[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87a5l7x6ru.fsf@all.your.base.are.belong.to.us>
Date: Thu, 02 May 2024 23:07:01 +0200
From: Björn Töpel <bjorn@...nel.org>
To: Jean-Philippe Brucker <jean-philippe@...aro.org>
Cc: Jiri Olsa <olsajiri@...il.com>, Arnaldo Carvalho de Melo
<acme@...hat.com>, linux-kernel@...r.kernel.org, Björn Töpel
<bjorn@...osinc.com>, Nathan Chancellor <nathan@...nel.org>, Nick
Desaulniers <ndesaulniers@...gle.com>, Tom Rix <trix@...hat.com>,
bpf@...r.kernel.org, Anders Roxell <anders.roxell@...aro.org>,
llvm@...ts.linux.dev
Subject: Re: [PATCH v2] tools/build: Add clang cross-compilation flags to
feature detection
Jean-Philippe Brucker <jean-philippe@...aro.org> writes:
> On Fri, Apr 26, 2024 at 12:31:17PM +0200, Björn Töpel wrote:
>> Jiri Olsa <olsajiri@...il.com> writes:
>>
>> > On Thu, Nov 02, 2023 at 11:32:52AM +0100, Björn Töpel wrote:
>> >> From: Björn Töpel <bjorn@...osinc.com>
>> >>
>> >> When a tool cross-build has LLVM=1 set, the clang cross-compilation
>> >> flags are not passed to the feature detection build system. This
>> >> results in the host's features are detected instead of the targets.
>> >>
>> >> E.g, triggering a cross-build of bpftool:
>> >>
>> >> cd tools/bpf/bpftool
>> >> make ARCH=riscv CROSS_COMPILE=riscv64-linux-gnu- LLVM=1
>> >>
>> >> would report the host's, and not the target's features.
>> >>
>> >> Correct the issue by passing the CLANG_CROSS_FLAGS variable to the
>> >> feature detection makefile.
>> >>
>> >> Fixes: cebdb7374577 ("tools: Help cross-building with clang")
>> >> Signed-off-by: Björn Töpel <bjorn@...osinc.com>
>> >
>> > Acked-by: Jiri Olsa <jolsa@...nel.org>
>>
>> Waking up the dead!
>>
>> Arnaldo, Jean-Philippe: I'm still stung by what this patch fixes. LMK
>> what you need from me/this patch to pick it up.
>
> I guess the problem is these files don't have a specific tree. Since you
> mention BPF maybe it should go through the BPF tree, in which case you
> could resend to the tools/bpf maintainers (and "PATCH bpf" subject prefix)
>
> FWIW the change looks good to me:
>
> Reviewed-by: Jean-Philippe Brucker <jean-philippe@...aro.org>
Thanks for the review. Yes, maybe it's a orphan files issue! It does
seem a bit weird to route it via BPF, since it's just a regular
cross-build fix -- Not directly related to BPF. Regardless; I'll do a
respin targetted at *some* tree. ;-)
Björn
Powered by blists - more mailing lists