time, add also blk_is_inserted as co_wrapper_mixed, since it
is called in both coroutine and non-coroutine contexts.
Because now this function creates a new coroutine and polls, we need to
take the AioContext lock where it is missing, for the only reason that
internally c_w_mixed_bdrv_rdlock calls AIO_WAIT_WHILE and it expects to
release the AioContext lock. Once the rwlock is ultimated and placed in
every place it needs to be, we will poll using AIO_WAIT_WHILE_UNLOCKED
and remove the AioContext lock.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230113204212.359076-5-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
It has only one caller---inline it and remove the function.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20221215130225.476477-2-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We have two inclusion loops:
block/block.h
-> block/block-global-state.h
-> block/block-common.h
-> block/blockjob.h
-> block/block.h
block/block.h
-> block/block-io.h
-> block/block-common.h
-> block/blockjob.h
-> block/block.h
I believe these go back to Emanuele's reorganization of the block API,
merged a few months ago in commit d7e2fe4aac.
Fortunately, breaking them is merely a matter of deleting unnecessary
includes from headers, and adding them back in places where they are
now missing.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20221221133551.3967339-2-armbru@redhat.com>
The generated coroutine wrappers already take care to take the lock in
the non-coroutine path, and assume that the lock is already taken in the
coroutine path.
The only thing we need to do for the wrapped function is adding the
GRAPH_RDLOCK annotation. Doing so also allows us to mark the
corresponding callbacks in BlockDriver as GRAPH_RDLOCK_PTR.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-19-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-16-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Remove the old assert_bdrv_graph_writable, and replace it with
the new version using graph-lock API.
See the function documentation for more information.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-14-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We don't protect bdrv->aio_context with the graph rwlock,
so these assertions are not needed
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-13-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Protect the main function where graph is modified.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-12-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_img_create() polls internally (when calling bdrv_create(), which is
a co_wrapper), so it can't be called while holding the lock of any
AioContext except the current one without causing deadlocks. Drop the
lock around the call in external_snapshot_prepare().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221207131838.239125-11-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This function is never called in coroutine context, therefore
instead of manually creating a new coroutine, delegate it to the
block-coroutine-wrapper script, defining it as co_wrapper.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20221128142337.657646-14-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
It is always called in coroutine_fn callbacks, therefore
it can directly call bdrv_co_create().
Rename it to bdrv_co_create_file too.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20221128142337.657646-9-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Call two different functions depending on whether bdrv_create
is in coroutine or not, following the same pattern as
generated_co_wrapper functions.
This allows to also call the coroutine function directly,
without using CreateCo or relying in bdrv_create().
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20221128142337.657646-8-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We know that the string will stay around until the function
returns, and the parameter of drv->bdrv_co_create_opts is const char*,
so it must not be modified either.
Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20221128142337.657646-7-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
All callers of bdrv_parent_drained_begin_single() pass poll=false now,
so we don't need the parameter any more.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221118174110.55183-16-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In order to make sure that bdrv_replace_child_noperm() doesn't have to
poll any more, get rid of the bdrv_parent_drained_begin_single() call.
This is possible now because we can require that the parent is already
drained through the child in question when the function is called and we
don't call the parent drain callbacks more than once.
The additional drain calls needed in callers cause the test case to run
its code in the drain handler too early (bdrv_attach_child() drains
now), so modify it to only enable the code after the test setup has
completed.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221118174110.55183-15-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
ignore_bds_parents is now ignored during drain_begin and drain_end, so
we can just remove it there. It is still a valid optimisation for
drain_all in bdrv_drained_poll(), so leave it around there.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221118174110.55183-13-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We only need to call both the BlockDriver's callback and the parent
callbacks when going from undrained to drained or vice versa. A second
drain section doesn't make a difference for the driver or the parent,
they weren't supposed to send new requests before and after the second
drain.
One thing that gets in the way is the 'ignore_bds_parents' parameter in
bdrv_do_drained_begin_quiesce() and bdrv_do_drained_end(): It means that
bdrv_drain_all_begin() increases bs->quiesce_counter, but does not
quiesce the parent through BdrvChildClass callbacks. If an additional
drain section is started now, bs->quiesce_counter will be non-zero, but
we would still need to quiesce the parent through BdrvChildClass in
order to keep things consistent (and unquiesce it on the matching
bdrv_drained_end(), even though the counter would not reach 0 yet as
long as the bdrv_drain_all() section is still active).
Instead of keeping track of this, let's just get rid of the parameter.
It was introduced in commit 6cd5c9d7b2 as an optimisation so that
during bdrv_drain_all(), we wouldn't recursively drain all parents up to
the root for each node, resulting in quadratic complexity. As it happens,
calling the callbacks only once solves the same problem, so as of this
patch, we'll still have O(n) complexity and ignore_bds_parents is not
needed any more.
This patch only ignores the 'ignore_bds_parents' parameter. It will be
removed in a separate patch.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-12-kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Subtree drains are not used any more. Remove them.
After this, BdrvChildClass.attach/detach() don't poll any more.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-11-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The subtree drain was introduced in commit b1e1af394d as a way to avoid
graph changes between finding the base node and changing the block graph
as necessary on completion of the image streaming job.
The block graph could change between these two points because
bdrv_set_backing_hd() first drains the parent node, which involved
polling and can do anything.
Subtree draining was an imperfect way to make this less likely (because
with it, fewer callbacks are called during this window). Everyone agreed
that it's not really the right solution, and it was only committed as a
stopgap solution.
This replaces the subtree drain with a solution that simply drains the
parent node before we try to find the base node, and then call a version
of bdrv_set_backing_hd() that doesn't drain, but just asserts that the
parent node is already drained.
This way, any graph changes caused by draining happen before we start
looking at the graph and things stay consistent between finding the base
node and changing the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-10-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Instead of using a subtree drain from the top node (which also drains
child nodes of base that we're not even interested in), use a normal
drain for base, which automatically drains all of the parents, too.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-9-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_reopen() and friends use subtree drains as a lazy way of covering
all the nodes they touch. Turns out that this lazy way is a lot more
complicated than just draining the nodes individually, even not
accounting for the additional complexity in the drain mechanism itself.
Simplify the code by switching to draining the individual nodes that are
already managed in the BlockReopenQueue anyway.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221118174110.55183-8-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Callers don't agree whether bdrv_reopen_queue_child() should be called
with the AioContext lock held or not. Standardise on holding the lock
(as done by QMP blockdev-reopen and the replication block driver) and
fix bdrv_reopen() to do the same.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221118174110.55183-7-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
drained_end_counter is unused now, nobody changes its value any more. It
can be removed.
In cases where we had two almost identical functions that only differed
in whether the caller passes drained_end_counter, or whether they would
poll for a local drained_end_counter to reach 0, these become a single
function.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20221118174110.55183-5-kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Polling during bdrv_drained_end() can be problematic (and in the future,
we may get cases for bdrv_drained_begin() where polling is forbidden,
and we don't care about already in-flight requests, but just want to
prevent new requests from arriving).
The .bdrv_drained_begin/end callbacks running in a coroutine is the only
reason why we have to do this polling, so make them non-coroutine
callbacks again. None of the callers actually yield any more.
This means that bdrv_drained_end() effectively doesn't poll any more,
even if AIO_WAIT_WHILE() loops are still there (their condition is false
from the beginning). This is generally not a problem, but in
test-bdrv-drain, some additional explicit aio_poll() calls need to be
added because the test case wants to verify the final state after BHs
have executed.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-4-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We are going to increase usage of collecting nodes in a list to then
update, and calling bdrv_topological_dfs() each time is not convenient,
and not correct as we are going to interleave graph modifying with
filling the node list.
So, let's switch to a function that takes any list of nodes, adds all
their subtrees and do topological sort. And finally, refresh
permissions.
