From: Dave Chinner <david@fromorbit.com>
To: linux-xfs@vger.kernel.org
Subject: [PATCH 1/4] xfs: make inode inactivation state changes atomic
Date: Tue, 19 Mar 2024 11:15:57 +1100 [thread overview]
Message-ID: <20240319001707.3430251-2-david@fromorbit.com> (raw)
In-Reply-To: <20240319001707.3430251-1-david@fromorbit.com>
From: Dave Chinner <dchinner@redhat.com>
We need the XFS_NEED_INACTIVE flag to correspond to whether the
inode is on the inodegc queues so that we can then use this state
for lazy removal.
To do this, move the addition of the inode to the inodegc queue
under the ip->i_flags_lock so that it is atomic w.r.t. setting
the XFS_NEED_INACTIVE flag.
Then, when we remove the inode from the inodegc list to actually run
inactivation, clear the XFS_NEED_INACTIVE at the same time we are
setting XFS_INACTIVATING to indicate that inactivation is in
progress.
These changes result in all the state changes and inodegc queuing
being atomic w.r.t. each other and inode lookups via the use of the
ip->i_flags lock.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
fs/xfs/xfs_icache.c | 16 ++++++++++++++--
fs/xfs/xfs_inode.h | 11 +++++++----
2 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index 6c87b90754c4..9a362964f656 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -1880,7 +1880,12 @@ xfs_inodegc_worker(
llist_for_each_entry_safe(ip, n, node, i_gclist) {
int error;
- xfs_iflags_set(ip, XFS_INACTIVATING);
+ /* Switch state to inactivating. */
+ spin_lock(&ip->i_flags_lock);
+ ip->i_flags |= XFS_INACTIVATING;
+ ip->i_flags &= ~XFS_NEED_INACTIVE;
+ spin_unlock(&ip->i_flags_lock);
+
error = xfs_inodegc_inactivate(ip);
if (error && !gc->error)
gc->error = error;
@@ -2075,9 +2080,14 @@ xfs_inodegc_queue(
unsigned long queue_delay = 1;
trace_xfs_inode_set_need_inactive(ip);
+
+ /*
+ * Put the addition of the inode to the gc list under the
+ * ip->i_flags_lock so that the state change and list addition are
+ * atomic w.r.t. lookup operations under the ip->i_flags_lock.
+ */
spin_lock(&ip->i_flags_lock);
ip->i_flags |= XFS_NEED_INACTIVE;
- spin_unlock(&ip->i_flags_lock);
cpu_nr = get_cpu();
gc = this_cpu_ptr(mp->m_inodegc);
@@ -2086,6 +2096,8 @@ xfs_inodegc_queue(
WRITE_ONCE(gc->items, items + 1);
shrinker_hits = READ_ONCE(gc->shrinker_hits);
+ spin_unlock(&ip->i_flags_lock);
+
/*
* Ensure the list add is always seen by anyone who finds the cpumask
* bit set. This effectively gives the cpumask bit set operation
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 94fa79ae1591..b0943d888f5c 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -349,10 +349,13 @@ static inline bool xfs_inode_has_forcealign(struct xfs_inode *ip)
/*
* If we need to update on-disk metadata before this IRECLAIMABLE inode can be
- * freed, then NEED_INACTIVE will be set. Once we start the updates, the
- * INACTIVATING bit will be set to keep iget away from this inode. After the
- * inactivation completes, both flags will be cleared and the inode is a
- * plain old IRECLAIMABLE inode.
+ * freed, then NEED_INACTIVE will be set. If the inode is accessed via iget
+ * whilst NEED_INACTIVE is set, the inode will be reactivated and become a
+ * normal inode again. Once we start the inactivation, the INACTIVATING bit will
+ * be set and the NEED_INACTIVE bit will be cleared. The INACTIVATING bit will
+ * keep iget away from this inode whilst inactivation is in progress. After the
+ * inactivation completes, INACTIVATING will be cleared and the inode
+ * transitions to a plain old IRECLAIMABLE inode.
*/
#define XFS_INACTIVATING (1 << 13)
--
2.43.0
next prev parent reply other threads:[~2024-03-19 0:17 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-19 0:15 [PATCH v2 0/4] xfs: recycle inactive inodes immediately Dave Chinner
2024-03-19 0:15 ` Dave Chinner [this message]
2024-03-19 18:01 ` [PATCH 1/4] xfs: make inode inactivation state changes atomic Darrick J. Wong
2024-03-19 0:15 ` [PATCH 2/4] xfs: prepare inode for i_gclist detection Dave Chinner
2024-03-19 0:15 ` [PATCH 3/4] xfs: allow lazy removal of inodes from the inodegc queues Dave Chinner
2024-03-19 0:16 ` [PATCH 4/4] xfs: reactivate XFS_NEED_INACTIVE inodes from xfs_iget Dave Chinner
2024-03-19 18:11 ` Darrick J. Wong
2024-03-20 8:39 ` Andre Noll
2024-03-20 14:53 ` Darrick J. Wong
2024-03-20 16:58 ` Andre Noll
2024-03-20 22:51 ` Dave Chinner
2024-03-21 9:59 ` Andre Noll
2024-03-22 1:09 ` Dave Chinner
2024-03-20 21:58 ` Dave Chinner
-- strict thread matches above, loose matches on Subject: below --
2024-02-01 0:30 [RFC] [PATCH 0/4] xfs: reactivate inodes immediately in xfs_iget Dave Chinner
2024-02-01 0:30 ` [PATCH 1/4] xfs: make inode inactivation state changes atomic Dave Chinner
2024-02-01 19:07 ` Darrick J. Wong
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=20240319001707.3430251-2-david@fromorbit.com \
--to=david@fromorbit.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).