Re: [PATCH v5 02/23] kernel: Define gettimeofday vdso common code

From: Arnd Bergmann <arnd@arndb.de>
To: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: linux-arch <linux-arch@vger.kernel.org>,
	Shuah Khan <shuah@kernel.org>,
	Rasmus Villemoes <linux@rasmusvillemoes.dk>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	Will Deacon <will.deacon@arm.com>,
	Russell King <linux@armlinux.org.uk>,
	Ralf Baechle <ralf@linux-mips.org>,
	Mark Salyzyn <salyzyn@android.com>,
	Paul Burton <paul.burton@mips.com>,
	Dmitry Safonov <0x7f454c46@gmail.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Peter Collingbourne <pcc@google.com>,
	Linux ARM <linux-arm-kernel@lists.infradead.org>
Subject: Re: [PATCH v5 02/23] kernel: Define gettimeofday vdso common code
Date: Thu, 28 Feb 2019 10:29:50 +0100	[thread overview]
Message-ID: <CAK8P3a2D44CDXbcKB-WtphYY9E74ovFPs+e4Ti3fUqRL77AYhw@mail.gmail.com> (raw)
In-Reply-To: <1f9b21cd-81f5-dfa9-70fb-af8228195cb4@arm.com>

On Wed, Feb 27, 2019 at 3:52 PM Vincenzo Frascino
<vincenzo.frascino@arm.com> wrote:
> On 22/02/2019 13:49, Arnd Bergmann wrote:
> > On Fri, Feb 22, 2019 at 1:25 PM Vincenzo Frascino
> > <vincenzo.frascino@arm.com> wrote:
> >
> >> +/*
> >> + * The definitions below are required to overcome the limitations
> >> + * of time_t on 32 bit architectures, which overflows in 2038.
> >> + * The new code should use the replacements based on time64_t and
> >> + * timespec64.
> >> + *
> >> + * The abstraction below will be updated once the migration to
> >> + * time64_t is complete.
> >> + */
> >> +#ifdef CONFIG_GENERIC_VDSO_32
> >> +#define __vdso_timespec                old_timespec32
> >> +#define __vdso_timeval         old_timeval32
> >> +#else
> >> +#ifdef ENABLE_COMPAT_VDSO
> >> +#define __vdso_timespec                old_timespec32
> >> +#define __vdso_timeval         old_timeval32
> >> +#else
> >> +#define __vdso_timespec                __kernel_timespec
> >> +#define __vdso_timeval         __kernel_old_timeval
> >> +#endif /* CONFIG_COMPAT_VDSO */
> >> +#endif /* CONFIG_GENERIC_VDSO_32 */

> > For __vdso_timespec, I see how you ended up with this
> > redefinition, and it makes the current version of your patches
> > easier, but I fear it will in turn make it harder to add the
> > __kernel_old_timeval based variant.
> >
>
> What is __kernel_old_timespec (based on you next email)? Why do you think it
> will make harder to add the new variants?

I mean you should ideally use the types that you have listed above
directly, without the abstraction.

In the long run I want one set of functions using __kernel_timespec that
implements the 64-bit time interfaces on both 32-bit and 64-bit kernels,
including the compat vdso, and another set using old_timespec32 for
just the native 32-bit and compat version.

This would match what we do in the normal system calls (in linux-5.1+),
where we always have the regular implementation use 64-bit types
only, and have an optional _time32 version that is used for existing
32-bit user space, on both native 32 bit kernels and compat syscalls
on 64-bit.

       Arnd

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel