[PATCH 01/12] scripts: subarch.include: fix SUBARCH on MacOS hosts

Sat Aug 24 19:34:58 UTC 2024

On Sat, Aug 24, 2024 at 12:14 AM Daniel Gomez (Samsung)
<d+samsung at kruces.com> wrote:
>
> On Fri, Aug 23, 2024 at 6:13 PM Masahiro Yamada <masahiroy at kernel.org> wrote:
> >
> > On Wed, Aug 7, 2024 at 8:10 AM Daniel Gomez via B4 Relay
> > <devnull+da.gomez.samsung.com at kernel.org> wrote:
> > >
> > > From: Nick Desaulniers <nick.desaulniers at gmail.com>
> > >
> > > When building the Linux kernel on an aarch64 MacOS based host, if we don't
> > > specify a value for ARCH when invoking make, we default to arm and thus
> > > multi_v7_defconfig rather than the expected arm64 and arm64's defconfig.
> > >
> > > This is because subarch.include invokes `uname -m` which on MacOS hosts
> > > evaluates to `arm64` but on Linux hosts evaluates to `aarch64`,
> > >
> > > This allows us to build ARCH=arm64 natively on MacOS (as in ARCH need
> > > not be specified on an aarch64-based system).
> > >
> > > Utilize a negative lookahead regular expression to avoid matching arm64.
> >
> >
> > Does sed support "negative lookahead regular expression"?
>
> I think they removed support for PCRE. I've found this:
>
> commit 261c7f145d015d9acb79dc650d27e4a23b839c23
> Author: Assaf Gordon <assafgordon at gmail.com>
> Date:   Tue Aug 21 14:25:57 2018 -0600
>
>     maint: remove REG_PERL code
>
>     Perl-regexp syntax (PCRE) in GNU Sed is shelved indefinitely.
>     See https://bugs.gnu.org/22801 , https://bugs.gnu.org/22647 .
>     Remove all (unused) REG_PERL related code.
>
>     * sed/sed.c, sed/sed.h, sed/regexp.c, sed/compile.c: Remove REG_PERL code.
>
> git tag --contains 261c7f145d015d9acb79dc650d27e4a23b839c23
> v4.6
> v4.7
> v4.8
> v4.9
>
> And my sed version is (Debian):
>
> sed --version
> sed (GNU sed) 4.9
> Packaged by Debian
> Copyright (C) 2022 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>.
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.
>
> Written by Jay Fenlason, Tom Lord, Ken Pizzini,
> Paolo Bonzini, Jim Meyering, and Assaf Gordon.
>
> This sed program was built with SELinux support.
> SELinux is disabled on this system.
>
> GNU sed home page: <https://www.gnu.org/software/sed/>.
> General help using GNU software: <https://www.gnu.org/gethelp/>.
> E-mail bug reports to: <bug-sed at gnu.org>.
>
> sed version (Homebrew):
> sed --version
> sed (GNU sed) 4.9
> Copyright (C) 2022 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>.
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.
>
> Written by Jay Fenlason, Tom Lord, Ken Pizzini,
> Paolo Bonzini, Jim Meyering, and Assaf Gordon.
>
> This sed program was built without SELinux support.
>
> GNU sed home page: <https://www.gnu.org/software/sed/>.
> General help using GNU software: <https://www.gnu.org/gethelp/>.
> E-mail bug reports to: <bug-sed at gnu.org>.
>
> >
> > >
> > > Add a separate expression to support for armv.* as per error reported by
> > > Nicolas Schier [1].
> > >
> > > [1] https://lore.kernel.org/all/Y3MRvtwdjIwMHvRo@bergen.fjasle.eu/#t
> > >
> > > Signed-off-by: Nick Desaulniers <nick.desaulniers at gmail.com>
> > > Signed-off-by: Daniel Gomez <da.gomez at samsung.com>
> > > ---
> > >  scripts/subarch.include | 3 ++-
> > >  1 file changed, 2 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/scripts/subarch.include b/scripts/subarch.include
> > > index 4bd327d0ae42..5d84ad8c0dee 100644
> > > --- a/scripts/subarch.include
> > > +++ b/scripts/subarch.include
> > > @@ -6,7 +6,8 @@
> > >
> > >  SUBARCH := $(shell uname -m | sed -e s/i.86/x86/ -e s/x86_64/x86/ \
> > >                                   -e s/sun4u/sparc64/ \
> > > -                                 -e s/arm.*/arm/ -e s/sa110/arm/ \
> > > +                                 -e s/armv.*/arm/ \
> > > +                                 -e s/arm$?:\(?!64$.*\)/arm/ -e s/sa110/arm/ \
> >
> >
> > s/arm$?:\(?!64$.*\)/arm/
> >
> > In sed, this expression does not seem to match anything.
>
> You are correct. I've removed the expression and saw no difference.
> See below with my test case:
> >
> > (or please give me some matching examples if I miss something)
>
> cat Makefile
> MACHINE ?= "aarch64"
> SUBARCH0 := $(shell echo $(MACHINE) | sed \
>                                   -e s/arm.*/arm/ \
>                                   -e s/aarch64.*/arm64/)
>
> SUBARCH1 := $(shell echo $(MACHINE) | sed \
>                                   -e s/armv.*/arm/ \
>                                   -e s/aarch64.*/arm64/)
>
> SUBARCH2 := $(shell echo $(MACHINE) | sed \
>                                   -e /^arm64$/!s/arm.*/arm/ \
>                                   -e s/aarch64.*/arm64/)
>
> test:
>         @echo "MACHINE=$(MACHINE)"
>         @echo "SUBARCH0=$(SUBARCH0)"
>         @echo "SUBARCH1=$(SUBARCH1)"
>         @echo "SUBARCH2=$(SUBARCH2)"
>         @echo "---"
>
> SUBARCH0 represents the current upstream expressions for arm/arm64.
> SUBARCH1 is my proposal in case we need to cover only armv* for 32-bit
> arm (I think that is incomplete?) and SUBARCH2 is Nicolas' proposal
> (which I can't make it work in the test Makefile).

To make Nicolas's expression work in Makefile I just need to pass 2 $ like this:

diff -u Makefile.old Makefile

--- Makefile.old  2024-08-24 21:25:28.525267566 +0200
+++ Makefile    2024-08-24 21:28:32.640477991 +0200
@@ -8,7 +8,7 @@
                                  -e s/aarch64.*/arm64/)

 SUBARCH2 := $(shell echo $(MACHINE) | sed \
-                                 -e /^arm64$/!s/arm.*/arm/ \
+                                 -e /^arm64$$/!s/arm.*/arm/ \
                                  -e s/aarch64.*/arm64/)

 test:

And all test cases passed. So, I will include this change for v2.

>
> Running the above Makefile, I get:
>
> make test MACHINE=armv4 && make test MACHINE=arm7 && make test
> MACHINE=armhf && make test MACHINE=aarch64 && make test MACHINE=arm64
> MACHINE=armv4
> SUBARCH0=arm
> SUBARCH1=arm
> SUBARCH2=armv4
> ---
> MACHINE=arm7
> SUBARCH0=arm
> SUBARCH1=arm7
> SUBARCH2=arm7
> ---
> MACHINE=armhf
> SUBARCH0=arm
> SUBARCH1=armhf
> SUBARCH2=armhf
> ---
> MACHINE=aarch64
> SUBARCH0=arm64
> SUBARCH1=arm64
> SUBARCH2=arm64
> ---
> MACHINE=arm64
> SUBARCH0=arm
> SUBARCH1=arm64
> SUBARCH2=arm64
> ---
> >
> >
> >
> >
> >
> > Nocolas already provided correct code:
> >
> > > [1] https://lore.kernel.org/all/Y3MRvtwdjIwMHvRo@bergen.fjasle.eu/#t
>
> I think it is even more simple if we just make this change:
>
> -                                 -e s/arm.*/arm/ -e s/sa110/arm/ \
> +                                 -e s/armv.*/arm/ \
>
> Does armv.* cover all arm32 machines? I see armhf, arm7, arm8 and
> armv*, is it correct?
>
> And thanks for checking!
>
> >
> >
> >
> >
> >
> >
> > >                                   -e s/s390x/s390/ \
> > >                                   -e s/ppc.*/powerpc/ -e s/mips.*/mips/ \
> > >                                   -e s/sh[234].*/sh/ -e s/aarch64.*/arm64/ \
> > >
> > > --
> > > Git-146)
> > >
> > >
> >
> >
> > --
> > Best Regards
> >
> >
> > Masahiro Yamada
