Btrfs devel - Aug 2012 - [PATCH 0/3] btrfs: extended inode refs

Currently btrfs has a limitation on the maximum number of hard links an
inode can have. Specifically, links are stored in an array of ref
items:

struct btrfs_inode_ref {
	__le64 index;
	__le16 name_len;
	/* name goes here */
} __attribute__ ((__packed__));

The ref arrays are found via key triple:

(inode objectid, BTRFS_INODE_EXTREF_KEY, parent dir objectid)

Since items can not exceed the size of a leaf, the total number of links
that can be stored for a given inode / parent dir pair is limited to under
4k. This works fine for the most common case of few to only a handful of
links. Once the link count gets higher however, we begin to return EMLINK.


The following patches fix this situation by introducing a new ref item:

struct btrfs_inode_extref {
	__le64 parent_objectid;
	__le64 index;
	__le16 name_len;
	__u8   name[0];
	/* name goes here */
} __attribute__ ((__packed__));

Extended refs use a different addressing scheme. Extended ref keys
look like:

(inode objectid, BTRFS_INODE_EXTREF_KEY, hash)

Where hash is defined as a function of the parent objectid and link name.

This effectively fixes the limitation, though we have a slightly less
efficient packing of link data. To keep the best of both worlds then, I
implemented the following behavior:

Extended refs don''t replace the existing ref array. An inode gets an
extended ref for a given link _only_ after the ref array has been filled.  So
the most common cases shouldn''t actually see any difference in
performance
or disk usage as they''ll never get to the point where we''re
using an
extended ref.

It''s important while reading the patches however that there''s
still the
possibility that we can have a set of operations that grow out an inode ref
array (adding some extended refs) and then remove only the refs in the
array.  I don''t really see this being common but it''s a case
we always have
to consider when coding these changes.

Extended refs handle the case of a hash collision by storing items with the
same key in an array just like the dir item code. This means we have to
search an array on rare occasion.

Testing wise, the basic namespace operations work well (link, unlink, etc).
The rest has gotten less debugging (and I really don''t have a great way
of
testing the code in tree-log.c) Attached to this e-mail are btrfs-progs
patches which make testing of the changes possible.

Finally, these patches are based off Linux v3.5.
	--Mark

Most recent review for this series can be found at:
http://thread.gmane.org/gmane.comp.file-systems.btrfs/17480

Thanks to Jan Schmidt for giving the patches thorough review. Most of the
changes are from his suggestions.

Changelog:

- rebased against 3.5

- many code style cleanups

- simplified ref_get_fields() and callers 

- count_inode_extrefs() return value is bubbled up to callers now

- Fixed up add_inode_ref() changes based on my understanding of the
  function. Hopefully I''ve hit upon the right changes :)


* Changes from the first version of this patch:

- Implemented collision handling.

- Standardized naming of extended ref variables (extref).

- moved hashing code to hash.h and gave the function a better name
  (btrfs_extref_hash).

- A few cleanups of error handling.

- Fixed a bug where btrfs_find_one_extref() was erroneously incrementing the
  extref offset before returning it.

- Moved btrfs_find_one_extref() into backref.c. This means that backref.c no
  longer has to include tree-log.h.

- Fixed a bug in iref_to_path() where we were looking for extended refs
  (this actually lead to other bugs). Since iref_to_path() only deals with
  directory inodes we would never have an extended ref.

- added some explicit locking calls in the backref.c changes

- Instead of adding a second iterate function for extended refs, I fixed up
  iterate_irefs_t arguments to take the raw information from whatever ref
  version we''re coming from. This removed a bunch of duplicated code.

- I am actually including a patch to btrfs-progs with this drop. :)



From: Mark Fasheh <mfasheh@suse.com>

[PATCH] btrfs-progs: basic support for extended inode refs

This patch adds enough mkfs support to turn on the superblock flag and
btrfs-debug-tree support so that we can visualize the state of extended refs
on disk.

Signed-off-by: Mark Fasheh <mfasheh@suse.de>
---
 ctree.h      |   27 ++++++++++++++++++++++++++-
 mkfs.c       |   14 +++++++++-----
 print-tree.c |   44 ++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 79 insertions(+), 6 deletions(-)

diff --git a/ctree.h b/ctree.h
index 6545c50..ebf38fe 100644
--- a/ctree.h
+++ b/ctree.h
@@ -115,6 +115,13 @@ struct btrfs_trans_handle;
  */
 #define BTRFS_NAME_LEN 255
 
+/*
+ * Theoretical limit is larger, but we keep this down to a sane
+ * value. That should limit greatly the possibility of collisions on
+ * inode ref items.
+ */
+#define	BTRFS_LINK_MAX	65535U
+
 /* 32 bytes in various csum fields */
 #define BTRFS_CSUM_SIZE 32
 
@@ -412,6 +419,7 @@ struct btrfs_super_block {
 #define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL	(1ULL << 1)
 #define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS	(1ULL << 2)
 #define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO	(1ULL << 3)
+
 /*
  * some patches floated around with a second compression method
  * lets save that incompat here for when they do get in
@@ -426,6 +434,7 @@ struct btrfs_super_block {
  */
 #define BTRFS_FEATURE_INCOMPAT_BIG_METADATA     (1ULL << 5)
 
+#define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF   (1ULL << 6)
 
 #define BTRFS_FEATURE_COMPAT_SUPP		0ULL
 #define BTRFS_FEATURE_COMPAT_RO_SUPP		0ULL
@@ -434,7 +443,8 @@ struct btrfs_super_block {
 	 BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL |	\
 	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO |		\
 	 BTRFS_FEATURE_INCOMPAT_BIG_METADATA |		\
-	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS)
+	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS |		\
+	 BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF)
 
 /*
  * A leaf is full of items. offset and size tell us where to find
@@ -573,6 +583,13 @@ struct btrfs_inode_ref {
 	/* name goes here */
 } __attribute__ ((__packed__));
 
+struct btrfs_inode_extref {
+	__le64 parent_objectid;
+	__le64 index;
+	__le16 name_len;
+	__u8   name[0]; /* name goes here */
+} __attribute__ ((__packed__));
+
 struct btrfs_timespec {
 	__le64 sec;
 	__le32 nsec;
@@ -866,6 +883,7 @@ struct btrfs_root {
  */
 #define BTRFS_INODE_ITEM_KEY		1
 #define BTRFS_INODE_REF_KEY		12
+#define BTRFS_INODE_EXTREF_KEY		13
 #define BTRFS_XATTR_ITEM_KEY		24
 #define BTRFS_ORPHAN_ITEM_KEY		48
 
@@ -1145,6 +1163,13 @@ BTRFS_SETGET_FUNCS(inode_ref_name_len, struct
btrfs_inode_ref, name_len, 16);
 BTRFS_SETGET_STACK_FUNCS(stack_inode_ref_name_len, struct btrfs_inode_ref,
name_len, 16);
 BTRFS_SETGET_FUNCS(inode_ref_index, struct btrfs_inode_ref, index, 64);
 
+/* struct btrfs_inode_extref */
+BTRFS_SETGET_FUNCS(inode_extref_parent, struct btrfs_inode_extref,
+		   parent_objectid, 64);
+BTRFS_SETGET_FUNCS(inode_extref_name_len, struct btrfs_inode_extref,
+		   name_len, 16);
+BTRFS_SETGET_FUNCS(inode_extref_index, struct btrfs_inode_extref, index, 64);
+
 /* struct btrfs_inode_item */
 BTRFS_SETGET_FUNCS(inode_generation, struct btrfs_inode_item, generation, 64);
 BTRFS_SETGET_FUNCS(inode_sequence, struct btrfs_inode_item, sequence, 64);
diff --git a/mkfs.c b/mkfs.c
index c531ef2..5c18a6d 100644
--- a/mkfs.c
+++ b/mkfs.c
@@ -1225,6 +1225,9 @@ int main(int ac, char **av)
 	u64 source_dir_size = 0;
 	char *pretty_buf;
 
+	struct btrfs_super_block *super;
+	u64 flags;
+
 	while(1) {
 		int c;
 		c = getopt_long(ac, av, "A:b:l:n:s:m:d:L:r:VM", long_options,
@@ -1426,13 +1429,14 @@ raid_groups:
 	ret = create_data_reloc_tree(trans, root);
 	BUG_ON(ret);
 
-	if (mixed) {
-		struct btrfs_super_block *super = &root->fs_info->super_copy;
-		u64 flags = btrfs_super_incompat_flags(super);
+	super = &root->fs_info->super_copy;
+	flags = btrfs_super_incompat_flags(super);
+	flags |= BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF;
 
+	if (mixed)
 		flags |= BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS;
-		btrfs_set_super_incompat_flags(super, flags);
-	}
+
+	btrfs_set_super_incompat_flags(super, flags);
 
 	printf("fs created label %s on %s\n\tnodesize %u leafsize %u "
 	    "sectorsize %u size %s\n",
diff --git a/print-tree.c b/print-tree.c
index fc134c0..6012df8 100644
--- a/print-tree.c
+++ b/print-tree.c
@@ -55,6 +55,42 @@ static int print_dir_item(struct extent_buffer *eb, struct
btrfs_item *item,
 	return 0;
 }
 
+static int print_inode_extref_item(struct extent_buffer *eb,
+				   struct btrfs_item *item,
+				   struct btrfs_inode_extref *extref)
+{
+	u32 total;
+	u32 cur = 0;
+	u32 len;
+	u32 name_len = 0;
+	u64 index = 0;
+	u64 parent_objid;
+	char namebuf[BTRFS_NAME_LEN];
+
+	total = btrfs_item_size(eb, item);
+
+	while (cur < total) {
+		index = btrfs_inode_extref_index(eb, extref);
+		name_len = btrfs_inode_extref_name_len(eb, extref);
+		parent_objid = btrfs_inode_extref_parent(eb, extref);
+
+		len = (name_len <= sizeof(namebuf))? name_len: sizeof(namebuf);
+
+		read_extent_buffer(eb, namebuf, (unsigned long)(extref->name), len);
+
+		printf("\t\tinode extref index %llu parent %llu namelen %u "
+		       "name: %.*s\n",
+		       (unsigned long long)index,
+		       (unsigned long long)parent_objid,
+		       name_len, len, namebuf);
+
+		len = sizeof(*extref) + name_len;
+		extref = (struct btrfs_inode_extref *)((char *)extref + len);
+		cur += len;
+	}
+	return 0;
+}
+
 static int print_inode_ref_item(struct extent_buffer *eb, struct btrfs_item
*item,
 				struct btrfs_inode_ref *ref)
 {
@@ -285,6 +321,9 @@ static void print_key_type(u8 type)
 	case BTRFS_INODE_REF_KEY:
 		printf("INODE_REF");
 		break;
+	case BTRFS_INODE_EXTREF_KEY:
+		printf("INODE_EXTREF");
+		break;
 	case BTRFS_DIR_ITEM_KEY:
 		printf("DIR_ITEM");
 		break;
@@ -454,6 +493,7 @@ void btrfs_print_leaf(struct btrfs_root *root, struct
extent_buffer *l)
 	struct btrfs_extent_data_ref *dref;
 	struct btrfs_shared_data_ref *sref;
 	struct btrfs_inode_ref *iref;
+	struct btrfs_inode_extref *iref2;
 	struct btrfs_dev_extent *dev_extent;
 	struct btrfs_disk_key disk_key;
 	struct btrfs_root_item root_item;
@@ -492,6 +532,10 @@ void btrfs_print_leaf(struct btrfs_root *root, struct
extent_buffer *l)
 			iref = btrfs_item_ptr(l, i, struct btrfs_inode_ref);
 			print_inode_ref_item(l, item, iref);
 			break;
+		case BTRFS_INODE_EXTREF_KEY:
+			iref2 = btrfs_item_ptr(l, i, struct btrfs_inode_extref);
+			print_inode_extref_item(l, item, iref2);
+			break;
 		case BTRFS_DIR_ITEM_KEY:
 		case BTRFS_DIR_INDEX_KEY:
 		case BTRFS_XATTR_ITEM_KEY:
-- 
1.7.7

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

This patch adds basic support for extended inode refs. This includes support
for link and unlink of the refs, which basically gets us support for rename
as well.

Inode creation does not need changing - extended refs are only added after
the ref array is full.

Signed-off-by: Mark Fasheh <mfasheh@suse.de>
---
 fs/btrfs/ctree.h      |   52 ++++++++--
 fs/btrfs/hash.h       |   10 ++
 fs/btrfs/inode-item.c |  281 +++++++++++++++++++++++++++++++++++++++++++++++--
 fs/btrfs/inode.c      |   23 +++--
 4 files changed, 339 insertions(+), 27 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index fa5c45b..9fa4ebe 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -151,6 +151,13 @@ struct btrfs_ordered_sum;
  */
 #define BTRFS_NAME_LEN 255
 
+/*
+ * Theoretical limit is larger, but we keep this down to a sane
+ * value. That should limit greatly the possibility of collisions on
+ * inode ref items.
+ */
+#define BTRFS_LINK_MAX 65535U
+
 /* 32 bytes in various csum fields */
 #define BTRFS_CSUM_SIZE 32
 
@@ -486,6 +493,8 @@ struct btrfs_super_block {
  */
 #define BTRFS_FEATURE_INCOMPAT_BIG_METADATA	(1ULL << 5)
 
+#define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF	(1ULL << 6)
+
 #define BTRFS_FEATURE_COMPAT_SUPP		0ULL
 #define BTRFS_FEATURE_COMPAT_RO_SUPP		0ULL
 #define BTRFS_FEATURE_INCOMPAT_SUPP			\
@@ -493,7 +502,8 @@ struct btrfs_super_block {
 	 BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL |	\
 	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS |		\
 	 BTRFS_FEATURE_INCOMPAT_BIG_METADATA |		\
-	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO)
+	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO |		\
+	 BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF)
 
 /*
  * A leaf is full of items. offset and size tell us where to find
@@ -640,6 +650,14 @@ struct btrfs_inode_ref {
 	/* name goes here */
 } __attribute__ ((__packed__));
 
+struct btrfs_inode_extref {
+	__le64 parent_objectid;
+	__le64 index;
+	__le16 name_len;
+	__u8   name[0];
+	/* name goes here */
+} __attribute__ ((__packed__));
+
 struct btrfs_timespec {
 	__le64 sec;
 	__le32 nsec;
@@ -1457,6 +1475,7 @@ struct btrfs_ioctl_defrag_range_args {
  */
 #define BTRFS_INODE_ITEM_KEY		1
 #define BTRFS_INODE_REF_KEY		12
+#define BTRFS_INODE_EXTREF_KEY		13
 #define BTRFS_XATTR_ITEM_KEY		24
 #define BTRFS_ORPHAN_ITEM_KEY		48
 /* reserve 2-15 close to the inode for later flexibility */
@@ -1778,6 +1797,13 @@ BTRFS_SETGET_STACK_FUNCS(block_group_flags,
 BTRFS_SETGET_FUNCS(inode_ref_name_len, struct btrfs_inode_ref, name_len, 16);
 BTRFS_SETGET_FUNCS(inode_ref_index, struct btrfs_inode_ref, index, 64);
 
