qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Kevin Wolf <kwolf@redhat.com>
To: qemu-block@nongnu.org
Cc: kwolf@redhat.com, stefanha@redhat.com, hreitz@redhat.com,
	eblake@redhat.com, qemu-devel@nongnu.org
Subject: [PATCH v2 07/11] block: Fix AioContext locking in bdrv_open_inherit()
Date: Mon,  5 Jun 2023 10:57:07 +0200	[thread overview]
Message-ID: <20230605085711.21261-8-kwolf@redhat.com> (raw)
In-Reply-To: <20230605085711.21261-1-kwolf@redhat.com>

bdrv_open_inherit() calls several functions for which it needs to hold
the AioContext lock, but currently doesn't. This includes calls in
bdrv_append_temp_snapshot(), for which bdrv_open_inherit() is the only
caller. Fix the locking in these places.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block.c | 25 ++++++++++++++++++++++++-
 1 file changed, 24 insertions(+), 1 deletion(-)

diff --git a/block.c b/block.c
index a0793b6435..954b9923d9 100644
--- a/block.c
+++ b/block.c
@@ -3794,6 +3794,7 @@ static BlockDriverState *bdrv_append_temp_snapshot(BlockDriverState *bs,
     int64_t total_size;
     QemuOpts *opts = NULL;
     BlockDriverState *bs_snapshot = NULL;
+    AioContext *ctx = bdrv_get_aio_context(bs);
     int ret;
 
     GLOBAL_STATE_CODE();
@@ -3802,7 +3803,10 @@ static BlockDriverState *bdrv_append_temp_snapshot(BlockDriverState *bs,
        instead of opening 'filename' directly */
 
     /* Get the required size from the image */
+    aio_context_acquire(ctx);
     total_size = bdrv_getlength(bs);
+    aio_context_release(ctx);
+
     if (total_size < 0) {
         error_setg_errno(errp, -total_size, "Could not get image size");
         goto out;
@@ -3836,7 +3840,10 @@ static BlockDriverState *bdrv_append_temp_snapshot(BlockDriverState *bs,
         goto out;
     }
 
+    aio_context_acquire(ctx);
     ret = bdrv_append(bs_snapshot, bs, errp);
+    aio_context_release(ctx);
+
     if (ret < 0) {
         bs_snapshot = NULL;
         goto out;
@@ -3880,6 +3887,7 @@ bdrv_open_inherit(const char *filename, const char *reference, QDict *options,
     Error *local_err = NULL;
     QDict *snapshot_options = NULL;
     int snapshot_flags = 0;
+    AioContext *ctx = qemu_get_aio_context();
 
     assert(!child_class || !flags);
     assert(!child_class == !parent);
@@ -4017,9 +4025,13 @@ bdrv_open_inherit(const char *filename, const char *reference, QDict *options,
             /* Not requesting BLK_PERM_CONSISTENT_READ because we're only
              * looking at the header to guess the image format. This works even
              * in cases where a guest would not see a consistent state. */
-            file = blk_new(bdrv_get_aio_context(file_bs), 0, BLK_PERM_ALL);
+            ctx = bdrv_get_aio_context(file_bs);
+            aio_context_acquire(ctx);
+            file = blk_new(ctx, 0, BLK_PERM_ALL);
             blk_insert_bs(file, file_bs, &local_err);
             bdrv_unref(file_bs);
+            aio_context_release(ctx);
+
             if (local_err) {
                 goto fail;
             }
@@ -4065,8 +4077,13 @@ bdrv_open_inherit(const char *filename, const char *reference, QDict *options,
         goto fail;
     }
 
+    /* The AioContext could have changed during bdrv_open_common() */
+    ctx = bdrv_get_aio_context(bs);
+
     if (file) {
+        aio_context_acquire(ctx);
         blk_unref(file);
+        aio_context_release(ctx);
         file = NULL;
     }
 
@@ -4124,13 +4141,16 @@ bdrv_open_inherit(const char *filename, const char *reference, QDict *options,
          * (snapshot_bs); thus, we have to drop the strong reference to bs
          * (which we obtained by calling bdrv_new()). bs will not be deleted,
          * though, because the overlay still has a reference to it. */
+        aio_context_acquire(ctx);
         bdrv_unref(bs);
+        aio_context_release(ctx);
         bs = snapshot_bs;
     }
 
     return bs;
 
 fail:
+    aio_context_acquire(ctx);
     blk_unref(file);
     qobject_unref(snapshot_options);
     qobject_unref(bs->explicit_options);
@@ -4139,11 +4159,14 @@ fail:
     bs->options = NULL;
     bs->explicit_options = NULL;
     bdrv_unref(bs);
+    aio_context_release(ctx);
     error_propagate(errp, local_err);
     return NULL;
 
 close_and_fail:
+    aio_context_acquire(ctx);
     bdrv_unref(bs);
+    aio_context_release(ctx);
     qobject_unref(snapshot_options);
     qobject_unref(options);
     error_propagate(errp, local_err);
-- 
2.40.1



  parent reply	other threads:[~2023-06-05  8:58 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-06-05  8:57 [PATCH v2 00/11] block: Re-enable the graph lock Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 01/11] iotests: Test active commit with iothread and background I/O Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 02/11] qdev-properties-system: Lock AioContext for blk_insert_bs() Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 03/11] test-block-iothread: " Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 04/11] block: Fix AioContext locking in bdrv_open_child() Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 05/11] block: Fix AioContext locking in bdrv_attach_child_common() Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 06/11] block: Fix AioContext locking in bdrv_reopen_parse_file_or_backing() Kevin Wolf
2023-06-05  8:57 ` Kevin Wolf [this message]
2023-06-05  8:57 ` [PATCH v2 08/11] block: Fix AioContext locking in bdrv_open_backing_file() Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 09/11] blockjob: Fix AioContext locking in block_job_add_bdrv() Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 10/11] graph-lock: Unlock the AioContext while polling Kevin Wolf
2023-06-05  8:57 ` [PATCH v2 11/11] Revert "graph-lock: Disable locking for now" Kevin Wolf
2023-06-05  9:06   ` Paolo Bonzini
2023-06-05 10:03     ` Kevin Wolf
2023-06-07 18:19 ` [PATCH v2 00/11] block: Re-enable the graph lock 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=20230605085711.21261-8-kwolf@redhat.com \
    --to=kwolf@redhat.com \
    --cc=eblake@redhat.com \
    --cc=hreitz@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    /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).