linux-kernel - Re: [PATCH] tools: fix unavoidable GCC call in Clang builds

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Yg6PousYSZXimFCS@dev-arch.archlinux-ax161>
Date:   Thu, 17 Feb 2022 11:10:42 -0700
From:   Nathan Chancellor <nathan@...nel.org>
To:     Adrian Ratiu <adrian.ratiu@...labora.com>
Cc:     llvm@...ts.linux.dev, linux-kernel@...r.kernel.org,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Manoj Gupta <manojgupta@...omium.com>
Subject: Re: [PATCH] tools: fix unavoidable GCC call in Clang builds

Hi Adrian

On Thu, Feb 17, 2022 at 05:36:48AM +0200, Adrian Ratiu wrote:
> In ChromeOS and Gentoo we catch any unwanted mixed Clang/LLVM
> and GCC/binutils usage via toolchain wrappers which fail builds.

Neat for hermetic builds.

> This has revealed that GCC is called unconditionally in Clang
> configured builds to populate GCC_TOOLCHAIN_DIR.
> 
> Allow overriding the variable to avoid the GCC call - in our
> case we can set GCC_TOOLCHAIN_DIR directly in the ebuild recipe.

Would you just set GCC_TOOLCHAIN_DIR to nothing to avoid triggering the
'gcc -print-sysroot' call?

An alternative might be allowing CLANG_CROSS_FLAGS to be supplied by the
user, so that you can stil benefit from cross compiling tools, but with
the flags and sysroot that you expect.

> Suggested-by: Manoj Gupta <manojgupta@...omium.com>
> Signed-off-by: Adrian Ratiu <adrian.ratiu@...labora.com>

Regardless, as long as it works for your use case:

Acked-by: Nathan Chancellor <nathan@...nel.org>

The change that this patch fixes [1] went via bpf-next [2] last cycle,
you could either target that tree to have it fixed for 5.18 or make a
case for having it merged in the 5.17 cycle via bpf [3]. I would resend
this change to the kernel/bpf maintainers, as Nick and I do not
currently pick up patches.

[1]: https://lore.kernel.org/r/20211216163842.829836-2-jean-philippe@linaro.org/
[2]: https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/
[3]: https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git/

> ---
>  tools/scripts/Makefile.include | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/scripts/Makefile.include b/tools/scripts/Makefile.include
> index 79d102304470..98c098c064dd 100644
> --- a/tools/scripts/Makefile.include
> +++ b/tools/scripts/Makefile.include
> @@ -90,7 +90,7 @@ EXTRA_WARNINGS += -Wstrict-aliasing=3
>  
>  else ifneq ($(CROSS_COMPILE),)
>  CLANG_CROSS_FLAGS := --target=$(notdir $(CROSS_COMPILE:%-=%))
> -GCC_TOOLCHAIN_DIR := $(dir $(shell which $(CROSS_COMPILE)gcc 2>/dev/null))
> +GCC_TOOLCHAIN_DIR ?= $(dir $(shell which $(CROSS_COMPILE)gcc 2>/dev/null))
>  ifneq ($(GCC_TOOLCHAIN_DIR),)
>  CLANG_CROSS_FLAGS += --prefix=$(GCC_TOOLCHAIN_DIR)$(notdir $(CROSS_COMPILE))
>  CLANG_CROSS_FLAGS += --sysroot=$(shell $(CROSS_COMPILE)gcc -print-sysroot)
> -- 
> 2.35.0
>