Merge tag 'pm-5.19-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull more power management updates from Rafael Wysocki:
"These update the ARM cpufreq drivers and fix up the CPPC cpufreq
driver after recent changes, update the OPP code and PM documentation
and add power sequences support to the system reboot and power off
code.

Specifics:

- Add Tegra234 cpufreq support (Sumit Gupta)

- Clean up and enhance the Mediatek cpufreq driver (Wan Jiabing,
Rex-BC Chen, and Jia-Wei Chang)

- Fix up the CPPC cpufreq driver after recent changes (Zheng Bin,
Pierre Gondois)

- Minor update to dt-binding for Qcom's opp-v2-kryo-cpu (Yassine
Oudjana)

- Use list iterator only inside the list_for_each_entry loop
(Xiaomeng Tong, and Jakob Koschel)

- New APIs related to finding OPP based on interconnect bandwidth
(Krzysztof Kozlowski)

- Fix the missing of_node_put() in _bandwidth_supported() (Dan
Carpenter)

- Cleanups (Krzysztof Kozlowski, and Viresh Kumar)

- Add Out of Band mode description to the intel-speed-select utility
documentation (Srinivas Pandruvada)

- Add power sequences support to the system reboot and power off code
and make related platform-specific changes for multiple platforms
(Dmitry Osipenko, Geert Uytterhoeven)"

* tag 'pm-5.19-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (60 commits)
cpufreq: CPPC: Fix unused-function warning
cpufreq: CPPC: Fix build error without CONFIG_ACPI_CPPC_CPUFREQ_FIE
Documentation: admin-guide: PM: Add Out of Band mode
kernel/reboot: Change registration order of legacy power-off handler
m68k: virt: Switch to new sys-off handler API
kernel/reboot: Add devm_register_restart_handler()
kernel/reboot: Add devm_register_power_off_handler()
soc/tegra: pmc: Use sys-off handler API to power off Nexus 7 properly
reboot: Remove pm_power_off_prepare()
regulator: pfuze100: Use devm_register_sys_off_handler()
ACPI: power: Switch to sys-off handler API
memory: emif: Use kernel_can_power_off()
mips: Use do_kernel_power_off()
ia64: Use do_kernel_power_off()
x86: Use do_kernel_power_off()
sh: Use do_kernel_power_off()
m68k: Switch to new sys-off handler API
powerpc: Use do_kernel_power_off()
xen/x86: Use do_kernel_power_off()
parisc: Use do_kernel_power_off()
...

Documentation/admin-guide/pm/intel-speed-select.rst

···
 level 0 to 2800 MHz at performance level 4. As a result, any workload, which can
 use fewer CPUs, can see a boost of 200 MHz compared to performance level 0.
 
+Changing performance level via BMC Interface
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+It is possible to change SST-PP level using out of band (OOB) agent (Via some
+remote management console, through BMC "Baseboard Management Controller"
+interface). This mode is supported from the Sapphire Rapids processor
+generation. The kernel and tool change to support this mode is added to Linux
+kernel version 5.18. To enable this feature, kernel config
+"CONFIG_INTEL_HFI_THERMAL" is required. The minimum version of the tool
+is "v1.12" to support this feature, which is part of Linux kernel version 5.18.
+
+To support such configuration, this tool can be used as a daemon. Add
+a command line option --oob::
+
+ # intel-speed-select --oob
+ Intel(R) Speed Select Technology
+ Executing on CPU model:143[0x8f]
+ OOB mode is enabled and will run as daemon
+
+In this mode the tool will online/offline CPUs based on the new performance
+level.
+
 Check presence of other Intel(R) SST features
 ---------------------------------------------
 

Documentation/devicetree/bindings/arm/tegra/nvidia,tegra-ccplex-cluster.yaml

···
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: "http://devicetree.org/schemas/arm/tegra/nvidia,tegra-ccplex-cluster.yaml#"
+$schema: "http://devicetree.org/meta-schemas/core.yaml#"
+
+title: NVIDIA Tegra CPU COMPLEX CLUSTER area device tree bindings
+
+maintainers:
+  - Sumit Gupta <[e-mail]>
+  - Mikko Perttunen <[e-mail]>
+  - Jon Hunter <[e-mail]>
+  - Thierry Reding <[e-mail]>
+
+description: |+
+  The Tegra CPU COMPLEX CLUSTER area contains memory-mapped
+  registers that initiate CPU frequency/voltage transitions.
+
+properties:
+  $nodename:
+    pattern: "ccplex@([0-9a-f]+)$"
+
+  compatible:
+    enum:
+      - nvidia,tegra186-ccplex-cluster
+      - nvidia,tegra234-ccplex-cluster
+
+  reg:
+    maxItems: 1
+
+  nvidia,bpmp:
+    $ref: '/schemas/types.yaml#/definitions/phandle'
+    description: |
+      Specifies the BPMP node that needs to be queried to get
+      operating point data for all CPUs.
+
+additionalProperties: false
+
+required:
+  - compatible
+  - reg
+  - nvidia,bpmp
+  - status
+
+examples:
+  - |
+    ccplex@e000000 {
+      compatible = "nvidia,tegra234-ccplex-cluster";
+      reg = <0x0e000000 0x5ffff>;
+      nvidia,bpmp = <&bpmp>;
+      status = "okay";
+    };

Documentation/devicetree/bindings/cpufreq/cpufreq-mediatek.txt

···
 	       Vsram to fit SoC specific needs. When absent, the voltage scaling
 	       flow is handled by hardware, hence no software "voltage tracking" is
 	       needed.
+- mediatek,cci:
+	Used to confirm the link status between cpufreq and mediatek cci. Because
+	cpufreq and mediatek cci could share the same regulator in some MediaTek SoCs.
+	To prevent the issue of high frequency and low voltage, we need to use this
+	property to make sure mediatek cci is ready.
+	For details of mediatek cci, please refer to
+	Documentation/devicetree/bindings/interconnect/mediatek,cci.yaml
 - #cooling-cells:
 	For details, please refer to
 	Documentation/devicetree/bindings/thermal/thermal-cooling-devices.yaml

Documentation/devicetree/bindings/opp/opp-v2-kryo-cpu.yaml

···
   the CPU frequencies subset and voltage value of each OPP varies based on
   the silicon variant in use.
   Qualcomm Technologies, Inc. Process Voltage Scaling Tables
-  defines the voltage and frequency value based on the msm-id in SMEM
-  and speedbin blown in the efuse combination.
-  The qcom-cpufreq-nvmem driver reads the msm-id and efuse value from the SoC
-  to provide the OPP framework with required information (existing HW bitmap).
+  defines the voltage and frequency value based on the speedbin blown in
+  the efuse combination.
+  The qcom-cpufreq-nvmem driver reads the efuse value from the SoC to provide
+  the OPP framework with required information (existing HW bitmap).
   This is used to determine the voltage and frequency value for each OPP of
   operating-points-v2 table when it is parsed by the OPP framework.
 
···
         description: |
           A single 32 bit bitmap value, representing compatible HW.
           Bitmap:
-          0:  MSM8996 V3, speedbin 0
-          1:  MSM8996 V3, speedbin 1
-          2:  MSM8996 V3, speedbin 2
-          3:  unused
-          4:  MSM8996 SG, speedbin 0
-          5:  MSM8996 SG, speedbin 1
-          6:  MSM8996 SG, speedbin 2
-          7-31:  unused
-        maximum: 0x77
+          0:  MSM8996, speedbin 0
+          1:  MSM8996, speedbin 1
+          2:  MSM8996, speedbin 2
+          3-31:  unused
+        maximum: 0x7
 
       clock-latency-ns: true
 
···
             opp-307200000 {
                 opp-hz = /bits/ 64 <307200000>;
                 opp-microvolt = <905000 905000 1140000>;
-                opp-supported-hw = <0x77>;
+                opp-supported-hw = <0x7>;
+                clock-latency-ns = <200000>;
+            };
+            opp-1401600000 {
+                opp-hz = /bits/ 64 <1401600000>;
+                opp-microvolt = <1140000 905000 1140000>;
+                opp-supported-hw = <0x5>;
                 clock-latency-ns = <200000>;
             };
             opp-1593600000 {
                 opp-hz = /bits/ 64 <1593600000>;
                 opp-microvolt = <1140000 905000 1140000>;
-                opp-supported-hw = <0x71>;
-                clock-latency-ns = <200000>;
-            };
-            opp-2188800000 {
-                opp-hz = /bits/ 64 <2188800000>;
-                opp-microvolt = <1140000 905000 1140000>;
-                opp-supported-hw = <0x10>;
+                opp-supported-hw = <0x1>;
                 clock-latency-ns = <200000>;
             };
         };
···
             opp-307200000 {
                 opp-hz = /bits/ 64 <307200000>;
                 opp-microvolt = <905000 905000 1140000>;
-                opp-supported-hw = <0x77>;
+                opp-supported-hw = <0x7>;
                 clock-latency-ns = <200000>;
             };
-            opp-1593600000 {
-                opp-hz = /bits/ 64 <1593600000>;
+            opp-1804800000 {
+                opp-hz = /bits/ 64 <1804800000>;
                 opp-microvolt = <1140000 905000 1140000>;
-                opp-supported-hw = <0x70>;
+                opp-supported-hw = <0x6>;
+                clock-latency-ns = <200000>;
+            };
+            opp-1900800000 {
+                opp-hz = /bits/ 64 <1900800000>;
+                opp-microvolt = <1140000 905000 1140000>;
+                opp-supported-hw = <0x4>;
                 clock-latency-ns = <200000>;
             };
             opp-2150400000 {
                 opp-hz = /bits/ 64 <2150400000>;
                 opp-microvolt = <1140000 905000 1140000>;
-                opp-supported-hw = <0x31>;
-                clock-latency-ns = <200000>;
-            };
-            opp-2342400000 {
-                opp-hz = /bits/ 64 <2342400000>;
-                opp-microvolt = <1140000 905000 1140000>;
-                opp-supported-hw = <0x10>;
+                opp-supported-hw = <0x1>;
                 clock-latency-ns = <200000>;
             };
         };

arch/arm/kernel/reboot.c

···
 {
 	local_irq_disable();
 	smp_send_stop();
-
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 }
 
 /*

arch/arm64/kernel/process.c

···
 {
 	local_irq_disable();
 	smp_send_stop();
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 }
 
 /*

arch/csky/kernel/power.c

···
 void machine_power_off(void)
 {
 	local_irq_disable();
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 	asm volatile ("bkpt");
 }
 
 void machine_halt(void)
 {
 	local_irq_disable();
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 	asm volatile ("bkpt");
 }
 

arch/ia64/kernel/process.c

···
 #include <linux/module.h>
 #include <linux/notifier.h>
 #include <linux/personality.h>
+#include <linux/reboot.h>
 #include <linux/sched.h>
 #include <linux/sched/debug.h>
 #include <linux/sched/hotplug.h>
···
 void
 machine_power_off (void)
 {
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 	machine_halt();
 }
 

arch/m68k/emu/natfeat.c

···
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/reboot.h>
 #include <linux/io.h>
 #include <asm/machdep.h>
 #include <asm/natfeat.h>
···
 	pr_info("NatFeats found (%s, %lu.%lu)\n", buf, version >> 16,
 		version & 0xffff);
 
-	mach_power_off = nf_poweroff;
+	register_platform_power_off(nf_poweroff);
 }

arch/m68k/include/asm/machdep.h

···
 extern int (*mach_set_rtc_pll)(struct rtc_pll_info *);
 extern void (*mach_reset)( void );
 extern void (*mach_halt)( void );
-extern void (*mach_power_off)( void );
 extern unsigned long (*mach_hd_init) (unsigned long, unsigned long);
 extern void (*mach_hd_setup)(char *, int *);
 extern void (*mach_heartbeat) (int);

arch/m68k/kernel/process.c

···
 
 void machine_power_off(void)
 {
-	if (mach_power_off)
-		mach_power_off();
+	do_kernel_power_off();
 	for (;;);
 }
 
-void (*pm_power_off)(void) = machine_power_off;
+void (*pm_power_off)(void);
 EXPORT_SYMBOL(pm_power_off);
 
 void show_regs(struct pt_regs * regs)

arch/m68k/kernel/setup_mm.c

···
 void (*mach_get_hardware_list) (struct seq_file *m);
 void (*mach_reset)( void );
 void (*mach_halt)( void );
-void (*mach_power_off)( void );
 #ifdef CONFIG_HEARTBEAT
 void (*mach_heartbeat) (int);
 EXPORT_SYMBOL(mach_heartbeat);

arch/m68k/kernel/setup_no.c

···
 /* machine dependent reboot functions */
 void (*mach_reset)(void);
 void (*mach_halt)(void);
-void (*mach_power_off)(void);
 
 #ifdef CONFIG_M68000
 #if defined(CONFIG_M68328)

arch/m68k/mac/config.c

···
 
 #include <linux/errno.h>
 #include <linux/module.h>
+#include <linux/reboot.h>
 #include <linux/types.h>
 #include <linux/mm.h>
 #include <linux/tty.h>
···
 	mach_hwclk = mac_hwclk;
 	mach_reset = mac_reset;
 	mach_halt = mac_poweroff;
-	mach_power_off = mac_poweroff;
 #if IS_ENABLED(CONFIG_INPUT_M68K_BEEP)
 	mach_beep = mac_mksound;
 #endif
···
 
 	if (macintosh_config->ident == MAC_MODEL_IICI)
 		mach_l2_flush = via_l2_flush;
+
+	register_platform_power_off(mac_poweroff);
 }
 
 

arch/m68k/virt/config.c

···
 // SPDX-License-Identifier: GPL-2.0
 
+#include <linux/reboot.h>
 #include <linux/serial_core.h>
 #include <clocksource/timer-goldfish.h>
 
···
 	mach_get_model = virt_get_model;
 	mach_reset = virt_reset;
 	mach_halt = virt_halt;
-	mach_power_off = virt_halt;
+
+	register_platform_power_off(virt_halt);
 }

arch/mips/kernel/reset.c

···
 
 void machine_power_off(void)
 {
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 
 #ifdef CONFIG_SMP
 	preempt_disable();

arch/parisc/kernel/process.c

···
 #include <linux/module.h>
 #include <linux/personality.h>
 #include <linux/ptrace.h>
+#include <linux/reboot.h>
 #include <linux/sched.h>
 #include <linux/sched/debug.h>
 #include <linux/sched/task.h>
···
 	pdc_chassis_send_status(PDC_CHASSIS_DIRECT_SHUTDOWN);
 
