Skip to contentSkip to conversation

xfs: use xfs_ialloc_blks_per_cluster in xfs_bulkstat

PATCH Superseded Jeff Liu 2 participants 3 messages Dec 11, 2013 – Dec 12, 2013 XFS filesystem

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Jeff Liu

author
#0 lore raw

PATCH 2/5 xfs: use xfs_ialloc_blks_per_cluster in xfs_bulkstat

From: Jie Liu <[email protected]>

Use xfs_ialloc_blks_per_cluster() in xfs_bulkstat(), make the related variables more meaningful and remove an unused variable nimask from it.

---
 fs/xfs/xfs_itable.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)
fs/xfs/xfs_itable.c +8 -12
index c237ad1..6671f5b 100644
@@ -209,9 +209,8 @@ xfs_bulkstat(
 	xfs_inobt_rec_incore_t	*irbuf;	/* start of irec buffer */
 	xfs_inobt_rec_incore_t	*irbufend; /* end of good irec buffer entries */
 	xfs_ino_t		lastino; /* last inode number returned */
-	int			nbcluster; /* # of blocks in a cluster */
-	int			nicluster; /* # of inodes in a cluster */
-	int			nimask;	/* mask for inode clusters */
+	int			blks_per_cluster; /* # of blocks in a cluster */
+	int			ninodes; /* # of inodes in a cluster */
 	int			nirbuf;	/* size of irbuf */
 	int			rval;	/* return value error code */
 	int			tmp;	/* result value from btree calls */
@@ -243,11 +242,8 @@ xfs_bulkstat(
 	*done = 0;
 	fmterror = 0;
 	ubufp = ubuffer;
-	nicluster = mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp) ?
-		mp->m_sb.sb_inopblock :
-		(XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog);
-	nimask = ~(nicluster - 1);
-	nbcluster = nicluster >> mp->m_sb.sb_inopblog;
+	blks_per_cluster = xfs_ialloc_blks_per_cluster(mp);
+	ninodes = blks_per_cluster << mp->m_sb.sb_inopblog;
 	irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4);
 	if (!irbuf)
 		return ENOMEM;
@@ -390,12 +386,12 @@ xfs_bulkstat(
 				agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino);
 				for (chunkidx = 0;
 				     chunkidx < XFS_INODES_PER_CHUNK;
-				     chunkidx += nicluster,
-				     agbno += nbcluster) {
-					if (xfs_inobt_maskn(chunkidx, nicluster)
+				     chunkidx += ninodes,
+				     agbno += blks_per_cluster) {
+					if (xfs_inobt_maskn(chunkidx, ninodes)
 							& ~r.ir_free)
 						xfs_btree_reada_bufs(mp, agno,
-							agbno, nbcluster,
+							agbno, blks_per_cluster,
 							&xfs_inode_buf_ops);
 				}
 				blk_finish_plug(&plug);
Signature · 6 lines
1.8.3.2

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
#1 lore raw
Quoted 16 lines · Jeff Liu

From: Jie Liu <[email protected]>

Use xfs_ialloc_blks_per_cluster() in xfs_bulkstat(), make the related variables more meaningful and remove an unused variable nimask from it.

---
 fs/xfs/xfs_itable.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)
fs/xfs/xfs_itable.c +0 -0
index c237ad1..6671f5b 100644

fs/xfs/xfs_itable.c xfs_bulkstat( · line 213

Quoted 8 lines
fs/xfs/xfs_itable.c +2 -3
@@ -209,9 +209,8 @@ xfs_bulkstat(
 	xfs_inobt_rec_incore_t	*irbuf;	/* start of irec buffer */
 	xfs_inobt_rec_incore_t	*irbufend; /* end of good irec buffer entries */
 	xfs_ino_t		lastino; /* last inode number returned */
-	int			nbcluster; /* # of blocks in a cluster */
-	int			nicluster; /* # of inodes in a cluster */
-	int			nimask;	/* mask for inode clusters */
+	int			blks_per_cluster; /* # of blocks in a cluster */
+	int			ninodes; /* # of inodes in a cluster */

inodes_per_cluster?

fs/xfs/xfs_itable.c xfs_bulkstat( · line 245

Quoted 13 lines
fs/xfs/xfs_itable.c +1 -5
 	int			nirbuf;	/* size of irbuf */
 	int			rval;	/* return value error code */
 	int			tmp;	/* result value from btree calls */
@@ -243,11 +242,8 @@ xfs_bulkstat(
 	*done = 0;
 	fmterror = 0;
 	ubufp = ubuffer;
-	nicluster = mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp) ?
-		mp->m_sb.sb_inopblock :
-		(XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog);
-	nimask = ~(nicluster - 1);
-	nbcluster = nicluster >> mp->m_sb.sb_inopblog;
+	blks_per_cluster = xfs_ialloc_blks_per_cluster(mp);

Ok, now I see the helper name, perhaps it shoul dbe more like all the other conversion functions. "ialloc" means nothing in the context, and "blks" doesn't tell us the unit we are converting to
- if could be basic blocks, it could be filesystem blocks. So, perhaps something like xfs_icluster_size_fsb()?

Otherwise this patch looks ok.

Cheers,

Dave.

Signature · 7 lines
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
author
#2 lore raw
Quoted 12 lines · Dave Chinner
Quoted 11 lines · Jeff Liu

From: Jie Liu <[email protected]>

Use xfs_ialloc_blks_per_cluster() in xfs_bulkstat(), make the related variables more meaningful and remove an unused variable nimask from it.

---
 fs/xfs/xfs_itable.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)
Quoted 15 lines

fs/xfs/xfs_itable.c xfs_bulkstat( · line 213

Quoted 13 lines
fs/xfs/xfs_itable.c +2 -3
index c237ad1..6671f5b 100644
@@ -209,9 +209,8 @@ xfs_bulkstat(
 	xfs_inobt_rec_incore_t	*irbuf;	/* start of irec buffer */
 	xfs_inobt_rec_incore_t	*irbufend; /* end of good irec buffer entries */
 	xfs_ino_t		lastino; /* last inode number returned */
-	int			nbcluster; /* # of blocks in a cluster */
-	int			nicluster; /* # of inodes in a cluster */
-	int			nimask;	/* mask for inode clusters */
+	int			blks_per_cluster; /* # of blocks in a cluster */
+	int			ninodes; /* # of inodes in a cluster */

inodes_per_cluster?

Ok.

Quoted 12 lines
Quoted 12 lines
fs/xfs/xfs_itable.c +0 -4
 	int			nirbuf;	/* size of irbuf */
 	int			rval;	/* return value error code */
 	int			tmp;	/* result value from btree calls */
@@ -243,11 +242,8 @@ xfs_bulkstat(
 	*done = 0;
 	fmterror = 0;
 	ubufp = ubuffer;
-	nicluster = mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp) ?
-		mp->m_sb.sb_inopblock :
-		(XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog);
-	nimask = ~(nicluster - 1);
Quoted 8 lines

fs/xfs/xfs_itable.c xfs_bulkstat( · line 245

Quoted 2 lines
fs/xfs/xfs_itable.c +1 -1
-	nbcluster = nicluster >> mp->m_sb.sb_inopblog;
+	blks_per_cluster = xfs_ialloc_blks_per_cluster(mp);

Ok, now I see the helper name, perhaps it shoul dbe more like all the other conversion functions. "ialloc" means nothing in the context, and "blks" doesn't tell us the unit we are converting to
- if could be basic blocks, it could be filesystem blocks. So, perhaps something like xfs_icluster_size_fsb()?

Sounds make sense to me, I'd like to adopt this name.

Thanks,
-Jeff

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs