If we don't have active request, that waiting for this handle to be
received, we should report an error.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210902103805.25686-6-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
OK, that's a big rewrite of the logic.
Pre-patch we have an always running coroutine - connection_co. It does
reply receiving and reconnecting. And it leads to a lot of difficult
and unobvious code around drained sections and context switch. We also
abuse bs->in_flight counter which is increased for connection_co and
temporary decreased in points where we want to allow drained section to
begin. One of these place is in another file: in nbd_read_eof() in
nbd/client.c.
We also cancel reconnect and requests waiting for reconnect on drained
begin which is not correct. And this patch fixes that.
Let's finally drop this always running coroutine and go another way:
do both reconnect and receiving in request coroutines.
The detailed list of changes below (in the sequence of diff hunks).
1. receiving coroutines are woken directly from nbd_channel_error, when
we change s->state
2. nbd_co_establish_connection_cancel(): we don't have drain_begin now,
and in nbd_teardown_connection() all requests should already be
finished (and reconnect is done from request). So
nbd_co_establish_connection_cancel() is called from
nbd_cancel_in_flight() (to cancel the request that is doing
nbd_co_establish_connection()) and from reconnect_delay_timer_cb()
(previously we didn't need it, as reconnect delay only should cancel
active requests not the reconnection itself). But now reconnection
itself is done in the separate thread (we now call
nbd_client_connection_enable_retry() in nbd_open()), and we need to
cancel the requests that wait in nbd_co_establish_connection()
now).
2A. We do receive headers in request coroutine. But we also should
dispatch replies for other pending requests. So,
nbd_connection_entry() is turned into nbd_receive_replies(), which
does reply dispatching while it receives other request headers, and
returns when it receives the requested header.
3. All old staff around drained sections and context switch is dropped.
In details:
- we don't need to move connection_co to new aio context, as we
don't have connection_co anymore
- we don't have a fake "request" of connection_co (extra increasing
in_flight), so don't care with it in drain_begin/end
- we don't stop reconnection during drained section anymore. This
means that drain_begin may wait for a long time (up to
reconnect_delay). But that's an improvement and more correct
behavior see below[*]
4. In nbd_teardown_connection() we don't have to wait for
connection_co, as it is dropped. And cleanup for s->ioc and nbd_yank
is moved here from removed connection_co.
5. In nbd_co_do_establish_connection() we now should handle
NBD_CLIENT_CONNECTING_NOWAIT: if new request comes when we are in
NBD_CLIENT_CONNECTING_NOWAIT, it still should call
nbd_co_establish_connection() (who knows, maybe the connection was
already established by another thread in the background). But we
shouldn't wait: if nbd_co_establish_connection() can't return new
channel immediately the request should fail (we are in
NBD_CLIENT_CONNECTING_NOWAIT state).
6. nbd_reconnect_attempt() is simplified: it's now easier to wait for
other requests in the caller, so here we just assert that fact.
Also delay time is now initialized here: we can easily detect first
attempt and start a timer.
7. nbd_co_reconnect_loop() is dropped, we don't need it. Reconnect
retries are fully handle by thread (nbd/client-connection.c), delay
timer we initialize in nbd_reconnect_attempt(), we don't have to
bother with s->drained and friends. nbd_reconnect_attempt() now
called from nbd_co_send_request().
8. nbd_connection_entry is dropped: reconnect is now handled by
nbd_co_send_request(), receiving reply is now handled by
nbd_receive_replies(): all handled from request coroutines.
9. So, welcome new nbd_receive_replies() called from request coroutine,
that receives reply header instead of nbd_connection_entry().
Like with sending requests, only one coroutine may receive in a
moment. So we introduce receive_mutex, which is locked around
nbd_receive_reply(). It also protects some related fields. Still,
full audit of thread-safety in nbd driver is a separate task.
New function waits for a reply with specified handle being received
and works rather simple:
Under mutex:
- if current handle is 0, do receive by hand. If another handle
received - switch to other request coroutine, release mutex and
yield. Otherwise return success
- if current handle == requested handle, we are done
- otherwise, release mutex and yield
10: in nbd_co_send_request() we now do nbd_reconnect_attempt() if
needed. Also waiting in free_sema queue we now wait for one of two
conditions:
- connectED, in_flight < MAX_NBD_REQUESTS (so we can start new one)
- connectING, in_flight == 0, so we can call
nbd_reconnect_attempt()
And this logic is protected by s->send_mutex
Also, on failure we don't have to care of removed s->connection_co
11. nbd_co_do_receive_one_chunk(): now instead of yield() and wait for
s->connection_co we just call new nbd_receive_replies().
12. nbd_co_receive_one_chunk(): place where s->reply.handle becomes 0,
which means that handling of the whole reply is finished. Here we
need to wake one of coroutines sleeping in nbd_receive_replies().
If none are sleeping - do nothing. That's another behavior change: we
don't have endless recv() in the idle time. It may be considered as
a drawback. If so, it may be fixed later.
13. nbd_reply_chunk_iter_receive(): don't care about removed
connection_co, just ping in_flight waiters.
14. Don't create connection_co, enable retry in the connection thread
(we don't have own reconnect loop anymore)
15. We now need to add a nbd_co_establish_connection_cancel() call in
nbd_cancel_in_flight(), to cancel the request that is doing a
connection attempt.
[*], ok, now we don't cancel reconnect on drain begin. That's correct:
reconnect feature leads to possibility of long-running requests (up
to reconnect delay). Still, drain begin is not a reason to kill
long requests. We should wait for them.
This also means, that we can again reproduce a dead-lock, described
in 8c517de24a.
Why we are OK with it:
1. Now this is not absolutely-dead dead-lock: the vm is unfrozen
after reconnect delay. Actually 8c517de24a fixed a bug in
NBD logic, that was not described in 8c517de24a and led to
forever dead-lock. The problem was that nobody woke the free_sema
queue, but drain_begin can't finish until there is a request in
free_sema queue. Now we have a reconnect delay timer that works
well.
2. It's not a problem of the NBD driver, but of the ide code,
because it does drain_begin under the global mutex; the problem
doesn't reproduce when using scsi instead of ide.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210902103805.25686-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: grammar and comment tweaks]
Signed-off-by: Eric Blake <eblake@redhat.com>
Split out nbd_recv_coroutine_wake_one(), as it will be used
separately.
Rename the function and add a possibility to wake only first found
sleeping coroutine.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210902103805.25686-4-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: grammar tweak]
Signed-off-by: Eric Blake <eblake@redhat.com>
We are going to use it in nbd_channel_error(), so move it up. Note,
that we are going also refactor and rename
nbd_recv_coroutines_wake_all() in future anyway, so keeping it where it
is and making forward declaration doesn't make real sense.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210902103805.25686-3-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Don't rely on connection being totally broken in case of -EIO. Safer
and more correct is to just shut down the channel anyway, since we
change the state and plan on reconnecting.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210902103805.25686-2-vsementsov@virtuozzo.com>
[eblake: grammar tweaks]
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 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>
Although unlikely, qemu might hang in nbd_send_request().
Allow recovery in this case by registering the yank function before
calling it.
Signed-off-by: Lukas Straub <lukasstraub2@web.de>
Message-Id: <20210704000730.1befb596@gecko.fritz.box>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Avoid accessing QCryptoTLSCreds internals by using
the qcrypto_tls_creds_check_endpoint() helper.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
req->receiving is a flag of request being in one concrete yield point
in nbd_co_do_receive_one_chunk().
Such kind of boolean flag is always better to unset before scheduling
the coroutine, to avoid double scheduling. So, let's be more careful.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-33-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We already have two similar helpers for other state. Let's add another
one for convenience.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-32-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
The only last step we need to reuse the function is coroutine-wrapper.
nbd_open() may be called from non-coroutine context. So, generate the
wrapper and use it.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-31-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We'll need a possibility of non-blocking nbd_co_establish_connection(),
so that it returns immediately, and it returns success only if a
connections was previously established in background.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-30-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Split out the part that we want to reuse for nbd_open().