 	/* ipmi_poweroff may have been installed. */
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 		
 	/* It seems we have no way to power the system off via
 	 * software. The user has to press the button himself. */

arch/powerpc/kernel/setup-common.c

···
 void machine_power_off(void)
 {
 	machine_shutdown();
-	if (pm_power_off)
-		pm_power_off();
-
+	do_kernel_power_off();
 	smp_send_stop();
 	machine_hang();
 }

arch/powerpc/xmon/xmon.c

···
 	} else if (cmd == 'h') {
 		ppc_md.halt();
 	} else if (cmd == 'p') {
-		if (pm_power_off)
-			pm_power_off();
+		do_kernel_power_off();
 	}
 }
 

arch/riscv/kernel/reset.c

···
 
 void machine_halt(void)
 {
-	if (pm_power_off != NULL)
-		pm_power_off();
-	else
-		default_power_off();
+	do_kernel_power_off();
+	default_power_off();
 }
 
 void machine_power_off(void)
 {
-	if (pm_power_off != NULL)
-		pm_power_off();
-	else
-		default_power_off();
+	do_kernel_power_off();
+	default_power_off();
 }

arch/sh/kernel/reboot.c

···
 
 static void native_machine_power_off(void)
 {
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 }
 
 static void native_machine_halt(void)

arch/x86/kernel/reboot.c

···
 
 static void native_machine_power_off(void)
 {
-	if (pm_power_off) {
+	if (kernel_can_power_off()) {
 		if (!reboot_force)
 			machine_shutdown();
-		pm_power_off();
+		do_kernel_power_off();
 	}
 	/* A fallback in case there is no PM info available */
 	tboot_shutdown(TB_SHUTDOWN_HALT);

arch/x86/xen/enlighten_pv.c

···
 #include <linux/pci.h>
 #include <linux/gfp.h>
 #include <linux/edd.h>
+#include <linux/reboot.h>
 
 #include <xen/xen.h>
 #include <xen/events.h>
···
 
 static void xen_machine_power_off(void)
 {
-	if (pm_power_off)
-		pm_power_off();
+	do_kernel_power_off();
 	xen_reboot(SHUTDOWN_poweroff);
 }
 

drivers/acpi/sleep.c

···
 static inline void acpi_sleep_hibernate_setup(void) {}
 #endif /* !CONFIG_HIBERNATION */
 
-static void acpi_power_off_prepare(void)
+static int acpi_power_off_prepare(struct sys_off_data *data)
 {
 	/* Prepare to power off the system */
 	acpi_sleep_prepare(ACPI_STATE_S5);
 	acpi_disable_all_gpes();
 	acpi_os_wait_events_complete();
+	return NOTIFY_DONE;
 }
 
-static void acpi_power_off(void)
+static int acpi_power_off(struct sys_off_data *data)
 {
 	/* acpi_sleep_prepare(ACPI_STATE_S5) should have already been called */
 	pr_debug("%s called\n", __func__);
 	local_irq_disable();
 	acpi_enter_sleep_state(ACPI_STATE_S5);
+	return NOTIFY_DONE;
 }
 
 int __init acpi_sleep_init(void)
···
 
 	if (acpi_sleep_state_supported(ACPI_STATE_S5)) {
 		sleep_states[ACPI_STATE_S5] = 1;
-		pm_power_off_prepare = acpi_power_off_prepare;
-		pm_power_off = acpi_power_off;
+
+		register_sys_off_handler(SYS_OFF_MODE_POWER_OFF_PREPARE,
+					 SYS_OFF_PRIO_FIRMWARE,
+					 acpi_power_off_prepare, NULL);
+
+		register_sys_off_handler(SYS_OFF_MODE_POWER_OFF,
+					 SYS_OFF_PRIO_FIRMWARE,
+					 acpi_power_off, NULL);
 	} else {
 		acpi_no_s5 = true;
 	}

drivers/cpufreq/cppc_cpufreq.c

···
 	}
 };
 
+static struct cpufreq_driver cppc_cpufreq_driver;
+
 #ifdef CONFIG_ACPI_CPPC_CPUFREQ_FIE
 
 /* Frequency invariance support */
···
 static DEFINE_PER_CPU(struct cppc_freq_invariance, cppc_freq_inv);
 static struct kthread_worker *kworker_fie;
 
-static struct cpufreq_driver cppc_cpufreq_driver;
 static unsigned int hisi_cppc_cpufreq_get_rate(unsigned int cpu);
 static int cppc_perf_from_fbctrs(struct cppc_cpudata *cpu_data,
 				 struct cppc_perf_fb_ctrs *fb_ctrs_t0,
···
 	}
 	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
 }
+#else
+static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu)
+{
+	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
+}
+#endif
+
+#if defined(CONFIG_ARM64) && defined(CONFIG_ENERGY_MODEL)
 
 static DEFINE_PER_CPU(unsigned int, efficiency_class);
 static void cppc_cpufreq_register_em(struct cpufreq_policy *policy);
···
 }
 
 #else
-
-static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu)
-{
-	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
-}
 static int populate_efficiency_class(void)
 {
 	return 0;
 }
-static void cppc_cpufreq_register_em(struct cpufreq_policy *policy)
-{
-}
 #endif
-
 
 static struct cppc_cpudata *cppc_cpufreq_get_cpu_data(unsigned int cpu)
 {

drivers/cpufreq/mediatek-cpufreq.c

···
 #include <linux/cpu.h>
 #include <linux/cpufreq.h>
 #include <linux/cpumask.h>
+#include <linux/minmax.h>
 #include <linux/module.h>
 #include <linux/of.h>
+#include <linux/of_platform.h>
 #include <linux/platform_device.h>
 #include <linux/pm_opp.h>
 #include <linux/regulator/consumer.h>
-#include <linux/slab.h>
-#include <linux/thermal.h>
 
-#define MIN_VOLT_SHIFT		(100000)
-#define MAX_VOLT_SHIFT		(200000)
-#define MAX_VOLT_LIMIT		(1150000)
-#define VOLT_TOL		(10000)
+struct mtk_cpufreq_platform_data {
+	int min_volt_shift;
+	int max_volt_shift;
+	int proc_max_volt;
+	int sram_min_volt;
+	int sram_max_volt;
+	bool ccifreq_supported;
+};
 
 /*
  * The struct mtk_cpu_dvfs_info holds necessary information for doing CPU DVFS
···
 struct mtk_cpu_dvfs_info {
 	struct cpumask cpus;
 	struct device *cpu_dev;
+	struct device *cci_dev;
 	struct regulator *proc_reg;
 	struct regulator *sram_reg;
 	struct clk *cpu_clk;
···
 	struct list_head list_head;
 	int intermediate_voltage;
 	bool need_voltage_tracking;
+	int vproc_on_boot;
+	int pre_vproc;
+	/* Avoid race condition for regulators between notify and policy */
+	struct mutex reg_lock;
+	struct notifier_block opp_nb;
+	unsigned int opp_cpu;
+	unsigned long current_freq;
+	const struct mtk_cpufreq_platform_data *soc_data;
+	int vtrack_max;
+	bool ccifreq_bound;
 };
+
+static struct platform_device *cpufreq_pdev;
 
 static LIST_HEAD(dvfs_info_list);
 
···
 static int mtk_cpufreq_voltage_tracking(struct mtk_cpu_dvfs_info *info,
 					int new_vproc)
 {
+	const struct mtk_cpufreq_platform_data *soc_data = info->soc_data;
 	struct regulator *proc_reg = info->proc_reg;
 	struct regulator *sram_reg = info->sram_reg;
-	int old_vproc, old_vsram, new_vsram, vsram, vproc, ret;
+	int pre_vproc, pre_vsram, new_vsram, vsram, vproc, ret;
+	int retry = info->vtrack_max;
 
-	old_vproc = regulator_get_voltage(proc_reg);
-	if (old_vproc < 0) {
-		pr_err("%s: invalid Vproc value: %d\n", __func__, old_vproc);
-		return old_vproc;
+	pre_vproc = regulator_get_voltage(proc_reg);
+	if (pre_vproc < 0) {
+		dev_err(info->cpu_dev,
+			"invalid Vproc value: %d\n", pre_vproc);
+		return pre_vproc;
 	}
-	/* Vsram should not exceed the maximum allowed voltage of SoC. */
-	new_vsram = min(new_vproc + MIN_VOLT_SHIFT, MAX_VOLT_LIMIT);
 
-	if (old_vproc < new_vproc) {
-		/*
-		 * When scaling up voltages, Vsram and Vproc scale up step
-		 * by step. At each step, set Vsram to (Vproc + 200mV) first,
-		 * then set Vproc to (Vsram - 100mV).
-		 * Keep doing it until Vsram and Vproc hit target voltages.
-		 */
-		do {
-			old_vsram = regulator_get_voltage(sram_reg);
-			if (old_vsram < 0) {
-				pr_err("%s: invalid Vsram value: %d\n",
-				       __func__, old_vsram);
-				return old_vsram;
-			}
-			old_vproc = regulator_get_voltage(proc_reg);
-			if (old_vproc < 0) {
-				pr_err("%s: invalid Vproc value: %d\n",
-				       __func__, old_vproc);
-				return old_vproc;
-			}
+	pre_vsram = regulator_get_voltage(sram_reg);
+	if (pre_vsram < 0) {
+		dev_err(info->cpu_dev, "invalid Vsram value: %d\n", pre_vsram);
+		return pre_vsram;
+	}
 
-			vsram = min(new_vsram, old_vproc + MAX_VOLT_SHIFT);
+	new_vsram = clamp(new_vproc + soc_data->min_volt_shift,
+			  soc_data->sram_min_volt, soc_data->sram_max_volt);
 
-			if (vsram + VOLT_TOL >= MAX_VOLT_LIMIT) {
-				vsram = MAX_VOLT_LIMIT;
+	do {
+		if (pre_vproc <= new_vproc) {
+			vsram = clamp(pre_vproc + soc_data->max_volt_shift,
+				      soc_data->sram_min_volt, new_vsram);
+			ret = regulator_set_voltage(sram_reg, vsram,
+						    soc_data->sram_max_volt);
 
-				/*
-				 * If the target Vsram hits the maximum voltage,
-				 * try to set the exact voltage value first.
-				 */
-				ret = regulator_set_voltage(sram_reg, vsram,
-							    vsram);
-				if (ret)
-					ret = regulator_set_voltage(sram_reg,
-							vsram - VOLT_TOL,
-							vsram);
-
-				vproc = new_vproc;
-			} else {
-				ret = regulator_set_voltage(sram_reg, vsram,
-							    vsram + VOLT_TOL);
-
-				vproc = vsram - MIN_VOLT_SHIFT;
-			}
 			if (ret)
 				return ret;
 
+			if (vsram == soc_data->sram_max_volt ||
+			    new_vsram == soc_data->sram_min_volt)
+				vproc = new_vproc;
+			else
+				vproc = vsram - soc_data->min_volt_shift;
+
 			ret = regulator_set_voltage(proc_reg, vproc,
-						    vproc + VOLT_TOL);
+						    soc_data->proc_max_volt);
 			if (ret) {
-				regulator_set_voltage(sram_reg, old_vsram,
-						      old_vsram);
+				regulator_set_voltage(sram_reg, pre_vsram,
+						      soc_data->sram_max_volt);
 				return ret;
 			}
-		} while (vproc < new_vproc || vsram < new_vsram);
-	} else if (old_vproc > new_vproc) {
-		/*
-		 * When scaling down voltages, Vsram and Vproc scale down step
-		 * by step. At each step, set Vproc to (Vsram - 200mV) first,
-		 * then set Vproc to (Vproc + 100mV).
-		 * Keep doing it until Vsram and Vproc hit target voltages.
-		 */
-		do {
-			old_vproc = regulator_get_voltage(proc_reg);
-			if (old_vproc < 0) {
-				pr_err("%s: invalid Vproc value: %d\n",
-				       __func__, old_vproc);
-				return old_vproc;
-			}
-			old_vsram = regulator_get_voltage(sram_reg);
-			if (old_vsram < 0) {
-				pr_err("%s: invalid Vsram value: %d\n",
-				       __func__, old_vsram);
-				return old_vsram;
-			}
-
-			vproc = max(new_vproc, old_vsram - MAX_VOLT_SHIFT);
+		} else if (pre_vproc > new_vproc) {
+			vproc = max(new_vproc,
+				    pre_vsram - soc_data->max_volt_shift);
 			ret = regulator_set_voltage(proc_reg, vproc,
-						    vproc + VOLT_TOL);
+						    soc_data->proc_max_volt);
 			if (ret)
 				return ret;
 
 			if (vproc == new_vproc)
 				vsram = new_vsram;
 			else
-				vsram = max(new_vsram, vproc + MIN_VOLT_SHIFT);
+				vsram = max(new_vsram,
+					    vproc + soc_data->min_volt_shift);
 
-			if (vsram + VOLT_TOL >= MAX_VOLT_LIMIT) {
-				vsram = MAX_VOLT_LIMIT;
-
-				/*
-				 * If the target Vsram hits the maximum voltage,
-				 * try to set the exact voltage value first.
-				 */
-				ret = regulator_set_voltage(sram_reg, vsram,
-							    vsram);
-				if (ret)
-					ret = regulator_set_voltage(sram_reg,
-							vsram - VOLT_TOL,
-							vsram);
-			} else {
-				ret = regulator_set_voltage(sram_reg, vsram,
-							    vsram + VOLT_TOL);
-			}
-
+			ret = regulator_set_voltage(sram_reg, vsram,
+						    soc_data->sram_max_volt);
 			if (ret) {
-				regulator_set_voltage(proc_reg, old_vproc,
-						      old_vproc);
+				regulator_set_voltage(proc_reg, pre_vproc,
+						      soc_data->proc_max_volt);
 				return ret;
 			}
-		} while (vproc > new_vproc + VOLT_TOL ||
-			 vsram > new_vsram + VOLT_TOL);
-	}
+		}
+
+		pre_vproc = vproc;
+		pre_vsram = vsram;
+
+		if (--retry < 0) {
+			dev_err(info->cpu_dev,
+				"over loop count, failed to set voltage\n");
+			return -EINVAL;
+		}
+	} while (vproc != new_vproc || vsram != new_vsram);
 
 	return 0;
 }
 
 static int mtk_cpufreq_set_voltage(struct mtk_cpu_dvfs_info *info, int vproc)
 {
+	const struct mtk_cpufreq_platform_data *soc_data = info->soc_data;
+	int ret;
+
 	if (info->need_voltage_tracking)
-		return mtk_cpufreq_voltage_tracking(info, vproc);
+		ret = mtk_cpufreq_voltage_tracking(info, vproc);
 	else
-		return regulator_set_voltage(info->proc_reg, vproc,
-					     vproc + VOLT_TOL);
+		ret = regulator_set_voltage(info->proc_reg, vproc,
+					    soc_data->proc_max_volt);
+	if (!ret)
+		info->pre_vproc = vproc;
+
+	return ret;
+}
+
+static bool is_ccifreq_ready(struct mtk_cpu_dvfs_info *info)
+{
+	struct device_link *sup_link;
+
+	if (info->ccifreq_bound)
+		return true;
+
+	sup_link = device_link_add(info->cpu_dev, info->cci_dev,
+				   DL_FLAG_AUTOREMOVE_CONSUMER);
+	if (!sup_link) {
+		dev_err(info->cpu_dev, "cpu%d: sup_link is NULL\n", info->opp_cpu);
+		return false;
+	}
+
+	if (sup_link->supplier->links.status != DL_DEV_DRIVER_BOUND)
+		return false;
+
+	info->ccifreq_bound = true;
+
+	return true;
 }
 
 static int mtk_cpufreq_set_target(struct cpufreq_policy *policy,
···
 	struct mtk_cpu_dvfs_info *info = policy->driver_data;
 	struct device *cpu_dev = info->cpu_dev;
 	struct dev_pm_opp *opp;
-	long freq_hz, old_freq_hz;
-	int vproc, old_vproc, inter_vproc, target_vproc, ret;
+	long freq_hz, pre_freq_hz;
+	int vproc, pre_vproc, inter_vproc, target_vproc, ret;
 
 	inter_vproc = info->intermediate_voltage;
 
-	old_freq_hz = clk_get_rate(cpu_clk);
-	old_vproc = regulator_get_voltage(info->proc_reg);
-	if (old_vproc < 0) {
-		pr_err("%s: invalid Vproc value: %d\n", __func__, old_vproc);
-		return old_vproc;
+	pre_freq_hz = clk_get_rate(cpu_clk);
+
+	mutex_lock(&info->reg_lock);
+
+	if (unlikely(info->pre_vproc <= 0))
+		pre_vproc = regulator_get_voltage(info->proc_reg);
+	else
+		pre_vproc = info->pre_vproc;
+
+	if (pre_vproc < 0) {
+		dev_err(cpu_dev, "invalid Vproc value: %d\n", pre_vproc);
+		ret = pre_vproc;
+		goto out;
 	}
 
 	freq_hz = freq_table[index].frequency * 1000;
 
 	opp = dev_pm_opp_find_freq_ceil(cpu_dev, &freq_hz);
 	if (IS_ERR(opp)) {
-		pr_err("cpu%d: failed to find OPP for %ld\n",
-		       policy->cpu, freq_hz);
-		return PTR_ERR(opp);
+		dev_err(cpu_dev, "cpu%d: failed to find OPP for %ld\n",
+			policy->cpu, freq_hz);
+		ret = PTR_ERR(opp);
+		goto out;
 	}
 	vproc = dev_pm_opp_get_voltage(opp);
 	dev_pm_opp_put(opp);
 
 	/*
+	 * If MediaTek cci is supported but is not ready, we will use the value
+	 * of max(target cpu voltage, booting voltage) to prevent high freqeuncy
+	 * low voltage crash.
+	 */
+	if (info->soc_data->ccifreq_supported && !is_ccifreq_ready(info))
+		vproc = max(vproc, info->vproc_on_boot);
+
+	/*
 	 * If the new voltage or the intermediate voltage is higher than the
 	 * current voltage, scale up voltage first.
 	 */
-	target_vproc = (inter_vproc > vproc) ? inter_vproc : vproc;
-	if (old_vproc < target_vproc) {
+	target_vproc = max(inter_vproc, vproc);
+	if (pre_vproc <= target_vproc) {
 		ret = mtk_cpufreq_set_voltage(info, target_vproc);
 		if (ret) {
-			pr_err("cpu%d: failed to scale up voltage!\n",
-			       policy->cpu);
-			mtk_cpufreq_set_voltage(info, old_vproc);
-			return ret;
+			dev_err(cpu_dev,
+				"cpu%d: failed to scale up voltage!\n", policy->cpu);
+			mtk_cpufreq_set_voltage(info, pre_vproc);
+			goto out;
 		}
 	}
 
 	/* Reparent the CPU clock to intermediate clock. */
 	ret = clk_set_parent(cpu_clk, info->inter_clk);
 	if (ret) {
-		pr_err("cpu%d: failed to re-parent cpu clock!\n",
-		       policy->cpu);
-		mtk_cpufreq_set_voltage(info, old_vproc);
-		WARN_ON(1);
-		return ret;
+		dev_err(cpu_dev,
+			"cpu%d: failed to re-parent cpu clock!\n", policy->cpu);
+		mtk_cpufreq_set_voltage(info, pre_vproc);
+		goto out;
 	}
 
 	/* Set the original PLL to target rate. */
 	ret = clk_set_rate(armpll, freq_hz);
 	if (ret) {
-		pr_err("cpu%d: failed to scale cpu clock rate!\n",
-		       policy->cpu);
+		dev_err(cpu_dev,
+			"cpu%d: failed to scale cpu clock rate!\n", policy->cpu);
 		clk_set_parent(cpu_clk, armpll);
-		mtk_cpufreq_set_voltage(info, old_vproc);
-		return ret;
+		mtk_cpufreq_set_voltage(info, pre_vproc);
+		goto out;
 	}
 
 	/* Set parent of CPU clock back to the original PLL. */
 	ret = clk_set_parent(cpu_clk, armpll);
 	if (ret) {
-		pr_err("cpu%d: failed to re-parent cpu clock!\n",
-		       policy->cpu);
+		dev_err(cpu_dev,
+			"cpu%d: failed to re-parent cpu clock!\n", policy->cpu);
 		mtk_cpufreq_set_voltage(info, inter_vproc);
-		WARN_ON(1);
-		return ret;
+		goto out;
 	}
 
 	/*
 	 * If the new voltage is lower than the intermediate voltage or the
 	 * original voltage, scale down to the new voltage.
 	 */
-	if (vproc < inter_vproc || vproc < old_vproc) {
+	if (vproc < inter_vproc || vproc < pre_vproc) {
 		ret = mtk_cpufreq_set_voltage(info, vproc);
 		if (ret) {
-			pr_err("cpu%d: failed to scale down voltage!\n",
-			       policy->cpu);
+			dev_err(cpu_dev,
+				"cpu%d: failed to scale down voltage!\n", policy->cpu);
 			clk_set_parent(cpu_clk, info->inter_clk);
-			clk_set_rate(armpll, old_freq_hz);
+			clk_set_rate(armpll, pre_freq_hz);
 			clk_set_parent(cpu_clk, armpll);
-			return ret;
+			goto out;
 		}
 	}
 
-	return 0;
+	info->current_freq = freq_hz;
+
+out:
+	mutex_unlock(&info->reg_lock);
+
+	return ret;
 }
 
 #define DYNAMIC_POWER "dynamic-power-coefficient"
 
+static int mtk_cpufreq_opp_notifier(struct notifier_block *nb,
+				    unsigned long event, void *data)
+{
+	struct dev_pm_opp *opp = data;
+	struct dev_pm_opp *new_opp;
+	struct mtk_cpu_dvfs_info *info;
+	unsigned long freq, volt;
+	struct cpufreq_policy *policy;
+	int ret = 0;
+
+	info = container_of(nb, struct mtk_cpu_dvfs_info, opp_nb);
+
+	if (event == OPP_EVENT_ADJUST_VOLTAGE) {
+		freq = dev_pm_opp_get_freq(opp);
+
+		mutex_lock(&info->reg_lock);
+		if (info->current_freq == freq) {
+			volt = dev_pm_opp_get_voltage(opp);
+			ret = mtk_cpufreq_set_voltage(info, volt);
+			if (ret)
+				dev_err(info->cpu_dev,
+					"failed to scale voltage: %d\n", ret);
+		}
+		mutex_unlock(&info->reg_lock);
+	} else if (event == OPP_EVENT_DISABLE) {
+		freq = dev_pm_opp_get_freq(opp);
+
+		/* case of current opp item is disabled */
+		if (info->current_freq == freq) {
+			freq = 1;
+			new_opp = dev_pm_opp_find_freq_ceil(info->cpu_dev,
+							    &freq);
+			if (IS_ERR(new_opp)) {
+				dev_err(info->cpu_dev,
+					"all opp items are disabled\n");
+				ret = PTR_ERR(new_opp);
+				return notifier_from_errno(ret);
+			}
+
+			dev_pm_opp_put(new_opp);
+			policy = cpufreq_cpu_get(info->opp_cpu);
+			if (policy) {
+				cpufreq_driver_target(policy, freq / 1000,
+						      CPUFREQ_RELATION_L);
+				cpufreq_cpu_put(policy);
+			}
+		}
+	}
+
+	return notifier_from_errno(ret);
+}
+
+static struct device *of_get_cci(struct device *cpu_dev)
+{
+	struct device_node *np;
+	struct platform_device *pdev;
+
+	np = of_parse_phandle(cpu_dev->of_node, "mediatek,cci", 0);
+	if (IS_ERR_OR_NULL(np))
+		return NULL;
+
+	pdev = of_find_device_by_node(np);
+	of_node_put(np);
+	if (IS_ERR_OR_NULL(pdev))
+		return NULL;
+
+	return &pdev->dev;
+}
+
 static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu)
 {
 	struct device *cpu_dev;
-	struct regulator *proc_reg = ERR_PTR(-ENODEV);
-	struct regulator *sram_reg = ERR_PTR(-ENODEV);
-	struct clk *cpu_clk = ERR_PTR(-ENODEV);
-	struct clk *inter_clk = ERR_PTR(-ENODEV);
 	struct dev_pm_opp *opp;
 	unsigned long rate;
 	int ret;
 
 	cpu_dev = get_cpu_device(cpu);
 	if (!cpu_dev) {
-		pr_err("failed to get cpu%d device\n", cpu);
+		dev_err(cpu_dev, "failed to get cpu%d device\n", cpu);
 		return -ENODEV;
 	}
+	info->cpu_dev = cpu_dev;
 
-	cpu_clk = clk_get(cpu_dev, "cpu");
-	if (IS_ERR(cpu_clk)) {
-		if (PTR_ERR(cpu_clk) == -EPROBE_DEFER)
-			pr_warn("cpu clk for cpu%d not ready, retry.\n", cpu);
-		else
-			pr_err("failed to get cpu clk for cpu%d\n", cpu);
-
-		ret = PTR_ERR(cpu_clk);
-		return ret;
+	info->ccifreq_bound = false;
+	if (info->soc_data->ccifreq_supported) {
+		info->cci_dev = of_get_cci(info->cpu_dev);
+		if (IS_ERR_OR_NULL(info->cci_dev)) {
+			ret = PTR_ERR(info->cci_dev);
+			dev_err(cpu_dev, "cpu%d: failed to get cci device\n", cpu);
+			return -ENODEV;
+		}
 	}
 
-	inter_clk = clk_get(cpu_dev, "intermediate");
-	if (IS_ERR(inter_clk)) {
-		if (PTR_ERR(inter_clk) == -EPROBE_DEFER)
-			pr_warn("intermediate clk for cpu%d not ready, retry.\n",
-				cpu);
-		else
-			pr_err("failed to get intermediate clk for cpu%d\n",
-			       cpu);
+	info->cpu_clk = clk_get(cpu_dev, "cpu");
+	if (IS_ERR(info->cpu_clk)) {
+		ret = PTR_ERR(info->cpu_clk);
+		return dev_err_probe(cpu_dev, ret,
+				     "cpu%d: failed to get cpu clk\n", cpu);
+	}
 
-		ret = PTR_ERR(inter_clk);
+	info->inter_clk = clk_get(cpu_dev, "intermediate");
+	if (IS_ERR(info->inter_clk)) {
+		ret = PTR_ERR(info->inter_clk);
+		dev_err_probe(cpu_dev, ret,
+			      "cpu%d: failed to get intermediate clk\n", cpu);
 		goto out_free_resources;
 	}
 
-	proc_reg = regulator_get_optional(cpu_dev, "proc");
-	if (IS_ERR(proc_reg)) {
-		if (PTR_ERR(proc_reg) == -EPROBE_DEFER)
-			pr_warn("proc regulator for cpu%d not ready, retry.\n",
-				cpu);
-		else
-			pr_err("failed to get proc regulator for cpu%d\n",
-			       cpu);
+	info->proc_reg = regulator_get_optional(cpu_dev, "proc");
+	if (IS_ERR(info->proc_reg)) {
+		ret = PTR_ERR(info->proc_reg);
+		dev_err_probe(cpu_dev, ret,
+			      "cpu%d: failed to get proc regulator\n", cpu);
+		goto out_free_resources;
+	}
 
-		ret = PTR_ERR(proc_reg);
+	ret = regulator_enable(info->proc_reg);
+	if (ret) {
+		dev_warn(cpu_dev, "cpu%d: failed to enable vproc\n", cpu);
 		goto out_free_resources;
 	}
 
 	/* Both presence and absence of sram regulator are valid cases. */
-	sram_reg = regulator_get_exclusive(cpu_dev, "sram");
+	info->sram_reg = regulator_get_optional(cpu_dev, "sram");
+	if (IS_ERR(info->sram_reg))
+		info->sram_reg = NULL;
+	else {
+		ret = regulator_enable(info->sram_reg);
+		if (ret) {
+			dev_warn(cpu_dev, "cpu%d: failed to enable vsram\n", cpu);
+			goto out_free_resources;
+		}
+	}
 
 	/* Get OPP-sharing information from "operating-points-v2" bindings */
 	ret = dev_pm_opp_of_get_sharing_cpus(cpu_dev, &info->cpus);
 	if (ret) {
-		pr_err("failed to get OPP-sharing information for cpu%d\n",
-		       cpu);
+		dev_err(cpu_dev,
+			"cpu%d: failed to get OPP-sharing information\n", cpu);
 		goto out_free_resources;
 	}
 
 	ret = dev_pm_opp_of_cpumask_add_table(&info->cpus);
 	if (ret) {
-		pr_warn("no OPP table for cpu%d\n", cpu);
+		dev_warn(cpu_dev, "cpu%d: no OPP table\n", cpu);
 		goto out_free_resources;
 	}
 
