[PATCH 21.02] ipq806x: backport cpufreq changes to 5.4
Shane Synan
digitalcircuit36939 at gmail.com
Wed Jun 16 22:33:04 PDT 2021
Date: Thu, 8 Apr 2021 16:20:36 +0200
From: Ansuel Smith <ansuelsmth at gmail.com>
The new cpufreq driver requires different dts bindings.
Backport the new driver to kernel 5.4
Signed-off-by: Ansuel Smith <ansuelsmth at gmail.com>
Tested-by: Shane Synan <digitalcircuit36939 at gmail.com>
(cherry picked from commit 6e411b8416388a9c8be1b2291be9b5adeeb07784)
---
I've tested these changes on the ZyXEL NBG6817 and they've resolved a
5.4 kernel crash for me, triggered by SFTP uploads via OpenSSH server
to an attached USB 3.0 drive, as documented in FS#3099.
https://bugs.openwrt.org/index.php?do=details&task_id=3099#comment9820
However, I'm not certain this resolves the crashes for others, hence
not mentioning a "Fixes" for the issue in the commit directly.
This is my first attempt at my first submission to the openwrt-devel
mailing list, let me know if anything's amiss! I'm unsure if
attribution is properly credited to the original patch author.
target/linux/ipq806x/config-5.4 | 1 +
...rt-adjusting-OPP-voltages-at-runtime.patch | 153 ----
...per-to-get-an-opp-regulator-for-devi.patch | 52 --
...e-voltage-tolerance-when-adjusting-t.patch | 47 --
...-dt-Handle-OPP-voltage-adjust-events.patch | 118 ---
...-dt-Add-L2-frequency-scaling-support.patch | 199 -----
...056-cpufreq-dt-Add-missing-rcu-locks.patch | 23 -
...qcom-cpufreq-nvmem-support-specific-.patch | 51 ++
...q-add-Krait-dedicated-scaling-driver.patch | 679 ++++++++++++++++++
...ufreq-add-qcom-krait-cpufreq-binding.patch | 237 ++++++
...dd-fab-scaling-support-with-cpufreq.patch} | 34 +-
11 files changed, 985 insertions(+), 609 deletions(-)
delete mode 100644 target/linux/ipq806x/patches-5.4/0049-PM-OPP-Support-adjusting-OPP-voltages-at-runtime.patch
delete mode 100644 target/linux/ipq806x/patches-5.4/0051-PM-OPP-Add-a-helper-to-get-an-opp-regulator-for-devi.patch
delete mode 100644 target/linux/ipq806x/patches-5.4/0052-PM-OPP-Update-the-voltage-tolerance-when-adjusting-t.patch
delete mode 100644 target/linux/ipq806x/patches-5.4/0054-cpufreq-dt-Handle-OPP-voltage-adjust-events.patch
delete mode 100644 target/linux/ipq806x/patches-5.4/0055-cpufreq-dt-Add-L2-frequency-scaling-support.patch
delete mode 100644 target/linux/ipq806x/patches-5.4/0056-cpufreq-dt-Add-missing-rcu-locks.patch
create mode 100644 target/linux/ipq806x/patches-5.4/093-drivers-cpufreq-qcom-cpufreq-nvmem-support-specific-.patch
create mode 100644 target/linux/ipq806x/patches-5.4/098-1-cpufreq-add-Krait-dedicated-scaling-driver.patch
create mode 100644 target/linux/ipq806x/patches-5.4/098-2-Documentation-cpufreq-add-qcom-krait-cpufreq-binding.patch
rename target/linux/ipq806x/patches-5.4/{0057-add-fab-scaling-support-with-cpufreq.patch => 098-3-add-fab-scaling-support-with-cpufreq.patch} (93%)
diff --git a/target/linux/ipq806x/config-5.4 b/target/linux/ipq806x/config-5.4
index 68ed6ec0c7..d410cd3829 100644
--- a/target/linux/ipq806x/config-5.4
+++ b/target/linux/ipq806x/config-5.4
@@ -62,6 +62,7 @@ CONFIG_ARM_MODULE_PLTS=y
CONFIG_ARM_PATCH_IDIV=y
CONFIG_ARM_PATCH_PHYS_VIRT=y
# CONFIG_ARM_QCOM_CPUFREQ_HW is not set
+CONFIG_ARM_QCOM_CPUFREQ_KRAIT=y
CONFIG_ARM_QCOM_CPUFREQ_NVMEM=y
CONFIG_ARM_QCOM_CPUIDLE=y
# CONFIG_ARM_SMMU is not set
diff --git a/target/linux/ipq806x/patches-5.4/0049-PM-OPP-Support-adjusting-OPP-voltages-at-runtime.patch b/target/linux/ipq806x/patches-5.4/0049-PM-OPP-Support-adjusting-OPP-voltages-at-runtime.patch
deleted file mode 100644
index 9efbd583b4..0000000000
--- a/target/linux/ipq806x/patches-5.4/0049-PM-OPP-Support-adjusting-OPP-voltages-at-runtime.patch
+++ /dev/null
@@ -1,153 +0,0 @@
-From: Sylwester Nawrocki <s.nawrocki at samsung.com>
-To: krzk at kernel.org, vireshk at kernel.org, robh+dt at kernel.org
-Cc: sboyd at kernel.org, roger.lu at mediatek.com,
- linux-pm at vger.kernel.org, linux-arm-kernel at lists.infradead.org,
- linux-samsung-soc at vger.kernel.org, devicetree at vger.kernel.org,
- b.zolnierkie at samsung.com, m.szyprowski at samsung.com,
- Stephen Boyd <sboyd at codeaurora.org>,
- Sylwester Nawrocki <s.nawrocki at samsung.com>
-Subject: [PATCH v5 1/4] PM / OPP: Support adjusting OPP voltages at runtime
-Date: Wed, 16 Oct 2019 16:57:53 +0200
-Message-ID: <20191016145756.16004-2-s.nawrocki at samsung.com> (raw)
-In-Reply-To: <20191016145756.16004-1-s.nawrocki at samsung.com>
-
-From: Stephen Boyd <sboyd at codeaurora.org>
-
-On some SoCs the Adaptive Voltage Scaling (AVS) technique is
-employed to optimize the operating voltage of a device. At a
-given frequency, the hardware monitors dynamic factors and either
-makes a suggestion for how much to adjust a voltage for the
-current frequency, or it automatically adjusts the voltage
-without software intervention. Add an API to the OPP library for
-the former case, so that AVS type devices can update the voltages
-for an OPP when the hardware determines the voltage should
-change. The assumption is that drivers like CPUfreq or devfreq
-will register for the OPP notifiers and adjust the voltage
-according to suggestions that AVS makes.
-
-This patch is derived from [1] submitted by Stephen.
-[1] https://lore.kernel.org/patchwork/patch/599279/
-
-Signed-off-by: Stephen Boyd <sboyd at codeaurora.org>
-Signed-off-by: Roger Lu <roger.lu at mediatek.com>
-[s.nawrocki at samsung.com: added handling of OPP min/max voltage]
-Signed-off-by: Sylwester Nawrocki <s.nawrocki at samsung.com>
----
- drivers/opp/core.c | 69 ++++++++++++++++++++++++++++++++++++++++++
- include/linux/pm_opp.h | 13 ++++++++
- 2 files changed, 82 insertions(+)
-
---- a/drivers/opp/core.c
-+++ b/drivers/opp/core.c
-@@ -2102,6 +2102,75 @@ put_table:
- }
-
- /**
-+ * dev_pm_opp_adjust_voltage() - helper to change the voltage of an OPP
-+ * @dev: device for which we do this operation
-+ * @freq: OPP frequency to adjust voltage of
-+ * @u_volt: new OPP target voltage
-+ * @u_volt_min: new OPP min voltage
-+ * @u_volt_max: new OPP max voltage
-+ *
-+ * Return: -EINVAL for bad pointers, -ENOMEM if no memory available for the
-+ * copy operation, returns 0 if no modifcation was done OR modification was
-+ * successful.
-+ */
-+int dev_pm_opp_adjust_voltage(struct device *dev, unsigned long freq,
-+ unsigned long u_volt, unsigned long u_volt_min,
-+ unsigned long u_volt_max)
-+
-+{
-+ struct opp_table *opp_table;
-+ struct dev_pm_opp *tmp_opp, *opp = ERR_PTR(-ENODEV);
-+ int r = 0;
-+
-+ /* Find the opp_table */
-+ opp_table = _find_opp_table(dev);
-+ if (IS_ERR(opp_table)) {
-+ r = PTR_ERR(opp_table);
-+ dev_warn(dev, "%s: Device OPP not found (%d)\n", __func__, r);
-+ return r;
-+ }
-+
-+ mutex_lock(&opp_table->lock);
-+
-+ /* Do we have the frequency? */
-+ list_for_each_entry(tmp_opp, &opp_table->opp_list, node) {
-+ if (tmp_opp->rate == freq) {
-+ opp = tmp_opp;
-+ break;
-+ }
-+ }
-+
-+ if (IS_ERR(opp)) {
-+ r = PTR_ERR(opp);
-+ goto adjust_unlock;
-+ }
-+
-+ /* Is update really needed? */
-+ if (opp->supplies->u_volt == u_volt)
-+ goto adjust_unlock;
-+
-+ opp->supplies->u_volt = u_volt;
-+ opp->supplies->u_volt_min = u_volt_min;
-+ opp->supplies->u_volt_max = u_volt_max;
-+
-+ dev_pm_opp_get(opp);
-+ mutex_unlock(&opp_table->lock);
-+
-+ /* Notify the voltage change of the OPP */
-+ blocking_notifier_call_chain(&opp_table->head, OPP_EVENT_ADJUST_VOLTAGE,
-+ opp);
-+
-+ dev_pm_opp_put(opp);
-+ goto adjust_put_table;
-+
-+adjust_unlock:
-+ mutex_unlock(&opp_table->lock);
-+adjust_put_table:
-+ dev_pm_opp_put_opp_table(opp_table);
-+ return r;
-+}
-+
-+/**
- * dev_pm_opp_enable() - Enable a specific OPP
- * @dev: device for which we do this operation
- * @freq: OPP frequency to enable
---- a/include/linux/pm_opp.h
-+++ b/include/linux/pm_opp.h
-@@ -22,6 +22,7 @@ struct opp_table;
-
- enum dev_pm_opp_event {
- OPP_EVENT_ADD, OPP_EVENT_REMOVE, OPP_EVENT_ENABLE, OPP_EVENT_DISABLE,
-+ OPP_EVENT_ADJUST_VOLTAGE,
- };
-
- /**
-@@ -113,6 +114,10 @@ int dev_pm_opp_add(struct device *dev, u
- void dev_pm_opp_remove(struct device *dev, unsigned long freq);
- void dev_pm_opp_remove_all_dynamic(struct device *dev);
-
-+int dev_pm_opp_adjust_voltage(struct device *dev, unsigned long freq,
-+ unsigned long u_volt, unsigned long u_volt_min,
-+ unsigned long u_volt_max);
-+
- int dev_pm_opp_enable(struct device *dev, unsigned long freq);
-
- int dev_pm_opp_disable(struct device *dev, unsigned long freq);
-@@ -242,6 +247,14 @@ static inline void dev_pm_opp_remove_all
- {
- }
-
-+static inline int
-+dev_pm_opp_adjust_voltage(struct device *dev, unsigned long freq,
-+ unsigned long u_volt, unsigned long u_volt_min,
-+ unsigned long u_volt_max)
-+{
-+ return 0;
-+}
-+
- static inline int dev_pm_opp_enable(struct device *dev, unsigned long freq)
- {
- return 0;
diff --git a/target/linux/ipq806x/patches-5.4/0051-PM-OPP-Add-a-helper-to-get-an-opp-regulator-for-devi.patch b/target/linux/ipq806x/patches-5.4/0051-PM-OPP-Add-a-helper-to-get-an-opp-regulator-for-devi.patch
deleted file mode 100644
index 35fe45fca7..0000000000
--- a/target/linux/ipq806x/patches-5.4/0051-PM-OPP-Add-a-helper-to-get-an-opp-regulator-for-devi.patch
+++ /dev/null
@@ -1,52 +0,0 @@
-From d06ca5e7a3cf726f5be5ffd96e93ccd798b8c09a Mon Sep 17 00:00:00 2001
-From: Georgi Djakov <georgi.djakov at linaro.org>
-Date: Thu, 12 May 2016 14:41:33 +0300
-Subject: [PATCH 51/69] PM / OPP: Add a helper to get an opp regulator for
- device
-
-Signed-off-by: Georgi Djakov <georgi.djakov at linaro.org>
----
- drivers/opp/core.c | 21 +++++++++++++++++++++
- include/linux/pm_opp.h | 1 +
- 2 files changed, 22 insertions(+)
-
---- a/drivers/opp/core.c
-+++ b/drivers/opp/core.c
-@@ -127,6 +127,27 @@ unsigned long dev_pm_opp_get_freq(struct
- }
- EXPORT_SYMBOL_GPL(dev_pm_opp_get_freq);
-
-+struct regulator *dev_pm_opp_get_regulator(struct device *dev)
-+{
-+ struct opp_table *opp_table;
-+ struct regulator *reg;
-+
-+ rcu_read_lock();
-+
-+ opp_table = _find_opp_table(dev);
-+ if (IS_ERR(opp_table)) {
-+ rcu_read_unlock();
-+ return ERR_CAST(opp_table);
-+ }
-+
-+ reg = opp_table->regulators[0];
-+
-+ rcu_read_unlock();
-+
-+ return reg;
-+}
-+EXPORT_SYMBOL_GPL(dev_pm_opp_get_regulator);
-+
- /**
- * dev_pm_opp_get_level() - Gets the level corresponding to an available opp
- * @opp: opp for which level value has to be returned for
---- a/include/linux/pm_opp.h
-+++ b/include/linux/pm_opp.h
-@@ -83,6 +83,7 @@ void dev_pm_opp_put_opp_table(struct opp
- unsigned long dev_pm_opp_get_voltage(struct dev_pm_opp *opp);
-
- unsigned long dev_pm_opp_get_freq(struct dev_pm_opp *opp);
-+struct regulator *dev_pm_opp_get_regulator(struct device *dev);
-
- unsigned int dev_pm_opp_get_level(struct dev_pm_opp *opp);
-
diff --git a/target/linux/ipq806x/patches-5.4/0052-PM-OPP-Update-the-voltage-tolerance-when-adjusting-t.patch b/target/linux/ipq806x/patches-5.4/0052-PM-OPP-Update-the-voltage-tolerance-when-adjusting-t.patch
deleted file mode 100644
index 8498a0b6df..0000000000
--- a/target/linux/ipq806x/patches-5.4/0052-PM-OPP-Update-the-voltage-tolerance-when-adjusting-t.patch
+++ /dev/null
@@ -1,47 +0,0 @@
-From 4533c285c2aedce6d4434d7b877066de3b1ecb33 Mon Sep 17 00:00:00 2001
-From: Georgi Djakov <georgi.djakov at linaro.org>
-Date: Thu, 25 Aug 2016 18:43:35 +0300
-Subject: [PATCH 52/69] PM / OPP: Update the voltage tolerance when adjusting
- the OPP
-
-When the voltage is adjusted, the voltage tolerance is not updated.
-This can lead to situations where the voltage min value is greater
-than the voltage max value. The final result is triggering a BUG()
-in the regulator core.
-Fix this by updating the voltage tolerance values too.
-
-Signed-off-by: Georgi Djakov <georgi.djakov at linaro.org>
----
- drivers/opp/core.c | 5 +++++
- 1 file changed, 5 insertions(+)
-
---- a/drivers/opp/core.c
-+++ b/drivers/opp/core.c
-@@ -2142,6 +2142,7 @@ int dev_pm_opp_adjust_voltage(struct dev
- struct opp_table *opp_table;
- struct dev_pm_opp *tmp_opp, *opp = ERR_PTR(-ENODEV);
- int r = 0;
-+ unsigned long tol;
-
- /* Find the opp_table */
- opp_table = _find_opp_table(dev);
-@@ -2171,8 +2172,17 @@ int dev_pm_opp_adjust_voltage(struct dev
- goto adjust_unlock;
-
- opp->supplies->u_volt = u_volt;
-- opp->supplies->u_volt_min = u_volt_min;
-- opp->supplies->u_volt_max = u_volt_max;
-+
-+ tol = u_volt * opp_table->voltage_tolerance_v1 / 100;
-+ if ( u_volt_min == u_volt )
-+ opp->supplies->u_volt_min = u_volt - tol;
-+ else
-+ opp->supplies->u_volt_min = u_volt_min;
-+
-+ if ( u_volt_max == u_volt )
-+ opp->supplies->u_volt_max = u_volt + tol;
-+ else
-+ opp->supplies->u_volt_max = u_volt_max;
-
- dev_pm_opp_get(opp);
- mutex_unlock(&opp_table->lock);
diff --git a/target/linux/ipq806x/patches-5.4/0054-cpufreq-dt-Handle-OPP-voltage-adjust-events.patch b/target/linux/ipq806x/patches-5.4/0054-cpufreq-dt-Handle-OPP-voltage-adjust-events.patch
deleted file mode 100644
index 2e9c101d94..0000000000
--- a/target/linux/ipq806x/patches-5.4/0054-cpufreq-dt-Handle-OPP-voltage-adjust-events.patch
+++ /dev/null
@@ -1,118 +0,0 @@
-From 10577f74c35bd395951d1b2382c8d821089b5745 Mon Sep 17 00:00:00 2001
-From: Stephen Boyd <sboyd at codeaurora.org>
-Date: Fri, 18 Sep 2015 17:52:08 -0700
-Subject: [PATCH 54/69] cpufreq-dt: Handle OPP voltage adjust events
-
-On some SoCs the Adaptive Voltage Scaling (AVS) technique is
-employed to optimize the operating voltage of a device. At a
-given frequency, the hardware monitors dynamic factors and either
-makes a suggestion for how much to adjust a voltage for the
-current frequency, or it automatically adjusts the voltage
-without software intervention.
-
-In the former case, an AVS driver will call
-dev_pm_opp_modify_voltage() and update the voltage for the
-particular OPP the CPUs are using. Add an OPP notifier to
-cpufreq-dt so that we can adjust the voltage of the CPU when AVS
-updates the OPP.
-
-Signed-off-by: Stephen Boyd <sboyd at codeaurora.org>
-Acked-by: Viresh Kumar <viresh.kumar at linaro.org>
-Signed-off-by: Georgi Djakov <georgi.djakov at linaro.org>
----
- drivers/cpufreq/cpufreq-dt.c | 68 ++++++++++++++++++++++++++++++++++++++++++--
- 1 file changed, 65 insertions(+), 3 deletions(-)
-
---- a/drivers/cpufreq/cpufreq-dt.c
-+++ b/drivers/cpufreq/cpufreq-dt.c
-@@ -27,6 +27,9 @@ struct private_data {
- struct opp_table *opp_table;
- struct device *cpu_dev;
- const char *reg_name;
-+ struct notifier_block opp_nb;
-+ struct mutex lock;
-+ unsigned long opp_freq;
- bool have_static_opps;
- };
-
-@@ -42,12 +45,15 @@ static int set_target(struct cpufreq_pol
- unsigned long freq = policy->freq_table[index].frequency;
- int ret;
-
-+ mutex_lock(&priv->lock);
- ret = dev_pm_opp_set_rate(priv->cpu_dev, freq * 1000);
-
- if (!ret) {
-+ priv->opp_freq = freq * 1000;
- arch_set_freq_scale(policy->related_cpus, freq,
- policy->cpuinfo.max_freq);
- }
-+ mutex_unlock(&priv->lock);
-
- return ret;
- }
-@@ -90,6 +96,39 @@ node_put:
- return name;
- }
-
-+static int opp_notifier(struct notifier_block *nb, unsigned long event,
-+ void *data)
-+{
-+ struct dev_pm_opp *opp = data;
-+ struct private_data *priv = container_of(nb, struct private_data,
-+ opp_nb);
-+ struct device *cpu_dev = priv->cpu_dev;
-+ struct regulator *cpu_reg;
-+ unsigned long volt, freq;
-+ int ret = 0;
-+
-+ if (event == OPP_EVENT_ADJUST_VOLTAGE) {
-+ cpu_reg = dev_pm_opp_get_regulator(cpu_dev);
-+ if (IS_ERR(cpu_reg)) {
-+ ret = PTR_ERR(cpu_reg);
-+ goto out;
-+ }
-+ volt = dev_pm_opp_get_voltage(opp);
-+ freq = dev_pm_opp_get_freq(opp);
-+
-+ mutex_lock(&priv->lock);
-+ if (freq == priv->opp_freq) {
-+ ret = regulator_set_voltage_triplet(cpu_reg, volt, volt, volt);
-+ }
-+ mutex_unlock(&priv->lock);
-+ if (ret)
-+ dev_err(cpu_dev, "failed to scale voltage: %d\n", ret);
-+ }
-+
-+out:
-+ return notifier_from_errno(ret);
-+}
-+
- static int resources_available(void)
- {
- struct device *cpu_dev;
-@@ -246,10 +285,14 @@ static int cpufreq_init(struct cpufreq_p
- __func__, ret);
- }
-
-+ mutex_init(&priv->lock);
-+ priv->opp_nb.notifier_call = opp_notifier;
-+ dev_pm_opp_register_notifier(cpu_dev, &priv->opp_nb);
-+
- ret = dev_pm_opp_init_cpufreq_table(cpu_dev, &freq_table);
- if (ret) {
- dev_err(cpu_dev, "failed to init cpufreq table: %d\n", ret);
-- goto out_free_opp;
-+ goto out_unregister_nb;
- }
-
- priv->cpu_dev = cpu_dev;
-@@ -281,6 +324,8 @@ static int cpufreq_init(struct cpufreq_p
-
- out_free_cpufreq_table:
- dev_pm_opp_free_cpufreq_table(cpu_dev, &freq_table);
-+out_unregister_nb:
-+ dev_pm_opp_unregister_notifier(cpu_dev, &priv->opp_nb);
- out_free_opp:
- if (priv->have_static_opps)
- dev_pm_opp_of_cpumask_remove_table(policy->cpus);
diff --git a/target/linux/ipq806x/patches-5.4/0055-cpufreq-dt-Add-L2-frequency-scaling-support.patch b/target/linux/ipq806x/patches-5.4/0055-cpufreq-dt-Add-L2-frequency-scaling-support.patch
deleted file mode 100644
index 910c2b1722..0000000000
--- a/target/linux/ipq806x/patches-5.4/0055-cpufreq-dt-Add-L2-frequency-scaling-support.patch
+++ /dev/null
@@ -1,199 +0,0 @@
---- a/drivers/cpufreq/cpufreq-dt.c
-+++ b/drivers/cpufreq/cpufreq-dt.c
-@@ -39,20 +39,85 @@ static struct freq_attr *cpufreq_dt_attr
- NULL,
- };
-
-+struct shared_data {
-+ unsigned long *curr_freq_table;
-+ unsigned long curr_l2_freq;
-+ unsigned long curr_l2_volt;
-+};
-+
-+static struct shared_data *cpus_shared_data;
-+
- static int set_target(struct cpufreq_policy *policy, unsigned int index)
- {
- struct private_data *priv = policy->driver_data;
- unsigned long freq = policy->freq_table[index].frequency;
-+ struct clk *l2_clk = policy->l2_clk;
-+ struct regulator *l2_regulator = policy->l2_regulator;
-+ unsigned long l2_freq, target_l2_freq;
-+ unsigned long l2_vol, target_l2_volt;
-+ unsigned long target_freq;
- int ret;
-
- mutex_lock(&priv->lock);
- ret = dev_pm_opp_set_rate(priv->cpu_dev, freq * 1000);
-
- if (!ret) {
-+ if (policy->l2_rate_set) {
-+ int cpu, l2_index, tol = 0;
-+ target_freq = freq * 1000;
-+
-+ cpus_shared_data->curr_freq_table[policy->cpu] = target_freq;
-+ for_each_present_cpu(cpu)
-+ if (cpu != policy->cpu)
-+ target_freq = max(target_freq, cpus_shared_data->curr_freq_table[cpu]);
-+
-+ for (l2_index = 2; l2_index > 0; l2_index--)
-+ if (target_freq >= policy->l2_cpufreq[l2_index])
-+ break;
-+
-+ l2_freq = cpus_shared_data->curr_l2_freq;
-+ target_l2_freq = policy->l2_rate[l2_index];
-+
-+ if (l2_freq != target_l2_freq) {
-+
-+ /*
-+ * Set to idle bin if switching from normal to high bin
-+ * or vice versa
-+ */
-+ if ( (l2_index == 2 && l2_freq == policy->l2_rate[1]) ||
-+ (l2_index == 1 && l2_freq == policy->l2_rate[2]) ) {
-+ ret = clk_set_rate(l2_clk, policy->l2_rate[0]);
-+ if (ret)
-+ goto exit;
-+ }
-+
-+ /* scale l2 with the core */
-+ ret = clk_set_rate(l2_clk, target_l2_freq);
-+ if (ret)
-+ goto exit;
-+ cpus_shared_data->curr_l2_freq = target_l2_freq;
-+
-+ if (policy->l2_volt_set) {
-+
-+ l2_vol = cpus_shared_data->curr_l2_volt;
-+ target_l2_volt = policy->l2_volt[l2_index];
-+
-+ if (l2_vol != target_l2_volt) {
-+ tol = target_l2_volt * policy->l2_volt_tol / 100;
-+ ret = regulator_set_voltage_tol(l2_regulator,target_l2_volt,tol);
-+ if (ret)
-+ goto exit;
-+ cpus_shared_data->curr_l2_volt = target_l2_volt;
-+ }
-+ }
-+ }
-+ }
- priv->opp_freq = freq * 1000;
- arch_set_freq_scale(policy->related_cpus, freq,
- policy->cpuinfo.max_freq);
- }
-+
-+exit:
- mutex_unlock(&priv->lock);
-
- return ret;
-@@ -195,6 +260,9 @@ static int cpufreq_init(struct cpufreq_p
- bool fallback = false;
- const char *name;
- int ret;
-+ struct device_node *np, *l2_np;
-+ struct clk *l2_clk = NULL;
-+ struct regulator *l2_regulator = NULL;
-
- cpu_dev = get_cpu_device(policy->cpu);
- if (!cpu_dev) {
-@@ -302,6 +370,57 @@ static int cpufreq_init(struct cpufreq_p
-
- policy->suspend_freq = dev_pm_opp_get_suspend_opp_freq(cpu_dev) / 1000;
-
-+ policy->l2_rate_set = false;
-+ policy->l2_volt_set = false;
-+
-+ l2_clk = clk_get(cpu_dev, "l2");
-+ if (!IS_ERR(l2_clk))
-+ policy->l2_clk = l2_clk;
-+
-+ l2_np = of_find_node_by_name(NULL, "qcom,l2");
-+ if (l2_np) {
-+ struct device_node *vdd;
-+ np = of_node_get(priv->cpu_dev->of_node);
-+
-+ if (np)
-+ of_property_read_u32(np, "voltage-tolerance", &policy->l2_volt_tol);
-+
-+ of_property_read_u32_array(l2_np, "qcom,l2-rates", policy->l2_rate, 3);
-+ if (policy->l2_rate[0] && policy->l2_rate[1] && policy->l2_rate[2]) {
-+ policy->l2_rate_set = true;
-+ of_property_read_u32_array(l2_np, "qcom,l2-cpufreq", policy->l2_cpufreq, 3);
-+ of_property_read_u32_array(l2_np, "qcom,l2-volt", policy->l2_volt, 3);
-+ } else
-+ pr_warn("L2: failed to parse L2 rates\n");
-+
-+ if (!policy->l2_cpufreq[0] && !policy->l2_cpufreq[1] &&
-+ !policy->l2_cpufreq[2] && policy->l2_rate_set) {
-+ int i;
-+
-+ pr_warn("L2: failed to parse target cpu freq, using defaults\n");
-+ for (i = 0; i < 3; i++)
-+ policy->l2_cpufreq[i] = policy->l2_rate[i];
-+ }
-+
-+ if (policy->l2_volt[0] && policy->l2_volt[1] && policy->l2_volt[2] &&
-+ policy->l2_volt_tol && policy->l2_rate_set) {
-+ vdd = of_parse_phandle(l2_np, "qcom,l2-supply", 0);
-+
-+ if (vdd) {
-+ l2_regulator = devm_regulator_get(cpu_dev, vdd->name);
-+ if (!IS_ERR(l2_regulator)) {
-+ policy->l2_regulator = l2_regulator;
-+ policy->l2_volt_set = true;
-+ } else {
-+ pr_warn("failed to get l2 supply\n");
-+ l2_regulator = NULL;
-+ }
-+
-+ of_node_put(vdd);
-+ }
-+ }
-+ }
-+
- /* Support turbo/boost mode */
- if (policy_has_boost_freq(policy)) {
- /* This gets disabled by core on driver unregister */
-@@ -401,6 +520,14 @@ static int dt_cpufreq_probe(struct platf
- if (ret)
- return ret;
-
-+ cpus_shared_data = kzalloc(sizeof(struct shared_data), GFP_KERNEL);
-+ if (!cpus_shared_data)
-+ return -ENOMEM;
-+
-+ cpus_shared_data->curr_freq_table = kzalloc(sizeof(int) * CONFIG_NR_CPUS, GFP_KERNEL);
-+ if (!cpus_shared_data->curr_freq_table)
-+ return -ENOMEM;
-+
- if (data) {
- if (data->have_governor_per_policy)
- dt_cpufreq_driver.flags |= CPUFREQ_HAVE_GOVERNOR_PER_POLICY;
-@@ -419,6 +546,8 @@ static int dt_cpufreq_probe(struct platf
-
- static int dt_cpufreq_remove(struct platform_device *pdev)
- {
-+ kfree(cpus_shared_data->curr_freq_table);
-+ kfree(cpus_shared_data);
- cpufreq_unregister_driver(&dt_cpufreq_driver);
- return 0;
- }
---- a/include/linux/cpufreq.h
-+++ b/include/linux/cpufreq.h
-@@ -58,7 +58,15 @@ struct cpufreq_policy {
- should set cpufreq */
- unsigned int cpu; /* cpu managing this policy, must be online */
-
-- struct clk *clk;
-+ struct clk *clk;
-+ struct clk *l2_clk; /* L2 clock */
-+ struct regulator *l2_regulator; /* L2 supply */
-+ unsigned int l2_rate[3]; /* L2 bus clock rate */
-+ bool l2_rate_set;
-+ unsigned int l2_cpufreq[3]; /* L2 target CPU frequency */
-+ unsigned int l2_volt[3]; /* L2 voltage array */
-+ bool l2_volt_set;
-+ unsigned int l2_volt_tol; /* L2 voltage tolerance */
- struct cpufreq_cpuinfo cpuinfo;/* see above */
-
- unsigned int min; /* in kHz */
diff --git a/target/linux/ipq806x/patches-5.4/0056-cpufreq-dt-Add-missing-rcu-locks.patch b/target/linux/ipq806x/patches-5.4/0056-cpufreq-dt-Add-missing-rcu-locks.patch
deleted file mode 100644
index 9b85839e19..0000000000
--- a/target/linux/ipq806x/patches-5.4/0056-cpufreq-dt-Add-missing-rcu-locks.patch
+++ /dev/null
@@ -1,23 +0,0 @@
-From 001a8dcb56ced58c518aaa10a4f0ba5e878705b6 Mon Sep 17 00:00:00 2001
-From: Georgi Djakov <georgi.djakov at linaro.org>
-Date: Tue, 17 May 2016 16:15:43 +0300
-Subject: [PATCH 56/69] cpufreq-dt: Add missing rcu locks
-
-Signed-off-by: Georgi Djakov <georgi.djakov at linaro.org>
----
- drivers/cpufreq/cpufreq-dt.c | 2 ++
- 1 file changed, 2 insertions(+)
-
---- a/drivers/cpufreq/cpufreq-dt.c
-+++ b/drivers/cpufreq/cpufreq-dt.c
-@@ -178,8 +178,10 @@ static int opp_notifier(struct notifier_
- ret = PTR_ERR(cpu_reg);
- goto out;
- }
-+ rcu_read_lock();
- volt = dev_pm_opp_get_voltage(opp);
- freq = dev_pm_opp_get_freq(opp);
-+ rcu_read_unlock();
-
- mutex_lock(&priv->lock);
- if (freq == priv->opp_freq) {
diff --git a/target/linux/ipq806x/patches-5.4/093-drivers-cpufreq-qcom-cpufreq-nvmem-support-specific-.patch b/target/linux/ipq806x/patches-5.4/093-drivers-cpufreq-qcom-cpufreq-nvmem-support-specific-.patch
new file mode 100644
index 0000000000..19c3d096c4
--- /dev/null
+++ b/target/linux/ipq806x/patches-5.4/093-drivers-cpufreq-qcom-cpufreq-nvmem-support-specific-.patch
@@ -0,0 +1,51 @@
+From a206d4061f1cc2c5cd17ee45c53a0ba711e48e6d Mon Sep 17 00:00:00 2001
+From: Ansuel Smith <ansuelsmth at gmail.com>
+Date: Sun, 7 Feb 2021 16:42:52 +0100
+Subject: [PATCH 3/3] drivers: cpufreq: qcom-cpufreq-nvmem: support specific
+ cpufreq driver
+
+Add support for specific cpufreq driver for qcom-cpufreq-nvmem driver.
+
+Signed-off-by: Ansuel Smith <ansuelsmth at gmail.com>
+---
+ drivers/cpufreq/qcom-cpufreq-nvmem.c | 15 +++++++++++++++
+ 1 file changed, 15 insertions(+)
+
+--- a/drivers/cpufreq/qcom-cpufreq-nvmem.c
++++ b/drivers/cpufreq/qcom-cpufreq-nvmem.c
+@@ -52,6 +52,7 @@ struct qcom_cpufreq_match_data {
+ char **pvs_name,
+ struct qcom_cpufreq_drv *drv);
+ const char **genpd_names;
++ const char *cpufreq_driver;
+ };
+
+ struct qcom_cpufreq_drv {
+@@ -250,6 +251,7 @@ static const struct qcom_cpufreq_match_d
+
+ static const struct qcom_cpufreq_match_data match_data_krait = {
+ .get_version = qcom_cpufreq_krait_name_version,
++ .cpufreq_driver = "krait-cpufreq",
+ };
+
+ static const char *qcs404_genpd_names[] = { "cpr", NULL };
+@@ -385,6 +387,19 @@ static int qcom_cpufreq_probe(struct pla
+ }
+ }
+
++ if (drv->data->cpufreq_driver) {
++ cpufreq_dt_pdev = platform_device_register_simple(
++ drv->data->cpufreq_driver, -1, NULL, 0);
++ if (!IS_ERR(cpufreq_dt_pdev)) {
++ platform_set_drvdata(pdev, drv);
++ return 0;
++ } else {
++ dev_err(cpu_dev,
++ "Failed to register dedicated %s cpufreq\n",
++ drv->data->cpufreq_driver);
++ }
++ }
++
+ cpufreq_dt_pdev = platform_device_register_simple("cpufreq-dt", -1,
+ NULL, 0);
+ if (!IS_ERR(cpufreq_dt_pdev)) {
diff --git a/target/linux/ipq806x/patches-5.4/098-1-cpufreq-add-Krait-dedicated-scaling-driver.patch b/target/linux/ipq806x/patches-5.4/098-1-cpufreq-add-Krait-dedicated-scaling-driver.patch
new file mode 100644
index 0000000000..eb9978154d
--- /dev/null
+++ b/target/linux/ipq806x/patches-5.4/098-1-cpufreq-add-Krait-dedicated-scaling-driver.patch
@@ -0,0 +1,679 @@
+From cc41a266280cad0b55319e614167c88dff344248 Mon Sep 17 00:00:00 2001
+From: Ansuel Smith <ansuelsmth at gmail.com>
+Date: Sat, 22 Feb 2020 16:33:10 +0100
+Subject: [PATCH 1/8] cpufreq: add Krait dedicated scaling driver
+
+This new driver is based on generic cpufreq-dt driver.
