projects/linux·commit e0ea4045

Merge tag 'xfs-for-linus-v3.12-rc1-2' of git://oss.sgi.com/xfs/xfs

e0ea4045 · Linus Torvalds <[email protected]> · 2013-09-12 16:13:41 -07:00

Pull xfs update #2 from Ben Myers:
 "Here we have defrag support for v5 superblock, a number of bugfixes
  and a cleanup or two.

   - defrag support for CRC filesystems
   - fix endian worning in xlog_recover_get_buf_lsn
   - fixes for sparse warnings
   - fix for assert in xfs_dir3_leaf_hdr_from_disk
   - fix for log recovery of remote symlinks
   - fix for log recovery of btree root splits
   - fixes formemory allocation failures with ACLs
   - fix for assert in xfs_buf_item_relse
   - fix for assert in xfs_inode_buf_verify
   - fix an assignment in an assert that should be a test in
     xfs_bmbt_change_owner
   - remove dead code in xlog_recover_inode_pass2"

* tag 'xfs-for-linus-v3.12-rc1-2' of git://oss.sgi.com/xfs/xfs:
  xfs: remove dead code from xlog_recover_inode_pass2
  xfs: = vs == typo in ASSERT()
  xfs: don't assert fail on bad inode numbers
  xfs: aborted buf items can be in the AIL.
  xfs: factor all the kmalloc-or-vmalloc fallback allocations
  xfs: fix memory allocation failures with ACLs
  xfs: ensure we copy buffer type in da btree root splits
  xfs: set remote symlink buffer type for recovery
  xfs: recovery of swap extents operations for CRC filesystems
  xfs: swap extents operations for CRC filesystems
  xfs: check magic numbers in dir3 leaf verifier first
  xfs: fix some minor sparse warnings
  xfs: fix endian warning in xlog_recover_get_buf_lsn()
Commit
e0ea4045bce3cee84e35746fb98946ca36781248
Parents
48efe453 · 08474ed6

Merge commit. Changes below are compared with the first parent: 48efe453.

Changes

25 files changed · +461 −166

Only the first 600 diff lines are shown. Download the complete patch.

fs/xfs/kmem.c

View file
diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c
index 4a7286c..a02cfb9 100644
--- a/fs/xfs/kmem.c
+++ b/fs/xfs/kmem.c
@@ -27,8 +27,6 @@
/*
* Greedy allocation. May fail and may return vmalloced memory.
- *
- * Must be freed using kmem_free_large.
*/
void *
kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize)
@@ -36,7 +34,7 @@ kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize)
void *ptr;
size_t kmsize = maxsize;
- while (!(ptr = kmem_zalloc_large(kmsize))) {
+ while (!(ptr = vzalloc(kmsize))) {
if ((kmsize >>= 1) <= minsize)
kmsize = minsize;
}
@@ -75,6 +73,17 @@ kmem_zalloc(size_t size, xfs_km_flags_t flags)
return ptr;
}
+void *
+kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
+{
+ void *ptr;
+
+ ptr = kmem_zalloc(size, flags | KM_MAYFAIL);
+ if (ptr)
+ return ptr;
+ return vzalloc(size);
+}
+
void
kmem_free(const void *ptr)
{

fs/xfs/kmem.h

View file
diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h
index b2f2620..3a7371c 100644
--- a/fs/xfs/kmem.h
+++ b/fs/xfs/kmem.h
@@ -57,17 +57,10 @@ kmem_flags_convert(xfs_km_flags_t flags)
extern void *kmem_alloc(size_t, xfs_km_flags_t);
extern void *kmem_zalloc(size_t, xfs_km_flags_t);
+extern void *kmem_zalloc_large(size_t size, xfs_km_flags_t);
extern void *kmem_realloc(const void *, size_t, size_t, xfs_km_flags_t);
extern void kmem_free(const void *);
-static inline void *kmem_zalloc_large(size_t size)
-{
- return vzalloc(size);
-}
-static inline void kmem_free_large(void *ptr)
-{
- vfree(ptr);
-}
extern void *kmem_zalloc_greedy(size_t *, size_t, size_t);

fs/xfs/xfs_acl.c

View file
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index 6951896..0e2f37e 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -152,7 +152,7 @@ xfs_get_acl(struct inode *inode, int type)
* go out to the disk.
*/
len = XFS_ACL_MAX_SIZE(ip->i_mount);
- xfs_acl = kzalloc(len, GFP_KERNEL);
+ xfs_acl = kmem_zalloc_large(len, KM_SLEEP);
if (!xfs_acl)
return ERR_PTR(-ENOMEM);
@@ -175,10 +175,10 @@ xfs_get_acl(struct inode *inode, int type)
if (IS_ERR(acl))
goto out;
- out_update_cache:
+out_update_cache:
set_cached_acl(inode, type, acl);
- out:
- kfree(xfs_acl);
+out:
+ kmem_free(xfs_acl);
return acl;
}
@@ -209,7 +209,7 @@ xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl)
struct xfs_acl *xfs_acl;
int len = XFS_ACL_MAX_SIZE(ip->i_mount);
- xfs_acl = kzalloc(len, GFP_KERNEL);
+ xfs_acl = kmem_zalloc_large(len, KM_SLEEP);
if (!xfs_acl)
return -ENOMEM;
@@ -222,7 +222,7 @@ xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl)
error = -xfs_attr_set(ip, ea_name, (unsigned char *)xfs_acl,
len, ATTR_ROOT);
- kfree(xfs_acl);
+ kmem_free(xfs_acl);
} else {
/*
* A NULL ACL argument means we want to remove the ACL.

fs/xfs/xfs_bmap.c

View file
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 92b8309..f47e65c 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -4450,7 +4450,7 @@ xfs_bmapi_write(
{
struct xfs_mount *mp = ip->i_mount;
struct xfs_ifork *ifp;
- struct xfs_bmalloca bma = { 0 }; /* args for xfs_bmap_alloc */
+ struct xfs_bmalloca bma = { NULL }; /* args for xfs_bmap_alloc */
xfs_fileoff_t end; /* end of mapped file region */
int eof; /* after the end of extents */
int error; /* error return */

fs/xfs/xfs_bmap_btree.c

View file
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index cf3bc76..bb8de8e 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -925,3 +925,47 @@ xfs_bmdr_maxrecs(
return blocklen / sizeof(xfs_bmdr_rec_t);
return blocklen / (sizeof(xfs_bmdr_key_t) + sizeof(xfs_bmdr_ptr_t));
}
+
+/*
+ * Change the owner of a btree format fork fo the inode passed in. Change it to
+ * the owner of that is passed in so that we can change owners before or after
+ * we switch forks between inodes. The operation that the caller is doing will
+ * determine whether is needs to change owner before or after the switch.
+ *
+ * For demand paged transactional modification, the fork switch should be done
+ * after reading in all the blocks, modifying them and pinning them in the
+ * transaction. For modification when the buffers are already pinned in memory,
+ * the fork switch can be done before changing the owner as we won't need to
+ * validate the owner until the btree buffers are unpinned and writes can occur
+ * again.
+ *
+ * For recovery based ownership change, there is no transactional context and
+ * so a buffer list must be supplied so that we can record the buffers that we
+ * modified for the caller to issue IO on.
+ */
+int
+xfs_bmbt_change_owner(
+ struct xfs_trans *tp,
+ struct xfs_inode *ip,
+ int whichfork,
+ xfs_ino_t new_owner,
+ struct list_head *buffer_list)
+{
+ struct xfs_btree_cur *cur;
+ int error;
+
+ ASSERT(tp || buffer_list);
+ ASSERT(!(tp && buffer_list));
+ if (whichfork == XFS_DATA_FORK)
+ ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_BTREE);
+ else
+ ASSERT(ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE);
+
+ cur = xfs_bmbt_init_cursor(ip->i_mount, tp, ip, whichfork);
+ if (!cur)
+ return ENOMEM;
+
+ error = xfs_btree_change_owner(cur, new_owner, buffer_list);
+ xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+ return error;
+}

fs/xfs/xfs_bmap_btree.h

View file
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index 1b726d6..e367461 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -236,6 +236,10 @@ extern int xfs_bmbt_get_maxrecs(struct xfs_btree_cur *, int level);
extern int xfs_bmdr_maxrecs(struct xfs_mount *, int blocklen, int leaf);
extern int xfs_bmbt_maxrecs(struct xfs_mount *, int blocklen, int leaf);
+extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
+ int whichfork, xfs_ino_t new_owner,
+ struct list_head *buffer_list);
+
extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
struct xfs_trans *, struct xfs_inode *, int);

fs/xfs/xfs_bmap_util.c

View file
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 541d59f..97f952c 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -612,13 +612,9 @@ xfs_getbmap(
if (bmv->bmv_count > ULONG_MAX / sizeof(struct getbmapx))
return XFS_ERROR(ENOMEM);
- out = kmem_zalloc(bmv->bmv_count * sizeof(struct getbmapx), KM_MAYFAIL);
- if (!out) {
- out = kmem_zalloc_large(bmv->bmv_count *
- sizeof(struct getbmapx));
- if (!out)
- return XFS_ERROR(ENOMEM);
- }
+ out = kmem_zalloc_large(bmv->bmv_count * sizeof(struct getbmapx), 0);
+ if (!out)
+ return XFS_ERROR(ENOMEM);
xfs_ilock(ip, XFS_IOLOCK_SHARED);
if (whichfork == XFS_DATA_FORK && !(iflags & BMV_IF_DELALLOC)) {
@@ -754,10 +750,7 @@ xfs_getbmap(
break;
}
- if (is_vmalloc_addr(out))
- kmem_free_large(out);
- else
- kmem_free(out);
+ kmem_free(out);
return error;
}
@@ -1789,14 +1782,6 @@ xfs_swap_extents(
int taforkblks = 0;
__uint64_t tmp;
- /*
- * We have no way of updating owner information in the BMBT blocks for
- * each inode on CRC enabled filesystems, so to avoid corrupting the
- * this metadata we simply don't allow extent swaps to occur.
- */
- if (xfs_sb_version_hascrc(&mp->m_sb))
- return XFS_ERROR(EINVAL);
-
tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL);
if (!tempifp) {
error = XFS_ERROR(ENOMEM);
@@ -1920,6 +1905,42 @@ xfs_swap_extents(
goto out_trans_cancel;
}
+ xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+ xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+
+ /*
+ * Before we've swapped the forks, lets set the owners of the forks
+ * appropriately. We have to do this as we are demand paging the btree
+ * buffers, and so the validation done on read will expect the owner
+ * field to be correctly set. Once we change the owners, we can swap the
+ * inode forks.
+ *
+ * Note the trickiness in setting the log flags - we set the owner log
+ * flag on the opposite inode (i.e. the inode we are setting the new
+ * owner to be) because once we swap the forks and log that, log
+ * recovery is going to see the fork as owned by the swapped inode,
+ * not the pre-swapped inodes.
+ */
+ src_log_flags = XFS_ILOG_CORE;
+ target_log_flags = XFS_ILOG_CORE;
+ if (ip->i_d.di_version == 3 &&
+ ip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+ target_log_flags |= XFS_ILOG_DOWNER;
+ error = xfs_bmbt_change_owner(tp, ip, XFS_DATA_FORK,
+ tip->i_ino, NULL);
+ if (error)
+ goto out_trans_cancel;
+ }
+
+ if (tip->i_d.di_version == 3 &&
+ tip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+ src_log_flags |= XFS_ILOG_DOWNER;
+ error = xfs_bmbt_change_owner(tp, tip, XFS_DATA_FORK,
+ ip->i_ino, NULL);
+ if (error)
+ goto out_trans_cancel;
+ }
+
/*
* Swap the data forks of the inodes
*/
@@ -1957,7 +1978,6 @@ xfs_swap_extents(
tip->i_delayed_blks = ip->i_delayed_blks;
ip->i_delayed_blks = 0;
- src_log_flags = XFS_ILOG_CORE;
switch (ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
@@ -1971,11 +1991,12 @@ xfs_swap_extents(
src_log_flags |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
+ ASSERT(ip->i_d.di_version < 3 ||
+ (src_log_flags & XFS_ILOG_DOWNER));
src_log_flags |= XFS_ILOG_DBROOT;
break;
}
- target_log_flags = XFS_ILOG_CORE;
switch (tip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
@@ -1990,13 +2011,11 @@ xfs_swap_extents(
break;
case XFS_DINODE_FMT_BTREE:
target_log_flags |= XFS_ILOG_DBROOT;
+ ASSERT(tip->i_d.di_version < 3 ||
+ (target_log_flags & XFS_ILOG_DOWNER));
break;
}
-
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
-
xfs_trans_log_inode(tp, ip, src_log_flags);
xfs_trans_log_inode(tp, tip, target_log_flags);