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-29-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
block/nbd doesn't need underlying sioc channel anymore. So, we can
update nbd/client-connection interface to return only one top-most io
channel, which is more straight forward.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-27-vsementsov@virtuozzo.com>
[eblake: squash in Vladimir's fixes for uninit usage caught by clang]
Signed-off-by: Eric Blake <eblake@redhat.com>
Currently sioc pointer is used just to pass from socket-connection to
nbd negotiation. Drop the field, and use local variables instead. With
next commit we'll update nbd/client-connection.c to behave
appropriately (return only top-most ioc, not two channels).
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-26-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Negotiation during reconnect is now done in a thread, and s->sioc is
not available during negotiation. Negotiation in thread will be
cancelled by nbd_client_connection_release() called from
nbd_clear_bdrvstate(). So, we don't need this code chunk anymore.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-25-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Now that we can opt in to negotiation as part of the client connection
thread, use that to simplify connection_co. This is another step on
the way to moving all reconnect code into NBDClientConnection.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-24-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
To be reused in the following patch.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-23-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Add arguments and logic to support nbd negotiation in the same thread
after successful connection.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-20-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We now have bs-independent connection API, which consists of four
functions:
nbd_client_connection_new()
nbd_client_connection_release()
nbd_co_establish_connection()
nbd_co_establish_connection_cancel()
Move them to a separate file together with NBDClientConnection
structure which becomes private to the new API.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-18-vsementsov@virtuozzo.com>
[eblake: comment tweaks]
Signed-off-by: Eric Blake <eblake@redhat.com>
This is a last step of creating bs-independent nbd connection
interface. With next commit we can finally move it to separate file.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-17-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
This is a step of creating bs-independent nbd connection interface.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-16-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We are going to move the connection code to its own file, and want
clear names and APIs first.
The structure is shared between user and (possibly) several runs of
connect-thread. So it's wrong to call it "thread". Let's rename to
something more generic.
Appropriately rename connect_thread and thr variables to conn.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-15-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
nbd_co_establish_connection_cancel() actually needs only pointer to
NBDConnectThread. So, make it clean.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-14-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We are going to split connection code to a separate file. Now we are
ready to give nbd_co_establish_connection() clean and bs-independent
interface.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-13-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We don't need all these states. The code refactored to use two boolean
variables looks simpler.
While moving the comment in nbd_co_establish_connection() rework it to
give better information. Also, we are going to move the connection code
to separate file and mentioning drained section would be confusing.
Improve also the comment in NBDConnectThread, while dropping removed
state names from it.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-12-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: comment tweak]
Signed-off-by: Eric Blake <eblake@redhat.com>
Instead of managing connect_bh, bh_ctx, and wait_connect fields, we
can use a single link to the waiting coroutine with proper mutex
protection.
So new logic is:
nbd_co_establish_connection() sets wait_co under the mutex, releases
the mutex, then yield()s. Note that wait_co may be scheduled by the
thread immediately after unlocking the mutex. Still, the main thread
(or iothread) will not reach the code for entering the coroutine until
the yield(), so we are safe.
connect_thread_func() and nbd_co_establish_connection_cancel() do
the following to handle wait_co:
Under the mutex, if thr->wait_co is not NULL, make it NULL and
schedule it. This way, we avoid scheduling the coroutine twice.
Still scheduling is a bit different:
In connect_thread_func() we can just call aio_co_wake under mutex,
after commit
[async: the main AioContext is only "current" if under the BQL]
we are sure that aio_co_wake() will not try to acquire the aio context
and do qemu_aio_coroutine_enter() but simply schedule the coroutine by
aio_co_schedule().
nbd_co_establish_connection_cancel() will be called from non-coroutine
context in further patch and will be able to go through
qemu_aio_coroutine_enter() path of aio_co_wake(). So keep current
behavior of waking the coroutine after the critical section.
Also, this commit reduces the dependence of
nbd_co_establish_connection() on the internals of bs (we now use a
generic pointer to the coroutine, instead of direct use of
s->connection_co). This is a step towards splitting the connection
API out of nbd.c.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-11-vsementsov@virtuozzo.com>
Reviewied-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
These fields are write-only. Drop them.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-10-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Message-Id: <20210610100802.5888-9-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Simplify lifetime management of BDRVNBDState->connect_thread by
delaying the possible cleanup of it until the BDRVNBDState itself goes
away.
This also reverts
0267101af6 "block/nbd: fix possible use after free of s->connect_thread"
as now s->connect_thread can't be cleared until the very end.
Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
[vsementsov: rebase, revert 0267101af6 changes]
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[eblake: tweak comment]
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-8-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Detecting monitor by current coroutine works bad when we are not in
coroutine context. And that's exactly so in nbd reconnect code, where
qio_channel_socket_connect_sync() is called from thread.
Monitor is needed only to parse named file descriptor. So, let's just
parse it during nbd_open(), so that all further users of s->saddr don't
need to access monitor.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-7-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
nbd_open() does it (through nbd_establish_connection()).
Actually we lost that call on reconnect path in 1dc4718d84
"block/nbd: use non-blocking connect: fix vm hang on connect()"
when we have introduced reconnect thread.
Fixes: 1dc4718d84
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We have two "return error" paths in nbd_open() after
nbd_process_options(). Actually we should call nbd_clear_bdrvstate()
on these paths. Interesting that nbd_process_options() calls
nbd_clear_bdrvstate() by itself.
Let's fix leaks and refactor things to be more obvious:
- intialize yank at top of nbd_open()
- move yank cleanup to nbd_clear_bdrvstate()
- refactor nbd_open() so that all failure paths except for
yank-register goes through nbd_clear_bdrvstate()
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-4-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
nbd_free_connect_thread leaks the channel object if it hasn't been
stolen.
Unref it and fix the leak.
Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210610100802.5888-3-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Right now, users of qemu_co_sleep_ns_wakeable are simply passing
a pointer to QemuCoSleepState by reference to the function. But
QemuCoSleepState really is just a Coroutine*; making the
content of the struct public is just as efficient and lets us
skip the user_state_pointer indirection.
Since the usage is changed, take the occasion to rename the
struct to QemuCoSleep.
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-6-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
All callers of qemu_co_sleep_wake are checking whether they are passing
a NULL argument inside the pointer-to-pointer: do the check in
qemu_co_sleep_wake itself.
As a side effect, qemu_co_sleep_wake can be called more than once and
it will only wake the coroutine once; after the first time, the argument
will be set to NULL via *sleep_state->user_state_pointer. However, this
would not be safe unless co_sleep_cb keeps using the QemuCoSleepState*
directly, so make it go through the pointer-to-pointer instead.
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20210517100548.28806-4-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
If on nbd_close() we detach the thread (in
nbd_co_establish_connection_cancel() thr->state becomes
CONNECT_THREAD_RUNNING_DETACHED), after that point we should not use
s->connect_thread (which is set to NULL), as running thread may free it
at any time.
Still nbd_co_establish_connection() does exactly this: it saves
s->connect_thread to local variable (just for better code style) and
use it even after yield point, when thread may be already detached.
Fix that. Also check thr to be non-NULL on
nbd_co_establish_connection() start for safety.
After this patch "case CONNECT_THREAD_RUNNING_DETACHED" becomes
impossible in the second switch in nbd_co_establish_connection().
