Adds a fast path on aio context setting preventing
unnecessary context setting routine.
Also, it prevents issues with cyclic walk of child
bds-es appeared because of registering aio walking
notifiers:
This can happen because of "new" context attachment to VM disk bds.
When attaching a new context the corresponding aio context handler is
called for each of aio_notifiers registered on the VM disk bds context.
Among those handlers, there is the block_job_attached_aio_context handler
which sets a new aio context for the block job bds. When doing so,
the old context is detached from all the block job bds children and one of
them is the VM disk bds, serving as backing store for the blockjob bds,
although the VM disk bds is actually the initializer of that process.
Since the VM disk bds is protected with walking_aio_notifiers flag
from double processing in recursive calls, the assert fires.
Signed-off-by: Denis Plotnikov <email address hidden>
Signed-off-by: Kevin Wolf <email address hidden>
I found the commit that introduced this regression.
commit 57830a499f7c815 bb0cb325c94a3d8 c910d13cfa (HEAD)
Author: Denis Plotnikov <email address hidden>
Date: Fri Feb 15 16:03:25 2019 +0300
block: don't set the same context
Adds a fast path on aio context setting preventing
unnecessary context setting routine.
Also, it prevents issues with cyclic walk of child
bds-es appeared because of registering aio walking
notifiers:
Call stack:
0 __GI_raise aio_context (bs=0x55f54d65c000) <<< aio_context (bs=0x55f54fc8a800) aio_context (bs=0x55f54fc8a800, ...) attached_ aio_context aio_context (bs=0x55f54d65c000, ...) <<< aio_context (bs=0x55f54d65c000) blk_data_ plane_stop bus_stop_ ioeventfd vmstate_ change STATE_SHUTDOWN) RUN_STATE_ SHUTDOWN, send_stop=true) RUN_STATE_ SHUTDOWN) should_ exit
1 __GI_abort
2 __assert_fail_base
3 __GI___assert_fail
4 bdrv_detach_
5 bdrv_detach_
6 bdrv_set_
7 block_job_
8 bdrv_attach_
9 bdrv_set_
10 blk_set_aio_context
11 virtio_
12 virtio_
13 virtio_
14 vm_state_notify (running=0, state=RUN_
15 do_vm_stop (state=
16 vm_stop (state=
17 main_loop_
18 main_loop
19 main
This can happen because of "new" context attachment to VM disk bds. attached_ aio_context handler aio_notifiers flag
When attaching a new context the corresponding aio context handler is
called for each of aio_notifiers registered on the VM disk bds context.
Among those handlers, there is the block_job_
which sets a new aio context for the block job bds. When doing so,
the old context is detached from all the block job bds children and one of
them is the VM disk bds, serving as backing store for the blockjob bds,
although the VM disk bds is actually the initializer of that process.
Since the VM disk bds is protected with walking_
from double processing in recursive calls, the assert fires.
Signed-off-by: Denis Plotnikov <email address hidden>
Signed-off-by: Kevin Wolf <email address hidden>
diff --git a/block.c b/block.c .0c12632661 100644 aio_context( BlockDriverStat e *bs, AioContext *new_context) aio_context( bs);
index 4ad0e90d7e.
--- a/block.c
+++ b/block.c
@@ -5265,6 +5265,10 @@ void bdrv_set_
{
AioContext *ctx = bdrv_get_
+ if (ctx == new_context) { disable_ external( ctx); parent_ drained_ begin(bs, NULL, false); drain(bs) ; /* ensure there are no in-flight requests */
+ return;
+ }
+
aio_
bdrv_
bdrv_