+/* struct btrfs_inode_extref */
+BTRFS_SETGET_FUNCS(inode_extref_parent, struct btrfs_inode_extref,
+		   parent_objectid, 64);
+BTRFS_SETGET_FUNCS(inode_extref_name_len, struct btrfs_inode_extref,
+		   name_len, 16);
+BTRFS_SETGET_FUNCS(inode_extref_index, struct btrfs_inode_extref, index, 64);
+
 /* struct btrfs_inode_item */
 BTRFS_SETGET_FUNCS(inode_generation, struct btrfs_inode_item, generation, 64);
 BTRFS_SETGET_FUNCS(inode_sequence, struct btrfs_inode_item, sequence, 64);
@@ -2884,12 +2910,12 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle
*trans,
 			   struct btrfs_root *root,
 			   const char *name, int name_len,
 			   u64 inode_objectid, u64 ref_objectid, u64 *index);
-struct btrfs_inode_ref *
-btrfs_lookup_inode_ref(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root,
-			struct btrfs_path *path,
-			const char *name, int name_len,
-			u64 inode_objectid, u64 ref_objectid, int mod);
+int btrfs_get_inode_ref_index(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      const char *name, int name_len,
+			      u64 inode_objectid, u64 ref_objectid, int mod,
+			      u64 *ret_index);
 int btrfs_insert_empty_inode(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     struct btrfs_path *path, u64 objectid);
@@ -2897,6 +2923,18 @@ int btrfs_lookup_inode(struct btrfs_trans_handle *trans,
struct btrfs_root
 		       *root, struct btrfs_path *path,
 		       struct btrfs_key *location, int mod);
 
+struct btrfs_inode_extref *
+btrfs_lookup_inode_extref(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  struct btrfs_path *path,
+			  const char *name, int name_len,
+			  u64 inode_objectid, u64 ref_objectid, int ins_len,
+			  int cow);
+
+int btrfs_find_name_in_ext_backref(struct btrfs_path *path, const char *name,
+				   int name_len,
+				   struct btrfs_inode_extref **extref_ret);
+
 /* file-item.c */
 int btrfs_del_csums(struct btrfs_trans_handle *trans,
 		    struct btrfs_root *root, u64 bytenr, u64 len);
diff --git a/fs/btrfs/hash.h b/fs/btrfs/hash.h
index db2ff97..1d98281 100644
--- a/fs/btrfs/hash.h
+++ b/fs/btrfs/hash.h
@@ -24,4 +24,14 @@ static inline u64 btrfs_name_hash(const char *name, int len)
 {
 	return crc32c((u32)~1, name, len);
 }
+
+/*
+ * Figure the key offset of an extended inode ref
+ */
+static inline u64 btrfs_extref_hash(u64 parent_objectid, const char *name,
+				    int len)
+{
+	return (u64) crc32c(parent_objectid, name, len);
+}
+
 #endif
diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c
index a13cf1a..a64f74b 100644
--- a/fs/btrfs/inode-item.c
+++ b/fs/btrfs/inode-item.c
@@ -18,6 +18,7 @@
 
 #include "ctree.h"
 #include "disk-io.h"
+#include "hash.h"
 #include "transaction.h"
 #include "print-tree.h"
 
@@ -50,18 +51,56 @@ static int find_name_in_backref(struct btrfs_path *path,
const char *name,
 	return 0;
 }
 
-struct btrfs_inode_ref *
+int btrfs_find_name_in_ext_backref(struct btrfs_path *path, const char *name,
+				   int name_len,
+				   struct btrfs_inode_extref **extref_ret)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_inode_extref *extref;
+	unsigned long ptr;
+	unsigned long name_ptr;
+	u32 item_size;
+	u32 cur_offset = 0;
+	int ref_name_len;
+
+	leaf = path->nodes[0];
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+
+	/*
+	 * Search all extended backrefs in this item. We''re only
+	 * looking through any collisions so most of the time this is
+	 * just going to compare against one buffer. If all is well,
+	 * we''ll return success and the inode ref object.
+	 */
+	while (cur_offset < item_size) {
+		extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
+		name_ptr = (unsigned long)(&extref->name);
+		ref_name_len = btrfs_inode_extref_name_len(leaf, extref);
+
+		if (ref_name_len == name_len
+		    && (memcmp_extent_buffer(leaf, name, name_ptr, name_len) == 0)) {
+			if (extref_ret)
+				*extref_ret = extref;
+			return 1;
+		}
+
+		cur_offset += ref_name_len + sizeof(*extref);
+	}
+	return 0;
+}
+
+static struct btrfs_inode_ref *
 btrfs_lookup_inode_ref(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root,
-			struct btrfs_path *path,
-			const char *name, int name_len,
-			u64 inode_objectid, u64 ref_objectid, int mod)
+		       struct btrfs_root *root,
+		       struct btrfs_path *path,
+		       const char *name, int name_len,
+		       u64 inode_objectid, u64 ref_objectid, int ins_len,
+		       int cow)
 {
+	int ret;
 	struct btrfs_key key;
 	struct btrfs_inode_ref *ref;
-	int ins_len = mod < 0 ? -1 : 0;
-	int cow = mod != 0;
-	int ret;
 
 	key.objectid = inode_objectid;
 	key.type = BTRFS_INODE_REF_KEY;
@@ -77,13 +116,149 @@ btrfs_lookup_inode_ref(struct btrfs_trans_handle *trans,
 	return ref;
 }
 
-int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
+/* Returns NULL if no extref found */
+struct btrfs_inode_extref *
+btrfs_lookup_inode_extref(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  struct btrfs_path *path,
+			  const char *name, int name_len,
+			  u64 inode_objectid, u64 ref_objectid, int ins_len,
+			  int cow)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_inode_extref *extref;
+
+	key.objectid = inode_objectid;
+	key.type = BTRFS_INODE_EXTREF_KEY;
+	key.offset = btrfs_extref_hash(ref_objectid, name, name_len);
+
+	ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
+	if (ret < 0)
+		return ERR_PTR(ret);
+	if (ret > 0)
+		return NULL;
+	if (!btrfs_find_name_in_ext_backref(path, name, name_len, &extref))
+		return NULL;
+	return extref;
+}
+
+int btrfs_get_inode_ref_index(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      const char *name, int name_len,
+			      u64 inode_objectid, u64 ref_objectid, int mod,
+			      u64 *ret_index)
+{
+	struct btrfs_inode_ref *ref;
+	struct btrfs_inode_extref *extref;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+
+	ref = btrfs_lookup_inode_ref(trans, root, path, name, name_len,
+				     inode_objectid, ref_objectid, ins_len,
+				     cow);
+	if (IS_ERR(ref))
+		return PTR_ERR(ref);
+
+	if (ref != NULL) {
+		*ret_index = btrfs_inode_ref_index(path->nodes[0], ref);
+		return 0;
+	}
+
+	btrfs_release_path(path);
+
+	extref = btrfs_lookup_inode_extref(trans, root, path, name,
+					   name_len, inode_objectid,
+					   ref_objectid, ins_len, cow);
+	if (IS_ERR(extref))
+		return PTR_ERR(extref);
+
+	if (extref) {
+		*ret_index = btrfs_inode_extref_index(path->nodes[0], extref);
+		return 0;
+	}
+
+	return -ENOENT;
+}
+
+int btrfs_del_inode_extref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   const char *name, int name_len,
 			   u64 inode_objectid, u64 ref_objectid, u64 *index)
 {
 	struct btrfs_path *path;
 	struct btrfs_key key;
+	struct btrfs_inode_extref *extref;
+	struct extent_buffer *leaf;
+	int ret;
+	int del_len = name_len + sizeof(*extref);
+	unsigned long ptr;
+	unsigned long item_start;
+	u32 item_size;
+
+	key.objectid = inode_objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
+	key.offset = btrfs_extref_hash(ref_objectid, name, name_len);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->leave_spinning = 1;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret > 0)
+		ret = -ENOENT;
+	if (ret < 0)
+		goto out;
+
+	/*
+	 * Sanity check - did we find the right item for this name?
+	 * This should always succeed so error here will make the FS
+	 * readonly.
+	 */
+	if (!btrfs_find_name_in_ext_backref(path, name, name_len, &extref)) {
+		btrfs_std_error(root->fs_info, -ENOENT);
+		ret = -EROFS;
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+	if (index)
+		*index = btrfs_inode_extref_index(leaf, extref);
+
+	if (del_len == item_size) {
+		/*
+		 * Common case only one ref in the item, remove the
+		 * whole item.
+		 */
+		ret = btrfs_del_item(trans, root, path);
+		goto out;
+	}
+
+	ptr = (unsigned long)extref;
+	item_start = btrfs_item_ptr_offset(leaf, path->slots[0]);
+
+	memmove_extent_buffer(leaf, ptr, ptr + del_len,
+			      item_size - (ptr + del_len - item_start));
+
+	btrfs_truncate_item(trans, root, path, item_size - del_len, 1);
+
+out:
+	btrfs_free_path(path);
+
+	return ret;
+}
+
+int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			const char *name, int name_len,
+			u64 inode_objectid, u64 ref_objectid, u64 *index)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
 	struct btrfs_inode_ref *ref;
 	struct extent_buffer *leaf;
 	unsigned long ptr;
@@ -91,6 +266,7 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 	u32 item_size;
 	u32 sub_item_len;
 	int ret;
+	int search_ext_refs = 0;
 	int del_len = name_len + sizeof(*ref);
 
 	key.objectid = inode_objectid;
@@ -106,12 +282,14 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 	if (ret > 0) {
 		ret = -ENOENT;
+		search_ext_refs = 1;
 		goto out;
 	} else if (ret < 0) {
 		goto out;
 	}
 	if (!find_name_in_backref(path, name, name_len, &ref)) {
 		ret = -ENOENT;
+		search_ext_refs = 1;
 		goto out;
 	}
 	leaf = path->nodes[0];
@@ -129,8 +307,76 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 	item_start = btrfs_item_ptr_offset(leaf, path->slots[0]);
 	memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
 			      item_size - (ptr + sub_item_len - item_start));
-	btrfs_truncate_item(trans, root, path,
-				  item_size - sub_item_len, 1);
+	btrfs_truncate_item(trans, root, path, item_size - sub_item_len, 1);
+out:
+	btrfs_free_path(path);
+
+	if (search_ext_refs) {
+		/*
+		 * No refs were found, or we could not find the
+		 * name in our ref array. Find and remove the extended
+		 * inode ref then.
+		 */
+		return btrfs_del_inode_extref(trans, root, name, name_len,
+					      inode_objectid, ref_objectid, index);
+	}
+
+	return ret;
+}
+
+/*
+ * btrfs_insert_inode_extref() - Inserts an extended inode ref into a tree.
+ *
+ * The caller must have checked against BTRFS_LINK_MAX already.
+ */
+static int btrfs_insert_inode_extref(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root,
+				     const char *name, int name_len,
+				     u64 inode_objectid, u64 ref_objectid, u64 index)
+{
+	struct btrfs_inode_extref *extref;
+	int ret;
+	int ins_len = name_len + sizeof(*extref);
+	unsigned long ptr;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+
+	key.objectid = inode_objectid;
+	key.type = BTRFS_INODE_EXTREF_KEY;
+	key.offset = btrfs_extref_hash(ref_objectid, name, name_len);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->leave_spinning = 1;
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      ins_len);
+	if (ret == -EEXIST) {
+		if (btrfs_find_name_in_ext_backref(path, name, name_len, NULL))
+			goto out;
+
+		btrfs_extend_item(trans, root, path, ins_len);
+	}
+	if (ret < 0)
+		goto out;
+
+	leaf = path->nodes[0];
+	item = btrfs_item_nr(leaf, path->slots[0]);
+	ptr = (unsigned long)btrfs_item_ptr(leaf, path->slots[0], char);
+	ptr += btrfs_item_size(leaf, item) - ins_len;
+	extref = (struct btrfs_inode_extref *)ptr;
+
+	btrfs_set_inode_extref_name_len(path->nodes[0], extref, name_len);
+	btrfs_set_inode_extref_index(path->nodes[0], extref, index);
+	btrfs_set_inode_extref_parent(path->nodes[0], extref, ref_objectid);
+
+	ptr = (unsigned long)&extref->name;
+	write_extent_buffer(path->nodes[0], name, ptr, name_len);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -191,6 +437,19 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle
*trans,
 
 out:
 	btrfs_free_path(path);
+
+	if (ret == -EMLINK) {
+		struct btrfs_super_block *disk_super = root->fs_info->super_copy;
+		/* We ran out of space in the ref array. Need to
+		 * add an extended ref. */
+		if (btrfs_super_incompat_flags(disk_super)
+		    & BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF)
+			ret = btrfs_insert_inode_extref(trans, root, name,
+							name_len,
+							inode_objectid,
+							ref_objectid, index);
+	}
+
 	return ret;
 }
 
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index a7d1921..b96f0dc 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2890,7 +2890,6 @@ static struct btrfs_trans_handle
*__unlink_start_trans(struct inode *dir,
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
 	struct btrfs_path *path;
-	struct btrfs_inode_ref *ref;
 	struct btrfs_dir_item *di;
 	struct inode *inode = dentry->d_inode;
 	u64 index;
@@ -3004,17 +3003,17 @@ static struct btrfs_trans_handle
*__unlink_start_trans(struct inode *dir,
 	}
 	btrfs_release_path(path);
 
-	ref = btrfs_lookup_inode_ref(trans, root, path,
-				dentry->d_name.name, dentry->d_name.len,
-				ino, dir_ino, 0);
-	if (IS_ERR(ref)) {
-		err = PTR_ERR(ref);
+	ret = btrfs_get_inode_ref_index(trans, root, path, dentry->d_name.name,
+					dentry->d_name.len, ino, dir_ino, 0,
+					&index);
+	if (ret) {
+		err = ret;
 		goto out;
 	}
-	BUG_ON(!ref); /* Logic error */
+
 	if (check_path_shared(root, path))
 		goto out;
-	index = btrfs_inode_ref_index(path->nodes[0], ref);
+
 	btrfs_release_path(path);
 
 	/*
@@ -4673,6 +4672,12 @@ static struct inode *btrfs_new_inode(struct
btrfs_trans_handle *trans,
 	btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
 	key[0].offset = 0;
 
+	/*
+	 * Start new inodes with an inode_ref. This is slightly more
+	 * efficient for small numbers of hard links since they will
+	 * be packed into one item. Extended refs will kick in if we
+	 * add more hard links than can fit in the ref item.
+	 */
 	key[1].objectid = objectid;
 	btrfs_set_key_type(&key[1], BTRFS_INODE_REF_KEY);
 	key[1].offset = ref_objectid;
@@ -4975,7 +4980,7 @@ static int btrfs_link(struct dentry *old_dentry, struct
inode *dir,
 	if (root->objectid != BTRFS_I(inode)->root->objectid)
 		return -EXDEV;
 
-	if (inode->i_nlink == ~0U)
+	if (inode->i_nlink >= BTRFS_LINK_MAX)
 		return -EMLINK;
 
 	err = btrfs_set_inode_index(dir, &index);
-- 
1.7.7

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Teach tree-log.c about extended inode refs. In particular, we have to adjust
the behavior of inode ref replay as well as log tree recovery to account for
the existence of extended refs.

Signed-off-by: Mark Fasheh <mfasheh@suse.de>
---
 fs/btrfs/backref.c  |   68 ++++++++++++
 fs/btrfs/backref.h  |    5 +
 fs/btrfs/tree-log.c |  297 ++++++++++++++++++++++++++++++++++++++++++---------
 3 files changed, 319 insertions(+), 51 deletions(-)

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index a383c18..658e09c 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -1111,6 +1111,74 @@ static int inode_ref_info(u64 inum, u64 ioff, struct
btrfs_root *fs_root,
 				found_key);
 }
 
+int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
+			  u64 start_off, struct btrfs_path *path,
+			  struct btrfs_inode_extref **ret_extref,
+			  u64 *found_off)
+{
+	int ret, slot;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_inode_extref *extref;
+	struct extent_buffer *leaf;
+	unsigned long ptr;
+
+	key.objectid = inode_objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
+	key.offset = start_off;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		return ret;
+
+	while (1) {
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(leaf)) {
+			/*
+			 * If the item at offset is not found,
+			 * btrfs_search_slot will point us to the slot
+			 * where it should be inserted. In our case
+			 * that will be the slot directly before the
+			 * next INODE_REF_KEY_V2 item. In the case
+			 * that we''re pointing to the last slot in a
+			 * leaf, we must move one leaf over.
+			 */
+			ret = btrfs_next_leaf(root, path);
+			if (ret) {
+				if (ret >= 1)
+					ret = -ENOENT;
+				break;
+			}
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		/*
+		 * Check that we''re still looking at an extended ref key for
+		 * this particular objectid. If we have different
+		 * objectid or type then there are no more to be found
+		 * in the tree and we can exit.
+		 */
+		ret = -ENOENT;
+		if (found_key.objectid != inode_objectid)
+			break;
+		if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
+			break;
+
+		ret = 0;
+		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+		extref = (struct btrfs_inode_extref *)ptr;
+		*ret_extref = extref;
+		if (found_off)
+			*found_off = found_key.offset;
+		break;
+	}
+
+	return ret;
+}
+
 /*
  * this iterates to turn a btrfs_inode_ref into a full filesystem path.
elements
  * of the path are separated by ''/'' and the path is
guaranteed to be
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index c18d8ac..9f3e251 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -66,4 +66,9 @@ struct inode_fs_paths *init_ipath(s32 total_bytes, struct
btrfs_root *fs_root,
 					struct btrfs_path *path);
 void free_ipath(struct inode_fs_paths *ipath);
 
+int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
+			  u64 start_off, struct btrfs_path *path,
+			  struct btrfs_inode_extref **ret_extref,
+			  u64 *found_off);
+
 #endif
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 8abeae4..e5ba0a4 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -23,8 +23,10 @@
 #include "disk-io.h"
 #include "locking.h"
 #include "print-tree.h"
+#include "backref.h"
 #include "compat.h"
 #include "tree-log.h"
+#include "hash.h"
 
 /* magic values for the inode_only field in btrfs_log_inode:
  *
@@ -764,8 +766,16 @@ static noinline int backref_in_log(struct btrfs_root *log,
 	if (ret != 0)
 		goto out;
 
-	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
 	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+
+	if (key->type == BTRFS_INODE_EXTREF_KEY) {
+		if (btrfs_find_name_in_ext_backref(path, name, namelen, NULL))
+			match = 1;
+
+		goto out;
+	}
+
+	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
 	ptr_end = ptr + item_size;
 	while (ptr < ptr_end) {
 		ref = (struct btrfs_inode_ref *)ptr;
@@ -786,6 +796,47 @@ out:
 	return match;
 }
 
+static int extref_get_fields(struct extent_buffer *eb, int slot,
+			     u32 *namelen, char **name, u64 *index,
+			     u64 *parent_objectid)
+{
+	struct btrfs_inode_extref *extref;
+
+	extref = (struct btrfs_inode_extref *)btrfs_item_ptr_offset(eb, slot);
+
+	*namelen = btrfs_inode_extref_name_len(eb, extref);
+	*name = kmalloc(*namelen, GFP_NOFS);
+	if (*name == NULL)
+		return -ENOMEM;
+
+	read_extent_buffer(eb, *name, (unsigned long)&extref->name,
+			   *namelen);
+
+	*index = btrfs_inode_extref_index(eb, extref);
+	if (parent_objectid)
+		*parent_objectid = btrfs_inode_extref_parent(eb, extref);
+
+	return 0;
+}
+
+static int ref_get_fields(struct extent_buffer *eb, int slot, u32 *namelen,
+			  char **name, u64 *index)
+{
+	struct btrfs_inode_ref *ref;
+
+	ref = (struct btrfs_inode_ref *)btrfs_item_ptr_offset(eb, slot);
+
+	*namelen = btrfs_inode_ref_name_len(eb, ref);
+	*name = kmalloc(*namelen, GFP_NOFS);	
+	if (*name == NULL)
+		return -ENOMEM;
+
+	read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);
+
+	*index = btrfs_inode_ref_index(eb, ref);
+
+	return 0;
+}
 
 /*
  * replay one inode back reference item found in the log tree.
@@ -801,15 +852,50 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
 				  struct btrfs_key *key)
 {
 	struct btrfs_inode_ref *ref;
+	struct btrfs_inode_extref *extref;
 	struct btrfs_dir_item *di;
+	struct btrfs_key search_key;
 	struct inode *dir;
 	struct inode *inode;
 	unsigned long ref_ptr;
 	unsigned long ref_end;
 	char *name;
+	char *victim_name;
 	int namelen;
+	int victim_name_len;
 	int ret;
 	int search_done = 0;
+	int log_ref_ver = 0;
+	u64 parent_objectid;
+	u64 inode_objectid;
+	u64 ref_index;
+	struct extent_buffer *leaf;
+	int ref_struct_size;
+
+
+	ref_ptr = btrfs_item_ptr_offset(eb, slot);
+	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
+
+	if (key->type == BTRFS_INODE_EXTREF_KEY) {	
+		ref_struct_size = sizeof(*extref);
+		log_ref_ver = 1;
+
+		ret = extref_get_fields(eb, slot, &namelen, &name, &ref_index,
+					&parent_objectid);
+		if (ret)
+			return ret;
+	} else {
+		ref_struct_size = sizeof(*ref);
+
+		ret = ref_get_fields(eb, slot, &namelen, &name, &ref_index);
+		if (ret)
+			return ret;
+
+
+		parent_objectid = key->offset;
+	}
+
+	inode_objectid = key->objectid;
 
 	/*
 	 * it is possible that we didn''t log all the parent directories
@@ -817,32 +903,20 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
 	 * copy the back ref in.  The link count fixup code will take
 	 * care of the rest
 	 */
-	dir = read_one_inode(root, key->offset);
+	dir = read_one_inode(root, parent_objectid);
 	if (!dir)
 		return -ENOENT;
 
-	inode = read_one_inode(root, key->objectid);
+	inode = read_one_inode(root, inode_objectid);
 	if (!inode) {
 		iput(dir);
 		return -EIO;
 	}
 
-	ref_ptr = btrfs_item_ptr_offset(eb, slot);
-	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
-
 again:
-	ref = (struct btrfs_inode_ref *)ref_ptr;
-
-	namelen = btrfs_inode_ref_name_len(eb, ref);
-	name = kmalloc(namelen, GFP_NOFS);
-	BUG_ON(!name);
-
-	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
-
 	/* if we already have a perfect match, we''re done */
 	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
-			 btrfs_inode_ref_index(eb, ref),
-			 name, namelen)) {
+			 ref_index, name, namelen)) {
 		goto out;
 	}
 
@@ -857,19 +931,23 @@ again:
 	if (search_done)
 		goto insert;
 
-	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+	/* Search old style refs */
+	search_key.objectid = inode_objectid;
+	search_key.type = BTRFS_INODE_REF_KEY;
+	search_key.offset = parent_objectid;
+
+	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
 	if (ret == 0) {
-		char *victim_name;
-		int victim_name_len;
 		struct btrfs_inode_ref *victim_ref;
 		unsigned long ptr;
 		unsigned long ptr_end;
-		struct extent_buffer *leaf = path->nodes[0];
+
+		leaf = path->nodes[0];
 
 		/* are we trying to overwrite a back ref for the root directory
 		 * if so, just jump out, we''re done
 		 */
-		if (key->objectid == key->offset)
+		if (search_key.objectid == search_key.offset)
 			goto out_nowrite;
 
 		/* check all the names in this back reference to see
@@ -889,7 +967,7 @@ again:
 					   (unsigned long)(victim_ref + 1),
 					   victim_name_len);
 
-			if (!backref_in_log(log, key, victim_name,
+			if (!backref_in_log(log, &search_key, victim_name,
 					    victim_name_len)) {
 				btrfs_inc_nlink(inode);
 				btrfs_release_path(path);
@@ -903,19 +981,61 @@ again:
 			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
 		}
 		BUG_ON(ret);
-
 		/*
 		 * NOTE: we have searched root tree and checked the
-		 * coresponding ref, it does not need to check again.
+		 * coresponding refs, it does not need to be checked again.
 		 */
 		search_done = 1;
 	}
 	btrfs_release_path(path);
 
+	/* Same search but for extended refs */
+	extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
+					   inode_objectid, parent_objectid, 0,
+					   0);
+	if (!IS_ERR_OR_NULL(extref)) {
+		u32 item_size;
+		u32 cur_offset = 0;
+		unsigned long base;
+
+		leaf = path->nodes[0];
+
+		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+		base = btrfs_item_ptr_offset(leaf, path->slots[0]);
+
+		while (cur_offset < item_size) {
+			extref = (struct btrfs_inode_extref *)base + cur_offset;
+
+			victim_name_len = btrfs_inode_extref_name_len(eb, extref);
+			victim_name = kmalloc(namelen, GFP_NOFS);
+			leaf = path->nodes[0];
+			read_extent_buffer(eb, name, (unsigned long)&extref->name, namelen);
+
+			search_key.objectid = inode_objectid;
+			search_key.type = BTRFS_INODE_EXTREF_KEY;
+			search_key.offset = btrfs_extref_hash(parent_objectid,
+							      name, namelen);
+			if (!backref_in_log(log, &search_key, victim_name,
+					    victim_name_len)) {
+				btrfs_inc_nlink(inode);
+				btrfs_release_path(path);
+
+				ret = btrfs_unlink_inode(trans, root, dir,
+							 inode, victim_name,
+							 victim_name_len);
+			}
+			kfree(victim_name);
+			BUG_ON(ret);
+
+			cur_offset += victim_name_len + sizeof(*extref);
+		}
+		search_done = 1;
+	}
+
+
 	/* look for a conflicting sequence number */
 	di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
-					 btrfs_inode_ref_index(eb, ref),
-					 name, namelen, 0);
+					 ref_index, name, namelen, 0);
 	if (di && !IS_ERR(di)) {
 		ret = drop_one_dir_item(trans, root, path, dir, di);
 		BUG_ON(ret);
@@ -933,17 +1053,25 @@ again:
 
 insert:
 	/* insert our name */
-	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
-			     btrfs_inode_ref_index(eb, ref));
+	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0, ref_index);
 	BUG_ON(ret);
 
 	btrfs_update_inode(trans, root, inode);
 
 out:
-	ref_ptr = (unsigned long)(ref + 1) + namelen;
+	ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
 	kfree(name);
-	if (ref_ptr < ref_end)
+	if (ref_ptr < ref_end) {
+		if (log_ref_ver) {
+			ret = extref_get_fields(eb, slot, &namelen, &name,
+						&ref_index, NULL);
+		} else {
+			ret = ref_get_fields(eb, slot, &namelen, &name,
+					     &ref_index);
+		}
+		BUG_ON(ret);
 		goto again;
+	}
 
 	/* finally write the back reference in the inode */
 	ret = overwrite_item(trans, root, path, eb, slot, key);
@@ -966,25 +1094,52 @@ static int insert_orphan_item(struct btrfs_trans_handle
*trans,
 	return ret;
 }
 
+static int count_inode_extrefs(struct btrfs_root *root,
+			       struct inode *inode, struct btrfs_path *path)
+{
+	int ret = 0;
+	int name_len;
+	unsigned int nlink = 0;
+	u32 item_size;
+	u32 cur_offset = 0;
+	u64 inode_objectid = btrfs_ino(inode);
+	u64 offset = 0;
+	unsigned long ptr;
+	struct btrfs_inode_extref *extref;
+	struct extent_buffer *leaf;
+
+	while (1) {
+		ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
+					    &extref, &offset);
+		if (ret)
+			break;
 
-/*
- * There are a few corners where the link count of the file can''t
- * be properly maintained during replay.  So, instead of adding
- * lots of complexity to the log code, we just scan the backrefs
- * for any file that has been through replay.
- *
- * The scan will update the link count on the inode to reflect the
- * number of back refs found.  If it goes down to zero, the iput
- * will free the inode.
- */
-static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
-					   struct btrfs_root *root,
-					   struct inode *inode)
+		leaf = path->nodes[0];
+		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+
+		while (cur_offset < item_size) {
+			extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
+			name_len = btrfs_inode_extref_name_len(leaf, extref);
+
+			nlink++;
+
+			cur_offset += name_len + sizeof(*extref);
+		}
+
+		offset++;
+	}
+	btrfs_release_path(path);
+
+	return (ret == 0) ? nlink : ret;
+}
+
+static int count_inode_refs(struct btrfs_root *root,
+			       struct inode *inode, struct btrfs_path *path)
 {
-	struct btrfs_path *path;
 	int ret;
 	struct btrfs_key key;
-	u64 nlink = 0;
+	unsigned int nlink = 0;
 	unsigned long ptr;
 	unsigned long ptr_end;
 	int name_len;
@@ -994,10 +1149,6 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
 	key.type = BTRFS_INODE_REF_KEY;
 	key.offset = (u64)-1;
 
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-
 	while (1) {
 		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 		if (ret < 0)
@@ -1031,6 +1182,45 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
 		btrfs_release_path(path);
 	}
 	btrfs_release_path(path);
+
+	return nlink;
+}
+
+/*
+ * There are a few corners where the link count of the file can''t
+ * be properly maintained during replay.  So, instead of adding
+ * lots of complexity to the log code, we just scan the backrefs
+ * for any file that has been through replay.
+ *
+ * The scan will update the link count on the inode to reflect the
+ * number of back refs found.  If it goes down to zero, the iput
+ * will free the inode.
+ */
+static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
+					   struct btrfs_root *root,
+					   struct inode *inode)
+{
+	struct btrfs_path *path;
+	int ret;
+	u64 nlink = 0;
+	u64 ino = btrfs_ino(inode);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = count_inode_refs(root, inode, path);
+	if (ret < 0)
+		goto out;
+
+	nlink = ret;
+
+	ret = count_inode_extrefs(root, inode, path);
+	if (ret < 0)
+		goto out;
+
+	nlink += ret;
+
 	if (nlink != inode->i_nlink) {
 		set_nlink(inode, nlink);
 		btrfs_update_inode(trans, root, inode);
@@ -1046,9 +1236,10 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
 		ret = insert_orphan_item(trans, root, ino);
 		BUG_ON(ret);
 	}
-	btrfs_free_path(path);
 
-	return 0;
+out:
+	btrfs_free_path(path);
+	return ret;
 }
 
 static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
@@ -1695,6 +1886,10 @@ static int replay_one_buffer(struct btrfs_root *log,
struct extent_buffer *eb,
 			ret = add_inode_ref(wc->trans, root, log, path,
 					    eb, i, &key);
 			BUG_ON(ret && ret != -ENOENT);
+		} else if (key.type == BTRFS_INODE_EXTREF_KEY) {
+			ret = add_inode_ref(wc->trans, root, log, path,
+					    eb, i, &key);
+			BUG_ON(ret && ret != -ENOENT);
 		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
 			ret = replay_one_extent(wc->trans, root, path,
 						eb, i, &key);
-- 
1.7.7

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

The iterate_irefs in backref.c is used to build path components from inode
refs. This patch adds code to iterate extended refs as well.

I had modify the callback function signature to abstract out some of the
differences between ref structures. iref_to_path() also needed similar
changes.

