Mingming Cao wrote:
> On Thu, 2008-03-27 at 19:15 +0900, Akira Fujita wrote:
>> Aneesh Kumar K.V wrote:
>>> On Mon, Mar 24, 2008 at 09:31:03PM +0900, Akira Fujita wrote:
>>>
>>>> Hi Aneesh,
>>>>
>>>>> On Mon, Mar 24, 2008 at 08:00:54PM +0900, Akira Fujita wrote:
>>>>>
>>>>>>>
>>>>>>>
>>>>>> Is EXT4_IOC_FIEMAP going to be added to the ext4 patch queue?
>>>>>> I will try to use EXT4_IOC_FIEMAP instead of EXT4_IOC_EXTENTS_INFO
>>>>>> in ext4 online defrag when kernel supports it.
>>>>>>
>>>>> Can you also look at the sparse warnings with the patches ?
>>>>>
>>>>>
>>>> I have already released the fixed patch before, but it is not in the ext4 patch queue yet.
>>>> Mingming, could you replace new ext4-online-defrag-free-space-fragmentation.patch
>>>> with old one in the ext4 patch queue?
>>>>
>>>>
>>> CHECK fs/ext4/defrag.c
>>> fs/ext4/defrag.c:181:7: warning: cast to restricted type
>>> fs/ext4/defrag.c:181:7: warning: cast from restricted type
>>> fs/ext4/defrag.c:613:21: warning: incorrect type in argument 1
>>> (different address spaces)
>>> fs/ext4/defrag.c:613:21: expected void [noderef] <asn:1>*to
>>> fs/ext4/defrag.c:613:21: got struct ext4_group_data_info *<noident>
>>> fs/ext4/defrag.c:629:24: warning: incorrect type in argument 1
>>> (different address spaces)
>>> fs/ext4/defrag.c:629:24: expected void [noderef] <asn:1>*to
>>> fs/ext4/defrag.c:629:24: got struct ext4_extents_info *<noident>
>>> fs/ext4/defrag.c:849:29: warning: potentially expensive pointer
>>> subtraction
>>> fs/ext4/defrag.c:912:22: warning: potentially expensive pointer
>>> subtraction
>>>
>>> you would need http://www.kernel.org/pub/software/devel/sparse/
>>> to find and fix these warnings.
>>>
>>>
>> Oops, there are a lot of warnings. =-O
>> Thank you for letting me know.
>> I will release fixed defrag patches in a few days.
>
> As part of transition to move ext4 related header files out of
> include/linux, and move to fs/ext4, I also updated the defrag patches to
> adapt this change. You might want to checkout the defrag patches in
> patch queue and merge the updates before release the new version.
>
I see.
I'll use the latest patches in the ext4 patch queue to fix warnings.
BTW, I have a slight question about this warning fix.
The compiler with sparse outputs the warning for the following line.
slots_range = o_end - o_start + 1;
<warning>
fs/ext4/defrag.c:912:22: warning: potentially expensive pointer subtraction
Then I tried to fix this line as follows.
slots_range = ((unsigned long)o_end - (unsigned long)o_start + 1)
/ sizeof(struct ext4_extent);
There is no warning but it seems hard to read a bit.
I have no idea whether I should fix this line or not in this case. :-\
Regards,
Akira