does 548b8b5168c9 qualify for stable?

does 548b8b5168c9 qualify for stable?

[Rasmus Villemoes]

Hi,

Please consider whether

commit 548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8
Author: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Date:   Thu Sep 17 08:56:11 2020 +0200

    scripts/setlocalversion: make git describe output more reliable

qualifies for -stable. It removes one potential source of binary
non-reproducibility that we have actually seen cause problems.

I'm fine with it not qualifying, but please let me know if so, because
then I'll go and add some workarounds to various customer projects.

In case it doesn't cherry-pick cleanly (I think there might have been
some shell-portability patches replacing $() by `` or something like
that) I am happy to provide backports to the still maintained -stable
branches.

Thanks,
Rasmus

Re: does 548b8b5168c9 qualify for stable?

[Greg KH]

On Fri, Oct 23, 2020 at 03:40:26PM +0200, Rasmus Villemoes wrote:
> Hi,
> 
> Please consider whether
> 
> commit 548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8
> Author: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> Date:   Thu Sep 17 08:56:11 2020 +0200
> 
>     scripts/setlocalversion: make git describe output more reliable
> 
> qualifies for -stable. It removes one potential source of binary
> non-reproducibility that we have actually seen cause problems.
> 
> I'm fine with it not qualifying, but please let me know if so, because
> then I'll go and add some workarounds to various customer projects.
> 
> In case it doesn't cherry-pick cleanly (I think there might have been
> some shell-portability patches replacing $() by `` or something like
> that) I am happy to provide backports to the still maintained -stable
> branches.

Looks like it qualifies, how far back do you want it to go?

And yes, backported patches always make it much easier to apply :)

thanks,

greg k-h

Re: does 548b8b5168c9 qualify for stable?

[Rasmus Villemoes]

On 23/10/2020 16.22, Greg KH wrote:
> On Fri, Oct 23, 2020 at 03:40:26PM +0200, Rasmus Villemoes wrote:
>> Hi,
>>
>> Please consider whether
>>
>> commit 548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8
>> Author: Rasmus Villemoes <linux@rasmusvillemoes.dk>
>> Date:   Thu Sep 17 08:56:11 2020 +0200
>>
>>     scripts/setlocalversion: make git describe output more reliable
>>
>> qualifies for -stable. 
> 
> Looks like it qualifies, how far back do you want it to go?
> 

Cool, thanks. I think we have a project using 4.9.y, certainly we have
projects based on 4.19 and 5.4 - so might as well make it all of the
ones listed on kernel.org currently.

> And yes, backported patches always make it much easier to apply :)

OK. How do you prefer to get those? Individual patch emails with [PATCH
X.Y-stable] in subject? Or should I put them in a git repo you can
cherry-pick them from? Should I include the "Commit
548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8 upstream" line? How about notes
on how it differs from the upstream commit (e.g. when just the context
uses `` instead of $() or similar)?

Rasmus

Re: does 548b8b5168c9 qualify for stable?

[Greg KH]

On Fri, Oct 23, 2020 at 04:37:25PM +0200, Rasmus Villemoes wrote:
> On 23/10/2020 16.22, Greg KH wrote:
> > On Fri, Oct 23, 2020 at 03:40:26PM +0200, Rasmus Villemoes wrote:
> >> Hi,
> >>
> >> Please consider whether
> >>
> >> commit 548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8
> >> Author: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> >> Date:   Thu Sep 17 08:56:11 2020 +0200
> >>
> >>     scripts/setlocalversion: make git describe output more reliable
> >>
> >> qualifies for -stable. 
> > 
> > Looks like it qualifies, how far back do you want it to go?
> > 
> 
> Cool, thanks. I think we have a project using 4.9.y, certainly we have
> projects based on 4.19 and 5.4 - so might as well make it all of the
> ones listed on kernel.org currently.
> 
> > And yes, backported patches always make it much easier to apply :)
> 
> OK. How do you prefer to get those? Individual patch emails with [PATCH
> X.Y-stable] in subject?

That works.

> Or should I put them in a git repo you can cherry-pick them from?

git repos don't work, email does :)

> Should I include the "Commit 548b8b5168c90c42e88f70fcf041b4ce0b8e7aa8
> upstream" line?

Yes please.

> How about notes on how it differs from the upstream commit (e.g. when
> just the context uses `` instead of $() or similar)?

That is also nice to have, if possible, whatever you feel like doing
here.

thanks,

greg k-h