Linux-XFS Archive on lore.kernel.org
 help / color / Atom feed
From: Gao Xiang <hsiangkao@redhat.com>
To: linux-xfs@vger.kernel.org
Cc: Dave Chinner <david@fromorbit.com>,
	"Darrick J. Wong" <darrick.wong@oracle.com>,
	Brian Foster <bfoster@redhat.com>,
	Gao Xiang <hsiangkao@redhat.com>
Subject: [RFC PATCH 1/2] xfs: arrange all unlinked inodes into one list
Date: Tue,  7 Jul 2020 21:57:40 +0800
Message-ID: <20200707135741.487-2-hsiangkao@redhat.com> (raw)
In-Reply-To: <20200707135741.487-1-hsiangkao@redhat.com>

There is no need to keep old multiple short unlink inode buckets
since we have an in-memory double linked list for all unlinked
inodes.

Apart from the perspective of the necessity, the main advantage
is that the log and AGI update can be reduced since each AG has
the only one head now, which is implemented in the following patch.

Therefore, this patch applies the new way in xfs_iunlink() and
keep the old approach in xfs_iunlink_remove_inode() path as well
so inode eviction can still work properly in recovery.

Signed-off-by: Gao Xiang <hsiangkao@redhat.com>
---
 fs/xfs/xfs_inode.c | 40 ++++++++++++++++++++--------------------
 1 file changed, 20 insertions(+), 20 deletions(-)

diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index ab288424764c..10565fa5ace4 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -33,6 +33,7 @@
 #include "xfs_symlink.h"
 #include "xfs_trans_priv.h"
 #include "xfs_log.h"
+#include "xfs_log_priv.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_reflink.h"
 #include "xfs_iunlink_item.h"
@@ -1955,25 +1956,32 @@ xfs_iunlink_update_bucket(
 	struct xfs_trans	*tp,
 	xfs_agnumber_t		agno,
 	struct xfs_buf		*agibp,
-	unsigned int		bucket_index,
+	xfs_agino_t		old_agino,
 	xfs_agino_t		new_agino)
 {
+	struct xlog		*log = tp->t_mountp->m_log;
 	struct xfs_agi		*agi = agibp->b_addr;
 	xfs_agino_t		old_value;
-	int			offset;
+	unsigned int		bucket_index;
+	int                     offset;
 
 	ASSERT(xfs_verify_agino_or_null(tp->t_mountp, agno, new_agino));
 
+	bucket_index = 0;
+	/* During recovery, the old multiple bucket index can be applied */
+	if (!log || log->l_flags & XLOG_RECOVERY_NEEDED) {
+		ASSERT(old_agino != NULLAGINO);
+
+		if (be32_to_cpu(agi->agi_unlinked[0]) != old_agino)
+			bucket_index = old_agino % XFS_AGI_UNLINKED_BUCKETS;
+	}
+
 	old_value = be32_to_cpu(agi->agi_unlinked[bucket_index]);
 	trace_xfs_iunlink_update_bucket(tp->t_mountp, agno, bucket_index,
 			old_value, new_agino);
 
-	/*
-	 * We should never find the head of the list already set to the value
-	 * passed in because either we're adding or removing ourselves from the
-	 * head of the list.
-	 */
-	if (old_value == new_agino) {
+	/* check if the old agi_unlinked head is as expected */
+	if (old_value != old_agino) {
 		xfs_buf_mark_corrupt(agibp);
 		return -EFSCORRUPTED;
 	}
@@ -2001,14 +2009,13 @@ xfs_iunlink_insert_inode(
 	xfs_agino_t		agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
 	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
 	xfs_agino_t		next_agino;
-	short			bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
 
 	/*
 	 * Get the index into the agi hash table for the list this inode will
 	 * go on.  Make sure the pointer isn't garbage and that this inode
 	 * isn't already on the list.
 	 */
-	next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]);
+	next_agino = be32_to_cpu(agi->agi_unlinked[0]);
 	if (next_agino == agino ||
 	    !xfs_verify_agino_or_null(mp, agno, next_agino)) {
 		xfs_buf_mark_corrupt(agibp);
@@ -2036,7 +2043,7 @@ xfs_iunlink_insert_inode(
 	}
 
 	/* Point the head of the list to point to this inode. */
-	return xfs_iunlink_update_bucket(tp, agno, agibp, bucket_index, agino);
+	return xfs_iunlink_update_bucket(tp, agno, agibp, next_agino, agino);
 }
 
