[Cluster-devel] [PATCH 00/17] GFS2: Pre-pull patch posting (merge window)

[Cluster-devel] [PATCH 00/17] GFS2: Pre-pull patch posting (merge window)

[Bob Peterson]

Hi,

We've got a total of 17 GFS2 patches for this merge window. The
patches are basically in three categories: (1) patches related to
broken xfstest cases, (2) patches related to improving iomap and
start using it in GFS2, and (3) general typos and clarifications.

Please note that one of the iomap patches extends beyond GFS2 and
affects other file systems, but it was publically reviewed by a
variety of file system people in the community.

1. Andreas has a patch that simply renames variable 'bsize' to 'factor'
   to clarify the logic related to gfs2_block_map.
2. He also has a patch to correctly set ctime in the setflags ioctl,
   which fixes broken xfstests test 277.
3. He also fixed broken xfstest 258, due to an atime initialization
   problem.
4. He also fixed broken xfstest 307, in which GFS2 was not setting
   ctime when setting acls.
5. He has a patch to switch general iomap code from blkno to disk
   offset for a variety of file systems.
6. He has a patch to add a new IOMAP_F_DATA_INLINE flag for iomap
   to indicate blocks that have data mixed with metadata.
7. I contributed a patch to make inode height info part of the
   'metapath' data structure to facilitate using iomap in GFS2.
8. I have a patch to start using iomap inside GFS2 and switch GFS2's
   block_map functions to use iomap under the covers.
9. I have a patch to switch GFS2's fiemap implementation from using
   block_map to using iomap under the covers.
10. Andreas has a patch to implement SEEK_HOLE and SEEK_DATA via
    iomap in GFS2.
11. I have a patch related to journaled data pages not being properly
    synced to media when writing inodes. This was caught with xfstests.
12. I have a patch to fix another failing xfstest case in which
    switching a file from ordered_write to journaled data via set_flags
    caused a deadlock.
13. Andreas has a patch to fix failing xfstest case 066, which was
    due to not properly syncing dirty inodes when changing extended
    attributes.
14. Andreas fixed a minor typo in a comment.
15. Andreas contributed a patch to partially fix xfstest 424, which
    involved GET_FLAGS and SET_FLAGS ioctl. This is also a cleanup
    and simplification of the translation of flags from fs flags to
    gfs2 flags.
16. He also added support for STATX_ATTR_ in statx, which fixed broken
    xfstest 424.
17. He also contributed a fix for failing xfstest 093 which fixes a
    recursive glock problem with gfs2_xattr_get and _set.

Regards,

Bob Peterson
---
Andreas Gruenbacher (12):
  gfs2: Clarify gfs2_block_map
  gfs2: Update ctime in setflags ioctl
  gfs2: Support negative atimes
  gfs2: Always update inode ctime in set_acl
  iomap: Switch from blkno to disk offset
  iomap: Add IOMAP_F_DATA_INLINE flag
  gfs2: Implement SEEK_HOLE / SEEK_DATA via iomap
  gfs2: Fix xattr fsync
  gfs2: Fix a harmless typo
  gfs2: Fix and clean up {GET,SET}FLAGS ioctl
  gfs2: Add support for statx inode flags
  gfs2: Allow gfs2_xattr_set to be called with the glock held

Bob Peterson (5):
  GFS2: Make height info part of metapath
  GFS2: Implement iomap for block_map
  GFS2: Switch fiemap implementation to use iomap
  GFS2: flush the log and all pages for jdata as we do for WB_SYNC_ALL
  GFS2: Take inode off order_write list when setting jdata flag

-- 
2.13.6

[Cluster-devel] [PATCH 01/17] gfs2: Clarify gfs2_block_map

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Add a comment about the logical block size for directories.  Rename
"bsize" in gfs2_block_map to "factor".  Fix a typo in the description of
metaptr1.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/bmap.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 3dd0cceefa43..8830e2903a34 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -248,7 +248,7 @@ static inline unsigned int metapath_branch_start(const struct metapath *mp)
 }
 
 /**
- * metaptr1 - Return the first possible metadata pointer in a metaath buffer
+ * metaptr1 - Return the first possible metadata pointer in a metapath buffer
  * @height: The metadata height (0 = dinode)
  * @mp: The metapath
  */
@@ -659,7 +659,7 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
-	unsigned int bsize = sdp->sd_sb.sb_bsize;
+	unsigned int factor = sdp->sd_sb.sb_bsize;
 	const size_t maxlen = bh_map->b_size >> inode->i_blkbits;
 	const u64 *arr = sdp->sd_heightsize;
 	__be64 *ptr;
@@ -679,8 +679,14 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 	clear_buffer_new(bh_map);
 	clear_buffer_boundary(bh_map);
 	trace_gfs2_bmap(ip, bh_map, lblock, create, 1);
+
+	/*
+	 * Directory data blocks have a struct gfs2_meta_header header, so the
+	 * remaining size is smaller than the filesystem block size.  Logical
+	 * block numbers for directories are in units of this remaining size!
+	 */
 	if (gfs2_is_dir(ip)) {
-		bsize = sdp->sd_jbsize;
+		factor = sdp->sd_jbsize;
 		arr = sdp->sd_jheightsize;
 	}
 
@@ -689,7 +695,7 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 		goto out;
 
 	height = ip->i_height;
-	size = (lblock + 1) * bsize;
+	size = (lblock + 1) * factor;
 	while (size > arr[height])
 		height++;
 	find_metapath(sdp, lblock, &mp, height);
-- 
2.13.6

[Cluster-devel] [PATCH 02/17] gfs2: Update ctime in setflags ioctl

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

The FS_IOC_SETFLAGS ioctl is supposed to update the inode ctime.
Fixes xfstests generic/277.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/file.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 33a0cb5701a3..c7a904a8fbb4 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -271,6 +271,7 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 	error = gfs2_meta_inode_buffer(ip, &bh);
 	if (error)
 		goto out_trans_end;
+	inode->i_ctime = current_time(inode);
 	gfs2_trans_add_meta(ip->i_gl, bh);
 	ip->i_diskflags = new_flags;
 	gfs2_dinode_out(ip, bh->b_data);
-- 
2.13.6

[Cluster-devel] [PATCH 03/17] gfs2: Support negative atimes

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

When inodes are read from disk, GFS2 will only update in-memory atimes
older than the on-disk atimes; this prevents atimes from going
backwards.  The atimes of newly allocated inodes are initialized to 0.
This means that when an atime is explicitly set to a negative value,
this value will not persist.

Fix by setting the atime of newly allocated inodes to the lowest
possible value instead of 0.

Fixes xfstest generic/258.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/inode.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 863749e29bf9..b288cf2f85e6 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -189,7 +189,8 @@ struct inode *gfs2_inode_lookup(struct super_block *sb, unsigned int type,
 
 		gfs2_set_iop(inode);
 
-		inode->i_atime.tv_sec = 0;
+		/* Lowest possible timestamp; will be overwritten in gfs2_dinode_in. */
+		inode->i_atime.tv_sec = 1LL << (8 * sizeof(inode->i_atime.tv_sec) - 1);
 		inode->i_atime.tv_nsec = 0;
 
 		unlock_new_inode(inode);
-- 
2.13.6

[Cluster-devel] [PATCH 04/17] gfs2: Always update inode ctime in set_acl

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Three-entry POSIX ACLs can be stored in the file mode permission bits,
with no need to store them in extended attributes.  When a process sets
such a minimal ACL, the kernel updates the file mode like chmod does,
and removes any existing extended attributes for that ACL.  Make sure
the ctime is always updated in that case.

Fixes xfstest generic/307.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/acl.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c
index 9d5eecb123de..776717f1eeea 100644
--- a/fs/gfs2/acl.c
+++ b/fs/gfs2/acl.c
@@ -141,6 +141,7 @@ int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 
 	ret = __gfs2_set_acl(inode, acl, type);
 	if (!ret && mode != inode->i_mode) {
+		inode->i_ctime = current_time(inode);
 		inode->i_mode = mode;
 		mark_inode_dirty(inode);
 	}
-- 
2.13.6

[Cluster-devel] [PATCH 05/17] iomap: Switch from blkno to disk offset

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Replace iomap->blkno, the sector number, with iomap->addr, the disk
offset in bytes.  For invalid disk offsets, use the special value
IOMAP_NULL_ADDR instead of IOMAP_NULL_BLOCK.

This allows to use iomap for mappings which are not block aligned, such
as inline data on ext4.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>  # iomap, xfs
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/buffer.c           |  4 ++--
 fs/dax.c              |  2 +-
 fs/ext2/inode.c       |  4 ++--
 fs/ext4/inode.c       |  4 ++--
 fs/iomap.c            | 11 +++++------
 fs/nfsd/blocklayout.c |  4 ++--
 fs/xfs/xfs_iomap.c    |  6 +++---
 include/linux/iomap.h | 10 +++++-----
 8 files changed, 22 insertions(+), 23 deletions(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index 170df856bdb9..bd4d0923cdce 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1978,8 +1978,8 @@ iomap_to_bh(struct inode *inode, sector_t block, struct buffer_head *bh,
 	case IOMAP_MAPPED:
 		if (offset >= i_size_read(inode))
 			set_buffer_new(bh);
-		bh->b_blocknr = (iomap->blkno >> (inode->i_blkbits - 9)) +
-				((offset - iomap->offset) >> inode->i_blkbits);
+		bh->b_blocknr = (iomap->addr + offset - iomap->offset) >>
+				inode->i_blkbits;
 		set_buffer_mapped(bh);
 		break;
 	}
diff --git a/fs/dax.c b/fs/dax.c
index f001d8c72a06..f3a44a7c14b3 100644
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -938,7 +938,7 @@ EXPORT_SYMBOL_GPL(__dax_zero_page_range);
 
 static sector_t dax_iomap_sector(struct iomap *iomap, loff_t pos)
 {
-	return iomap->blkno + (((pos & PAGE_MASK) - iomap->offset) >> 9);
+	return (iomap->addr + (pos & PAGE_MASK) - iomap->offset) >> 9;
 }
 
 static loff_t
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index 4dca6f348714..1b8fc73de4a1 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -820,11 +820,11 @@ static int ext2_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 
 	if (ret == 0) {
 		iomap->type = IOMAP_HOLE;
-		iomap->blkno = IOMAP_NULL_BLOCK;
+		iomap->addr = IOMAP_NULL_ADDR;
 		iomap->length = 1 << blkbits;
 	} else {
 		iomap->type = IOMAP_MAPPED;
-		iomap->blkno = (sector_t)bno << (blkbits - 9);
+		iomap->addr = (u64)bno << blkbits;
 		iomap->length = (u64)ret << blkbits;
 		iomap->flags |= IOMAP_F_MERGED;
 	}
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 31db875bc7a1..d9e633c12aae 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -3472,7 +3472,7 @@ static int ext4_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 
 	if (ret == 0) {
 		iomap->type = IOMAP_HOLE;
-		iomap->blkno = IOMAP_NULL_BLOCK;
+		iomap->addr = IOMAP_NULL_ADDR;
 		iomap->length = (u64)map.m_len << blkbits;
 	} else {
 		if (map.m_flags & EXT4_MAP_MAPPED) {
@@ -3483,7 +3483,7 @@ static int ext4_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
 			WARN_ON_ONCE(1);
 			return -EIO;
 		}
-		iomap->blkno = (sector_t)map.m_pblk << (blkbits - 9);
+		iomap->addr = (u64)map.m_pblk << blkbits;
 		iomap->length = (u64)map.m_len << blkbits;
 	}
 
diff --git a/fs/iomap.c b/fs/iomap.c
index be61cf742b5e..d25db039ee13 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -350,8 +350,8 @@ static int iomap_zero(struct inode *inode, loff_t pos, unsigned offset,
 static int iomap_dax_zero(loff_t pos, unsigned offset, unsigned bytes,
 		struct iomap *iomap)
 {
-	sector_t sector = iomap->blkno +
-		(((pos & ~(PAGE_SIZE - 1)) - iomap->offset) >> 9);
+	sector_t sector = (iomap->addr +
+			   (pos & PAGE_MASK) - iomap->offset) >> 9;
 
 	return __dax_zero_page_range(iomap->bdev, iomap->dax_dev, sector,
 			offset, bytes);
@@ -512,9 +512,8 @@ static int iomap_to_fiemap(struct fiemap_extent_info *fi,
 		flags |= FIEMAP_EXTENT_SHARED;
 
 	return fiemap_fill_next_extent(fi, iomap->offset,
-			iomap->blkno != IOMAP_NULL_BLOCK ? iomap->blkno << 9: 0,
+			iomap->addr != IOMAP_NULL_ADDR ? iomap->addr : 0,
 			iomap->length, flags);
-
 }
 
 static loff_t
@@ -823,7 +822,7 @@ iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 	bio = bio_alloc(GFP_KERNEL, 1);
 	bio_set_dev(bio, iomap->bdev);
 	bio->bi_iter.bi_sector =
-		iomap->blkno + ((pos - iomap->offset) >> 9);
+		(iomap->addr + pos - iomap->offset) >> 9;
 	bio->bi_private = dio;
 	bio->bi_end_io = iomap_dio_bio_end_io;
 
@@ -902,7 +901,7 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 		bio = bio_alloc(GFP_KERNEL, nr_pages);
 		bio_set_dev(bio, iomap->bdev);
 		bio->bi_iter.bi_sector =
-			iomap->blkno + ((pos - iomap->offset) >> 9);
+			(iomap->addr + pos - iomap->offset) >> 9;
 		bio->bi_write_hint = dio->iocb->ki_hint;
 		bio->bi_private = dio;
 		bio->bi_end_io = iomap_dio_bio_end_io;
diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c
index c862c2489df0..2d1d37b27dc7 100644
--- a/fs/nfsd/blocklayout.c
+++ b/fs/nfsd/blocklayout.c
@@ -65,7 +65,7 @@ nfsd4_block_proc_layoutget(struct inode *inode, const struct svc_fh *fhp,
 			bex->es = PNFS_BLOCK_READ_DATA;
 		else
 			bex->es = PNFS_BLOCK_READWRITE_DATA;
-		bex->soff = (iomap.blkno << 9);
+		bex->soff = iomap.addr;
 		break;
 	case IOMAP_UNWRITTEN:
 		if (seg->iomode & IOMODE_RW) {
@@ -78,7 +78,7 @@ nfsd4_block_proc_layoutget(struct inode *inode, const struct svc_fh *fhp,
 			}
 
 			bex->es = PNFS_BLOCK_INVALID_DATA;
-			bex->soff = (iomap.blkno << 9);
+			bex->soff = iomap.addr;
 			break;
 		}
 		/*FALLTHRU*/
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index f179bdf1644d..9744b4819e0d 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -54,13 +54,13 @@ xfs_bmbt_to_iomap(
 	struct xfs_mount	*mp = ip->i_mount;
 
 	if (imap->br_startblock == HOLESTARTBLOCK) {
-		iomap->blkno = IOMAP_NULL_BLOCK;
+		iomap->addr = IOMAP_NULL_ADDR;
 		iomap->type = IOMAP_HOLE;
 	} else if (imap->br_startblock == DELAYSTARTBLOCK) {
-		iomap->blkno = IOMAP_NULL_BLOCK;
+		iomap->addr = IOMAP_NULL_ADDR;
 		iomap->type = IOMAP_DELALLOC;
 	} else {
-		iomap->blkno = xfs_fsb_to_db(ip, imap->br_startblock);
+		iomap->addr = BBTOB(xfs_fsb_to_db(ip, imap->br_startblock));
 		if (imap->br_state == XFS_EXT_UNWRITTEN)
 			iomap->type = IOMAP_UNWRITTEN;
 		else
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index f64dc6ce5161..7b8a615fa021 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -15,8 +15,8 @@ struct vm_fault;
  */
 #define IOMAP_HOLE	0x01	/* no blocks allocated, need allocation */
 #define IOMAP_DELALLOC	0x02	/* delayed allocation blocks */
-#define IOMAP_MAPPED	0x03	/* blocks allocated @blkno */
-#define IOMAP_UNWRITTEN	0x04	/* blocks allocated @blkno in unwritten state */
+#define IOMAP_MAPPED	0x03	/* blocks allocated at @addr */
+#define IOMAP_UNWRITTEN	0x04	/* blocks allocated at @addr in unwritten state */
 
 /*
  * Flags for all iomap mappings:
@@ -30,12 +30,12 @@ struct vm_fault;
 #define IOMAP_F_SHARED	0x20	/* block shared with another file */
 
 /*
- * Magic value for blkno:
+ * Magic value for addr:
  */
-#define IOMAP_NULL_BLOCK -1LL	/* blkno is not valid */
+#define IOMAP_NULL_ADDR -1ULL	/* addr is not valid */
 
 struct iomap {
-	sector_t		blkno;	/* 1st sector of mapping, 512b units */
+	u64			addr; /* disk offset of mapping, bytes */
 	loff_t			offset;	/* file offset of mapping, bytes */
 	u64			length;	/* length of mapping, bytes */
 	u16			type;	/* type of mapping */
-- 
2.13.6

[Cluster-devel] [PATCH 06/17] iomap: Add IOMAP_F_DATA_INLINE flag

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Add a new IOMAP_F_DATA_INLINE flag to indicate that a mapping is in a
disk area that contains data as well as metadata.  In iomap_fiemap, map
this flag to FIEMAP_EXTENT_DATA_INLINE.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/iomap.c            | 2 ++
 include/linux/iomap.h | 5 +++--
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/fs/iomap.c b/fs/iomap.c
index d25db039ee13..379eb9896093 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -510,6 +510,8 @@ static int iomap_to_fiemap(struct fiemap_extent_info *fi,
 		flags |= FIEMAP_EXTENT_MERGED;
 	if (iomap->flags & IOMAP_F_SHARED)
 		flags |= FIEMAP_EXTENT_SHARED;
+	if (iomap->flags & IOMAP_F_DATA_INLINE)
+		flags |= FIEMAP_EXTENT_DATA_INLINE;
 
 	return fiemap_fill_next_extent(fi, iomap->offset,
 			iomap->addr != IOMAP_NULL_ADDR ? iomap->addr : 0,
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 7b8a615fa021..2b0790dbd6ea 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -26,8 +26,9 @@ struct vm_fault;
 /*
  * Flags that only need to be reported for IOMAP_REPORT requests:
  */
-#define IOMAP_F_MERGED	0x10	/* contains multiple blocks/extents */
-#define IOMAP_F_SHARED	0x20	/* block shared with another file */
+#define IOMAP_F_MERGED		0x10	/* contains multiple blocks/extents */
+#define IOMAP_F_SHARED		0x20	/* block shared with another file */
+#define IOMAP_F_DATA_INLINE	0x40	/* data inline in the inode */
 
 /*
  * Magic value for addr:
-- 
2.13.6

[Cluster-devel] [PATCH 07/17] GFS2: Make height info part of metapath

[Bob Peterson]

This patch eliminates height parameters from function gfs2_bmap_alloc.
Function find_metapath determines the metapath's "find height", also
known as the desired height. Function lookup_metapath determines the
metapath's "actual height", previously known as starting height or
sheight. Function gfs2_bmap_alloc now gets both height values from
the metapath. This simplification was done as a step toward switching
the block_map functions to using iomap. The bh_map responsibilities
are also removed from function gfs2_bmap_alloc for the same reason.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/bmap.c | 109 ++++++++++++++++++++++++++++++++-------------------------
 1 file changed, 62 insertions(+), 47 deletions(-)

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 8830e2903a34..03badc8417d7 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -36,6 +36,8 @@
 struct metapath {
 	struct buffer_head *mp_bh[GFS2_MAX_META_HEIGHT];
 	__u16 mp_list[GFS2_MAX_META_HEIGHT];
+	int mp_fheight; /* find_metapath height */
+	int mp_aheight; /* actual height (lookup height) */
 };
 
 /**
@@ -235,9 +237,9 @@ static void find_metapath(const struct gfs2_sbd *sdp, u64 block,
 {
 	unsigned int i;
 
+	mp->mp_fheight = height;
 	for (i = height; i--;)
 		mp->mp_list[i] = do_div(block, sdp->sd_inptrs);
-
 }
 
 static inline unsigned int metapath_branch_start(const struct metapath *mp)
@@ -345,10 +347,13 @@ static int lookup_metapath(struct gfs2_inode *ip, struct metapath *mp)
 	for (x = 0; x < end_of_metadata; x++) {
 		ret = lookup_mp_height(ip, mp, x);
 		if (ret)
-			return ret;
+			goto out;
 	}
 
-	return ip->i_height;
+	ret = ip->i_height;
+out:
+	mp->mp_aheight = ret;
+	return ret;
 }
 
 /**
@@ -480,10 +485,11 @@ static inline unsigned int hptrs(struct gfs2_sbd *sdp, const unsigned int hgt)
  * @inode: The GFS2 inode
  * @lblock: The logical starting block of the extent
  * @bh_map: This is used to return the mapping details
- * @mp: The metapath
- * @sheight: The starting height (i.e. whats already mapped)
- * @height: The height to build to
+ * @zero_new: True if newly allocated blocks should be zeroed
+ * @mp: The metapath, with proper height information calculated
  * @maxlen: The max number of data blocks to alloc
+ * @dblock: Pointer to return the resulting new block
+ * @dblks: Pointer to return the number of blocks allocated
  *
  * In this routine we may have to alloc:
  *   i) Indirect blocks to grow the metadata tree height
@@ -500,62 +506,63 @@ static inline unsigned int hptrs(struct gfs2_sbd *sdp, const unsigned int hgt)
  */
 
 static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
-			   struct buffer_head *bh_map, struct metapath *mp,
-			   const unsigned int sheight,
-			   const unsigned int height,
-			   const size_t maxlen)
+			   bool zero_new, struct metapath *mp,
+			   const size_t maxlen, sector_t *dblock,
+			   unsigned *dblks)
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	struct super_block *sb = sdp->sd_vfs;
 	struct buffer_head *dibh = mp->mp_bh[0];
-	u64 bn, dblock = 0;
+	u64 bn;
 	unsigned n, i, blks, alloced = 0, iblks = 0, branch_start = 0;
-	unsigned dblks = 0;
 	unsigned ptrs_per_blk;
-	const unsigned end_of_metadata = height - 1;
+	const unsigned end_of_metadata = mp->mp_fheight - 1;
 	int ret;
 	int eob = 0;
 	enum alloc_state state;
 	__be64 *ptr;
 	__be64 zero_bn = 0;
 
-	BUG_ON(sheight < 1);
+	BUG_ON(mp->mp_aheight < 1);
 	BUG_ON(dibh == NULL);
 
+	*dblock = 0;
+	*dblks = 0;
 	gfs2_trans_add_meta(ip->i_gl, dibh);
 
-	if (height == sheight) {
+	if (mp->mp_fheight == mp->mp_aheight) {
 		struct buffer_head *bh;
 		/* Bottom indirect block exists, find unalloced extent size */
 		ptr = metapointer(end_of_metadata, mp);
 		bh = mp->mp_bh[end_of_metadata];
-		dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen,
-					   &eob);
-		BUG_ON(dblks < 1);
+		*dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr,
+					    maxlen, &eob);
+		BUG_ON(*dblks < 1);
 		state = ALLOC_DATA;
 	} else {
 		/* Need to allocate indirect blocks */
-		ptrs_per_blk = height > 1 ? sdp->sd_inptrs : sdp->sd_diptrs;
-		dblks = min(maxlen, (size_t)(ptrs_per_blk -
-					     mp->mp_list[end_of_metadata]));
-		if (height == ip->i_height) {
+		ptrs_per_blk = mp->mp_fheight > 1 ? sdp->sd_inptrs :
+			sdp->sd_diptrs;
+		*dblks = min(maxlen, (size_t)(ptrs_per_blk -
+					      mp->mp_list[end_of_metadata]));
+		if (mp->mp_fheight == ip->i_height) {
 			/* Writing into existing tree, extend tree down */
-			iblks = height - sheight;
+			iblks = mp->mp_fheight - mp->mp_aheight;
 			state = ALLOC_GROW_DEPTH;
 		} else {
 			/* Building up tree height */
 			state = ALLOC_GROW_HEIGHT;
-			iblks = height - ip->i_height;
+			iblks = mp->mp_fheight - ip->i_height;
 			branch_start = metapath_branch_start(mp);
-			iblks += (height - branch_start);
+			iblks += (mp->mp_fheight - branch_start);
 		}
 	}
 
 	/* start of the second part of the function (state machine) */
 
-	blks = dblks + iblks;
-	i = sheight;
+	blks = *dblks + iblks;
+	i = mp->mp_aheight;
 	do {
 		int error;
 		n = blks - alloced;
@@ -573,9 +580,10 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 						 sizeof(struct gfs2_dinode));
 				zero_bn = *ptr;
 			}
-			for (; i - 1 < height - ip->i_height && n > 0; i++, n--)
+			for (; i - 1 < mp->mp_fheight - ip->i_height && n > 0;
+			     i++, n--)
 				gfs2_indirect_init(mp, ip->i_gl, i, 0, bn++);
-			if (i - 1 == height - ip->i_height) {
+			if (i - 1 == mp->mp_fheight - ip->i_height) {
 				i--;
 				gfs2_buffer_copy_tail(mp->mp_bh[i],
 						sizeof(struct gfs2_meta_header),
@@ -587,7 +595,7 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 					sizeof(struct gfs2_meta_header));
 				*ptr = zero_bn;
 				state = ALLOC_GROW_DEPTH;
-				for(i = branch_start; i < height; i++) {
+				for(i = branch_start; i < mp->mp_fheight; i++) {
 					if (mp->mp_bh[i] == NULL)
 						break;
 					brelse(mp->mp_bh[i]);
@@ -599,44 +607,40 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 				break;
 		/* Branching from existing tree */
 		case ALLOC_GROW_DEPTH:
-			if (i > 1 && i < height)
+			if (i > 1 && i < mp->mp_fheight)
 				gfs2_trans_add_meta(ip->i_gl, mp->mp_bh[i-1]);
-			for (; i < height && n > 0; i++, n--)
+			for (; i < mp->mp_fheight && n > 0; i++, n--)
 				gfs2_indirect_init(mp, ip->i_gl, i,
 						   mp->mp_list[i-1], bn++);
-			if (i == height)
+			if (i == mp->mp_fheight)
 				state = ALLOC_DATA;
 			if (n == 0)
 				break;
 		/* Tree complete, adding data blocks */
 		case ALLOC_DATA:
-			BUG_ON(n > dblks);
+			BUG_ON(n > *dblks);
 			BUG_ON(mp->mp_bh[end_of_metadata] == NULL);
 			gfs2_trans_add_meta(ip->i_gl, mp->mp_bh[end_of_metadata]);
-			dblks = n;
+			*dblks = n;
 			ptr = metapointer(end_of_metadata, mp);
-			dblock = bn;
+			*dblock = bn;
 			while (n-- > 0)
 				*ptr++ = cpu_to_be64(bn++);
-			if (buffer_zeronew(bh_map)) {
-				ret = sb_issue_zeroout(sb, dblock, dblks,
+			if (zero_new) {
+				ret = sb_issue_zeroout(sb, *dblock, *dblks,
 						       GFP_NOFS);
 				if (ret) {
 					fs_err(sdp,
 					       "Failed to zero data buffers\n");
-					clear_buffer_zeronew(bh_map);
 				}
 			}
 			break;
 		}
-	} while ((state != ALLOC_DATA) || !dblock);
+	} while ((state != ALLOC_DATA) || !(*dblock));
 
-	ip->i_height = height;
+	ip->i_height = mp->mp_fheight;
 	gfs2_add_inode_blocks(&ip->i_inode, alloced);
 	gfs2_dinode_out(ip, mp->mp_bh[0]->b_data);
-	map_bh(bh_map, inode->i_sb, dblock);
-	bh_map->b_size = dblks << inode->i_blkbits;
-	set_buffer_new(bh_map);
 	return 0;
 }
 
@@ -670,6 +674,9 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 	unsigned int len;
 	struct buffer_head *bh;
 	u8 height;
+	bool zero_new = false;
+	sector_t dblock = 0;
+	unsigned dblks;
 
 	BUG_ON(maxlen == 0);
 
@@ -699,13 +706,13 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 	while (size > arr[height])
 		height++;
 	find_metapath(sdp, lblock, &mp, height);
-	ret = 1;
+	mp.mp_aheight = 1;
 	if (height > ip->i_height || gfs2_is_stuffed(ip))
 		goto do_alloc;
 	ret = lookup_metapath(ip, &mp);
 	if (ret < 0)
 		goto out;
-	if (ret != ip->i_height)
+	if (mp.mp_aheight != ip->i_height)
 		goto do_alloc;
 	ptr = metapointer(ip->i_height - 1, &mp);
 	if (*ptr == 0)
@@ -732,7 +739,15 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 	}
 
 	/* At this point ret is the tree depth of already allocated blocks */
-	ret = gfs2_bmap_alloc(inode, lblock, bh_map, &mp, ret, height, maxlen);
+	if (buffer_zeronew(bh_map))
+		zero_new = true;
+	ret = gfs2_bmap_alloc(inode, lblock, zero_new, &mp, maxlen, &dblock,
+			      &dblks);
+	if (ret == 0) {
+		map_bh(bh_map, inode->i_sb, dblock);
+		bh_map->b_size = dblks << inode->i_blkbits;
+		set_buffer_new(bh_map);
+	}
 	goto out;
 }
 
-- 
2.13.6

[Cluster-devel] [PATCH 08/17] GFS2: Implement iomap for block_map

[Bob Peterson]

This patch implements iomap for block mapping, and switches the
block_map function to use it under the covers.

The additional IOMAP_F_BOUNDARY iomap flag indicates when iomap has
reached a "metadata boundary" and fetching the next mapping is likely to
incur an additional I/O.  This flag is used for setting the bh buffer
boundary flag.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/bmap.c        | 273 +++++++++++++++++++++++++++++++++++++-------------
 fs/gfs2/bmap.h        |   4 +
 fs/gfs2/trace_gfs2.h  |  65 ++++++++++++
 include/linux/iomap.h |   3 +-
 4 files changed, 276 insertions(+), 69 deletions(-)

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 03badc8417d7..d5f0d96169c5 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -13,6 +13,7 @@
 #include <linux/blkdev.h>
 #include <linux/gfs2_ondisk.h>
 #include <linux/crc32.h>
+#include <linux/iomap.h>
 
 #include "gfs2.h"
 #include "incore.h"
@@ -505,10 +506,8 @@ static inline unsigned int hptrs(struct gfs2_sbd *sdp, const unsigned int hgt)
  * Returns: errno on error
  */
 
-static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
-			   bool zero_new, struct metapath *mp,
-			   const size_t maxlen, sector_t *dblock,
-			   unsigned *dblks)
+static int gfs2_iomap_alloc(struct inode *inode, struct iomap *iomap,
+			    unsigned flags, struct metapath *mp)
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
@@ -516,36 +515,37 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 	struct buffer_head *dibh = mp->mp_bh[0];
 	u64 bn;
 	unsigned n, i, blks, alloced = 0, iblks = 0, branch_start = 0;
+	unsigned dblks = 0;
 	unsigned ptrs_per_blk;
 	const unsigned end_of_metadata = mp->mp_fheight - 1;
 	int ret;
-	int eob = 0;
 	enum alloc_state state;
 	__be64 *ptr;
 	__be64 zero_bn = 0;
+	size_t maxlen = iomap->length >> inode->i_blkbits;
 
 	BUG_ON(mp->mp_aheight < 1);
 	BUG_ON(dibh == NULL);
 
-	*dblock = 0;
-	*dblks = 0;
 	gfs2_trans_add_meta(ip->i_gl, dibh);
 
 	if (mp->mp_fheight == mp->mp_aheight) {
 		struct buffer_head *bh;
+		int eob;
+
 		/* Bottom indirect block exists, find unalloced extent size */
 		ptr = metapointer(end_of_metadata, mp);
 		bh = mp->mp_bh[end_of_metadata];
-		*dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr,
-					    maxlen, &eob);
-		BUG_ON(*dblks < 1);
+		dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr,
+					   maxlen, &eob);
+		BUG_ON(dblks < 1);
 		state = ALLOC_DATA;
 	} else {
 		/* Need to allocate indirect blocks */
 		ptrs_per_blk = mp->mp_fheight > 1 ? sdp->sd_inptrs :
 			sdp->sd_diptrs;
-		*dblks = min(maxlen, (size_t)(ptrs_per_blk -
-					      mp->mp_list[end_of_metadata]));
+		dblks = min(maxlen, (size_t)(ptrs_per_blk -
+					     mp->mp_list[end_of_metadata]));
 		if (mp->mp_fheight == ip->i_height) {
 			/* Writing into existing tree, extend tree down */
 			iblks = mp->mp_fheight - mp->mp_aheight;
@@ -561,7 +561,7 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 
 	/* start of the second part of the function (state machine) */
 
-	blks = *dblks + iblks;
+	blks = dblks + iblks;
 	i = mp->mp_aheight;
 	do {
 		int error;
@@ -618,26 +618,29 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 				break;
 		/* Tree complete, adding data blocks */
 		case ALLOC_DATA:
-			BUG_ON(n > *dblks);
+			BUG_ON(n > dblks);
 			BUG_ON(mp->mp_bh[end_of_metadata] == NULL);
 			gfs2_trans_add_meta(ip->i_gl, mp->mp_bh[end_of_metadata]);
-			*dblks = n;
+			dblks = n;
 			ptr = metapointer(end_of_metadata, mp);
-			*dblock = bn;
+			iomap->addr = bn << inode->i_blkbits;
+			iomap->flags |= IOMAP_F_NEW;
 			while (n-- > 0)
 				*ptr++ = cpu_to_be64(bn++);
-			if (zero_new) {
-				ret = sb_issue_zeroout(sb, *dblock, *dblks,
-						       GFP_NOFS);
+			if (flags & IOMAP_ZERO) {
+				ret = sb_issue_zeroout(sb, iomap->addr >> inode->i_blkbits,
+						       dblks, GFP_NOFS);
 				if (ret) {
 					fs_err(sdp,
 					       "Failed to zero data buffers\n");
+					flags &= ~IOMAP_ZERO;
 				}
 			}
 			break;
 		}
-	} while ((state != ALLOC_DATA) || !(*dblock));
+	} while (iomap->addr == IOMAP_NULL_ADDR);
 
+	iomap->length = (u64)dblks << inode->i_blkbits;
 	ip->i_height = mp->mp_fheight;
 	gfs2_add_inode_blocks(&ip->i_inode, alloced);
 	gfs2_dinode_out(ip, mp->mp_bh[0]->b_data);
@@ -645,47 +648,123 @@ static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock,
 }
 
 /**
- * gfs2_block_map - Map a block from an inode to a disk block
+ * hole_size - figure out the size of a hole
  * @inode: The inode
- * @lblock: The logical block number
- * @bh_map: The bh to be mapped
- * @create: True if its ok to alloc blocks to satify the request
+ * @lblock: The logical starting block number
+ * @mp: The metapath
  *
- * Sets buffer_mapped() if successful, sets buffer_boundary() if a
- * read of metadata will be required before the next block can be
- * mapped. Sets buffer_new() if new blocks were allocated.
+ * Returns: The hole size in bytes
  *
- * Returns: errno
  */
+static u64 hole_size(struct inode *inode, sector_t lblock, struct metapath *mp)
+{
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct metapath mp_eof;
+	u64 factor = 1;
+	int hgt;
+	u64 holesz = 0;
+	const __be64 *first, *end, *ptr;
+	const struct buffer_head *bh;
+	u64 lblock_stop = (i_size_read(inode) - 1) >> inode->i_blkbits;
+	int zeroptrs;
+	bool done = false;
+
+	/* Get another metapath, to the very last byte */
+	find_metapath(sdp, lblock_stop, &mp_eof, ip->i_height);
+	for (hgt = ip->i_height - 1; hgt >= 0 && !done; hgt--) {
+		bh = mp->mp_bh[hgt];
+		if (bh) {
+			zeroptrs = 0;
+			first = metapointer(hgt, mp);
+			end = (const __be64 *)(bh->b_data + bh->b_size);
+
+			for (ptr = first; ptr < end; ptr++) {
+				if (*ptr) {
+					done = true;
+					break;
+				} else {
+					zeroptrs++;
+				}
+			}
+		} else {
+			zeroptrs = sdp->sd_inptrs;
+		}
+		if (factor * zeroptrs >= lblock_stop - lblock + 1) {
+			holesz = lblock_stop - lblock + 1;
+			break;
+		}
+		holesz += factor * zeroptrs;
 
-int gfs2_block_map(struct inode *inode, sector_t lblock,
-		   struct buffer_head *bh_map, int create)
+		factor *= sdp->sd_inptrs;
+		if (hgt && (mp->mp_list[hgt - 1] < mp_eof.mp_list[hgt - 1]))
+			(mp->mp_list[hgt - 1])++;
+	}
+	return holesz << inode->i_blkbits;
+}
+
+static void gfs2_stuffed_iomap(struct inode *inode, struct iomap *iomap)
+{
+	struct gfs2_inode *ip = GFS2_I(inode);
+
+	iomap->addr = (ip->i_no_addr << inode->i_blkbits) +
+		      sizeof(struct gfs2_dinode);
+	iomap->offset = 0;
+	iomap->length = i_size_read(inode);
+	iomap->type = IOMAP_MAPPED;
+	iomap->flags = IOMAP_F_DATA_INLINE;
+}
+
+/**
+ * gfs2_iomap_begin - Map blocks from an inode to disk blocks
+ * @inode: The inode
+ * @pos: Starting position in bytes
+ * @length: Length to map, in bytes
+ * @flags: iomap flags
+ * @iomap: The iomap structure
+ *
+ * Returns: errno
+ */
+int gfs2_iomap_begin(struct inode *inode, loff_t pos, loff_t length,
+		     unsigned flags, struct iomap *iomap)
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct metapath mp = { .mp_aheight = 1, };
 	unsigned int factor = sdp->sd_sb.sb_bsize;
-	const size_t maxlen = bh_map->b_size >> inode->i_blkbits;
 	const u64 *arr = sdp->sd_heightsize;
 	__be64 *ptr;
-	u64 size;
-	struct metapath mp;
+	sector_t lblock;
+	sector_t lend;
 	int ret;
 	int eob;
 	unsigned int len;
 	struct buffer_head *bh;
 	u8 height;
-	bool zero_new = false;
-	sector_t dblock = 0;
-	unsigned dblks;
 
-	BUG_ON(maxlen == 0);
+	trace_gfs2_iomap_start(ip, pos, length, flags);
+	if (!length) {
+		ret = -EINVAL;
+		goto out;
+	}
 
