backup-top failure path fix
v2 cover backup-top failure path fix
Hi all!
Here is a small crash fix.
v2: rebase on Max's block tree
01: add Max's r-b
02: improve wording
Vladimir Sementsov-Ogievskiy (2):
block/backup-top: fix failure path
iotests: add test for backup-top failure on permission activation
block/backup-top.c | 21 +++++---- tests/qemu-iotests/283 | 92 ++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/283.out | 8 ++++ tests/qemu-iotests/group | 1 + 4 files changed, 113 insertions(+), 9 deletions(-) create mode 100644 tests/qemu-iotests/283 create mode 100644 tests/qemu-iotests/283.out
Signature · 1 line
2.21.0
PATCH v2 1/2 block/backup-top: fix failure path
We can't access top after call bdrv_backup_top_drop, as it is already freed at this time.
Also, no needs to unref target child by hand, it will be unrefed on bdrv_close() automatically.
So, just do bdrv_backup_top_drop if append succeed and one bdrv_unref otherwise.
Note, that in !appended case bdrv_unref(top) moved into drained section on source. It doesn't really matter, but just for code simplicity.
--- block/backup-top.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-)
block/backup-top.c +12 -9
@@ -190,6 +190,7 @@ BlockDriverState *bdrv_backup_top_append(BlockDriverState *source,BlockDriverState *top =bdrv_new_open_driver (&bdrv_backup_top_filter, filter_node_name,BDRV_O_RDWR , errp); +bool appended = false;if (!top) {return NULL ;
@@ -212,8 +213,9 @@ BlockDriverState *bdrv_backup_top_append(BlockDriverState *source,bdrv_append (top, source, &local_err);if (local_err) {error_prepend (&local_err,"Cannot append backup-top filter: " ); - goto append_failed; + goto fail; } + appended = true;/* * bdrv_append() finished successfully, now we can require permissions
@@ -224,14 +226,14 @@ BlockDriverState *bdrv_backup_top_append(BlockDriverState *source,if (local_err) {error_prepend (&local_err,"Cannot set permissions for backup-top filter: " ); - goto failed_after_append; + goto fail; } state->bcs =block_copy_state_new (top->backing, state->target, cluster_size, write_flags, &local_err);if (local_err) {error_prepend (&local_err,"Cannot create block-copy-state: " ); - goto failed_after_append; + goto fail; } *bcs = state->bcs;
@@ -239,14 +241,15 @@ BlockDriverState *bdrv_backup_top_append(BlockDriverState *source,return top; -failed_after_append: - state->active = false; -bdrv_backup_top_drop (top); +fail: +if (appended) { + state->active = false; +bdrv_backup_top_drop (top); + }else { +bdrv_unref (top); + } -append_failed:bdrv_drained_end (source); -bdrv_unref_child (top, state->target); -bdrv_unref (top);error_propagate (errp, local_err);return NULL ;
Signature · 1 line
2.21.0
PATCH v2 2/2 iotests: add test for backup-top failure on permission activation
In reply to Vladimir Sementsov-OgievskiyThis test checks that bug is really fixed by previous commit.
--- tests/qemu-iotests/283 | 92 ++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/283.out | 8 ++++ tests/qemu-iotests/group | 1 + 3 files changed, 101 insertions(+) create mode 100644 tests/qemu-iotests/283 create mode 100644 tests/qemu-iotests/283.out
tests/qemu-iotests/283 new file +92 -0
@@ -0,0 +1,92 @@ +#!/usr/bin/env python +# +# Test for backup-top filter permission activation failure +# +# Copyright (c) 2019 Virtuozzo International GmbH. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# + +import iotests + +# The test is unrelated to formats, restrict it to qcow2 to avoid extra runs +iotests.verify_image_format(supported_fmts=['qcow2']) + +size = 1024 * 1024 + +""" Test description + +When performing a backup, all writes on the source subtree must go through the +backup-top filter so it can copy all data to the target before it is changed. +backup-top filter is appended above source node, to achieve this thing, so all +parents of source node are handled. A configuration with side parents of source +sub-tree with write permission is unsupported (we'd have append several +backup-top filter like nodes to handle such parents). The test create an +example of such configuration and checks that a backup is then not allowed +(blockdev-backup command should fail). + +The configuration: + + ┌────────┐ target ┌─────────────┐ + │ target │ ◀─────── │ backup_top │ + └────────┘ └─────────────┘ + │ + │ backing + ▼ + ┌─────────────┐ + │ source │ + └─────────────┘ + │ + │ file + ▼ + ┌─────────────┐ write perm ┌───────┐ + │ base │ ◀──────────── │ other │ + └─────────────┘ └───────┘ + +On activation (see .active field of backup-top state in block/backup-top.c), +backup-top is going to unshare write permission on its source child. Write +unsharing will be propagated to the "source->base" link and will conflict with +other node write permission. So permission update will fail and backup job will +not be started. + +Note, that the only thing which prevents backup of running on such +configuration is default