linux-kernel - Re: kbuild: check the minimum compiler version in Kconfig

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <CAKwvOdnNA7pFGNuord-yiArE55oPNCCwCOHdrVQiRiATtYC23g@mail.gmail.com>
Date:   Wed, 16 Nov 2022 15:10:59 -0800
From:   Nick Desaulniers <ndesaulniers@...gle.com>
To:     Florian Fainelli <f.fainelli@...il.com>
Cc:     Masahiro Yamada <masahiroy@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Paul Gortmaker <paul.gortmaker@...driver.com>,
        Nathan Chancellor <natechancellor@...il.com>, ojeda@...nel.org,
        sedat.dilek@...il.com, clang-built-linux <llvm@...ts.linux.dev>,
        Manoj Gupta <manojgupta@...gle.com>
Subject: Re: kbuild: check the minimum compiler version in Kconfig

On Wed, Nov 16, 2022 at 2:56 PM Florian Fainelli <f.fainelli@...il.com> wrote:
>
> On 11/16/22 14:48, Nick Desaulniers wrote:
> > On Wed, Nov 16, 2022 at 1:48 PM Florian Fainelli <f.fainelli@...il.com> wrote:
> >>
> >> Hi Masahiro, Nick,
> >>
> >> The compiler version check performed with
> >> aec6c60a01d3a3170242d6a99372a388e1136dc6 ("kbuild: check the minimum
> >> compiler version in Kconfig") can be defeated and prevent running the
> >> *config targets if specifying LLVM=1 on the command line, in that way:
> >>
> >> BR_BINARIES_DIR=/local/users/fainelli/buildroot-llvm/output/arm64/images
> >> LLVM=1 LLVM_IAS=1 PKG_CONFIG_PATH="" /usr/bin/make -j49 -C
> >> /local/users/fainelli/buildroot-llvm/output/arm64/build/linux-custom
> >> HOSTCC="/usr/bin/gcc"
> >> HOSTCC="/local/users/fainelli/buildroot-llvm/output/arm64/host/bin/ccache
> >> /usr/bin/gcc -O2
> >> -I/local/users/fainelli/buildroot-llvm/output/arm64/host/include
> >> -DNDEBUG -L/local/users/fainelli/buildroot-llvm/output/arm64/host/lib
> >> -Wl,-rpath,/local/users/fainelli/buildroot-llvm/output/arm64/host/lib"
> >> ARCH=arm64
> >> INSTALL_MOD_PATH=/local/users/fainelli/buildroot-llvm/output/arm64/target
> >> CROSS_COMPILE="/local/users/fainelli/buildroot-llvm/output/arm64/host/bin/aarch64-linux-"
> >> WERROR=0
> >> DEPMOD=/local/users/fainelli/buildroot-llvm/output/arm64/host/sbin/depmod
> >> INSTALL_MOD_STRIP=1 HOSTCC="/usr/bin/gcc" menuconfig
> >>     UPD     scripts/kconfig/mconf-cfg
> >>     HOSTCC  scripts/kconfig/mconf.o
> >>     HOSTCC  scripts/kconfig/lxdialog/checklist.o
> >>     HOSTCC  scripts/kconfig/lxdialog/inputbox.o
> >>     HOSTCC  scripts/kconfig/lxdialog/menubox.o
> >>     HOSTCC  scripts/kconfig/lxdialog/textbox.o
> >>     HOSTCC  scripts/kconfig/lxdialog/util.o
> >>     HOSTCC  scripts/kconfig/lxdialog/yesno.o
> >>     HOSTLD  scripts/kconfig/mconf
> >> ***
> >> *** Compiler is too old.
> >> ***   Your Clang version:    10.0.0
> >> ***   Minimum Clang version: 10.0.1
> >> ***
> >> scripts/Kconfig.include:44: Sorry, this compiler is not supported.
> >>
> >> Here, the compiler check is actually checking the host compiler clang
> >> version installed on my Ubuntu 20.04 system, as opposed to the cross
> >> compiler clang version that is being used.
> >
> > LLVM=1 will use `clang` as found by your $PATH.  Where did you express
> > to make what the "cross compiler clang version" is? (And why do you
> > set HOSTCC three times)
>
> We are setting CROSS_COMPILE to express the cross compiler clang prefix
> to use.

Can you try setting CLANG_PREFIX rather than CROSS_COMPILE (to the same value)?

CLANG_PREFIX=/local/users/fainelli/buildroot-llvm/output/arm64/host/bin/aarch64-linux-


>
> Setting HOSTCC three times is done by buildroot, it does it through make
> flags, and then variables passed to make etc. clearly we could change
> that, but that's out of scope is not it?
> --
> Florian
>


--
Thanks,
~Nick Desaulniers