Re: Patch "kbuild: rust: force `alloc` extern to allow "empty" Rust files" has been added to the 6.1-stable tree

Re: Patch "kbuild: rust: force `alloc` extern to allow "empty" Rust files" has been added to the 6.1-stable tree

[Miguel Ojeda]

On Sun, Jun 16, 2024 at 4:11 AM Sasha Levin <sashal@kernel.org> wrote:
>
>     Cc: stable@vger.kernel.org # v6.6+

This cannot be backported to 6.1 unless we upgrade the compiler, since
the feature is not available in old versions.

For future reference, this patch got picked for 6.1 a couple more times:

    https://lore.kernel.org/stable/CANiq72=V1=D-X5ncqN1pyfE4L1bz5zFRdBot6HpkCYie-EQnPA@mail.gmail.com/
    https://lore.kernel.org/stable/CANiq72ndLzts-KzUv_22vHF0tYkPvROv=oG+KP2KhbCvHkn60g@mail.gmail.com/

If I had known it would cause your scripts to pick it up repeatedly,
then I would have probably avoided `Fixes`/`Cc` -- it is a very minor
issue.

Cheers,
Miguel

Re: Patch "kbuild: rust: force `alloc` extern to allow "empty" Rust files" has been added to the 6.1-stable tree

[Greg KH]

On Sun, Jun 16, 2024 at 12:38:41PM +0200, Miguel Ojeda wrote:
> On Sun, Jun 16, 2024 at 4:11 AM Sasha Levin <sashal@kernel.org> wrote:
> >
> >     Cc: stable@vger.kernel.org # v6.6+
> 
> This cannot be backported to 6.1 unless we upgrade the compiler, since
> the feature is not available in old versions.
> 
> For future reference, this patch got picked for 6.1 a couple more times:
> 
>     https://lore.kernel.org/stable/CANiq72=V1=D-X5ncqN1pyfE4L1bz5zFRdBot6HpkCYie-EQnPA@mail.gmail.com/
>     https://lore.kernel.org/stable/CANiq72ndLzts-KzUv_22vHF0tYkPvROv=oG+KP2KhbCvHkn60g@mail.gmail.com/
> 
> If I had known it would cause your scripts to pick it up repeatedly,
> then I would have probably avoided `Fixes`/`Cc` -- it is a very minor
> issue.

Now dropped, sorry for the problems.

greg k-h

Re: Patch "kbuild: rust: force `alloc` extern to allow "empty" Rust files" has been added to the 6.1-stable tree

[Miguel Ojeda]

On Fri, May 3, 2024 at 6:42 PM Sasha Levin <sashal@kernel.org> wrote:
>
> If you, or anyone else, feels it should not be added to the stable tree,
> please let <stable@vger.kernel.org> know about it.

I don't think it should be added since it requires upgrading the
compiler to Rust 1.71 (from 1.62) at least, given:

>     be fairly confusing for kernel developers [1], thus use the unstable
>     `force` option of `--extern` [2] (added in Rust 1.71 [3]) to force the
>     compiler to resolve `alloc`.

Now, we have upgraded the compiler in the past (in 6.6 LTS), so it
could be done, but the issue here was small enough (it should only
really affect kernel developers if they happen to create a new file or
similar) that it felt too minor to warrant it (especially since it
would a bigger compiler jump this time, with more changes required
too), so I asked for doing it only in 6.6 and 6.8 since those were
straightforward:

    https://lore.kernel.org/stable/2024042909-whimsical-drapery-40d1@gregkh/

If someone is actually doing development in 6.1 LTS with Rust enabled,
we may create bigger problems for them (even if it is just time used)
by upgrading the compiler than what this fix fixes here (which is an
issue they may not even care about or ever notice).

Cheers,
Miguel

Re: Patch "kbuild: rust: force `alloc` extern to allow "empty" Rust files" has been added to the 6.1-stable tree

[Sasha Levin]

On Fri, May 03, 2024 at 07:10:35PM +0200, Miguel Ojeda wrote:
>On Fri, May 3, 2024 at 6:42 PM Sasha Levin <sashal@kernel.org> wrote:
>>
>> If you, or anyone else, feels it should not be added to the stable tree,
>> please let <stable@vger.kernel.org> know about it.
>
>I don't think it should be added since it requires upgrading the
>compiler to Rust 1.71 (from 1.62) at least, given:
>
>>     be fairly confusing for kernel developers [1], thus use the unstable
>>     `force` option of `--extern` [2] (added in Rust 1.71 [3]) to force the
>>     compiler to resolve `alloc`.
>
>Now, we have upgraded the compiler in the past (in 6.6 LTS), so it
>could be done, but the issue here was small enough (it should only
>really affect kernel developers if they happen to create a new file or
>similar) that it felt too minor to warrant it (especially since it
>would a bigger compiler jump this time, with more changes required
>too), so I asked for doing it only in 6.6 and 6.8 since those were
>straightforward:
>
>    https://lore.kernel.org/stable/2024042909-whimsical-drapery-40d1@gregkh/
>
>If someone is actually doing development in 6.1 LTS with Rust enabled,
>we may create bigger problems for them (even if it is just time used)
>by upgrading the compiler than what this fix fixes here (which is an
>issue they may not even care about or ever notice).

Dropped, thanks!

In general, it would be nice to have a mechanism that matches supported
Rust compilers with whatever is in the kernel tree. This logic of "6.6
is ok but 6.1 is too old" feels so 90s.

-- 
Thanks,
Sasha