[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190729144831.GA21120@linux.intel.com>
Date: Mon, 29 Jul 2019 07:48:31 -0700
From: Sean Christopherson <sean.j.christopherson@...el.com>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
Andy Lutomirski <luto@...nel.org>,
Vincenzo Frascino <vincenzo.frascino@....com>,
Kees Cook <keescook@...omium.org>,
Paul Bolle <pebolle@...cali.nl>, Will Deacon <will@...nel.org>
Subject: Re: [patch 3/5] lib/vdso/32: Provide legacy syscall fallbacks
On Sun, Jul 28, 2019 at 03:12:54PM +0200, Thomas Gleixner wrote:
> To address the regression which causes seccomp to deny applications the
> access to clock_gettime64() and clock_getres64() syscalls because they
> are not enabled in the existing filters.
>
> That trips over the fact that 32bit VDSOs use the new clock_gettime64() and
> clock_getres64() syscalls in the fallback path.
>
> Implement a __cvdso_clock_get*time32() variants which invokes the legacy
> 32bit syscalls when the architecture requests it.
>
> The conditional can go away once all architectures are converted.
>
> Fixes: 00b26474c2f1 ("lib/vdso: Provide generic VDSO implementation")
> Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
> ---
> lib/vdso/gettimeofday.c | 46 +++++++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 45 insertions(+), 1 deletion(-)
>
> --- a/lib/vdso/gettimeofday.c
> +++ b/lib/vdso/gettimeofday.c
> @@ -117,6 +117,8 @@ static __maybe_unused int
> return 0;
> }
>
> +#ifndef VDSO_HAS_32BIT_FALLBACK
> +
> static __maybe_unused int
> __cvdso_clock_gettime32(clockid_t clock, struct old_timespec32 *res)
> {
> @@ -132,10 +134,29 @@ static __maybe_unused int
> res->tv_sec = ts.tv_sec;
> res->tv_nsec = ts.tv_nsec;
> }
> -
> return ret;
> }
>
> +#else
> +
> +static __maybe_unused int
> +__cvdso_clock_gettime32(clockid_t clock, struct old_timespec32 *res)
> +{
> + struct __kernel_timespec ts;
> + int ret;
> +
> + ret = __cvdso_clock_gettime_common(clock, &ts);
> +
> + if (likely(!ret)) {
> + res->tv_sec = ts.tv_sec;
> + res->tv_nsec = ts.tv_nsec;
> + return 0;
> + }
> + return clock_gettime32_fallback(clock, res);
> +}
> +
> +#endif
> +
> static __maybe_unused int
> __cvdso_gettimeofday(struct __kernel_old_timeval *tv, struct timezone *tz)
> {
> @@ -225,6 +246,8 @@ int __cvdso_clock_getres(clockid_t clock
> return 0;
> }
>
> +#ifndef VDSO_HAS_32BIT_FALLBACK
> +
> static __maybe_unused int
> __cvdso_clock_getres_time32(clockid_t clock, struct old_timespec32 *res)
> {
> @@ -241,4 +264,25 @@ static __maybe_unused int
> }
> return ret;
> }
> +
> +#else
> +
> +static __maybe_unused int
> +__cvdso_clock_getres_time32(clockid_t clock, struct old_timespec32 *res)
> +{
> + struct __kernel_timespec ts;
> + int ret;
> +
> + ret = __cvdso_clock_getres_common(clock, &ts);
> +
> + if (likely(!ret)) {
> + res->tv_sec = ts.tv_sec;
> + res->tv_nsec = ts.tv_nsec;
> + return 0;
> + }
> +
> + return clock_getres32_fallback(clock, res);
> +}
> +#endif
> +
> #endif /* VDSO_HAS_CLOCK_GETRES */
Any reason not to have the #ifndef apply only to the fallback? Wrapping
the entire function and flipping the order of handling 'ret' makes it a bit
difficult to discern that the only difference is the fallback invocation.
static __maybe_unused int
__cvdso_clock_gettime32(clockid_t clock, struct old_timespec32 *res)
{
struct __kernel_timespec ts;
int ret;
ret = __cvdso_clock_gettime_common(clock, &ts);
if (unlikely(ret))
#ifndef VDSO_HAS_32BIT_FALLBACK
ret = clock_gettime_fallback(clock, &ts);
#else
return clock_gettime32_fallback(clock, res);
#endif
if (likely(!ret)) {
res->tv_sec = ts.tv_sec;
res->tv_nsec = ts.tv_nsec;
}
return ret;
}
static __maybe_unused int
__cvdso_clock_getres_time32(clockid_t clock, struct old_timespec32 *res)
{
struct __kernel_timespec ts;
int ret;
ret = __cvdso_clock_getres_common(clock, &ts);
if (unlikely(ret))
#ifndef VDSO_HAS_32BIT_FALLBACK
ret = clock_getres_fallback(clock, &ts);
#else
return clock_getres32_fallback(clock, res);
#endif
if (likely(!ret)) {
res->tv_sec = ts.tv_sec;
res->tv_nsec = ts.tv_nsec;
}
return ret;
}
Powered by blists - more mailing lists