Still, don't add extra abort() just before the release. If it somehow
possible to reach this "case:" it won't hurt. Anyway, good refactoring
of all this reconnect mess will come soon.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210406155114.1057355-1-vsementsov@virtuozzo.com>
Reviewed-by: Roman Kagan <rvkagan@yandex-team.ru>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Just stop waiting for connection in existing requests.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20210205163720.887197-3-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
When an NBD block driver state is moved from one aio_context to another
(e.g. when doing a drain in a migration thread),
nbd_client_attach_aio_context_bh is executed that enters the connection
coroutine.
However, the assumption that ->connection_co is always present here
appears incorrect: the connection may have encountered an error other
than -EIO in the underlying transport, and thus may have decided to quit
rather than keep trying to reconnect, and therefore it may have
terminated the connection coroutine. As a result an attempt to reassign
the client in this state (NBD_CLIENT_QUIT) to a different aio_context
leads to a null pointer dereference:
#0 qio_channel_detach_aio_context (ioc=0x0)
at /build/qemu-gYtjVn/qemu-5.0.1/io/channel.c:452
#1 0x0000562a242824b3 in bdrv_detach_aio_context (bs=0x562a268d6a00)
at /build/qemu-gYtjVn/qemu-5.0.1/block.c:6151
#2 bdrv_set_aio_context_ignore (bs=bs@entry=0x562a268d6a00,
new_context=new_context@entry=0x562a260c9580,
ignore=ignore@entry=0x7feeadc9b780)
at /build/qemu-gYtjVn/qemu-5.0.1/block.c:6230
#3 0x0000562a24282969 in bdrv_child_try_set_aio_context
(bs=bs@entry=0x562a268d6a00, ctx=0x562a260c9580,
ignore_child=<optimized out>, errp=<optimized out>)
at /build/qemu-gYtjVn/qemu-5.0.1/block.c:6332
#4 0x0000562a242bb7db in blk_do_set_aio_context (blk=0x562a2735d0d0,
new_context=0x562a260c9580,
update_root_node=update_root_node@entry=true, errp=errp@entry=0x0)
at /build/qemu-gYtjVn/qemu-5.0.1/block/block-backend.c:1989
#5 0x0000562a242be0bd in blk_set_aio_context (blk=<optimized out>,
new_context=<optimized out>, errp=errp@entry=0x0)
at /build/qemu-gYtjVn/qemu-5.0.1/block/block-backend.c:2010
#6 0x0000562a23fbd953 in virtio_blk_data_plane_stop (vdev=<optimized
out>)
at /build/qemu-gYtjVn/qemu-5.0.1/hw/block/dataplane/virtio-blk.c:292
#7 0x0000562a241fc7bf in virtio_bus_stop_ioeventfd (bus=0x562a260dbf08)
at /build/qemu-gYtjVn/qemu-5.0.1/hw/virtio/virtio-bus.c:245
#8 0x0000562a23fefb2e in virtio_vmstate_change (opaque=0x562a260dbf90,
running=0, state=<optimized out>)
at /build/qemu-gYtjVn/qemu-5.0.1/hw/virtio/virtio.c:3220
#9 0x0000562a2402ebfd in vm_state_notify (running=running@entry=0,
state=state@entry=RUN_STATE_FINISH_MIGRATE)
at /build/qemu-gYtjVn/qemu-5.0.1/softmmu/vl.c:1275
#10 0x0000562a23f7bc02 in do_vm_stop (state=RUN_STATE_FINISH_MIGRATE,
send_stop=<optimized out>)
at /build/qemu-gYtjVn/qemu-5.0.1/cpus.c:1032
#11 0x0000562a24209765 in migration_completion (s=0x562a260e83a0)
at /build/qemu-gYtjVn/qemu-5.0.1/migration/migration.c:2914
#12 migration_iteration_run (s=0x562a260e83a0)
at /build/qemu-gYtjVn/qemu-5.0.1/migration/migration.c:3275
#13 migration_thread (opaque=opaque@entry=0x562a260e83a0)
at /build/qemu-gYtjVn/qemu-5.0.1/migration/migration.c:3439
#14 0x0000562a2435ca96 in qemu_thread_start (args=<optimized out>)
at /build/qemu-gYtjVn/qemu-5.0.1/util/qemu-thread-posix.c:519
#15 0x00007feed31466ba in start_thread (arg=0x7feeadc9c700)
at pthread_create.c:333
#16 0x00007feed2e7c41d in __GI___sysctl (name=0x0, nlen=608471908,
oldval=0x562a2452b138, oldlenp=0x0, newval=0x562a2452c5e0
<__func__.28102>, newlen=0)
at ../sysdeps/unix/sysv/linux/sysctl.c:30
#17 0x0000000000000000 in ?? ()
Fix it by checking that the connection coroutine is non-null before
trying to enter it. If it is null, no entering is needed, as the
connection is probably going down anyway.
Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210129073859.683063-3-rvkagan@yandex-team.ru>
Signed-off-by: Eric Blake <eblake@redhat.com>
When the reconnect in NBD client is in progress, the iochannel used for
NBD connection doesn't exist. Therefore an attempt to detach it from
the aio_context of the parent BlockDriverState results in a NULL pointer
dereference.
The problem is triggerable, in particular, when an outgoing migration is
about to finish, and stopping the dataplane tries to move the
BlockDriverState from the iothread aio_context to the main loop. If the
NBD connection is lost before this point, and the NBD client has entered
the reconnect procedure, QEMU crashes:
#0 qemu_aio_coroutine_enter (ctx=0x5618056c7580, co=0x0)
at /build/qemu-6MF7tq/qemu-5.0.1/util/qemu-coroutine.c:109
#1 0x00005618034b1b68 in nbd_client_attach_aio_context_bh (
opaque=0x561805ed4c00) at /build/qemu-6MF7tq/qemu-5.0.1/block/nbd.c:164
#2 0x000056180353116b in aio_wait_bh (opaque=0x7f60e1e63700)
at /build/qemu-6MF7tq/qemu-5.0.1/util/aio-wait.c:55
#3 0x0000561803530633 in aio_bh_call (bh=0x7f60d40a7e80)
at /build/qemu-6MF7tq/qemu-5.0.1/util/async.c:136
#4 aio_bh_poll (ctx=ctx@entry=0x5618056c7580)
at /build/qemu-6MF7tq/qemu-5.0.1/util/async.c:164
#5 0x0000561803533e5a in aio_poll (ctx=ctx@entry=0x5618056c7580,
blocking=blocking@entry=true)
at /build/qemu-6MF7tq/qemu-5.0.1/util/aio-posix.c:650
#6 0x000056180353128d in aio_wait_bh_oneshot (ctx=0x5618056c7580,
cb=<optimized out>, opaque=<optimized out>)
at /build/qemu-6MF7tq/qemu-5.0.1/util/aio-wait.c:71
#7 0x000056180345c50a in bdrv_attach_aio_context (new_context=0x5618056c7580,
bs=0x561805ed4c00) at /build/qemu-6MF7tq/qemu-5.0.1/block.c:6172
#8 bdrv_set_aio_context_ignore (bs=bs@entry=0x561805ed4c00,
new_context=new_context@entry=0x5618056c7580,
ignore=ignore@entry=0x7f60e1e63780)
at /build/qemu-6MF7tq/qemu-5.0.1/block.c:6237
#9 0x000056180345c969 in bdrv_child_try_set_aio_context (
bs=bs@entry=0x561805ed4c00, ctx=0x5618056c7580,
ignore_child=<optimized out>, errp=<optimized out>)
at /build/qemu-6MF7tq/qemu-5.0.1/block.c:6332
#10 0x00005618034957db in blk_do_set_aio_context (blk=0x56180695b3f0,
new_context=0x5618056c7580, update_root_node=update_root_node@entry=true,
errp=errp@entry=0x0)
at /build/qemu-6MF7tq/qemu-5.0.1/block/block-backend.c:1989
#11 0x00005618034980bd in blk_set_aio_context (blk=<optimized out>,
new_context=<optimized out>, errp=errp@entry=0x0)
at /build/qemu-6MF7tq/qemu-5.0.1/block/block-backend.c:2010
#12 0x0000561803197953 in virtio_blk_data_plane_stop (vdev=<optimized out>)
at /build/qemu-6MF7tq/qemu-5.0.1/hw/block/dataplane/virtio-blk.c:292
#13 0x00005618033d67bf in virtio_bus_stop_ioeventfd (bus=0x5618056d9f08)
at /build/qemu-6MF7tq/qemu-5.0.1/hw/virtio/virtio-bus.c:245
#14 0x00005618031c9b2e in virtio_vmstate_change (opaque=0x5618056d9f90,
running=0, state=<optimized out>)
at /build/qemu-6MF7tq/qemu-5.0.1/hw/virtio/virtio.c:3220
#15 0x0000561803208bfd in vm_state_notify (running=running@entry=0,
state=state@entry=RUN_STATE_FINISH_MIGRATE)
at /build/qemu-6MF7tq/qemu-5.0.1/softmmu/vl.c:1275
#16 0x0000561803155c02 in do_vm_stop (state=RUN_STATE_FINISH_MIGRATE,
send_stop=<optimized out>) at /build/qemu-6MF7tq/qemu-5.0.1/cpus.c:1032
#17 0x00005618033e3765 in migration_completion (s=0x5618056e6960)
at /build/qemu-6MF7tq/qemu-5.0.1/migration/migration.c:2914
#18 migration_iteration_run (s=0x5618056e6960)
at /build/qemu-6MF7tq/qemu-5.0.1/migration/migration.c:3275
#19 migration_thread (opaque=opaque@entry=0x5618056e6960)
at /build/qemu-6MF7tq/qemu-5.0.1/migration/migration.c:3439
#20 0x0000561803536ad6 in qemu_thread_start (args=<optimized out>)
at /build/qemu-6MF7tq/qemu-5.0.1/util/qemu-thread-posix.c:519
#21 0x00007f61085d06ba in start_thread ()
from /lib/x86_64-linux-gnu/libpthread.so.0
#22 0x00007f610830641d in sysctl () from /lib/x86_64-linux-gnu/libc.so.6
#23 0x0000000000000000 in ?? ()
Fix it by checking that the iochannel is non-null before trying to
detach it from the aio_context. If it is null, no detaching is needed,
and it will get reattached in the proper aio_context once the connection
is reestablished.
Signed-off-by: Roman Kagan <rvkagan@yandex-team.ru>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210129073859.683063-2-rvkagan@yandex-team.ru>
Signed-off-by: Eric Blake <eblake@redhat.com>
Register a yank function which shuts down the socket and sets
s->state = NBD_CLIENT_QUIT. This is the same behaviour as if an
error occured.
Signed-off-by: Lukas Straub <lukasstraub2@web.de>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <b73eb07db6d1fcd00667beb13ae6117260f002c3.1609167865.git.lukasstraub2@web.de>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
This commit is the result of running the timer-del-timer-free.cocci
script on the whole source tree.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Corey Minyard <cminyard@mvista.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-4-peter.maydell@linaro.org
Allow the server to expose an additional metacontext to be requested
by savvy clients. qemu-nbd adds a new option -A to expose the
qemu:allocation-depth metacontext through NBD_CMD_BLOCK_STATUS; this
can also be set via QMP when using block-export-add.
qemu as client is hacked into viewing the key aspects of this new
context by abusing the already-experimental x-dirty-bitmap option to
collapse all depths greater than 2, which results in a tri-state value
visible in the output of 'qemu-img map --output=json' (yes, that means
x-dirty-bitmap is now a bit of a misnomer, but I didn't feel like
renaming it as it would introduce a needless break of back-compat,
even though we make no compat guarantees with x- members):
unallocated (depth 0) => "zero":false, "data":true
local (depth 1) => "zero":false, "data":false
backing (depth 2+) => "zero":true, "data":true
libnbd as client is probably a nicer way to get at the information
without having to decipher such hacks in qemu as client. ;)
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20201027050556.269064-11-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
In a recent commit 12c75e20a2 we've improved
nbd_co_reconnect_loop() to not make drain wait for additional sleep.
Similarly, we shouldn't try to connect, if previous sleep was
interrupted by drain begin, otherwise drain_begin will have to wait for
the whole connection attempt.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200903190301.367620-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
reconnect-delay has a design flaw: we handle it in the same loop where
we do connection attempt. So, reconnect-delay may be exceeded by
unpredictable time of connection attempt.
Let's instead use separate timer.
How to reproduce the bug:
1. Create an image on node1:
qemu-img create -f qcow2 xx 100M
2. Start NBD server on node1:
qemu-nbd xx
3. On node2 start qemu-io:
./build/qemu-io --image-opts \
driver=nbd,server.type=inet,server.host=192.168.100.5,server.port=10809,reconnect-delay=15
4. Type 'read 0 512' in qemu-io interface to check that connection
works
Be careful: you should make steps 5-7 in a short time, less than 15
seconds.
5. Kill nbd server on node1
6. Run 'read 0 512' in qemu-io interface again, to be sure that nbd
client goes to reconnect loop.
7. On node1 run the following command
sudo iptables -A INPUT -p tcp --dport 10809 -j DROP
This will make the connect() call of qemu-io at node2 take a long time.
And you'll see that read command in qemu-io will hang for a long time,
more than 15 seconds specified by reconnect-delay parameter. It's the
bug.
8. Don't forget to drop iptables rule on node1:
sudo iptables -D INPUT -p tcp --dport 10809 -j DROP
Important note: Step [5] is necessary to reproduce _this_ bug. If we
miss step [5], the read command (step 6) will hang for a long time and
this commit doesn't help, because there will be not long connect() to
unreachable host, but long sendmsg() to unreachable host, which should
be fixed by enabling and adjusting keep-alive on the socket, which is a
thing for further patch set.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200903190301.367620-4-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Don't use nbd_client_detach_aio_context() driver handler where we want
to finalize the connection. We should directly use
qio_channel_detach_aio_context() in such cases. Driver handler may (and
will) contain another things, unrelated to the qio channel.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200903190301.367620-3-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
We pause reconnect process during drained section. So, if we have some
requests, waiting for reconnect we should cancel them, otherwise they
deadlock the drained section.
How to reproduce:
1. Create an image:
qemu-img create -f qcow2 xx 100M
2. Start NBD server:
qemu-nbd xx
3. Start vm with second nbd disk on node2, like this:
./build/x86_64-softmmu/qemu-system-x86_64 -nodefaults -drive \
file=/work/images/cent7.qcow2 -drive \
driver=nbd,server.type=inet,server.host=192.168.100.5,server.port=10809,reconnect-delay=60 \
-vnc :0 -m 2G -enable-kvm -vga std
4. Access the vm through vnc (or some other way?), and check that NBD
drive works:
dd if=/dev/sdb of=/dev/null bs=1M count=10
- the command should succeed.
5. Now, kill the nbd server, and run dd in the guest again:
dd if=/dev/sdb of=/dev/null bs=1M count=10
Now Qemu is trying to reconnect, and dd-generated requests are waiting
for the connection (they will wait up to 60 seconds (see
reconnect-delay option above) and than fail). But suddenly, vm may
totally hang in the deadlock. You may need to increase reconnect-delay
period to catch the dead-lock.
VM doesn't respond because drain dead-lock happens in cpu thread with
global mutex taken. That's not good thing by itself and is not fixed
by this commit (true way is using iothreads). Still this commit fixes
drain dead-lock itself.
Note: probably, we can instead continue to reconnect during drained
section. To achieve this, we may move negotiation to the connect thread
to make it independent of bs aio context. But expanding drained section
doesn't seem good anyway. So, let's now fix the bug the simplest way.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200903190301.367620-2-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