+	ret = clk_prepare_enable(info->cpu_clk);
+	if (ret)
+		goto out_free_opp_table;
+
+	ret = clk_prepare_enable(info->inter_clk);
+	if (ret)
+		goto out_disable_mux_clock;
+
+	if (info->soc_data->ccifreq_supported) {
+		info->vproc_on_boot = regulator_get_voltage(info->proc_reg);
+		if (info->vproc_on_boot < 0) {
+			dev_err(info->cpu_dev,
+				"invalid Vproc value: %d\n", info->vproc_on_boot);
+			goto out_disable_inter_clock;
+		}
+	}
+
 	/* Search a safe voltage for intermediate frequency. */
-	rate = clk_get_rate(inter_clk);
+	rate = clk_get_rate(info->inter_clk);
 	opp = dev_pm_opp_find_freq_ceil(cpu_dev, &rate);
 	if (IS_ERR(opp)) {
-		pr_err("failed to get intermediate opp for cpu%d\n", cpu);
+		dev_err(cpu_dev, "cpu%d: failed to get intermediate opp\n", cpu);
 		ret = PTR_ERR(opp);
-		goto out_free_opp_table;
+		goto out_disable_inter_clock;
 	}
 	info->intermediate_voltage = dev_pm_opp_get_voltage(opp);
 	dev_pm_opp_put(opp);
 
-	info->cpu_dev = cpu_dev;
-	info->proc_reg = proc_reg;
-	info->sram_reg = IS_ERR(sram_reg) ? NULL : sram_reg;
-	info->cpu_clk = cpu_clk;
-	info->inter_clk = inter_clk;
+	mutex_init(&info->reg_lock);
+	info->current_freq = clk_get_rate(info->cpu_clk);
+
+	info->opp_cpu = cpu;
+	info->opp_nb.notifier_call = mtk_cpufreq_opp_notifier;
+	ret = dev_pm_opp_register_notifier(cpu_dev, &info->opp_nb);
+	if (ret) {
+		dev_err(cpu_dev, "cpu%d: failed to register opp notifier\n", cpu);
+		goto out_disable_inter_clock;
+	}
 
 	/*
 	 * If SRAM regulator is present, software "voltage tracking" is needed
 	 * for this CPU power domain.
 	 */
-	info->need_voltage_tracking = !IS_ERR(sram_reg);
+	info->need_voltage_tracking = (info->sram_reg != NULL);
+
+	/*
+	 * We assume min voltage is 0 and tracking target voltage using
+	 * min_volt_shift for each iteration.
+	 * The vtrack_max is 3 times of expeted iteration count.
+	 */
+	info->vtrack_max = 3 * DIV_ROUND_UP(max(info->soc_data->sram_max_volt,
+						info->soc_data->proc_max_volt),
+					    info->soc_data->min_volt_shift);
 
 	return 0;
+
+out_disable_inter_clock:
+	clk_disable_unprepare(info->inter_clk);
+
+out_disable_mux_clock:
+	clk_disable_unprepare(info->cpu_clk);
 
 out_free_opp_table:
 	dev_pm_opp_of_cpumask_remove_table(&info->cpus);
 
 out_free_resources:
-	if (!IS_ERR(proc_reg))
-		regulator_put(proc_reg);
-	if (!IS_ERR(sram_reg))
-		regulator_put(sram_reg);
-	if (!IS_ERR(cpu_clk))
-		clk_put(cpu_clk);
-	if (!IS_ERR(inter_clk))
-		clk_put(inter_clk);
+	if (regulator_is_enabled(info->proc_reg))
+		regulator_disable(info->proc_reg);
+	if (info->sram_reg && regulator_is_enabled(info->sram_reg))
+		regulator_disable(info->sram_reg);
 
-	return ret;
-}
-
-static void mtk_cpu_dvfs_info_release(struct mtk_cpu_dvfs_info *info)
-{
 	if (!IS_ERR(info->proc_reg))
 		regulator_put(info->proc_reg);
 	if (!IS_ERR(info->sram_reg))
···
 	if (!IS_ERR(info->inter_clk))
 		clk_put(info->inter_clk);
 
+	return ret;
+}
+
+static void mtk_cpu_dvfs_info_release(struct mtk_cpu_dvfs_info *info)
+{
+	if (!IS_ERR(info->proc_reg)) {
+		regulator_disable(info->proc_reg);
+		regulator_put(info->proc_reg);
+	}
+	if (!IS_ERR(info->sram_reg)) {
+		regulator_disable(info->sram_reg);
+		regulator_put(info->sram_reg);
+	}
+	if (!IS_ERR(info->cpu_clk)) {
+		clk_disable_unprepare(info->cpu_clk);
+		clk_put(info->cpu_clk);
+	}
+	if (!IS_ERR(info->inter_clk)) {
+		clk_disable_unprepare(info->inter_clk);
+		clk_put(info->inter_clk);
+	}
+
 	dev_pm_opp_of_cpumask_remove_table(&info->cpus);
+	dev_pm_opp_unregister_notifier(info->cpu_dev, &info->opp_nb);
 }
 
 static int mtk_cpufreq_init(struct cpufreq_policy *policy)
···
 	info = mtk_cpu_dvfs_info_lookup(policy->cpu);
 	if (!info) {
 		pr_err("dvfs info for cpu%d is not initialized.\n",
-		       policy->cpu);
+			policy->cpu);
 		return -EINVAL;
 	}
 
 	ret = dev_pm_opp_init_cpufreq_table(info->cpu_dev, &freq_table);
 	if (ret) {
-		pr_err("failed to init cpufreq table for cpu%d: %d\n",
-		       policy->cpu, ret);
+		dev_err(info->cpu_dev,
+			"failed to init cpufreq table for cpu%d: %d\n",
+			policy->cpu, ret);
 		return ret;
 	}
 
···
 
 static int mtk_cpufreq_probe(struct platform_device *pdev)
 {
+	const struct mtk_cpufreq_platform_data *data;
 	struct mtk_cpu_dvfs_info *info, *tmp;
 	int cpu, ret;
+
+	data = dev_get_platdata(&pdev->dev);
+	if (!data) {
+		dev_err(&pdev->dev,
+			"failed to get mtk cpufreq platform data\n");
+		return -ENODEV;
+	}
 
 	for_each_possible_cpu(cpu) {
 		info = mtk_cpu_dvfs_info_lookup(cpu);
···
 			goto release_dvfs_info_list;
 		}
 
+		info->soc_data = data;
 		ret = mtk_cpu_dvfs_info_init(info, cpu);
 		if (ret) {
 			dev_err(&pdev->dev,
···
 	.probe		= mtk_cpufreq_probe,
 };
 
+static const struct mtk_cpufreq_platform_data mt2701_platform_data = {
+	.min_volt_shift = 100000,
+	.max_volt_shift = 200000,
+	.proc_max_volt = 1150000,
+	.sram_min_volt = 0,
+	.sram_max_volt = 1150000,
+	.ccifreq_supported = false,
+};
+
+static const struct mtk_cpufreq_platform_data mt8183_platform_data = {
+	.min_volt_shift = 100000,
+	.max_volt_shift = 200000,
+	.proc_max_volt = 1150000,
+	.sram_min_volt = 0,
+	.sram_max_volt = 1150000,
+	.ccifreq_supported = true,
+};
+
+static const struct mtk_cpufreq_platform_data mt8186_platform_data = {
+	.min_volt_shift = 100000,
+	.max_volt_shift = 250000,
+	.proc_max_volt = 1118750,
+	.sram_min_volt = 850000,
+	.sram_max_volt = 1118750,
+	.ccifreq_supported = true,
+};
+
 /* List of machines supported by this driver */
 static const struct of_device_id mtk_cpufreq_machines[] __initconst = {
-	{ .compatible = "mediatek,mt2701", },
-	{ .compatible = "mediatek,mt2712", },
-	{ .compatible = "mediatek,mt7622", },
-	{ .compatible = "mediatek,mt7623", },
-	{ .compatible = "mediatek,mt8167", },
-	{ .compatible = "mediatek,mt817x", },
-	{ .compatible = "mediatek,mt8173", },
-	{ .compatible = "mediatek,mt8176", },
-	{ .compatible = "mediatek,mt8183", },
-	{ .compatible = "mediatek,mt8365", },
-	{ .compatible = "mediatek,mt8516", },
-
+	{ .compatible = "mediatek,mt2701", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt2712", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt7622", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt7623", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt8167", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt817x", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt8173", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt8176", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt8183", .data = &mt8183_platform_data },
+	{ .compatible = "mediatek,mt8186", .data = &mt8186_platform_data },
+	{ .compatible = "mediatek,mt8365", .data = &mt2701_platform_data },
+	{ .compatible = "mediatek,mt8516", .data = &mt2701_platform_data },
 	{ }
 };
 MODULE_DEVICE_TABLE(of, mtk_cpufreq_machines);
···
 {
 	struct device_node *np;
 	const struct of_device_id *match;
-	struct platform_device *pdev;
+	const struct mtk_cpufreq_platform_data *data;
 	int err;
 
 	np = of_find_node_by_path("/");
···
 		pr_debug("Machine is not compatible with mtk-cpufreq\n");
 		return -ENODEV;
 	}
+	data = match->data;
 
 	err = platform_driver_register(&mtk_cpufreq_platdrv);
 	if (err)
···
 	 * and the device registration codes are put here to handle defer
 	 * probing.
 	 */
-	pdev = platform_device_register_simple("mtk-cpufreq", -1, NULL, 0);
-	if (IS_ERR(pdev)) {
+	cpufreq_pdev = platform_device_register_data(NULL, "mtk-cpufreq", -1,
+						     data, sizeof(*data));
+	if (IS_ERR(cpufreq_pdev)) {
 		pr_err("failed to register mtk-cpufreq platform device\n");
 		platform_driver_unregister(&mtk_cpufreq_platdrv);
-		return PTR_ERR(pdev);
+		return PTR_ERR(cpufreq_pdev);
 	}
 
 	return 0;
 }
-device_initcall(mtk_cpufreq_driver_init);
+module_init(mtk_cpufreq_driver_init)
+
+static void __exit mtk_cpufreq_driver_exit(void)
+{
+	platform_device_unregister(cpufreq_pdev);
+	platform_driver_unregister(&mtk_cpufreq_platdrv);
+}
+module_exit(mtk_cpufreq_driver_exit)
 
 MODULE_DESCRIPTION("MediaTek CPUFreq driver");
 MODULE_AUTHOR("Pi-Cheng Chen <[e-mail]>");

drivers/cpufreq/tegra194-cpufreq.c

···
 // SPDX-License-Identifier: GPL-2.0
 /*
- * Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved
+ * Copyright (c) 2020 - 2022, NVIDIA CORPORATION. All rights reserved
  */
 
 #include <linux/cpu.h>
···
 #define CPUFREQ_TBL_STEP_HZ     (50 * KHZ * KHZ)
 #define MAX_CNT                 ~0U
 
+#define NDIV_MASK              0x1FF
+
+#define CORE_OFFSET(cpu)			(cpu * 8)
+#define CMU_CLKS_BASE				0x2000
+#define SCRATCH_FREQ_CORE_REG(data, cpu)	(data->regs + CMU_CLKS_BASE + CORE_OFFSET(cpu))
+
+#define MMCRAB_CLUSTER_BASE(cl)			(0x30000 + (cl * 0x10000))
+#define CLUSTER_ACTMON_BASE(data, cl) \
+			(data->regs + (MMCRAB_CLUSTER_BASE(cl) + data->soc->actmon_cntr_base))
+#define CORE_ACTMON_CNTR_REG(data, cl, cpu)	(CLUSTER_ACTMON_BASE(data, cl) + CORE_OFFSET(cpu))
+
 /* cpufreq transisition latency */
 #define TEGRA_CPUFREQ_TRANSITION_LATENCY (300 * 1000) /* unit in nanoseconds */
 
···
 	CLUSTER2,
 	CLUSTER3,
 	MAX_CLUSTERS,
-};
-
-struct tegra194_cpufreq_data {
-	void __iomem *regs;
-	size_t num_clusters;
-	struct cpufreq_frequency_table **tables;
 };
 
 struct tegra_cpu_ctr {
···
 	struct tegra_cpu_ctr c;
 };
 
+struct tegra_cpufreq_ops {
+	void (*read_counters)(struct tegra_cpu_ctr *c);
+	void (*set_cpu_ndiv)(struct cpufreq_policy *policy, u64 ndiv);
+	void (*get_cpu_cluster_id)(u32 cpu, u32 *cpuid, u32 *clusterid);
+	int (*get_cpu_ndiv)(u32 cpu, u32 cpuid, u32 clusterid, u64 *ndiv);
+};
+
+struct tegra_cpufreq_soc {
+	struct tegra_cpufreq_ops *ops;
+	int maxcpus_per_cluster;
+	phys_addr_t actmon_cntr_base;
+};
+
+struct tegra194_cpufreq_data {
+	void __iomem *regs;
+	size_t num_clusters;
+	struct cpufreq_frequency_table **tables;
+	const struct tegra_cpufreq_soc *soc;
+};
+
 static struct workqueue_struct *read_counters_wq;
 
-static void get_cpu_cluster(void *cluster)
+static void tegra_get_cpu_mpidr(void *mpidr)
 {
-	u64 mpidr = read_cpuid_mpidr() & MPIDR_HWID_BITMASK;
+	*((u64 *)mpidr) = read_cpuid_mpidr() & MPIDR_HWID_BITMASK;
+}
 
-	*((uint32_t *)cluster) = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+static void tegra234_get_cpu_cluster_id(u32 cpu, u32 *cpuid, u32 *clusterid)
+{
+	u64 mpidr;
+
+	smp_call_function_single(cpu, tegra_get_cpu_mpidr, &mpidr, true);
+
+	if (cpuid)
+		*cpuid = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+	if (clusterid)
+		*clusterid = MPIDR_AFFINITY_LEVEL(mpidr, 2);
+}
+
+static int tegra234_get_cpu_ndiv(u32 cpu, u32 cpuid, u32 clusterid, u64 *ndiv)
+{
+	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
+	void __iomem *freq_core_reg;
+	u64 mpidr_id;
+
+	/* use physical id to get address of per core frequency register */
+	mpidr_id = (clusterid * data->soc->maxcpus_per_cluster) + cpuid;
+	freq_core_reg = SCRATCH_FREQ_CORE_REG(data, mpidr_id);
+
+	*ndiv = readl(freq_core_reg) & NDIV_MASK;
+
+	return 0;
+}
+
+static void tegra234_set_cpu_ndiv(struct cpufreq_policy *policy, u64 ndiv)
+{
+	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
+	void __iomem *freq_core_reg;
+	u32 cpu, cpuid, clusterid;
+	u64 mpidr_id;
+
+	for_each_cpu_and(cpu, policy->cpus, cpu_online_mask) {
+		data->soc->ops->get_cpu_cluster_id(cpu, &cpuid, &clusterid);
+
+		/* use physical id to get address of per core frequency register */
+		mpidr_id = (clusterid * data->soc->maxcpus_per_cluster) + cpuid;
+		freq_core_reg = SCRATCH_FREQ_CORE_REG(data, mpidr_id);
+
+		writel(ndiv, freq_core_reg);
+	}
+}
+
+/*
+ * This register provides access to two counter values with a single
+ * 64-bit read. The counter values are used to determine the average
+ * actual frequency a core has run at over a period of time.
+ *     [63:32] PLLP counter: Counts at fixed frequency (408 MHz)
+ *     [31:0] Core clock counter: Counts on every core clock cycle
+ */
+static void tegra234_read_counters(struct tegra_cpu_ctr *c)
+{
+	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
+	void __iomem *actmon_reg;
+	u32 cpuid, clusterid;
+	u64 val;
+
+	data->soc->ops->get_cpu_cluster_id(c->cpu, &cpuid, &clusterid);
+	actmon_reg = CORE_ACTMON_CNTR_REG(data, clusterid, cpuid);
+
+	val = readq(actmon_reg);
+	c->last_refclk_cnt = upper_32_bits(val);
+	c->last_coreclk_cnt = lower_32_bits(val);
+	udelay(US_DELAY);
+	val = readq(actmon_reg);
+	c->refclk_cnt = upper_32_bits(val);
+	c->coreclk_cnt = lower_32_bits(val);
+}
+
+static struct tegra_cpufreq_ops tegra234_cpufreq_ops = {
+	.read_counters = tegra234_read_counters,
+	.get_cpu_cluster_id = tegra234_get_cpu_cluster_id,
+	.get_cpu_ndiv = tegra234_get_cpu_ndiv,
+	.set_cpu_ndiv = tegra234_set_cpu_ndiv,
+};
+
+const struct tegra_cpufreq_soc tegra234_cpufreq_soc = {
+	.ops = &tegra234_cpufreq_ops,
+	.actmon_cntr_base = 0x9000,
+	.maxcpus_per_cluster = 4,
+};
+
+static void tegra194_get_cpu_cluster_id(u32 cpu, u32 *cpuid, u32 *clusterid)
+{
+	u64 mpidr;
+
+	smp_call_function_single(cpu, tegra_get_cpu_mpidr, &mpidr, true);
+
+	if (cpuid)
+		*cpuid = MPIDR_AFFINITY_LEVEL(mpidr, 0);
+	if (clusterid)
+		*clusterid = MPIDR_AFFINITY_LEVEL(mpidr, 1);
 }
 
 /*
···
 	return nltbl->ref_clk_hz / KHZ * ndiv / (nltbl->pdiv * nltbl->mdiv);
 }
 
+static void tegra194_read_counters(struct tegra_cpu_ctr *c)
+{
+	u64 val;
+
+	val = read_freq_feedback();
+	c->last_refclk_cnt = lower_32_bits(val);
+	c->last_coreclk_cnt = upper_32_bits(val);
+	udelay(US_DELAY);
+	val = read_freq_feedback();
+	c->refclk_cnt = lower_32_bits(val);
+	c->coreclk_cnt = upper_32_bits(val);
+}
+
 static void tegra_read_counters(struct work_struct *work)
 {
+	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
 	struct read_counters_work *read_counters_work;
 	struct tegra_cpu_ctr *c;
-	u64 val;
 
 	/*
 	 * ref_clk_counter(32 bit counter) runs on constant clk,
···
 					  work);
 	c = &read_counters_work->c;
 
-	val = read_freq_feedback();
-	c->last_refclk_cnt = lower_32_bits(val);
-	c->last_coreclk_cnt = upper_32_bits(val);
-	udelay(US_DELAY);
-	val = read_freq_feedback();
-	c->refclk_cnt = lower_32_bits(val);
-	c->coreclk_cnt = upper_32_bits(val);
+	data->soc->ops->read_counters(c);
 }
 
 /*
···
 	return (rate_mhz * KHZ); /* in KHz */
 }
 
-static void get_cpu_ndiv(void *ndiv)
+static void tegra194_get_cpu_ndiv_sysreg(void *ndiv)
 {
 	u64 ndiv_val;
 
···
 	*(u64 *)ndiv = ndiv_val;
 }
 
-static void set_cpu_ndiv(void *data)
+static int tegra194_get_cpu_ndiv(u32 cpu, u32 cpuid, u32 clusterid, u64 *ndiv)
 {
-	struct cpufreq_frequency_table *tbl = data;
-	u64 ndiv_val = (u64)tbl->driver_data;
+	int ret;
+
+	ret = smp_call_function_single(cpu, tegra194_get_cpu_ndiv_sysreg, &ndiv, true);
+
+	return ret;
+}
+
+static void tegra194_set_cpu_ndiv_sysreg(void *data)
+{
+	u64 ndiv_val = *(u64 *)data;
 
 	asm volatile("msr s3_0_c15_c0_4, %0" : : "r" (ndiv_val));
+}
+
+static void tegra194_set_cpu_ndiv(struct cpufreq_policy *policy, u64 ndiv)
+{
+	on_each_cpu_mask(policy->cpus, tegra194_set_cpu_ndiv_sysreg, &ndiv, true);
 }
 
 static unsigned int tegra194_get_speed(u32 cpu)
 {
 	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
 	struct cpufreq_frequency_table *pos;
+	u32 cpuid, clusterid;
 	unsigned int rate;
 	u64 ndiv;
 	int ret;
-	u32 cl;
 
-	smp_call_function_single(cpu, get_cpu_cluster, &cl, true);
+	data->soc->ops->get_cpu_cluster_id(cpu, &cpuid, &clusterid);
 
 	/* reconstruct actual cpu freq using counters */
 	rate = tegra194_calculate_speed(cpu);
 
 	/* get last written ndiv value */
-	ret = smp_call_function_single(cpu, get_cpu_ndiv, &ndiv, true);
+	ret = data->soc->ops->get_cpu_ndiv(cpu, cpuid, clusterid, &ndiv);
 	if (WARN_ON_ONCE(ret))
 		return rate;
 
···
 	 * to the last written ndiv value from freq_table. This is
 	 * done to return consistent value.
 	 */
-	cpufreq_for_each_valid_entry(pos, data->tables[cl]) {
+	cpufreq_for_each_valid_entry(pos, data->tables[clusterid]) {
 		if (pos->driver_data != ndiv)
 			continue;
 
···
 static int tegra194_cpufreq_init(struct cpufreq_policy *policy)
 {
 	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
-	u32 cpu;
-	u32 cl;
+	int maxcpus_per_cluster = data->soc->maxcpus_per_cluster;
+	u32 start_cpu, cpu;
+	u32 clusterid;
 
-	smp_call_function_single(policy->cpu, get_cpu_cluster, &cl, true);
+	data->soc->ops->get_cpu_cluster_id(policy->cpu, NULL, &clusterid);
 
-	if (cl >= data->num_clusters || !data->tables[cl])
+	if (clusterid >= data->num_clusters || !data->tables[clusterid])
 		return -EINVAL;
 
+	start_cpu = rounddown(policy->cpu, maxcpus_per_cluster);
 	/* set same policy for all cpus in a cluster */
-	for (cpu = (cl * 2); cpu < ((cl + 1) * 2); cpu++)
-		cpumask_set_cpu(cpu, policy->cpus);
-
-	policy->freq_table = data->tables[cl];
+	for (cpu = start_cpu; cpu < (start_cpu + maxcpus_per_cluster); cpu++) {
+		if (cpu_possible(cpu))
+			cpumask_set_cpu(cpu, policy->cpus);
+	}
+	policy->freq_table = data->tables[clusterid];
 	policy->cpuinfo.transition_latency = TEGRA_CPUFREQ_TRANSITION_LATENCY;
 
 	return 0;
···
 				       unsigned int index)
 {
 	struct cpufreq_frequency_table *tbl = policy->freq_table + index;
+	struct tegra194_cpufreq_data *data = cpufreq_get_driver_data();
 
 	/*
 	 * Each core writes frequency in per core register. Then both cores
 	 * in a cluster run at same frequency which is the maximum frequency
 	 * request out of the values requested by both cores in that cluster.
 	 */
-	on_each_cpu_mask(policy->cpus, set_cpu_ndiv, tbl, true);
+	data->soc->ops->set_cpu_ndiv(policy, (u64)tbl->driver_data);
 
 	return 0;
 }
···
 	.get = tegra194_get_speed,
 	.init = tegra194_cpufreq_init,
 	.attr = cpufreq_generic_attr,
+};
+
+static struct tegra_cpufreq_ops tegra194_cpufreq_ops = {
+	.read_counters = tegra194_read_counters,
+	.get_cpu_cluster_id = tegra194_get_cpu_cluster_id,
+	.get_cpu_ndiv = tegra194_get_cpu_ndiv,
+	.set_cpu_ndiv = tegra194_set_cpu_ndiv,
+};
+
+const struct tegra_cpufreq_soc tegra194_cpufreq_soc = {
+	.ops = &tegra194_cpufreq_ops,
+	.maxcpus_per_cluster = 2,
 };
 
 static void tegra194_cpufreq_free_resources(void)
···
 
 static int tegra194_cpufreq_probe(struct platform_device *pdev)
 {
+	const struct tegra_cpufreq_soc *soc;
 	struct tegra194_cpufreq_data *data;
 	struct tegra_bpmp *bpmp;
 	int err, i;
···
 	if (!data)
 		return -ENOMEM;
 
+	soc = of_device_get_match_data(&pdev->dev);
+
+	if (soc->ops && soc->maxcpus_per_cluster) {
+		data->soc = soc;
+	} else {
+		dev_err(&pdev->dev, "soc data missing\n");
+		return -EINVAL;
+	}
+
 	data->num_clusters = MAX_CLUSTERS;
 	data->tables = devm_kcalloc(&pdev->dev, data->num_clusters,
 				    sizeof(*data->tables), GFP_KERNEL);
 	if (!data->tables)
 		return -ENOMEM;
+
+	if (soc->actmon_cntr_base) {
+		/* mmio registers are used for frequency request and re-construction */
+		data->regs = devm_platform_ioremap_resource(pdev, 0);
+		if (IS_ERR(data->regs))
+			return PTR_ERR(data->regs);
+	}
 
 	platform_set_drvdata(pdev, data);
 
···
 }
 
 static const struct of_device_id tegra194_cpufreq_of_match[] = {
-	{ .compatible = "nvidia,tegra194-ccplex", },
+	{ .compatible = "nvidia,tegra194-ccplex", .data = &tegra194_cpufreq_soc },
+	{ .compatible = "nvidia,tegra234-ccplex-cluster", .data = &tegra234_cpufreq_soc },
 	{ /* sentinel */ }
 };
-MODULE_DEVICE_TABLE(of, tegra194_cpufreq_of_match);
 
 static struct platform_driver tegra194_ccplex_driver = {
 	.driver = {

drivers/memory/emif.c

···
 		dev_emerg(emif->dev, "SDRAM temperature exceeds operating limit.. Needs shut down!!!\n");
 
 		/* If we have Power OFF ability, use it, else try restarting */
-		if (pm_power_off) {
+		if (kernel_can_power_off()) {
 			kernel_power_off();
 		} else {
 			WARN(1, "FIXME: NO pm_power_off!!! trying restart\n");

drivers/opp/core.c

···
 }
 EXPORT_SYMBOL_GPL(dev_pm_opp_find_freq_exact);
 
-/**
- * dev_pm_opp_find_level_exact() - search for an exact level
- * @dev:		device for which we do this operation
- * @level:		level to search for
- *
- * Return: Searches for exact match in the opp table and returns pointer to the
- * matching opp if found, else returns ERR_PTR in case of error and should
- * be handled using IS_ERR. Error return values can be:
- * EINVAL:	for bad pointer
- * ERANGE:	no match found for search
- * ENODEV:	if device not found in list of registered devices
- *
- * The callers are required to call dev_pm_opp_put() for the returned OPP after
- * use.
- */
-struct dev_pm_opp *dev_pm_opp_find_level_exact(struct device *dev,
-					       unsigned int level)
-{
-	struct opp_table *opp_table;
-	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
-
-	opp_table = _find_opp_table(dev);
-	if (IS_ERR(opp_table)) {
-		int r = PTR_ERR(opp_table);
-
-		dev_err(dev, "%s: OPP table not found (%d)\n", __func__, r);
-		return ERR_PTR(r);
-	}
-
-	mutex_lock(&opp_table->lock);
-
-	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
-		if (temp_opp->level == level) {
-			opp = temp_opp;
-
-			/* Increment the reference count of OPP */
-			dev_pm_opp_get(opp);
-			break;
-		}
-	}
-
-	mutex_unlock(&opp_table->lock);
-	dev_pm_opp_put_opp_table(opp_table);
-
-	return opp;
-}
-EXPORT_SYMBOL_GPL(dev_pm_opp_find_level_exact);
-
-/**
- * dev_pm_opp_find_level_ceil() - search for an rounded up level
- * @dev:		device for which we do this operation
- * @level:		level to search for
- *
- * Return: Searches for rounded up match in the opp table and returns pointer
- * to the  matching opp if found, else returns ERR_PTR in case of error and
- * should be handled using IS_ERR. Error return values can be:
- * EINVAL:	for bad pointer
- * ERANGE:	no match found for search
- * ENODEV:	if device not found in list of registered devices
- *
- * The callers are required to call dev_pm_opp_put() for the returned OPP after
- * use.
- */
-struct dev_pm_opp *dev_pm_opp_find_level_ceil(struct device *dev,
-					      unsigned int *level)
-{
-	struct opp_table *opp_table;
-	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
-
-	opp_table = _find_opp_table(dev);
-	if (IS_ERR(opp_table)) {
-		int r = PTR_ERR(opp_table);
-
-		dev_err(dev, "%s: OPP table not found (%d)\n", __func__, r);
-		return ERR_PTR(r);
-	}
-
-	mutex_lock(&opp_table->lock);
-
-	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
-		if (temp_opp->available && temp_opp->level >= *level) {
-			opp = temp_opp;
-			*level = opp->level;
-
-			/* Increment the reference count of OPP */
-			dev_pm_opp_get(opp);
-			break;
-		}
-	}
-
-	mutex_unlock(&opp_table->lock);
-	dev_pm_opp_put_opp_table(opp_table);
-
-	return opp;
-}
-EXPORT_SYMBOL_GPL(dev_pm_opp_find_level_ceil);
-
 static noinline struct dev_pm_opp *_find_freq_ceil(struct opp_table *opp_table,
 						   unsigned long *freq)
 {
···
 	return opp;
 }
 EXPORT_SYMBOL_GPL(dev_pm_opp_find_freq_ceil_by_volt);
+
+/**
+ * dev_pm_opp_find_level_exact() - search for an exact level
+ * @dev:		device for which we do this operation
+ * @level:		level to search for
+ *
+ * Return: Searches for exact match in the opp table and returns pointer to the
+ * matching opp if found, else returns ERR_PTR in case of error and should
+ * be handled using IS_ERR. Error return values can be:
+ * EINVAL:	for bad pointer
+ * ERANGE:	no match found for search
+ * ENODEV:	if device not found in list of registered devices
+ *
+ * The callers are required to call dev_pm_opp_put() for the returned OPP after
+ * use.
+ */
+struct dev_pm_opp *dev_pm_opp_find_level_exact(struct device *dev,
+					       unsigned int level)
+{
+	struct opp_table *opp_table;
+	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
+
+	opp_table = _find_opp_table(dev);
+	if (IS_ERR(opp_table)) {
+		int r = PTR_ERR(opp_table);
+
+		dev_err(dev, "%s: OPP table not found (%d)\n", __func__, r);
+		return ERR_PTR(r);
+	}
+
+	mutex_lock(&opp_table->lock);
+
+	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
+		if (temp_opp->level == level) {
+			opp = temp_opp;
+
+			/* Increment the reference count of OPP */
+			dev_pm_opp_get(opp);
+			break;
+		}
+	}
+
+	mutex_unlock(&opp_table->lock);
+	dev_pm_opp_put_opp_table(opp_table);
+
+	return opp;
+}
+EXPORT_SYMBOL_GPL(dev_pm_opp_find_level_exact);
+
+/**
+ * dev_pm_opp_find_level_ceil() - search for an rounded up level
+ * @dev:		device for which we do this operation
+ * @level:		level to search for
+ *
+ * Return: Searches for rounded up match in the opp table and returns pointer
+ * to the  matching opp if found, else returns ERR_PTR in case of error and
+ * should be handled using IS_ERR. Error return values can be:
+ * EINVAL:	for bad pointer
+ * ERANGE:	no match found for search
+ * ENODEV:	if device not found in list of registered devices
+ *
+ * The callers are required to call dev_pm_opp_put() for the returned OPP after
+ * use.
+ */
+struct dev_pm_opp *dev_pm_opp_find_level_ceil(struct device *dev,
+					      unsigned int *level)
+{
+	struct opp_table *opp_table;
+	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
+
+	opp_table = _find_opp_table(dev);
+	if (IS_ERR(opp_table)) {
+		int r = PTR_ERR(opp_table);
+
+		dev_err(dev, "%s: OPP table not found (%d)\n", __func__, r);
+		return ERR_PTR(r);
+	}
+
+	mutex_lock(&opp_table->lock);
+
+	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
+		if (temp_opp->available && temp_opp->level >= *level) {
+			opp = temp_opp;
+			*level = opp->level;
+
+			/* Increment the reference count of OPP */
+			dev_pm_opp_get(opp);
+			break;
+		}
+	}
+
+	mutex_unlock(&opp_table->lock);
+	dev_pm_opp_put_opp_table(opp_table);
+
+	return opp;
+}
+EXPORT_SYMBOL_GPL(dev_pm_opp_find_level_ceil);
+
+/**
+ * dev_pm_opp_find_bw_ceil() - Search for a rounded ceil bandwidth
+ * @dev:	device for which we do this operation
+ * @freq:	start bandwidth
+ * @index:	which bandwidth to compare, in case of OPPs with several values
+ *
+ * Search for the matching floor *available* OPP from a starting bandwidth
+ * for a device.
+ *
+ * Return: matching *opp and refreshes *bw accordingly, else returns
+ * ERR_PTR in case of error and should be handled using IS_ERR. Error return
+ * values can be:
+ * EINVAL:	for bad pointer
+ * ERANGE:	no match found for search
+ * ENODEV:	if device not found in list of registered devices
+ *
+ * The callers are required to call dev_pm_opp_put() for the returned OPP after
+ * use.
+ */
+struct dev_pm_opp *dev_pm_opp_find_bw_ceil(struct device *dev,
+					   unsigned int *bw, int index)
+{
+	struct opp_table *opp_table;
+	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
+
+	if (!dev || !bw) {
+		dev_err(dev, "%s: Invalid argument bw=%p\n", __func__, bw);
+		return ERR_PTR(-EINVAL);
+	}
+
+	opp_table = _find_opp_table(dev);
+	if (IS_ERR(opp_table))
+		return ERR_CAST(opp_table);
+
+	if (index >= opp_table->path_count)
+		return ERR_PTR(-EINVAL);
+
+	mutex_lock(&opp_table->lock);
+
+	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
+		if (temp_opp->available && temp_opp->bandwidth) {
+			if (temp_opp->bandwidth[index].peak >= *bw) {
+				opp = temp_opp;
+				*bw = opp->bandwidth[index].peak;
+
+				/* Increment the reference count of OPP */
+				dev_pm_opp_get(opp);
+				break;
+			}
+		}
+	}
+
+	mutex_unlock(&opp_table->lock);
+	dev_pm_opp_put_opp_table(opp_table);
+
+	return opp;
+}
+EXPORT_SYMBOL_GPL(dev_pm_opp_find_bw_ceil);
+
+/**
+ * dev_pm_opp_find_bw_floor() - Search for a rounded floor bandwidth
+ * @dev:	device for which we do this operation
+ * @freq:	start bandwidth
+ * @index:	which bandwidth to compare, in case of OPPs with several values
+ *
+ * Search for the matching floor *available* OPP from a starting bandwidth
+ * for a device.
+ *
+ * Return: matching *opp and refreshes *bw accordingly, else returns
+ * ERR_PTR in case of error and should be handled using IS_ERR. Error return
+ * values can be:
+ * EINVAL:	for bad pointer
+ * ERANGE:	no match found for search
+ * ENODEV:	if device not found in list of registered devices
+ *
+ * The callers are required to call dev_pm_opp_put() for the returned OPP after
+ * use.
+ */
+struct dev_pm_opp *dev_pm_opp_find_bw_floor(struct device *dev,
+					    unsigned int *bw, int index)
+{
+	struct opp_table *opp_table;
+	struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE);
+
+	if (!dev || !bw) {
+		dev_err(dev, "%s: Invalid argument bw=%p\n", __func__, bw);
+		return ERR_PTR(-EINVAL);
+	}
+
+	opp_table = _find_opp_table(dev);
+	if (IS_ERR(opp_table))
+		return ERR_CAST(opp_table);
+
+	if (index >= opp_table->path_count)
+		return ERR_PTR(-EINVAL);
+
+	mutex_lock(&opp_table->lock);
+
+	list_for_each_entry(temp_opp, &opp_table->opp_list, node) {
+		if (temp_opp->available && temp_opp->bandwidth) {
+			/* go to the next node, before choosing prev */
+			if (temp_opp->bandwidth[index].peak > *bw)
+				break;
+			opp = temp_opp;
+		}
+	}
+
+	/* Increment the reference count of OPP */
+	if (!IS_ERR(opp))
+		dev_pm_opp_get(opp);
+	mutex_unlock(&opp_table->lock);
+	dev_pm_opp_put_opp_table(opp_table);
+
+	if (!IS_ERR(opp))
+		*bw = opp->bandwidth[index].peak;
+
+	return opp;
+}
+EXPORT_SYMBOL_GPL(dev_pm_opp_find_bw_floor);
 
 static int _set_opp_voltage(struct device *dev, struct regulator *reg,
 			    struct dev_pm_opp_supply *supply)
···
  */
 void dev_pm_opp_remove(struct device *dev, unsigned long freq)
 {
-	struct dev_pm_opp *opp;
+	struct dev_pm_opp *opp = NULL, *iter;
 	struct opp_table *opp_table;
-	bool found = false;
 
 	opp_table = _find_opp_table(dev);
 	if (IS_ERR(opp_table))
···
 
 	mutex_lock(&opp_table->lock);
 
-	list_for_each_entry(opp, &opp_table->opp_list, node) {
-		if (opp->rate == freq) {
-			found = true;
+	list_for_each_entry(iter, &opp_table->opp_list, node) {
+		if (iter->rate == freq) {
+			opp = iter;
 			break;
 		}
 	}
 
 	mutex_unlock(&opp_table->lock);
 
-	if (found) {
+	if (opp) {
 		dev_pm_opp_put(opp);
 
 		/* Drop the reference taken by dev_pm_opp_add() */
···
 	for (i = 0; i < count; i++) {
 		reg = regulator_get_optional(dev, names[i]);
 		if (IS_ERR(reg)) {
-			ret = PTR_ERR(reg);
-			if (ret != -EPROBE_DEFER)
-				dev_err(dev, "%s: no regulator (%s) found: %d\n",
-					__func__, names[i], ret);
+			ret = dev_err_probe(dev, PTR_ERR(reg),
+					    "%s: no regulator (%s) found\n",
+					    __func__, names[i]);
 			goto free_regulators;
 		}
 
···
 	/* Find clk for the device */
 	opp_table->clk = clk_get(dev, name);
 	if (IS_ERR(opp_table->clk)) {
-		ret = PTR_ERR(opp_table->clk);
-		if (ret != -EPROBE_DEFER) {
-			dev_err(dev, "%s: Couldn't find clock: %d\n", __func__,
-				ret);
-		}
+		ret = dev_err_probe(dev, PTR_ERR(opp_table->clk),
+				    "%s: Couldn't find clock\n", __func__);
 		goto err;
 	}
 

drivers/opp/debugfs.c

···
 static void opp_migrate_dentry(struct opp_device *opp_dev,
 			       struct opp_table *opp_table)
 {
-	struct opp_device *new_dev;
+	struct opp_device *new_dev = NULL, *iter;
 	const struct device *dev;
 	struct dentry *dentry;
 
 	/* Look for next opp-dev */
-	list_for_each_entry(new_dev, &opp_table->dev_list, node)
-		if (new_dev != opp_dev)
+	list_for_each_entry(iter, &opp_table->dev_list, node)
+		if (iter != opp_dev) {
+			new_dev = iter;
 			break;
+		}
+
+	BUG_ON(!new_dev);
 
 	/* new_dev is guaranteed to be valid here */
 	dev = new_dev->dev;

drivers/opp/of.c

···
 
 	/* Checking only first OPP is sufficient */
 	np = of_get_next_available_child(opp_np, NULL);
+	of_node_put(opp_np);
 	if (!np) {
 		dev_err(dev, "OPP table empty\n");
 		return -EINVAL;
 	}
-	of_node_put(opp_np);
 
 	prop = of_find_property(np, "opp-peak-kBps", NULL);
 	of_node_put(np);

drivers/regulator/pfuze100-regulator.c

···
 #include <linux/of_device.h>
 #include <linux/regulator/of_regulator.h>
 #include <linux/platform_device.h>
+#include <linux/reboot.h>
 #include <linux/regulator/driver.h>
 #include <linux/regulator/machine.h>
 #include <linux/regulator/pfuze100.h>
···
 	return pfuze_matches[index].of_node;
 }
 
-static struct pfuze_chip *syspm_pfuze_chip;
-
-static void pfuze_power_off_prepare(void)
+static int pfuze_power_off_prepare(struct sys_off_data *data)
 {
+	struct pfuze_chip *syspm_pfuze_chip = data->cb_data;
+
 	dev_info(syspm_pfuze_chip->dev, "Configure standby mode for power off");
 
 	/* Switch from default mode: APS/APS to APS/Off */
···
 	regmap_update_bits(syspm_pfuze_chip->regmap, PFUZE100_VGEN6VOL,
 			   PFUZE100_VGENxLPWR | PFUZE100_VGENxSTBY,
 			   PFUZE100_VGENxSTBY);
+
+	return NOTIFY_DONE;
 }
 
 static int pfuze_power_off_prepare_init(struct pfuze_chip *pfuze_chip)
 {
+	int err;
+
 	if (pfuze_chip->chip_id != PFUZE100) {
 		dev_warn(pfuze_chip->dev, "Requested pm_power_off_prepare handler for not supported chip\n");
 		return -ENODEV;
 	}
 
-	if (pm_power_off_prepare) {
-		dev_warn(pfuze_chip->dev, "pm_power_off_prepare is already registered.\n");
-		return -EBUSY;
+	err = devm_register_sys_off_handler(pfuze_chip->dev,
+					    SYS_OFF_MODE_POWER_OFF_PREPARE,
+					    SYS_OFF_PRIO_DEFAULT,
+					    pfuze_power_off_prepare,
+					    pfuze_chip);
+	if (err) {
+		dev_err(pfuze_chip->dev, "failed to register sys-off handler: %d\n",
+			err);
+		return err;
 	}
-
-	if (syspm_pfuze_chip) {
-		dev_warn(pfuze_chip->dev, "syspm_pfuze_chip is already set.\n");
-		return -EBUSY;
-	}
-
-	syspm_pfuze_chip = pfuze_chip;
-	pm_power_off_prepare = pfuze_power_off_prepare;
 
 	return 0;
 }
···
 	return 0;
 }
 
-static int pfuze100_regulator_remove(struct i2c_client *client)
-{
-	if (syspm_pfuze_chip) {
-		syspm_pfuze_chip = NULL;
-		pm_power_off_prepare = NULL;
-	}
-
-	return 0;
-}
-
 static struct i2c_driver pfuze_driver = {
 	.driver = {
 		.name = "pfuze100-regulator",
 		.of_match_table = pfuze_dt_ids,
 	},
 	.probe = pfuze100_regulator_probe,
-	.remove = pfuze100_regulator_remove,
 };
 module_i2c_driver(pfuze_driver);
 

drivers/soc/tegra/pmc.c

···
 #include <linux/platform_device.h>
 #include <linux/pm_domain.h>
 #include <linux/pm_opp.h>
+#include <linux/power_supply.h>
 #include <linux/reboot.h>
 #include <linux/regmap.h>
 #include <linux/reset.h>
···
 #define PMC_USB_DEBOUNCE_DEL		0xec
 #define PMC_USB_AO			0xf0
 
+#define PMC_SCRATCH37			0x130
 #define PMC_SCRATCH41			0x140
 
 #define PMC_WAKE2_MASK			0x160
···
 	.notifier_call = tegra_pmc_reboot_notify,
 };
 
-static int tegra_pmc_restart_notify(struct notifier_block *this,
-				    unsigned long action, void *data)
+static void tegra_pmc_restart(void)
 {
 	u32 value;
 
