From: Kevin Wolf <kwolf@redhat.com>
To: qemu-block@nongnu.org
Cc: kwolf@redhat.com, peter.maydell@linaro.org, qemu-devel@nongnu.org
Subject: [PULL 13/51] stream: Replace subtree drain with a single node drain
Date: Wed, 14 Dec 2022 14:44:15 +0100 [thread overview]
Message-ID: <20221214134453.31665-14-kwolf@redhat.com> (raw)
In-Reply-To: <20221214134453.31665-1-kwolf@redhat.com>
The subtree drain was introduced in commit b1e1af394d9 as a way to avoid
graph changes between finding the base node and changing the block graph
as necessary on completion of the image streaming job.
The block graph could change between these two points because
bdrv_set_backing_hd() first drains the parent node, which involved
polling and can do anything.
Subtree draining was an imperfect way to make this less likely (because
with it, fewer callbacks are called during this window). Everyone agreed
that it's not really the right solution, and it was only committed as a
stopgap solution.
This replaces the subtree drain with a solution that simply drains the
parent node before we try to find the base node, and then call a version
of bdrv_set_backing_hd() that doesn't drain, but just asserts that the
parent node is already drained.
This way, any graph changes caused by draining happen before we start
looking at the graph and things stay consistent between finding the base
node and changing the graph.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Hanna Reitz <hreitz@redhat.com>
Message-Id: <20221118174110.55183-10-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
include/block/block-global-state.h | 3 +++
block.c | 17 ++++++++++++++---
block/stream.c | 26 ++++++++++++++++----------
3 files changed, 33 insertions(+), 13 deletions(-)
diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h
index c7bd4a2088..00e0cf8aea 100644
--- a/include/block/block-global-state.h
+++ b/include/block/block-global-state.h
@@ -82,6 +82,9 @@ int bdrv_open_file_child(const char *filename,
BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp);
int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd,
Error **errp);
+int bdrv_set_backing_hd_drained(BlockDriverState *bs,
+ BlockDriverState *backing_hd,
+ Error **errp);
int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options,
const char *bdref_key, Error **errp);
BlockDriverState *bdrv_open(const char *filename, const char *reference,
diff --git a/block.c b/block.c
index db043346d8..97bfb1494f 100644
--- a/block.c
+++ b/block.c
@@ -3426,14 +3426,15 @@ static int bdrv_set_backing_noperm(BlockDriverState *bs,
return bdrv_set_file_or_backing_noperm(bs, backing_hd, true, tran, errp);
}
-int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd,
- Error **errp)
+int bdrv_set_backing_hd_drained(BlockDriverState *bs,
+ BlockDriverState *backing_hd,
+ Error **errp)
{
int ret;
Transaction *tran = tran_new();
GLOBAL_STATE_CODE();
- bdrv_drained_begin(bs);
+ assert(bs->quiesce_counter > 0);
ret = bdrv_set_backing_noperm(bs, backing_hd, tran, errp);
if (ret < 0) {
@@ -3443,7 +3444,17 @@ int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd,
ret = bdrv_refresh_perms(bs, tran, errp);
out:
tran_finalize(tran, ret);
+ return ret;
+}
+int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd,
+ Error **errp)
+{
+ int ret;
+ GLOBAL_STATE_CODE();
+
+ bdrv_drained_begin(bs);
+ ret = bdrv_set_backing_hd_drained(bs, backing_hd, errp);
bdrv_drained_end(bs);
return ret;
diff --git a/block/stream.c b/block/stream.c
index 694709bd25..8744ad103f 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -64,13 +64,16 @@ static int stream_prepare(Job *job)
bdrv_cor_filter_drop(s->cor_filter_bs);
s->cor_filter_bs = NULL;
- bdrv_subtree_drained_begin(s->above_base);
+ /*
+ * bdrv_set_backing_hd() requires that unfiltered_bs is drained. Drain
+ * already here and use bdrv_set_backing_hd_drained() instead because
+ * the polling during drained_begin() might change the graph, and if we do
+ * this only later, we may end up working with the wrong base node (or it
+ * might even have gone away by the time we want to use it).
+ */
+ bdrv_drained_begin(unfiltered_bs);
base = bdrv_filter_or_cow_bs(s->above_base);
- if (base) {
- bdrv_ref(base);
- }
-
unfiltered_base = bdrv_skip_filters(base);
if (bdrv_cow_child(unfiltered_bs)) {
@@ -82,7 +85,13 @@ static int stream_prepare(Job *job)
}
}
- bdrv_set_backing_hd(unfiltered_bs, base, &local_err);
+ bdrv_set_backing_hd_drained(unfiltered_bs, base, &local_err);
+
+ /*
+ * This call will do I/O, so the graph can change again from here on.
+ * We have already completed the graph change, so we are not in danger
+ * of operating on the wrong node any more if this happens.
+ */
ret = bdrv_change_backing_file(unfiltered_bs, base_id, base_fmt, false);
if (local_err) {
error_report_err(local_err);
@@ -92,10 +101,7 @@ static int stream_prepare(Job *job)
}
out:
- if (base) {
- bdrv_unref(base);
- }
- bdrv_subtree_drained_end(s->above_base);
+ bdrv_drained_end(unfiltered_bs);
return ret;
}
--
2.38.1
next prev parent reply other threads:[~2022-12-14 13:56 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-14 13:44 [PULL 00/51] Block layer patches Kevin Wolf
2022-12-14 13:44 ` [PULL 01/51] block: Inline bdrv_detach_child() Kevin Wolf
2022-12-14 13:44 ` [PULL 02/51] block: drop bdrv_remove_filter_or_cow_child Kevin Wolf
2022-12-14 13:44 ` [PULL 03/51] block: bdrv_refresh_perms(): allow external tran Kevin Wolf
2022-12-14 13:44 ` [PULL 04/51] block: refactor bdrv_list_refresh_perms to allow any list of nodes Kevin Wolf
2022-12-14 13:44 ` [PULL 05/51] qed: Don't yield in bdrv_qed_co_drain_begin() Kevin Wolf
2022-12-14 13:44 ` [PULL 06/51] test-bdrv-drain: Don't yield in .bdrv_co_drained_begin/end() Kevin Wolf
2022-12-14 13:44 ` [PULL 07/51] block: Revert .bdrv_drained_begin/end to non-coroutine_fn Kevin Wolf
2022-12-14 13:44 ` [PULL 08/51] block: Remove drained_end_counter Kevin Wolf
2022-12-14 13:44 ` [PULL 09/51] block: Inline bdrv_drain_invoke() Kevin Wolf
2022-12-14 13:44 ` [PULL 10/51] block: Fix locking for bdrv_reopen_queue_child() Kevin Wolf
2022-12-14 13:44 ` [PULL 11/51] block: Drain individual nodes during reopen Kevin Wolf
2022-12-14 13:44 ` [PULL 12/51] block: Don't use subtree drains in bdrv_drop_intermediate() Kevin Wolf
2022-12-14 13:44 ` Kevin Wolf [this message]
2022-12-14 13:44 ` [PULL 14/51] block: Remove subtree drains Kevin Wolf
2022-12-14 13:44 ` [PULL 15/51] block: Call drain callbacks only once Kevin Wolf
2022-12-14 13:44 ` [PULL 16/51] block: Remove ignore_bds_parents parameter from drain_begin/end Kevin Wolf
2022-12-14 13:44 ` [PULL 17/51] block: Drop out of coroutine in bdrv_do_drained_begin_quiesce() Kevin Wolf
2022-12-14 13:44 ` [PULL 18/51] block: Don't poll in bdrv_replace_child_noperm() Kevin Wolf
2022-12-14 13:44 ` [PULL 19/51] block: Remove poll parameter from bdrv_parent_drained_begin_single() Kevin Wolf
2022-12-14 13:44 ` [PULL 20/51] block-io: introduce coroutine_fn duplicates for bdrv_common_block_status_above callers Kevin Wolf
2022-12-14 13:44 ` [PULL 21/51] block-copy: add coroutine_fn annotations Kevin Wolf
