* Re: [PATCH] PM / OPP: Update performance state when freq == old_freq
2019-03-12 4:57 [PATCH] PM / OPP: Update performance state when freq == old_freq Viresh Kumar
@ 2019-03-12 6:27 ` Jorge Ramirez
2019-03-12 12:06 ` Niklas Cassel
2019-03-13 9:56 ` Rafael J. Wysocki
2 siblings, 0 replies; 4+ messages in thread
From: Jorge Ramirez @ 2019-03-12 6:27 UTC (permalink / raw)
To: Viresh Kumar, Rafael Wysocki, Viresh Kumar, Nishanth Menon, Stephen Boyd
Cc: linux-pm, Vincent Guittot, v5 . 0, Niklas Cassel, linux-kernel
On 3/12/19 05:57, Viresh Kumar wrote:
> At boot up, CPUfreq core performs a sanity check to see if the system is
> running at a frequency defined in the frequency table of the CPU. If so,
> we try to find a valid frequency (lowest frequency greater than the
> currently programmed frequency) from the table and set it. When the call
> reaches dev_pm_opp_set_rate(), it calls _find_freq_ceil(opp_table,
> &old_freq) to find the previously configured OPP and this call also
> updates the old_freq. This eventually sets the old_freq == freq (new
> target requested by cpufreq core) and we skip updating the performance
> state in this case.
>
> Fix this by also updating the performance state when the old_freq ==
> freq.
>
> Fixes: ca1b5d77b1c6 ("OPP: Configure all required OPPs")
> Cc: v5.0 <stable@vger.kernel.org> # v5.0
> Reported-by: Niklas Cassel <niklas.cassel@linaro.org>
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org>
> ---
> drivers/opp/core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/opp/core.c b/drivers/opp/core.c
> index d7f97167cac3..0420f7e8ad5b 100644
> --- a/drivers/opp/core.c
> +++ b/drivers/opp/core.c
> @@ -760,7 +760,7 @@ int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq)
> old_freq, freq);
>
> /* Scaling up? Configure required OPPs before frequency */
> - if (freq > old_freq) {
> + if (freq >= old_freq) {
> ret = _set_required_opps(dev, opp_table, opp);
> if (ret)
> goto put_opp;
>
[ 3.048579] Loading compiled-in X.509 certificates
[ 3.064282] cpufreq: cpufreq_online: CPU0: Running at unlisted freq:
1113600 KHz
[ 3.064510] cpufreq: cpufreq_online: CPU0: Unlisted initial frequency
changed to: 1248000 KHz
[ 3.072072] hctosys: unable to open rtc device (rtc0)
...
and the ftrace on the registered driver handling the performance change
to state 2
kworker/3:0-26 [003] .... 3.063824: cpr_pd_attach_dev:
reg[0x100] = 0x00000014
kworker/3:0-26 [003] .... 3.063845: cpr_corner_restore:
reg[0x100] = 0x00000014
kworker/3:0-26 [003] .... 3.063848: cpr_corner_restore: gcnt
= 0x000132ab, ctl = 0x31000428, irq = 0x00000014
kworker/3:0-26 [003] .... 3.063851: cpr_ctl_enable.part.15:
reg[0x100] = 0x00000014
kworker/3:0-26 [003] .... 3.064352: cpr_set_performance:
cpr_set_performance: setting performance state: 2 (prev state: 0)
kworker/3:0-26 [003] .... 3.064355: cpr_ctl_disable.part.16:
reg[0x100] = 0x00000000
kworker/3:0-26 [003] .... 3.064410: cpr_corner_restore:
reg[0x100] = 0x00000014
kworker/3:0-26 [003] .... 3.064413: cpr_corner_restore: gcnt
= 0x000132f5, ctl = 0x31000428, irq = 0x00000014
kworker/3:0-26 [003] .... 3.064416: cpr_ctl_enable.part.15:
reg[0x100] = 0x00000014
irq/29-cpr-1320 [000] .... 3.079189: cpr_irq_handler:
IRQ_STATUS = 0x5
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH] PM / OPP: Update performance state when freq == old_freq
2019-03-12 4:57 [PATCH] PM / OPP: Update performance state when freq == old_freq Viresh Kumar
2019-03-12 6:27 ` Jorge Ramirez
@ 2019-03-12 12:06 ` Niklas Cassel
2019-03-13 9:56 ` Rafael J. Wysocki
2 siblings, 0 replies; 4+ messages in thread
From: Niklas Cassel @ 2019-03-12 12:06 UTC (permalink / raw)
To: Viresh Kumar
Cc: Rafael Wysocki, Viresh Kumar, Nishanth Menon, Stephen Boyd,
linux-pm, Vincent Guittot, jorge.ramirez-ortiz, v5 . 0,
linux-kernel
On Tue, Mar 12, 2019 at 10:27:18AM +0530, Viresh Kumar wrote:
> At boot up, CPUfreq core performs a sanity check to see if the system is
> running at a frequency defined in the frequency table of the CPU. If so,
> we try to find a valid frequency (lowest frequency greater than the
> currently programmed frequency) from the table and set it. When the call
> reaches dev_pm_opp_set_rate(), it calls _find_freq_ceil(opp_table,
> &old_freq) to find the previously configured OPP and this call also
> updates the old_freq. This eventually sets the old_freq == freq (new
> target requested by cpufreq core) and we skip updating the performance
> state in this case.
>
> Fix this by also updating the performance state when the old_freq ==
> freq.
>
> Fixes: ca1b5d77b1c6 ("OPP: Configure all required OPPs")
> Cc: v5.0 <stable@vger.kernel.org> # v5.0
> Reported-by: Niklas Cassel <niklas.cassel@linaro.org>
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> ---
> drivers/opp/core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/opp/core.c b/drivers/opp/core.c
> index d7f97167cac3..0420f7e8ad5b 100644
> --- a/drivers/opp/core.c
> +++ b/drivers/opp/core.c
> @@ -760,7 +760,7 @@ int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq)
> old_freq, freq);
>
> /* Scaling up? Configure required OPPs before frequency */
> - if (freq > old_freq) {
> + if (freq >= old_freq) {
> ret = _set_required_opps(dev, opp_table, opp);
> if (ret)
> goto put_opp;
> --
> 2.21.0.rc0.269.g1a574e7a288b
>
Tested-by: Niklas Cassel <niklas.cassel@linaro.org>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH] PM / OPP: Update performance state when freq == old_freq
2019-03-12 4:57 [PATCH] PM / OPP: Update performance state when freq == old_freq Viresh Kumar
2019-03-12 6:27 ` Jorge Ramirez
2019-03-12 12:06 ` Niklas Cassel
@ 2019-03-13 9:56 ` Rafael J. Wysocki
2 siblings, 0 replies; 4+ messages in thread
From: Rafael J. Wysocki @ 2019-03-13 9:56 UTC (permalink / raw)
To: Viresh Kumar
Cc: Viresh Kumar, Nishanth Menon, Stephen Boyd, linux-pm,
Vincent Guittot, jorge.ramirez-ortiz, v5 . 0, Niklas Cassel,
linux-kernel
On Tuesday, March 12, 2019 5:57:18 AM CET Viresh Kumar wrote:
> At boot up, CPUfreq core performs a sanity check to see if the system is
> running at a frequency defined in the frequency table of the CPU. If so,
> we try to find a valid frequency (lowest frequency greater than the
> currently programmed frequency) from the table and set it. When the call
> reaches dev_pm_opp_set_rate(), it calls _find_freq_ceil(opp_table,
> &old_freq) to find the previously configured OPP and this call also
> updates the old_freq. This eventually sets the old_freq == freq (new
> target requested by cpufreq core) and we skip updating the performance
> state in this case.
>
> Fix this by also updating the performance state when the old_freq ==
> freq.
>
> Fixes: ca1b5d77b1c6 ("OPP: Configure all required OPPs")
> Cc: v5.0 <stable@vger.kernel.org> # v5.0
> Reported-by: Niklas Cassel <niklas.cassel@linaro.org>
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> ---
> drivers/opp/core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/opp/core.c b/drivers/opp/core.c
> index d7f97167cac3..0420f7e8ad5b 100644
> --- a/drivers/opp/core.c
> +++ b/drivers/opp/core.c
> @@ -760,7 +760,7 @@ int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq)
> old_freq, freq);
>
> /* Scaling up? Configure required OPPs before frequency */
> - if (freq > old_freq) {
> + if (freq >= old_freq) {
> ret = _set_required_opps(dev, opp_table, opp);
> if (ret)
> goto put_opp;
>
Applied, thanks!
^ permalink raw reply [flat|nested] 4+ messages in thread