···
 	value = tegra_pmc_readl(pmc, PMC_CNTRL);
 	value |= PMC_CNTRL_MAIN_RST;
 	tegra_pmc_writel(pmc, value, PMC_CNTRL);
+}
+
+static int tegra_pmc_restart_handler(struct sys_off_data *data)
+{
+	tegra_pmc_restart();
 
 	return NOTIFY_DONE;
 }
 
-static struct notifier_block tegra_pmc_restart_handler = {
-	.notifier_call = tegra_pmc_restart_notify,
-	.priority = 128,
-};
+static int tegra_pmc_power_off_handler(struct sys_off_data *data)
+{
+	/*
+	 * Reboot Nexus 7 into special bootloader mode if USB cable is
+	 * connected in order to display battery status and power off.
+	 */
+	if (of_machine_is_compatible("asus,grouper") &&
+	    power_supply_is_system_supplied()) {
+		const u32 go_to_charger_mode = 0xa5a55a5a;
+
+		tegra_pmc_writel(pmc, go_to_charger_mode, PMC_SCRATCH37);
+		tegra_pmc_restart();
+	}
+
+	return NOTIFY_DONE;
+}
 
 static int powergate_show(struct seq_file *s, void *data)
 {
···
 	}
 
 	/*
+	 * PMC should be last resort for restarting since it soft-resets
+	 * CPU without resetting everything else.
+	 */
+	err = devm_register_reboot_notifier(&pdev->dev,
+					    &tegra_pmc_reboot_notifier);
+	if (err) {
+		dev_err(&pdev->dev, "unable to register reboot notifier, %d\n",
+			err);
+		return err;
+	}
+
+	err = devm_register_sys_off_handler(&pdev->dev,
+					    SYS_OFF_MODE_RESTART,
+					    SYS_OFF_PRIO_LOW,
+					    tegra_pmc_restart_handler, NULL);
+	if (err) {
+		dev_err(&pdev->dev, "failed to register sys-off handler: %d\n",
+			err);
+		return err;
+	}
+
+	/*
+	 * PMC should be primary power-off method if it soft-resets CPU,
+	 * asking bootloader to shutdown hardware.
+	 */
+	err = devm_register_sys_off_handler(&pdev->dev,
+					    SYS_OFF_MODE_POWER_OFF,
+					    SYS_OFF_PRIO_FIRMWARE,
+					    tegra_pmc_power_off_handler, NULL);
+	if (err) {
+		dev_err(&pdev->dev, "failed to register sys-off handler: %d\n",
+			err);
+		return err;
+	}
+
+	/*
 	 * PCLK clock rate can't be retrieved using CLK API because it
 	 * causes lockup if CPU enters LP2 idle state from some other
 	 * CLK notifier, hence we're caching the rate's value locally.
···
 			goto cleanup_sysfs;
 	}
 
-	err = devm_register_reboot_notifier(&pdev->dev,
-					    &tegra_pmc_reboot_notifier);
-	if (err) {
-		dev_err(&pdev->dev, "unable to register reboot notifier, %d\n",
-			err);
-		goto cleanup_debugfs;
-	}
-
-	err = register_restart_handler(&tegra_pmc_restart_handler);
-	if (err) {
-		dev_err(&pdev->dev, "unable to register restart handler, %d\n",
-			err);
-		goto cleanup_debugfs;
-	}
-
 	err = tegra_pmc_pinctrl_init(pmc);
 	if (err)
-		goto cleanup_restart_handler;
+		goto cleanup_debugfs;
 
 	err = tegra_pmc_regmap_init(pmc);
 	if (err < 0)
-		goto cleanup_restart_handler;
+		goto cleanup_debugfs;
 
 	err = tegra_powergate_init(pmc, pdev->dev.of_node);
 	if (err < 0)
···
 
 cleanup_powergates:
 	tegra_powergate_remove_all(pdev->dev.of_node);
-cleanup_restart_handler:
-	unregister_restart_handler(&tegra_pmc_restart_handler);
 cleanup_debugfs:
 	debugfs_remove(pmc->debugfs);
 cleanup_sysfs:

include/linux/notifier.h

···
 extern int srcu_notifier_chain_register(struct srcu_notifier_head *nh,
 		struct notifier_block *nb);
 
+extern int atomic_notifier_chain_register_unique_prio(
+		struct atomic_notifier_head *nh, struct notifier_block *nb);
+extern int blocking_notifier_chain_register_unique_prio(
+		struct blocking_notifier_head *nh, struct notifier_block *nb);
+
 extern int atomic_notifier_chain_unregister(struct atomic_notifier_head *nh,
 		struct notifier_block *nb);
 extern int blocking_notifier_chain_unregister(struct blocking_notifier_head *nh,
···
 		unsigned long val_up, unsigned long val_down, void *v);
 extern int raw_notifier_call_chain_robust(struct raw_notifier_head *nh,
 		unsigned long val_up, unsigned long val_down, void *v);
+
+extern bool atomic_notifier_call_chain_is_empty(struct atomic_notifier_head *nh);
 
 #define NOTIFY_DONE		0x0000		/* Don't care */
 #define NOTIFY_OK		0x0001		/* Suits me */

include/linux/pm.h

···
  * Callbacks for platform drivers to implement.
  */
 extern void (*pm_power_off)(void);
-extern void (*pm_power_off_prepare)(void);
 
 struct device; /* we have a circular dep with device.h */
 #ifdef CONFIG_VT_CONSOLE_SLEEP

include/linux/pm_opp.h

···
 struct dev_pm_opp *dev_pm_opp_find_freq_exact(struct device *dev,
 					      unsigned long freq,
 					      bool available);
-struct dev_pm_opp *dev_pm_opp_find_level_exact(struct device *dev,
-					       unsigned int level);
-struct dev_pm_opp *dev_pm_opp_find_level_ceil(struct device *dev,
-					      unsigned int *level);
-
 struct dev_pm_opp *dev_pm_opp_find_freq_floor(struct device *dev,
 					      unsigned long *freq);
 struct dev_pm_opp *dev_pm_opp_find_freq_ceil_by_volt(struct device *dev,
 						     unsigned long u_volt);
 
+struct dev_pm_opp *dev_pm_opp_find_level_exact(struct device *dev,
+					       unsigned int level);
+struct dev_pm_opp *dev_pm_opp_find_level_ceil(struct device *dev,
+					      unsigned int *level);
+
 struct dev_pm_opp *dev_pm_opp_find_freq_ceil(struct device *dev,
 					     unsigned long *freq);
+
+struct dev_pm_opp *dev_pm_opp_find_bw_ceil(struct device *dev,
+					   unsigned int *bw, int index);
+
+struct dev_pm_opp *dev_pm_opp_find_bw_floor(struct device *dev,
+					   unsigned int *bw, int index);
+
 void dev_pm_opp_put(struct dev_pm_opp *opp);
 
 int dev_pm_opp_add(struct device *dev, unsigned long freq,
···
 	return 0;
 }
 
-static inline struct dev_pm_opp *dev_pm_opp_find_freq_exact(struct device *dev,
-					unsigned long freq, bool available)
-{
-	return ERR_PTR(-EOPNOTSUPP);
-}
-
 static inline struct dev_pm_opp *dev_pm_opp_find_level_exact(struct device *dev,
 					unsigned int level)
 {
···
 
 static inline struct dev_pm_opp *dev_pm_opp_find_level_ceil(struct device *dev,
 					unsigned int *level)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline struct dev_pm_opp *dev_pm_opp_find_freq_exact(struct device *dev,
+					unsigned long freq, bool available)
 {
 	return ERR_PTR(-EOPNOTSUPP);
 }
···
 
 static inline struct dev_pm_opp *dev_pm_opp_find_freq_ceil(struct device *dev,
 					unsigned long *freq)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline struct dev_pm_opp *dev_pm_opp_find_bw_ceil(struct device *dev,
+					unsigned int *bw, int index)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline struct dev_pm_opp *dev_pm_opp_find_bw_floor(struct device *dev,
+					unsigned int *bw, int index)
 {
 	return ERR_PTR(-EOPNOTSUPP);
 }

include/linux/reboot.h

···
 #include <uapi/linux/reboot.h>
 
 struct device;
+struct sys_off_handler;
 
 #define SYS_DOWN	0x0001	/* Notify of system down */
 #define SYS_RESTART	SYS_DOWN
···
 struct pt_regs;
 extern void machine_crash_shutdown(struct pt_regs *);
 
+void do_kernel_power_off(void);
+
+/*
+ * sys-off handler API.
+ */
+
+/*
+ * Standard sys-off priority levels. Users are expected to set priorities
+ * relative to the standard levels.
+ *
+ * SYS_OFF_PRIO_PLATFORM:	Use this for platform-level handlers.
+ *
+ * SYS_OFF_PRIO_LOW:		Use this for handler of last resort.
+ *
+ * SYS_OFF_PRIO_DEFAULT:	Use this for normal handlers.
+ *
+ * SYS_OFF_PRIO_HIGH:		Use this for higher priority handlers.
+ *
+ * SYS_OFF_PRIO_FIRMWARE:	Use this if handler uses firmware call.
+ */
+#define SYS_OFF_PRIO_PLATFORM		-256
+#define SYS_OFF_PRIO_LOW		-128
+#define SYS_OFF_PRIO_DEFAULT		0
+#define SYS_OFF_PRIO_HIGH		192
+#define SYS_OFF_PRIO_FIRMWARE		224
+
+enum sys_off_mode {
+	/**
+	 * @SYS_OFF_MODE_POWER_OFF_PREPARE:
+	 *
+	 * Handlers prepare system to be powered off. Handlers are
+	 * allowed to sleep.
+	 */
+	SYS_OFF_MODE_POWER_OFF_PREPARE,
+
+	/**
+	 * @SYS_OFF_MODE_POWER_OFF:
+	 *
+	 * Handlers power-off system. Handlers are disallowed to sleep.
+	 */
+	SYS_OFF_MODE_POWER_OFF,
+
+	/**
+	 * @SYS_OFF_MODE_RESTART:
+	 *
+	 * Handlers restart system. Handlers are disallowed to sleep.
+	 */
+	SYS_OFF_MODE_RESTART,
+};
+
+/**
+ * struct sys_off_data - sys-off callback argument
+ *
+ * @mode: Mode ID. Currently used only by the sys-off restart mode,
+ *        see enum reboot_mode for the available modes.
+ * @cb_data: User's callback data.
+ * @cmd: Command string. Currently used only by the sys-off restart mode,
+ *       NULL otherwise.
+ */
+struct sys_off_data {
+	int mode;
+	void *cb_data;
+	const char *cmd;
+};
+
+struct sys_off_handler *
+register_sys_off_handler(enum sys_off_mode mode,
+			 int priority,
+			 int (*callback)(struct sys_off_data *data),
+			 void *cb_data);
+void unregister_sys_off_handler(struct sys_off_handler *handler);
+
+int devm_register_sys_off_handler(struct device *dev,
+				  enum sys_off_mode mode,
+				  int priority,
+				  int (*callback)(struct sys_off_data *data),
+				  void *cb_data);
+
+int devm_register_power_off_handler(struct device *dev,
+				    int (*callback)(struct sys_off_data *data),
+				    void *cb_data);
+
+int devm_register_restart_handler(struct device *dev,
+				  int (*callback)(struct sys_off_data *data),
+				  void *cb_data);
+
+int register_platform_power_off(void (*power_off)(void));
+void unregister_platform_power_off(void (*power_off)(void));
+
 /*
  * Architecture independent implemenations of sys_reboot commands.
  */
···
 extern void kernel_restart(char *cmd);
 extern void kernel_halt(void);
 extern void kernel_power_off(void);
+extern bool kernel_can_power_off(void);
 
 void ctrl_alt_del(void);
 

kernel/notifier.c

···
  */
 
 static int notifier_chain_register(struct notifier_block **nl,
-				   struct notifier_block *n)
+				   struct notifier_block *n,
+				   bool unique_priority)
 {
 	while ((*nl) != NULL) {
 		if (unlikely((*nl) == n)) {
···
 		}
 		if (n->priority > (*nl)->priority)
 			break;
+		if (n->priority == (*nl)->priority && unique_priority)
+			return -EBUSY;
 		nl = &((*nl)->next);
 	}
 	n->next = *nl;
···
 	int ret;
 
 	spin_lock_irqsave(&nh->lock, flags);
-	ret = notifier_chain_register(&nh->head, n);
+	ret = notifier_chain_register(&nh->head, n, false);
 	spin_unlock_irqrestore(&nh->lock, flags);
 	return ret;
 }
 EXPORT_SYMBOL_GPL(atomic_notifier_chain_register);
+
+/**
+ *	atomic_notifier_chain_register_unique_prio - Add notifier to an atomic notifier chain
+ *	@nh: Pointer to head of the atomic notifier chain
+ *	@n: New entry in notifier chain
+ *
+ *	Adds a notifier to an atomic notifier chain if there is no other
+ *	notifier registered using the same priority.
+ *
+ *	Returns 0 on success, %-EEXIST or %-EBUSY on error.
+ */
+int atomic_notifier_chain_register_unique_prio(struct atomic_notifier_head *nh,
+					       struct notifier_block *n)
+{
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&nh->lock, flags);
+	ret = notifier_chain_register(&nh->head, n, true);
+	spin_unlock_irqrestore(&nh->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(atomic_notifier_chain_register_unique_prio);
 
 /**
  *	atomic_notifier_chain_unregister - Remove notifier from an atomic notifier chain
···
 EXPORT_SYMBOL_GPL(atomic_notifier_call_chain);
 NOKPROBE_SYMBOL(atomic_notifier_call_chain);
 
+/**
+ *	atomic_notifier_call_chain_is_empty - Check whether notifier chain is empty
+ *	@nh: Pointer to head of the atomic notifier chain
+ *
+ *	Checks whether notifier chain is empty.
+ *
+ *	Returns true is notifier chain is empty, false otherwise.
+ */
+bool atomic_notifier_call_chain_is_empty(struct atomic_notifier_head *nh)
+{
+	return !rcu_access_pointer(nh->head);
+}
+
 /*
  *	Blocking notifier chain routines.  All access to the chain is
  *	synchronized by an rwsem.
  */
+
+static int __blocking_notifier_chain_register(struct blocking_notifier_head *nh,
+					      struct notifier_block *n,
+					      bool unique_priority)
+{
+	int ret;
+
+	/*
+	 * This code gets used during boot-up, when task switching is
+	 * not yet working and interrupts must remain disabled.  At
+	 * such times we must not call down_write().
+	 */
+	if (unlikely(system_state == SYSTEM_BOOTING))
+		return notifier_chain_register(&nh->head, n, unique_priority);
+
+	down_write(&nh->rwsem);
+	ret = notifier_chain_register(&nh->head, n, unique_priority);
+	up_write(&nh->rwsem);
+	return ret;
+}
 
 /**
  *	blocking_notifier_chain_register - Add notifier to a blocking notifier chain
···
 int blocking_notifier_chain_register(struct blocking_notifier_head *nh,
 		struct notifier_block *n)
 {
-	int ret;
-
-	/*
-	 * This code gets used during boot-up, when task switching is
-	 * not yet working and interrupts must remain disabled.  At
-	 * such times we must not call down_write().
-	 */
-	if (unlikely(system_state == SYSTEM_BOOTING))
-		return notifier_chain_register(&nh->head, n);
-
-	down_write(&nh->rwsem);
-	ret = notifier_chain_register(&nh->head, n);
-	up_write(&nh->rwsem);
-	return ret;
+	return __blocking_notifier_chain_register(nh, n, false);
 }
 EXPORT_SYMBOL_GPL(blocking_notifier_chain_register);
+
+/**
+ *	blocking_notifier_chain_register_unique_prio - Add notifier to a blocking notifier chain
+ *	@nh: Pointer to head of the blocking notifier chain
+ *	@n: New entry in notifier chain
+ *
+ *	Adds a notifier to an blocking notifier chain if there is no other
+ *	notifier registered using the same priority.
+ *
+ *	Returns 0 on success, %-EEXIST or %-EBUSY on error.
+ */
+int blocking_notifier_chain_register_unique_prio(struct blocking_notifier_head *nh,
+						 struct notifier_block *n)
+{
+	return __blocking_notifier_chain_register(nh, n, true);
+}
+EXPORT_SYMBOL_GPL(blocking_notifier_chain_register_unique_prio);
 
 /**
  *	blocking_notifier_chain_unregister - Remove notifier from a blocking notifier chain
···
 int raw_notifier_chain_register(struct raw_notifier_head *nh,
 		struct notifier_block *n)
 {
-	return notifier_chain_register(&nh->head, n);
+	return notifier_chain_register(&nh->head, n, false);
 }
 EXPORT_SYMBOL_GPL(raw_notifier_chain_register);
 
···
 	 * such times we must not call mutex_lock().
 	 */
 	if (unlikely(system_state == SYSTEM_BOOTING))
-		return notifier_chain_register(&nh->head, n);
+		return notifier_chain_register(&nh->head, n, false);
 
 	mutex_lock(&nh->mutex);
-	ret = notifier_chain_register(&nh->head, n);
+	ret = notifier_chain_register(&nh->head, n, false);
 	mutex_unlock(&nh->mutex);
 	return ret;
 }

kernel/reboot.c

···
 enum reboot_type reboot_type = BOOT_ACPI;
 int reboot_force;
 
-/*
- * If set, this is used for preparing the system to power off.
- */
+struct sys_off_handler {
+	struct notifier_block nb;
+	int (*sys_off_cb)(struct sys_off_data *data);
+	void *cb_data;
+	enum sys_off_mode mode;
+	bool blocking;
+	void *list;
+};
 
-void (*pm_power_off_prepare)(void);
-EXPORT_SYMBOL_GPL(pm_power_off_prepare);
+/*
+ * Temporary stub that prevents linkage failure while we're in process
+ * of removing all uses of legacy pm_power_off() around the kernel.
+ */
+void __weak (*pm_power_off)(void);
 
 /**
  *	emergency_restart - reboot the system
···
 }
 EXPORT_SYMBOL_GPL(kernel_halt);
 
+/*
+ *	Notifier list for kernel code which wants to be called
+ *	to prepare system for power off.
+ */
+static BLOCKING_NOTIFIER_HEAD(power_off_prep_handler_list);
+
+/*
+ *	Notifier list for kernel code which wants to be called
+ *	to power off system.
+ */
+static ATOMIC_NOTIFIER_HEAD(power_off_handler_list);
+
+static int sys_off_notify(struct notifier_block *nb,
+			  unsigned long mode, void *cmd)
+{
+	struct sys_off_handler *handler;
+	struct sys_off_data data = {};
+
+	handler = container_of(nb, struct sys_off_handler, nb);
+	data.cb_data = handler->cb_data;
+	data.mode = mode;
+	data.cmd = cmd;
+
+	return handler->sys_off_cb(&data);
+}
+
+/**
+ *	register_sys_off_handler - Register sys-off handler
+ *	@mode: Sys-off mode
+ *	@priority: Handler priority
+ *	@callback: Callback function
+ *	@cb_data: Callback argument
+ *
+ *	Registers system power-off or restart handler that will be invoked
+ *	at the step corresponding to the given sys-off mode. Handler's callback
+ *	should return NOTIFY_DONE to permit execution of the next handler in
+ *	the call chain or NOTIFY_STOP to break the chain (in error case for
+ *	example).
+ *
+ *	Multiple handlers can be registered at the default priority level.
+ *
+ *	Only one handler can be registered at the non-default priority level,
+ *	otherwise ERR_PTR(-EBUSY) is returned.
+ *
+ *	Returns a new instance of struct sys_off_handler on success, or
+ *	an ERR_PTR()-encoded error code otherwise.
+ */
+struct sys_off_handler *
+register_sys_off_handler(enum sys_off_mode mode,
+			 int priority,
+			 int (*callback)(struct sys_off_data *data),
+			 void *cb_data)
+{
+	struct sys_off_handler *handler;
+	int err;
+
+	handler = kzalloc(sizeof(*handler), GFP_KERNEL);
+	if (!handler)
+		return ERR_PTR(-ENOMEM);
+
+	switch (mode) {
+	case SYS_OFF_MODE_POWER_OFF_PREPARE:
+		handler->list = &power_off_prep_handler_list;
+		handler->blocking = true;
+		break;
+
+	case SYS_OFF_MODE_POWER_OFF:
+		handler->list = &power_off_handler_list;
+		break;
+
+	case SYS_OFF_MODE_RESTART:
+		handler->list = &restart_handler_list;
+		break;
+
+	default:
+		kfree(handler);
+		return ERR_PTR(-EINVAL);
+	}
+
+	handler->nb.notifier_call = sys_off_notify;
+	handler->nb.priority = priority;
+	handler->sys_off_cb = callback;
+	handler->cb_data = cb_data;
+	handler->mode = mode;
+
+	if (handler->blocking) {
+		if (priority == SYS_OFF_PRIO_DEFAULT)
+			err = blocking_notifier_chain_register(handler->list,
+							       &handler->nb);
+		else
+			err = blocking_notifier_chain_register_unique_prio(handler->list,
+									   &handler->nb);
+	} else {
+		if (priority == SYS_OFF_PRIO_DEFAULT)
+			err = atomic_notifier_chain_register(handler->list,
+							     &handler->nb);
+		else
+			err = atomic_notifier_chain_register_unique_prio(handler->list,
+									 &handler->nb);
+	}
+
+	if (err) {
+		kfree(handler);
+		return ERR_PTR(err);
+	}
+
+	return handler;
+}
+EXPORT_SYMBOL_GPL(register_sys_off_handler);
+
+/**
+ *	unregister_sys_off_handler - Unregister sys-off handler
+ *	@handler: Sys-off handler
+ *
+ *	Unregisters given sys-off handler.
+ */
+void unregister_sys_off_handler(struct sys_off_handler *handler)
+{
+	int err;
+
+	if (!handler)
+		return;
+
+	if (handler->blocking)
+		err = blocking_notifier_chain_unregister(handler->list,
+							 &handler->nb);
+	else
+		err = atomic_notifier_chain_unregister(handler->list,
+						       &handler->nb);
+
+	/* sanity check, shall never happen */
+	WARN_ON(err);
+
+	kfree(handler);
+}
+EXPORT_SYMBOL_GPL(unregister_sys_off_handler);
+
+static void devm_unregister_sys_off_handler(void *data)
+{
+	struct sys_off_handler *handler = data;
+
+	unregister_sys_off_handler(handler);
+}
+
+/**
+ *	devm_register_sys_off_handler - Register sys-off handler
+ *	@dev: Device that registers handler
+ *	@mode: Sys-off mode
+ *	@priority: Handler priority
+ *	@callback: Callback function
+ *	@cb_data: Callback argument
+ *
+ *	Registers resource-managed sys-off handler.
+ *
+ *	Returns zero on success, or error code on failure.
+ */
+int devm_register_sys_off_handler(struct device *dev,
+				  enum sys_off_mode mode,
+				  int priority,
+				  int (*callback)(struct sys_off_data *data),
+				  void *cb_data)
+{
+	struct sys_off_handler *handler;
+
+	handler = register_sys_off_handler(mode, priority, callback, cb_data);
+	if (IS_ERR(handler))
+		return PTR_ERR(handler);
+
+	return devm_add_action_or_reset(dev, devm_unregister_sys_off_handler,
+					handler);
+}
+EXPORT_SYMBOL_GPL(devm_register_sys_off_handler);
+
+/**
+ *	devm_register_power_off_handler - Register power-off handler
+ *	@dev: Device that registers callback
+ *	@callback: Callback function
+ *	@cb_data: Callback's argument
+ *
+ *	Registers resource-managed sys-off handler with a default priority
+ *	and using power-off mode.
+ *
+ *	Returns zero on success, or error code on failure.
+ */
+int devm_register_power_off_handler(struct device *dev,
+				    int (*callback)(struct sys_off_data *data),
+				    void *cb_data)
+{
+	return devm_register_sys_off_handler(dev,
+					     SYS_OFF_MODE_POWER_OFF,
+					     SYS_OFF_PRIO_DEFAULT,
+					     callback, cb_data);
+}
+EXPORT_SYMBOL_GPL(devm_register_power_off_handler);
+
+/**
+ *	devm_register_restart_handler - Register restart handler
+ *	@dev: Device that registers callback
+ *	@callback: Callback function
+ *	@cb_data: Callback's argument
+ *
+ *	Registers resource-managed sys-off handler with a default priority
+ *	and using restart mode.
+ *
+ *	Returns zero on success, or error code on failure.
+ */
+int devm_register_restart_handler(struct device *dev,
+				  int (*callback)(struct sys_off_data *data),
+				  void *cb_data)
+{
+	return devm_register_sys_off_handler(dev,
+					     SYS_OFF_MODE_RESTART,
+					     SYS_OFF_PRIO_DEFAULT,
+					     callback, cb_data);
+}
+EXPORT_SYMBOL_GPL(devm_register_restart_handler);
+
+static struct sys_off_handler *platform_power_off_handler;
+
+static int platform_power_off_notify(struct sys_off_data *data)
+{
+	void (*platform_power_power_off_cb)(void) = data->cb_data;
+
+	platform_power_power_off_cb();
+
+	return NOTIFY_DONE;
+}
+
+/**
+ *	register_platform_power_off - Register platform-level power-off callback
+ *	@power_off: Power-off callback
+ *
+ *	Registers power-off callback that will be called as last step
+ *	of the power-off sequence. This callback is expected to be invoked
+ *	for the last resort. Only one platform power-off callback is allowed
+ *	to be registered at a time.
+ *
+ *	Returns zero on success, or error code on failure.
+ */
+int register_platform_power_off(void (*power_off)(void))
+{
+	struct sys_off_handler *handler;
+
+	handler = register_sys_off_handler(SYS_OFF_MODE_POWER_OFF,
+					   SYS_OFF_PRIO_PLATFORM,
+					   platform_power_off_notify,
+					   power_off);
+	if (IS_ERR(handler))
+		return PTR_ERR(handler);
+
+	platform_power_off_handler = handler;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(register_platform_power_off);
+
+/**
+ *	unregister_platform_power_off - Unregister platform-level power-off callback
+ *	@power_off: Power-off callback
+ *
+ *	Unregisters previously registered platform power-off callback.
+ */
+void unregister_platform_power_off(void (*power_off)(void))
+{
+	if (platform_power_off_handler &&
+	    platform_power_off_handler->cb_data == power_off) {
+		unregister_sys_off_handler(platform_power_off_handler);
+		platform_power_off_handler = NULL;
+	}
+}
+EXPORT_SYMBOL_GPL(unregister_platform_power_off);
+
+static int legacy_pm_power_off(struct sys_off_data *data)
+{
+	if (pm_power_off)
+		pm_power_off();
+
+	return NOTIFY_DONE;
+}
+
+static void do_kernel_power_off_prepare(void)
+{
+	blocking_notifier_call_chain(&power_off_prep_handler_list, 0, NULL);
+}
+
+/**
+ *	do_kernel_power_off - Execute kernel power-off handler call chain
+ *
+ *	Expected to be called as last step of the power-off sequence.
+ *
+ *	Powers off the system immediately if a power-off handler function has
+ *	been registered. Otherwise does nothing.
+ */
+void do_kernel_power_off(void)
+{
+	atomic_notifier_call_chain(&power_off_handler_list, 0, NULL);
+}
+
+/**
+ *	kernel_can_power_off - check whether system can be powered off
+ *
+ *	Returns true if power-off handler is registered and system can be
+ *	powered off, false otherwise.
+ */
+bool kernel_can_power_off(void)
+{
+	return !atomic_notifier_call_chain_is_empty(&power_off_handler_list);
+}
+EXPORT_SYMBOL_GPL(kernel_can_power_off);
+
 /**
  *	kernel_power_off - power_off the system
  *
···
 void kernel_power_off(void)
 {
 	kernel_shutdown_prepare(SYSTEM_POWER_OFF);
-	if (pm_power_off_prepare)
-		pm_power_off_prepare();
+	do_kernel_power_off_prepare();
 	migrate_to_reboot_cpu();
 	syscore_shutdown();
 	pr_emerg("Power down\n");
···
 		void __user *, arg)
 {
 	struct pid_namespace *pid_ns = task_active_pid_ns(current);
+	struct sys_off_handler *sys_off = NULL;
 	char buffer[256];
 	int ret = 0;
 
···
 	if (ret)
 		return ret;
 
+	/*
+	 * Register sys-off handlers for legacy PM callback. This allows
+	 * legacy PM callbacks temporary co-exist with the new sys-off API.
+	 *
+	 * TODO: Remove legacy handlers once all legacy PM users will be
+	 *       switched to the sys-off based APIs.
+	 */
+	if (pm_power_off) {
+		sys_off = register_sys_off_handler(SYS_OFF_MODE_POWER_OFF,
+						   SYS_OFF_PRIO_DEFAULT,
+						   legacy_pm_power_off, NULL);
+		if (IS_ERR(sys_off))
+			return PTR_ERR(sys_off);
+	}
+
 	/* Instead of trying to make the power_off code look like
 	 * halt when pm_power_off is not set do it the easy way.
 	 */
-	if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
+	if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !kernel_can_power_off())
 		cmd = LINUX_REBOOT_CMD_HALT;
 
 	mutex_lock(&system_transition_mutex);
···
 		break;
 	}
 	mutex_unlock(&system_transition_mutex);
+	unregister_sys_off_handler(sys_off);
 	return ret;
 }