While being here, make the function public, as we'll want to use it
from blockdev.c in near future.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221107163558.618889-5-vsementsov@yandex-team.ru>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Allow passing external Transaction pointer, stop creating extra
Transaction objects.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221107163558.618889-4-vsementsov@yandex-team.ru>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Drop this simple wrapper used only in one place. We have too many graph
modifying functions even without it.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221107163558.618889-3-vsementsov@yandex-team.ru>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The only caller is bdrv_root_unref_child(), let's just do the logic
directly in it. It simplifies further conversion of
bdrv_root_unref_child() to transaction actions.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221107163558.618889-2-vsementsov@yandex-team.ru>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We want to use bdrv_child_get_parent_aio_context() from
bdrv_parent_drained_{begin,end}_single(), both of which are "I/O or GS"
functions.
Prior to 3ed4f708fe, all the implementations were I/O code anyway.
3ed4f708fe has put block jobs' AioContext field under the job mutex, so
to make child_job_get_parent_aio_context() work in an I/O context, we
need to take that lock there.
Furthermore, blk_root_get_parent_aio_context() is not marked as
anything, but is safe to run in an I/O context, so mark it that way now.
(blk_get_aio_context() is an I/O code function.)
With that done, all implementations explicitly are I/O code, so we can
mark bdrv_child_get_parent_aio_context() as I/O code, too, so callers
know it is safe to run from both GS and I/O contexts.
Signed-off-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221107151321.211175-2-hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
improve error handling during module load, by changing:
bool module_load(const char *prefix, const char *lib_name);
void module_load_qom(const char *type);
to:
int module_load(const char *prefix, const char *name, Error **errp);
int module_load_qom(const char *type, Error **errp);
where the return value is:
-1 on module load error, and errp is set with the error
0 on module or one of its dependencies are not installed
1 on module load success
2 on module load success (module already loaded or built-in)
module_load_qom_one has been introduced in:
commit 28457744c3 ("module: qom module support"), which built on top of
module_load_one, but discarded the bool return value. Restore it.
Adapt all callers to emit errors, or ignore them, or fail hard,
as appropriate in each context.
Replace the previous emission of errors via fprintf in _some_ error
conditions with Error and error_report, so as to emit to the appropriate
target.
A memory leak is also fixed as part of the module_load changes.
audio: when attempting to load an audio module, report module load errors.
Note that still for some callers, a single issue may generate multiple
error reports, and this could be improved further.
Regarding the audio code itself, audio_add() seems to ignore errors,
and this should probably be improved.
block: when attempting to load a block module, report module load errors.
For the code paths that already use the Error API, take advantage of those
to report module load errors into the Error parameter.
For the other code paths, we currently emit the error, but this could be
improved further by adding Error parameters to all possible code paths.
console: when attempting to load a display module, report module load errors.
qdev: when creating a new qdev Device object (DeviceState), report load errors.
If a module cannot be loaded to create that device, now abort execution
(if no CONFIG_MODULE) or exit (if CONFIG_MODULE).
qom/object.c: when initializing a QOM object, or looking up class_by_name,
report module load errors.
qtest: when processing the "module_load" qtest command, report errors
in the load of the module.
Signed-off-by: Claudio Fontana <cfontana@suse.de>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220929093035.4231-4-cfontana@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alberto Faria <afaria@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20221013123711.620631-16-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
No functional change intended.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-11-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
No functional changes intended.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-10-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Together with all _can_set_ and _set_ APIs, as they are not needed
anymore.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-9-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Replace all direct usage of ->can_set_aio_ctx and ->set_aio_ctx,
and call bdrv_child_try_change_aio_context() in
bdrv_try_set_aio_context(), the main function called through
the whole block layer.
From this point onwards, ->can_set_aio_ctx and ->set_aio_ctx
won't be used anymore.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-8-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bdrv_child_cb_change_aio_ctx() is identical to
bdrv_child_cb_can_set_aio_ctx(), as we only need
to recursively go on the parent bs.
Note: bdrv_child_try_change_aio_context() is not called by
anyone at this point.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-6-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Minor performance improvement, but given that we have hash tables
available, avoid iterating in the visited nodes list every time just
to check if a node has been already visited.
The data structure is not actually a proper hash map, but an hash set,
as we are just adding nodes and not key,value pairs.
Suggested-by: Hanna Reitz <hreitz@redhat.com>
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20221025084952.2139888-4-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Simplify the way the aiocontext can be changed in a BDS graph.
There are currently two problems in bdrv_try_set_aio_context:
- There is a confusion of AioContext locks taken and released, because
we assume that old aiocontext is always taken and new one is
taken inside.
- It doesn't look very safe to call bdrv_drained_begin while some
nodes have already switched to the new aiocontext and others haven't.
This could be especially dangerous because bdrv_drained_begin polls, so
something else could be executed while graph is in an inconsistent
state.
Additional minor nitpick: can_set and set_ callbacks both traverse the
graph, both using the ignored list of visited nodes in a different way.
Therefore, get rid of all of this and introduce a new callback,
change_aio_context, that uses transactions to efficiently, cleanly
and most importantly safely change the aiocontext of a graph.
This new callback is a "merge" of the two previous ones:
- Just like can_set_aio_context, recursively traverses the graph.
Marks all nodes that are visited using a GList, and checks if
they *could* change the aio_context.
- For each node that passes the above check, drain it and add a new transaction
that implements a callback that effectively changes the aiocontext.
- Once done, the recursive function returns if *all* nodes can change
the AioContext. If so, commit the above transactions.
Regardless of the outcome, call transaction.clean() to undo all drains
done in the recursion.
- The transaction list is scanned only after all nodes are being drained, so
we are sure that they all are in the same context, and then
we switch their AioContext, concluding the drain only after all nodes
switched to the new AioContext. In this way we make sure that
bdrv_drained_begin() is always called under the old AioContext, and
bdrv_drained_end() under the new one.
- Because of the above, we don't need to release and re-acquire the
old AioContext every time, as everything is done once (and not
per-node drain and aiocontext change).
Note that the "change" API is not yet invoked anywhere.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20221025084952.2139888-3-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Also here ->aio_context is read by I/O threads and written
under BQL.
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221025084952.2139888-2-eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Now the function can remove any child, so give it more common name.
Drop assertions and drop bs argument which becomes unused. Function
would be reused in a further commit.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220726201134.924743-16-vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
bs->file and bs->backing are a kind of duplication of part of
bs->children. But very useful diplication, so let's not drop them at
all:)
We should manage bs->file and bs->backing in same place, where we
manage bs->children, to keep them in sync.
Moreover, generic io paths are unprepared to BdrvChild without a bs, so
it's double good to clear bs->file / bs->backing when we detach the
child.
Detach is simple: if we detach bs->file or bs->backing child, just
set corresponding field to NULL.
Attach is a bit more complicated. But we still can precisely detect
should we set one of bs->file / bs->backing or not:
- if role is BDRV_CHILD_COW, we definitely deal with bs->backing
- else, if role is BDRV_CHILD_FILTERED (it must be also
BDRV_CHILD_PRIMARY), it's a filtered child. Use
bs->drv->filtered_child_is_backing to chose the pointer field to
modify.
- else, if role is BDRV_CHILD_PRIMARY, we deal with bs->file
- in all other cases, it's neither bs->backing nor bs->file. It's some
other child and we shouldn't care
OK. This change brings one more good thing: we can (and should) get rid
of all indirect pointers in the block-graph-change transactions:
bdrv_attach_child_common() stores BdrvChild** into transaction to clear
it on abort.
bdrv_attach_child_common() has two callers: bdrv_attach_child_noperm()
just pass-through this feature, bdrv_root_attach_child() doesn't need
the feature.
Look at bdrv_attach_child_noperm() callers:
- bdrv_attach_child() doesn't need the feature
- bdrv_set_file_or_backing_noperm() uses the feature to manage
bs->file and bs->backing, we don't want it anymore
- bdrv_append() uses the feature to manage bs->backing, again we
don't want it anymore
So, we should drop this stuff! Great!
We could probably keep BdrvChild** argument to keep the int return
value, but it seems not worth the complexity.
Finally, we now set .file / .backing automatically in generic code and
want to restring setting them by hand outside of .attach/.detach.
So, this patch cleanups all remaining places where they were set.
To find such places I use:
git grep '\->file ='
git grep '\->backing ='
git grep '&.*\<backing\>'
git grep '&.*\<file\>'
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20220726201134.924743-14-vsementsov@yandex-team.ru>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
