Re: [PATCH V2] ext3: Always set dx_node's fake_dirent explicitly.

[Eric Sandeen <sandeen@redhat.com>]

On 3/7/11 10:26 AM, Jan Kara wrote:
> On Fri 04-03-11 16:04:08, Eric Sandeen wrote:
>> (crossport of 1f7bebb9e911d870fa8f997ddff838e82b5715ea
>> by Andreas Schlick <schlick@lavabit.com>)
>>
>> When ext3_dx_add_entry() has to split an index node, it has to ensure that
>> name_len of dx_node's fake_dirent is also zero, because otherwise e2fsck
>> won't recognise it as an intermediate htree node and consider the htree to
>> be corrupted.
>>     
>> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
>   Thanks. Merged - and added CC to stable@kernel.org as Andreas suggested.

Thanks!

-Eric

>> While it's a simple one liner, this took me forever to (re-)find.
>   I feel with you :).
> 
> 								Honza
>>
>> diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
>> index b27ba71..75c968e 100644
>> --- a/fs/ext3/namei.c
>> +++ b/fs/ext3/namei.c
>> @@ -1540,8 +1540,8 @@ static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry,
>>  			goto cleanup;
>>  		node2 = (struct dx_node *)(bh2->b_data);
>>  		entries2 = node2->entries;
>> +		memset(&node2->fake, 0, sizeof(struct fake_dirent));
>>  		node2->fake.rec_len = ext3_rec_len_to_disk(sb->s_blocksize);
>> -		node2->fake.inode = 0;
>>  		BUFFER_TRACE(frame->bh, "get_write_access");
>>  		err = ext3_journal_get_write_access(handle, frame->bh);
>>  		if (err)