Signed-off-by: Mark Fasheh <mfasheh@suse.de>
---
 fs/btrfs/backref.c |  134 +++++++++++++++++++++++++++++++++++++++++++---------
 fs/btrfs/backref.h |    2 -
 2 files changed, 112 insertions(+), 24 deletions(-)

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 658e09c..4a01f7c 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -1194,11 +1194,10 @@ int btrfs_find_one_extref(struct btrfs_root *root, u64
inode_objectid,
  * value will be smaller than dest. callers must check this!
  */
 static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
-				struct btrfs_inode_ref *iref,
-				struct extent_buffer *eb_in, u64 parent,
-				char *dest, u32 size)
+			  u32 name_len, unsigned long name_off,
+			  struct extent_buffer *eb_in, u64 parent,
+			  char *dest, u32 size)
 {
-	u32 len;
 	int slot;
 	u64 next_inum;
 	int ret;
@@ -1206,17 +1205,17 @@ static char *iref_to_path(struct btrfs_root *fs_root,
struct btrfs_path *path,
 	struct extent_buffer *eb = eb_in;
 	struct btrfs_key found_key;
 	int leave_spinning = path->leave_spinning;
+	struct btrfs_inode_ref *iref;
 
 	if (bytes_left >= 0)
 		dest[bytes_left] = ''\0'';
 
 	path->leave_spinning = 1;
 	while (1) {
-		len = btrfs_inode_ref_name_len(eb, iref);
-		bytes_left -= len;
+		bytes_left -= name_len;
 		if (bytes_left >= 0)
 			read_extent_buffer(eb, dest + bytes_left,
-						(unsigned long)(iref + 1), len);
+					   name_off, name_len);
 		if (eb != eb_in) {
 			btrfs_tree_read_unlock_blocking(eb);
 			free_extent_buffer(eb);
@@ -1226,6 +1225,7 @@ static char *iref_to_path(struct btrfs_root *fs_root,
struct btrfs_path *path,
 			ret = -ENOENT;
 		if (ret)
 			break;
+
 		next_inum = found_key.offset;
 
 		/* regular exit ahead */
@@ -1241,8 +1241,11 @@ static char *iref_to_path(struct btrfs_root *fs_root,
struct btrfs_path *path,
 			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
 		}
 		btrfs_release_path(path);
-
 		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
+
+		name_len = btrfs_inode_ref_name_len(eb, iref);
+		name_off = (unsigned long)(iref + 1);
+
 		parent = next_inum;
 		--bytes_left;
 		if (bytes_left >= 0)
@@ -1531,9 +1534,12 @@ int iterate_inodes_from_logical(u64 logical, struct
btrfs_fs_info *fs_info,
 	return ret;
 }
 
-static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
-				struct btrfs_path *path,
-				iterate_irefs_t *iterate, void *ctx)
+typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off,
+			      struct extent_buffer *eb, void *ctx);
+
+static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
+			      struct btrfs_path *path,
+			      iterate_irefs_t *iterate, void *ctx)
 {
 	int ret = 0;
 	int slot;
@@ -1550,7 +1556,7 @@ static int iterate_irefs(u64 inum, struct btrfs_root
*fs_root,
 	while (!ret) {
 		path->leave_spinning = 1;
 		ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
-					&found_key);
+				     &found_key);
 		if (ret < 0)
 			break;
 		if (ret) {
@@ -1578,7 +1584,8 @@ static int iterate_irefs(u64 inum, struct btrfs_root
*fs_root,
 				 "tree %llu\n", cur,
 				 (unsigned long long)found_key.objectid,
 				 (unsigned long long)fs_root->objectid);
-			ret = iterate(parent, iref, eb, ctx);
+			ret = iterate(parent, name_len,
+				      (unsigned long)(iref + 1), eb, ctx);
 			if (ret)
 				break;
 			len = sizeof(*iref) + name_len;
@@ -1593,12 +1600,98 @@ static int iterate_irefs(u64 inum, struct btrfs_root
*fs_root,
 	return ret;
 }
 
+static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
+				 struct btrfs_path *path,
+				 iterate_irefs_t *iterate, void *ctx)
+{
+	int ret;
+	int slot;
+	u64 offset = 0;
+	u64 parent;
+	int found = 0;
+	struct extent_buffer *eb;
+	struct btrfs_inode_extref *extref;
+	struct extent_buffer *leaf;
+	u32 item_size;
+	u32 cur_offset;
+	unsigned long ptr;
+
+	while (1) {
+		ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
+					    &offset);
+		if (ret < 0)
+			break;
+		if (ret) {
+			ret = found ? 0 : -ENOENT;
+			break;
+		}
+		++found;
+
+		slot = path->slots[0];
+		eb = path->nodes[0];
+		/* make sure we can use eb after releasing the path */
+		atomic_inc(&eb->refs);
+
+		btrfs_tree_read_lock(eb);
+		btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
+		btrfs_release_path(path);
+
+		leaf = path->nodes[0];
+		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+		cur_offset = 0;
+
+		while (cur_offset < item_size) {
+			u32 name_len;
+
+			extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
+			parent = btrfs_inode_extref_parent(eb, extref);
+			name_len = btrfs_inode_extref_name_len(eb, extref);
+			ret = iterate(parent, name_len,
+				      (unsigned long)&extref->name, eb, ctx);
+			if (ret)
+				break;
+
+			cur_offset += btrfs_inode_extref_name_len(leaf, extref);
+			cur_offset += sizeof(*extref);
+		}
+		btrfs_tree_read_unlock_blocking(eb);
+		free_extent_buffer(eb);
+
+		offset++;
+	}
+
+	btrfs_release_path(path);
+
+	return ret;
+}
+
+static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
+			 struct btrfs_path *path, iterate_irefs_t *iterate,
+			 void *ctx)
+{
+	int ret;
+	int found_refs = 0;
+
+	ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
+	if (!ret)
+		++found_refs;
+	else if (ret != -ENOENT)
+		return ret;
+
+	ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
+	if (ret == -ENOENT && found_refs)
+		return 0;
+
+	return ret;
+}
+
 /*
  * returns 0 if the path could be dumped (probably truncated)
  * returns <0 in case of an error
  */
-static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
-				struct extent_buffer *eb, void *ctx)
+static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
+			 struct extent_buffer *eb, void *ctx)
 {
 	struct inode_fs_paths *ipath = ctx;
 	char *fspath;
@@ -1611,20 +1704,17 @@ static int inode_to_path(u64 inum, struct
btrfs_inode_ref *iref,
 					ipath->fspath->bytes_left - s_ptr : 0;
 
 	fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
-	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
-				inum, fspath_min, bytes_left);
+	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
+			      name_off, eb, inum, fspath_min,
+			      bytes_left);
 	if (IS_ERR(fspath))
 		return PTR_ERR(fspath);
 
 	if (fspath > fspath_min) {
-		pr_debug("path resolved: %s\n", fspath);
 		ipath->fspath->val[i] = (u64)(unsigned long)fspath;
 		++ipath->fspath->elem_cnt;
 		ipath->fspath->bytes_left = fspath - fspath_min;
 	} else {
-		pr_debug("missed path, not enough space. missing bytes: %lu, "
-			 "constructed so far: %s\n",
-			 (unsigned long)(fspath_min - fspath), fspath_min);
 		++ipath->fspath->elem_missed;
 		ipath->fspath->bytes_missing += fspath_min - fspath;
 		ipath->fspath->bytes_left = 0;
@@ -1646,7 +1736,7 @@ static int inode_to_path(u64 inum, struct btrfs_inode_ref
*iref,
 int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
 {
 	return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
-				inode_to_path, ipath);
+			     inode_to_path, ipath);
 }
 
 struct btrfs_data_container *init_data_container(u32 total_bytes)
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index 9f3e251..c80889a 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -32,8 +32,6 @@ struct inode_fs_paths {
 
 typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root,
 		void *ctx);
-typedef int (iterate_irefs_t)(u64 parent, struct btrfs_inode_ref *iref,
-				struct extent_buffer *eb, void *ctx);
 
 int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
 			struct btrfs_path *path);
-- 
1.7.7

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Wed, August 08, 2012 at 20:55 (+0200), Mark Fasheh
wrote:
> +/*
> + * btrfs_insert_inode_extref() - Inserts an extended inode ref into a
tree.
> + *
> + * The caller must have checked against BTRFS_LINK_MAX already.
> + */
> +static int btrfs_insert_inode_extref(struct btrfs_trans_handle *trans,
> +				     struct btrfs_root *root,
> +				     const char *name, int name_len,
> +				     u64 inode_objectid, u64 ref_objectid, u64 index)
> +{
> +	struct btrfs_inode_extref *extref;
> +	int ret;
> +	int ins_len = name_len + sizeof(*extref);
> +	unsigned long ptr;
> +	struct btrfs_path *path;
> +	struct btrfs_key key;
> +	struct extent_buffer *leaf;
> +	struct btrfs_item *item;
> +
> +	key.objectid = inode_objectid;
> +	key.type = BTRFS_INODE_EXTREF_KEY;
> +	key.offset = btrfs_extref_hash(ref_objectid, name, name_len);
> +
> +	path = btrfs_alloc_path();
> +	if (!path)
> +		return -ENOMEM;
> +
> +	path->leave_spinning = 1;
> +	ret = btrfs_insert_empty_item(trans, root, path, &key,
> +				      ins_len);
> +	if (ret == -EEXIST) {
> +		if (btrfs_find_name_in_ext_backref(path, name, name_len, NULL))
> +			goto out;
> +
> +		btrfs_extend_item(trans, root, path, ins_len);
> +	}
> +	if (ret < 0)
> +		goto out;
This doesn''t look right. Did you actually test it? I haven''t,
but I claim that
with this version of the patch, you won''t be able to add a hash
collision link.

Jeff changed btrfs_extend_item from int to void in March, so we''re no
longer
setting ret there. I suggest adding "ret = 0" within the EEXIST-block.

The rest of this patch looks good to me.

-Jan
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Tue, Aug 14, 2012 at 11:32:43AM +0200, Jan Schmidt
wrote:
> On Wed, August 08, 2012 at 20:55 (+0200), Mark Fasheh wrote:
> > +/*
> > + * btrfs_insert_inode_extref() - Inserts an extended inode ref into a
tree.
> > + *
> > + * The caller must have checked against BTRFS_LINK_MAX already.
> > + */
> > +static int btrfs_insert_inode_extref(struct btrfs_trans_handle
*trans,
> > +				     struct btrfs_root *root,
> > +				     const char *name, int name_len,
> > +				     u64 inode_objectid, u64 ref_objectid, u64 index)
> > +{
> > +	struct btrfs_inode_extref *extref;
> > +	int ret;
> > +	int ins_len = name_len + sizeof(*extref);
> > +	unsigned long ptr;
> > +	struct btrfs_path *path;
> > +	struct btrfs_key key;
> > +	struct extent_buffer *leaf;
> > +	struct btrfs_item *item;
> > +
> > +	key.objectid = inode_objectid;
> > +	key.type = BTRFS_INODE_EXTREF_KEY;
> > +	key.offset = btrfs_extref_hash(ref_objectid, name, name_len);
> > +
> > +	path = btrfs_alloc_path();
> > +	if (!path)
> > +		return -ENOMEM;
> > +
> > +	path->leave_spinning = 1;
> > +	ret = btrfs_insert_empty_item(trans, root, path, &key,
> > +				      ins_len);
> > +	if (ret == -EEXIST) {
> > +		if (btrfs_find_name_in_ext_backref(path, name, name_len, NULL))
> > +			goto out;
> > +
> > +		btrfs_extend_item(trans, root, path, ins_len);
> > +	}
> > +	if (ret < 0)
> > +		goto out;
> 
> This doesn''t look right. Did you actually test it? I
haven''t, but I claim that
> with this version of the patch, you won''t be able to add a hash
collision link.
> 
> Jeff changed btrfs_extend_item from int to void in March, so we''re
no longer
> setting ret there. I suggest adding "ret = 0" within the
EEXIST-block.
Doh, yeah I didn''t test collisions on this version of the patch (as you
can
obviously see, I messed up the merge of Jeff''s change).

> The rest of this patch looks good to me.
Awesome, I''ll send the above as a fix to this patch series (I have at
least
one other fix to send soon too)
	--Mark

--
Mark Fasheh
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Wed, August 08, 2012 at 20:55 (+0200), Mark Fasheh
wrote:
> The iterate_irefs in backref.c is used to build path components from inode
> refs. This patch adds code to iterate extended refs as well.
> 
> I had modify the callback function signature to abstract out some of the
> differences between ref structures. iref_to_path() also needed similar
> changes.
> 
> Signed-off-by: Mark Fasheh <mfasheh@suse.de>
> ---
>  fs/btrfs/backref.c |  134
+++++++++++++++++++++++++++++++++++++++++++---------
>  fs/btrfs/backref.h |    2 -
>  2 files changed, 112 insertions(+), 24 deletions(-)
> 
> diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
> index 658e09c..4a01f7c 100644
> --- a/fs/btrfs/backref.c
> +++ b/fs/btrfs/backref.c
> @@ -1194,11 +1194,10 @@ int btrfs_find_one_extref(struct btrfs_root *root,
u64 inode_objectid,
>   * value will be smaller than dest. callers must check this!
>   */
>  static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path
*path,
> -				struct btrfs_inode_ref *iref,
> -				struct extent_buffer *eb_in, u64 parent,
> -				char *dest, u32 size)
> +			  u32 name_len, unsigned long name_off,
> +			  struct extent_buffer *eb_in, u64 parent,
> +			  char *dest, u32 size)
>  {
> -	u32 len;
>  	int slot;
>  	u64 next_inum;
>  	int ret;
> @@ -1206,17 +1205,17 @@ static char *iref_to_path(struct btrfs_root
*fs_root, struct btrfs_path *path,
>  	struct extent_buffer *eb = eb_in;
>  	struct btrfs_key found_key;
>  	int leave_spinning = path->leave_spinning;
> +	struct btrfs_inode_ref *iref;
>  
>  	if (bytes_left >= 0)
>  		dest[bytes_left] = ''\0'';
>  
>  	path->leave_spinning = 1;
>  	while (1) {
> -		len = btrfs_inode_ref_name_len(eb, iref);
> -		bytes_left -= len;
> +		bytes_left -= name_len;
>  		if (bytes_left >= 0)
>  			read_extent_buffer(eb, dest + bytes_left,
> -						(unsigned long)(iref + 1), len);
> +					   name_off, name_len);
>  		if (eb != eb_in) {
>  			btrfs_tree_read_unlock_blocking(eb);
>  			free_extent_buffer(eb);
> @@ -1226,6 +1225,7 @@ static char *iref_to_path(struct btrfs_root *fs_root,
struct btrfs_path *path,
>  			ret = -ENOENT;
>  		if (ret)
>  			break;
> +
>  		next_inum = found_key.offset;
>  
>  		/* regular exit ahead */
> @@ -1241,8 +1241,11 @@ static char *iref_to_path(struct btrfs_root
*fs_root, struct btrfs_path *path,
>  			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
>  		}
>  		btrfs_release_path(path);
> -
>  		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
> +
> +		name_len = btrfs_inode_ref_name_len(eb, iref);
> +		name_off = (unsigned long)(iref + 1);
> +
>  		parent = next_inum;
>  		--bytes_left;
>  		if (bytes_left >= 0)
> @@ -1531,9 +1534,12 @@ int iterate_inodes_from_logical(u64 logical, struct
btrfs_fs_info *fs_info,
>  	return ret;
>  }
>  
> -static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
> -				struct btrfs_path *path,
> -				iterate_irefs_t *iterate, void *ctx)
> +typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long
name_off,
> +			      struct extent_buffer *eb, void *ctx);
> +
> +static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
> +			      struct btrfs_path *path,
> +			      iterate_irefs_t *iterate, void *ctx)
>  {
>  	int ret = 0;
>  	int slot;
> @@ -1550,7 +1556,7 @@ static int iterate_irefs(u64 inum, struct btrfs_root
*fs_root,
>  	while (!ret) {
>  		path->leave_spinning = 1;
>  		ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
> -					&found_key);
> +				     &found_key);
>  		if (ret < 0)
>  			break;
>  		if (ret) {
> @@ -1578,7 +1584,8 @@ static int iterate_irefs(u64 inum, struct btrfs_root
*fs_root,
>  				 "tree %llu\n", cur,
>  				 (unsigned long long)found_key.objectid,
>  				 (unsigned long long)fs_root->objectid);
> -			ret = iterate(parent, iref, eb, ctx);
> +			ret = iterate(parent, name_len,
> +				      (unsigned long)(iref + 1), eb, ctx);
>  			if (ret)
>  				break;
>  			len = sizeof(*iref) + name_len;
> @@ -1593,12 +1600,98 @@ static int iterate_irefs(u64 inum, struct
btrfs_root *fs_root,
>  	return ret;
>  }
>  
> +static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
> +				 struct btrfs_path *path,
> +				 iterate_irefs_t *iterate, void *ctx)
> +{
> +	int ret;
> +	int slot;
> +	u64 offset = 0;
> +	u64 parent;
> +	int found = 0;
> +	struct extent_buffer *eb;
> +	struct btrfs_inode_extref *extref;
> +	struct extent_buffer *leaf;
> +	u32 item_size;
> +	u32 cur_offset;
> +	unsigned long ptr;
> +
> +	while (1) {
> +		ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
> +					    &offset);
> +		if (ret < 0)
> +			break;
> +		if (ret) {
> +			ret = found ? 0 : -ENOENT;
> +			break;
> +		}
> +		++found;
> +
> +		slot = path->slots[0];
> +		eb = path->nodes[0];
> +		/* make sure we can use eb after releasing the path */
> +		atomic_inc(&eb->refs);
> +
> +		btrfs_tree_read_lock(eb);
> +		btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
> +		btrfs_release_path(path);
> +
> +		leaf = path->nodes[0];
> +		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
> +		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
> +		cur_offset = 0;
> +
> +		while (cur_offset < item_size) {
> +			u32 name_len;
> +
> +			extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
> +			parent = btrfs_inode_extref_parent(eb, extref);
> +			name_len = btrfs_inode_extref_name_len(eb, extref);
> +			ret = iterate(parent, name_len,
> +				      (unsigned long)&extref->name, eb, ctx);
> +			if (ret)
> +				break;
> +
> +			cur_offset += btrfs_inode_extref_name_len(leaf, extref);
> +			cur_offset += sizeof(*extref);
> +		}
> +		btrfs_tree_read_unlock_blocking(eb);
> +		free_extent_buffer(eb);
> +
> +		offset++;
> +	}
> +
> +	btrfs_release_path(path);
> +
> +	return ret;
> +}
> +
> +static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
> +			 struct btrfs_path *path, iterate_irefs_t *iterate,
> +			 void *ctx)
> +{
> +	int ret;
> +	int found_refs = 0;
> +
> +	ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
> +	if (!ret)
> +		++found_refs;
> +	else if (ret != -ENOENT)
> +		return ret;
> +
> +	ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
> +	if (ret == -ENOENT && found_refs)
> +		return 0;
> +
> +	return ret;
> +}
> +
>  /*
>   * returns 0 if the path could be dumped (probably truncated)
>   * returns <0 in case of an error
>   */
> -static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
> -				struct extent_buffer *eb, void *ctx)
> +static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
> +			 struct extent_buffer *eb, void *ctx)
>  {
>  	struct inode_fs_paths *ipath = ctx;
>  	char *fspath;
> @@ -1611,20 +1704,17 @@ static int inode_to_path(u64 inum, struct
btrfs_inode_ref *iref,
>  					ipath->fspath->bytes_left - s_ptr : 0;
>  
>  	fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
> -	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
> -				inum, fspath_min, bytes_left);
> +	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
> +			      name_off, eb, inum, fspath_min,
> +			      bytes_left);
>  	if (IS_ERR(fspath))
>  		return PTR_ERR(fspath);
>  
>  	if (fspath > fspath_min) {
> -		pr_debug("path resolved: %s\n", fspath);
>  		ipath->fspath->val[i] = (u64)(unsigned long)fspath;
>  		++ipath->fspath->elem_cnt;
>  		ipath->fspath->bytes_left = fspath - fspath_min;
>  	} else {
> -		pr_debug("missed path, not enough space. missing bytes: %lu, "
> -			 "constructed so far: %s\n",
> -			 (unsigned long)(fspath_min - fspath), fspath_min);
>  		++ipath->fspath->elem_missed;
>  		ipath->fspath->bytes_missing += fspath_min - fspath;
>  		ipath->fspath->bytes_left = 0;
> @@ -1646,7 +1736,7 @@ static int inode_to_path(u64 inum, struct
btrfs_inode_ref *iref,
>  int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
>  {
>  	return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
> -				inode_to_path, ipath);
> +			     inode_to_path, ipath);
>  }
>  
>  struct btrfs_data_container *init_data_container(u32 total_bytes)
> diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
> index 9f3e251..c80889a 100644
> --- a/fs/btrfs/backref.h
> +++ b/fs/btrfs/backref.h
> @@ -32,8 +32,6 @@ struct inode_fs_paths {
>  
>  typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root,
>  		void *ctx);
> -typedef int (iterate_irefs_t)(u64 parent, struct btrfs_inode_ref *iref,
> -				struct extent_buffer *eb, void *ctx);
>  
>  int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
>  			struct btrfs_path *path);
Reviewed-by: Jan Schmidt <list.btrfs@jan-o-sch.net>

-Jan
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

When applying this patch I get:

warning: 2 lines add whitespace errors.

More comments inline.

On Wed, August 08, 2012 at 20:55 (+0200), Mark Fasheh
wrote:
> Teach tree-log.c about extended inode refs. In particular, we have to
adjust
> the behavior of inode ref replay as well as log tree recovery to account
for
> the existence of extended refs.
> 
> Signed-off-by: Mark Fasheh <mfasheh@suse.de>
> ---
>  fs/btrfs/backref.c  |   68 ++++++++++++
>  fs/btrfs/backref.h  |    5 +
>  fs/btrfs/tree-log.c |  297
++++++++++++++++++++++++++++++++++++++++++---------
>  3 files changed, 319 insertions(+), 51 deletions(-)
> 
> diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
> index a383c18..658e09c 100644
> --- a/fs/btrfs/backref.c
> +++ b/fs/btrfs/backref.c
> @@ -1111,6 +1111,74 @@ static int inode_ref_info(u64 inum, u64 ioff, struct
btrfs_root *fs_root,
>  				found_key);
>  }
>  
> +int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
> +			  u64 start_off, struct btrfs_path *path,
> +			  struct btrfs_inode_extref **ret_extref,
> +			  u64 *found_off)
> +{
> +	int ret, slot;
> +	struct btrfs_key key;
> +	struct btrfs_key found_key;
> +	struct btrfs_inode_extref *extref;
> +	struct extent_buffer *leaf;
> +	unsigned long ptr;
> +
> +	key.objectid = inode_objectid;
> +	btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
> +	key.offset = start_off;
> +
> +	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
> +	if (ret < 0)
> +		return ret;
> +
> +	while (1) {
> +		leaf = path->nodes[0];
> +		slot = path->slots[0];
> +		if (slot >= btrfs_header_nritems(leaf)) {
> +			/*
> +			 * If the item at offset is not found,
> +			 * btrfs_search_slot will point us to the slot
> +			 * where it should be inserted. In our case
> +			 * that will be the slot directly before the
> +			 * next INODE_REF_KEY_V2 item. In the case
> +			 * that we''re pointing to the last slot in a
> +			 * leaf, we must move one leaf over.
> +			 */
> +			ret = btrfs_next_leaf(root, path);
> +			if (ret) {
> +				if (ret >= 1)
> +					ret = -ENOENT;
> +				break;
> +			}
We can finally replace this with btrfs_search_slot_for_read, according to my
first suggestion. It''s merged now. Saves that long comment and the
whole
while-1-continue construct, which is quite cumbersome to read.
> +			continue;
> +		}
> +
> +		btrfs_item_key_to_cpu(leaf, &found_key, slot);
> +
> +		/*
> +		 * Check that we''re still looking at an extended ref key for
> +		 * this particular objectid. If we have different
> +		 * objectid or type then there are no more to be found
> +		 * in the tree and we can exit.
> +		 */
> +		ret = -ENOENT;
> +		if (found_key.objectid != inode_objectid)
> +			break;
> +		if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
> +			break;
> +
> +		ret = 0;
> +		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
> +		extref = (struct btrfs_inode_extref *)ptr;
> +		*ret_extref = extref;
> +		if (found_off)
> +			*found_off = found_key.offset;
> +		break;
> +	}
> +
> +	return ret;
> +}
> +
>  /*
>   * this iterates to turn a btrfs_inode_ref into a full filesystem path.
elements
>   * of the path are separated by ''/'' and the path is
guaranteed to be
> diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
> index c18d8ac..9f3e251 100644
> --- a/fs/btrfs/backref.h
> +++ b/fs/btrfs/backref.h
> @@ -66,4 +66,9 @@ struct inode_fs_paths *init_ipath(s32 total_bytes, struct
btrfs_root *fs_root,
>  					struct btrfs_path *path);
>  void free_ipath(struct inode_fs_paths *ipath);
>  
> +int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
> +			  u64 start_off, struct btrfs_path *path,
> +			  struct btrfs_inode_extref **ret_extref,
> +			  u64 *found_off);
> +
>  #endif
> diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
> index 8abeae4..e5ba0a4 100644
> --- a/fs/btrfs/tree-log.c
> +++ b/fs/btrfs/tree-log.c
> @@ -23,8 +23,10 @@
>  #include "disk-io.h"
>  #include "locking.h"
>  #include "print-tree.h"
> +#include "backref.h"
>  #include "compat.h"
>  #include "tree-log.h"
> +#include "hash.h"
>  
>  /* magic values for the inode_only field in btrfs_log_inode:
>   *
> @@ -764,8 +766,16 @@ static noinline int backref_in_log(struct btrfs_root
*log,
>  	if (ret != 0)
>  		goto out;
>  
> -	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
>  	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
> +
> +	if (key->type == BTRFS_INODE_EXTREF_KEY) {
> +		if (btrfs_find_name_in_ext_backref(path, name, namelen, NULL))
> +			match = 1;
> +
> +		goto out;
> +	}
> +
> +	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
>  	ptr_end = ptr + item_size;
>  	while (ptr < ptr_end) {
>  		ref = (struct btrfs_inode_ref *)ptr;
> @@ -786,6 +796,47 @@ out:
>  	return match;
>  }
>  
> +static int extref_get_fields(struct extent_buffer *eb, int slot,
> +			     u32 *namelen, char **name, u64 *index,
> +			     u64 *parent_objectid)
> +{
> +	struct btrfs_inode_extref *extref;
> +
> +	extref = (struct btrfs_inode_extref *)btrfs_item_ptr_offset(eb, slot);
> +
> +	*namelen = btrfs_inode_extref_name_len(eb, extref);
> +	*name = kmalloc(*namelen, GFP_NOFS);
> +	if (*name == NULL)
> +		return -ENOMEM;
> +
> +	read_extent_buffer(eb, *name, (unsigned long)&extref->name,
> +			   *namelen);
> +
> +	*index = btrfs_inode_extref_index(eb, extref);
> +	if (parent_objectid)
> +		*parent_objectid = btrfs_inode_extref_parent(eb, extref);
> +
> +	return 0;
> +}
> +
> +static int ref_get_fields(struct extent_buffer *eb, int slot, u32
*namelen,
> +			  char **name, u64 *index)
> +{
> +	struct btrfs_inode_ref *ref;
> +
> +	ref = (struct btrfs_inode_ref *)btrfs_item_ptr_offset(eb, slot);
> +
> +	*namelen = btrfs_inode_ref_name_len(eb, ref);
> +	*name = kmalloc(*namelen, GFP_NOFS);	
> +	if (*name == NULL)
> +		return -ENOMEM;
> +
> +	read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);
> +
> +	*index = btrfs_inode_ref_index(eb, ref);
> +
> +	return 0;
> +}
>  
>  /*
>   * replay one inode back reference item found in the log tree.
> @@ -801,15 +852,50 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
>  				  struct btrfs_key *key)
Control flow in this function is horrible. I finally failed to understand it.
That''s why I made a patch that replaces most gotos by if blocks and
uses a sub
function (see the end of this email). I hope that it''s getting easier
to extend
that patched version. Cannot comment on changes to code I don''t
understand, sorry.

If you find my patch helps understanding that function and makes patching
easier, you can include that patch in your patch set as patch 2/4 and then do
your extension afterwards.
>  {
>  	struct btrfs_inode_ref *ref;
> +	struct btrfs_inode_extref *extref;
>  	struct btrfs_dir_item *di;
> +	struct btrfs_key search_key;
>  	struct inode *dir;
>  	struct inode *inode;
>  	unsigned long ref_ptr;
>  	unsigned long ref_end;
>  	char *name;
> +	char *victim_name;
>  	int namelen;
> +	int victim_name_len;
>  	int ret;
>  	int search_done = 0;
> +	int log_ref_ver = 0;
> +	u64 parent_objectid;
> +	u64 inode_objectid;
> +	u64 ref_index;
> +	struct extent_buffer *leaf;
> +	int ref_struct_size;
> +
> +
> +	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> +	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> +
> +	if (key->type == BTRFS_INODE_EXTREF_KEY) {	
> +		ref_struct_size = sizeof(*extref);
> +		log_ref_ver = 1;
> +
> +		ret = extref_get_fields(eb, slot, &namelen, &name,
&ref_index,
> +					&parent_objectid);
> +		if (ret)
> +			return ret;
> +	} else {
> +		ref_struct_size = sizeof(*ref);
> +
> +		ret = ref_get_fields(eb, slot, &namelen, &name, &ref_index);
> +		if (ret)
> +			return ret;
> +
> +
> +		parent_objectid = key->offset;
> +	}
> +
> +	inode_objectid = key->objectid;
>  
>  	/*
>  	 * it is possible that we didn''t log all the parent directories
> @@ -817,32 +903,20 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
>  	 * copy the back ref in.  The link count fixup code will take
>  	 * care of the rest
>  	 */
> -	dir = read_one_inode(root, key->offset);
> +	dir = read_one_inode(root, parent_objectid);
>  	if (!dir)
>  		return -ENOENT;
>  
> -	inode = read_one_inode(root, key->objectid);
> +	inode = read_one_inode(root, inode_objectid);
>  	if (!inode) {
>  		iput(dir);
>  		return -EIO;
>  	}
>  
> -	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> -	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> -
>  again:
> -	ref = (struct btrfs_inode_ref *)ref_ptr;
> -
> -	namelen = btrfs_inode_ref_name_len(eb, ref);
> -	name = kmalloc(namelen, GFP_NOFS);
> -	BUG_ON(!name);
> -
> -	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
> -
>  	/* if we already have a perfect match, we''re done */
>  	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
> -			 btrfs_inode_ref_index(eb, ref),
> -			 name, namelen)) {
> +			 ref_index, name, namelen)) {
>  		goto out;
>  	}
>  
> @@ -857,19 +931,23 @@ again:
>  	if (search_done)
>  		goto insert;
>  
> -	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
> +	/* Search old style refs */
> +	search_key.objectid = inode_objectid;
> +	search_key.type = BTRFS_INODE_REF_KEY;
> +	search_key.offset = parent_objectid;
> +
> +	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
>  	if (ret == 0) {
> -		char *victim_name;
> -		int victim_name_len;
>  		struct btrfs_inode_ref *victim_ref;
>  		unsigned long ptr;
>  		unsigned long ptr_end;
> -		struct extent_buffer *leaf = path->nodes[0];
> +
> +		leaf = path->nodes[0];
>  
>  		/* are we trying to overwrite a back ref for the root directory
>  		 * if so, just jump out, we''re done
>  		 */
> -		if (key->objectid == key->offset)
> +		if (search_key.objectid == search_key.offset)
>  			goto out_nowrite;
>  
>  		/* check all the names in this back reference to see
> @@ -889,7 +967,7 @@ again:
>  					   (unsigned long)(victim_ref + 1),
>  					   victim_name_len);
>  
> -			if (!backref_in_log(log, key, victim_name,
> +			if (!backref_in_log(log, &search_key, victim_name,
>  					    victim_name_len)) {
>  				btrfs_inc_nlink(inode);
>  				btrfs_release_path(path);
> @@ -903,19 +981,61 @@ again:
>  			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
>  		}
>  		BUG_ON(ret);
> -
>  		/*
>  		 * NOTE: we have searched root tree and checked the
> -		 * coresponding ref, it does not need to check again.
> +		 * coresponding refs, it does not need to be checked again.
>  		 */
>  		search_done = 1;
>  	}
>  	btrfs_release_path(path);
>  
> +	/* Same search but for extended refs */
> +	extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
> +					   inode_objectid, parent_objectid, 0,
> +					   0);
> +	if (!IS_ERR_OR_NULL(extref)) {
Okay, one obvious comment at least: You don''t execute this block on
IS_ERR(extref), but you don''t act on that situation either. Is that on
purpose?
> +		u32 item_size;
> +		u32 cur_offset = 0;
> +		unsigned long base;
> +
> +		leaf = path->nodes[0];
> +
> +		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
> +		base = btrfs_item_ptr_offset(leaf, path->slots[0]);
> +
> +		while (cur_offset < item_size) {
> +			extref = (struct btrfs_inode_extref *)base + cur_offset;
> +
> +			victim_name_len = btrfs_inode_extref_name_len(eb, extref);
> +			victim_name = kmalloc(namelen, GFP_NOFS);
> +			leaf = path->nodes[0];
> +			read_extent_buffer(eb, name, (unsigned long)&extref->name,
namelen);
> +
> +			search_key.objectid = inode_objectid;
> +			search_key.type = BTRFS_INODE_EXTREF_KEY;
> +			search_key.offset = btrfs_extref_hash(parent_objectid,
> +							      name, namelen);
> +			if (!backref_in_log(log, &search_key, victim_name,
> +					    victim_name_len)) {
> +				btrfs_inc_nlink(inode);
> +				btrfs_release_path(path);
> +
> +				ret = btrfs_unlink_inode(trans, root, dir,
> +							 inode, victim_name,
> +							 victim_name_len);
> +			}
> +			kfree(victim_name);
> +			BUG_ON(ret);
> +
> +			cur_offset += victim_name_len + sizeof(*extref);
> +		}
> +		search_done = 1;
> +	}
> +
> +
>  	/* look for a conflicting sequence number */
>  	di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
> -					 btrfs_inode_ref_index(eb, ref),
> -					 name, namelen, 0);
> +					 ref_index, name, namelen, 0);
>  	if (di && !IS_ERR(di)) {
>  		ret = drop_one_dir_item(trans, root, path, dir, di);
>  		BUG_ON(ret);
> @@ -933,17 +1053,25 @@ again:
>  
>  insert:
>  	/* insert our name */
> -	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
> -			     btrfs_inode_ref_index(eb, ref));
> +	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0, ref_index);
>  	BUG_ON(ret);
>  
>  	btrfs_update_inode(trans, root, inode);
>  
>  out:
> -	ref_ptr = (unsigned long)(ref + 1) + namelen;
> +	ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
>  	kfree(name);
> -	if (ref_ptr < ref_end)
> +	if (ref_ptr < ref_end) {
> +		if (log_ref_ver) {
> +			ret = extref_get_fields(eb, slot, &namelen, &name,
> +						&ref_index, NULL);
> +		} else {
> +			ret = ref_get_fields(eb, slot, &namelen, &name,
> +					     &ref_index);
> +		}
> +		BUG_ON(ret);
>  		goto again;
> +	}
>  
>  	/* finally write the back reference in the inode */
>  	ret = overwrite_item(trans, root, path, eb, slot, key);
> @@ -966,25 +1094,52 @@ static int insert_orphan_item(struct
btrfs_trans_handle *trans,
>  	return ret;
>  }
>  
> +static int count_inode_extrefs(struct btrfs_root *root,
> +			       struct inode *inode, struct btrfs_path *path)
> +{
> +	int ret = 0;
> +	int name_len;
> +	unsigned int nlink = 0;
> +	u32 item_size;
> +	u32 cur_offset = 0;
> +	u64 inode_objectid = btrfs_ino(inode);
> +	u64 offset = 0;
> +	unsigned long ptr;
> +	struct btrfs_inode_extref *extref;
> +	struct extent_buffer *leaf;
> +
> +	while (1) {
> +		ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
> +					    &extref, &offset);
> +		if (ret)
> +			break;
>  
> -/*
> - * There are a few corners where the link count of the file can''t
> - * be properly maintained during replay.  So, instead of adding
> - * lots of complexity to the log code, we just scan the backrefs
> - * for any file that has been through replay.
> - *
> - * The scan will update the link count on the inode to reflect the
> - * number of back refs found.  If it goes down to zero, the iput
> - * will free the inode.
> - */
> -static noinline int fixup_inode_link_count(struct btrfs_trans_handle
*trans,
> -					   struct btrfs_root *root,
> -					   struct inode *inode)
> +		leaf = path->nodes[0];
> +		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
> +		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
> +
> +		while (cur_offset < item_size) {
> +			extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
> +			name_len = btrfs_inode_extref_name_len(leaf, extref);
> +
> +			nlink++;
> +
> +			cur_offset += name_len + sizeof(*extref);
> +		}
> +
> +		offset++;
> +	}
> +	btrfs_release_path(path);
> +
> +	return (ret == 0) ? nlink : ret;
You depend on btrfs_find_one_extref never returning >0 here. I''d
make that ...

 if (ret < 0)
 	return ret;
 return nlink;

This ignores ret > 0 explicitly, which is in my opinion better than a
spurious
link count.
> +}
> +
> +static int count_inode_refs(struct btrfs_root *root,
> +			       struct inode *inode, struct btrfs_path *path)
>  {
> -	struct btrfs_path *path;
>  	int ret;
>  	struct btrfs_key key;
> -	u64 nlink = 0;
> +	unsigned int nlink = 0;
>  	unsigned long ptr;
>  	unsigned long ptr_end;
>  	int name_len;
> @@ -994,10 +1149,6 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
>  	key.type = BTRFS_INODE_REF_KEY;
>  	key.offset = (u64)-1;
>  
> -	path = btrfs_alloc_path();
> -	if (!path)
> -		return -ENOMEM;
> -
>  	while (1) {
>  		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
>  		if (ret < 0)
> @@ -1031,6 +1182,45 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
>  		btrfs_release_path(path);
>  	}
>  	btrfs_release_path(path);
> +
> +	return nlink;
> +}
> +
> +/*
> + * There are a few corners where the link count of the file can''t
> + * be properly maintained during replay.  So, instead of adding
> + * lots of complexity to the log code, we just scan the backrefs
> + * for any file that has been through replay.
> + *
> + * The scan will update the link count on the inode to reflect the
> + * number of back refs found.  If it goes down to zero, the iput
> + * will free the inode.
> + */
> +static noinline int fixup_inode_link_count(struct btrfs_trans_handle
*trans,
> +					   struct btrfs_root *root,
> +					   struct inode *inode)
> +{
> +	struct btrfs_path *path;
> +	int ret;
> +	u64 nlink = 0;
> +	u64 ino = btrfs_ino(inode);
> +
> +	path = btrfs_alloc_path();
> +	if (!path)
> +		return -ENOMEM;
> +
> +	ret = count_inode_refs(root, inode, path);
> +	if (ret < 0)
> +		goto out;
> +
> +	nlink = ret;
> +
> +	ret = count_inode_extrefs(root, inode, path);
> +	if (ret < 0)
> +		goto out;
> +
> +	nlink += ret;
> +
>  	if (nlink != inode->i_nlink) {
>  		set_nlink(inode, nlink);
>  		btrfs_update_inode(trans, root, inode);
> @@ -1046,9 +1236,10 @@ static noinline int fixup_inode_link_count(struct
btrfs_trans_handle *trans,
>  		ret = insert_orphan_item(trans, root, ino);
>  		BUG_ON(ret);
>  	}
> -	btrfs_free_path(path);
>  
> -	return 0;
> +out:
> +	btrfs_free_path(path);
> +	return ret;
>  }
>  
>  static noinline int fixup_inode_link_counts(struct btrfs_trans_handle
*trans,
> @@ -1695,6 +1886,10 @@ static int replay_one_buffer(struct btrfs_root *log,
struct extent_buffer *eb,
>  			ret = add_inode_ref(wc->trans, root, log, path,
>  					    eb, i, &key);
>  			BUG_ON(ret && ret != -ENOENT);
> +		} else if (key.type == BTRFS_INODE_EXTREF_KEY) {
> +			ret = add_inode_ref(wc->trans, root, log, path,
> +					    eb, i, &key);
> +			BUG_ON(ret && ret != -ENOENT);
>  		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
>  			ret = replay_one_extent(wc->trans, root, path,
>  						eb, i, &key);
That''s it for this round, here comes the promised patch. Should apply
with "git
am --scissors".

Can you please add a version to the subject before sending the revised patches?
Like "git format-patch --subject-prefix v4".

Thanks,
-Jan

-- >8 --
Subject: [PATCH] Btrfs: improved readablity for add_inode_ref

Moved part of the code into a sub function and replaced most of the gotos
by ifs, hoping that it will be easier to read now.

Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
---
 fs/btrfs/tree-log.c |  178 ++++++++++++++++++++++++++++-----------------------
 1 files changed, 97 insertions(+), 81 deletions(-)

diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index c86670f..59f3071 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -786,76 +786,18 @@ out:
 	return match;
 }

-
-/*
- * replay one inode back reference item found in the log tree.
- * eb, slot and key refer to the buffer and key found in the log tree.
- * root is the destination we are replaying into, and path is for temp
- * use by this function.  (it should be released on return).
- */
-static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
+static inline int __add_inode_ref(struct btrfs_trans_handle *trans,
 				  struct btrfs_root *root,
-				  struct btrfs_root *log,
 				  struct btrfs_path *path,
-				  struct extent_buffer *eb, int slot,
-				  struct btrfs_key *key)
+				  struct btrfs_root *log_root,
+				  struct inode *dir, struct inode *inode,
+				  struct btrfs_key *key,
+				  struct extent_buffer *eb,
+				  struct btrfs_inode_ref *ref,
+				  char *name, int namelen, int *search_done)
 {
-	struct btrfs_inode_ref *ref;
-	struct btrfs_dir_item *di;
-	struct inode *dir;
-	struct inode *inode;
-	unsigned long ref_ptr;
-	unsigned long ref_end;
-	char *name;
-	int namelen;
 	int ret;
-	int search_done = 0;
-
-	/*
-	 * it is possible that we didn''t log all the parent directories
-	 * for a given inode.  If we don''t find the dir, just don''t
-	 * copy the back ref in.  The link count fixup code will take
-	 * care of the rest
-	 */
-	dir = read_one_inode(root, key->offset);
-	if (!dir)
-		return -ENOENT;
-
-	inode = read_one_inode(root, key->objectid);
-	if (!inode) {
-		iput(dir);
-		return -EIO;
-	}
-
-	ref_ptr = btrfs_item_ptr_offset(eb, slot);
-	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
-
-again:
-	ref = (struct btrfs_inode_ref *)ref_ptr;
-
-	namelen = btrfs_inode_ref_name_len(eb, ref);
-	name = kmalloc(namelen, GFP_NOFS);
-	BUG_ON(!name);
-
-	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
-
-	/* if we already have a perfect match, we''re done */
-	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
-			 btrfs_inode_ref_index(eb, ref),
-			 name, namelen)) {
-		goto out;
-	}
-
-	/*
-	 * look for a conflicting back reference in the metadata.
-	 * if we find one we have to unlink that name of the file
-	 * before we add our new link.  Later on, we overwrite any
-	 * existing back reference, and we don''t want to create
-	 * dangling pointers in the directory.
-	 */
-
-	if (search_done)
-		goto insert;
+	struct btrfs_dir_item *di;

 	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
 	if (ret == 0) {
@@ -870,7 +812,7 @@ again:
 		 * if so, just jump out, we''re done
 		 */
 		if (key->objectid == key->offset)
-			goto out_nowrite;
+			return 1;

 		/* check all the names in this back reference to see
 		 * if they are in the log.  if so, we allow them to stay
@@ -889,7 +831,7 @@ again:
 					   (unsigned long)(victim_ref + 1),
 					   victim_name_len);

-			if (!backref_in_log(log, key, victim_name,
+			if (!backref_in_log(log_root, key, victim_name,
 					    victim_name_len)) {
 				btrfs_inc_nlink(inode);
 				btrfs_release_path(path);
@@ -908,7 +850,7 @@ again:
 		 * NOTE: we have searched root tree and checked the
 		 * coresponding ref, it does not need to check again.
 		 */
-		search_done = 1;
+		*search_done = 1;
 	}
 	btrfs_release_path(path);

@@ -931,25 +873,99 @@ again:
 	}
 	btrfs_release_path(path);

-insert:
-	/* insert our name */
-	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
-			     btrfs_inode_ref_index(eb, ref));
-	BUG_ON(ret);
+	return 0;
+}
+
+/*
+ * replay one inode back reference item found in the log tree.
+ * eb, slot and key refer to the buffer and key found in the log tree.
+ * root is the destination we are replaying into, and path is for temp
+ * use by this function.  (it should be released on return).
+ */
+static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  struct btrfs_root *log,
+				  struct btrfs_path *path,
+				  struct extent_buffer *eb, int slot,
+				  struct btrfs_key *key)
+{
+	struct btrfs_inode_ref *ref;
+	struct inode *dir;
+	struct inode *inode;
+	unsigned long ref_ptr;
+	unsigned long ref_end;
+	char *name;
+	int namelen;
+	int ret;
+	int search_done = 0;
+
+	/*
+	 * it is possible that we didn''t log all the parent directories
+	 * for a given inode.  If we don''t find the dir, just don''t
+	 * copy the back ref in.  The link count fixup code will take
+	 * care of the rest
+	 */
+	dir = read_one_inode(root, key->offset);
+	if (!dir)
+		return -ENOENT;

-	btrfs_update_inode(trans, root, inode);
+	inode = read_one_inode(root, key->objectid);
+	if (!inode) {
+		iput(dir);
+		return -EIO;
+	}

-out:
-	ref_ptr = (unsigned long)(ref + 1) + namelen;
-	kfree(name);
-	if (ref_ptr < ref_end)
-		goto again;
+	ref_ptr = btrfs_item_ptr_offset(eb, slot);
+	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
+
+	while (ref_ptr < ref_end) {
+		ref = (struct btrfs_inode_ref *)ref_ptr;
+
+		namelen = btrfs_inode_ref_name_len(eb, ref);
+		name = kmalloc(namelen, GFP_NOFS);
+		BUG_ON(!name);
+
+		read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
+
+		/* if we already have a perfect match, we''re done */
+		if (!inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
+				  btrfs_inode_ref_index(eb, ref),
+				  name, namelen)) {
+			/*
+			 * look for a conflicting back reference in the
+			 * metadata. if we find one we have to unlink that name
+			 * of the file before we add our new link.  Later on, we
+			 * overwrite any existing back reference, and we don''t
+			 * want to create dangling pointers in the directory.
+			 */
+
+			if (!search_done) {
+				ret = __add_inode_ref(trans, root, path, log,
+						      dir, inode, key, eb, ref,
+						      name, namelen,
+						      &search_done);
+				if (ret == 1)
+					goto out;
+				BUG_ON(ret);
+			}
+
+			/* insert our name */
+			ret = btrfs_add_link(trans, dir, inode, name, namelen,
+					     0, btrfs_inode_ref_index(eb, ref));
+			BUG_ON(ret);
+
+			btrfs_update_inode(trans, root, inode);
+		}
+
+		ref_ptr = (unsigned long)(ref + 1) + namelen;
+		kfree(name);
+	}

 	/* finally write the back reference in the inode */
 	ret = overwrite_item(trans, root, path, eb, slot, key);
 	BUG_ON(ret);

-out_nowrite:
+out:
 	btrfs_release_path(path);
 	iput(dir);
 	iput(inode);
-- 
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Wed, Aug 15, 2012 at 05:04:48PM +0200, Jan Schmidt
wrote:
> When applying this patch I get:
> 
> warning: 2 lines add whitespace errors.
> 
Oop, I''ll fix that up.

> More comments inline.
> 
> On Wed, August 08, 2012 at 20:55 (+0200), Mark Fasheh wrote:
> > Teach tree-log.c about extended inode refs. In particular, we have to
adjust
> > the behavior of inode ref replay as well as log tree recovery to
account for
> > the existence of extended refs.
> > 
> > Signed-off-by: Mark Fasheh <mfasheh@suse.de>
> > ---
> >  fs/btrfs/backref.c  |   68 ++++++++++++
> >  fs/btrfs/backref.h  |    5 +
> >  fs/btrfs/tree-log.c |  297
++++++++++++++++++++++++++++++++++++++++++---------
> >  3 files changed, 319 insertions(+), 51 deletions(-)
> > 
> > diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
> > index a383c18..658e09c 100644
> > --- a/fs/btrfs/backref.c
> > +++ b/fs/btrfs/backref.c
> > @@ -1111,6 +1111,74 @@ static int inode_ref_info(u64 inum, u64 ioff,
struct btrfs_root *fs_root,
> >  				found_key);
> >  }
> >  
> > +int btrfs_find_one_extref(struct btrfs_root *root, u64
inode_objectid,
> > +			  u64 start_off, struct btrfs_path *path,
> > +			  struct btrfs_inode_extref **ret_extref,
> > +			  u64 *found_off)
> > +{
> > +	int ret, slot;
> > +	struct btrfs_key key;
> > +	struct btrfs_key found_key;
> > +	struct btrfs_inode_extref *extref;
> > +	struct extent_buffer *leaf;
> > +	unsigned long ptr;
> > +
> > +	key.objectid = inode_objectid;
> > +	btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
> > +	key.offset = start_off;
> > +
> > +	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
> > +	if (ret < 0)
> > +		return ret;
> > +
> > +	while (1) {
> > +		leaf = path->nodes[0];
> > +		slot = path->slots[0];
> > +		if (slot >= btrfs_header_nritems(leaf)) {
> > +			/*
> > +			 * If the item at offset is not found,
> > +			 * btrfs_search_slot will point us to the slot
> > +			 * where it should be inserted. In our case
> > +			 * that will be the slot directly before the
> > +			 * next INODE_REF_KEY_V2 item. In the case
> > +			 * that we''re pointing to the last slot in a
> > +			 * leaf, we must move one leaf over.
> > +			 */
> > +			ret = btrfs_next_leaf(root, path);
> > +			if (ret) {
> > +				if (ret >= 1)
> > +					ret = -ENOENT;
> > +				break;
> > +			}
> 
> We can finally replace this with btrfs_search_slot_for_read, according to
my
> first suggestion. It''s merged now. Saves that long comment and the
whole
> while-1-continue construct, which is quite cumbersome to read.
Alright, I''ll take a look at this. I need to read
btrfs_search_slot_for_read() first to make sure it all fits correctly.

> > +			continue;
> > +		}
> > +
> > +		btrfs_item_key_to_cpu(leaf, &found_key, slot);
> > +
> > +		/*
> > +		 * Check that we''re still looking at an extended ref key
for
> > +		 * this particular objectid. If we have different
> > +		 * objectid or type then there are no more to be found
> > +		 * in the tree and we can exit.
> > +		 */
> > +		ret = -ENOENT;
> > +		if (found_key.objectid != inode_objectid)
> > +			break;
> > +		if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
> > +			break;
> > +
> > +		ret = 0;
> > +		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
> > +		extref = (struct btrfs_inode_extref *)ptr;
> > +		*ret_extref = extref;
> > +		if (found_off)
> > +			*found_off = found_key.offset;
> > +		break;
> > +	}
> > +
> > +	return ret;
> > +}
> > +
> >  /*
> >   * this iterates to turn a btrfs_inode_ref into a full filesystem
path. elements
> >   * of the path are separated by ''/'' and the path is
guaranteed to be
> > diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
> > index c18d8ac..9f3e251 100644
> > --- a/fs/btrfs/backref.h
> > +++ b/fs/btrfs/backref.h
> > @@ -66,4 +66,9 @@ struct inode_fs_paths *init_ipath(s32 total_bytes,
struct btrfs_root *fs_root,
> >  					struct btrfs_path *path);
> >  void free_ipath(struct inode_fs_paths *ipath);
> >  
> > +int btrfs_find_one_extref(struct btrfs_root *root, u64
inode_objectid,
> > +			  u64 start_off, struct btrfs_path *path,
> > +			  struct btrfs_inode_extref **ret_extref,
> > +			  u64 *found_off);
> > +
> >  #endif
> > diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
> > index 8abeae4..e5ba0a4 100644
> > --- a/fs/btrfs/tree-log.c
> > +++ b/fs/btrfs/tree-log.c
> > @@ -23,8 +23,10 @@
> >  #include "disk-io.h"
> >  #include "locking.h"
> >  #include "print-tree.h"
> > +#include "backref.h"
> >  #include "compat.h"
> >  #include "tree-log.h"
> > +#include "hash.h"
> >  
> >  /* magic values for the inode_only field in btrfs_log_inode:
> >   *
> > @@ -764,8 +766,16 @@ static noinline int backref_in_log(struct
btrfs_root *log,
> >  	if (ret != 0)
> >  		goto out;
> >  
> > -	item_size = btrfs_item_size_nr(path->nodes[0],
path->slots[0]);
> >  	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
> > +
> > +	if (key->type == BTRFS_INODE_EXTREF_KEY) {
> > +		if (btrfs_find_name_in_ext_backref(path, name, namelen, NULL))
> > +			match = 1;
> > +
> > +		goto out;
> > +	}
> > +
> > +	item_size = btrfs_item_size_nr(path->nodes[0],
path->slots[0]);
> >  	ptr_end = ptr + item_size;
> >  	while (ptr < ptr_end) {
> >  		ref = (struct btrfs_inode_ref *)ptr;
> > @@ -786,6 +796,47 @@ out:
> >  	return match;
> >  }
> >  
> > +static int extref_get_fields(struct extent_buffer *eb, int slot,
> > +			     u32 *namelen, char **name, u64 *index,
> > +			     u64 *parent_objectid)
> > +{
> > +	struct btrfs_inode_extref *extref;
> > +
> > +	extref = (struct btrfs_inode_extref *)btrfs_item_ptr_offset(eb,
slot);
> > +
> > +	*namelen = btrfs_inode_extref_name_len(eb, extref);
> > +	*name = kmalloc(*namelen, GFP_NOFS);
> > +	if (*name == NULL)
> > +		return -ENOMEM;
> > +
> > +	read_extent_buffer(eb, *name, (unsigned long)&extref->name,
> > +			   *namelen);
> > +
> > +	*index = btrfs_inode_extref_index(eb, extref);
> > +	if (parent_objectid)
> > +		*parent_objectid = btrfs_inode_extref_parent(eb, extref);
> > +
> > +	return 0;
> > +}
> > +
> > +static int ref_get_fields(struct extent_buffer *eb, int slot, u32
*namelen,
> > +			  char **name, u64 *index)
> > +{
> > +	struct btrfs_inode_ref *ref;
> > +
> > +	ref = (struct btrfs_inode_ref *)btrfs_item_ptr_offset(eb, slot);
> > +
> > +	*namelen = btrfs_inode_ref_name_len(eb, ref);
> > +	*name = kmalloc(*namelen, GFP_NOFS);	
> > +	if (*name == NULL)
> > +		return -ENOMEM;
> > +
> > +	read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);
> > +
> > +	*index = btrfs_inode_ref_index(eb, ref);
> > +
> > +	return 0;
> > +}
> >  
> >  /*
> >   * replay one inode back reference item found in the log tree.
> > @@ -801,15 +852,50 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
> >  				  struct btrfs_key *key)
> 
> Control flow in this function is horrible. I finally failed to understand
it.
> That''s why I made a patch that replaces most gotos by if blocks
and uses a sub
> function (see the end of this email). I hope that it''s getting
easier to extend
> that patched version. Cannot comment on changes to code I don''t
understand, sorry.
I *think* I understand most of it, after much much reading and re-reading. I
really am not confident enough though to say anything on the matter until
Chris weighs in.

> If you find my patch helps understanding that function and makes patching
> easier, you can include that patch in your patch set as patch 2/4 and then
do
> your extension afterwards.
Thanks, I''ll take a look at your patch. I think any attempt to clean it
up
will be nicely recieved.

> >  {
> >  	struct btrfs_inode_ref *ref;
> > +	struct btrfs_inode_extref *extref;
> >  	struct btrfs_dir_item *di;
> > +	struct btrfs_key search_key;
> >  	struct inode *dir;
> >  	struct inode *inode;
> >  	unsigned long ref_ptr;
> >  	unsigned long ref_end;
> >  	char *name;
> > +	char *victim_name;
> >  	int namelen;
> > +	int victim_name_len;
> >  	int ret;
> >  	int search_done = 0;
> > +	int log_ref_ver = 0;
> > +	u64 parent_objectid;
> > +	u64 inode_objectid;
> > +	u64 ref_index;
> > +	struct extent_buffer *leaf;
> > +	int ref_struct_size;
> > +
> > +
> > +	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> > +	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> > +
> > +	if (key->type == BTRFS_INODE_EXTREF_KEY) {	
> > +		ref_struct_size = sizeof(*extref);
> > +		log_ref_ver = 1;
> > +
> > +		ret = extref_get_fields(eb, slot, &namelen, &name,
&ref_index,
> > +					&parent_objectid);
> > +		if (ret)
> > +			return ret;
> > +	} else {
> > +		ref_struct_size = sizeof(*ref);
> > +
> > +		ret = ref_get_fields(eb, slot, &namelen, &name,
&ref_index);
> > +		if (ret)
> > +			return ret;
> > +
> > +
> > +		parent_objectid = key->offset;
> > +	}
> > +
> > +	inode_objectid = key->objectid;
> >  
> >  	/*
> >  	 * it is possible that we didn''t log all the parent
directories
> > @@ -817,32 +903,20 @@ static noinline int add_inode_ref(struct
btrfs_trans_handle *trans,
> >  	 * copy the back ref in.  The link count fixup code will take
> >  	 * care of the rest
> >  	 */
> > -	dir = read_one_inode(root, key->offset);
> > +	dir = read_one_inode(root, parent_objectid);
> >  	if (!dir)
> >  		return -ENOENT;
> >  
> > -	inode = read_one_inode(root, key->objectid);
> > +	inode = read_one_inode(root, inode_objectid);
> >  	if (!inode) {
> >  		iput(dir);
> >  		return -EIO;
> >  	}
> >  
> > -	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> > -	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> > -
> >  again:
> > -	ref = (struct btrfs_inode_ref *)ref_ptr;
> > -
> > -	namelen = btrfs_inode_ref_name_len(eb, ref);
> > -	name = kmalloc(namelen, GFP_NOFS);
> > -	BUG_ON(!name);
> > -
> > -	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
> > -
> >  	/* if we already have a perfect match, we''re done */
> >  	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
> > -			 btrfs_inode_ref_index(eb, ref),
> > -			 name, namelen)) {
> > +			 ref_index, name, namelen)) {
> >  		goto out;
> >  	}
> >  
> > @@ -857,19 +931,23 @@ again:
> >  	if (search_done)
> >  		goto insert;
> >  
> > -	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
> > +	/* Search old style refs */
> > +	search_key.objectid = inode_objectid;
> > +	search_key.type = BTRFS_INODE_REF_KEY;
> > +	search_key.offset = parent_objectid;
> > +
> > +	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
> >  	if (ret == 0) {
> > -		char *victim_name;
> > -		int victim_name_len;
> >  		struct btrfs_inode_ref *victim_ref;
> >  		unsigned long ptr;
> >  		unsigned long ptr_end;
> > -		struct extent_buffer *leaf = path->nodes[0];
> > +
> > +		leaf = path->nodes[0];
> >  
> >  		/* are we trying to overwrite a back ref for the root directory
> >  		 * if so, just jump out, we''re done
> >  		 */
> > -		if (key->objectid == key->offset)
> > +		if (search_key.objectid == search_key.offset)
> >  			goto out_nowrite;
> >  
> >  		/* check all the names in this back reference to see
> > @@ -889,7 +967,7 @@ again:
> >  					   (unsigned long)(victim_ref + 1),
> >  					   victim_name_len);
> >  
> > -			if (!backref_in_log(log, key, victim_name,
> > +			if (!backref_in_log(log, &search_key, victim_name,
> >  					    victim_name_len)) {
> >  				btrfs_inc_nlink(inode);
> >  				btrfs_release_path(path);
> > @@ -903,19 +981,61 @@ again:
> >  			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
> >  		}
> >  		BUG_ON(ret);
> > -
> >  		/*
> >  		 * NOTE: we have searched root tree and checked the
> > -		 * coresponding ref, it does not need to check again.
> > +		 * coresponding refs, it does not need to be checked again.
> >  		 */
> >  		search_done = 1;
> >  	}
> >  	btrfs_release_path(path);
> >  
> > +	/* Same search but for extended refs */
> > +	extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
> > +					   inode_objectid, parent_objectid, 0,
> > +					   0);
> > +	if (!IS_ERR_OR_NULL(extref)) {
> 
> Okay, one obvious comment at least: You don''t execute this block
on
> IS_ERR(extref), but you don''t act on that situation either. Is
that on purpose?
This is just mirroring the broken (or maybe misunderstood?) error handling
that this function has (unfortunately, all over the place). Basically, we
ignore errors from the lookups for either ref type.

> > +		u32 item_size;
> > +		u32 cur_offset = 0;
> > +		unsigned long base;
> > +
> > +		leaf = path->nodes[0];
> > +
> > +		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
> > +		base = btrfs_item_ptr_offset(leaf, path->slots[0]);
> > +
> > +		while (cur_offset < item_size) {
> > +			extref = (struct btrfs_inode_extref *)base + cur_offset;
> > +
> > +			victim_name_len = btrfs_inode_extref_name_len(eb, extref);
> > +			victim_name = kmalloc(namelen, GFP_NOFS);
> > +			leaf = path->nodes[0];
> > +			read_extent_buffer(eb, name, (unsigned long)&extref->name,
namelen);
> > +
> > +			search_key.objectid = inode_objectid;
> > +			search_key.type = BTRFS_INODE_EXTREF_KEY;
> > +			search_key.offset = btrfs_extref_hash(parent_objectid,
> > +							      name, namelen);
> > +			if (!backref_in_log(log, &search_key, victim_name,
> > +					    victim_name_len)) {
> > +				btrfs_inc_nlink(inode);
> > +				btrfs_release_path(path);
> > +
> > +				ret = btrfs_unlink_inode(trans, root, dir,
> > +							 inode, victim_name,
> > +							 victim_name_len);
> > +			}
> > +			kfree(victim_name);
> > +			BUG_ON(ret);
> > +
> > +			cur_offset += victim_name_len + sizeof(*extref);
> > +		}
> > +		search_done = 1;
> > +	}
> > +
> > +
> >  	/* look for a conflicting sequence number */
> >  	di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
> > -					 btrfs_inode_ref_index(eb, ref),
> > -					 name, namelen, 0);
> > +					 ref_index, name, namelen, 0);
> >  	if (di && !IS_ERR(di)) {
> >  		ret = drop_one_dir_item(trans, root, path, dir, di);
> >  		BUG_ON(ret);
> > @@ -933,17 +1053,25 @@ again:
> >  
> >  insert:
> >  	/* insert our name */
> > -	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
> > -			     btrfs_inode_ref_index(eb, ref));
> > +	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
ref_index);
> >  	BUG_ON(ret);
> >  
> >  	btrfs_update_inode(trans, root, inode);
> >  
> >  out:
> > -	ref_ptr = (unsigned long)(ref + 1) + namelen;
> > +	ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
> >  	kfree(name);
> > -	if (ref_ptr < ref_end)
> > +	if (ref_ptr < ref_end) {
> > +		if (log_ref_ver) {
> > +			ret = extref_get_fields(eb, slot, &namelen, &name,
> > +						&ref_index, NULL);
> > +		} else {
> > +			ret = ref_get_fields(eb, slot, &namelen, &name,
> > +					     &ref_index);
> > +		}
> > +		BUG_ON(ret);
> >  		goto again;
> > +	}
> >  
> >  	/* finally write the back reference in the inode */
> >  	ret = overwrite_item(trans, root, path, eb, slot, key);
> > @@ -966,25 +1094,52 @@ static int insert_orphan_item(struct
btrfs_trans_handle *trans,
> >  	return ret;
> >  }
> >  
> > +static int count_inode_extrefs(struct btrfs_root *root,
> > +			       struct inode *inode, struct btrfs_path *path)
> > +{
> > +	int ret = 0;
> > +	int name_len;
> > +	unsigned int nlink = 0;
> > +	u32 item_size;
> > +	u32 cur_offset = 0;
> > +	u64 inode_objectid = btrfs_ino(inode);
> > +	u64 offset = 0;
> > +	unsigned long ptr;
> > +	struct btrfs_inode_extref *extref;
> > +	struct extent_buffer *leaf;
> > +
> > +	while (1) {
> > +		ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
> > +					    &extref, &offset);
> > +		if (ret)
> > +			break;
> >  
> > -/*
> > - * There are a few corners where the link count of the file
can''t
> > - * be properly maintained during replay.  So, instead of adding
> > - * lots of complexity to the log code, we just scan the backrefs
> > - * for any file that has been through replay.
> > - *
> > - * The scan will update the link count on the inode to reflect the
> > - * number of back refs found.  If it goes down to zero, the iput
> > - * will free the inode.
> > - */
> > -static noinline int fixup_inode_link_count(struct btrfs_trans_handle
*trans,
> > -					   struct btrfs_root *root,
> > -					   struct inode *inode)
> > +		leaf = path->nodes[0];
> > +		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
> > +		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
> > +
> > +		while (cur_offset < item_size) {
> > +			extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
> > +			name_len = btrfs_inode_extref_name_len(leaf, extref);
> > +
> > +			nlink++;
> > +
> > +			cur_offset += name_len + sizeof(*extref);
> > +		}
> > +
> > +		offset++;
> > +	}
> > +	btrfs_release_path(path);
> > +
> > +	return (ret == 0) ? nlink : ret;
> 
> You depend on btrfs_find_one_extref never returning >0 here.
I''d make that ...
> 
>  if (ret < 0)
>  	return ret;
>  return nlink;
> 
> This ignores ret > 0 explicitly, which is in my opinion better than a
spurious
> link count.
Sounds good.

> > +}
> > +
> > +static int count_inode_refs(struct btrfs_root *root,
> > +			       struct inode *inode, struct btrfs_path *path)
> >  {
> > -	struct btrfs_path *path;
> >  	int ret;
> >  	struct btrfs_key key;
> > -	u64 nlink = 0;
> > +	unsigned int nlink = 0;
> >  	unsigned long ptr;
> >  	unsigned long ptr_end;
> >  	int name_len;
> > @@ -994,10 +1149,6 @@ static noinline int
fixup_inode_link_count(struct btrfs_trans_handle *trans,
> >  	key.type = BTRFS_INODE_REF_KEY;
> >  	key.offset = (u64)-1;
> >  
> > -	path = btrfs_alloc_path();
> > -	if (!path)
> > -		return -ENOMEM;
> > -
> >  	while (1) {
> >  		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
> >  		if (ret < 0)
> > @@ -1031,6 +1182,45 @@ static noinline int
fixup_inode_link_count(struct btrfs_trans_handle *trans,
> >  		btrfs_release_path(path);
> >  	}
> >  	btrfs_release_path(path);
> > +
> > +	return nlink;
> > +}
> > +
> > +/*
> > + * There are a few corners where the link count of the file
can''t
> > + * be properly maintained during replay.  So, instead of adding
> > + * lots of complexity to the log code, we just scan the backrefs
> > + * for any file that has been through replay.
> > + *
> > + * The scan will update the link count on the inode to reflect the
> > + * number of back refs found.  If it goes down to zero, the iput
> > + * will free the inode.
> > + */
> > +static noinline int fixup_inode_link_count(struct btrfs_trans_handle
*trans,
> > +					   struct btrfs_root *root,
> > +					   struct inode *inode)
> > +{
> > +	struct btrfs_path *path;
> > +	int ret;
> > +	u64 nlink = 0;
> > +	u64 ino = btrfs_ino(inode);
> > +
> > +	path = btrfs_alloc_path();
> > +	if (!path)
> > +		return -ENOMEM;
> > +
> > +	ret = count_inode_refs(root, inode, path);
> > +	if (ret < 0)
> > +		goto out;
> > +
> > +	nlink = ret;
> > +
> > +	ret = count_inode_extrefs(root, inode, path);
> > +	if (ret < 0)
> > +		goto out;
> > +
> > +	nlink += ret;
> > +
> >  	if (nlink != inode->i_nlink) {
> >  		set_nlink(inode, nlink);
> >  		btrfs_update_inode(trans, root, inode);
> > @@ -1046,9 +1236,10 @@ static noinline int
fixup_inode_link_count(struct btrfs_trans_handle *trans,
> >  		ret = insert_orphan_item(trans, root, ino);
> >  		BUG_ON(ret);
> >  	}
> > -	btrfs_free_path(path);
> >  
> > -	return 0;
> > +out:
> > +	btrfs_free_path(path);
> > +	return ret;
> >  }
> >  
> >  static noinline int fixup_inode_link_counts(struct btrfs_trans_handle
*trans,
> > @@ -1695,6 +1886,10 @@ static int replay_one_buffer(struct btrfs_root
*log, struct extent_buffer *eb,
> >  			ret = add_inode_ref(wc->trans, root, log, path,
> >  					    eb, i, &key);
> >  			BUG_ON(ret && ret != -ENOENT);
> > +		} else if (key.type == BTRFS_INODE_EXTREF_KEY) {
> > +			ret = add_inode_ref(wc->trans, root, log, path,
> > +					    eb, i, &key);
> > +			BUG_ON(ret && ret != -ENOENT);
> >  		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
> >  			ret = replay_one_extent(wc->trans, root, path,
> >  						eb, i, &key);
> 
> That''s it for this round, here comes the promised patch. Should
apply with "git
> am --scissors".
Fantastic, thanks.
> Can you please add a version to the subject before sending the revised
patches?
> Like "git format-patch --subject-prefix v4".
Sure, no problem. I''ll start the next round at v4 unless you''d
like some other starting
value :)


Thanks again for such valuable review Jan!
	--Mark
> 
> Thanks,
> -Jan
> 
> -- >8 --
> Subject: [PATCH] Btrfs: improved readablity for add_inode_ref
> 
> Moved part of the code into a sub function and replaced most of the gotos
> by ifs, hoping that it will be easier to read now.
> 
> Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
> ---
>  fs/btrfs/tree-log.c |  178
++++++++++++++++++++++++++++-----------------------
>  1 files changed, 97 insertions(+), 81 deletions(-)
> 
> diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
> index c86670f..59f3071 100644
> --- a/fs/btrfs/tree-log.c
> +++ b/fs/btrfs/tree-log.c
> @@ -786,76 +786,18 @@ out:
>  	return match;
>  }
> 
> -
> -/*
> - * replay one inode back reference item found in the log tree.
> - * eb, slot and key refer to the buffer and key found in the log tree.
> - * root is the destination we are replaying into, and path is for temp
> - * use by this function.  (it should be released on return).
> - */
> -static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
> +static inline int __add_inode_ref(struct btrfs_trans_handle *trans,
>  				  struct btrfs_root *root,
> -				  struct btrfs_root *log,
>  				  struct btrfs_path *path,
> -				  struct extent_buffer *eb, int slot,
> -				  struct btrfs_key *key)
> +				  struct btrfs_root *log_root,
> +				  struct inode *dir, struct inode *inode,
> +				  struct btrfs_key *key,
> +				  struct extent_buffer *eb,
> +				  struct btrfs_inode_ref *ref,
> +				  char *name, int namelen, int *search_done)
>  {
> -	struct btrfs_inode_ref *ref;
> -	struct btrfs_dir_item *di;
> -	struct inode *dir;
> -	struct inode *inode;
> -	unsigned long ref_ptr;
> -	unsigned long ref_end;
> -	char *name;
> -	int namelen;
>  	int ret;
> -	int search_done = 0;
> -
> -	/*
> -	 * it is possible that we didn''t log all the parent directories
> -	 * for a given inode.  If we don''t find the dir, just
don''t
> -	 * copy the back ref in.  The link count fixup code will take
> -	 * care of the rest
> -	 */
> -	dir = read_one_inode(root, key->offset);
> -	if (!dir)
> -		return -ENOENT;
> -
> -	inode = read_one_inode(root, key->objectid);
> -	if (!inode) {
> -		iput(dir);
> -		return -EIO;
> -	}
> -
> -	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> -	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> -
> -again:
> -	ref = (struct btrfs_inode_ref *)ref_ptr;
> -
> -	namelen = btrfs_inode_ref_name_len(eb, ref);
> -	name = kmalloc(namelen, GFP_NOFS);
> -	BUG_ON(!name);
> -
> -	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
> -
> -	/* if we already have a perfect match, we''re done */
> -	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
> -			 btrfs_inode_ref_index(eb, ref),
> -			 name, namelen)) {
> -		goto out;
> -	}
> -
> -	/*
> -	 * look for a conflicting back reference in the metadata.
> -	 * if we find one we have to unlink that name of the file
> -	 * before we add our new link.  Later on, we overwrite any
> -	 * existing back reference, and we don''t want to create
> -	 * dangling pointers in the directory.
> -	 */
> -
> -	if (search_done)
> -		goto insert;
> +	struct btrfs_dir_item *di;
> 
>  	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
>  	if (ret == 0) {
> @@ -870,7 +812,7 @@ again:
>  		 * if so, just jump out, we''re done
>  		 */
>  		if (key->objectid == key->offset)
> -			goto out_nowrite;
> +			return 1;
> 
>  		/* check all the names in this back reference to see
>  		 * if they are in the log.  if so, we allow them to stay
> @@ -889,7 +831,7 @@ again:
>  					   (unsigned long)(victim_ref + 1),
>  					   victim_name_len);
> 
> -			if (!backref_in_log(log, key, victim_name,
> +			if (!backref_in_log(log_root, key, victim_name,
>  					    victim_name_len)) {
>  				btrfs_inc_nlink(inode);
>  				btrfs_release_path(path);
> @@ -908,7 +850,7 @@ again:
>  		 * NOTE: we have searched root tree and checked the
>  		 * coresponding ref, it does not need to check again.
>  		 */
> -		search_done = 1;
> +		*search_done = 1;
>  	}
>  	btrfs_release_path(path);
> 
> @@ -931,25 +873,99 @@ again:
>  	}
>  	btrfs_release_path(path);
> 
> -insert:
> -	/* insert our name */
> -	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
> -			     btrfs_inode_ref_index(eb, ref));
> -	BUG_ON(ret);
> +	return 0;
> +}
> +
> +/*
> + * replay one inode back reference item found in the log tree.
> + * eb, slot and key refer to the buffer and key found in the log tree.
> + * root is the destination we are replaying into, and path is for temp
> + * use by this function.  (it should be released on return).
> + */
> +static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
> +				  struct btrfs_root *root,
> +				  struct btrfs_root *log,
> +				  struct btrfs_path *path,
> +				  struct extent_buffer *eb, int slot,
> +				  struct btrfs_key *key)
> +{
> +	struct btrfs_inode_ref *ref;
> +	struct inode *dir;
> +	struct inode *inode;
> +	unsigned long ref_ptr;
> +	unsigned long ref_end;
> +	char *name;
> +	int namelen;
> +	int ret;
> +	int search_done = 0;
> +
> +	/*
> +	 * it is possible that we didn''t log all the parent directories
> +	 * for a given inode.  If we don''t find the dir, just
don''t
> +	 * copy the back ref in.  The link count fixup code will take
> +	 * care of the rest
> +	 */
> +	dir = read_one_inode(root, key->offset);
> +	if (!dir)
> +		return -ENOENT;
> 
> -	btrfs_update_inode(trans, root, inode);
> +	inode = read_one_inode(root, key->objectid);
> +	if (!inode) {
> +		iput(dir);
> +		return -EIO;
> +	}
> 
> -out:
> -	ref_ptr = (unsigned long)(ref + 1) + namelen;
> -	kfree(name);
> -	if (ref_ptr < ref_end)
> -		goto again;
> +	ref_ptr = btrfs_item_ptr_offset(eb, slot);
> +	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
> +
> +	while (ref_ptr < ref_end) {
> +		ref = (struct btrfs_inode_ref *)ref_ptr;
> +
> +		namelen = btrfs_inode_ref_name_len(eb, ref);
> +		name = kmalloc(namelen, GFP_NOFS);
> +		BUG_ON(!name);
> +
> +		read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
> +
> +		/* if we already have a perfect match, we''re done */
> +		if (!inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
> +				  btrfs_inode_ref_index(eb, ref),
> +				  name, namelen)) {
> +			/*
> +			 * look for a conflicting back reference in the
> +			 * metadata. if we find one we have to unlink that name
> +			 * of the file before we add our new link.  Later on, we
> +			 * overwrite any existing back reference, and we don''t
> +			 * want to create dangling pointers in the directory.
> +			 */
> +
> +			if (!search_done) {
> +				ret = __add_inode_ref(trans, root, path, log,
> +						      dir, inode, key, eb, ref,
> +						      name, namelen,
> +						      &search_done);
> +				if (ret == 1)
> +					goto out;
> +				BUG_ON(ret);
> +			}
> +
> +			/* insert our name */
> +			ret = btrfs_add_link(trans, dir, inode, name, namelen,
> +					     0, btrfs_inode_ref_index(eb, ref));
> +			BUG_ON(ret);
> +
> +			btrfs_update_inode(trans, root, inode);
> +		}
> +
> +		ref_ptr = (unsigned long)(ref + 1) + namelen;
> +		kfree(name);
> +	}
> 
>  	/* finally write the back reference in the inode */
>  	ret = overwrite_item(trans, root, path, eb, slot, key);
>  	BUG_ON(ret);
> 
> -out_nowrite:
> +out:
>  	btrfs_release_path(path);
>  	iput(dir);
>  	iput(inode);
> -- 
> 1.7.1
--
Mark Fasheh
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html