The networking subsystem may wish to use defer_call(), so move the code
to util/ where it can be reused.
As a reminder of what defer_call() does:
This API defers a function call within a defer_call_begin()/defer_call_end()
section, allowing multiple calls to batch up. This is a performance
optimization that is used in the block layer to submit several I/O requests
at once instead of individually:
defer_call_begin(); <-- start of section
...
defer_call(my_func, my_obj); <-- deferred my_func(my_obj) call
defer_call(my_func, my_obj); <-- another
defer_call(my_func, my_obj); <-- another
...
defer_call_end(); <-- end of section, my_func(my_obj) is called once
Suggested-by: Ilya Maximets <i.maximets@ovn.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230913200045.1024233-3-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Prepare to move the blk_io_plug_call() API out of the block layer so
that other subsystems call use this deferred call mechanism. Rename it
to defer_call() but leave the code in block/plug.c.
The next commit will move the code out of the block layer.
Suggested-by: Ilya Maximets <i.maximets@ovn.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230913200045.1024233-2-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_io_plug_call() is invoked outside a blk_io_plug()/blk_io_unplug()
section while opening the NVMe drive from:
nvme_file_open() ->
nvme_init() ->
nvme_identify() ->
nvme_admin_cmd_sync() ->
nvme_submit_command() ->
blk_io_plug_call()
blk_io_plug_call() immediately invokes the given callback when the
current thread is not plugged, as is the case during nvme_file_open().
Unfortunately, nvme_submit_command() calls blk_io_plug_call() with
q->lock still held:
...
q->sq.tail = (q->sq.tail + 1) % NVME_QUEUE_SIZE;
q->need_kick++;
blk_io_plug_call(nvme_unplug_fn, q);
qemu_mutex_unlock(&q->lock);
^^^^^^^^^^^^^^^^^^^^^^^^^^^
nvme_unplug_fn() deadlocks trying to acquire q->lock because the lock is
already acquired by the same thread. The symptom is that QEMU hangs
during startup while opening the NVMe drive.
Fix this by moving the blk_io_plug_call() outside q->lock. This is safe
because no other thread runs code related to this queue and
blk_io_plug_call()'s internal state is immune to thread safety issues
since it is thread-local.
Reported-by: Lukáš Doktor <ldoktor@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Lukas Doktor <ldoktor@redhat.com>
Message-id: 20230712191628.252806-1-stefanha@redhat.com
Fixes: f2e590002b ("block/nvme: convert to blk_io_plug_call() API")
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>
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>
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>
BlockDriver->bdrv_io_unplug 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.
The only caller of this function is blk_io_unplug(), therefore make
blk_io_unplug() a co_wrapper, so that we're always running in 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: <20230113204212.359076-4-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
BlockDriver->bdrv_io_plug 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.
The only caller of this function is blk_io_plug(), therefore make
blk_io_plug() a co_wrapper, so that we're always running in 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: <20230113204212.359076-3-kwolf@redhat.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@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>
Registering an I/O buffer is only a performance optimization hint but it
is still necessary to return errors when it fails.
Later patches will need to detect errors when registering buffers but an
immediate advantage is that error_report() calls are no longer needed in
block driver .bdrv_register_buf() functions.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20221013185908.1297568-8-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The only implementor of bdrv_register_buf() is block/nvme.c, where the
size is not needed when unregistering a buffer. This is because
util/vfio-helpers.c can look up mappings by address.
Future block drivers that implement bdrv_register_buf() may not be able
to do their job given only the buffer address. Add a size argument to
bdrv_unregister_buf().
Also document the assumptions about
bdrv_register_buf()/bdrv_unregister_buf() calls. The same <host, size>
values that were given to bdrv_register_buf() must be given to
bdrv_unregister_buf().
gcc 11.2.1 emits a spurious warning that img_bench()'s buf_size local
variable might be uninitialized, so it's necessary to silence the
compiler.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20221013185908.1297568-5-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Callers of coroutine_fn must be coroutine_fn themselves, or the call
must be within "if (qemu_in_coroutine())". Apply coroutine_fn to
functions where this holds.
Reviewed-by: Alberto Faria <afaria@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220922084924.201610-13-pbonzini@redhat.com>
[kwolf: Fixed up coding style]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
nvme_get_free_req has very difference semantics when called in
coroutine context (where it waits) and in non-coroutine context
(where it doesn't). Split the two cases to make it clear what
is being requested.
Cc: qemu-block@nongnu.org
Reviewed-by: Alberto Faria <afaria@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20220922084924.201610-2-pbonzini@redhat.com>
[kwolf: Fixed up coding style]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Replace the global variables with inlined helper functions. getpagesize() is very
likely annotated with a "const" function attribute (at least with glibc), and thus
optimization should apply even better.
This avoids the need for a constructor initialization too.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20220323155743.1585078-12-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move the various memalign-related functions out of osdep.h and into
their own header, which we include only where they are used.
While we're doing this, add some brief documentation comments.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20220226180723.1706285-10-peter.maydell@linaro.org
Adaptive polling measures the execution time of the polling check plus
handlers called when a polled event becomes ready. Handlers can take a
significant amount of time, making it look like polling was running for
a long time when in fact the event handler was running for a long time.
For example, on Linux the io_submit(2) syscall invoked when a virtio-blk
device's virtqueue becomes ready can take 10s of microseconds. This
can exceed the default polling interval (32 microseconds) and cause
adaptive polling to stop polling.
By excluding the handler's execution time from the polling check we make
the adaptive polling calculation more accurate. As a result, the event
loop now stays in polling mode where previously it would have fallen
back to file descriptor monitoring.
The following data was collected with virtio-blk num-queues=2
event_idx=off using an IOThread. Before:
168k IOPS, IOThread syscalls:
9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16
9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8
9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8
9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3
9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8
9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8
9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8
9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32
174k IOPS (+3.6%), IOThread syscalls:
9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32
9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8
9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8
9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32
Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because
the IOThread stays in polling mode instead of falling back to file
descriptor monitoring.
As usual, polling is not implemented on Windows so this patch ignores
the new io_poll_read() callback in aio-win32.c.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Message-id: 20211207132336.36627-2-stefanha@redhat.com
[Fixed up aio_set_event_notifier() calls in
tests/unit/test-fdmon-epoll.c added after this series was queued.
--Stefan]
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
When the request free list is exhausted the coroutine waits on
q->free_req_queue for the next free request. Whenever a request is
completed a BH is scheduled to invoke nvme_free_req_queue_cb() and wake
up waiting coroutines.
1. nvme_get_free_req() waits for a free request:
while (q->free_req_head == -1) {
...
trace_nvme_free_req_queue_wait(q->s, q->index);
qemu_co_queue_wait(&q->free_req_queue, &q->lock);
...
}
2. nvme_free_req_queue_cb() wakes up the coroutine:
while (qemu_co_enter_next(&q->free_req_queue, &q->lock)) {
^--- infinite loop when free_req_head == -1
}
nvme_free_req_queue_cb() and the coroutine form an infinite loop when
q->free_req_head == -1. Fix this by checking q->free_req_head in
nvme_free_req_queue_cb(). If the free request list is exhausted, don't
wake waiting coroutines. Eventually an in-flight request will complete
and the BH will be scheduled again, guaranteeing forward progress.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20211208152246.244585-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Instead of duplicating code, extract the common helper to free
a single queue.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211006164931.172349-4-philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For debugging purpose it is helpful to know the CQ/SQ pointers.
We already have a trace event in nvme_free_queue_pair(), extend
it to report these pointer addresses.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211006164931.172349-3-philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Since commit 4d324c0bf6 ("introduce QEMU_AUTO_VFREE") buffers
allocated by qemu_memalign() can automatically freed when using
the QEMU_AUTO_VFREE macro. Use it to simplify a bit.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20211006164931.172349-2-philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver discard handlers bytes parameter to int64_t.
The only caller of all updated function is bdrv_co_pdiscard in
block/io.c. It is already prepared to work with 64bit requests, but
pass at most max(bs->bl.max_pdiscard, INT_MAX) to the driver.