-	memset(&mp, 0, sizeof(mp));
-	bmap_lock(ip, create);
-	clear_buffer_mapped(bh_map);
-	clear_buffer_new(bh_map);
-	clear_buffer_boundary(bh_map);
-	trace_gfs2_bmap(ip, bh_map, lblock, create, 1);
+	if ((flags & IOMAP_REPORT) && gfs2_is_stuffed(ip)) {
+		gfs2_stuffed_iomap(inode, iomap);
+		if (pos >= iomap->length)
+			return -ENOENT;
+		ret = 0;
+		goto out;
+	}
+
+	lblock = pos >> inode->i_blkbits;
+	lend = (pos + length + sdp->sd_sb.sb_bsize - 1) >> inode->i_blkbits;
+
+	iomap->offset = lblock << inode->i_blkbits;
+	iomap->addr = IOMAP_NULL_ADDR;
+	iomap->type = IOMAP_HOLE;
+	iomap->length = (u64)(lend - lblock) << inode->i_blkbits;
+	iomap->flags = IOMAP_F_MERGED;
+	bmap_lock(ip, 0);
 
 	/*
 	 * Directory data blocks have a struct gfs2_meta_header header, so the
@@ -699,56 +778,114 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 
 	ret = gfs2_meta_inode_buffer(ip, &mp.mp_bh[0]);
 	if (ret)
-		goto out;
+		goto out_release;
 
 	height = ip->i_height;
-	size = (lblock + 1) * factor;
-	while (size > arr[height])
+	while ((lblock + 1) * factor > arr[height])
 		height++;
 	find_metapath(sdp, lblock, &mp, height);
-	mp.mp_aheight = 1;
 	if (height > ip->i_height || gfs2_is_stuffed(ip))
 		goto do_alloc;
+
 	ret = lookup_metapath(ip, &mp);
 	if (ret < 0)
-		goto out;
+		goto out_release;
+
 	if (mp.mp_aheight != ip->i_height)
 		goto do_alloc;
+
 	ptr = metapointer(ip->i_height - 1, &mp);
 	if (*ptr == 0)
 		goto do_alloc;
-	map_bh(bh_map, inode->i_sb, be64_to_cpu(*ptr));
+
+	iomap->type = IOMAP_MAPPED;
+	iomap->addr = be64_to_cpu(*ptr) << inode->i_blkbits;
+
 	bh = mp.mp_bh[ip->i_height - 1];
-	len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, &eob);
-	bh_map->b_size = (len << inode->i_blkbits);
+	len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, lend - lblock, &eob);
 	if (eob)
-		set_buffer_boundary(bh_map);
+		iomap->flags |= IOMAP_F_BOUNDARY;
+	iomap->length = (u64)len << inode->i_blkbits;
+
 	ret = 0;
-out:
+
+out_release:
 	release_metapath(&mp);
-	trace_gfs2_bmap(ip, bh_map, lblock, create, ret);
-	bmap_unlock(ip, create);
+	bmap_unlock(ip, 0);
+out:
+	trace_gfs2_iomap_end(ip, iomap, ret);
 	return ret;
 
 do_alloc:
-	/* All allocations are done here, firstly check create flag */
-	if (!create) {
-		BUG_ON(gfs2_is_stuffed(ip));
+	if (!(flags & IOMAP_WRITE)) {
+		if (pos >= i_size_read(inode)) {
+			ret = -ENOENT;
+			goto out_release;
+		}
 		ret = 0;
-		goto out;
+		iomap->length = hole_size(inode, lblock, &mp);
+		goto out_release;
 	}
 
-	/* At this point ret is the tree depth of already allocated blocks */
+	ret = gfs2_iomap_alloc(inode, iomap, flags, &mp);
+	goto out_release;
+}
+
+/**
+ * gfs2_block_map - Map a block from an inode to a disk block
+ * @inode: The inode
+ * @lblock: The logical block number
+ * @bh_map: The bh to be mapped
+ * @create: True if its ok to alloc blocks to satify the request
+ *
+ * Sets buffer_mapped() if successful, sets buffer_boundary() if a
+ * read of metadata will be required before the next block can be
+ * mapped. Sets buffer_new() if new blocks were allocated.
+ *
+ * Returns: errno
+ */
+
+int gfs2_block_map(struct inode *inode, sector_t lblock,
+		   struct buffer_head *bh_map, int create)
+{
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct iomap iomap;
+	int ret, flags = 0;
+
+	clear_buffer_mapped(bh_map);
+	clear_buffer_new(bh_map);
+	clear_buffer_boundary(bh_map);
+	trace_gfs2_bmap(ip, bh_map, lblock, create, 1);
+
+	if (create)
+		flags |= IOMAP_WRITE;
 	if (buffer_zeronew(bh_map))
-		zero_new = true;
-	ret = gfs2_bmap_alloc(inode, lblock, zero_new, &mp, maxlen, &dblock,
-			      &dblks);
-	if (ret == 0) {
-		map_bh(bh_map, inode->i_sb, dblock);
-		bh_map->b_size = dblks << inode->i_blkbits;
-		set_buffer_new(bh_map);
+		flags |= IOMAP_ZERO;
+	ret = gfs2_iomap_begin(inode, (loff_t)lblock << inode->i_blkbits,
+			       bh_map->b_size, flags, &iomap);
+	if (ret) {
+		if (!create && ret == -ENOENT) {
+			/* Return unmapped buffer beyond the end of file.  */
+			ret = 0;
+		}
+		goto out;
+	}
+
+	if (iomap.length > bh_map->b_size) {
+		iomap.length = bh_map->b_size;
+		iomap.flags &= ~IOMAP_F_BOUNDARY;
 	}
-	goto out;
+	if (iomap.addr != IOMAP_NULL_ADDR)
+		map_bh(bh_map, inode->i_sb, iomap.addr >> inode->i_blkbits);
+	bh_map->b_size = iomap.length;
+	if (iomap.flags & IOMAP_F_BOUNDARY)
+		set_buffer_boundary(bh_map);
+	if (iomap.flags & IOMAP_F_NEW)
+		set_buffer_new(bh_map);
+
+out:
+	trace_gfs2_bmap(ip, bh_map, lblock, create, ret);
+	return ret;
 }
 
 /*
diff --git a/fs/gfs2/bmap.h b/fs/gfs2/bmap.h
index 81ded5e2aaa2..443cc182cf18 100644
--- a/fs/gfs2/bmap.h
+++ b/fs/gfs2/bmap.h
@@ -10,6 +10,8 @@
 #ifndef __BMAP_DOT_H__
 #define __BMAP_DOT_H__
 
+#include <linux/iomap.h>
+
 #include "inode.h"
 
 struct inode;
@@ -47,6 +49,8 @@ static inline void gfs2_write_calc_reserv(const struct gfs2_inode *ip,
 extern int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page);
 extern int gfs2_block_map(struct inode *inode, sector_t lblock,
 			  struct buffer_head *bh, int create);
+extern int gfs2_iomap_begin(struct inode *inode, loff_t pos, loff_t length,
+			    unsigned flags, struct iomap *iomap);
 extern int gfs2_extent_map(struct inode *inode, u64 lblock, int *new,
 			   u64 *dblock, unsigned *extlen);
 extern int gfs2_setattr_size(struct inode *inode, u64 size);
diff --git a/fs/gfs2/trace_gfs2.h b/fs/gfs2/trace_gfs2.h
index 49ac55da4e33..3c91ae3cf0b2 100644
--- a/fs/gfs2/trace_gfs2.h
+++ b/fs/gfs2/trace_gfs2.h
@@ -12,6 +12,7 @@
 #include <linux/gfs2_ondisk.h>
 #include <linux/writeback.h>
 #include <linux/ktime.h>
+#include <linux/iomap.h>
 #include "incore.h"
 #include "glock.h"
 #include "rgrp.h"
@@ -469,6 +470,70 @@ TRACE_EVENT(gfs2_bmap,
 		  __entry->errno)
 );
 
+TRACE_EVENT(gfs2_iomap_start,
+
+	TP_PROTO(const struct gfs2_inode *ip, loff_t pos, ssize_t length,
+		 u16 flags),
+
+	TP_ARGS(ip, pos, length, flags),
+
+	TP_STRUCT__entry(
+		__field(        dev_t,  dev                     )
+		__field(	u64,	inum			)
+		__field(	loff_t, pos			)
+		__field(	ssize_t, length			)
+		__field(	u16,	flags			)
+	),
+
+	TP_fast_assign(
+		__entry->dev            = ip->i_gl->gl_name.ln_sbd->sd_vfs->s_dev;
+		__entry->inum		= ip->i_no_addr;
+		__entry->pos		= pos;
+		__entry->length		= length;
+		__entry->flags		= flags;
+	),
+
+	TP_printk("%u,%u bmap %llu iomap start %llu/%lu flags:%08x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->inum,
+		  (unsigned long long)__entry->pos,
+		  (unsigned long)__entry->length, (u16)__entry->flags)
+);
+
+TRACE_EVENT(gfs2_iomap_end,
+
+	TP_PROTO(const struct gfs2_inode *ip, struct iomap *iomap, int ret),
+
+	TP_ARGS(ip, iomap, ret),
+
+	TP_STRUCT__entry(
+		__field(        dev_t,  dev                     )
+		__field(	u64,	inum			)
+		__field(	loff_t, offset			)
+		__field(	ssize_t, length			)
+		__field(	u16,	flags			)
+		__field(	u16,	type			)
+		__field(	int,	ret			)
+	),
+
+	TP_fast_assign(
+		__entry->dev            = ip->i_gl->gl_name.ln_sbd->sd_vfs->s_dev;
+		__entry->inum		= ip->i_no_addr;
+		__entry->offset		= iomap->offset;
+		__entry->length		= iomap->length;
+		__entry->flags		= iomap->flags;
+		__entry->type		= iomap->type;
+		__entry->ret		= ret;
+	),
+
+	TP_printk("%u,%u bmap %llu iomap end %llu/%lu ty:%d flags:%08x rc:%d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->inum,
+		  (unsigned long long)__entry->offset,
+		  (unsigned long)__entry->length, (u16)__entry->type,
+		  (u16)__entry->flags, __entry->ret)
+);
+
 /* Keep track of blocks as they are allocated/freed */
 TRACE_EVENT(gfs2_block_alloc,
 
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 2b0790dbd6ea..a61be86710b5 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -21,7 +21,8 @@ struct vm_fault;
 /*
  * Flags for all iomap mappings:
  */
-#define IOMAP_F_NEW	0x01	/* blocks have been newly allocated */
+#define IOMAP_F_NEW		0x01	/* blocks have been newly allocated */
+#define IOMAP_F_BOUNDARY	0x02	/* mapping ends at metadata boundary */
 
 /*
  * Flags that only need to be reported for IOMAP_REPORT requests:
-- 
2.13.6

[Cluster-devel] [PATCH 09/17] GFS2: Switch fiemap implementation to use iomap

[Bob Peterson]

This patch switches GFS2's implementation of fiemap from the old
block_map code to the new iomap interface.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/Kconfig |  1 +
 fs/gfs2/inode.c | 34 +++++++++-------------------------
 2 files changed, 10 insertions(+), 25 deletions(-)

diff --git a/fs/gfs2/Kconfig b/fs/gfs2/Kconfig
index 90c6a8faaecb..43c827a7cce5 100644
--- a/fs/gfs2/Kconfig
+++ b/fs/gfs2/Kconfig
@@ -4,6 +4,7 @@ config GFS2_FS
 	select FS_POSIX_ACL
 	select CRC32
 	select QUOTACTL
+	select FS_IOMAP
 	help
 	  A cluster filesystem.
 
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index b288cf2f85e6..321da48ca123 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -18,7 +18,7 @@
 #include <linux/posix_acl.h>
 #include <linux/gfs2_ondisk.h>
 #include <linux/crc32.h>
-#include <linux/fiemap.h>
+#include <linux/iomap.h>
 #include <linux/security.h>
 #include <linux/uaccess.h>
 
@@ -2003,6 +2003,10 @@ static int gfs2_getattr(const struct path *path, struct kstat *stat,
 	return 0;
 }
 
+const struct iomap_ops gfs2_iomap_ops = {
+	.iomap_begin = gfs2_iomap_begin,
+};
+
 static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		       u64 start, u64 len)
 {
@@ -2010,38 +2014,18 @@ static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	struct gfs2_holder gh;
 	int ret;
 
-	ret = fiemap_check_flags(fieinfo, FIEMAP_FLAG_SYNC);
-	if (ret)
-		return ret;
-
-	inode_lock(inode);
+	inode_lock_shared(inode);
 
 	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
 	if (ret)
 		goto out;
 
-	if (gfs2_is_stuffed(ip)) {
-		u64 phys = ip->i_no_addr << inode->i_blkbits;
-		u64 size = i_size_read(inode);
-		u32 flags = FIEMAP_EXTENT_LAST|FIEMAP_EXTENT_NOT_ALIGNED|
-			    FIEMAP_EXTENT_DATA_INLINE;
-		phys += sizeof(struct gfs2_dinode);
-		phys += start;
-		if (start + len > size)
-			len = size - start;
-		if (start < size)
-			ret = fiemap_fill_next_extent(fieinfo, start, phys,
-						      len, flags);
-		if (ret == 1)
-			ret = 0;
-	} else {
-		ret = __generic_block_fiemap(inode, fieinfo, start, len,
-					     gfs2_block_map);
-	}
+	ret = iomap_fiemap(inode, fieinfo, start, len, &gfs2_iomap_ops);
 
 	gfs2_glock_dq_uninit(&gh);
+
 out:
-	inode_unlock(inode);
+	inode_unlock_shared(inode);
 	return ret;
 }
 
-- 
2.13.6

[Cluster-devel] [PATCH 10/17] gfs2: Implement SEEK_HOLE / SEEK_DATA via iomap

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

So far, lseek on gfs2 did not report holes.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/file.c  | 17 ++++++++++++++---
 fs/gfs2/inode.c | 38 ++++++++++++++++++++++++++++++++++++++
 fs/gfs2/inode.h |  2 ++
 3 files changed, 54 insertions(+), 3 deletions(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index c7a904a8fbb4..8fefb80fe830 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -60,9 +60,7 @@ static loff_t gfs2_llseek(struct file *file, loff_t offset, int whence)
 	loff_t error;
 
 	switch (whence) {
-	case SEEK_END: /* These reference inode->i_size */
-	case SEEK_DATA:
-	case SEEK_HOLE:
+	case SEEK_END:
 		error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
 					   &i_gh);
 		if (!error) {
@@ -70,8 +68,21 @@ static loff_t gfs2_llseek(struct file *file, loff_t offset, int whence)
 			gfs2_glock_dq_uninit(&i_gh);
 		}
 		break;
+
+	case SEEK_DATA:
+		error = gfs2_seek_data(file, offset);
+		break;
+
+	case SEEK_HOLE:
+		error = gfs2_seek_hole(file, offset);
+		break;
+
 	case SEEK_CUR:
 	case SEEK_SET:
+		/*
+		 * These don't reference inode->i_size and don't depend on the
+		 * block mapping, so we don't need the glock.
+		 */
 		error = generic_file_llseek(file, offset, whence);
 		break;
 	default:
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 321da48ca123..4749a6b8e4dd 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -2029,6 +2029,44 @@ static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	return ret;
 }
 
+loff_t gfs2_seek_data(struct file *file, loff_t offset)
+{
+	struct inode *inode = file->f_mapping->host;
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	loff_t ret;
+
+	inode_lock_shared(inode);
+	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
+	if (!ret)
+		ret = iomap_seek_data(inode, offset, &gfs2_iomap_ops);
+	gfs2_glock_dq_uninit(&gh);
+	inode_unlock_shared(inode);
+
+	if (ret < 0)
+		return ret;
+	return vfs_setpos(file, ret, inode->i_sb->s_maxbytes);
+}
+
+loff_t gfs2_seek_hole(struct file *file, loff_t offset)
+{
+	struct inode *inode = file->f_mapping->host;
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	loff_t ret;
+
+	inode_lock_shared(inode);
+	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
+	if (!ret)
+		ret = iomap_seek_hole(inode, offset, &gfs2_iomap_ops);
+	gfs2_glock_dq_uninit(&gh);
+	inode_unlock_shared(inode);
+
+	if (ret < 0)
+		return ret;
+	return vfs_setpos(file, ret, inode->i_sb->s_maxbytes);
+}
+
 const struct inode_operations gfs2_file_iops = {
 	.permission = gfs2_permission,
 	.setattr = gfs2_setattr,
diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h
index aace8ce34a18..b5b6341a4f5c 100644
--- a/fs/gfs2/inode.h
+++ b/fs/gfs2/inode.h
@@ -109,6 +109,8 @@ extern int gfs2_setattr_simple(struct inode *inode, struct iattr *attr);
 extern struct inode *gfs2_lookup_simple(struct inode *dip, const char *name);
 extern void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf);
 extern int gfs2_open_common(struct inode *inode, struct file *file);
+extern loff_t gfs2_seek_data(struct file *file, loff_t offset);
+extern loff_t gfs2_seek_hole(struct file *file, loff_t offset);
 
 extern const struct inode_operations gfs2_file_iops;
 extern const struct inode_operations gfs2_dir_iops;
-- 
2.13.6

[Cluster-devel] [PATCH 11/17] GFS2: flush the log and all pages for jdata as we do for WB_SYNC_ALL

[Bob Peterson]

In function gfs2_write_inode, starting with patch a9185b41a4f84, we
only flush the log and call filemap_fdatawait if we're passed in a
wbc sync_mode of WB_SYNC_ALL. We also need to do these things if
we're evicting a jdata inode, because we might have jdata pages
still attached to bufdata descriptors that need to be revoked, but
by the time it gets to evict() it's too late to start a new
transaction. This patch changes it to treat jdata inodes as if
WB_SYNC_ALL had been specified.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Acked-by: Abhijith Das <adas@redhat.com>
---
 fs/gfs2/super.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 8e54f2e3a304..9cb5c9a97d69 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -754,14 +754,15 @@ static int gfs2_write_inode(struct inode *inode, struct writeback_control *wbc)
 	struct address_space *metamapping = gfs2_glock2aspace(ip->i_gl);
 	struct backing_dev_info *bdi = inode_to_bdi(metamapping->host);
 	int ret = 0;
+	bool flush_all = (wbc->sync_mode == WB_SYNC_ALL || gfs2_is_jdata(ip));
 
-	if (wbc->sync_mode == WB_SYNC_ALL)
+	if (flush_all)
 		gfs2_log_flush(GFS2_SB(inode), ip->i_gl, NORMAL_FLUSH);
 	if (bdi->wb.dirty_exceeded)
 		gfs2_ail1_flush(sdp, wbc);
 	else
 		filemap_fdatawrite(metamapping);
-	if (wbc->sync_mode == WB_SYNC_ALL)
+	if (flush_all)
 		ret = filemap_fdatawait(metamapping);
 	if (ret)
 		mark_inode_dirty_sync(inode);
-- 
2.13.6

[Cluster-devel] [PATCH 12/17] GFS2: Take inode off order_write list when setting jdata flag

[Bob Peterson]

This patch fixes a deadlock caused when the jdata flag is set for
inodes that are already on the ordered write list. Since it is
on the ordered write list, log_flush calls gfs2_ordered_write which
calls filemap_fdatawrite. But since the inode had the jdata flag
set, that calls gfs2_jdata_writepages, which tries to start a new
transaction. A new transaction cannot be started because it tries
to acquire the log_flush rwsem which is already locked by the log
flush operation.

The bottom line is: We cannot switch an inode from ordered to jdata
until we eliminate any ordered data pages (via log flush) or any
log_flush operation afterward will create the circular dependency
above. So we need to flush the log before setting the diskflags to
switch the file mode, then we need to remove the inode from the
ordered writes list.

Before this patch, the log flush was done for jdata->ordered, but
that's wrong. If we're going from jdata to ordered, we don't need
to call gfs2_log_flush because the call to filemap_fdatawrite will
do it for us:

   filemap_fdatawrite() -> __filemap_fdatawrite_range()
      __filemap_fdatawrite_range() -> do_writepages()
         do_writepages() -> gfs2_jdata_writepages()
            gfs2_jdata_writepages() -> gfs2_log_flush()

This patch modifies function do_gfs2_set_flags so that if a file
has its jdata flag set, and it's already on the ordered write list,
the log will be flushed and it will be removed from the list
before setting the flag.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Acked-by: Abhijith Das <adas@redhat.com>
---
 fs/gfs2/file.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 8fefb80fe830..c7aea96144b4 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -267,7 +267,7 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 			goto out;
 	}
 	if ((flags ^ new_flags) & GFS2_DIF_JDATA) {
-		if (flags & GFS2_DIF_JDATA)
+		if (new_flags & GFS2_DIF_JDATA)
 			gfs2_log_flush(sdp, ip->i_gl, NORMAL_FLUSH);
 		error = filemap_fdatawrite(inode->i_mapping);
 		if (error)
@@ -275,6 +275,8 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 		error = filemap_fdatawait(inode->i_mapping);
 		if (error)
 			goto out;
+		if (new_flags & GFS2_DIF_JDATA)
+			gfs2_ordered_del_inode(ip);
 	}
 	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
 	if (error)
-- 
2.13.6

[Cluster-devel] [PATCH 13/17] gfs2: Fix xattr fsync

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Make sure that changing xattrs marks the corresponding inode dirty so
that a subsequent fsync will sync those changes to disk.  We set
I_DIRTY_SYNC as well as I_DIRTY_DATASYNC so that both fsync and
fdatasync will sync xattr changes: xattrs can contain information
critical to how the data can be accessed, so we don't want fdatasync
to skip them.

Fixes xfstest generic/066.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Andrew Price <anprice@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/xattr.c | 40 ++++++++--------------------------------
 1 file changed, 8 insertions(+), 32 deletions(-)

diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c
index ea09e41dbb49..3e96dce21c1c 100644
--- a/fs/gfs2/xattr.c
+++ b/fs/gfs2/xattr.c
@@ -231,7 +231,6 @@ static int ea_dealloc_unstuffed(struct gfs2_inode *ip, struct buffer_head *bh,
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
 	struct gfs2_rgrpd *rgd;
 	struct gfs2_holder rg_gh;
-	struct buffer_head *dibh;
 	__be64 *dataptrs;
 	u64 bn = 0;
 	u64 bstart = 0;
@@ -308,13 +307,8 @@ static int ea_dealloc_unstuffed(struct gfs2_inode *ip, struct buffer_head *bh,
 		ea->ea_num_ptrs = 0;
 	}
 
-	error = gfs2_meta_inode_buffer(ip, &dibh);
-	if (!error) {
-		ip->i_inode.i_ctime = current_time(&ip->i_inode);
-		gfs2_trans_add_meta(ip->i_gl, dibh);
-		gfs2_dinode_out(ip, dibh->b_data);
-		brelse(dibh);
-	}
+	ip->i_inode.i_ctime = current_time(&ip->i_inode);
+	__mark_inode_dirty(&ip->i_inode, I_DIRTY_SYNC | I_DIRTY_DATASYNC);
 
 	gfs2_trans_end(sdp);
 
@@ -749,7 +743,6 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 			     ea_skeleton_call_t skeleton_call, void *private)
 {
 	struct gfs2_alloc_parms ap = { .target = blks };
-	struct buffer_head *dibh;
 	int error;
 
 	error = gfs2_rindex_update(GFS2_SB(&ip->i_inode));
@@ -774,13 +767,8 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 	if (error)
 		goto out_end_trans;
 
-	error = gfs2_meta_inode_buffer(ip, &dibh);
-	if (!error) {
-		ip->i_inode.i_ctime = current_time(&ip->i_inode);
-		gfs2_trans_add_meta(ip->i_gl, dibh);
-		gfs2_dinode_out(ip, dibh->b_data);
-		brelse(dibh);
-	}
+	ip->i_inode.i_ctime = current_time(&ip->i_inode);
+	__mark_inode_dirty(&ip->i_inode, I_DIRTY_SYNC | I_DIRTY_DATASYNC);
 
 out_end_trans:
 	gfs2_trans_end(GFS2_SB(&ip->i_inode));
@@ -891,7 +879,6 @@ static int ea_set_simple_noalloc(struct gfs2_inode *ip, struct buffer_head *bh,
 				 struct gfs2_ea_header *ea, struct ea_set *es)
 {
 	struct gfs2_ea_request *er = es->es_er;
-	struct buffer_head *dibh;
 	int error;
 
 	error = gfs2_trans_begin(GFS2_SB(&ip->i_inode), RES_DINODE + 2 * RES_EATTR, 0);
@@ -908,14 +895,9 @@ static int ea_set_simple_noalloc(struct gfs2_inode *ip, struct buffer_head *bh,
 	if (es->es_el)
 		ea_set_remove_stuffed(ip, es->es_el);
 
-	error = gfs2_meta_inode_buffer(ip, &dibh);
-	if (error)
-		goto out;
 	ip->i_inode.i_ctime = current_time(&ip->i_inode);
-	gfs2_trans_add_meta(ip->i_gl, dibh);
-	gfs2_dinode_out(ip, dibh->b_data);
-	brelse(dibh);
-out:
+	__mark_inode_dirty(&ip->i_inode, I_DIRTY_SYNC | I_DIRTY_DATASYNC);
+
 	gfs2_trans_end(GFS2_SB(&ip->i_inode));
 	return error;
 }
@@ -1111,7 +1093,6 @@ static int ea_remove_stuffed(struct gfs2_inode *ip, struct gfs2_ea_location *el)
 {
 	struct gfs2_ea_header *ea = el->el_ea;
 	struct gfs2_ea_header *prev = el->el_prev;
-	struct buffer_head *dibh;
 	int error;
 
 	error = gfs2_trans_begin(GFS2_SB(&ip->i_inode), RES_DINODE + RES_EATTR, 0);
@@ -1132,13 +1113,8 @@ static int ea_remove_stuffed(struct gfs2_inode *ip, struct gfs2_ea_location *el)
 		ea->ea_type = GFS2_EATYPE_UNUSED;
 	}
 
-	error = gfs2_meta_inode_buffer(ip, &dibh);
-	if (!error) {
-		ip->i_inode.i_ctime = current_time(&ip->i_inode);
-		gfs2_trans_add_meta(ip->i_gl, dibh);
-		gfs2_dinode_out(ip, dibh->b_data);
-		brelse(dibh);
-	}
+	ip->i_inode.i_ctime = current_time(&ip->i_inode);
+	__mark_inode_dirty(&ip->i_inode, I_DIRTY_SYNC | I_DIRTY_DATASYNC);
 
 	gfs2_trans_end(GFS2_SB(&ip->i_inode));
 
-- 
2.13.6

[Cluster-devel] [PATCH 14/17] gfs2: Fix a harmless typo

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/trans.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/gfs2/trans.c b/fs/gfs2/trans.c
index affef3c066e0..a85ca8b2c9ba 100644
--- a/fs/gfs2/trans.c
+++ b/fs/gfs2/trans.c
@@ -145,7 +145,7 @@ static struct gfs2_bufdata *gfs2_alloc_bufdata(struct gfs2_glock *gl,
  *
  * This is used in two distinct cases:
  * i) In ordered write mode
- *    We put the data buffer on a list so that we can ensure that its
+ *    We put the data buffer on a list so that we can ensure that it's
  *    synced to disk at the right time
  * ii) In journaled data mode
  *    We need to journal the data block in the same way as metadata in
-- 
2.13.6

[Cluster-devel] [PATCH 15/17] gfs2: Fix and clean up {GET, SET}FLAGS ioctl

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Switch to a simple array for mapping between the FS_*_FL and GFS_DIF_*
flags.  Clarify how the mapping between FS_JOURNAL_DATA_FL and the
filesystem flags works.  The GFS2_DIF_SYSTEM flag cannot be set from
user space, so remove it from GFS2_FLAGS_USER_SET.  Fail with -EINVAL
when trying to set flags that are not supported instead of silently
ignoring those flags.

