2017-07-17 12:50:39

by Tahsin Erdogan

[permalink] [raw]
Subject: [PATCH] e2fsck: eliminate bad update link count attempt

Extended attribute inodes have a link count of 1 but they are not
attached to any directories. When an xattr inode with zero ea
references is found, the remedy is to reconnect it to lost+found dir.
Since reconnect operation increments the link count, it would normally
become 2 but to avoid that, check_ea_inode() sets the link count to
zero in anticipation of reconnect operation. And it does it even when
e2fsck is invoked with -n option which causes a fatal e2fsck failure
as can be demonstrated with the following test script:

mke2fs -b 4096 -I 256 -F -O extent,ea_inode /dev/sdb 1G
mount /dev/sdb /mnt/sdb
touch /mnt/sdb/x
setfattr -n user.ie1 -v $(python -c "print 'v'*5000") /mnt/sdb/x
umount /dev/sdb
debugfs -w -R 'rm x' /dev/sdb
e2fsck -f -n /dev/sdb

In this case, e2fsck fails with the following error:

e2fsck 1.43.5-WIP (17-Feb-2017)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
ext2fs_write_inode: Attempt to write to filesystem opened read-only
while writing inode 13 in check_ea_inode
e2fsck: aborted

Setting link count to zero is not really required because e2fsck will
detect the discrepancy in link count and offer to fix it.

Signed-off-by: Tahsin Erdogan <[email protected]>
---
e2fsck/pass4.c | 11 +----------
1 file changed, 1 insertion(+), 10 deletions(-)

diff --git a/e2fsck/pass4.c b/e2fsck/pass4.c
index 663f87ab59c0..b23347cdb6b0 100644
--- a/e2fsck/pass4.c
+++ b/e2fsck/pass4.c
@@ -122,17 +122,8 @@ static void check_ea_inode(e2fsck_t ctx, ext2_ino_t i,

if (ctx->ea_inode_refs)
ea_refcount_fetch(ctx->ea_inode_refs, i, &actual_refs);
- if (!actual_refs) {
- /*
- * There are no attribute references to the ea_inode.
- * Zero the link count so that when inode is linked to
- * lost+found it has correct link count.
- */
- inode->i_links_count = 0;
- e2fsck_write_inode(ctx, i, EXT2_INODE(inode), "check_ea_inode");
- ext2fs_icount_store(ctx->inode_link_info, i, 0);
+ if (!actual_refs)
return;
- }

/*
* There are some attribute references, link_counted is now considered
--
2.13.2.932.g7449e964c-goog


2017-07-24 04:11:54

by Theodore Ts'o

[permalink] [raw]
Subject: Re: [PATCH] e2fsck: eliminate bad update link count attempt

On Mon, Jul 17, 2017 at 05:50:35AM -0700, Tahsin Erdogan wrote:
> Extended attribute inodes have a link count of 1 but they are not
> attached to any directories. When an xattr inode with zero ea
> references is found, the remedy is to reconnect it to lost+found dir.
> Since reconnect operation increments the link count, it would normally
> become 2 but to avoid that, check_ea_inode() sets the link count to
> zero in anticipation of reconnect operation. And it does it even when
> e2fsck is invoked with -n option which causes a fatal e2fsck failure
> as can be demonstrated with the following test script:
>
> mke2fs -b 4096 -I 256 -F -O extent,ea_inode /dev/sdb 1G
> mount /dev/sdb /mnt/sdb
> touch /mnt/sdb/x
> setfattr -n user.ie1 -v $(python -c "print 'v'*5000") /mnt/sdb/x
> umount /dev/sdb
> debugfs -w -R 'rm x' /dev/sdb
> e2fsck -f -n /dev/sdb
>
> In this case, e2fsck fails with the following error:
>
> e2fsck 1.43.5-WIP (17-Feb-2017)
> Pass 1: Checking inodes, blocks, and sizes
> Pass 2: Checking directory structure
> Pass 3: Checking directory connectivity
> Pass 4: Checking reference counts
> ext2fs_write_inode: Attempt to write to filesystem opened read-only
> while writing inode 13 in check_ea_inode
> e2fsck: aborted
>
> Setting link count to zero is not really required because e2fsck will
> detect the discrepancy in link count and offer to fix it.
>
> Signed-off-by: Tahsin Erdogan <[email protected]>

Thanks, applied.

- Ted