+Krait SoCs have 2-4 cpu and one shared L2 cache that can
+operate at different frequency based on the maximum cpu clk
+across all core.
+L2 frequency and voltage are scaled on every frequency change
+if needed. On Krait SoCs is present a bug that can cause
+transition problem between frequency bin, to workaround this
+on more than one transition, the L2 frequency is first set to the
+base rate and then to the target rate.
+The L2 frequency use the OPP framework and use the opp-level
+bindings to link the l2 freq to different cpu freq. This is needed
+as the Krait l2 clk are note mapped 1:1 to the core clks and some
+of the l2 clk is set based on a range of the cpu clks. If the driver
+find a broken config (for example no opp-level set) the l2 scaling is
+skipped.
+
+Signed-off-by: Ansuel Smith <ansuelsmth at gmail.com>
+---
+ drivers/cpufreq/Kconfig.arm | 14 +-
+ drivers/cpufreq/Makefile | 2 +
+ drivers/cpufreq/qcom-cpufreq-krait.c | 589 +++++++++++++++++++++++++++
+ 3 files changed, 604 insertions(+), 1 deletion(-)
+ create mode 100644 drivers/cpufreq/qcom-cpufreq-krait.c
+
+--- a/drivers/cpufreq/Kconfig.arm
++++ b/drivers/cpufreq/Kconfig.arm
+@@ -150,6 +150,18 @@ config ARM_QCOM_CPUFREQ_HW
+ The driver implements the cpufreq interface for this HW engine.
+ Say Y if you want to support CPUFreq HW.
+
++config ARM_QCOM_CPUFREQ_KRAIT
++ tristate "CPU Frequency scaling support for Krait SoCs"
++ depends on ARCH_QCOM || COMPILE_TEST
++ select PM_OPP
++ select ARM_QCOM_CPUFREQ_NVMEM
++ help
++ This adds the CPUFreq driver for Qualcomm Krait SoC based boards.
++ This scale the cache clk and regulator based on the different cpu
++ clks when scaling the different cores clk.
++
++ If in doubt, say N.
++
+ config ARM_RASPBERRYPI_CPUFREQ
+ tristate "Raspberry Pi cpufreq support"
+ depends on CLK_RASPBERRYPI || COMPILE_TEST
+@@ -339,4 +351,4 @@ config ARM_PXA2xx_CPUFREQ
+ help
+ This add the CPUFreq driver support for Intel PXA2xx SOCs.
+
+- If in doubt, say N.
++ If in doubt, say N.
+\ No newline at end of file
+--- a/drivers/cpufreq/Makefile
++++ b/drivers/cpufreq/Makefile
+@@ -63,6 +63,7 @@ obj-$(CONFIG_ARM_PXA2xx_CPUFREQ) += pxa2
+ obj-$(CONFIG_PXA3xx) += pxa3xx-cpufreq.o
+ obj-$(CONFIG_ARM_QCOM_CPUFREQ_HW) += qcom-cpufreq-hw.o
+ obj-$(CONFIG_ARM_QCOM_CPUFREQ_NVMEM) += qcom-cpufreq-nvmem.o
++obj-$(CONFIG_ARM_QCOM_CPUFREQ_KRAIT) += qcom-cpufreq-krait.o
+ obj-$(CONFIG_ARM_RASPBERRYPI_CPUFREQ) += raspberrypi-cpufreq.o
+ obj-$(CONFIG_ARM_S3C2410_CPUFREQ) += s3c2410-cpufreq.o
+ obj-$(CONFIG_ARM_S3C2412_CPUFREQ) += s3c2412-cpufreq.o
+@@ -86,6 +87,7 @@ obj-$(CONFIG_ARM_TEGRA186_CPUFREQ) += te
+ obj-$(CONFIG_ARM_TEGRA194_CPUFREQ) += tegra194-cpufreq.o
+ obj-$(CONFIG_ARM_TI_CPUFREQ) += ti-cpufreq.o
+ obj-$(CONFIG_ARM_VEXPRESS_SPC_CPUFREQ) += vexpress-spc-cpufreq.o
++obj-$(CONFIG_ARM_KRAIT_CPUFREQ) += krait-cpufreq.o
+
+
+ ##################################################################################
+--- /dev/null
++++ b/drivers/cpufreq/qcom-cpufreq-krait.c
+@@ -0,0 +1,601 @@
++// SPDX-License-Identifier: GPL-2.0
++
++#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
++
++#include <linux/clk.h>
++#include <linux/cpu.h>
++#include <linux/cpufreq.h>
++#include <linux/cpumask.h>
++#include <linux/err.h>
++#include <linux/module.h>
++#include <linux/of.h>
++#include <linux/of_device.h>
++#include <linux/pm_opp.h>
++#include <linux/platform_device.h>
++#include <linux/regulator/consumer.h>
++#include <linux/slab.h>
++#include <linux/thermal.h>
++
++#include "cpufreq-dt.h"
++
++static struct platform_device *l2_pdev;
++
++struct private_data {
++ struct opp_table *opp_table;
++ struct device *cpu_dev;
++ const char *reg_name;
++ bool have_static_opps;
++};
++
++static int set_target(struct cpufreq_policy *policy, unsigned int index)
++{
++ struct private_data *priv = policy->driver_data;
++ unsigned long freq = policy->freq_table[index].frequency;
++ unsigned long target_freq = freq * 1000;
++ struct dev_pm_opp *opp;
++ unsigned int level;
++ int cpu, ret;
++
++ if (l2_pdev) {
++ /* find the max freq across all core */
++ for_each_present_cpu(cpu)
++ if (cpu != index)
++ target_freq = max(
++ target_freq,
++ (unsigned long)cpufreq_quick_get(cpu));
++
++ opp = dev_pm_opp_find_freq_exact(priv->cpu_dev, target_freq,
++ true);
++ if (IS_ERR(opp)) {
++ dev_err(&l2_pdev->dev, "failed to find OPP for %ld\n",
++ target_freq);
++ return PTR_ERR(opp);
++ }
++ level = dev_pm_opp_get_level(opp);
++ dev_pm_opp_put(opp);
++
++ opp = dev_pm_opp_find_level_exact(&l2_pdev->dev, level);
++ if (IS_ERR(opp)) {
++ dev_err(&l2_pdev->dev,
++ "failed to find level OPP for %d\n", level);
++ return PTR_ERR(opp);
++ }
++ target_freq = dev_pm_opp_get_freq(opp);
++ dev_pm_opp_put(opp);
++
++ ret = dev_pm_opp_set_rate(&l2_pdev->dev, target_freq);
++ if (ret)
++ return ret;
++
++ /*
++ * Hardware constraint:
++ * Krait CPU cannot operate at 384MHz with L2 at 1Ghz.
++ * Assume index 0 with the idle freq and level > 0 as
++ * any L2 freq > 384MHz.
++ * Skip CPU freq change in this corner case.
++ */
++ if (unlikely(index == 0 && level != 0)) {
++ dev_err(priv->cpu_dev, "Krait CPU can't operate at idle freq with L2 at 1GHz");
++ return -EINVAL;
++ }
++ }
++
++ ret = dev_pm_opp_set_rate(priv->cpu_dev, freq * 1000);
++ if (ret)
++ return ret;
++
++ arch_set_freq_scale(policy->related_cpus, freq,
++ policy->cpuinfo.max_freq);
++
++ return 0;
++}
++
++/*
++ * An earlier version of opp-v1 bindings used to name the regulator
++ * "cpu0-supply", we still need to handle that for backwards compatibility.
++ */
++static const char *find_supply_name(struct device *dev)
++{
++ struct device_node *np;
++ struct property *pp;
++ int cpu = dev->id;
++ const char *name = NULL;
++
++ np = of_node_get(dev->of_node);
++
++ /* This must be valid for sure */
++ if (WARN_ON(!np))
++ return NULL;
++
++ /* Try "cpu0" for older DTs */
++ if (!cpu) {
++ pp = of_find_property(np, "cpu0-supply", NULL);
++ if (pp) {
++ name = "cpu0";
++ goto node_put;
++ }
++ }
++
++ pp = of_find_property(np, "cpu-supply", NULL);
++ if (pp) {
++ name = "cpu";
++ goto node_put;
++ }
++
++ dev_dbg(dev, "no regulator for cpu%d\n", cpu);
++node_put:
++ of_node_put(np);
++ return name;
++}
++
++static int resources_available(void)
++{
++ struct device *cpu_dev;
++ struct regulator *cpu_reg;
++ struct clk *cpu_clk;
++ int ret = 0;
++ const char *name;
++
++ cpu_dev = get_cpu_device(0);
++ if (!cpu_dev) {
++ pr_err("failed to get cpu0 device\n");
++ return -ENODEV;
++ }
++
++ cpu_clk = clk_get(cpu_dev, NULL);
++ ret = PTR_ERR_OR_ZERO(cpu_clk);
++ if (ret) {
++ /*
++ * If cpu's clk node is present, but clock is not yet
++ * registered, we should try defering probe.
++ */
++ if (ret == -EPROBE_DEFER)
++ dev_dbg(cpu_dev, "clock not ready, retry\n");
++ else
++ dev_err(cpu_dev, "failed to get clock: %d\n", ret);
++
++ return ret;
++ }
++
++ clk_put(cpu_clk);
++
++ name = find_supply_name(cpu_dev);
++ /* Platform doesn't require regulator */
++ if (!name)
++ return 0;
++
++ cpu_reg = regulator_get_optional(cpu_dev, name);
++ ret = PTR_ERR_OR_ZERO(cpu_reg);
++ if (ret) {
++ /*
++ * If cpu's regulator supply node is present, but regulator is
++ * not yet registered, we should try defering probe.
++ */
++ if (ret == -EPROBE_DEFER)
++ dev_dbg(cpu_dev, "cpu0 regulator not ready, retry\n");
++ else
++ dev_dbg(cpu_dev, "no regulator for cpu0: %d\n", ret);
++
++ return ret;
++ }
++
++ regulator_put(cpu_reg);
++ return 0;
++}
++
++static int cpufreq_init(struct cpufreq_policy *policy)
++{
++ struct cpufreq_frequency_table *freq_table;
++ struct opp_table *opp_table = NULL;
++ unsigned int transition_latency;
++ struct private_data *priv;
++ struct device *cpu_dev;
++ bool fallback = false;
++ struct clk *cpu_clk;
++ const char *name;
++ int ret;
++
++ cpu_dev = get_cpu_device(policy->cpu);
++ if (!cpu_dev) {
++ pr_err("failed to get cpu%d device\n", policy->cpu);
++ return -ENODEV;
++ }
++
++ cpu_clk = clk_get(cpu_dev, NULL);
++ if (IS_ERR(cpu_clk)) {
++ ret = PTR_ERR(cpu_clk);
++ dev_err(cpu_dev, "%s: failed to get clk: %d\n", __func__, ret);
++ return ret;
++ }
++
++ /* Get OPP-sharing information from "operating-points-v2" bindings */
++ ret = dev_pm_opp_of_get_sharing_cpus(cpu_dev, policy->cpus);
++ if (ret) {
++ if (ret != -ENOENT)
++ goto out_put_clk;
++
++ /*
++ * operating-points-v2 not supported, fallback to old method of
++ * finding shared-OPPs for backward compatibility if the
++ * platform hasn't set sharing CPUs.
++ */
++ if (dev_pm_opp_get_sharing_cpus(cpu_dev, policy->cpus))
++ fallback = true;
++ }
++
++ /*
++ * OPP layer will be taking care of regulators now, but it needs to know
++ * the name of the regulator first.
++ */
++ name = find_supply_name(cpu_dev);
++ if (name) {
++ opp_table = dev_pm_opp_set_regulators(cpu_dev, &name, 1);
++ if (IS_ERR(opp_table)) {
++ ret = PTR_ERR(opp_table);
++ dev_err(cpu_dev,
++ "Failed to set regulator for cpu%d: %d\n",
++ policy->cpu, ret);
++ goto out_put_clk;
++ }
++ }
++
++ priv = kzalloc(sizeof(*priv), GFP_KERNEL);
++ if (!priv) {
++ ret = -ENOMEM;
++ goto out_put_regulator;
++ }
++
++ priv->reg_name = name;
++ priv->opp_table = opp_table;
++
++ /*
++ * Initialize OPP tables for all policy->cpus. They will be shared by
++ * all CPUs which have marked their CPUs shared with OPP bindings.
++ *
++ * For platforms not using operating-points-v2 bindings, we do this
++ * before updating policy->cpus. Otherwise, we will end up creating
++ * duplicate OPPs for policy->cpus.
++ *
++ * OPPs might be populated at runtime, don't check for error here
++ */
++ if (!dev_pm_opp_of_cpumask_add_table(policy->cpus))
++ priv->have_static_opps = true;
++
++ /*
++ * But we need OPP table to function so if it is not there let's
++ * give platform code chance to provide it for us.
++ */
++ ret = dev_pm_opp_get_opp_count(cpu_dev);
++ if (ret <= 0) {
++ dev_dbg(cpu_dev, "OPP table is not ready, deferring probe\n");
++ ret = -EPROBE_DEFER;
++ goto out_free_opp;
++ }
++
++ if (fallback) {
++ cpumask_setall(policy->cpus);
++
++ /*
++ * OPP tables are initialized only for policy->cpu, do it for
++ * others as well.
++ */
++ ret = dev_pm_opp_set_sharing_cpus(cpu_dev, policy->cpus);
++ if (ret)
++ dev_err(cpu_dev,
++ "%s: failed to mark OPPs as shared: %d\n",
++ __func__, ret);
++ }
++
++ ret = dev_pm_opp_init_cpufreq_table(cpu_dev, &freq_table);
++ if (ret) {
++ dev_err(cpu_dev, "failed to init cpufreq table: %d\n", ret);
++ goto out_free_opp;
++ }
++
++ priv->cpu_dev = cpu_dev;
++
++ policy->driver_data = priv;
++ policy->clk = cpu_clk;
++ policy->freq_table = freq_table;
++
++ policy->suspend_freq = dev_pm_opp_get_suspend_opp_freq(cpu_dev) / 1000;
++
++ transition_latency = dev_pm_opp_get_max_transition_latency(cpu_dev);
++ if (!transition_latency)
++ transition_latency = CPUFREQ_ETERNAL;
++
++ policy->cpuinfo.transition_latency = transition_latency;
++ policy->dvfs_possible_from_any_cpu = true;
++
++ dev_pm_opp_of_register_em(policy->cpus);
++
++ return 0;
++
++out_free_opp:
++ if (priv->have_static_opps)
++ dev_pm_opp_of_cpumask_remove_table(policy->cpus);
++ kfree(priv);
++out_put_regulator:
++ if (name)
++ dev_pm_opp_put_regulators(opp_table);
++out_put_clk:
++ clk_put(cpu_clk);
++
++ return ret;
++}
++
++static int cpufreq_online(struct cpufreq_policy *policy)
++{
++ /* We did light-weight tear down earlier, nothing to do here */
++ return 0;
++}
++
++static int cpufreq_offline(struct cpufreq_policy *policy)
++{
++ /*
++ * Preserve policy->driver_data and don't free resources on light-weight
++ * tear down.
++ */
++ return 0;
++}
++
++static int cpufreq_exit(struct cpufreq_policy *policy)
++{
++ struct private_data *priv = policy->driver_data;
++
++ dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table);
++ if (priv->have_static_opps)
++ dev_pm_opp_of_cpumask_remove_table(policy->related_cpus);
++ if (priv->reg_name)
++ dev_pm_opp_put_regulators(priv->opp_table);
++
++ clk_put(policy->clk);
++ kfree(priv);
++
++ return 0;
++}
++
++static struct cpufreq_driver krait_cpufreq_driver = {
++ .flags = CPUFREQ_STICKY | CPUFREQ_NEED_INITIAL_FREQ_CHECK |
++ CPUFREQ_IS_COOLING_DEV,
++ .verify = cpufreq_generic_frequency_table_verify,
++ .target_index = set_target,
++ .get = cpufreq_generic_get,
++ .init = cpufreq_init,
++ .exit = cpufreq_exit,
++ .online = cpufreq_online,
++ .offline = cpufreq_offline,
++ .name = "krait-cpufreq",
++ .suspend = cpufreq_generic_suspend,
++};
++
++struct krait_data {
++ unsigned long idle_freq;
++ bool regulator_enabled;
++};
++
++static int krait_cache_set_opp(struct dev_pm_set_opp_data *data)
++{
++ unsigned long old_freq = data->old_opp.rate, freq = data->new_opp.rate;
++ struct dev_pm_opp_supply *supply = &data->new_opp.supplies[0];
++ struct regulator *reg = data->regulators[0];
++ struct clk *clk = data->clk;
++ struct krait_data *kdata;
++ unsigned long idle_freq;
++ int ret;
++
++ kdata = (struct krait_data *)dev_get_drvdata(data->dev);
++ idle_freq = kdata->idle_freq;
++
++ /* Scaling up? Scale voltage before frequency */
++ if (freq >= old_freq) {
++ ret = regulator_set_voltage_triplet(reg, supply->u_volt_min,
++ supply->u_volt,
++ supply->u_volt_max);
++ if (ret)
++ goto exit;
++ }
++
++ /*
++ * Set to idle bin if switching from normal to high bin
++ * or vice versa. It has been notice that a bug is triggered
++ * in cache scaling when more than one bin is scaled, to fix
++ * this we first need to transition to the base rate and then
++ * to target rate
++ */
++ if (likely(freq != idle_freq && old_freq != idle_freq)) {
++ ret = clk_set_rate(clk, idle_freq);
++ if (ret)
++ goto exit;
++ }
++
++ ret = clk_set_rate(clk, freq);
++ if (ret)
++ goto exit;
++
++ /* Scaling down? Scale voltage after frequency */
++ if (freq < old_freq) {
++ ret = regulator_set_voltage_triplet(reg, supply->u_volt_min,
++ supply->u_volt,
++ supply->u_volt_max);
++ }
++
++ if (unlikely(!kdata->regulator_enabled)) {
++ ret = regulator_enable(reg);
++ if (ret < 0)
++ dev_warn(data->dev, "Failed to enable regulator: %d", ret);
++ else
++ kdata->regulator_enabled = true;
++ }
++
++exit:
++ return ret;
++};
++
++static int krait_cache_probe(struct platform_device *pdev)
++{
++ struct device *dev = &pdev->dev;
++ struct krait_data *data;
++ struct opp_table *table;
++ struct dev_pm_opp *opp;
++ struct device *cpu_dev;
++ int ret;
++
++ data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
++ if (!data)
++ return -ENOMEM;
++
++ table = dev_pm_opp_set_regulators(dev, (const char *[]){ "l2" }, 1);
++ if (IS_ERR(table)) {
++ ret = PTR_ERR(table);
++ if (ret != -EPROBE_DEFER)
++ dev_err(dev, "failed to set regulators %d\n", ret);
++
++ return ret;
++ }
++
++ ret = PTR_ERR_OR_ZERO(
++ dev_pm_opp_register_set_opp_helper(dev, krait_cache_set_opp));
++ if (ret)
++ return ret;
++
++ ret = dev_pm_opp_of_add_table(dev);
++ if (ret) {
++ dev_err(dev, "failed to parse L2 freq thresholds\n");
++ return ret;
++ }
++
++ opp = dev_pm_opp_find_freq_ceil(dev, &data->idle_freq);
++ dev_pm_opp_put(opp);
++
++ /*
++ * Check opp-level configuration
++ * At least 2 level must be set or the cache will always be scaled
++ * the idle freq causing some performance problem
++ *
++ * In case of invalid configuration, the l2 scaling is skipped
++ */
++ cpu_dev = get_cpu_device(0);
++ if (!cpu_dev) {
++ pr_err("failed to get cpu0 device\n");
++ return -ENODEV;
++ }
++
++ /*
++ * Check if we have at least opp-level 1, 0 should always be set to
++ * the idle freq
++ */
++ opp = dev_pm_opp_find_level_exact(dev, 1);
++ if (IS_ERR(opp)) {
++ dev_err(dev,
++ "Invalid configuration found of l2 opp. Can't find opp-level 1");
++ goto invalid_conf;
++ }
++ dev_pm_opp_put(opp);
++
++ /*
++ * Check if we have at least opp-level 1 in the cpu opp, 0 should always
++ * be set to the idle freq
++ */
++ opp = dev_pm_opp_find_level_exact(cpu_dev, 1);
++ if (IS_ERR(opp)) {
++ dev_err(dev,
++ "Invalid configuration found of cpu opp. Can't find opp-level 1");
++ goto invalid_conf;
++ }
++ dev_pm_opp_put(opp);
++
++ platform_set_drvdata(pdev, data);
++
++ /* The l2 scaling is enabled by linking the cpufreq driver */
++ l2_pdev = pdev;
++
++ return 0;
++
++invalid_conf:
++ dev_pm_opp_remove_table(dev);
++ dev_pm_opp_put_regulators(table);
++ dev_pm_opp_unregister_set_opp_helper(table);
++
++ return -EINVAL;
++};
++
++static int krait_cache_remove(struct platform_device *pdev)
++{
++ struct device *dev = &pdev->dev;
++ struct opp_table *table = dev_pm_opp_get_opp_table(dev);
++
++ dev_pm_opp_remove_table(dev);
++ dev_pm_opp_put_regulators(table);
++ dev_pm_opp_unregister_set_opp_helper(table);
++
++ return 0;
++};
++
++static const struct of_device_id krait_cache_match_table[] = {
++ { .compatible = "qcom,krait-cache" },
++ {}
++};
++
++static struct platform_driver krait_cache_driver = {
++ .driver = {
++ .name = "krait-cache",
++ .of_match_table = krait_cache_match_table,
++ },
++ .probe = krait_cache_probe,
++ .remove = krait_cache_remove,
++};
++module_platform_driver(krait_cache_driver);
++
++static int krait_cpufreq_probe(struct platform_device *pdev)
++{
++ struct cpufreq_dt_platform_data *data = dev_get_platdata(&pdev->dev);
++ int ret;
++
++ /*
++ * All per-cluster (CPUs sharing clock/voltages) initialization is done
++ * from ->init(). In probe(), we just need to make sure that clk and
++ * regulators are available. Else defer probe and retry.
++ *
++ * FIXME: Is checking this only for CPU0 sufficient ?
++ */
++ ret = resources_available();
++ if (ret)
++ return ret;
++
++ if (data) {
++ if (data->have_governor_per_policy)
++ krait_cpufreq_driver.flags |=
++ CPUFREQ_HAVE_GOVERNOR_PER_POLICY;
++
++ krait_cpufreq_driver.resume = data->resume;
++ if (data->suspend)
++ krait_cpufreq_driver.suspend = data->suspend;
++ }
++
++ ret = cpufreq_register_driver(&krait_cpufreq_driver);
++ if (ret)
++ dev_err(&pdev->dev, "failed register driver: %d\n", ret);
++
++ return ret;
++}
++
++static int krait_cpufreq_remove(struct platform_device *pdev)
++{
++ cpufreq_unregister_driver(&krait_cpufreq_driver);
++ return 0;
++}
++
++static struct platform_driver krait_cpufreq_platdrv = {
++ .driver = {
++ .name = "krait-cpufreq",
++ },
++ .probe = krait_cpufreq_probe,
++ .remove = krait_cpufreq_remove,
++};
++module_platform_driver(krait_cpufreq_platdrv);
++
++MODULE_ALIAS("platform:krait-cpufreq");
++MODULE_AUTHOR("Ansuel Smith <ansuelsmth at gmail.com>");
++MODULE_DESCRIPTION("Dedicated Krait SoC cpufreq driver");
++MODULE_LICENSE("GPL");
diff --git a/target/linux/ipq806x/patches-5.4/098-2-Documentation-cpufreq-add-qcom-krait-cpufreq-binding.patch b/target/linux/ipq806x/patches-5.4/098-2-Documentation-cpufreq-add-qcom-krait-cpufreq-binding.patch
new file mode 100644
index 0000000000..316e18b790
--- /dev/null
+++ b/target/linux/ipq806x/patches-5.4/098-2-Documentation-cpufreq-add-qcom-krait-cpufreq-binding.patch
@@ -0,0 +1,237 @@
+From c9ecd920324a647bf1f2b47f771c8f599cc7b551 Mon Sep 17 00:00:00 2001
+From: Ansuel Smith <ansuelsmth at gmail.com>
+Date: Sat, 22 Feb 2020 18:02:17 +0100
+Subject: [PATCH 2/8] Documentation: cpufreq: add qcom,krait-cache bindings
+
+Document dedicated cpufreq for Krait CPUs.
+
+Signed-off-by: Ansuel Smith <ansuelsmth at gmail.com>
+---
+ .../bindings/cpufreq/qcom-cpufreq-krait.yaml | 221 ++++++++++++++++++
+ 1 file changed, 221 insertions(+)
+ create mode 100644 Documentation/devicetree/bindings/cpufreq/qcom-cpufreq-krait.yaml
+
+--- /dev/null
++++ b/Documentation/devicetree/bindings/cpufreq/qcom-cpufreq-krait.yaml
+@@ -0,0 +1,221 @@
++# SPDX-License-Identifier: GPL-2.0
++%YAML 1.2
++---
++$id: http://devicetree.org/schemas/cpufreq/qcom-cpufreq-krait.yaml#
++$schema: http://devicetree.org/meta-schemas/core.yaml#
++
++title: CPU Frequency scaling driver for Krait SoCs
++
++maintainers:
++ - Ansuel Smith <ansuelsmth at gmail.com>
++
++description: |
++ The krait cpufreq driver is a dedicated frequency scaling driver
++ based on cpufreq-dt generic driver that scale L2 cache and the
++ cores. TEST
++
++ The L2 cache is scaled based on the max clk across all cores and
++ the clock is decided based on the opp-level set in the device tree.