permission propagation scheme. It may be altered by +different block drivers, so backup will run in invalid configuration. But +something is better than nothing. Also, before the previous commit (commit +preceding this test creation), starting backup on such configuration led to +crash, so current "something" is a lot better, and this test actual goal is +to check that crash is fixed :) +""" + +vm = iotests.VM() +vm.launch() + +vm.qmp_log('blockdev-add', **{'node-name': 'target', 'driver': 'null-co'}) + +vm.qmp_log('blockdev-add', **{ + 'node-name': 'source', + 'driver': 'blkdebug', + 'image': {'node-name': 'base', 'driver': 'null-co', 'size': size} +}) + +vm.qmp_log('blockdev-add', **{ + 'node-name': 'other', + 'driver': 'blkdebug', + 'image': 'base', + 'take-child-perms': ['write'] +}) + +vm.qmp_log('blockdev-backup', sync='full', device='source', target='target') + +vm.shutdown()
tests/qemu-iotests/283.out new file +8 -0
@@ -0,0 +1,8 @@ +{"execute": "blockdev-add", "arguments": {"driver": "null-co", "node-name": "target"}} +{"return": {}} +{"execute": "blockdev-add", "arguments": {"driver": "blkdebug", "image": {"driver": "null-co", "node-name": "base", "size": 1048576}, "node-name": "source"}} +{"return": {}} +{"execute": "blockdev-add", "arguments": {"driver": "blkdebug", "image": "base", "node-name": "other", "take-child-perms": ["write"]}} +{"return": {}} +{"execute": "blockdev-backup", "arguments": {"device": "source", "sync": "full", "target": "target"}} +{"error": {"class": "GenericError", "desc": "Cannot set permissions for backup-top filter: Conflicts with use by other as 'image', which uses 'write' on base"}}
tests/qemu-iotests/group +1 -0
@@ -289,3 +289,4 @@ 279 rw backing quick 280 rw migration quick 281 rw quick +283 auto quick
Signature · 1 line
2.21.0
Re: [PATCH v2 2/2] iotests: add test for backup-top failure on permission activation
Quoted 18 lines · Vladimir Sementsov-Ogievskiy
This test checks that bug is really fixed by previous commit.
Cc [email protected] # v4.2.0Signed-off-by Vladimir Sementsov-Ogievskiy <[email protected]>--- tests/qemu-iotests/283 | 92 ++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/283.out | 8 ++++ tests/qemu-iotests/group | 1 + 3 files changed, 101 insertions(+) create mode 100644 tests/qemu-iotests/283 create mode 100644 tests/qemu-iotests/283.outtests/qemu-iotests/283 new file +0
-0@@ -0,0 +1,92 @@
[...]
Quoted 32 lines
+""" Test description
tests/qemu-iotests/283 new file +31
-0+ +When performing a backup, all writes on the source subtree must go through the +backup-top filter so it can copy all data to the target before it is changed. +backup-top filter is appended above source node, to achieve this thing, so all +parents of source node are handled. A configuration with side parents of source +sub-tree with write permission is unsupported (we'd have append several +backup-top filter like nodes to handle such parents). The test create an +example of such configuration and checks that a backup is then not allowed +(blockdev-backup command should fail). + +The configuration: + + ┌────────┐ target ┌─────────────┐ + │ target │ ◀─────── │ backup_top │ + └────────┘ └─────────────┘ + │ + │ backing + ▼ + ┌─────────────┐ + │ source │ + └─────────────┘ + │ + │ file + ▼ + ┌─────────────┐ write perm ┌───────┐ + │ base │ ◀──────────── │ other │ + └─────────────┘ └───────┘ + +On activation (see .active field of backup-top state in block/backup-top.c), +backup-top is going to unshare write permission on its source child. Write +unsharing will be propagated to the "source->base" link and will conflict with
tests/qemu-iotests/283 line 69
Quoted 10 lines
tests/qemu-iotests/283 new file +10
-0+other node write permission. So permission update will fail and backup job will +not be started. + +Note, that the only thing which prevents backup of running on such +configuration is default permission propagation scheme. It may be altered by +different block drivers, so backup will run in invalid configuration. But +something is better than nothing. Also, before the previous commit (commit +preceding this test creation), starting backup on such configuration led to +crash, so current "something" is a lot better, and this test actual goal is +to check that crash is fixed :)
Thanks a lot for bearing with me!
I was wondering whether this is the first smiley in our code, but it isn’t. (Not unfortunately, I think. :-)) It’s also not the first smiley in the iotests, but the second one! (As far as I can tell.)
Max
Max Reitz
AppliedQuoted 3 lines · Vladimir Sementsov-Ogievskiy
Hi all!
Here is a small crash fix.
Thanks, applied to my block branch:
https://git.xanclic.moe/XanClic/qemu/commits/branch/block
Max
Quoted 9 lines · Max Reitz
Quoted 3 lines · Vladimir Sementsov-Ogievskiy
Hi all!
Here is a small crash fix.
Thanks, applied to my block branch:
Thank you Max!
Signature · 2 lines
Best regards, Vladimir