Re: [PATCH 05/26] Unexport ticks_per_sec variable. Create get_ticks_per_sec() function

malc wrote:
> And generalizations are always true. Anyhow, i'm explicitly against the
> patch, so first obtain the express acknowledgment from the leaders,
> otherwise i'll revert it should it go in.
>   

I'm adding the following patch to Juan's series.  The result is that 
get_ticks_per_sec() should be optimized to a literal value.  The result 
being that uses of it are faster than they were before (not it should 
matter).

I think the result of this patch is that the refactoring is an 
undeniable improvement.

Regards,

Anthony Liguori

Anthony Liguori <anthony@codemonkey.ws> wrote:
> malc wrote:
>> And generalizations are always true. Anyhow, i'm explicitly against the
>> patch, so first obtain the express acknowledgment from the leaders,
>> otherwise i'll revert it should it go in.
>>   
>
> I'm adding the following patch to Juan's series.  The result is that
> get_ticks_per_sec() should be optimized to a literal value.  The
> result being that uses of it are faster than they were before (not it
> should matter).
>
> I think the result of this patch is that the refactoring is an
> undeniable improvement.
>
> Regards,
>
> Anthony Liguori

Nice.

Acked-by: Juan Quintela <quintela@redhat.com>

>
> commit 1c7aff17af0ca9e1803b952ce455f096c5da8847
> Author: Anthony Liguori <aliguori@us.ibm.com>
> Date:   Fri Sep 11 10:28:26 2009 -0500
>
>     Make get_ticks_per_sec() a static inline
>     
>     ticks_per_sec is a constant.  There's no need to store it as a variable as it
>     never changes since our time is based on units.
>     
>     Convert get_ticks_per_sec() to a static inline and move the constant into
>     qemu-timer.h.  Remove all references to QEMU_TIMER_BASE so that we consistently
>     use this interface.
>     
>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
>
> diff --git a/qemu-timer.h b/qemu-timer.h
> index 00b166d..e44c334 100644
> --- a/qemu-timer.h
> +++ b/qemu-timer.h
> @@ -26,7 +26,10 @@ void qemu_mod_timer(QEMUTimer *ts, int64_t expire_time);
>  int qemu_timer_pending(QEMUTimer *ts);
>  int qemu_timer_expired(QEMUTimer *timer_head, int64_t current_time);
>  
> -int64_t get_ticks_per_sec(void);
> +static inline int64_t get_ticks_per_sec(void)
> +{
> +    return 1000000000LL;
> +}
>  
>  void qemu_get_timer(QEMUFile *f, QEMUTimer *ts);
>  void qemu_put_timer(QEMUFile *f, QEMUTimer *ts);
> diff --git a/vl.c b/vl.c
> index 6052b1c..c3c874d 100644
> --- a/vl.c
> +++ b/vl.c
> @@ -528,8 +528,6 @@ uint64_t muldiv64(uint64_t a, uint32_t b, uint32_t c)
>  /***********************************************************/
>  /* real time host monotonic timer */
>  
> -#define QEMU_TIMER_BASE 1000000000LL
> -
>  #ifdef WIN32
>  
>  static int64_t clock_freq;
> @@ -550,7 +548,7 @@ static int64_t get_clock(void)
>  {
>      LARGE_INTEGER ti;
>      QueryPerformanceCounter(&ti);
> -    return muldiv64(ti.QuadPart, QEMU_TIMER_BASE, clock_freq);
> +    return muldiv64(ti.QuadPart, get_ticks_per_sec(), clock_freq);
>  }
>  
>  #else
> @@ -758,7 +756,7 @@ static void rtc_stop_timer(struct qemu_alarm_timer *t);
>     fairly approximate, so ignore small variation.
>     When the guest is idle real and virtual time will be aligned in
>     the IO wait loop.  */
> -#define ICOUNT_WOBBLE (QEMU_TIMER_BASE / 10)
> +#define ICOUNT_WOBBLE (get_ticks_per_sec() / 10)
>  
>  static void icount_adjust(void)
>  {
> @@ -800,7 +798,7 @@ static void icount_adjust_rt(void * opaque)
>  static void icount_adjust_vm(void * opaque)
>  {
>      qemu_mod_timer(icount_vm_timer,
> -                   qemu_get_clock(vm_clock) + QEMU_TIMER_BASE / 10);
> +                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
>      icount_adjust();
>  }
>  
> @@ -816,7 +814,7 @@ static void init_icount_adjust(void)
>                     qemu_get_clock(rt_clock) + 1000);
>      icount_vm_timer = qemu_new_timer(vm_clock, icount_adjust_vm, NULL);
>      qemu_mod_timer(icount_vm_timer,
> -                   qemu_get_clock(vm_clock) + QEMU_TIMER_BASE / 10);
> +                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
>  }
>  
>  static struct qemu_alarm_timer alarm_timers[] = {
> @@ -1036,15 +1034,10 @@ int64_t qemu_get_clock(QEMUClock *clock)
>      }
>  }
>  
> -int64_t get_ticks_per_sec(void)
> -{
> -    return timers_state.ticks_per_sec;
> -}
> -
>  static void init_timers(void)
>  {
>      init_get_clock();
> -    timers_state.ticks_per_sec = QEMU_TIMER_BASE;
> +    timers_state.ticks_per_sec = get_ticks_per_sec();
>      rt_clock = qemu_new_clock(QEMU_TIMER_REALTIME);
>      vm_clock = qemu_new_clock(QEMU_TIMER_VIRTUAL);
>  }

Anthony Liguori wrote:
> malc wrote:
>> And generalizations are always true. Anyhow, i'm explicitly against the
>> patch, so first obtain the express acknowledgment from the leaders,
>> otherwise i'll revert it should it go in.
>>   
> 
> I'm adding the following patch to Juan's series.  The result is that
> get_ticks_per_sec() should be optimized to a literal value.  The result
> being that uses of it are faster than they were before (not it should
> matter).

Perfect!

> 
> I think the result of this patch is that the refactoring is an
> undeniable improvement.
> 
> Regards,
> 
> Anthony Liguori
> 
> 

Jan

On Friday 11 September 2009, Anthony Liguori wrote:
> malc wrote:
> > And generalizations are always true. Anyhow, i'm explicitly against the
> > patch, so first obtain the express acknowledgment from the leaders,
> > otherwise i'll revert it should it go in.
> 
> I'm adding the following patch to Juan's series.  The result is that
> get_ticks_per_sec() should be optimized to a literal value.  The result
> being that uses of it are faster than they were before (not it should
> matter).

Having this as a function/variable is completely misleading. It is and always 
will be 1000000000. You'd be better off using QEMU_TIMER_BASE directly.

Paul