 /*
@@ -2051,27 +2058,20 @@ xfs_iunlink_remove_inode(
 	struct xfs_inode	*ip)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
-	struct xfs_agi		*agi = agibp->b_addr;
 	xfs_agino_t		agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
 	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
 	xfs_agino_t		next_agino = ip->i_next_unlinked;
-	short			bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
 	int			error;
 
 	if (ip->i_prev_unlinked == NULLAGINO) {
 		/* remove from head of list */
-		if (be32_to_cpu(agi->agi_unlinked[bucket_index]) != agino) {
-			xfs_buf_mark_corrupt(agibp);
-			return -EFSCORRUPTED;
-		}
 		if (next_agino == agino ||
 		    !xfs_verify_agino_or_null(mp, agno, next_agino))
 			return -EFSCORRUPTED;
 
-		error = xfs_iunlink_update_bucket(tp, agno, agibp,
-					bucket_index, next_agino);
+		error = xfs_iunlink_update_bucket(tp, agno, agibp, agino, next_agino);
 		if (error)
-			return -EFSCORRUPTED;
+			return error;
 	} else {
 		/* lookup previous inode and update to point at next */
 		struct xfs_inode	*pip;
-- 
2.18.1


  reply index

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-07 13:57 [RFC PATCH 0/2] xfs: more unlinked inode list optimization v1 Gao Xiang
2020-07-07 13:57 ` Gao Xiang [this message]
2020-07-08 22:33   ` [RFC PATCH 1/2] xfs: arrange all unlinked inodes into one list Dave Chinner
2020-07-09  0:17     ` Gao Xiang
2020-07-07 13:57 ` [RFC PATCH 2/2] xfs: don't access AGI on unlinked inodes if it can Gao Xiang
2020-07-08 17:03   ` Darrick J. Wong
2020-07-08 23:40     ` Gao Xiang
2020-07-08 23:33   ` Dave Chinner
2020-07-09  0:55     ` Gao Xiang
2020-07-09  2:32       ` Dave Chinner
2020-07-09 10:36         ` Gao Xiang
2020-07-09 10:47           ` Gao Xiang
2020-07-09 22:36           ` Dave Chinner
2020-07-24  6:12 ` [RFC PATCH v2 0/3] xfs: more unlinked inode list optimization v2 Gao Xiang
2020-07-24  6:12   ` [RFC PATCH v2 1/3] xfs: arrange all unlinked inodes into one list Gao Xiang
2020-07-24  6:12   ` [RFC PATCH v2 2/3] xfs: introduce perag iunlink lock Gao Xiang
2020-07-24  6:12   ` [RFC PATCH v2 3/3] xfs: insert unlinked inodes from tail Gao Xiang
2020-08-18 13:30   ` [RFC PATCH v4 0/3] xfs: more unlinked inode list optimization v4 Gao Xiang
2020-08-18 13:30     ` [RFC PATCH v4 1/3] xfs: get rid of unused pagi_unlinked_hash Gao Xiang
2020-08-19  0:54       ` Darrick J. Wong
2020-08-21  1:09         ` Dave Chinner
2020-08-18 13:30     ` [RFC PATCH v4 2/3] xfs: introduce perag iunlink lock Gao Xiang
2020-08-19  1:06       ` Darrick J. Wong
2020-08-19  1:23         ` Gao Xiang
2020-08-18 13:30     ` [RFC PATCH v4 3/3] xfs: insert unlinked inodes from tail Gao Xiang
2020-08-19  0:53     ` [RFC PATCH v4 0/3] xfs: more unlinked inode list optimization v4 Darrick J. Wong
2020-08-19  1:14       ` Gao Xiang
2020-08-20  2:46     ` Darrick J. Wong
2020-08-20  4:01       ` Gao Xiang

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=20200707135741.487-2-hsiangkao@redhat.com \
    --to=hsiangkao@redhat.com \
    --cc=bfoster@redhat.com \
    --cc=darrick.wong@oracle.com \
    --cc=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

Linux-XFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-xfs/0 linux-xfs/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-xfs linux-xfs/ https://lore.kernel.org/linux-xfs \
		linux-xfs@vger.kernel.org
	public-inbox-index linux-xfs

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-xfs


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git