2022-12-14 13:44 ` [PULL 22/51] nbd/server.c: " Kevin Wolf
2022-12-14 13:44 ` [PULL 23/51] block-backend: replace bdrv_*_above with blk_*_above Kevin Wolf
2022-12-14 13:44 ` [PULL 24/51] block/vmdk: add coroutine_fn annotations Kevin Wolf
2022-12-14 13:44 ` [PULL 25/51] block: avoid duplicating filename string in bdrv_create Kevin Wolf
2022-12-14 13:44 ` [PULL 26/51] block: distinguish between bdrv_create running in coroutine and not Kevin Wolf
2022-12-14 13:44 ` [PULL 27/51] block: bdrv_create_file is a coroutine_fn Kevin Wolf
2022-12-14 13:44 ` [PULL 28/51] block: rename generated_co_wrapper in co_wrapper_mixed Kevin Wolf
2022-12-14 13:44 ` [PULL 29/51] block-coroutine-wrapper.py: introduce co_wrapper Kevin Wolf
2022-12-14 13:44 ` [PULL 30/51] block-coroutine-wrapper.py: support functions without bs arg Kevin Wolf
2022-12-14 13:44 ` [PULL 31/51] block-coroutine-wrapper.py: support also basic return types Kevin Wolf
2022-12-14 13:44 ` [PULL 32/51] block: convert bdrv_create to co_wrapper Kevin Wolf
2022-12-14 13:44 ` [PULL 33/51] block/dirty-bitmap: convert coroutine-only functions " Kevin Wolf
2022-12-14 13:44 ` [PULL 34/51] block: Factor out bdrv_drain_all_begin_nopoll() Kevin Wolf
2022-12-14 13:44 ` [PULL 35/51] graph-lock: Introduce a lock to protect block graph operations Kevin Wolf
2022-12-14 13:44 ` [PULL 36/51] graph-lock: Implement guard macros Kevin Wolf
2022-12-14 13:44 ` [PULL 37/51] async: Register/unregister aiocontext in graph lock list Kevin Wolf
2022-12-14 13:44 ` [PULL 38/51] Import clang-tsa.h Kevin Wolf
2022-12-14 13:44 ` [PULL 39/51] clang-tsa: Add TSA_ASSERT() macro Kevin Wolf
2022-12-14 13:44 ` [PULL 40/51] clang-tsa: Add macros for shared locks Kevin Wolf
2022-12-14 13:44 ` [PULL 41/51] configure: Enable -Wthread-safety if present Kevin Wolf
2022-12-14 13:44 ` [PULL 42/51] test-bdrv-drain: Fix incorrrect drain assumptions Kevin Wolf
2022-12-14 13:44 ` [PULL 43/51] block: Fix locking in external_snapshot_prepare() Kevin Wolf
2022-12-14 13:44 ` [PULL 44/51] block: wrlock in bdrv_replace_child_noperm Kevin Wolf
2022-12-14 13:44 ` [PULL 45/51] block: remove unnecessary assert_bdrv_graph_writable() Kevin Wolf
2022-12-14 13:44 ` [PULL 46/51] block: assert that graph read and writes are performed correctly Kevin Wolf
2022-12-14 13:44 ` [PULL 47/51] graph-lock: TSA annotations for lock/unlock functions Kevin Wolf
2022-12-14 13:44 ` [PULL 48/51] Mark assert_bdrv_graph_readable/writable() GRAPH_RD/WRLOCK Kevin Wolf
2022-12-14 13:44 ` [PULL 49/51] block-coroutine-wrapper.py: introduce annotations that take the graph rdlock Kevin Wolf
2022-12-14 13:44 ` [PULL 50/51] block: use co_wrapper_mixed_bdrv_rdlock in functions taking the rdlock Kevin Wolf
2022-12-14 13:44 ` [PULL 51/51] block: GRAPH_RDLOCK for functions only called by co_wrappers Kevin Wolf
2022-12-14 22:35 ` [PULL 00/51] Block layer patches Peter Maydell
2022-12-15 9:44 ` Kevin Wolf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20221214134453.31665-14-kwolf@redhat.com \
--to=kwolf@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).