From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Carlos Maiolino <cmaiolino@redhat.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH] xfs: Do not free xfs_extent_busy from inside a spinlock
Date: Tue, 23 Jul 2019 08:11:02 -0700 [thread overview]
Message-ID: <20190723151102.GA1561054@magnolia> (raw)
In-Reply-To: <20190723150017.31891-1-cmaiolino@redhat.com>
On Tue, Jul 23, 2019 at 05:00:17PM +0200, Carlos Maiolino wrote:
> xfs_extent_busy_clear_one() calls kmem_free() with the pag spinlock
> locked.
Er, what problem does this solve? Does holding on to the pag spinlock
too long while memory freeing causes everything else to stall? When is
memory freeing slow enough to cause a noticeable impact?
> Fix this by adding a new temporary list, and, make
> xfs_extent_busy_clear_one() to move the extent_busy items to this new
> list, instead of freeing them.
>
> Free the objects in the temporary list after we drop the pagb_lock
>
> Reported-by: Jeff Layton <jlayton@kernel.org>
> Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
> ---
> fs/xfs/xfs_extent_busy.c | 14 ++++++++++----
> 1 file changed, 10 insertions(+), 4 deletions(-)
>
> diff --git a/fs/xfs/xfs_extent_busy.c b/fs/xfs/xfs_extent_busy.c
> index 0ed68379e551..0a7dcf03340b 100644
> --- a/fs/xfs/xfs_extent_busy.c
> +++ b/fs/xfs/xfs_extent_busy.c
> @@ -523,7 +523,8 @@ STATIC void
> xfs_extent_busy_clear_one(
> struct xfs_mount *mp,
> struct xfs_perag *pag,
> - struct xfs_extent_busy *busyp)
> + struct xfs_extent_busy *busyp,
> + struct list_head *list)
> {
> if (busyp->length) {
> trace_xfs_extent_busy_clear(mp, busyp->agno, busyp->bno,
> @@ -531,8 +532,7 @@ xfs_extent_busy_clear_one(
> rb_erase(&busyp->rb_node, &pag->pagb_tree);
> }
>
> - list_del_init(&busyp->list);
> - kmem_free(busyp);
> + list_move(&busyp->list, list);
> }
>
> static void
> @@ -565,6 +565,7 @@ xfs_extent_busy_clear(
> struct xfs_perag *pag = NULL;
> xfs_agnumber_t agno = NULLAGNUMBER;
> bool wakeup = false;
> + LIST_HEAD(busy_list);
>
> list_for_each_entry_safe(busyp, n, list, list) {
> if (busyp->agno != agno) {
> @@ -580,13 +581,18 @@ xfs_extent_busy_clear(
> !(busyp->flags & XFS_EXTENT_BUSY_SKIP_DISCARD)) {
> busyp->flags = XFS_EXTENT_BUSY_DISCARDED;
> } else {
> - xfs_extent_busy_clear_one(mp, pag, busyp);
> + xfs_extent_busy_clear_one(mp, pag, busyp, &busy_list);
...and why not just put the busyp on the busy_list here so you don't
have to pass the list_head pointer around?
--D
> wakeup = true;
> }
> }
>
> if (pag)
> xfs_extent_busy_put_pag(pag, wakeup);
> +
> + list_for_each_entry_safe(busyp, n, &busy_list, list) {
> + list_del_init(&busyp->list);
> + kmem_free(busyp);
> + }
> }
>
> /*
> --
> 2.20.1
>
next prev parent reply other threads:[~2019-07-23 15:11 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-23 15:00 [PATCH] xfs: Do not free xfs_extent_busy from inside a spinlock Carlos Maiolino
2019-07-23 15:11 ` Darrick J. Wong [this message]
2019-07-23 15:31 ` Carlos Maiolino
2019-07-23 15:51 ` Christoph Hellwig
2019-07-23 17:07 ` Jeff Layton
2019-07-23 17:08 ` Christoph Hellwig
2019-07-23 17:38 ` Jeff Layton
2019-07-23 17:41 ` Christoph Hellwig
2019-07-23 15:13 ` Carlos Maiolino
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=20190723151102.GA1561054@magnolia \
--to=darrick.wong@oracle.com \
--cc=cmaiolino@redhat.com \
--cc=linux-xfs@vger.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox