From: Hanna Czenczek <hre...@redhat.com> Since AioContext locks have been removed, a BlockBackend's AioContext may really change at any time (only exception is that it is often confined to a drained section, as noted in this patch). Therefore, blk_get_aio_context() cannot rely on its root node's context always matching that of the BlockBackend.
In practice, whether they match does not matter anymore anyway: Requests can be sent to BDSs from any context, so anyone who requests the BB's context should have no reason to require the root node to have the same context. Therefore, we can and should remove the assertion to that effect. In addition, because the context can be set and queried from different threads concurrently, it has to be accessed with atomic operations. Buglink: https://issues.redhat.com/browse/RHEL-19381 Suggested-by: Kevin Wolf <kw...@redhat.com> Signed-off-by: Hanna Czenczek <hre...@redhat.com> Message-ID: <20240202144755.671354-2-hre...@redhat.com> Reviewed-by: Stefan Hajnoczi <stefa...@redhat.com> Reviewed-by: Kevin Wolf <kw...@redhat.com> Signed-off-by: Kevin Wolf <kw...@redhat.com> --- block/block-backend.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/block/block-backend.c b/block/block-backend.c index 209eb07528..9c4de79e6b 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -44,7 +44,7 @@ struct BlockBackend { char *name; int refcnt; BdrvChild *root; - AioContext *ctx; + AioContext *ctx; /* access with atomic operations only */ DriveInfo *legacy_dinfo; /* null unless created by drive_new() */ QTAILQ_ENTRY(BlockBackend) link; /* for block_backends */ QTAILQ_ENTRY(BlockBackend) monitor_link; /* for monitor_block_backends */ @@ -2414,22 +2414,22 @@ void blk_op_unblock_all(BlockBackend *blk, Error *reason) } } +/** + * Return BB's current AioContext. Note that this context may change + * concurrently at any time, with one exception: If the BB has a root node + * attached, its context will only change through bdrv_try_change_aio_context(), + * which creates a drained section. Therefore, incrementing such a BB's + * in-flight counter will prevent its context from changing. + */ AioContext *blk_get_aio_context(BlockBackend *blk) { - BlockDriverState *bs; IO_CODE(); if (!blk) { return qemu_get_aio_context(); } - bs = blk_bs(blk); - if (bs) { - AioContext *ctx = bdrv_get_aio_context(blk_bs(blk)); - assert(ctx == blk->ctx); - } - - return blk->ctx; + return qatomic_read(&blk->ctx); } int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, @@ -2442,7 +2442,7 @@ int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, GLOBAL_STATE_CODE(); if (!bs) { - blk->ctx = new_context; + qatomic_set(&blk->ctx, new_context); return 0; } @@ -2471,7 +2471,7 @@ static void blk_root_set_aio_ctx_commit(void *opaque) AioContext *new_context = s->new_ctx; ThrottleGroupMember *tgm = &blk->public.throttle_group_member; - blk->ctx = new_context; + qatomic_set(&blk->ctx, new_context); if (tgm->throttle_state) { throttle_group_detach_aio_context(tgm); throttle_group_attach_aio_context(tgm, new_context); -- 2.43.0
