[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 15/17] migration/block: Fix possible race with block_inactive
From: |
Fabiano Rosas |
Subject: |
[PULL 15/17] migration/block: Fix possible race with block_inactive |
Date: |
Tue, 17 Dec 2024 14:48:53 -0300 |
From: Peter Xu <peterx@redhat.com>
Src QEMU sets block_inactive=true very early before the invalidation takes
place. It means if something wrong happened during setting the flag but
before reaching qemu_savevm_state_complete_precopy_non_iterable() where it
did the invalidation work, it'll make block_inactive flag inconsistent.
For example, think about when qemu_savevm_state_complete_precopy_iterable()
can fail: it will have block_inactive set to true even if all block drives
are active.
Fix that by only update the flag after the invalidation is done.
No Fixes for any commit, because it's not an issue if bdrv_activate_all()
is re-entrant upon all-active disks - false positive block_inactive can
bring nothing more than "trying to active the blocks but they're already
active". However let's still do it right to avoid the inconsistent flag
v.s. reality.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Message-Id: <20241206230838.1111496-6-peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
---
migration/migration.c | 9 +++------
migration/savevm.c | 2 ++
2 files changed, 5 insertions(+), 6 deletions(-)
diff --git a/migration/migration.c b/migration/migration.c
index e6db9cfc50..ba5deec5bc 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -2749,14 +2749,11 @@ static int migration_completion_precopy(MigrationState
*s,
goto out_unlock;
}
- /*
- * Inactivate disks except in COLO, and track that we have done so in order
- * to remember to reactivate them if migration fails or is cancelled.
- */
- s->block_inactive = !migrate_colo();
migration_rate_set(RATE_LIMIT_DISABLED);
+
+ /* Inactivate disks except in COLO */
ret = qemu_savevm_state_complete_precopy(s->to_dst_file, false,
- s->block_inactive);
+ !migrate_colo());
out_unlock:
bql_unlock();
return ret;
diff --git a/migration/savevm.c b/migration/savevm.c
index 80726726fe..706b77ffab 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1558,6 +1558,8 @@ int
qemu_savevm_state_complete_precopy_non_iterable(QEMUFile *f,
qemu_file_set_error(f, ret);
return ret;
}
+ /* Remember that we did this */
+ s->block_inactive = true;
}
if (!in_postcopy) {
/* Postcopy stream will still be going */
--
2.35.3
- [PULL 01/17] migration/multifd: Fix compile error caused by page_size usage, (continued)
- [PULL 01/17] migration/multifd: Fix compile error caused by page_size usage, Fabiano Rosas, 2024/12/17
- [PULL 03/17] migration/multifd: Allow to sync with sender threads only, Fabiano Rosas, 2024/12/17
- [PULL 06/17] migration/multifd: Remove sync processing on postcopy, Fabiano Rosas, 2024/12/17
- [PULL 09/17] migration/multifd: Fix compat with QEMU < 9.0, Fabiano Rosas, 2024/12/17
- [PULL 02/17] migration/multifd: Further remove the SYNC on complete, Fabiano Rosas, 2024/12/17
- [PULL 10/17] s390x: Fix CSS migration, Fabiano Rosas, 2024/12/17
- [PULL 08/17] migration/multifd: Document the reason to sync for save_setup(), Fabiano Rosas, 2024/12/17
- [PULL 11/17] migration: Add helper to get target runstate, Fabiano Rosas, 2024/12/17
- [PULL 12/17] qmp/cont: Only activate disks if migration completed, Fabiano Rosas, 2024/12/17
- [PULL 13/17] migration/block: Make late-block-active the default, Fabiano Rosas, 2024/12/17
- [PULL 15/17] migration/block: Fix possible race with block_inactive,
Fabiano Rosas <=
- [PULL 17/17] tests/qtest/migration: Fix compile errors when CONFIG_UADK is set, Fabiano Rosas, 2024/12/17
- [PULL 14/17] migration/block: Apply late-block-active behavior to postcopy, Fabiano Rosas, 2024/12/17
- [PULL 16/17] migration/block: Rewrite disk activation, Fabiano Rosas, 2024/12/17
- Re: [PULL 00/17] Migration patches for 2024-12-17, Stefan Hajnoczi, 2024/12/19