[Ocfs2-tools-devel] [PATCH 07/10] fsck.ocfs2 now know inline data.

[Tao Ma]

Hi joel,
Joel Becker wrote:
>> +		if (di->i_clusters > 0 &&
>> +		    prompt(ost, PY, PR_INODE_CLUSTERS,
>> +			   "Inode %"PRIu64" has %"PRIu32" clusters but it has "
>> +			   "inline data flag set. "
>> +			   "Correct the number of clusters?",
>> +			   (uint64_t)di->i_blkno, di->i_clusters)) {
>> +			di->i_clusters = 0;
>> +			o2fsck_write_inode(ost, blkno, di);
>> +		}
> 
> 	The way you have this working, you trust INLINE_DATA_FL and
> truncate clusters.  That is, if you have INLINE_DATA_FL, but i_size is
> greater than expected *and* i_clusters != 0, you reduce i_size and clear
> i_clusters.  But what if INLINE_DATA_FL is the corruption?  i_size and
> i_clusters are correct, and there is really an extent list on there.
> The result is an inode with garbage in id_data, and the extents are
> returned to the global bitmap.
> 	I'm not sure what the right answer is here.  Perhaps if *both*
> i_size>expected and i_clusters!=0 you then see if i_list looks sane?  If
> it's sane, you ask about clearing INLINE_DATA_FL?
> 	Mark, does this make any sense?
You suggestion make me to think more of it. So how should we trust this 
flag? As you have said, yes, there is a possibility that the 
INLINE_DATA_FL is happened to be the corruption, then there is also one 
possibility that ~INLINE_DATA_FL is also the corruption, what should we 
handle this also?

So now my thought is that take INLINE_DATA_FL, i_size, i_clusters as one 
compound judgement. So if the 2 things happens to be right, we will 
regard them as right.
1) if INLINE_DATA_FL is set:
    a) if either i_size <= max_inline_data or i_clusters=0, we trust 
this flag and clear the error i_size or i_clusters.
    b) if both i_size > max_inline_data and i_clusters>0, clear the flag 
and go on with check_el. The only problem is how should we handle the 
error in check_el(any error in check_el will not make us reset 
INLINE_DATA_FL or if there is any error, we will trust INLINE_DATA_FL 
and set it back?).

2) if INLINE_DATA_FL isn't set:
    a) if both i_size <= max_inline_data and i_clusters=0, we set this 
flag.
    b) if either i_size > max_inline_data or i_clusters>0, we trust the 
file isn't inline. so go with check_el.

Does it make sense? Mark, do you have any suggestion about this?

Regards,
Tao