Hi Jaegeuk,

On 2022-04-13 18:22, Jaegeuk Kim wrote:
>>> The section size actually supports multiple 2MBs, not PO2.
>>>
>> Thanks a lot for the clarification. I will remove this statement in the
>> next revision.
>>
>> I was partially misled by [1] where it is stated "Segments are collected
>> into sections. There is genuine flexibility in the size of a section,
>> though it must be a power of two.".
>>
>> Just FYI, when I did a quick check, there are some assumptions in the
>> zoned support for f2fs which assumes the zoned device size is a power of
>> 2 such as in the __f2fs_issue_discard_zone. So if I am not wrong, when
>> we remove those assumptions in f2fs for zone size, then everything
>> should work fine provided the zone size is a multiple of 2MB. Am I
>> missing something here?
> 
> All the implementaion assumes PO2 by block layer in kernel, but basically
Yeah, at the moment it is not an issue as kernel rejects non power of 2
zoned devices as a block device. But we have been having some
conversation around this topic [1] to remove this constraint from the
block layer.

> f2fs could support 2MBs. So, I remember there's no PO2 check in f2fs as such.
> 
>>
>> I am new to f2fs but is there testsuite that I can run for f2fs apart
>> from the two tests listed in (x)fstests?
> 
> I usually run 1) full xfstests, 2) loop of fsstress + shutdown. You can find
> a script here. :)
> 
> https://github.com/jaegeuk/xfstests-f2fs/blob/f2fs/run.sh
>
Awesome, thanks a lot :)

[1] https://lore.kernel.org/all/20220315135245.eqf4tqngxxb7ymqa@unifi/


_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Reply via email to