diff mbox series

[V2,12/15] cpufreq: mediatek: Use maximum voltage in init stage

Message ID 20220408045908.21671-13-rex-bc.chen@mediatek.com (mailing list archive)
State Superseded, archived
Headers show
Series cpufreq: mediatek: Cleanup and support MT8183 and MT8186 | expand

Commit Message

Rex-BC Chen (陳柏辰) April 8, 2022, 4:59 a.m. UTC
From: Jia-Wei Chang <jia-wei.chang@mediatek.com>

Two or more clients may use the same regulator, and it could cause the
issue of high-freqeuncy-low-voltage.
To prevent this, we use maximum voltage in mtk_cpu_dvfs_info_init().

Signed-off-by: Jia-Wei Chang <jia-wei.chang@mediatek.com>
---
 drivers/cpufreq/mediatek-cpufreq.c | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

Comments

AngeloGioacchino Del Regno April 8, 2022, 1:37 p.m. UTC | #1
Il 08/04/22 06:59, Rex-BC Chen ha scritto:
> From: Jia-Wei Chang <jia-wei.chang@mediatek.com>
> 
> Two or more clients may use the same regulator, and it could cause the
> issue of high-freqeuncy-low-voltage.
> To prevent this, we use maximum voltage in mtk_cpu_dvfs_info_init().
> 
> Signed-off-by: Jia-Wei Chang <jia-wei.chang@mediatek.com>

Is this happening for proc-supply (proc_reg)?
...because it looks like it is, so you should send this commit separately
and with an appropriate Fixes: tag.
Rex-BC Chen (陳柏辰) April 12, 2022, 11:24 a.m. UTC | #2
On Fri, 2022-04-08 at 15:37 +0200, AngeloGioacchino Del Regno wrote:
> Il 08/04/22 06:59, Rex-BC Chen ha scritto:
> > From: Jia-Wei Chang <jia-wei.chang@mediatek.com>
> > 
> > Two or more clients may use the same regulator, and it could cause
> > the
> > issue of high-freqeuncy-low-voltage.
> > To prevent this, we use maximum voltage in
> > mtk_cpu_dvfs_info_init().
> > 
> > Signed-off-by: Jia-Wei Chang <jia-wei.chang@mediatek.com>
> 
> Is this happening for proc-supply (proc_reg)?
> ...because it looks like it is, so you should send this commit
> separately
> and with an appropriate Fixes: tag.

Hello Angelo,

The commit message is not clear.
The clients are cpufreq and mediatek cci.
I think there is no any error before introducing mediatek cci into
cpufreq.

The order of this patch should be after "cpufreq: mediatek: Link CCI
device to CPU".

If so, I will also correct the commit message in next version.

BRs,
Rex
Rex-BC Chen (陳柏辰) April 14, 2022, 3:40 a.m. UTC | #3
On Fri, 2022-04-08 at 15:37 +0200, AngeloGioacchino Del Regno wrote:
> Il 08/04/22 06:59, Rex-BC Chen ha scritto:
> > From: Jia-Wei Chang <jia-wei.chang@mediatek.com>
> > 
> > Two or more clients may use the same regulator, and it could cause
> > the
> > issue of high-freqeuncy-low-voltage.
> > To prevent this, we use maximum voltage in
> > mtk_cpu_dvfs_info_init().
> > 
> > Signed-off-by: Jia-Wei Chang <jia-wei.chang@mediatek.com>
> 
> Is this happening for proc-supply (proc_reg)?
> ...because it looks like it is, so you should send this commit
> separately
> and with an appropriate Fixes: tag.

Hello all,

If we use the value of max(booting voltage, target cpufreq voltage)
when cci is not ready, I think we don't need this patch to prevent high
frequenc low voltage issue.
(mentioned in [1])

I will drop this patch in next version.

Thanks.

[1]:https://patchwork.kernel.org/comment/24816091/

BRs,
Rex
diff mbox series

Patch

diff --git a/drivers/cpufreq/mediatek-cpufreq.c b/drivers/cpufreq/mediatek-cpufreq.c
index e69b16a6541e..b08ab7c14818 100644
--- a/drivers/cpufreq/mediatek-cpufreq.c
+++ b/drivers/cpufreq/mediatek-cpufreq.c
@@ -333,7 +333,7 @@  static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu)
 {
 	struct device *cpu_dev;
 	struct dev_pm_opp *opp;
-	unsigned long rate;
+	unsigned long rate, opp_volt;
 	int ret;
 
 	cpu_dev = get_cpu_device(cpu);
@@ -417,6 +417,24 @@  static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu)
 	info->intermediate_voltage = dev_pm_opp_get_voltage(opp);
 	dev_pm_opp_put(opp);
 
+	/* Use highest opp voltage in the init stage */
+	rate = U32_MAX;
+	opp = dev_pm_opp_find_freq_floor(info->cpu_dev, &rate);
+	if (IS_ERR(opp)) {
+		ret = PTR_ERR(opp);
+		dev_err(cpu_dev, "cpu%d: failed to get opp\n", info->opp_cpu);
+		goto out_disable_inter_clock;
+	}
+
+	opp_volt = dev_pm_opp_get_voltage(opp);
+	dev_pm_opp_put(opp);
+	ret = mtk_cpufreq_set_voltage(info, opp_volt);
+	if (ret) {
+		dev_err(cpu_dev, "cpu%d: failed to scale to highest voltage %lu in proc_reg\n",
+			info->opp_cpu, opp_volt);
+		goto out_disable_inter_clock;
+	}
+
 	info->opp_cpu = cpu;
 	info->opp_nb.notifier_call = mtk_cpufreq_opp_notifier;
 	ret = dev_pm_opp_register_notifier(cpu_dev, &info->opp_nb);