Let's look at all updated functions:
blkdebug: all calculations are still OK, thanks to
bdrv_check_qiov_request().
both rule_check and bdrv_co_pdiscard are 64bit
blklogwrites: pass to blk_loc_writes_co_log which is 64bit
blkreplay, copy-on-read, filter-compress: pass to bdrv_co_pdiscard, OK
copy-before-write: pass to bdrv_co_pdiscard which is 64bit and to
cbw_do_copy_before_write which is 64bit
file-posix: one handler calls raw_account_discard() is 64bit and both
handlers calls raw_do_pdiscard(). Update raw_do_pdiscard, which pass
to RawPosixAIOData::aio_nbytes, which is 64bit (and calls
raw_account_discard())
gluster: somehow, third argument of glfs_discard_async is size_t.
Let's set max_pdiscard accordingly.
iscsi: iscsi_allocmap_set_invalid is 64bit,
!is_byte_request_lun_aligned is 64bit.
list.num is uint32_t. Let's clarify max_pdiscard and
pdiscard_alignment.
mirror_top: pass to bdrv_mirror_top_do_write() which is
64bit
nbd: protocol limitation. max_pdiscard is alredy set strict enough,
keep it as is for now.
nvme: buf.nlb is uint32_t and we do shift. So, add corresponding limits
to nvme_refresh_limits().
preallocate: pass to bdrv_co_pdiscard() which is 64bit.
rbd: pass to qemu_rbd_start_co() which is 64bit.
qcow2: calculations are still OK, thanks to bdrv_check_qiov_request(),
qcow2_cluster_discard() is 64bit.
raw-format: raw_adjust_offset() is 64bit, bdrv_co_pdiscard too.
throttle: pass to bdrv_co_pdiscard() which is 64bit and to
throttle_group_co_io_limits_intercept() which is 64bit as well.
test-block-iothread: bytes argument is unused
Great! Now all drivers are prepared to handle 64bit discard requests,
or else have explicit max_pdiscard limits.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-11-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver write_zeroes handlers bytes parameter to int64_t.
The only caller of all updated function is bdrv_co_do_pwrite_zeroes().
bdrv_co_do_pwrite_zeroes() itself is of course OK with widening of
callee parameter type. Also, bdrv_co_do_pwrite_zeroes()'s
max_write_zeroes is limited to INT_MAX. So, updated functions all are
safe, they will not get "bytes" larger than before.
Still, let's look through all updated functions, and add assertions to
the ones which are actually unprepared to values larger than INT_MAX.
For these drivers also set explicit max_pwrite_zeroes limit.
Let's go:
blkdebug: calculations can't overflow, thanks to
bdrv_check_qiov_request() in generic layer. rule_check() and
bdrv_co_pwrite_zeroes() both have 64bit argument.
blklogwrites: pass to blk_log_writes_co_log() with 64bit argument.
blkreplay, copy-on-read, filter-compress: pass to
bdrv_co_pwrite_zeroes() which is OK
copy-before-write: Calls cbw_do_copy_before_write() and
bdrv_co_pwrite_zeroes, both have 64bit argument.
file-posix: both handler calls raw_do_pwrite_zeroes, which is updated.
In raw_do_pwrite_zeroes() calculations are OK due to
bdrv_check_qiov_request(), bytes go to RawPosixAIOData::aio_nbytes
which is uint64_t.
Check also where that uint64_t gets handed:
handle_aiocb_write_zeroes_block() passes a uint64_t[2] to
ioctl(BLKZEROOUT), handle_aiocb_write_zeroes() calls do_fallocate()
which takes off_t (and we compile to always have 64-bit off_t), as
does handle_aiocb_write_zeroes_unmap. All look safe.
gluster: bytes go to GlusterAIOCB::size which is int64_t and to
glfs_zerofill_async works with off_t.
iscsi: Aha, here we deal with iscsi_writesame16_task() that has
uint32_t num_blocks argument and iscsi_writesame16_task() has
uint16_t argument. Make comments, add assertions and clarify
max_pwrite_zeroes calculation.
