Update Linux to v5.4.2
Change-Id: Idf6911045d9d382da2cfe01b1edff026404ac8fd
diff --git a/fs/ext4/block_validity.c b/fs/ext4/block_validity.c
index 913061c..d4d4fdf 100644
--- a/fs/ext4/block_validity.c
+++ b/fs/ext4/block_validity.c
@@ -38,6 +38,7 @@
void ext4_exit_system_zone(void)
{
+ rcu_barrier();
kmem_cache_destroy(ext4_system_zone_cachep);
}
@@ -49,17 +50,26 @@
return 0;
}
+static void release_system_zone(struct ext4_system_blocks *system_blks)
+{
+ struct ext4_system_zone *entry, *n;
+
+ rbtree_postorder_for_each_entry_safe(entry, n,
+ &system_blks->root, node)
+ kmem_cache_free(ext4_system_zone_cachep, entry);
+}
+
/*
* Mark a range of blocks as belonging to the "system zone" --- that
* is, filesystem metadata blocks which should never be used by
* inodes.
*/
-static int add_system_zone(struct ext4_sb_info *sbi,
+static int add_system_zone(struct ext4_system_blocks *system_blks,
ext4_fsblk_t start_blk,
unsigned int count)
{
struct ext4_system_zone *new_entry = NULL, *entry;
- struct rb_node **n = &sbi->system_blks.rb_node, *node;
+ struct rb_node **n = &system_blks->root.rb_node, *node;
struct rb_node *parent = NULL, *new_node = NULL;
while (*n) {
@@ -91,7 +101,7 @@
new_node = &new_entry->node;
rb_link_node(new_node, parent, n);
- rb_insert_color(new_node, &sbi->system_blks);
+ rb_insert_color(new_node, &system_blks->root);
}
/* Can we merge to the left? */
@@ -101,7 +111,7 @@
if (can_merge(entry, new_entry)) {
new_entry->start_blk = entry->start_blk;
new_entry->count += entry->count;
- rb_erase(node, &sbi->system_blks);
+ rb_erase(node, &system_blks->root);
kmem_cache_free(ext4_system_zone_cachep, entry);
}
}
@@ -112,7 +122,7 @@
entry = rb_entry(node, struct ext4_system_zone, node);
if (can_merge(new_entry, entry)) {
new_entry->count += entry->count;
- rb_erase(node, &sbi->system_blks);
+ rb_erase(node, &system_blks->root);
kmem_cache_free(ext4_system_zone_cachep, entry);
}
}
@@ -126,7 +136,7 @@
int first = 1;
printk(KERN_INFO "System zones: ");
- node = rb_first(&sbi->system_blks);
+ node = rb_first(&sbi->system_blks->root);
while (node) {
entry = rb_entry(node, struct ext4_system_zone, node);
printk(KERN_CONT "%s%llu-%llu", first ? "" : ", ",
@@ -137,68 +147,18 @@
printk(KERN_CONT "\n");
}
-int ext4_setup_system_zone(struct super_block *sb)
-{
- ext4_group_t ngroups = ext4_get_groups_count(sb);
- struct ext4_sb_info *sbi = EXT4_SB(sb);
- struct ext4_group_desc *gdp;
- ext4_group_t i;
- int flex_size = ext4_flex_bg_size(sbi);
- int ret;
-
- if (!test_opt(sb, BLOCK_VALIDITY)) {
- if (sbi->system_blks.rb_node)
- ext4_release_system_zone(sb);
- return 0;
- }
- if (sbi->system_blks.rb_node)
- return 0;
-
- for (i=0; i < ngroups; i++) {
- if (ext4_bg_has_super(sb, i) &&
- ((i < 5) || ((i % flex_size) == 0)))
- add_system_zone(sbi, ext4_group_first_block_no(sb, i),
- ext4_bg_num_gdb(sb, i) + 1);
- gdp = ext4_get_group_desc(sb, i, NULL);
- ret = add_system_zone(sbi, ext4_block_bitmap(sb, gdp), 1);
- if (ret)
- return ret;
- ret = add_system_zone(sbi, ext4_inode_bitmap(sb, gdp), 1);
- if (ret)
- return ret;
- ret = add_system_zone(sbi, ext4_inode_table(sb, gdp),
- sbi->s_itb_per_group);
- if (ret)
- return ret;
- }
-
- if (test_opt(sb, DEBUG))
- debug_print_tree(sbi);
- return 0;
-}
-
-/* Called when the filesystem is unmounted */
-void ext4_release_system_zone(struct super_block *sb)
-{
- struct ext4_system_zone *entry, *n;
-
- rbtree_postorder_for_each_entry_safe(entry, n,
- &EXT4_SB(sb)->system_blks, node)
- kmem_cache_free(ext4_system_zone_cachep, entry);
-
- EXT4_SB(sb)->system_blks = RB_ROOT;
-}
-
/*
* Returns 1 if the passed-in block region (start_blk,
* start_blk+count) is valid; 0 if some part of the block region
* overlaps with filesystem metadata blocks.
*/
-int ext4_data_block_valid(struct ext4_sb_info *sbi, ext4_fsblk_t start_blk,
- unsigned int count)
+static int ext4_data_block_valid_rcu(struct ext4_sb_info *sbi,
+ struct ext4_system_blocks *system_blks,
+ ext4_fsblk_t start_blk,
+ unsigned int count)
{
struct ext4_system_zone *entry;
- struct rb_node *n = sbi->system_blks.rb_node;
+ struct rb_node *n;
if ((start_blk <= le32_to_cpu(sbi->s_es->s_first_data_block)) ||
(start_blk + count < start_blk) ||
@@ -206,6 +166,11 @@
sbi->s_es->s_last_error_block = cpu_to_le64(start_blk);
return 0;
}
+
+ if (system_blks == NULL)
+ return 1;
+
+ n = system_blks->root.rb_node;
while (n) {
entry = rb_entry(n, struct ext4_system_zone, node);
if (start_blk + count - 1 < entry->start_blk)
@@ -220,6 +185,178 @@
return 1;
}
+static int ext4_protect_reserved_inode(struct super_block *sb,
+ struct ext4_system_blocks *system_blks,
+ u32 ino)
+{
+ struct inode *inode;
+ struct ext4_sb_info *sbi = EXT4_SB(sb);
+ struct ext4_map_blocks map;
+ u32 i = 0, num;
+ int err = 0, n;
+
+ if ((ino < EXT4_ROOT_INO) ||
+ (ino > le32_to_cpu(sbi->s_es->s_inodes_count)))
+ return -EINVAL;
+ inode = ext4_iget(sb, ino, EXT4_IGET_SPECIAL);
+ if (IS_ERR(inode))
+ return PTR_ERR(inode);
+ num = (inode->i_size + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
+ while (i < num) {
+ map.m_lblk = i;
+ map.m_len = num - i;
+ n = ext4_map_blocks(NULL, inode, &map, 0);
+ if (n < 0) {
+ err = n;
+ break;
+ }
+ if (n == 0) {
+ i++;
+ } else {
+ if (!ext4_data_block_valid_rcu(sbi, system_blks,
+ map.m_pblk, n)) {
+ ext4_error(sb, "blocks %llu-%llu from inode %u "
+ "overlap system zone", map.m_pblk,
+ map.m_pblk + map.m_len - 1, ino);
+ err = -EFSCORRUPTED;
+ break;
+ }
+ err = add_system_zone(system_blks, map.m_pblk, n);
+ if (err < 0)
+ break;
+ i += n;
+ }
+ }
+ iput(inode);
+ return err;
+}
+
+static void ext4_destroy_system_zone(struct rcu_head *rcu)
+{
+ struct ext4_system_blocks *system_blks;
+
+ system_blks = container_of(rcu, struct ext4_system_blocks, rcu);
+ release_system_zone(system_blks);
+ kfree(system_blks);
+}
+
+/*
+ * Build system zone rbtree which is used for block validity checking.
+ *
+ * The update of system_blks pointer in this function is protected by
+ * sb->s_umount semaphore. However we have to be careful as we can be
+ * racing with ext4_data_block_valid() calls reading system_blks rbtree
+ * protected only by RCU. That's why we first build the rbtree and then
+ * swap it in place.
+ */
+int ext4_setup_system_zone(struct super_block *sb)
+{
+ ext4_group_t ngroups = ext4_get_groups_count(sb);
+ struct ext4_sb_info *sbi = EXT4_SB(sb);
+ struct ext4_system_blocks *system_blks;
+ struct ext4_group_desc *gdp;
+ ext4_group_t i;
+ int flex_size = ext4_flex_bg_size(sbi);
+ int ret;
+
+ if (!test_opt(sb, BLOCK_VALIDITY)) {
+ if (sbi->system_blks)
+ ext4_release_system_zone(sb);
+ return 0;
+ }
+ if (sbi->system_blks)
+ return 0;
+
+ system_blks = kzalloc(sizeof(*system_blks), GFP_KERNEL);
+ if (!system_blks)
+ return -ENOMEM;
+
+ for (i=0; i < ngroups; i++) {
+ cond_resched();
+ if (ext4_bg_has_super(sb, i) &&
+ ((i < 5) || ((i % flex_size) == 0)))
+ add_system_zone(system_blks,
+ ext4_group_first_block_no(sb, i),
+ ext4_bg_num_gdb(sb, i) + 1);
+ gdp = ext4_get_group_desc(sb, i, NULL);
+ ret = add_system_zone(system_blks,
+ ext4_block_bitmap(sb, gdp), 1);
+ if (ret)
+ goto err;
+ ret = add_system_zone(system_blks,
+ ext4_inode_bitmap(sb, gdp), 1);
+ if (ret)
+ goto err;
+ ret = add_system_zone(system_blks,
+ ext4_inode_table(sb, gdp),
+ sbi->s_itb_per_group);
+ if (ret)
+ goto err;
+ }
+ if (ext4_has_feature_journal(sb) && sbi->s_es->s_journal_inum) {
+ ret = ext4_protect_reserved_inode(sb, system_blks,
+ le32_to_cpu(sbi->s_es->s_journal_inum));
+ if (ret)
+ goto err;
+ }
+
+ /*
+ * System blks rbtree complete, announce it once to prevent racing
+ * with ext4_data_block_valid() accessing the rbtree at the same
+ * time.
+ */
+ rcu_assign_pointer(sbi->system_blks, system_blks);
+
+ if (test_opt(sb, DEBUG))
+ debug_print_tree(sbi);
+ return 0;
+err:
+ release_system_zone(system_blks);
+ kfree(system_blks);
+ return ret;
+}
+
+/*
+ * Called when the filesystem is unmounted or when remounting it with
+ * noblock_validity specified.
+ *
+ * The update of system_blks pointer in this function is protected by
+ * sb->s_umount semaphore. However we have to be careful as we can be
+ * racing with ext4_data_block_valid() calls reading system_blks rbtree
+ * protected only by RCU. So we first clear the system_blks pointer and
+ * then free the rbtree only after RCU grace period expires.
+ */
+void ext4_release_system_zone(struct super_block *sb)
+{
+ struct ext4_system_blocks *system_blks;
+
+ system_blks = rcu_dereference_protected(EXT4_SB(sb)->system_blks,
+ lockdep_is_held(&sb->s_umount));
+ rcu_assign_pointer(EXT4_SB(sb)->system_blks, NULL);
+
+ if (system_blks)
+ call_rcu(&system_blks->rcu, ext4_destroy_system_zone);
+}
+
+int ext4_data_block_valid(struct ext4_sb_info *sbi, ext4_fsblk_t start_blk,
+ unsigned int count)
+{
+ struct ext4_system_blocks *system_blks;
+ int ret;
+
+ /*
+ * Lock the system zone to prevent it being released concurrently
+ * when doing a remount which inverse current "[no]block_validity"
+ * mount option.
+ */
+ rcu_read_lock();
+ system_blks = rcu_dereference(sbi->system_blks);
+ ret = ext4_data_block_valid_rcu(sbi, system_blks, start_blk,
+ count);
+ rcu_read_unlock();
+ return ret;
+}
+
int ext4_check_blockref(const char *function, unsigned int line,
struct inode *inode, __le32 *p, unsigned int max)
{
@@ -227,6 +364,11 @@
__le32 *bref = p;
unsigned int blk;
+ if (ext4_has_feature_journal(inode->i_sb) &&
+ (inode->i_ino ==
+ le32_to_cpu(EXT4_SB(inode->i_sb)->s_es->s_journal_inum)))
+ return 0;
+
while (bref < p+max) {
blk = le32_to_cpu(*bref++);
if (blk &&