Linux Btrfs filesystem development
 help / color / mirror / Atom feed
From: Qu Wenruo <quwenruo.btrfs@gmx.com>
To: Josef Bacik <josef@toxicpanda.com>,
	linux-btrfs@vger.kernel.org, kernel-team@fb.com
Cc: lei lu <llfamsec@gmail.com>
Subject: Re: [PATCH] btrfs: make sure that WRITTEN is set on all metadata blocks
Date: Thu, 2 May 2024 07:15:36 +0930	[thread overview]
Message-ID: <2aec5fb2-f881-416a-b558-cb265886dad7@gmx.com> (raw)
In-Reply-To: <d82bd6cef76e7beaa0d33ef48f9292f3779d015c.1714395805.git.josef@toxicpanda.com>



在 2024/4/29 22:33, Josef Bacik 写道:
> We previously would call btrfs_check_leaf() if we had the check
> integrity code enabled, which meant that we could only run the extended
> leaf checks if we had WRITTEN set on the header flags.
>
> This leaves a gap in our checking, because we could end up with
> corruption on disk where WRITTEN isn't set on the leaf, and then the
> extended leaf checks don't get run which we rely on to validate all of
> the item pointers to make sure we don't access memory outside of the
> extent buffer.
>
> However, since 732fab95abe2 ("btrfs: check-integrity: remove
> CONFIG_BTRFS_FS_CHECK_INTEGRITY option") we no longer call
> btrfs_check_leaf() from btrfs_mark_buffer_dirty(), which means we only
> ever call it on blocks that are being written out, and thus have WRITTEN
> set, or that are being read in, which should have WRITTEN set.
>
> Add checks to make sure we have WRITTEN set appropriately, and then make
> sure __btrfs_check_leaf() always does the item checking.  This will
> protect us from file systems that have been corrupted and no longer have
> WRITTEN set on some of the blocks.
>
> Reported-by: lei lu <llfamsec@gmail.com>
> Signed-off-by: Josef Bacik <josef@toxicpanda.com>

Reviewed-by: Qu Wenruo <wqu@suse.com>

Is there any real world bug report on this? Or just some code reading
exposed this problem?

Thanks,
Qu
> ---
>   fs/btrfs/tree-checker.c | 30 +++++++++++++++---------------
>   fs/btrfs/tree-checker.h |  1 +
>   2 files changed, 16 insertions(+), 15 deletions(-)
>
> diff --git a/fs/btrfs/tree-checker.c b/fs/btrfs/tree-checker.c
> index a127abbc09c3..5a7e869da230 100644
> --- a/fs/btrfs/tree-checker.c
> +++ b/fs/btrfs/tree-checker.c
> @@ -1797,6 +1797,11 @@ enum btrfs_tree_block_status __btrfs_check_leaf(struct extent_buffer *leaf)
>   		return BTRFS_TREE_BLOCK_INVALID_LEVEL;
>   	}
>
> +	if (unlikely(!btrfs_header_flag(leaf, BTRFS_HEADER_FLAG_WRITTEN))) {
> +		generic_err(leaf, 0, "invalid flag for leaf, WRITTEN not set");
> +		return BTRFS_TREE_BLOCK_WRITTEN_NOT_SET;
> +	}
> +
>   	/*
>   	 * Extent buffers from a relocation tree have a owner field that
>   	 * corresponds to the subvolume tree they are based on. So just from an
> @@ -1858,6 +1863,7 @@ enum btrfs_tree_block_status __btrfs_check_leaf(struct extent_buffer *leaf)
>   	for (slot = 0; slot < nritems; slot++) {
>   		u32 item_end_expected;
>   		u64 item_data_end;
> +		enum btrfs_tree_block_status ret;
>
>   		btrfs_item_key_to_cpu(leaf, &key, slot);
>
> @@ -1913,21 +1919,10 @@ enum btrfs_tree_block_status __btrfs_check_leaf(struct extent_buffer *leaf)
>   			return BTRFS_TREE_BLOCK_INVALID_OFFSETS;
>   		}
>
> -		/*
> -		 * We only want to do this if WRITTEN is set, otherwise the leaf
> -		 * may be in some intermediate state and won't appear valid.
> -		 */
> -		if (btrfs_header_flag(leaf, BTRFS_HEADER_FLAG_WRITTEN)) {
> -			enum btrfs_tree_block_status ret;
> -
> -			/*
> -			 * Check if the item size and content meet other
> -			 * criteria
> -			 */
> -			ret = check_leaf_item(leaf, &key, slot, &prev_key);
> -			if (unlikely(ret != BTRFS_TREE_BLOCK_CLEAN))
> -				return ret;
> -		}
> +		/* Check if the item size and content meet other criteria */
> +		ret = check_leaf_item(leaf, &key, slot, &prev_key);
> +		if (unlikely(ret != BTRFS_TREE_BLOCK_CLEAN))
> +			return ret;
>
>   		prev_key.objectid = key.objectid;
>   		prev_key.type = key.type;
> @@ -1957,6 +1952,11 @@ enum btrfs_tree_block_status __btrfs_check_node(struct extent_buffer *node)
>   	int level = btrfs_header_level(node);
>   	u64 bytenr;
>
> +	if (unlikely(!btrfs_header_flag(node, BTRFS_HEADER_FLAG_WRITTEN))) {
> +		generic_err(node, 0, "invalid flag for node, WRITTEN not set");
> +		return BTRFS_TREE_BLOCK_WRITTEN_NOT_SET;
> +	}
> +
>   	if (unlikely(level <= 0 || level >= BTRFS_MAX_LEVEL)) {
>   		generic_err(node, 0,
>   			"invalid level for node, have %d expect [1, %d]",
> diff --git a/fs/btrfs/tree-checker.h b/fs/btrfs/tree-checker.h
> index 5c809b50b2d0..01669cfa6578 100644
> --- a/fs/btrfs/tree-checker.h
> +++ b/fs/btrfs/tree-checker.h
> @@ -53,6 +53,7 @@ enum btrfs_tree_block_status {
>   	BTRFS_TREE_BLOCK_INVALID_BLOCKPTR,
>   	BTRFS_TREE_BLOCK_INVALID_ITEM,
>   	BTRFS_TREE_BLOCK_INVALID_OWNER,
> +	BTRFS_TREE_BLOCK_WRITTEN_NOT_SET,
>   };
>
>   /*

  parent reply	other threads:[~2024-05-01 21:45 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-04-29 13:03 [PATCH] btrfs: make sure that WRITTEN is set on all metadata blocks Josef Bacik
2024-04-30 20:37 ` David Sterba
2024-05-01 21:45 ` Qu Wenruo [this message]
2024-05-02 12:27   ` David Sterba
2024-05-02 21:32     ` Qu Wenruo

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=2aec5fb2-f881-416a-b558-cb265886dad7@gmx.com \
    --to=quwenruo.btrfs@gmx.com \
    --cc=josef@toxicpanda.com \
    --cc=kernel-team@fb.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=llfamsec@gmail.com \
    /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