Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1632400 > unrolled thread

Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block

Started byRichard Weinberger <richard@nod.at>
First post2017-04-28 00:10 +0200
Last post2017-05-09 09:10 +0200
Articles 12 — 3 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Richard Weinberger <richard@nod.at> - 2017-04-28 00:10 +0200
    Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Amir Goldstein <amir73il@gmail.com> - 2017-04-28 11:00 +0200
      Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Oleksij Rempel <ore@pengutronix.de> - 2017-05-02 07:40 +0200
    Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Amir Goldstein <amir73il@gmail.com> - 2017-05-02 09:30 +0200
      Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Richard Weinberger <richard@nod.at> - 2017-05-02 09:40 +0200
        Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Oleksij Rempel <ore@pengutronix.de> - 2017-05-09 06:20 +0200
          Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Oleksij Rempel <ore@pengutronix.de> - 2017-05-09 08:00 +0200
            Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Amir Goldstein <amir73il@gmail.com> - 2017-05-09 09:10 +0200
              Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Oleksij Rempel <ore@pengutronix.de> - 2017-05-09 09:40 +0200
              Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Richard Weinberger <richard@nod.at> - 2017-05-09 09:40 +0200
                Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Amir Goldstein <amir73il@gmail.com> - 2017-05-09 10:00 +0200
            Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block Richard Weinberger <richard@nod.at> - 2017-05-09 09:10 +0200

#1632400 — Re: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block

FromRichard Weinberger <richard@nod.at>
Date2017-04-28 00:10 +0200
SubjectRe: [PATCH v2 3/3] fs: ubifs: set s_uuid in super block
Message-ID<tAZLP-6S4-3@gated-at.bofh.it>
Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>> So, if some flag should be implemented, who should do it? :)
> 
> I'll not do it for you. ;)

Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2

Thanks,
//richard

[toc] | [next] | [standalone]

#1632700

FromAmir Goldstein <amir73il@gmail.com>
Date2017-04-28 11:00 +0200
Message-ID<tB9US-53z-37@gated-at.bofh.it>
In reply to#1632400
On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger <richard@nod.at> wrote:
> Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>>> So, if some flag should be implemented, who should do it? :)
>>
>> I'll not do it for you. ;)
>
> Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>

Perhaps you meant this:
https://marc.info/?l=linux-fsdevel&m=149328358909709&w=2

There does not seem to be much objections to adding the flag,
so hopefully, we can merge it for v.12 and filesystems and consumers
will pick it up whenever.

Amir.

[toc] | [prev] | [next] | [standalone]

#1634140

FromOleksij Rempel <ore@pengutronix.de>
Date2017-05-02 07:40 +0200
Message-ID<tCyHv-27S-1@gated-at.bofh.it>
In reply to#1632700
On 04/28/2017 10:53 AM, Amir Goldstein wrote:
> On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger <richard@nod.at> wrote:
>> Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>>>> So, if some flag should be implemented, who should do it? :)
>>>
>>> I'll not do it for you. ;)
>>
>> Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>>
>
> Perhaps you meant this:
> https://marc.info/?l=linux-fsdevel&m=149328358909709&w=2
>
> There does not seem to be much objections to adding the flag,
> so hopefully, we can merge it for v.12 and filesystems and consumers
> will pick it up whenever.

Ok, thanks.

Then i will need to wait untill your patches is merged and then resent 
updated patch to avoid merge race condition.

[toc] | [prev] | [next] | [standalone]

#1634184

FromAmir Goldstein <amir73il@gmail.com>
Date2017-05-02 09:30 +0200
Message-ID<tCApX-3rN-1@gated-at.bofh.it>
In reply to#1632400
On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger <richard@nod.at> wrote:
> Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>>> So, if some flag should be implemented, who should do it? :)
>>
>> I'll not do it for you. ;)
>
> Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>

Richard,

Considering the facts that:
1. I proposed the said flag and Al didn't think it was needed [1]
2. ext4 already sets s_uuid without any flag for a long time now
3. A similar patch was queued for v4.12 to set s_uuid for xfs without any flag

I think it would be right to take Oleksij's patch as is.

FYI, my current work on 'constant inode numbers for overlayfs' requires that
underlying filesystem had set a non-zero s_uuid. Not sure if that matters for
ubifs+overlayfs users.

Amir.

[1] https://marc.info/?l=linux-unionfs&m=149352864527985&w=2

[toc] | [prev] | [next] | [standalone]

#1634192

FromRichard Weinberger <richard@nod.at>
Date2017-05-02 09:40 +0200
Message-ID<tCAzD-3uY-7@gated-at.bofh.it>
In reply to#1634184
Amir,

Am 02.05.2017 um 09:19 schrieb Amir Goldstein:
> On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger <richard@nod.at> wrote:
>> Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>>>> So, if some flag should be implemented, who should do it? :)
>>>
>>> I'll not do it for you. ;)
>>
>> Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>>
> 
> Richard,
> 
> Considering the facts that:
> 1. I proposed the said flag and Al didn't think it was needed [1]
> 2. ext4 already sets s_uuid without any flag for a long time now
> 3. A similar patch was queued for v4.12 to set s_uuid for xfs without any flag
> 
> I think it would be right to take Oleksij's patch as is.
> 
> FYI, my current work on 'constant inode numbers for overlayfs' requires that
> underlying filesystem had set a non-zero s_uuid. Not sure if that matters for
> ubifs+overlayfs users.

If VFS maintainers are fine with that, I'll take it.
From UBIFS' POV it does not matter much. :-)

Thanks
//richard

[toc] | [prev] | [next] | [standalone]

#1637856

FromOleksij Rempel <ore@pengutronix.de>
Date2017-05-09 06:20 +0200
Message-ID<tF4MW-4Yx-3@gated-at.bofh.it>
In reply to#1634192

On 05/02/2017 09:37 AM, Richard Weinberger wrote:
> Amir,
>
> Am 02.05.2017 um 09:19 schrieb Amir Goldstein:
>> On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger <richard@nod.at> wrote:
>>> Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>>>>> So, if some flag should be implemented, who should do it? :)
>>>>
>>>> I'll not do it for you. ;)
>>>
>>> Please also see http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>>>
>>
>> Richard,
>>
>> Considering the facts that:
>> 1. I proposed the said flag and Al didn't think it was needed [1]
>> 2. ext4 already sets s_uuid without any flag for a long time now
>> 3. A similar patch was queued for v4.12 to set s_uuid for xfs without any flag
>>
>> I think it would be right to take Oleksij's patch as is.
>>
>> FYI, my current work on 'constant inode numbers for overlayfs' requires that
>> underlying filesystem had set a non-zero s_uuid. Not sure if that matters for
>> ubifs+overlayfs users.
>
> If VFS maintainers are fine with that, I'll take it.
> From UBIFS' POV it does not matter much. :-)

Ping to VFS maintainers?

[toc] | [prev] | [next] | [standalone]

#1637880

FromOleksij Rempel <ore@pengutronix.de>
Date2017-05-09 08:00 +0200
Message-ID<tF6lH-5Sx-7@gated-at.bofh.it>
In reply to#1637856

On 05/09/2017 07:37 AM, Amir Goldstein wrote:
>
>
> On Tue, May 9, 2017 at 7:13 AM, Oleksij Rempel <ore@pengutronix.de
> <mailto:ore@pengutronix.de>> wrote:
>
>
>
>     On 05/02/2017 09:37 AM, Richard Weinberger wrote:
>
>         Amir,
>
>         Am 02.05.2017 um 09:19 schrieb Amir Goldstein:
>
>             On Fri, Apr 28, 2017 at 1:03 AM, Richard Weinberger
>             <richard@nod.at <mailto:richard@nod.at>> wrote:
>
>                 Am 24.04.2017 um 17:47 schrieb Richard Weinberger:
>
>                         So, if some flag should be implemented, who
>                         should do it? :)
>
>
>                     I'll not do it for you. ;)
>
>
>                 Please also see
>                 http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2
>                 <http://marc.info/?l=linux-fsdevel&m=149327990608749&w=2>
>
>
>             Richard,
>
>             Considering the facts that:
>             1. I proposed the said flag and Al didn't think it was
>             needed [1]
>             2. ext4 already sets s_uuid without any flag for a long time now
>             3. A similar patch was queued for v4.12 to set s_uuid for
>             xfs without any flag
>
>             I think it would be right to take Oleksij's patch as is.
>
>             FYI, my current work on 'constant inode numbers for
>             overlayfs' requires that
>             underlying filesystem had set a non-zero s_uuid. Not sure if
>             that matters for
>             ubifs+overlayfs users.
>
>
>         If VFS maintainers are fine with that, I'll take it.
>         From UBIFS' POV it does not matter much. :-)
>
>
>     Ping to VFS maintainers?
>
>
> What ping? Al made it clear that a flag is not needed.
> BTW, xfs s_uuid patch was merged to master.


I'm talking about ubifs patch.

[toc] | [prev] | [next] | [standalone]

#1637911

FromAmir Goldstein <amir73il@gmail.com>
Date2017-05-09 09:10 +0200
Message-ID<tF7rr-6Pv-3@gated-at.bofh.it>
In reply to#1637880
On Tue, May 9, 2017 at 10:01 AM, Richard Weinberger <richard@nod.at> wrote:
>
> Oleksij,
>
> Am 09.05.2017 um 07:52 schrieb Oleksij Rempel:
> >>
> >>         If VFS maintainers are fine with that, I'll take it.
> >>         From UBIFS' POV it does not matter much. :-)
> >>
> >>
> >>     Ping to VFS maintainers?
> >>
> >>
> >> What ping? Al made it clear that a flag is not needed.
> >> BTW, xfs s_uuid patch was merged to master.
> >
> >
> > I'm talking about ubifs patch.

