Re: [PATCH RFC 1/4] syscalls: Create unified partial table for all archs

"Arnd Bergmann" <[email protected]> Fri, 24 Jul 2026 22:25:05 +0200
Newsgroups org.kernel.vger.linux-kbuild,org.kernel.vger.linux-kernel
Message-ID <[email protected]>
On Fri, Jul 24, 2026, at 22:18, Andr=C3=A9 Almeida wrote:
> Em 24/07/2026 17:09, Arnd Bergmann escreveu:
>> On Fri, Jul 24, 2026, at 22:00, Andr=C3=A9 Almeida wrote:
>> It should really be anything starting at 403, not just 442 and higher.
>>=20
>> There are some trivial differences that have crept in over time,
>> but we should be able to work around those.
>>=20
>
> The thing that prevent me starting before 442 was how to deal with=20
> compat. If I add to the table
>
> 441	common	epoll_pwait2		sys_epoll_pwait2		compat_sys_epoll_pwait2
>
> It will wrongly add
>
> __SYSCALL_WITH_COMPAT(441, sys_epoll_pwait2, compat_sys_epoll_pwait2)
>
> to arch/x86/include/generated/asm/syscalls_64.h.
>
> Maybe I could add a --ignore-compat to scripts/syscalltbl.sh, and add=20
> this flag for 64 builds. What do you think?

IIRC, we can just use compat_sys_epoll_pwait2 for both i386
and x32, because compat_sigset_t is compatible with sigset_t --
they are only really different on big-endian targets, which
count the bits in a u64 different from two u32.

This problem also goes away once x32 is gone.

      Arnd