qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Kevin Wolf <kwolf@redhat.com>
To: Fiona Ebner <f.ebner@proxmox.com>
Cc: qemu-block@nongnu.org, qemu-devel@nongnu.org, den@virtuozzo.com,
	andrey.drobyshev@virtuozzo.com, hreitz@redhat.com,
	stefanha@redhat.com, eblake@redhat.com, jsnow@redhat.com,
	vsementsov@yandex-team.ru
Subject: Re: [PATCH 05/11] block: move drain outside of read-locked bdrv_inactivate_recurse()
Date: Wed, 14 May 2025 19:06:02 +0200	[thread overview]
Message-ID: <aCTNek1ugvseHAQq@redhat.com> (raw)
In-Reply-To: <20250508140936.3344485-6-f.ebner@proxmox.com>

Am 08.05.2025 um 16:09 hat Fiona Ebner geschrieben:
> This is in preparation to mark bdrv_drained_begin() as GRAPH_UNLOCKED.
> 
> More granular draining is not trivially possible, because
> bdrv_inactivate_recurse() can recursively call itself.
> 
> Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
> ---
>  block.c | 19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
> 
> diff --git a/block.c b/block.c
> index c7c26533c9..10052027cd 100644
> --- a/block.c
> +++ b/block.c
> @@ -6991,6 +6991,8 @@ bdrv_inactivate_recurse(BlockDriverState *bs, bool top_level)
>      int ret;
>      uint64_t cumulative_perms, cumulative_shared_perms;
>  
> +    assert(qatomic_read(&bs->quiesce_counter) > 0);
> +
>      GLOBAL_STATE_CODE();
>  
>      if (!bs->drv) {
> @@ -7036,9 +7038,7 @@ 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.
> @@ -7063,14 +7063,20 @@ int bdrv_inactivate(BlockDriverState *bs, Error **errp)
>      int ret;
>  
>      GLOBAL_STATE_CODE();
> -    GRAPH_RDLOCK_GUARD_MAINLOOP();
> +
> +    bdrv_drain_all_begin();
> +    bdrv_graph_rdlock_main_loop();
>  
>      if (bdrv_has_bds_parent(bs, true)) {
>          error_setg(errp, "Node has active parent node");
> +        bdrv_graph_rdunlock_main_loop();
> +        bdrv_drain_all_end();
>          return -EPERM;
>      }
>  
>      ret = bdrv_inactivate_recurse(bs, true);
> +    bdrv_graph_rdunlock_main_loop();
> +    bdrv_drain_all_end();
>      if (ret < 0) {
>          error_setg_errno(errp, -ret, "Failed to inactivate node");
>          return ret;

This might be another one that would become a bit clearer if you put the
unlock/drain_end at the end and use a goto out for the error cases.

Kevin



  reply	other threads:[~2025-05-14 17:07 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-08 14:09 [RFC 00/11] do not drain while holding the graph lock Fiona Ebner
2025-05-08 14:09 ` [PATCH 01/11] block: remove outdated comments about AioContext locking Fiona Ebner
2025-05-14 16:22   ` Kevin Wolf
2025-05-08 14:09 ` [PATCH 02/11] block: move drain outside of read-locked bdrv_reopen_queue_child() Fiona Ebner
2025-05-14 16:36   ` Kevin Wolf
2025-05-15 11:48     ` Fiona Ebner
2025-05-15 12:28       ` Kevin Wolf
2025-05-15 13:41         ` Fiona Ebner
2025-05-08 14:09 ` [PATCH 03/11] block/snapshot: move drain outside of read-locked bdrv_snapshot_delete() Fiona Ebner
2025-05-14 16:52   ` Kevin Wolf
2025-05-15 12:03     ` Fiona Ebner
2025-05-08 14:09 ` [PATCH 04/11] block: drain while unlocked in bdrv_reopen_parse_file_or_backing() Fiona Ebner
2025-05-14 16:59   ` Kevin Wolf
2025-05-08 14:09 ` [PATCH 05/11] block: move drain outside of read-locked bdrv_inactivate_recurse() Fiona Ebner
2025-05-14 17:06   ` Kevin Wolf [this message]
2025-05-08 14:09 ` [PATCH 06/11] blockdev: drain while unlocked in internal_snapshot_action() Fiona Ebner
2025-05-14 17:26   ` Kevin Wolf
2025-05-19 12:37     ` Fiona Ebner
2025-05-08 14:09 ` [PATCH 07/11] blockdev: drain while unlocked in external_snapshot_action() Fiona Ebner
2025-05-08 14:09 ` [PATCH 08/11] block: mark bdrv_drained_begin() as GRAPH_UNLOCKED Fiona Ebner
2025-05-14 17:50   ` Kevin Wolf
2025-05-08 14:09 ` [PATCH 09/11] block: move drain out of bdrv_change_aio_context() Fiona Ebner
2025-05-14 19:45   ` Kevin Wolf
2025-05-08 14:09 ` [PATCH 10/11] block/graph-lock: add drain flag to bdrv_graph_wr{, un}lock Fiona Ebner
2025-05-14 19:54   ` [PATCH 10/11] block/graph-lock: add drain flag to bdrv_graph_wr{,un}lock Kevin Wolf
2025-05-19 12:10     ` Fiona Ebner
2025-05-20  6:09       ` Fiona Ebner
2025-05-20  8:42         ` Kevin Wolf
2025-05-08 14:09 ` [PATCH 11/11] iotests/graph-changes-while-io: add test case with removal of lower snapshot Fiona Ebner
2025-05-14 20:14   ` 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=aCTNek1ugvseHAQq@redhat.com \
    --to=kwolf@redhat.com \
    --cc=andrey.drobyshev@virtuozzo.com \
    --cc=den@virtuozzo.com \
    --cc=eblake@redhat.com \
    --cc=f.ebner@proxmox.com \
    --cc=hreitz@redhat.com \
    --cc=jsnow@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=vsementsov@yandex-team.ru \
    /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).