++
++ Different core freq can be linked to a specific l2 freq and the driver
++ on frequency change will scale the core and the l2 clk based of the
++ linked freq.
++
++ On Krait SoC is present a bug and on every L2 clk change the driver
++ needs to set the clk to the idle freq before changing it to the new value.
++
++ This requires the qcom cpufreq nvmem driver to parse the different opp
++ core clk and an additional opp table for the l2 scaling.
++
++ If the driver detect broken config (for example missing opp-level) the
++ cpufreq driver skips the l2 scaling
++
++ Referring to this example opp-level can be used to link a range of cpu freq
++ to a specific l2 freq:
++ cpu opp freq 384000000 has opp-level 0
++ l2 opp freq 384000000 has opp-level 0
++ The driver will scale l2 to 384000000
++
++ cpu opp freq 600000000-1000000000 has opp-level 1
++ l2 opp freq 1000000000 has opp-level 1
++ The driver will scale l2 to 1000000000
++
++allOf:
++ - $ref: /schemas/cache-controller.yaml#
++
++select:
++ properties:
++ compatible:
++ items:
++ - enum:
++ - qcom,krait-cache
++
++ required:
++ - compatible
++
++properties:
++ compatible:
++ items:
++ - const: qcom,krait-cache
++ - const: cache
++
++ cache-level:
++ const: 2
++
++ clocks:
++ maxItems: 1
++
++ clock-names:
++ const: l2
++
++ l2-supply: true
++
++ operating-points-v2: true
++
++required:
++ - compatible
++ - cache-level
++ - clocks
++ - clock-names
++ - l2-supply
++ - operating-points-v2
++
++additionalProperties: false
++
++examples:
++ - |
++ cpus {
++ #address-cells = <1>;
++ #size-cells = <0>;
++
++ cpu0: cpu at 0 {
++ compatible = "qcom,krait";
++ enable-method = "qcom,kpss-acc-v1";
++ device_type = "cpu";
++ reg = <0>;
++ next-level-cache = <&L2>;
++ qcom,acc = <&acc0>;
++ qcom,saw = <&saw0>;
++ clocks = <&kraitcc 0>, <&kraitcc 4>;
++ clock-names = "cpu", "l2";
++ clock-latency = <100000>;
++ cpu-supply = <&smb208_s2a>;
++ operating-points-v2 = <&opp_table0>;
++ voltage-tolerance = <5>;
++ cooling-min-state = <0>;
++ cooling-max-state = <10>;
++ #cooling-cells = <2>;
++ cpu-idle-states = <&CPU_SPC>;
++ };
++
++ /* ... */
++
++ };
++
++ opp_table0: opp_table0 {
++ compatible = "operating-points-v2-kryo-cpu";
++ nvmem-cells = <&speedbin_efuse>;
++
++ opp-384000000 {
++ opp-hz = /bits/ 64 <384000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1000000>;
++ opp-microvolt-speed0-pvs1-v0 = <925000>;
++ opp-microvolt-speed0-pvs2-v0 = <875000>;
++ opp-microvolt-speed0-pvs3-v0 = <800000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <0>;
++ };
++
++ opp-600000000 {
++ opp-hz = /bits/ 64 <600000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1050000>;
++ opp-microvolt-speed0-pvs1-v0 = <975000>;
++ opp-microvolt-speed0-pvs2-v0 = <925000>;
++ opp-microvolt-speed0-pvs3-v0 = <850000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <1>;
++ };
++
++ opp-800000000 {
++ opp-hz = /bits/ 64 <800000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1100000>;
++ opp-microvolt-speed0-pvs1-v0 = <1025000>;
++ opp-microvolt-speed0-pvs2-v0 = <995000>;
++ opp-microvolt-speed0-pvs3-v0 = <900000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <1>;
++ };
++
++ opp-1000000000 {
++ opp-hz = /bits/ 64 <1000000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1150000>;
++ opp-microvolt-speed0-pvs1-v0 = <1075000>;
++ opp-microvolt-speed0-pvs2-v0 = <1025000>;
++ opp-microvolt-speed0-pvs3-v0 = <950000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <1>;
++ };
++
++ opp-1200000000 {
++ opp-hz = /bits/ 64 <1200000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1200000>;
++ opp-microvolt-speed0-pvs1-v0 = <1125000>;
++ opp-microvolt-speed0-pvs2-v0 = <1075000>;
++ opp-microvolt-speed0-pvs3-v0 = <1000000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <2>;
++ };
++
++ opp-1400000000 {
++ opp-hz = /bits/ 64 <1400000000>;
++ opp-microvolt-speed0-pvs0-v0 = <1250000>;
++ opp-microvolt-speed0-pvs1-v0 = <1175000>;
++ opp-microvolt-speed0-pvs2-v0 = <1125000>;
++ opp-microvolt-speed0-pvs3-v0 = <1050000>;
++ opp-supported-hw = <0x1>;
++ clock-latency-ns = <100000>;
++ opp-level = <2>;
++ };
++ };
++
++ opp_table_l2: opp_table_l2 {
++ compatible = "operating-points-v2";
++
++ opp-384000000 {
++ opp-hz = /bits/ 64 <384000000>;
++ opp-microvolt = <1100000>;
++ clock-latency-ns = <100000>;
++ opp-level = <0>;
++ };
++ opp-1000000000 {
++ opp-hz = /bits/ 64 <1000000000>;
++ opp-microvolt = <1100000>;
++ clock-latency-ns = <100000>;
++ opp-level = <1>;
++ };
++ opp-1200000000 {
++ opp-hz = /bits/ 64 <1200000000>;
++ opp-microvolt = <1150000>;
++ clock-latency-ns = <100000>;
++ opp-level = <2>;
++ };
++ };
++
++ soc {
++ L2: l2-cache {
++ compatible = "qcom,krait-cache", "cache";
++ cache-level = <2>;
++
++ clocks = <&kraitcc 4>;
++ clock-names = "l2";
++ l2-supply = <&smb208_s1a>;
++ operating-points-v2 = <&opp_table_l2>;
++ };
++ };
++
++...
diff --git a/target/linux/ipq806x/patches-5.4/0057-add-fab-scaling-support-with-cpufreq.patch b/target/linux/ipq806x/patches-5.4/098-3-add-fab-scaling-support-with-cpufreq.patch
similarity index 93%
rename from target/linux/ipq806x/patches-5.4/0057-add-fab-scaling-support-with-cpufreq.patch
rename to target/linux/ipq806x/patches-5.4/098-3-add-fab-scaling-support-with-cpufreq.patch
index 441500df79..fe9143218f 100644
--- a/target/linux/ipq806x/patches-5.4/0057-add-fab-scaling-support-with-cpufreq.patch
+++ b/target/linux/ipq806x/patches-5.4/098-3-add-fab-scaling-support-with-cpufreq.patch
@@ -217,9 +217,9 @@
+int scale_fabrics(unsigned long max_cpu_freq);
+
+#endif
---- a/drivers/cpufreq/cpufreq-dt.c
-+++ b/drivers/cpufreq/cpufreq-dt.c
-@@ -20,6 +20,7 @@
+--- a/drivers/cpufreq/qcom-cpufreq-krait.c
++++ b/drivers/cpufreq/qcom-cpufreq-krait.c
+@@ -15,6 +15,7 @@
#include <linux/regulator/consumer.h>
#include <linux/slab.h>
#include <linux/thermal.h>
@@ -227,17 +227,17 @@
#include "cpufreq-dt.h"
-@@ -111,6 +112,13 @@ static int set_target(struct cpufreq_pol
- }
- }
- }
-+
-+ /*
-+ * Scale fabrics with max freq across all cores
-+ */
-+ ret = scale_fabrics(target_freq);
-+ if (ret)
-+ goto exit;
- }
- priv->opp_freq = freq * 1000;
- arch_set_freq_scale(policy->related_cpus, freq,
+@@ -54,6 +55,13 @@ static int set_target(struct cpufreq_pol
+ level = dev_pm_opp_get_level(opp);
+ dev_pm_opp_put(opp);
+
++ /*
++ * Scale fabrics with max freq across all cores
++ */
++ ret = scale_fabrics(target_freq);
++ if (ret)
++ return ret;
++
+ opp = dev_pm_opp_find_level_exact(&l2_pdev->dev, level);
+ if (IS_ERR(opp)) {
+ dev_err(&l2_pdev->dev,
--
2.25.1
More information about the openwrt-devel
mailing list