Re: [PATCH v2 00/24] erofs: promote erofs from staging

From: Gao Xiang
Date: Mon Jul 15 2019 - 04:37:21 EST

On 2019/7/15 15:56, Pavel Machek wrote:
> Hi!
>>>> Changelog from v1:
>>>> o resend the whole filesystem into a patchset suggested by Greg;
>>>> o code is more cleaner, especially for decompression frontend.
>>>> --8<----------
>>>> Hi,
>>>> EROFS file system has been in Linux-staging for about a year.
>>>> It has been proved to be stable enough to move out of staging
>>>> by 10+ millions of HUAWEI Android mobile phones on the market
>>>> from EMUI 9.0.1, and it was promoted as one of the key features
>>>> of EMUI 9.1 [1], including P30(pro).
>>> Ok, maybe it is ready to be moved to kernel proper, but as git can
>>> do moves, would it be better to do it as one commit?
>>> Separate patches are still better for review, I guess.
>> Thanks for you reply. Either form is OK for me... The first step could
>> be that I hope someone could kindly take some time to look into these
>> patches... :)
>> The patch v2 is slightly different for the current code in the staging
>> tree since I did some code cleanup these days (mainly renaming / moving,
>> including rename unzip_vle.{c,h} to zdata.{c,h} and some confusing
>> structure names and clean up internal.h...). No functional chance and I
>> can submit cleanup patches to staging as well if doing moves by git...
> I believe you should get those committed to staging/, yes. Then you
> ask Al Viro to do pull the git move, I guess, and you follow his
> instructions at that point...
> FILESYSTEMS (VFS and infrastructure)
> M: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx>
> L: linux-fsdevel@xxxxxxxxxxxxxxx

OK, I will send the incremental patches as well later if the above approach
can be done in practice...

Actually I'd like to get fs people Acked-by about EROFS stuffes, e.g. Al, Ted, etc...

It seems rare filesystems upstreamed these years, but I think EROFS is more useful
after moving out of staging. If some people really care about compression ratio,
I can add multi-block fixed-output compression support later (Not very hard, it's
already on my TODO list), although my current company HUAWEI doesn't have any
interest in that way in the near future...

In the long term, I'd like to spend my personal free time to decouple code like
fscrypt and introduce fscompr for other generic fs to compress unmodified files
as well then...

That is another stuff. Anyway, EROFS is one of optimal read-only performance
solutions for consumer electronics compared with others (Note that block storage
has been improved a lot in the past decade...)

Thank you very much,
Gao Xiang

> Best regards,
> Pavel