qemu/include/block
Greg Kurz 1a6d3bd229 block: End quiescent sections when a BDS is deleted
If a BDS gets deleted during blk_drain_all(), it might miss a
call to bdrv_do_drained_end(). This means missing a call to
aio_enable_external() and the AIO context remains disabled for
ever. This can cause a device to become irresponsive and to
disrupt the guest execution, ie. hang, loop forever or worse.

This scenario is quite easy to encounter with virtio-scsi
on POWER when punching multiple blockdev-create QMP commands
while the guest is booting and it is still running the SLOF
firmware. This happens because SLOF disables/re-enables PCI
devices multiple times via IO/MEM/MASTER bits of PCI_COMMAND
register after the initial probe/feature negotiation, as it
tends to work with a single device at a time at various stages
like probing and running block/network bootloaders without
doing a full reset in-between. This naturally generates many
dataplane stops and starts, and thus many drain sections that
can race with blockdev_create_run(). In the end, SLOF bails
out.

It is somehow reproducible on x86 but it requires to generate
articial dataplane start/stop activity with stop/cont QMP
commands. In this case, seabios ends up looping for ever,
waiting for the virtio-scsi device to send a response to
a command it never received.

Add a helper that pairs all previously called bdrv_do_drained_begin()
with a bdrv_do_drained_end() and call it from bdrv_close().
While at it, update the "/bdrv-drain/graph-change/drain_all"
test in test-bdrv-drain so that it can catch the issue.

BugId: https://bugzilla.redhat.com/show_bug.cgi?id=1874441
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <160346526998.272601.9045392804399803158.stgit@bahia.lan>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2020-10-27 15:26:20 +01:00
..
accounting.h block: add empty account cookie type 2019-10-10 10:56:18 +02:00
aio-wait.h qemu/atomic.h: rename atomic_ to qatomic_ 2020-09-23 16:07:44 +01:00
aio.h util/async: Add aio_co_reschedule_self() 2020-10-09 07:08:20 +02:00
aio_task.h block: introduce aio task pool 2019-10-10 10:56:17 +02:00
block-copy.h block/block-copy: hide structure definitions 2020-03-11 12:42:30 +01:00
block-hmp-cmds.h monitor/hmp: move hmp_info_block* to block-hmp-cmds.c 2020-03-09 18:20:21 +00:00
block.h block: End quiescent sections when a BDS is deleted 2020-10-27 15:26:20 +01:00
block_backup.h block/backup: drop unused synchronization interface 2018-12-14 11:52:40 +01:00
block_int.h block: Leave BDS.backing_{file,format} constant 2020-09-07 12:31:31 +02:00
blockjob.h blockdev: blockdev_mark_auto_del: drop usage of bs->job 2019-06-18 16:41:10 +02:00
blockjob_int.h job: drop job_drain 2019-09-10 08:58:43 +02:00
dirty-bitmap.h block/dirty-bitmap: add bdrv_has_named_bitmaps helper 2020-05-28 13:15:22 -05:00
export.h block/export: Move blk to BlockExport 2020-10-02 15:46:40 +02:00
nbd.h nbd: Merge nbd_export_new() and nbd_export_create() 2020-10-02 15:46:40 +02:00
nvme.h hw/block/nvme: be consistent about zeros vs zeroes 2020-09-02 08:48:50 +02:00
qapi.h qapi: Allow getting flat output from 'query-named-block-nodes' 2020-02-20 16:43:42 +01:00
qdict.h block: Factor out qobject_input_visitor_new_flat_confused() 2018-06-15 14:49:44 +02:00
raw-aio.h block/io_uring: implements interfaces for io_uring 2020-01-30 20:59:41 +00:00
snapshot.h migration: introduce icount field for snapshots 2020-10-06 08:34:49 +02:00
thread-pool.h Clean up decorations and whitespace around header guards 2016-07-12 16:20:46 +02:00
throttle-groups.h Use OBJECT_DECLARE_SIMPLE_TYPE when possible 2020-09-18 14:12:32 -04:00
write-threshold.h include: Make headers more self-contained 2019-08-16 13:31:51 +02:00