summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--MAINTAINERS6
-rw-r--r--fs/btrfs/backref.c84
-rw-r--r--fs/btrfs/backref.h1
-rw-r--r--fs/btrfs/block-group.c2
-rw-r--r--fs/btrfs/extent-io-tree.c15
-rw-r--r--fs/btrfs/send.c5
-rw-r--r--fs/btrfs/send.h5
-rw-r--r--fs/erofs/fscache.c3
-rw-r--r--fs/erofs/zdata.c17
-rw-r--r--fs/erofs/zdata.h6
-rw-r--r--fs/erofs/zmap.c22
11 files changed, 114 insertions, 52 deletions
diff --git a/MAINTAINERS b/MAINTAINERS
index 07938eab80dd..ca063a504026 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -4459,13 +4459,15 @@ M: Josef Bacik <josef@toxicpanda.com>
M: David Sterba <dsterba@suse.com>
L: linux-btrfs@vger.kernel.org
S: Maintained
-W: http://btrfs.wiki.kernel.org/
-Q: http://patchwork.kernel.org/project/linux-btrfs/list/
+W: https://btrfs.readthedocs.io
+W: https://btrfs.wiki.kernel.org/
+Q: https://patchwork.kernel.org/project/linux-btrfs/list/
C: irc://irc.libera.chat/btrfs
T: git git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux.git
F: Documentation/filesystems/btrfs.rst
F: fs/btrfs/
F: include/linux/btrfs*
+F: include/trace/events/btrfs.h
F: include/uapi/linux/btrfs*
BTTV VIDEO4LINUX DRIVER
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index dce3a16996b9..4ec18ceb2f21 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -138,6 +138,7 @@ struct share_check {
u64 root_objectid;
u64 inum;
int share_count;
+ bool have_delayed_delete_refs;
};
static inline int extent_is_shared(struct share_check *sc)
@@ -820,16 +821,11 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
struct preftrees *preftrees, struct share_check *sc)
{
struct btrfs_delayed_ref_node *node;
- struct btrfs_delayed_extent_op *extent_op = head->extent_op;
struct btrfs_key key;
- struct btrfs_key tmp_op_key;
struct rb_node *n;
int count;
int ret = 0;
- if (extent_op && extent_op->update_key)
- btrfs_disk_key_to_cpu(&tmp_op_key, &extent_op->key);
-
spin_lock(&head->lock);
for (n = rb_first_cached(&head->ref_tree); n; n = rb_next(n)) {
node = rb_entry(n, struct btrfs_delayed_ref_node,
@@ -855,10 +851,16 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
case BTRFS_TREE_BLOCK_REF_KEY: {
/* NORMAL INDIRECT METADATA backref */
struct btrfs_delayed_tree_ref *ref;
+ struct btrfs_key *key_ptr = NULL;
+
+ if (head->extent_op && head->extent_op->update_key) {
+ btrfs_disk_key_to_cpu(&key, &head->extent_op->key);
+ key_ptr = &key;
+ }
ref = btrfs_delayed_node_to_tree_ref(node);
ret = add_indirect_ref(fs_info, preftrees, ref->root,
- &tmp_op_key, ref->level + 1,
+ key_ptr, ref->level + 1,
node->bytenr, count, sc,
GFP_ATOMIC);
break;
@@ -884,13 +886,22 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
key.offset = ref->offset;
/*
- * Found a inum that doesn't match our known inum, we
- * know it's shared.
+ * If we have a share check context and a reference for
+ * another inode, we can't exit immediately. This is
+ * because even if this is a BTRFS_ADD_DELAYED_REF
+ * reference we may find next a BTRFS_DROP_DELAYED_REF
+ * which cancels out this ADD reference.
+ *
+ * If this is a DROP reference and there was no previous
+ * ADD reference, then we need to signal that when we
+ * process references from the extent tree (through
+ * add_inline_refs() and add_keyed_refs()), we should
+ * not exit early if we find a reference for another
+ * inode, because one of the delayed DROP references
+ * may cancel that reference in the extent tree.
*/
- if (sc && sc->inum && ref->objectid != sc->inum) {
- ret = BACKREF_FOUND_SHARED;
- goto out;
- }
+ if (sc && count < 0)
+ sc->have_delayed_delete_refs = true;
ret = add_indirect_ref(fs_info, preftrees, ref->root,
&key, 0, node->bytenr, count, sc,
@@ -920,7 +931,7 @@ static int add_delayed_refs(const struct btrfs_fs_info *fs_info,
}
if (!ret)
ret = extent_is_shared(sc);
-out:
+
spin_unlock(&head->lock);
return ret;
}
@@ -1023,7 +1034,8 @@ static int add_inline_refs(const struct btrfs_fs_info *fs_info,
key.type = BTRFS_EXTENT_DATA_KEY;
key.offset = btrfs_extent_data_ref_offset(leaf, dref);
- if (sc && sc->inum && key.objectid != sc->inum) {
+ if (sc && sc->inum && key.objectid != sc->inum &&
+ !sc->have_delayed_delete_refs) {
ret = BACKREF_FOUND_SHARED;
break;
}
@@ -1033,6 +1045,7 @@ static int add_inline_refs(const struct btrfs_fs_info *fs_info,
ret = add_indirect_ref(fs_info, preftrees, root,
&key, 0, bytenr, count,
sc, GFP_NOFS);
+
break;
}
default:
@@ -1122,7 +1135,8 @@ static int add_keyed_refs(struct btrfs_root *extent_root,
key.type = BTRFS_EXTENT_DATA_KEY;
key.offset = btrfs_extent_data_ref_offset(leaf, dref);
- if (sc && sc->inum && key.objectid != sc->inum) {
+ if (sc && sc->inum && key.objectid != sc->inum &&
+ !sc->have_delayed_delete_refs) {
ret = BACKREF_FOUND_SHARED;
break;
}
@@ -1522,6 +1536,9 @@ static bool lookup_backref_shared_cache(struct btrfs_backref_shared_cache *cache
{
struct btrfs_backref_shared_cache_entry *entry;
+ if (!cache->use_cache)
+ return false;
+
if (WARN_ON_ONCE(level >= BTRFS_MAX_LEVEL))
return false;
@@ -1557,6 +1574,19 @@ static bool lookup_backref_shared_cache(struct btrfs_backref_shared_cache *cache
return false;
*is_shared = entry->is_shared;
+ /*
+ * If the node at this level is shared, than all nodes below are also
+ * shared. Currently some of the nodes below may be marked as not shared
+ * because we have just switched from one leaf to another, and switched
+ * also other nodes above the leaf and below the current level, so mark
+ * them as shared.
+ */
+ if (*is_shared) {
+ for (int i = 0; i < level; i++) {
+ cache->entries[i].is_shared = true;
+ cache->entries[i].gen = entry->gen;
+ }
+ }
return true;
}
@@ -1573,6 +1603,9 @@ static void store_backref_shared_cache(struct btrfs_backref_shared_cache *cache,
struct btrfs_backref_shared_cache_entry *entry;
u64 gen;
+ if (!cache->use_cache)
+ return;
+
if (WARN_ON_ONCE(level >= BTRFS_MAX_LEVEL))
return;
@@ -1648,6 +1681,7 @@ int btrfs_is_data_extent_shared(struct btrfs_root *root, u64 inum, u64 bytenr,
.root_objectid = root->root_key.objectid,
.inum = inum,
.share_count = 0,
+ .have_delayed_delete_refs = false,
};
int level;
@@ -1669,6 +1703,7 @@ int btrfs_is_data_extent_shared(struct btrfs_root *root, u64 inum, u64 bytenr,
/* -1 means we are in the bytenr of the data extent. */
level = -1;
ULIST_ITER_INIT(&uiter);
+ cache->use_cache = true;
while (1) {
bool is_shared;
bool cached;
@@ -1698,6 +1733,24 @@ int btrfs_is_data_extent_shared(struct btrfs_root *root, u64 inum, u64 bytenr,
extent_gen > btrfs_root_last_snapshot(&root->root_item))
break;
+ /*
+ * If our data extent was not directly shared (without multiple
+ * reference items), than it might have a single reference item
+ * with a count > 1 for the same offset, which means there are 2
+ * (or more) file extent items that point to the data extent -
+ * this happens when a file extent item needs to be split and
+ * then one item gets moved to another leaf due to a b+tree leaf
+ * split when inserting some item. In this case the file extent
+ * items may be located in different leaves and therefore some
+ * of the leaves may be referenced through shared subtrees while
+ * others are not. Since our extent buffer cache only works for
+ * a single path (by far the most common case and simpler to
+ * deal with), we can not use it if we have multiple leaves
+ * (which implies multiple paths).
+ */
+ if (level == -1 && tmp->nnodes > 1)
+ cache->use_cache = false;
+
if (level >= 0)
store_backref_shared_cache(cache, root, bytenr,
level, false);
@@ -1713,6 +1766,7 @@ int btrfs_is_data_extent_shared(struct btrfs_root *root, u64 inum, u64 bytenr,
break;
}
shared.share_count = 0;
+ shared.have_delayed_delete_refs = false;
cond_resched();
}
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index 52ae6957b414..8e69584d538d 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -29,6 +29,7 @@ struct btrfs_backref_shared_cache {
* a given data extent should never exceed the maximum b+tree height.
*/
struct btrfs_backref_shared_cache_entry entries[BTRFS_MAX_LEVEL];
+ bool use_cache;
};
typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root,
diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 32c415cfbdfe..deebc8ddbd93 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -774,10 +774,8 @@ int btrfs_cache_block_group(struct btrfs_block_group *cache, bool wait)
btrfs_queue_work(fs_info->caching_workers, &caching_ctl->work);
out:
- /* REVIEW */
if (wait && caching_ctl)
ret = btrfs_caching_ctl_wait_done(cache, caching_ctl);
- /* wait_event(caching_ctl->wait, space_cache_v1_done(cache)); */
if (caching_ctl)
btrfs_put_caching_control(caching_ctl);
diff --git a/fs/btrfs/extent-io-tree.c b/fs/btrfs/extent-io-tree.c
index 618275af19c4..83cb0378096f 100644
--- a/fs/btrfs/extent-io-tree.c
+++ b/fs/btrfs/extent-io-tree.c
@@ -1641,16 +1641,17 @@ int lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
int err;
u64 failed_start;
- while (1) {
+ err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, &failed_start,
+ cached_state, NULL, GFP_NOFS);
+ while (err == -EEXIST) {
+ if (failed_start != start)
+ clear_extent_bit(tree, start, failed_start - 1,
+ EXTENT_LOCKED, cached_state);
+
+ wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
err = __set_extent_bit(tree, start, end, EXTENT_LOCKED,
&failed_start, cached_state, NULL,
GFP_NOFS);
- if (err == -EEXIST) {
- wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
- start = failed_start;
- } else
- break;
- WARN_ON(start > end);
}
return err;
}
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 4ef4167072b8..ec6e1752af2c 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -348,6 +348,7 @@ static bool proto_cmd_ok(const struct send_ctx *sctx, int cmd)
switch (sctx->proto) {
case 1: return cmd <= BTRFS_SEND_C_MAX_V1;
case 2: return cmd <= BTRFS_SEND_C_MAX_V2;
+ case 3: return cmd <= BTRFS_SEND_C_MAX_V3;
default: return false;
}
}
@@ -6469,7 +6470,9 @@ static int finish_inode_if_needed(struct send_ctx *sctx, int at_end)
if (ret < 0)
goto out;
}
- if (sctx->cur_inode_needs_verity) {
+
+ if (proto_cmd_ok(sctx, BTRFS_SEND_C_ENABLE_VERITY)
+ && sctx->cur_inode_needs_verity) {
ret = process_verity(sctx);
if (ret < 0)
goto out;
diff --git a/fs/btrfs/send.h b/fs/btrfs/send.h
index 0a4537775e0c..f7585cfa7e52 100644
--- a/fs/btrfs/send.h
+++ b/fs/btrfs/send.h
@@ -10,7 +10,12 @@
#include <linux/types.h>
#define BTRFS_SEND_STREAM_MAGIC "btrfs-stream"
+/* Conditional support for the upcoming protocol version. */
+#ifdef CONFIG_BTRFS_DEBUG
+#define BTRFS_SEND_STREAM_VERSION 3
+#else
#define BTRFS_SEND_STREAM_VERSION 2
+#endif
/*
* In send stream v1, no command is larger than 64K. In send stream v2, no limit
diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index 998cd26a1b3b..fe05bc51f9f2 100644
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -590,14 +590,17 @@ struct erofs_fscache *erofs_domain_register_cookie(struct super_block *sb,
struct super_block *psb = erofs_pseudo_mnt->mnt_sb;
mutex_lock(&erofs_domain_cookies_lock);
+ spin_lock(&psb->s_inode_list_lock);
list_for_each_entry(inode, &psb->s_inodes, i_sb_list) {
ctx = inode->i_private;
if (!ctx || ctx->domain != domain || strcmp(ctx->name, name))
continue;
igrab(inode);
+ spin_unlock(&psb->s_inode_list_lock);
mutex_unlock(&erofs_domain_cookies_lock);
return ctx;
}
+ spin_unlock(&psb->s_inode_list_lock);
ctx = erofs_fscache_domain_init_cookie(sb, name, need_inode);
mutex_unlock(&erofs_domain_cookies_lock);
return ctx;
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index 559380a535af..c7f24fc7efd5 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -813,15 +813,14 @@ retry:
++spiltted;
if (fe->pcl->pageofs_out != (map->m_la & ~PAGE_MASK))
fe->pcl->multibases = true;
-
- if ((map->m_flags & EROFS_MAP_FULL_MAPPED) &&
- !(map->m_flags & EROFS_MAP_PARTIAL_REF) &&
- fe->pcl->length == map->m_llen)
- fe->pcl->partial = false;
if (fe->pcl->length < offset + end - map->m_la) {
fe->pcl->length = offset + end - map->m_la;
fe->pcl->pageofs_out = map->m_la & ~PAGE_MASK;
}
+ if ((map->m_flags & EROFS_MAP_FULL_MAPPED) &&
+ !(map->m_flags & EROFS_MAP_PARTIAL_REF) &&
+ fe->pcl->length == map->m_llen)
+ fe->pcl->partial = false;
next_part:
/* shorten the remaining extent to update progress */
map->m_llen = offset + cur - map->m_la;
@@ -888,15 +887,13 @@ static void z_erofs_do_decompressed_bvec(struct z_erofs_decompress_backend *be,
if (!((bvec->offset + be->pcl->pageofs_out) & ~PAGE_MASK)) {
unsigned int pgnr;
- struct page *oldpage;
pgnr = (bvec->offset + be->pcl->pageofs_out) >> PAGE_SHIFT;
DBG_BUGON(pgnr >= be->nr_pages);
- oldpage = be->decompressed_pages[pgnr];
- be->decompressed_pages[pgnr] = bvec->page;
-
- if (!oldpage)
+ if (!be->decompressed_pages[pgnr]) {
+ be->decompressed_pages[pgnr] = bvec->page;
return;
+ }
}
/* (cold path) one pcluster is requested multiple times */
diff --git a/fs/erofs/zdata.h b/fs/erofs/zdata.h
index e7f04c4fbb81..d98c95212985 100644
--- a/fs/erofs/zdata.h
+++ b/fs/erofs/zdata.h
@@ -126,10 +126,10 @@ static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
}
/*
- * bit 31: I/O error occurred on this page
- * bit 0 - 30: remaining parts to complete this page
+ * bit 30: I/O error occurred on this page
+ * bit 0 - 29: remaining parts to complete this page
*/
-#define Z_EROFS_PAGE_EIO (1 << 31)
+#define Z_EROFS_PAGE_EIO (1 << 30)
static inline void z_erofs_onlinepage_init(struct page *page)
{
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 44c27ef39c43..0bb66927e3d0 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -57,8 +57,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
pos = ALIGN(iloc(EROFS_SB(sb), vi->nid) + vi->inode_isize +
vi->xattr_isize, 8);
- kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos),
- EROFS_KMAP_ATOMIC);
+ kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos), EROFS_KMAP);
if (IS_ERR(kaddr)) {
err = PTR_ERR(kaddr);
goto out_unlock;
@@ -73,7 +72,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
vi->z_advise = Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
vi->z_fragmentoff = le64_to_cpu(*(__le64 *)h) ^ (1ULL << 63);
vi->z_tailextent_headlcn = 0;
- goto unmap_done;
+ goto done;
}
vi->z_advise = le16_to_cpu(h->h_advise);
vi->z_algorithmtype[0] = h->h_algorithmtype & 15;
@@ -85,7 +84,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
erofs_err(sb, "unknown HEAD%u format %u for nid %llu, please upgrade kernel",
headnr + 1, vi->z_algorithmtype[headnr], vi->nid);
err = -EOPNOTSUPP;
- goto unmap_done;
+ goto out_put_metabuf;
}
vi->z_logical_clusterbits = LOG_BLOCK_SIZE + (h->h_clusterbits & 7);
@@ -95,7 +94,7 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
erofs_err(sb, "per-inode big pcluster without sb feature for nid %llu",
vi->nid);
err = -EFSCORRUPTED;
- goto unmap_done;
+ goto out_put_metabuf;
}
if (vi->datalayout == EROFS_INODE_FLAT_COMPRESSION &&
!(vi->z_advise & Z_EROFS_ADVISE_BIG_PCLUSTER_1) ^
@@ -103,12 +102,8 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
erofs_err(sb, "big pcluster head1/2 of compact indexes should be consistent for nid %llu",
vi->nid);
err = -EFSCORRUPTED;
- goto unmap_done;
+ goto out_put_metabuf;
}
-unmap_done:
- erofs_put_metabuf(&buf);
- if (err)
- goto out_unlock;
if (vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER) {
struct erofs_map_blocks map = {
@@ -127,7 +122,7 @@ unmap_done:
err = -EFSCORRUPTED;
}
if (err < 0)
- goto out_unlock;
+ goto out_put_metabuf;
}
if (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER &&
@@ -141,11 +136,14 @@ unmap_done:
EROFS_GET_BLOCKS_FINDTAIL);
erofs_put_metabuf(&map.buf);
if (err < 0)
- goto out_unlock;
+ goto out_put_metabuf;
}
+done:
/* paired with smp_mb() at the beginning of the function */
smp_mb();
set_bit(EROFS_I_Z_INITED_BIT, &vi->flags);
+out_put_metabuf:
+ erofs_put_metabuf(&buf);
out_unlock:
clear_and_wake_up_bit(EROFS_I_BL_Z_BIT, &vi->flags);
return err;
generated by cgit v1.2.3-70-g09d2 (git 2.47.1) at 2024-12-29 07:57:35 +0000