Re: [PATCH] kbuild: add resolve_btfids to pacman PKGBUILD

[Peter Jung <admin@ptr1337.dev>]

On 03.11.24 10:47, Masahiro Yamada wrote:
> On Sat, Nov 2, 2024 at 9:06 PM Peter Jung<admin@ptr1337.dev> wrote:
>> If the config is using DEBUG_INFO_BTF, it is required to,
>> package resolve_btfids with.
>> Compiling dkms modules will fail otherwise.
>>
>> Add a check, if resolve_btfids is present and then package it, if required.
>>
>> Signed-off-by: Peter Jung<admin@ptr1337.dev>
>> ---
>>   scripts/package/PKGBUILD | 5 +++++
>>   1 file changed, 5 insertions(+)
>>
>> diff --git a/scripts/package/PKGBUILD b/scripts/package/PKGBUILD
>> index f83493838cf9..4010899652b8 100644
>> --- a/scripts/package/PKGBUILD
>> +++ b/scripts/package/PKGBUILD
>> @@ -91,6 +91,11 @@ _package-headers() {
>>                  "${srctree}/scripts/package/install-extmod-build" "${builddir}"
>>          fi
>>
>> +       # required when DEBUG_INFO_BTF_MODULES is enabled
>> +       if [ -f tools/bpf/resolve_btfids/resolve_btfids ]; then
>> +               install -Dt "$builddir/tools/bpf/resolve_btfids" tools/bpf/resolve_btfids/resolve_btfids
>> +       fi
>> +
> This is not the right place.
>
> scripts/package/install-extmod-build is a script to set up
> the build environment to build external modules.
> It is shared by rpm-pkg, deb-pkg, and pacman-pkg.
>
>
> https://github.com/torvalds/linux/blob/v6.12-rc5/scripts/package/install-extmod-build#L34
>
> You will see how objtool is copied.
>
>
>
>
> (Anyway, it depends on your urgency.
> My hope is to support objtool and resolve_btfids in more generic ways.)
>

Thanks Masahiro for the suggestion. I will look into and likely bring a v2.
I did not know about other distribution/package managers, if this is 
also a problem at them.

At archlinux we have included this since a while already in the 
PKGBUILD, see here:
https://gitlab.archlinux.org/archlinux/packaging/packages/linux/-/blob/main/PKGBUILD?ref_type=heads#L151-152

I will also make the change to grep for DEBUG_INFO_BTF in the config 
with the is_enabled function, instead of checking the path


Regards,

Peter