Me too.

>
> Then we can queue this patch for 4.13.
> Please resend and make sure it addresses everything what was also
> suggested for the xfs s_uuid patch.
>

Just to be clear, the xfs s_uuid patch is just a memcpy,
no different from Oleksij's patch.

Thanks,
Amir.

[toc] | [prev] | [next] | [standalone]

#1637929

FromOleksij Rempel <ore@pengutronix.de>
Date2017-05-09 09:40 +0200
Message-ID<tF7Uu-72H-3@gated-at.bofh.it>
In reply to#1637911

On 05/09/2017 09:08 AM, Amir Goldstein wrote:
> On Tue, May 9, 2017 at 10:01 AM, Richard Weinberger <richard@nod.at> wrote:
>>
>> Oleksij,
>>
>> Am 09.05.2017 um 07:52 schrieb Oleksij Rempel:
>>>>
>>>>         If VFS maintainers are fine with that, I'll take it.
>>>>         From UBIFS' POV it does not matter much. :-)
>>>>
>>>>
>>>>     Ping to VFS maintainers?
>>>>
>>>>
>>>> What ping? Al made it clear that a flag is not needed.
>>>> BTW, xfs s_uuid patch was merged to master.
>>>
>>>
>>> I'm talking about ubifs patch.
>
> Me too.

:) ok

>>
>> Then we can queue this patch for 4.13.
>> Please resend and make sure it addresses everything what was also
>> suggested for the xfs s_uuid patch.
>>
>
> Just to be clear, the xfs s_uuid patch is just a memcpy,
> no different from Oleksij's patch.

So, should i change something?

here is the patch:
https://patchwork.kernel.org/patch/9674817/

[toc] | [prev] | [next] | [standalone]

#1637930

FromRichard Weinberger <richard@nod.at>
Date2017-05-09 09:40 +0200
Message-ID<tF7Uu-72H-5@gated-at.bofh.it>
In reply to#1637911
Amir,

Am 09.05.2017 um 09:08 schrieb Amir Goldstein:
>> Then we can queue this patch for 4.13.
>> Please resend and make sure it addresses everything what was also
>> suggested for the xfs s_uuid patch.
>>
> 
> Just to be clear, the xfs s_uuid patch is just a memcpy,
> no different from Oleksij's patch.

Wasn't there a huge discussion about LE/BE/uniqueness and more details
on UUID that hurt my brain.

Thanks,
//richard

[toc] | [prev] | [next] | [standalone]

#1637957

FromAmir Goldstein <amir73il@gmail.com>
Date2017-05-09 10:00 +0200
Message-ID<tF8dQ-7bU-19@gated-at.bofh.it>
In reply to#1637930
On Tue, May 9, 2017 at 10:35 AM, Richard Weinberger <richard@nod.at> wrote:
> Amir,
>
> Am 09.05.2017 um 09:08 schrieb Amir Goldstein:
>>> Then we can queue this patch for 4.13.
>>> Please resend and make sure it addresses everything what was also
>>> suggested for the xfs s_uuid patch.
>>>
>>
>> Just to be clear, the xfs s_uuid patch is just a memcpy,
>> no different from Oleksij's patch.

See upstream commit
8f720d9 xfs: publish UUID in struct super_block


>
> Wasn't there a huge discussion about LE/BE/uniqueness and more details
> on UUID that hurt my brain.
>

LE/BE discussions are more about which variants of uuid helpers should be
created, among other things, for consumers of s_uuid to check that s_uuid
was filled by fs.
Converting s_uuid type to uuid_t or whatever is for the far future.

uniqueness of s_uuid does not exist with current filesystems,
so no reason whatsoever to act differently with ubifs.
Fixing uniqueness of s_uuid (if at all is needed) is a future VFS task.

Bottom line, for Oleksij's original patch:

Reviewed-by: Amir Goldstein <amir73il@gmail.com>

If it's not too late for 4.12 that could be nice, because then
ubifs+overlayfs would gain a new feature (constant inode numbers)

[toc] | [prev] | [next] | [standalone]

#1637916

FromRichard Weinberger <richard@nod.at>
Date2017-05-09 09:10 +0200
Message-ID<tF7rr-6Pv-5@gated-at.bofh.it>
In reply to#1637880
Oleksij,

Am 09.05.2017 um 07:52 schrieb Oleksij Rempel:
>>
>>         If VFS maintainers are fine with that, I'll take it.
>>         From UBIFS' POV it does not matter much. :-)
>>
>>
>>     Ping to VFS maintainers?
>>
>>
>> What ping? Al made it clear that a flag is not needed.
>> BTW, xfs s_uuid patch was merged to master.
> 
> 
> I'm talking about ubifs patch.

Then we can queue this patch for 4.13.
Please resend and make sure it addresses everything what was also
suggested for the xfs s_uuid patch.

Thanks,
//richard

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web