iscsi_allocmap_() functions already has int64_t argument
is_byte_request_lun_aligned is simple to update, do it.
mirror_top: pass to bdrv_mirror_top_do_write which has uint64_t
argument
nbd: Aha, here we have protocol limitation, and NBDRequest::len is
uint32_t. max_pwrite_zeroes is cleanly set to 32bit value, so we are
OK for now.
nvme: Again, protocol limitation. And no inherent limit for
write-zeroes at all. But from code that calculates cdw12 it's obvious
that we do have limit and alignment. Let's clarify it. Also,
obviously the code is not prepared to handle bytes=0. Let's handle
this case too.
trace events already 64bit
preallocate: pass to handle_write() and bdrv_co_pwrite_zeroes(), both
64bit.
rbd: pass to qemu_rbd_start_co() which is 64bit.
qcow2: offset + bytes and alignment still works good (thanks to
bdrv_check_qiov_request()), so tail calculation is OK
qcow2_subcluster_zeroize() has 64bit argument, should be OK
trace events updated
qed: qed_co_request wants int nb_sectors. Also in code we have size_t
used for request length which may be 32bit. So, let's just keep
INT_MAX as a limit (aligning it down to pwrite_zeroes_alignment) and
don't care.
raw-format: Is OK. raw_adjust_offset and bdrv_co_pwrite_zeroes are both
64bit.
throttle: Both throttle_group_co_io_limits_intercept() and
bdrv_co_pwrite_zeroes() are 64bit.
vmdk: pass to vmdk_pwritev which is 64bit
quorum: pass to quorum_co_pwritev() which is 64bit
Hooray!
At this point all block drivers are prepared to support 64bit
write-zero requests, or have explicitly set max_pwrite_zeroes.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-8-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: use <= rather than < in assertions relying on max_pwrite_zeroes]
Signed-off-by: Eric Blake <eblake@redhat.com>
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver write handlers parameters which are already 64bit to
signed type.
While being here, convert also flags parameter to be BdrvRequestFlags.
Now let's consider all callers. Simple
git grep '\->bdrv_\(aio\|co\)_pwritev\(_part\)\?'
shows that's there three callers of driver function:
bdrv_driver_pwritev() and bdrv_driver_pwritev_compressed() in
block/io.c, both pass int64_t, checked by bdrv_check_qiov_request() to
be non-negative.
qcow2_save_vmstate() does bdrv_check_qiov_request().
Still, the functions may be called directly, not only by drv->...
Let's check:
git grep '\.bdrv_\(aio\|co\)_pwritev\(_part\)\?\s*=' | \
awk '{print $4}' | sed 's/,//' | sed 's/&//' | sort | uniq | \
while read func; do git grep "$func(" | \
grep -v "$func(BlockDriverState"; done
shows several callers:
qcow2:
qcow2_co_truncate() write at most up to @offset, which is checked in
generic qcow2_co_truncate() by bdrv_check_request().
qcow2_co_pwritev_compressed_task() pass the request (or part of the
request) that already went through normal write path, so it should
be OK
qcow:
qcow_co_pwritev_compressed() pass int64_t, it's updated by this patch
quorum:
quorum_co_pwrite_zeroes() pass int64_t and int - OK
throttle:
throttle_co_pwritev_compressed() pass int64_t, it's updated by this
patch
vmdk:
vmdk_co_pwritev_compressed() pass int64_t, it's updated by this
patch
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver read handlers parameters which are already 64bit to
signed type.
While being here, convert also flags parameter to be BdrvRequestFlags.
Now let's consider all callers. Simple
git grep '\->bdrv_\(aio\|co\)_preadv\(_part\)\?'
shows that's there three callers of driver function:
bdrv_driver_preadv() in block/io.c, passes int64_t, checked by
bdrv_check_qiov_request() to be non-negative.
qcow2_load_vmstate() does bdrv_check_qiov_request().
do_perform_cow_read() has uint64_t argument. And a lot of things in
qcow2 driver are uint64_t, so converting it is big job. But we must
not work with requests that don't satisfy bdrv_check_qiov_request(),
so let's just assert it here.
Still, the functions may be called directly, not only by drv->...
Let's check:
git grep '\.bdrv_\(aio\|co\)_preadv\(_part\)\?\s*=' | \
awk '{print $4}' | sed 's/,//' | sed 's/&//' | sort | uniq | \
while read func; do git grep "$func(" | \
grep -v "$func(BlockDriverState"; done
The only one such caller:
QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, &data, 1);
...
ret = bdrv_replace_test_co_preadv(bs, 0, 1, &qiov, 0);
in tests/unit/test-bdrv-drain.c, and it's OK obviously.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-4-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: fix typos]
Signed-off-by: Eric Blake <eblake@redhat.com>
We expect the first qemu_vfio_dma_map() to fail (indicating
DMA mappings exhaustion, see commit 15a730e7a3). Do not
report the first failure as error, since we are going to
flush the mappings and retry.
This removes spurious error message displayed on the monitor:
(qemu) c
(qemu) qemu-kvm: VFIO_MAP_DMA failed: No space left on device
(qemu) info status
VM status: running
Reported-by: Tingting Mao <timao@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20210902070025.197072-12-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Currently qemu_vfio_dma_map() displays errors on stderr.
When using management interface, this information is simply
lost. Pass qemu_vfio_dma_map() an Error** handle so it can
propagate the error to callers.
Reviewed-by: Fam Zheng <fam@euphon.net>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20210902070025.197072-7-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
nvme_create_queue_pair() does not return a boolean value (indicating
eventual error) but a pointer, and is inconsistent in how it fills the
error handler. To fulfill callers expectations, always set an error
message on failure.
Reported-by: Auger Eric <eric.auger@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20210902070025.197072-6-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
When the NVMe block driver was introduced (see commit bdd6a90a9e,
January 2018), Linux VFIO_IOMMU_MAP_DMA ioctl was only returning
-ENOMEM in case of error. The driver was correctly handling the
error path to recycle its volatile IOVA mappings.
To fix CVE-2019-3882, Linux commit 492855939bdb ("vfio/type1: Limit
DMA mappings per container", April 2019) added the -ENOSPC error to
signal the user exhausted the DMA mappings available for a container.
The block driver started to mis-behave:
qemu-system-x86_64: VFIO_MAP_DMA failed: No space left on device
(qemu)
(qemu) info status
VM status: paused (io-error)
(qemu) c
VFIO_MAP_DMA failed: No space left on device
(qemu) c
VFIO_MAP_DMA failed: No space left on device
(The VM is not resumable from here, hence stuck.)
Fix by handling the new -ENOSPC error (when DMA mappings are
exhausted) without any distinction to the current -ENOMEM error,
so we don't change the behavior on old kernels where the CVE-2019-3882
fix is not present.
An easy way to reproduce this bug is to restrict the DMA mapping
limit (65535 by default) when loading the VFIO IOMMU module:
# modprobe vfio_iommu_type1 dma_entry_limit=666
Cc: qemu-stable@nongnu.org
Cc: Fam Zheng <fam@euphon.net>
Cc: Maxim Levitsky <mlevitsk@redhat.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Reported-by: Michal Prívozník <mprivozn@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20210723195843.1032825-1-philmd@redhat.com
Fixes: bdd6a90a9e ("block: Add VFIO based NVMe driver")
Buglink: https://bugs.launchpad.net/qemu/+bug/1863333
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/65
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
NVMe controllers implement different versions of the spec,
and different features of it. It is useful to gather this
information when debugging.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210127212137.3482291-3-philmd@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Commit 15b2260bef ("block/nvme: Trace controller capabilities")
misunderstood the doorbell stride value from the datasheet, use
the correct one. The 'doorbell_scale' variable used few lines
later is correct.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210127212137.3482291-2-philmd@redhat.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
NVMe drive cannot be shrunk.
Since commit c80d8b06cf we can use the @exact parameter (set
to false) to return success if the block device is larger than
the requested offset (even if we can not be shrunk).
Use this parameter to implement the NVMe truncate() coroutine,
similarly how it is done for the iscsi and file-posix drivers
(see commit 82325ae5f2 "Evaluate @exact in protocol drivers").
Reported-by: Xueqiang Wei <xuwei@redhat.com>
Suggested-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201210125202.858656-1-philmd@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
The Completion Queue Command Identifier is a 16-bit value,
so nvme_submit_command() is unlikely to work on big-endian
hosts, as the relevant bits are truncated.
Fix by using the correct byte-swap function.
Fixes: bdd6a90a9e ("block: Add VFIO based NVMe driver")
Reported-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-25-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
qemu_vfio_pci_map_bar() calls mmap(), and mmap(2) states:
'offset' must be a multiple of the page size as returned
by sysconf(_SC_PAGE_SIZE).
In commit f68453237b we started to use an offset of 4K which
broke this contract on Aarch64 arch.
Fix by mapping at offset 0, and and accessing doorbells at offset=4K.
Fixes: f68453237b ("block/nvme: Map doorbells pages write-only")
Reported-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-24-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Make sure iov's va and size are properly aligned on the
host page size.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-23-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
In preparation of 64kB host page support, let's change the size
and alignment of the prp_list_pages so that the VFIO DMA MAP succeeds
with 64kB host page size. We align on the host page size.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-22-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
In preparation of 64kB host page support, let's change the size
and alignment of the queue so that the VFIO DMA MAP succeeds.
We align on the host page size.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-21-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
In preparation of 64kB host page support, let's change the size
and alignment of the IDENTIFY command response buffer so that
the VFIO DMA MAP succeeds. We align on the host page size.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-20-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
While trying to simplify the code using a macro, we forgot
the 12-bit shift... Correct that.
Fixes: fad1eb6886 ("block/nvme: Use register definitions from 'block/nvme.h'")
Reported-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-19-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Commit bdd6a90a9e ("block: Add VFIO based NVMe driver")
sets the request_alignment in nvme_refresh_limits().
For consistency, also set it during initialization.
Reported-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-18-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
As all commands use the ADMIN queue, it is pointless to pass
it as argument each time. Remove the argument, and rename the
function as nvme_admin_cmd_sync() to make this new behavior
clearer.
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-17-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
We don't need to dereference from BDRVNVMeState each time.
Use a NVMeQueuePair pointer on the admin queue.
The nvme_init() becomes easier to review, matching the style
of nvme_add_io_queue().
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-16-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
From the specification chapter 3.1.8 "AQA - Admin Queue Attributes"
the Admin Submission Queue Size field is a 0’s based value:
Admin Submission Queue Size (ASQS):
Defines the size of the Admin Submission Queue in entries.
Enabling a controller while this field is cleared to 00h
produces undefined results. The minimum size of the Admin
Submission Queue is two entries. The maximum size of the
Admin Submission Queue is 4096 entries.
This is a 0’s based value.
This bug has never been hit because the device initialization
uses a single command synchronously :)
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-15-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Replace magic values by definitions, and simplifiy since the
number of queues will never reach 64K.
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-14-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Just for consistency, following the example documented since
commit e3fe3988d7 ("error: Document Error API usage rules"),
return a boolean value indicating an error is set or not.
Directly pass errp as the local_err is not requested in our
case. This simplifies a bit nvme_create_queue_pair().
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-12-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Just for consistency, following the example documented since
commit e3fe3988d7 ("error: Document Error API usage rules"),
return a boolean value indicating an error is set or not.
Directly pass errp as the local_err is not requested in our
case.
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20201029093306.1063879-11-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
We can not have negative queue count/size/index, use unsigned type.
Rename 'nr_queues' as 'queue_count' to match the spec naming.
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-10-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
To be able to use some definitions in structure declarations,
move them earlier. No logical change.
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-9-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-8-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
What we want to trace is the block driver state and the queue index.
Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-7-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
As we want to enable multiple queues, report the event
in each nvme_poll_queue() call, rather than once in
the callback calling nvme_poll_queues().
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20201029093306.1063879-6-philmd@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
