Re: [PATCH V4] fs: New zonefs file system

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2019/09/04 6:56, Darrick J. Wong wrote:
> On Tue, Sep 03, 2019 at 03:49:11AM +0000, Damien Le Moal wrote:
>> On 2019/09/03 12:26, Darrick J. Wong wrote:
>>> On Tue, Sep 03, 2019 at 12:59:17AM +0000, Damien Le Moal wrote:
>>>> Hi Darrick,
>>>>
>>>> Any comments on this new version ?
>>>
>>> I took a brief glance a few days ago and it looked ok wrt the iomap
>>> parts.  I'm assuming you received the same complaint from the kbuild
>>> robot as I did?
>>
>> Yes, I did receive the same complaints.
>>
>>>
>>>> Should I wait for the iomap code to make it to 5.4 first before trying to get
>>>> this new FS included ?
>>>
>>> Given that the merge window apparently won't close until Sept. 29, that
>>> gives us more time to make any more minor tweaks.
>>
>> I keep monitoring your iomap-for-next branch for any change, rebasing and
>> testing if anything changes there. So far, no problems.
>>
>>> (That means 80% of a "Go for it" but I'll look more closely tomorrow ;))
>>
>> OK. Thank you. One question I was wondering about: All my code until now has
>> gone through a maintainer, even for the little parts I maintain myself
>> (dm-zoned). However, it seems customary for file systems to each have their own
>> maintainer and sending PRs to Linus. So should I prepare myself for having a
>> tree specifically for zonefs and getting a signed GPG key for sending pull
>> requests ?
> 
> That is a good idea, particularly if you're at LPC and can leverage that
> to get key signatures.  It's awkward to try to push changes to zonefs
> that aren't related to iomap through the iomap tree, and you're in a
> much better position to integrate & test zonefs changes given that
> xfstests doesn't translate too well to it.

OK. Understood.

>> Or given how small zonefs is, would you be willing to take zonefs
>> through the iomap tree ? The last option would be easier for me, but I do not
>> want to put on you any overhead :)
> 
> Most file systems are such big piles of ... code that it makes more
> sense to have the maintainer(s) be the specialists in that filesystem.
> AFAICT most dm targets are much simpler by comparison.

Yes indeed. I will prepare everything needed to send PRs myself then.

Thanks !

-- 
Damien Le Moal
Western Digital Research




[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux