All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ilya Dryomov <idryomov@gmail.com>
To: Jeff Layton <jlayton@kernel.org>
Cc: Ceph Development <ceph-devel@vger.kernel.org>
Subject: Re: [PATCH 2/3] ceph: clean up locking annotation for ceph_get_snap_realm and __lookup_snap_realm
Date: Mon, 21 Jun 2021 22:48:39 +0200	[thread overview]
Message-ID: <CAOi1vP83Za9rW3wK-XvOW8k=UXObczyQeQgTmrQRRXJ0yOmXsw@mail.gmail.com> (raw)
In-Reply-To: <20210603165231.110559-3-jlayton@kernel.org>

On Thu, Jun 3, 2021 at 6:52 PM Jeff Layton <jlayton@kernel.org> wrote:
>
> They both say that the snap_rwsem must be held for write, but I don't
> see any real reason for it, and it's not currently always called that
> way.
>
> The lookup is just walking the rbtree, so holding it for read should be
> fine there. The "get" is bumping the refcount and (possibly) removing
> it from the empty list. I see no need to hold the snap_rwsem for write
> for that.
>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>  fs/ceph/snap.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c
> index bc6c33d485e6..f8cac2abab3f 100644
> --- a/fs/ceph/snap.c
> +++ b/fs/ceph/snap.c
> @@ -60,12 +60,12 @@
>  /*
>   * increase ref count for the realm
>   *
> - * caller must hold snap_rwsem for write.
> + * caller must hold snap_rwsem.
>   */
>  void ceph_get_snap_realm(struct ceph_mds_client *mdsc,
>                          struct ceph_snap_realm *realm)
>  {
> -       lockdep_assert_held_write(&mdsc->snap_rwsem);
> +       lockdep_assert_held(&mdsc->snap_rwsem);
>
>         dout("get_realm %p %d -> %d\n", realm,
>              atomic_read(&realm->nref), atomic_read(&realm->nref)+1);
> @@ -139,7 +139,7 @@ static struct ceph_snap_realm *ceph_create_snap_realm(
>  /*
>   * lookup the realm rooted at @ino.
>   *
> - * caller must hold snap_rwsem for write.
> + * caller must hold snap_rwsem.
>   */
>  static struct ceph_snap_realm *__lookup_snap_realm(struct ceph_mds_client *mdsc,
>                                                    u64 ino)
> @@ -147,7 +147,7 @@ static struct ceph_snap_realm *__lookup_snap_realm(struct ceph_mds_client *mdsc,
>         struct rb_node *n = mdsc->snap_realms.rb_node;
>         struct ceph_snap_realm *r;
>
> -       lockdep_assert_held_write(&mdsc->snap_rwsem);
> +       lockdep_assert_held(&mdsc->snap_rwsem);
>
>         while (n) {
>                 r = rb_entry(n, struct ceph_snap_realm, node);
> --
> 2.31.1
>

Ah, since you are relaxing the requirement some of those lockdep
asserts from the previous patch aren't actually redundant.  This seems
fine to me: lookup definitely doesn't need the write lock and allowing
concurrent gets should be OK.  The write lock made some sense back when
get was an atomic_read followed by atomic_inc but not now.

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

Thanks,

                Ilya

  reply	other threads:[~2021-06-21 20:48 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-03 16:52 [PATCH 0/3] ceph: locking fixes for snaprealm handling Jeff Layton
2021-06-03 16:52 ` [PATCH 1/3] ceph: add some lockdep assertions around " Jeff Layton
2021-06-21 20:05   ` Ilya Dryomov
2021-06-21 20:51     ` Jeff Layton
2021-06-21 21:27       ` Ilya Dryomov
2021-06-03 16:52 ` [PATCH 2/3] ceph: clean up locking annotation for ceph_get_snap_realm and __lookup_snap_realm Jeff Layton
2021-06-21 20:48   ` Ilya Dryomov [this message]
2021-06-03 16:52 ` [PATCH 3/3] ceph: must hold snap_rwsem when filling inode for async create Jeff Layton
2021-06-21 19:45   ` Ilya Dryomov

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='CAOi1vP83Za9rW3wK-XvOW8k=UXObczyQeQgTmrQRRXJ0yOmXsw@mail.gmail.com' \
    --to=idryomov@gmail.com \
    --cc=ceph-devel@vger.kernel.org \
    --cc=jlayton@kernel.org \
    --subject='Re: [PATCH 2/3] ceph: clean up locking annotation for ceph_get_snap_realm and __lookup_snap_realm' \
    /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

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.