* [PATCH] btrfs: pass a valid btrfs_tree_parent_check when possible
@ 2026-04-09 23:16 Qu Wenruo
2026-04-10 0:03 ` Boris Burkov
0 siblings, 1 reply; 2+ messages in thread
From: Qu Wenruo @ 2026-04-09 23:16 UTC (permalink / raw)
To: linux-btrfs
Commit 6e181cfe2409 ("btrfs: revalidate cached tree blocks on the
uptodate path") introduced the @check parameter for
btrfs_buffer_uptodate() to allow re-validation of a cached extent
buffer.
But there are still call sites that doesn't utilize this parameter,
which exposes them to possible corrupted tree blocks, e.g. an empty
child leaf of a parent node, which should be rejected by
btrfs_verify_level_key() but if @check is NULL such check will be
skipped and cause problems.
Thankfully for a lot of cases there is already an existing @check
structure around and we can pass it directly to btrfs_buffer_uptodate().
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
fs/btrfs/ctree.c | 16 +++++-----------
fs/btrfs/extent-tree.c | 11 ++++++++---
fs/btrfs/extent_io.c | 2 +-
3 files changed, 14 insertions(+), 15 deletions(-)
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index d70da290bedf..829d8be7f423 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -1497,17 +1497,11 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
if (p->reada == READA_FORWARD_ALWAYS)
reada_for_search(fs_info, p, parent_level, slot, key->objectid);
- /* first we do an atomic uptodate check */
- if (btrfs_buffer_uptodate(tmp, check.transid, NULL) > 0) {
- /*
- * Do extra check for first_key, eb can be stale due to
- * being cached, read from scrub, or have multiple
- * parents (shared tree blocks).
- */
- if (unlikely(btrfs_verify_level_key(tmp, &check))) {
- ret = -EUCLEAN;
- goto out;
- }
+ /* Check if the cached eb is uptodate. */
+ ret = btrfs_buffer_uptodate(tmp, check.transid, &check);
+ if (unlikely(ret < 0))
+ goto out;
+ if (ret > 0) {
*eb_ret = tmp;
tmp = NULL;
ret = 0;
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 391fad41c3b6..5d5b42ea4ade 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -5781,16 +5781,21 @@ static int check_next_block_uptodate(struct btrfs_trans_handle *trans,
generation = btrfs_node_ptr_generation(path->nodes[level], path->slots[level]);
- if (btrfs_buffer_uptodate(next, generation, NULL))
- return 0;
-
check.level = level - 1;
check.transid = generation;
check.owner_root = btrfs_root_id(root);
check.has_first_key = true;
btrfs_node_key_to_cpu(path->nodes[level], &check.first_key, path->slots[level]);
+ ret = btrfs_buffer_uptodate(next, generation, &check);
+ if (ret > 0)
+ return 0;
btrfs_tree_unlock(next);
+ if (ret < 0) {
+ free_extent_buffer(next);
+ return ret;
+ }
+
if (level == 1)
reada_walk_down(trans, root, wc, path);
ret = btrfs_read_extent_buffer(next, &check);
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index a8887285deda..3f59fb0a9ee5 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -4648,7 +4648,7 @@ void btrfs_readahead_tree_block(struct btrfs_fs_info *fs_info,
if (IS_ERR(eb))
return;
- if (btrfs_buffer_uptodate(eb, gen, NULL)) {
+ if (btrfs_buffer_uptodate(eb, gen, &check)) {
free_extent_buffer(eb);
return;
}
--
2.53.0
^ permalink raw reply related [flat|nested] 2+ messages in thread* Re: [PATCH] btrfs: pass a valid btrfs_tree_parent_check when possible
2026-04-09 23:16 [PATCH] btrfs: pass a valid btrfs_tree_parent_check when possible Qu Wenruo
@ 2026-04-10 0:03 ` Boris Burkov
0 siblings, 0 replies; 2+ messages in thread
From: Boris Burkov @ 2026-04-10 0:03 UTC (permalink / raw)
To: Qu Wenruo; +Cc: linux-btrfs
On Fri, Apr 10, 2026 at 08:46:40AM +0930, Qu Wenruo wrote:
> Commit 6e181cfe2409 ("btrfs: revalidate cached tree blocks on the
> uptodate path") introduced the @check parameter for
> btrfs_buffer_uptodate() to allow re-validation of a cached extent
> buffer.
