[dpdk-dev] [PATCH v5 2/2] test/power: round cpuinfo cur freq value in cpufreq autotest
David Hunt
david.hunt at intel.com
Tue Apr 20 16:01:38 CEST 2021
On 15/4/2021 6:59 AM, Richael Zhuang wrote:
> The value in "/sys/.../cpuinfo_cur_freq" may not be exactly the
> same as what was set. For example, if "2400000" is written to
> "/sys/.../cpufreq/scaling_setspeed" to set the frequency, then the
> value in "/sys/.../cpuinfo_cur_freq" may be "2401222". So need to
> round the value.
>
> Fixes: ed7c51a6a680 ("app/test: vm power management")
> Cc: alan.carew at intel.com
> Cc: stable at dpdk.org
>
> Signed-off-by: Richael Zhuang <richael.zhuang at arm.com>
> ---
> app/test/test_power_cpufreq.c | 16 +++++++++++++++-
> 1 file changed, 15 insertions(+), 1 deletion(-)
>
> diff --git a/app/test/test_power_cpufreq.c b/app/test/test_power_cpufreq.c
> index d47b3e0a1..f753d24ac 100644
> --- a/app/test/test_power_cpufreq.c
> +++ b/app/test/test_power_cpufreq.c
> @@ -35,6 +35,10 @@ test_power_caps(void)
> #define TEST_POWER_LCORE_INVALID ((unsigned)RTE_MAX_LCORE)
> #define TEST_POWER_FREQS_NUM_MAX ((unsigned)RTE_MAX_LCORE_FREQS)
>
> +/* macros used for rounding frequency to nearest 100000 */
> +#define TEST_FREQ_ROUNDING_DELTA 50000
> +#define TEST_ROUND_FREQ_TO_N_100000 100000
> +
> #define TEST_POWER_SYSFILE_CUR_FREQ \
> "/sys/devices/system/cpu/cpu%u/cpufreq/cpuinfo_cur_freq"
>
> @@ -67,7 +71,17 @@ check_cur_freq(unsigned lcore_id, uint32_t idx)
> goto fail_all;
>
> cur_freq = strtoul(buf, NULL, TEST_POWER_CONVERT_TO_DECIMAL);
> - ret = (freqs[idx] == cur_freq ? 0 : -1);
> +
> + /* convert the frequency to nearest 100000 value
> + * Ex: if cur_freq=1396789 then freq_conv=1400000
> + * Ex: if cur_freq=800030 then freq_conv=800000
> + */
> + unsigned int freq_conv = 0;
> + freq_conv = (cur_freq + TEST_FREQ_ROUNDING_DELTA)
> + / TEST_ROUND_FREQ_TO_N_100000;
> + freq_conv = freq_conv * TEST_ROUND_FREQ_TO_N_100000;
> +
> + ret = (freqs[idx] == freq_conv ? 0 : -1);
>
> if (ret == 0)
> break;
Hi Richael,
I played around with this, rounds nicely. I found an unrelated issue
around turbo mode that I need to look at but this patch looks fine.
Reviewed-by: David Hunt <david.hunt at intel.com>
More information about the dev
mailing list