From: Kevin Wolf <kwolf@redhat.com>
To: qemu-block@nongnu.org
Cc: kwolf@redhat.com, qemu-devel@nongnu.org
Subject: [PULL 19/26] qcow2: Mark qcow2_inactivate() and callers GRAPH_RDLOCK
Date: Thu, 12 Oct 2023 18:22:17 +0200 [thread overview]
Message-ID: <20231012162224.240535-20-kwolf@redhat.com> (raw)
In-Reply-To: <20231012162224.240535-1-kwolf@redhat.com>
This adds GRAPH_RDLOCK annotations to declare that callers of
qcow2_inactivate() need to hold a reader lock for the graph because it
calls bdrv_get_device_or_node_name(), which accesses the parents list of
a node.
qcow2_do_close() is a bit strange because it is called from different
contexts. If close_data_file = true, we know that we were called from
non-coroutine main loop context (more specifically, we're coming from
qcow2_close()) and can safely drop the reader lock temporarily with
bdrv_graph_rdunlock_main_loop() and acquire the writer lock.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20230929145157.45443-16-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block/qcow2.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/block/qcow2.c b/block/qcow2.c
index 137c036ea4..aa01d9e7b5 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -2737,14 +2737,12 @@ fail_nometa:
return ret;
}
-static int qcow2_inactivate(BlockDriverState *bs)
+static int GRAPH_RDLOCK qcow2_inactivate(BlockDriverState *bs)
{
BDRVQcow2State *s = bs->opaque;
int ret, result = 0;
Error *local_err = NULL;
- assume_graph_lock(); /* FIXME */
-
qcow2_store_persistent_dirty_bitmaps(bs, true, &local_err);
if (local_err != NULL) {
result = -EINVAL;
@@ -2774,7 +2772,8 @@ static int qcow2_inactivate(BlockDriverState *bs)
return result;
}
-static void qcow2_do_close(BlockDriverState *bs, bool close_data_file)
+static void coroutine_mixed_fn GRAPH_RDLOCK
+qcow2_do_close(BlockDriverState *bs, bool close_data_file)
{
BDRVQcow2State *s = bs->opaque;
qemu_vfree(s->l1_table);
@@ -2801,18 +2800,24 @@ static void qcow2_do_close(BlockDriverState *bs, bool close_data_file)
g_free(s->image_backing_format);
if (close_data_file && has_data_file(bs)) {
+ GLOBAL_STATE_CODE();
+ bdrv_graph_rdunlock_main_loop();
bdrv_graph_wrlock(NULL);
bdrv_unref_child(bs, s->data_file);
bdrv_graph_wrunlock();
s->data_file = NULL;
+ bdrv_graph_rdlock_main_loop();
}
qcow2_refcount_close(bs);
qcow2_free_snapshots(bs);
}
-static void qcow2_close(BlockDriverState *bs)
+static void GRAPH_UNLOCKED qcow2_close(BlockDriverState *bs)
{
+ GLOBAL_STATE_CODE();
+ GRAPH_RDLOCK_GUARD_MAINLOOP();
+
qcow2_do_close(bs, true);
}
--
2.41.0
next prev parent reply other threads:[~2023-10-12 16:31 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-12 16:21 [PULL 00/26] Block layer patches Kevin Wolf
2023-10-12 16:21 ` [PULL 01/26] block: rename the bdrv_co_block_status static function Kevin Wolf
2023-10-12 16:22 ` [PULL 02/26] block: complete public block status API Kevin Wolf
2023-10-12 16:22 ` [PULL 03/26] block: switch to co_wrapper for bdrv_is_allocated_* Kevin Wolf
2023-10-12 16:22 ` [PULL 04/26] block: convert more bdrv_is_allocated* and bdrv_block_status* calls to coroutine versions Kevin Wolf
2023-10-12 16:22 ` [PULL 05/26] test-bdrv-drain: Don't call bdrv_graph_wrlock() in coroutine context Kevin Wolf
2023-10-12 16:22 ` [PULL 06/26] block-coroutine-wrapper: Add no_co_wrapper_bdrv_rdlock functions Kevin Wolf
2023-10-12 16:22 ` [PULL 07/26] block: Take graph rdlock in bdrv_inactivate_all() Kevin Wolf
2023-10-12 16:22 ` [PULL 08/26] block: Mark bdrv_first_blk() and bdrv_is_root_node() GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 09/26] block: Mark drain related functions GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 10/26] block: Mark bdrv_parent_cb_resize() and callers GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 11/26] block: Mark bdrv_snapshot_fallback() " Kevin Wolf
2023-10-12 16:22 ` [PULL 12/26] block: Take graph rdlock in parts of reopen Kevin Wolf
2023-10-12 16:22 ` [PULL 13/26] block: Mark bdrv_get_xdbg_block_graph() and callers GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 14/26] block: Mark bdrv_refresh_filename() " Kevin Wolf
2023-10-12 16:22 ` [PULL 15/26] block: Mark bdrv_primary_child() " Kevin Wolf
2023-10-12 16:22 ` [PULL 16/26] block: Mark bdrv_get_parent_name() " Kevin Wolf
2023-10-12 16:22 ` [PULL 17/26] block: Mark bdrv_amend_options() " Kevin Wolf
2023-10-12 16:22 ` [PULL 18/26] qcow2: Mark qcow2_signal_corruption() " Kevin Wolf
2023-10-12 16:22 ` Kevin Wolf [this message]
2023-10-12 16:22 ` [PULL 20/26] qcow2: Mark check_constraints_on_bitmap() GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 21/26] block: Mark bdrv_op_is_blocked() and callers GRAPH_RDLOCK Kevin Wolf
2023-10-12 16:22 ` [PULL 22/26] block: Mark bdrv_apply_auto_read_only() " Kevin Wolf
2023-10-12 16:22 ` [PULL 23/26] block: Mark bdrv_get_specific_info() " Kevin Wolf
2023-10-12 16:22 ` [PULL 24/26] block: Protect bs->parents with graph_lock Kevin Wolf
2023-10-12 16:22 ` [PULL 25/26] block: Protect bs->children " Kevin Wolf
2023-10-12 16:22 ` [PULL 26/26] block: Add assertion for bdrv_graph_wrlock() Kevin Wolf
2023-10-16 19:20 ` [PULL 00/26] Block layer patches Stefan Hajnoczi
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=20231012162224.240535-20-kwolf@redhat.com \
--to=kwolf@redhat.com \
--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).