>
> But there are still call sites that doesn't utilize this parameter,
> which exposes them to possible corrupted tree blocks, e.g. an empty
> child leaf of a parent node, which should be rejected by
> btrfs_verify_level_key() but if @check is NULL such check will be
> skipped and cause problems.
>
> Thankfully for a lot of cases there is already an existing @check
> structure around and we can pass it directly to btrfs_buffer_uptodate().
>
Reviewed-by: Boris Burkov <boris@bur.io>
> Signed-off-by: Qu Wenruo <wqu@suse.com>
> ---
> fs/btrfs/ctree.c | 16 +++++-----------
> fs/btrfs/extent-tree.c | 11 ++++++++---
> fs/btrfs/extent_io.c | 2 +-
> 3 files changed, 14 insertions(+), 15 deletions(-)
>
> diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
> index d70da290bedf..829d8be7f423 100644
> --- a/fs/btrfs/ctree.c
> +++ b/fs/btrfs/ctree.c
> @@ -1497,17 +1497,11 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
> if (p->reada == READA_FORWARD_ALWAYS)
> reada_for_search(fs_info, p, parent_level, slot, key->objectid);
>
> - /* first we do an atomic uptodate check */
> - if (btrfs_buffer_uptodate(tmp, check.transid, NULL) > 0) {
> - /*
> - * Do extra check for first_key, eb can be stale due to
> - * being cached, read from scrub, or have multiple
> - * parents (shared tree blocks).
> - */
> - if (unlikely(btrfs_verify_level_key(tmp, &check))) {
> - ret = -EUCLEAN;
> - goto out;
> - }
> + /* Check if the cached eb is uptodate. */
> + ret = btrfs_buffer_uptodate(tmp, check.transid, &check);
> + if (unlikely(ret < 0))
> + goto out;
> + if (ret > 0) {
> *eb_ret = tmp;
> tmp = NULL;
> ret = 0;
> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
> index 391fad41c3b6..5d5b42ea4ade 100644
> --- a/fs/btrfs/extent-tree.c
> +++ b/fs/btrfs/extent-tree.c
> @@ -5781,16 +5781,21 @@ static int check_next_block_uptodate(struct btrfs_trans_handle *trans,
>
> generation = btrfs_node_ptr_generation(path->nodes[level], path->slots[level]);
>
> - if (btrfs_buffer_uptodate(next, generation, NULL))
> - return 0;
> -
> check.level = level - 1;
> check.transid = generation;
> check.owner_root = btrfs_root_id(root);
> check.has_first_key = true;
> btrfs_node_key_to_cpu(path->nodes[level], &check.first_key, path->slots[level]);
>
> + ret = btrfs_buffer_uptodate(next, generation, &check);
> + if (ret > 0)
> + return 0;
> btrfs_tree_unlock(next);
> + if (ret < 0) {
> + free_extent_buffer(next);
> + return ret;
> + }
> +
> if (level == 1)
> reada_walk_down(trans, root, wc, path);
> ret = btrfs_read_extent_buffer(next, &check);
> diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
> index a8887285deda..3f59fb0a9ee5 100644
> --- a/fs/btrfs/extent_io.c
> +++ b/fs/btrfs/extent_io.c
> @@ -4648,7 +4648,7 @@ void btrfs_readahead_tree_block(struct btrfs_fs_info *fs_info,
> if (IS_ERR(eb))
> return;
>
> - if (btrfs_buffer_uptodate(eb, gen, NULL)) {
> + if (btrfs_buffer_uptodate(eb, gen, &check)) {
> free_extent_buffer(eb);
> return;
> }
> --
> 2.53.0
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-04-10 0:03 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-04-09 23:16 [PATCH] btrfs: pass a valid btrfs_tree_parent_check when possible Qu Wenruo
2026-04-10 0:03 ` Boris Burkov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox