Re: [Buildroot] [PATCH 1/1] linux: fix LINUX_VERSION_PROBED content

[Giulio Benetti <giulio.benetti@benettiengineering.com>]

Hello Thomas, Vincent,

On 07/02/23 15:16, Thomas Petazzoni via buildroot wrote:
> Hello Giulio,
> 
> On Wed,  4 Jan 2023 03:05:58 +0100
> Giulio Benetti <giulio.benetti@benettiengineering.com> wrote:
> 
>> Actually LINUX_VERSION_PROBED is set with the string sorrounded by the
>> backquotes including them. So when we evaluate $(LINUX_VERSION_PROBED) we
>> end up having the command itself and instead we only need the Linux version
>> probed in the style x.y.z.
>>
>> Verifying it turns out that:
>> 1) --no-print-directory and -s make flags must be passed before -C flag,
>> The error was not visible because errors were suppressed with
>> '2>/dev/null'
>> 2) backquotes don't evaluate the expression they sorround but become part
>> of the string they sorround and get assigned to the variable
>>
>> To fix this let's move --no-print-directory right after $(MAKE) and
>> use $(shell ) instead of backquotes.
>>
>> Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
> 
> What is this fixing exactly? Do you have a specific problem that this
> fixes? Or is this just fixing the fact that LINUX_VERSION_PROBED
> contains `a very long command` instead of the result of the command
> itself?

While building the various rtl* wifi drivers I've noticed KVER='a very 
long command` but it gets evaluated correctly, indeed the .ko module is
installed in the correct "lib/modules/$(LINUX_VERSION_PROBED)" path.

> Indeed, switching back to $(shell ...) like you're proposing would
> result in basically backporting:
> 
> commit 28f57bb863ba48df054cec0057c844bf23d5969f
> Author: Arnout Vandecappelle <arnout@mind.be>
> Date:   Sun Jul 12 16:35:27 2015 +0200
> 
>      linux: use backtick instead of $(shell ...) make function
>      
>      Verified that LINUX_VERSION_PROBED is only used in "-quoted commands
>      (actually, usually it's not quoted).
>      
>      Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
>      Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> 
> which is not desirable, especially if you don't have a good
> justification of why the current situation doesn't work.
> 
> So for now, I'll mark as Rejected, but if you have a good
> justification, you can resend with a better commit log with this
> justification.

Let's keep it as rejected, sorry for the noise!

Best regards
-- 
Giulio Benetti
Benetti Engineering sas
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot