[PATCH 2/4] arm64: Fix the update_vsyscall() prototype

[John Stultz]

On 10/16/2012 09:46 AM, Catalin Marinas wrote:
> With commit 576094b7 (time: Introduce new GENERIC_TIME_VSYSCALL) the old
> update_vsyscall() prototype is no longer available. This patch updates
> the arm64 port.
>
> Signed-off-by: Catalin Marinas <catalin.marinas at arm.com>
> Acked-by: Will Deacon <will.deacon at arm.com>
> Cc: John Stultz <john.stultz at linaro.org>
> ---
>   arch/arm64/kernel/vdso.c | 20 ++++++++++----------
>   1 file changed, 10 insertions(+), 10 deletions(-)
>
> diff --git a/arch/arm64/kernel/vdso.c b/arch/arm64/kernel/vdso.c
> index 17948fc..ba45794 100644
> --- a/arch/arm64/kernel/vdso.c
> +++ b/arch/arm64/kernel/vdso.c
> @@ -28,6 +28,7 @@
>   #include <linux/sched.h>
>   #include <linux/signal.h>
>   #include <linux/slab.h>
> +#include <linux/timekeeper_internal.h>
>   #include <linux/vmalloc.h>
>   
>   #include <asm/cacheflush.h>
> @@ -222,11 +223,10 @@ struct vm_area_struct *get_gate_vma(struct mm_struct *mm)
>   /*
>    * Update the vDSO data page to keep in sync with kernel timekeeping.
>    */
> -void update_vsyscall(struct timespec *ts, struct timespec *wtm,
> -		     struct clocksource *clock, u32 mult)
> +void update_vsyscall(struct timekeeper *tk)
>   {
>   	struct timespec xtime_coarse;
> -	u32 use_syscall = strcmp(clock->name, "arch_sys_counter");
> +	u32 use_syscall = strcmp(tk->clock->name, "arch_sys_counter");
>   
>   	++vdso_data->tb_seq_count;
>   	smp_wmb();
> @@ -237,13 +237,13 @@ void update_vsyscall(struct timespec *ts, struct timespec *wtm,
>   	vdso_data->xtime_coarse_nsec		= xtime_coarse.tv_nsec;
>   
>   	if (!use_syscall) {
> -		vdso_data->cs_cycle_last	= clock->cycle_last;
> -		vdso_data->xtime_clock_sec	= ts->tv_sec;
> -		vdso_data->xtime_clock_nsec	= ts->tv_nsec;
> -		vdso_data->cs_mult		= mult;
> -		vdso_data->cs_shift		= clock->shift;
> -		vdso_data->wtm_clock_sec	= wtm->tv_sec;
> -		vdso_data->wtm_clock_nsec	= wtm->tv_nsec;
> +		vdso_data->cs_cycle_last	= tk->clock->cycle_last;
> +		vdso_data->xtime_clock_sec	= tk->xtime_sec;
> +		vdso_data->xtime_clock_nsec	= tk->xtime_nsec >> tk->shift;
Sorry for the collision here!

One word of warning:  Truncating the sub-ns base like this can cause 
small single ns inconsistencies from the vsyscalls.

So either use GENERIC_TIME_VSYSCALL_OLD to enable the 1ns generic 
rounding-up code, or rework the vsyscall code to use the sub-ns 
granularity as done in 650ea02475106e8d6bdf561896d2ffe0d1c0ebb4 for 
x86_64.  I hope to drop the _OLD config eventually, so this will need to 
be done at some point.

thanks
-john