From e7fa06c5ad14fc4df863265ca4723c9b74368682 Mon Sep 17 00:00:00 2001 From: Kevin Wolf Date: Tue, 4 Feb 2025 22:14:02 +0100 Subject: [PATCH 17/22] block: Drain nodes before inactivating them RH-Author: Kevin Wolf RH-MergeRequest: 340: QMP command for block device reactivation after migration RH-Jira: RHEL-54670 RH-Acked-by: Eric Blake RH-Acked-by: Stefan Hajnoczi RH-Commit: [17/22] 64b1f9168b91793dcf3cf1df39e4799c7dc0f73c (kmwolf/centos-qemu-kvm) So far the assumption has always been that if we try to inactivate a node, it is already idle. This doesn't hold true any more if we allow inactivating exported nodes because we can't know when new external requests come in. Drain the node around setting BDRV_O_INACTIVE so that requests can't start operating on an active node and then in the middle it suddenly becomes inactive. With this change, it's enough for exports to check for new requests that they operate on an active node (or, like reads, are allowed even on an inactive node). Signed-off-by: Kevin Wolf Acked-by: Fabiano Rosas Message-ID: <20250204211407.381505-12-kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi Reviewed-by: Eric Blake Signed-off-by: Kevin Wolf (cherry picked from commit 2849092a0024405e74c96f0a5ec41bb182ec8538) Signed-off-by: Kevin Wolf --- block.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/block.c b/block.c index 2140a5d3b7..38cb8481a8 100644 --- a/block.c +++ b/block.c @@ -7032,7 +7032,9 @@ bdrv_inactivate_recurse(BlockDriverState *bs, bool top_level) return -EPERM; } + bdrv_drained_begin(bs); bs->open_flags |= BDRV_O_INACTIVE; + bdrv_drained_end(bs); /* * Update permissions, they may differ for inactive nodes. -- 2.39.3