If the processor support dynamic cpu frequency and the support is
enabled in kernel, this sched_clock() implementation will be broken(and
If the frequency of the MIPS CP0 counter is related to the cpu's
frequency).
So, some extra resitrictions should be added to it.
arch/mips/Kconfig
config CPU_HAS_FIXED_CP0_COUNTER
bool
config SYS_SUPPORTS_HRES_SCHED_CLOCK
bool
depends on CPU_HAS_FIXED_CP0_COUNTER || !CPU_FREQ
arch/mips/kernel/csrc-r4k.c
#ifdef SYS_SUPPORTS_HRES_SCHED_CLOCK
/* The high resolution version of sched_clock() */
#endif
And I'm not sure whether the cavium octeon support dynamic cpu
frequency, if yes, it's high resolution version of sched_clock() also
should be wrapped with the above macro to ensure it is not broken:
arch/mips/cavium-octeon/csrc-octeon.c
Regards,
Wu Zhangjin
On Fri, 2009-11-27 at 18:51 +0800, Wu Zhangjin wrote:
> From: Wu Zhangjin <wuzhangjin@xxxxxxxxx>
>
> (The changes of this v6 revision from v5 revision:
>
> o hard-codes the cycle2ns_scale_factor as 8 for 30(cs->shift) is too
> big. With 30, the return value of sched_clock() will also overflow quickly.
> o moves the sched_clock() back into csrc-r4k.c as David and Sergei
> recommended.
> o inits c0 count as zero for PRINTK_TIME=y.
> o drops the HR_SCHED_CLCOK option for the current sched_clock() is stable
> enough to replace the jiffies based one.
> )
>
> This patch adds a cnt32_to_63() and MIPS c0 count based sched_clock(),
> which provides high resolution.
>
> Without it, the Ftrace for MIPS will give useless timestamp information.
>
> Because cnt32_to_63() needs to be called at least once per half period
> to work properly, Differ from the old version, this v2 revision set up a
> kernel timer to ensure the requirement of some MIPSs which have short c0
> count period.
>
> And also, we init the c0 count as ZERO(just as jiffies does) in
> time_init() before plat_time_init(), without it, PRINTK_TIME=y will get
> wrong timestamp information. (NOTE: some platforms have initiazlied c0
> count as zero, but some not, this may introduce some duplication,
> perhaps a new patch is needed to remove the initialized of c0 count in
> the platforms later?)
>
> This is originally from arch/arm/plat-orion/time.c
>
> This revision works well for function graph tracer now, and also,
> PRINTK_TIME=y will get normal timestamp informatin.
>
> Signed-off-by: Wu Zhangjin <wuzhangjin@xxxxxxxxx>
> ---
> arch/mips/kernel/csrc-r4k.c | 54 +++++++++++++++++++++++++++++++++++++++++++
> arch/mips/kernel/time.c | 3 ++
> 2 files changed, 57 insertions(+), 0 deletions(-)
>
> diff --git a/arch/mips/kernel/csrc-r4k.c b/arch/mips/kernel/csrc-r4k.c
> index e95a3cd..12755f2 100644
> --- a/arch/mips/kernel/csrc-r4k.c
> +++ b/arch/mips/kernel/csrc-r4k.c
> @@ -6,10 +6,62 @@
> * Copyright (C) 2007 by Ralf Baechle
> */
> #include <linux/clocksource.h>
> +#include <linux/cnt32_to_63.h>
> #include <linux/init.h>
> +#include <linux/timer.h>
>
> #include <asm/time.h>
>
> +/*
> + * MIPS sched_clock implementation.
> + *
> + * Because the hardware timer period is quite short and because cnt32_to_63()
> + * needs to be called at least once per half period to work properly, a kernel
> + * timer is set up to ensure this requirement is always met.
> + *
> + * Please refer to include/linux/cnt32_to_63.h and arch/arm/plat-orion/time.c
> + */
> +#define CLOCK2NS_SCALE_FACTOR 8
> +
> +static unsigned long clock2ns_scale;
> +
> +unsigned long long notrace sched_clock(void)
> +{
> + unsigned long long v = cnt32_to_63(read_c0_count());
> + return (v * clock2ns_scale) >> CLOCK2NS_SCALE_FACTOR;
> +}
> +
> +static struct timer_list cnt32_to_63_keepwarm_timer;
> +
> +static void cnt32_to_63_keepwarm(unsigned long data)
> +{
> + mod_timer(&cnt32_to_63_keepwarm_timer, round_jiffies(jiffies + data));
> + sched_clock();
> +}
> +
> +static void __init setup_hres_sched_clock(unsigned long clock)
> +{
> + unsigned long long v;
> + unsigned long data;
> +
> + v = NSEC_PER_SEC;
> + v <<= CLOCK2NS_SCALE_FACTOR;
> + v += clock/2;
> + do_div(v, clock);
> + /*
> + * We want an even value to automatically clear the top bit
> + * returned by cnt32_to_63() without an additional run time
> + * instruction. So if the LSB is 1 then round it up.
> + */
> + if (v & 1)
> + v++;
> + clock2ns_scale = v;
> +
> + data = 0x80000000UL / clock * HZ;
> + setup_timer(&cnt32_to_63_keepwarm_timer, cnt32_to_63_keepwarm, data);
> + mod_timer(&cnt32_to_63_keepwarm_timer, round_jiffies(jiffies + data));
> +}
> +
> static cycle_t c0_hpt_read(struct clocksource *cs)
> {
> return read_c0_count();
> @@ -27,6 +79,8 @@ int __init init_r4k_clocksource(void)
> if (!cpu_has_counter || !mips_hpt_frequency)
> return -ENXIO;
>
> + setup_hres_sched_clock(mips_hpt_frequency);
> +
> /* Calculate a somewhat reasonable rating value */
> clocksource_mips.rating = 200 + mips_hpt_frequency / 10000000;
>
> diff --git a/arch/mips/kernel/time.c b/arch/mips/kernel/time.c
> index 1f467d5..4b5e93c 100644
> --- a/arch/mips/kernel/time.c
> +++ b/arch/mips/kernel/time.c
> @@ -152,6 +152,9 @@ static __init int cpu_has_mfc0_count_bug(void)
>
> void __init time_init(void)
> {
> + if (!mips_clockevent_init() || !cpu_has_mfc0_count_bug())
> + write_c0_count(0);
> +
> plat_time_init();
>
> if (!mips_clockevent_init() || !cpu_has_mfc0_count_bug())
