On 30 March 2013 06:25, Brad Smith <b...@comstyle.com> wrote:
> init_get_clock() already indicates to get_clock() to utilize
> clock_gettime() with a monotonic clock but OpenBSD was missing
> from the list of OS's to use a monotonic clock on.
>
> What I don't understand is why is init_get_clock() not called
> anywhere?
It's marked as a constructor so the C runtime will call it.
> diff --git a/include/qemu/timer.h b/include/qemu/timer.h
> index 1766b2d..8b86fa5 100644
> --- a/include/qemu/timer.h
> +++ b/include/qemu/timer.h
> @@ -118,7 +118,8 @@ extern int use_rt_clock;
> static inline int64_t get_clock(void)
> {
> #if defined(__linux__) || (defined(__FreeBSD__) && __FreeBSD_version >=
> 500000) \
> - || defined(__DragonFly__) || defined(__FreeBSD_kernel__)
> + || defined(__DragonFly__) || defined(__FreeBSD_kernel__) \
> + || defined(__OpenBSD__)
> if (use_rt_clock) {
> struct timespec ts;
> clock_gettime(CLOCK_MONOTONIC, &ts);
Rather than extending this already long OS dependent #if,
could you write a configure test for whatever we're using
and then just have a simple #ifdef CONFIG_WHATEVER ?
thanks
-- PMM
