From: Gu Jinxiang <gujx@cn.fujitsu.com>
To: <linux-btrfs@vger.kernel.org>
Cc: <lakshmipathi.g@gmail.com>
Subject: [PATCH 2/5] btrfs-progs: Add logic to judge the level between parent and child
Date: Fri, 24 Nov 2017 18:41:29 +0800 [thread overview]
Message-ID: <1511520092-37101-2-git-send-email-gujx@cn.fujitsu.com> (raw)
In-Reply-To: <1511520092-37101-1-git-send-email-gujx@cn.fujitsu.com>
The following test failed becasuse the level of child node is not
correct. In repair mode, when update parent's ptr in
btrfs_search_slot->btrfs_cow_block->__btrfs_cow_block,
it use path[level+1] to get the parent, and not judge whether
path[level+1] is NULL.
$sudo TEST=003\* make test-fuzz
failed (ignored, ret=139): /home/adam/btrfs/btrfs-progs/btrfs check --init-csum-tree /home/adam/btrfs/btrfs-progs/tests/fuzz-tests/images/bko-161821.raw.restored
mayfail: returned code 139 (SEGFAULT), not ignored
Signed-off-by: Gu Jinxiang <gujx@cn.fujitsu.com>
---
cmds-check.c | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)
diff --git a/cmds-check.c b/cmds-check.c
index 71b15de4..ac0375e5 100644
--- a/cmds-check.c
+++ b/cmds-check.c
@@ -2085,6 +2085,7 @@ static void reada_walk_down(struct btrfs_root *root,
* in parent.
* 2. block in parent node should match the child node/leaf.
* 3. generation of parent node and child's header should be consistent.
+ * 4. level of parent node should not smaller than the child node/leaf.
*
* Or the child node/leaf pointed by the key in parent is not valid.
*
@@ -2125,6 +2126,17 @@ static int check_child_node(struct extent_buffer *parent, int slot,
btrfs_header_generation(child),
btrfs_node_ptr_generation(parent, slot));
}
+ /* If level of child is not correct, it will lead to
+ * Segmentation fault when repair parent.
+ * Because when update parent's ptr in
+ * btrfs_search_slot->btrfs_cow_block->__btrfs_cow_block,
+ * it use path[level+1] to get the parent
+ */
+ if (btrfs_header_level(parent) <= btrfs_header_level(child)) {
+ ret = -EFAULT;
+ fprintf(stderr, "Wrong level of child node/leaf, wanted: %d, have: %d\n",
+ btrfs_header_level(parent)-1, btrfs_header_level(child));
+ }
return ret;
}
@@ -4067,6 +4079,12 @@ static int check_fs_root(struct btrfs_root *root,
wret = walk_down_tree(root, &path, wc, &level, &nrefs);
if (wret < 0)
ret = wret;
+ /* When return value is -EFAULT, it indicate that level
+ * in path is incorrect. And it will lead to Segmentation fault
+ * in repair mod.
+ */
+ if (wret == -EFAULT && repair)
+ goto out;
if (wret != 0)
break;
@@ -4123,6 +4141,7 @@ skip_walking:
if (!ret)
ret = err;
+out:
free_corrupt_blocks_tree(&corrupt_blocks);
root->fs_info->corrupt_blocks = NULL;
free_orphan_data_extents(&root->orphan_data_extents);
--
2.14.3
next prev parent reply other threads:[~2017-11-24 10:41 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-24 10:41 [PATCH 1/5] btrfs-progs: Add location check when process share_data_ref item Gu Jinxiang
2017-11-24 10:41 ` Gu Jinxiang [this message]
2017-11-24 10:41 ` [PATCH 3/5] btrfs-progs: return error to upper caller instead of BUG_ON Gu Jinxiang
2017-11-24 10:41 ` [PATCH 4/5] btrfs-progs: check null pointer before use it Gu Jinxiang
2017-11-24 10:41 ` [PATCH 5/5] btrfs-progs: return error to caller instead of BUG_ON Gu Jinxiang
2017-11-28 19:02 ` [PATCH 1/5] btrfs-progs: Add location check when process share_data_ref item David Sterba
2017-11-30 6:55 ` Qu Wenruo
2017-11-30 18:25 ` David Sterba
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1511520092-37101-2-git-send-email-gujx@cn.fujitsu.com \
--to=gujx@cn.fujitsu.com \
--cc=lakshmipathi.g@gmail.com \
--cc=linux-btrfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).