fs/xfs/xfs_btree.c

View file
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index 7a2b4da..5690e10 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -855,6 +855,41 @@ xfs_btree_readahead(
return xfs_btree_readahead_sblock(cur, lr, block);
}
+STATIC xfs_daddr_t
+xfs_btree_ptr_to_daddr(
+ struct xfs_btree_cur *cur,
+ union xfs_btree_ptr *ptr)
+{
+ if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+ ASSERT(ptr->l != cpu_to_be64(NULLDFSBNO));
+
+ return XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l));
+ } else {
+ ASSERT(cur->bc_private.a.agno != NULLAGNUMBER);
+ ASSERT(ptr->s != cpu_to_be32(NULLAGBLOCK));
+
+ return XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_private.a.agno,
+ be32_to_cpu(ptr->s));
+ }
+}
+
+/*
+ * Readahead @count btree blocks at the given @ptr location.
+ *
+ * We don't need to care about long or short form btrees here as we have a
+ * method of converting the ptr directly to a daddr available to us.
+ */
+STATIC void
+xfs_btree_readahead_ptr(
+ struct xfs_btree_cur *cur,
+ union xfs_btree_ptr *ptr,
+ xfs_extlen_t count)
+{
+ xfs_buf_readahead(cur->bc_mp->m_ddev_targp,
+ xfs_btree_ptr_to_daddr(cur, ptr),
+ cur->bc_mp->m_bsize * count, cur->bc_ops->buf_ops);
+}
+
/*
* Set the buffer for level "lev" in the cursor to bp, releasing
* any previous buffer.
@@ -1073,24 +1108,6 @@ xfs_btree_buf_to_ptr(
}
}
-STATIC xfs_daddr_t
-xfs_btree_ptr_to_daddr(
- struct xfs_btree_cur *cur,
- union xfs_btree_ptr *ptr)
-{
- if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
- ASSERT(ptr->l != cpu_to_be64(NULLDFSBNO));
-
- return XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l));
- } else {
- ASSERT(cur->bc_private.a.agno != NULLAGNUMBER);
- ASSERT(ptr->s != cpu_to_be32(NULLAGBLOCK));
-
- return XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_private.a.agno,
- be32_to_cpu(ptr->s));
- }
-}
-
STATIC void
xfs_btree_set_refs(
struct xfs_btree_cur *cur,
@@ -3869,3 +3886,120 @@ xfs_btree_get_rec(
*stat = 1;
return 0;
}
+
+/*
+ * Change the owner of a btree.
+ *
+ * The mechanism we use here is ordered buffer logging. Because we don't know
+ * how many buffers were are going to need to modify, we don't really want to
+ * have to make transaction reservations for the worst case of every buffer in a
+ * full size btree as that may be more space that we can fit in the log....
+ *
+ * We do the btree walk in the most optimal manner possible - we have sibling
+ * pointers so we can just walk all the blocks on each level from left to right
+ * in a single pass, and then move to the next level and do the same. We can
+ * also do readahead on the sibling pointers to get IO moving more quickly,
+ * though for slow disks this is unlikely to make much difference to performance
+ * as the amount of CPU work we have to do before moving to the next block is
+ * relatively small.
+ *
+ * For each btree block that we load, modify the owner appropriately, set the
+ * buffer as an ordered buffer and log it appropriately. We need to ensure that
+ * we mark the region we change dirty so that if the buffer is relogged in
+ * a subsequent transaction the changes we make here as an ordered buffer are
+ * correctly relogged in that transaction. If we are in recovery context, then
+ * just queue the modified buffer as delayed write buffer so the transaction
+ * recovery completion writes the changes to disk.
+ */
+static int
+xfs_btree_block_change_owner(
+ struct xfs_btree_cur *cur,
+ int level,
+ __uint64_t new_owner,
+ struct list_head *buffer_list)
+{
+ struct xfs_btree_block *block;
+ struct xfs_buf *bp;
+ union xfs_btree_ptr rptr;
+
+ /* do right sibling readahead */
+ xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
+
+ /* modify the owner */
+ block = xfs_btree_get_block(cur, level, &bp);
+ if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+ block->bb_u.l.bb_owner = cpu_to_be64(new_owner);
+ else
+ block->bb_u.s.bb_owner = cpu_to_be32(new_owner);
+
+ /*
+ * If the block is a root block hosted in an inode, we might not have a
+ * buffer pointer here and we shouldn't attempt to log the change as the
+ * information is already held in the inode and discarded when the root
+ * block is formatted into the on-disk inode fork. We still change it,
+ * though, so everything is consistent in memory.
+ */
+ if (bp) {
+ if (cur->bc_tp) {
+ xfs_trans_ordered_buf(cur->bc_tp, bp);
+ xfs_btree_log_block(cur, bp, XFS_BB_OWNER);
+ } else {
+ xfs_buf_delwri_queue(bp, buffer_list);
+ }
+ } else {
+ ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+ ASSERT(level == cur->bc_nlevels - 1);
+ }
+
+ /* now read rh sibling block for next iteration */
+ xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+ if (xfs_btree_ptr_is_null(cur, &rptr))
+ return ENOENT;
+
+ return xfs_btree_lookup_get_block(cur, level, &rptr, &block);
+}
+
+int
+xfs_btree_change_owner(
+ struct xfs_btree_cur *cur,
+ __uint64_t new_owner,
+ struct list_head *buffer_list)
+{
+ union xfs_btree_ptr lptr;
+ int level;
+ struct xfs_btree_block *block = NULL;
+ int error = 0;
+
+ cur->bc_ops->init_ptr_from_cur(cur, &lptr);
+
+ /* for each level */
+ for (level = cur->bc_nlevels - 1; level >= 0; level--) {
+ /* grab the left hand block */
+ error = xfs_btree_lookup_get_block(cur, level, &lptr, &block);
+ if (error)
+ return error;
+
+ /* readahead the left most block for the next level down */
+ if (level > 0) {
+ union xfs_btree_ptr *ptr;
+
+ ptr = xfs_btree_ptr_addr(cur, 1, block);
+ xfs_btree_readahead_ptr(cur, ptr, 1);
+
+ /* save for the next iteration of the loop */
+ lptr = *ptr;
+ }
+
+ /* for each buffer in the level */
+ do {
+ error = xfs_btree_block_change_owner(cur, level,
+ new_owner,
+ buffer_list);
+ } while (!error);
+
+ if (error != ENOENT)
+ return error;
+ }
+
+ return 0;
+}

fs/xfs/xfs_btree.h

View file
diff --git a/fs/xfs/xfs_btree.h b/fs/xfs/xfs_btree.h
index c8473c7..06729b6 100644
--- a/fs/xfs/xfs_btree.h
+++ b/fs/xfs/xfs_btree.h
@@ -121,15 +121,18 @@ union xfs_btree_rec {
/*
* For logging record fields.
*/
-#define XFS_BB_MAGIC 0x01
-#define XFS_BB_LEVEL 0x02
-#define XFS_BB_NUMRECS 0x04
-#define XFS_BB_LEFTSIB 0x08
-#define XFS_BB_RIGHTSIB 0x10
-#define XFS_BB_BLKNO 0x20
+#define XFS_BB_MAGIC (1 << 0)
+#define XFS_BB_LEVEL (1 << 1)
+#define XFS_BB_NUMRECS (1 << 2)
+#define XFS_BB_LEFTSIB (1 << 3)
+#define XFS_BB_RIGHTSIB (1 << 4)
+#define XFS_BB_BLKNO (1 << 5)
+#define XFS_BB_LSN (1 << 6)
+#define XFS_BB_UUID (1 << 7)
+#define XFS_BB_OWNER (1 << 8)
#define XFS_BB_NUM_BITS 5
#define XFS_BB_ALL_BITS ((1 << XFS_BB_NUM_BITS) - 1)
-#define XFS_BB_NUM_BITS_CRC 8
+#define XFS_BB_NUM_BITS_CRC 9
#define XFS_BB_ALL_BITS_CRC ((1 << XFS_BB_NUM_BITS_CRC) - 1)
/*
@@ -442,6 +445,8 @@ int xfs_btree_new_iroot(struct xfs_btree_cur *, int *, int *);
int xfs_btree_insert(struct xfs_btree_cur *, int *);
int xfs_btree_delete(struct xfs_btree_cur *, int *);
int xfs_btree_get_rec(struct xfs_btree_cur *, union xfs_btree_rec **, int *);
+int xfs_btree_change_owner(struct xfs_btree_cur *cur, __uint64_t new_owner,
+ struct list_head *buffer_list);
/*
* btree block CRC helpers

fs/xfs/xfs_buf_item.c

View file
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 3a944b1..88c5ea7 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -613,13 +613,27 @@ xfs_buf_item_unlock(
}
}
}
- if (clean || aborted) {
- if (atomic_dec_and_test(&bip->bli_refcount)) {
- ASSERT(!aborted || XFS_FORCED_SHUTDOWN(lip->li_mountp));
+
+ /*
+ * Clean buffers, by definition, cannot be in the AIL. However, aborted
+ * buffers may be dirty and hence in the AIL. Therefore if we are
+ * aborting a buffer and we've just taken the last refernce away, we
+ * have to check if it is in the AIL before freeing it. We need to free
+ * it in this case, because an aborted transaction has already shut the
+ * filesystem down and this is the last chance we will have to do so.
+ */
+ if (atomic_dec_and_test(&bip->bli_refcount)) {
+ if (clean)
+ xfs_buf_item_relse(bp);
+ else if (aborted) {
+ ASSERT(XFS_FORCED_SHUTDOWN(lip->li_mountp));
+ if (lip->li_flags & XFS_LI_IN_AIL) {
+ xfs_trans_ail_delete(lip->li_ailp, lip,
+ SHUTDOWN_LOG_IO_ERROR);
+ }
xfs_buf_item_relse(bp);
}
- } else
- atomic_dec(&bip->bli_refcount);
+ }
if (!(flags & XFS_BLI_HOLD))
xfs_buf_relse(bp);

fs/xfs/xfs_da_btree.c

View file
diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c
index d4e59a4..069537c 100644
--- a/fs/xfs/xfs_da_btree.c
+++ b/fs/xfs/xfs_da_btree.c
@@ -635,6 +635,7 @@ xfs_da3_root_split(
xfs_trans_log_buf(tp, bp, 0, size - 1);
bp->b_ops = blk1->bp->b_ops;
+ xfs_trans_buf_copy_type(bp, blk1->bp);
blk1->bp = bp;
blk1->blkno = blkno;

fs/xfs/xfs_dir2_leaf.c

View file
diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c
index 08984ee..1021c83 100644
--- a/fs/xfs/xfs_dir2_leaf.c

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_dquot_item.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_extent_busy.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_icache.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_icache.h

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_inode_buf.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_inode_buf.h

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_ioctl.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_ioctl32.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_itable.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_log.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_log_format.h

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_log_recover.c

View file

This diff is incomplete. Download the complete patch.

fs/xfs/xfs_symlink.c

View file

This diff is incomplete. Download the complete patch.