On 2017/1/23 上午2:01, Avi Kivity wrote:
> Hello,
> 
> 
> On 11/27/2016 05:24 PM, Avi Kivity wrote:
>> mkfs /dev/md0 can take a very long time, if /dev/md0 is a very large
>> disk that supports TRIM/DISCARD (erase whichever is inappropriate). 
>> That is because mkfs issues a TRIM/DISCARD (erase whichever is
>> inappropriate) for the entire partition. As far as I can tell, md
>> converts the large TRIM/DISCARD (erase whichever is inappropriate)
>> into a large number of TRIM/DISCARD (erase whichever is inappropriate)
>> requests, one per chunk-size worth of disk, and issues them to the
>> RAID components individually.
>>
>>
>> It seems to me that md can convert the large TRIM/DISCARD (erase
>> whichever is inappropriate) request it gets into one TRIM/DISCARD
>> (erase whichever is inappropriate) per RAID component, converting an
>> O(disk size / chunk size) operation into an O(number of RAID
>> components) operation, which is much faster.
>>
>>
>> I observed this with mkfs.xfs on a RAID0 of four 3TB NVMe devices,
>> with the operation taking about a quarter of an hour, continuously
>> pushing half-megabyte TRIM/DISCARD (erase whichever is inappropriate)
>> requests to the disk. Linux 4.1.12.
>>
> 
> Did anyone pick this up by any chance?  The only thing I could find is
> more people complaining about the same issue.

Hi Avi,

I proposed a POC patch, Shaohua and Neil provide review comments,
suggest me to simplify this patch.

If you notice, there is a patch I sent out on Dec 9, 2016, in this email
thread. This patch works, but not the final version to be accepted by
upstream. I quote the performance number here,

“ On 4x3TB NVMe raid0, format it with mkfs.xfs. Current upstream kernel
spends 306 seconds, the patched kernel spends 15 seconds. I see average
request size increases from 1 chunk (1024 sectors) to 2048 chunks
(2097152 sectors).”

I call this patch as RFC v2 patch, and attach it again in this email.
Now I am working on a new version by the suggestion from Shaohua and
Neil, but any testing feed back of RFC v2 patch is welcome. It works,
just too complex to review.

Thanks.

Coly