Re: [PATCH -next v5 06/32] ext4: recheck extent status tree before block allocation

[Zhang Yi <yizhang089@gmail.com>]

On 8/14/2026 6:15 PM, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] Reading uninitialized `map->m_seq` on error paths leads to an infinite retry loop in `ext4_map_blocks()`, causing a kernel hard lockup.
> --
> 
> --- Patch [6]: [PATCH -next v5 06/32] ext4: recheck extent status tree before block allocation ---
> commit 6ae9b515f2027351280fce82734e7e8361a281e7
> Author: Zhang Yi <yi.zhang@huawei.com>
> 
> ext4: recheck extent status tree before block allocation
> 
> After acquiring i_data_sem in write mode, recheck that the mapping
> found via the extent status tree or disk query has not changed. If
> we find a valid extent and the sequence number has changed, retry
> the entire lookup to obtain the correct trimmed mapping.
> 
>> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
>> index 9dbece14ae569..548a3968c5a7d 100644
>> --- a/fs/ext4/inode.c
>> +++ b/fs/ext4/inode.c
> 
> [ ... ]
> 
>> @@ -820,6 +821,19 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
>>   	 * with create == 1 flag.
>>   	 */
>>   	down_write(&EXT4_I(inode)->i_data_sem);
>> +
>> +	/*
>> +	 * Check the validity of the mapping found via the extent status
>> +	 * tree or the disk query. A racing truncate may have changed the
>> +	 * extent, since writeback does not hold i_rwsem or the folio locks
>> +	 * covering the full extent.
>> +	 */
>> +	if (map->m_seq != READ_ONCE(EXT4_I(inode)->i_es_seq)) {
>> +		up_write(&EXT4_I(inode)->i_data_sem);
>> +		map->m_flags = 0;
>> +		map->m_len = orig_mlen;
>> +		goto create_retry;
>> +	}
> 
> [Severity: High]
> Could this cause an infinite loop and hard lockup on disk error?
> 
> If ext4_map_query_blocks() encounters an error like -EIO or -EFSCORRUPTED,
> it returns < 0 without setting map->m_seq.
> 
> However, if EXT4_GET_BLOCKS_CREATE is set, the function does not return
> early on error and falls through to this TOCTOU check.
> 
> Since callers like _ext4_get_block() allocate struct ext4_map_blocks
> on the stack:
> 
> fs/ext4/inode.c:_ext4_get_block() {
>          struct ext4_map_blocks map;
>          ...
>          map.m_lblk = iblock;
>          map.m_len = bh->b_size >> inode->i_blkbits;
> 
>          ret = ext4_map_blocks(..., &map, flags);
> }
> 
> Will map->m_seq contain uninitialized stack garbage, causing the sequence
> check to mismatch and jump to create_retry, which then encounters the exact
> same disk error indefinitely?
> 

This is a valid report. I think ext4_map_query_blocks() should bail
out immediately on error, rather than trying to allocate blocks
afterwards.

Thanks,
Yi.