[PATCH 1/2] erofs: avoid infinite loops due to corrupted subpage compact indexes

Hongbo Li lihongbo22 at huawei.com
Wed Oct 22 00:11:50 AEDT 2025



On 2025/10/17 15:05, Gao Xiang wrote:
> Robert reported an infinite loop observed by two crafted images.
> 
> The root cause is that `clusterofs` can be larger than `lclustersize`
> for !NONHEAD `lclusters` in corrupted subpage compact indexes, e.g.:
> 
>    blocksize = lclustersize = 512   lcn = 6   clusterofs = 515
> 
> Move the corresponding check for full compress indexes to
> `z_erofs_load_lcluster_from_disk()` to also cover subpage compact
> compress indexes.
> 
> It also fixes the position of `m->type >= Z_EROFS_LCLUSTER_TYPE_MAX`
> check, since it should be placed right after
> `z_erofs_load_{compact,full}_lcluster()`.
> 
> Fixes: 8d2517aaeea3 ("erofs: fix up compacted indexes for block size < 4096")
> Fixes: 1a5223c182fd ("erofs: do sanity check on m->type in z_erofs_load_compact_lcluster()")
> Reported-by: Robert Morris <rtm at csail.mit.edu>
> Closes: https://lore.kernel.org/r/35167.1760645886@localhost
> Signed-off-by: Gao Xiang <hsiangkao at linux.alibaba.com>

The clusterofs is unused in NONHEAD lcluster, so it is easy to understand.

Reviewed-by: Hongbo Li <lihongbo22 at huawei.com>

Thanks,
Hongbo

> ---
>   fs/erofs/zmap.c | 32 ++++++++++++++++++--------------
>   1 file changed, 18 insertions(+), 14 deletions(-)
> 
> diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
> index e5581dbeb4c2..6aca228cd2a5 100644
> --- a/fs/erofs/zmap.c
> +++ b/fs/erofs/zmap.c
> @@ -55,10 +55,6 @@ static int z_erofs_load_full_lcluster(struct z_erofs_maprecorder *m,
>   	} else {
>   		m->partialref = !!(advise & Z_EROFS_LI_PARTIAL_REF);
>   		m->clusterofs = le16_to_cpu(di->di_clusterofs);
> -		if (m->clusterofs >= 1 << vi->z_lclusterbits) {
> -			DBG_BUGON(1);
> -			return -EFSCORRUPTED;
> -		}
>   		m->pblk = le32_to_cpu(di->di_u.blkaddr);
>   	}
>   	return 0;
> @@ -240,21 +236,29 @@ static int z_erofs_load_compact_lcluster(struct z_erofs_maprecorder *m,
>   static int z_erofs_load_lcluster_from_disk(struct z_erofs_maprecorder *m,
>   					   unsigned int lcn, bool lookahead)
>   {
> +	struct erofs_inode *vi = EROFS_I(m->inode);
> +	int err;
> +
> +	if (vi->datalayout == EROFS_INODE_COMPRESSED_COMPACT) {
> +		err = z_erofs_load_compact_lcluster(m, lcn, lookahead);
> +	} else {
> +		DBG_BUGON(vi->datalayout != EROFS_INODE_COMPRESSED_FULL);
> +		err = z_erofs_load_full_lcluster(m, lcn);
> +	}
> +	if (err)
> +		return err;
> +
>   	if (m->type >= Z_EROFS_LCLUSTER_TYPE_MAX) {
>   		erofs_err(m->inode->i_sb, "unknown type %u @ lcn %u of nid %llu",
> -				m->type, lcn, EROFS_I(m->inode)->nid);
> +			  m->type, lcn, EROFS_I(m->inode)->nid);
>   		DBG_BUGON(1);
>   		return -EOPNOTSUPP;
> +	} else if (m->type != Z_EROFS_LCLUSTER_TYPE_NONHEAD &&
> +		   m->clusterofs >= (1 << vi->z_lclusterbits)) {
> +		DBG_BUGON(1);
> +		return -EFSCORRUPTED;
>   	}
> -
> -	switch (EROFS_I(m->inode)->datalayout) {
> -	case EROFS_INODE_COMPRESSED_FULL:
> -		return z_erofs_load_full_lcluster(m, lcn);
> -	case EROFS_INODE_COMPRESSED_COMPACT:
> -		return z_erofs_load_compact_lcluster(m, lcn, lookahead);
> -	default:
> -		return -EINVAL;
> -	}
> +	return 0;
>   }
>   
>   static int z_erofs_extent_lookback(struct z_erofs_maprecorder *m,


More information about the Linux-erofs mailing list