All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Yan, Zheng" <ukernel@gmail.com>
To: Jeff Layton <jlayton@kernel.org>
Cc: ceph-devel <ceph-devel@vger.kernel.org>,
	Ilya Dryomov <idryomov@gmail.com>
Subject: Re: [PATCH] ceph: wake waiters on any IMPORT that grants new caps
Date: Sat, 29 Jan 2022 11:14:53 +0800	[thread overview]
Message-ID: <CAAM7YAmcj4JQ64EHWRTAVnEGnhfSN1OSUCSuOoi2PhOT8s_cHg@mail.gmail.com> (raw)
In-Reply-To: <20220127200849.96580-1-jlayton@kernel.org>

On Sat, Jan 29, 2022 at 2:32 AM Jeff Layton <jlayton@kernel.org> wrote:
>
> I've noticed an intermittent hang waiting for caps in some testing. What
> I see is that the client will try to get caps for an operation (e.g. a
> read), and ends up waiting on the waitqueue forever. The caps debugfs
> file however shows that the caps it's waiting on have already been
> granted.
>
> The current grant handling code will wake the waitqueue when it sees
> that there are newly-granted caps in the issued set. On an import
> however, we'll end up adding a new cap first, which fools the logic into
> thinking that nothing has changed. A later hack in the code works around
> this, but only for auth caps.

not right. handle_cap_import() saves old issued to extra_info->issued.

>
> Ensure we wake the waiters whenever we get an IMPORT that grants new
> caps for the inode.
>
> URL: https://tracker.ceph.com/issues/54044
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>  fs/ceph/caps.c | 23 ++++++++++++-----------
>  1 file changed, 12 insertions(+), 11 deletions(-)
>
> I'm still testing this patch, but I think this may be the cause of some
> mysterious hangs I've hit in testing.
>
> diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
> index e668cdb9c99e..06b65a68e920 100644
> --- a/fs/ceph/caps.c
> +++ b/fs/ceph/caps.c
> @@ -3541,21 +3541,22 @@ static void handle_cap_grant(struct inode *inode,
>                         fill_inline = true;
>         }
>
> -       if (ci->i_auth_cap == cap &&
> -           le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) {
> +       if (le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) {
>                 if (newcaps & ~extra_info->issued)
>                         wake = true;
>
> -               if (ci->i_requested_max_size > max_size ||
> -                   !(le32_to_cpu(grant->wanted) & CEPH_CAP_ANY_FILE_WR)) {
> -                       /* re-request max_size if necessary */
> -                       ci->i_requested_max_size = 0;
> -                       wake = true;
> -               }
> +               if (ci->i_auth_cap == cap) {
> +                       if (ci->i_requested_max_size > max_size ||
> +                           !(le32_to_cpu(grant->wanted) & CEPH_CAP_ANY_FILE_WR)) {
> +                               /* re-request max_size if necessary */
> +                               ci->i_requested_max_size = 0;
> +                               wake = true;
> +                       }
>
> -               ceph_kick_flushing_inode_caps(session, ci);
> -               spin_unlock(&ci->i_ceph_lock);
> -               up_read(&session->s_mdsc->snap_rwsem);
> +                       ceph_kick_flushing_inode_caps(session, ci);
> +                       spin_unlock(&ci->i_ceph_lock);
> +                       up_read(&session->s_mdsc->snap_rwsem);
> +               }
>         } else {
>                 spin_unlock(&ci->i_ceph_lock);
>         }
> --
> 2.34.1
>

  reply	other threads:[~2022-01-29  3:15 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-27 20:08 [PATCH] ceph: wake waiters on any IMPORT that grants new caps Jeff Layton
2022-01-29  3:14 ` Yan, Zheng [this message]
2022-01-29 10:36   ` Jeff Layton
     [not found]     ` <CAAM7YAn+8H5nv_Y3hBDtEGA5=jbtsG8-RX=dGU6wB3RRH-HGmw@mail.gmail.com>
2022-02-02 12:40       ` Jeff Layton

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=CAAM7YAmcj4JQ64EHWRTAVnEGnhfSN1OSUCSuOoi2PhOT8s_cHg@mail.gmail.com \
    --to=ukernel@gmail.com \
    --cc=ceph-devel@vger.kernel.org \
    --cc=idryomov@gmail.com \
    --cc=jlayton@kernel.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 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.