Partially fixes xfstest generic/424.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Andrew Price <anprice@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/file.c | 102 ++++++++++++++++++++++++++-------------------------------
 1 file changed, 47 insertions(+), 55 deletions(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index c7aea96144b4..58705ef8643a 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -119,45 +119,22 @@ static int gfs2_readdir(struct file *file, struct dir_context *ctx)
 }
 
 /**
- * fsflags_cvt
- * @table: A table of 32 u32 flags
- * @val: a 32 bit value to convert
+ * fsflag_gfs2flag
  *
- * This function can be used to convert between fsflags values and
- * GFS2's own flags values.
- *
- * Returns: the converted flags
+ * The FS_JOURNAL_DATA_FL flag maps to GFS2_DIF_INHERIT_JDATA for directories,
+ * and to GFS2_DIF_JDATA for non-directories.
  */
-static u32 fsflags_cvt(const u32 *table, u32 val)
-{
-	u32 res = 0;
-	while(val) {
-		if (val & 1)
-			res |= *table;
-		table++;
-		val >>= 1;
-	}
-	return res;
-}
-
-static const u32 fsflags_to_gfs2[32] = {
-	[3] = GFS2_DIF_SYNC,
-	[4] = GFS2_DIF_IMMUTABLE,
-	[5] = GFS2_DIF_APPENDONLY,
-	[7] = GFS2_DIF_NOATIME,
-	[12] = GFS2_DIF_EXHASH,
-	[14] = GFS2_DIF_INHERIT_JDATA,
-	[17] = GFS2_DIF_TOPDIR,
-};
-
-static const u32 gfs2_to_fsflags[32] = {
-	[gfs2fl_Sync] = FS_SYNC_FL,
-	[gfs2fl_Immutable] = FS_IMMUTABLE_FL,
-	[gfs2fl_AppendOnly] = FS_APPEND_FL,
-	[gfs2fl_NoAtime] = FS_NOATIME_FL,
-	[gfs2fl_ExHash] = FS_INDEX_FL,
-	[gfs2fl_TopLevel] = FS_TOPDIR_FL,
-	[gfs2fl_InheritJdata] = FS_JOURNAL_DATA_FL,
+static struct {
+	u32 fsflag;
+	u32 gfsflag;
+} fsflag_gfs2flag[] = {
+	{FS_SYNC_FL, GFS2_DIF_SYNC},
+	{FS_IMMUTABLE_FL, GFS2_DIF_IMMUTABLE},
+	{FS_APPEND_FL, GFS2_DIF_APPENDONLY},
+	{FS_NOATIME_FL, GFS2_DIF_NOATIME},
+	{FS_INDEX_FL, GFS2_DIF_EXHASH},
+	{FS_TOPDIR_FL, GFS2_DIF_TOPDIR},
+	{FS_JOURNAL_DATA_FL, GFS2_DIF_JDATA | GFS2_DIF_INHERIT_JDATA},
 };
 
 static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
@@ -165,17 +142,23 @@ static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
 	struct inode *inode = file_inode(filp);
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_holder gh;
-	int error;
-	u32 fsflags;
+	int i, error;
+	u32 gfsflags, fsflags = 0;
 
 	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
 	error = gfs2_glock_nq(&gh);
 	if (error)
 		goto out_uninit;
 
-	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags);
-	if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA)
-		fsflags |= FS_JOURNAL_DATA_FL;
+	gfsflags = ip->i_diskflags;
+	if (S_ISDIR(inode->i_mode))
+		gfsflags &= ~GFS2_DIF_JDATA;
+	else
+		gfsflags &= ~GFS2_DIF_INHERIT_JDATA;
+	for (i = 0; i < ARRAY_SIZE(fsflag_gfs2flag); i++)
+		if (gfsflags & fsflag_gfs2flag[i].gfsflag)
+			fsflags |= fsflag_gfs2flag[i].fsflag;
+
 	if (put_user(fsflags, ptr))
 		error = -EFAULT;
 
@@ -210,7 +193,6 @@ void gfs2_set_inode_flags(struct inode *inode)
 			     GFS2_DIF_APPENDONLY|		\
 			     GFS2_DIF_NOATIME|			\
 			     GFS2_DIF_SYNC|			\
-			     GFS2_DIF_SYSTEM|			\
 			     GFS2_DIF_TOPDIR|			\
 			     GFS2_DIF_INHERIT_JDATA)
 
@@ -249,10 +231,6 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 	if ((new_flags ^ flags) == 0)
 		goto out;
 
-	error = -EINVAL;
-	if ((new_flags ^ flags) & ~GFS2_FLAGS_USER_SET)
-		goto out;
-
 	error = -EPERM;
 	if (IS_IMMUTABLE(inode) && (new_flags & GFS2_DIF_IMMUTABLE))
 		goto out;
@@ -303,19 +281,33 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
 {
 	struct inode *inode = file_inode(filp);
-	u32 fsflags, gfsflags;
+	u32 fsflags, gfsflags = 0;
+	u32 mask;
+	int i;
 
 	if (get_user(fsflags, ptr))
 		return -EFAULT;
 
-	gfsflags = fsflags_cvt(fsflags_to_gfs2, fsflags);
-	if (!S_ISDIR(inode->i_mode)) {
-		gfsflags &= ~GFS2_DIF_TOPDIR;
-		if (gfsflags & GFS2_DIF_INHERIT_JDATA)
-			gfsflags ^= (GFS2_DIF_JDATA | GFS2_DIF_INHERIT_JDATA);
-		return do_gfs2_set_flags(filp, gfsflags, ~GFS2_DIF_SYSTEM);
+	for (i = 0; i < ARRAY_SIZE(fsflag_gfs2flag); i++) {
+		if (fsflags & fsflag_gfs2flag[i].fsflag) {
+			fsflags &= ~fsflag_gfs2flag[i].fsflag;
+			gfsflags |= fsflag_gfs2flag[i].gfsflag;
+		}
+	}
+	if (fsflags || gfsflags & ~GFS2_FLAGS_USER_SET)
+		return -EINVAL;
+
+	mask = GFS2_FLAGS_USER_SET;
+	if (S_ISDIR(inode->i_mode)) {
+		mask &= ~GFS2_DIF_JDATA;
+	} else {
+		/* The GFS2_DIF_TOPDIR flag is only valid for directories. */
+		if (gfsflags & GFS2_DIF_TOPDIR)
+			return -EINVAL;
+		mask &= ~(GFS2_DIF_TOPDIR | GFS2_DIF_INHERIT_JDATA);
 	}
-	return do_gfs2_set_flags(filp, gfsflags, ~(GFS2_DIF_SYSTEM | GFS2_DIF_JDATA));
+
+	return do_gfs2_set_flags(filp, gfsflags, mask);
 }
 
 static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
-- 
2.13.6

[Cluster-devel] [PATCH 16/17] gfs2: Add support for statx inode flags

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

Add support for the STATX_ATTR_ flags in statx.  (Compression,
encryption, and the nodump flag are not supported by gfs2.)

Partially fixes xfstest generic/424.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Andrew Price <anprice@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/inode.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 4749a6b8e4dd..4e971b1c7f92 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -1987,6 +1987,7 @@ static int gfs2_getattr(const struct path *path, struct kstat *stat,
 	struct inode *inode = d_inode(path->dentry);
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_holder gh;
+	u32 gfsflags;
 	int error;
 
 	gfs2_holder_mark_uninitialized(&gh);
@@ -1996,7 +1997,20 @@ static int gfs2_getattr(const struct path *path, struct kstat *stat,
 			return error;
 	}
 
+	gfsflags = ip->i_diskflags;
+	if (gfsflags & GFS2_DIF_APPENDONLY)
+		stat->attributes |= STATX_ATTR_APPEND;
+	if (gfsflags & GFS2_DIF_IMMUTABLE)
+		stat->attributes |= STATX_ATTR_IMMUTABLE;
+
+	stat->attributes_mask |= (STATX_ATTR_APPEND |
+				  STATX_ATTR_COMPRESSED |
+				  STATX_ATTR_ENCRYPTED |
+				  STATX_ATTR_IMMUTABLE |
+				  STATX_ATTR_NODUMP);
+
 	generic_fillattr(inode, stat);
+
 	if (gfs2_holder_initialized(&gh))
 		gfs2_glock_dq_uninit(&gh);
 
-- 
2.13.6

[Cluster-devel] [PATCH 17/17] gfs2: Allow gfs2_xattr_set to be called with the glock held

[Bob Peterson]

From: Andreas Gruenbacher <agruenba@redhat.com>

On the following call path:

  gfs2_setattr -> setattr_prepare -> ... ->
    cap_inode_killpriv -> ... ->
      gfs2_xattr_set

the glock is locked in gfs2_setattr, so check for recursive locking in
gfs2_xattr_set as gfs2_xattr_get already does.  While at it, get rid of
need_unlock in gfs2_xattr_get.

Fixes xfstest generic/093.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Acked-by: Abhijith Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
---
 fs/gfs2/xattr.c | 23 ++++++++++++++++-------
 1 file changed, 16 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c
index 3e96dce21c1c..05de20954659 100644
--- a/fs/gfs2/xattr.c
+++ b/fs/gfs2/xattr.c
@@ -610,7 +610,6 @@ static int gfs2_xattr_get(const struct xattr_handler *handler,
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_holder gh;
-	bool need_unlock = false;
 	int ret;
 
 	/* During lookup, SELinux calls this function with the glock locked. */
@@ -619,10 +618,11 @@ static int gfs2_xattr_get(const struct xattr_handler *handler,
 		ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &gh);
 		if (ret)
 			return ret;
-		need_unlock = true;
+	} else {
+		gfs2_holder_mark_uninitialized(&gh);
 	}
 	ret = __gfs2_xattr_get(inode, name, buffer, size, handler->flags);
-	if (need_unlock)
+	if (gfs2_holder_initialized(&gh))
 		gfs2_glock_dq_uninit(&gh);
 	return ret;
 }
@@ -1244,11 +1244,20 @@ static int gfs2_xattr_set(const struct xattr_handler *handler,
 	if (ret)
 		return ret;
 
-	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
-	if (ret)
-		return ret;
+	/* May be called from gfs_setattr with the glock locked. */
+
+	if (!gfs2_glock_is_locked_by_me(ip->i_gl)) {
+		ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
+		if (ret)
+			return ret;
+	} else {
+		if (WARN_ON_ONCE(ip->i_gl->gl_state != LM_ST_EXCLUSIVE))
+			return -EIO;
+		gfs2_holder_mark_uninitialized(&gh);
+	}
 	ret = __gfs2_xattr_set(inode, name, value, size, flags, handler->flags);
-	gfs2_glock_dq_uninit(&gh);
+	if (gfs2_holder_initialized(&gh))
+		gfs2_glock_dq_uninit(&gh);
 	return ret;
 }
 
-- 
2.13.6