Merge tag 'pm-6.16-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull more power management updates from Rafael Wysocki:
 "These revert an x86 commit that introduced a nasty power regression on
  some systems, fix PSCI cpuidle driver and ACPI cpufreq driver
  regressions, add Rust abstractions for cpufreq, OPP, clk, and
  cpumasks, add a Rust-based cpufreq-dt driver, and do a minor SCMI
  cpufreq driver cleanup:

   - Revert an x86 commit that went into 6.15 and caused idle power,
     including power in suspend-to-idle, to rise rather dramatically on
     systems booting with "nosmt" in the kernel command line (Rafael
     Wysocki)

   - Prevent freeing an uninitialized pointer in error path of
     dt_idle_state_present() in the PSCI cpuidle driver (Dan Carpenter)

   - Use KHz as the nominal_freq units in get_max_boost_ratio() in the
     ACPI cpufreq driver (iGautham Shenoy)

   - Add Rust abstractions for CPUFreq framework (Viresh Kumar)

   - Add Rust abstractions for OPP framework (Viresh Kumar)

   - Add basic Rust abstractions for Clk and Cpumask frameworks (Viresh
     Kumar)

   - Clean up the SCMI cpufreq driver somewhat (Mike Tipton)"

* tag 'pm-6.16-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (21 commits)
  Revert "x86/smp: Eliminate mwait_play_dead_cpuid_hint()"
  acpi-cpufreq: Fix nominal_freq units to KHz in get_max_boost_ratio()
  rust: opp: Move `cfg(CONFIG_OF)` attribute to the top of doc test
  cpuidle: psci: Fix uninitialized variable in dt_idle_state_present()
  rust: opp: Make the doctest example depend on CONFIG_OF
  cpufreq: scmi: Skip SCMI devices that aren't used by the CPUs
  cpufreq: Add Rust-based cpufreq-dt driver
  rust: opp: Extend OPP abstractions with cpufreq support
  rust: cpufreq: Extend abstractions for driver registration
  rust: cpufreq: Extend abstractions for policy and driver ops
  rust: cpufreq: Add initial abstractions for cpufreq framework
  rust: opp: Add abstractions for the configuration options
  rust: opp: Add abstractions for the OPP table
  rust: opp: Add initial abstractions for OPP framework
  rust: cpu: Add from_cpu()
  rust: macros: enable use of hyphens in module names
  rust: clk: Add initial abstractions
  rust: clk: Add helpers for Rust code
  MAINTAINERS: Add entry for Rust cpumask API
  rust: cpumask: Add initial abstractions
  ...

Author: torvalds

MAINTAINERS

@@ -5952,6 +5952,8 @@ F:	include/dt-bindings/clock/
 F:	include/linux/clk-pr*
 F:	include/linux/clk/
 F:	include/linux/of_clk.h
+F:	rust/helpers/clk.c
+F:	rust/kernel/clk.rs
 X:	drivers/clk/clkdev.c
 
 COMMON INTERNET FILE SYSTEM CLIENT (CIFS and SMB3)
@@ -6211,6 +6213,7 @@ F:	drivers/cpufreq/
 F:	include/linux/cpufreq.h
 F:	include/linux/sched/cpufreq.h
 F:	kernel/sched/cpufreq*.c
+F:	rust/kernel/cpufreq.rs
 F:	tools/testing/selftests/cpufreq/
 
 CPU HOTPLUG
@@ -6224,6 +6227,7 @@ F:	include/linux/cpuhotplug.h
 F:	include/linux/smpboot.h
 F:	kernel/cpu.c
 F:	kernel/smpboot.*
+F:	rust/kernel/cpu.rs
 
 CPU IDLE TIME MANAGEMENT FRAMEWORK
 M:	"Rafael J. Wysocki" <[email protected]>
@@ -6308,6 +6312,12 @@ L:	[email protected]
 S:	Maintained
 F:	drivers/cpuidle/cpuidle-riscv-sbi.c
 
+CPUMASK API [RUST]
+M:	Viresh Kumar <[email protected]>
+R:	Yury Norov <[email protected]>
+S:	Maintained
+F:	rust/kernel/cpumask.rs
+
 CRAMFS FILESYSTEM
 M:	Nicolas Pitre <[email protected]>
 S:	Maintained
@@ -18514,6 +18524,7 @@ F:	Documentation/devicetree/bindings/opp/
 F:	Documentation/power/opp.rst
 F:	drivers/opp/
 F:	include/linux/pm_opp.h
+F:	rust/kernel/opp.rs
 
 OPL4 DRIVER
 M:	Clemens Ladisch <[email protected]>

arch/x86/kernel/smpboot.c

@@ -1244,10 +1244,6 @@ void play_dead_common(void)
 	local_irq_disable();
 }
 
-/*
- * We need to flush the caches before going to sleep, lest we have
- * dirty data in our caches when we come back up.
- */
 void __noreturn mwait_play_dead(unsigned int eax_hint)
 {
 	struct mwait_cpu_dead *md = this_cpu_ptr(&mwait_cpu_dead);
@@ -1293,6 +1289,50 @@ void __noreturn mwait_play_dead(unsigned int eax_hint)
 	}
 }
 
+/*
+ * We need to flush the caches before going to sleep, lest we have
+ * dirty data in our caches when we come back up.
+ */
+static inline void mwait_play_dead_cpuid_hint(void)
+{
+	unsigned int eax, ebx, ecx, edx;
+	unsigned int highest_cstate = 0;
+	unsigned int highest_subcstate = 0;
+	int i;
+
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD ||
+	    boot_cpu_data.x86_vendor == X86_VENDOR_HYGON)
+		return;
+	if (!this_cpu_has(X86_FEATURE_MWAIT))
+		return;
+	if (!this_cpu_has(X86_FEATURE_CLFLUSH))
+		return;
+
+	eax = CPUID_LEAF_MWAIT;
+	ecx = 0;
+	native_cpuid(&eax, &ebx, &ecx, &edx);
+
+	/*
+	 * eax will be 0 if EDX enumeration is not valid.
+	 * Initialized below to cstate, sub_cstate value when EDX is valid.
+	 */
+	if (!(ecx & CPUID5_ECX_EXTENSIONS_SUPPORTED)) {
+		eax = 0;
+	} else {
+		edx >>= MWAIT_SUBSTATE_SIZE;
+		for (i = 0; i < 7 && edx; i++, edx >>= MWAIT_SUBSTATE_SIZE) {
+			if (edx & MWAIT_SUBSTATE_MASK) {
+				highest_cstate = i;
+				highest_subcstate = edx & MWAIT_SUBSTATE_MASK;
+			}
+		}
+		eax = (highest_cstate << MWAIT_SUBSTATE_SIZE) |
+			(highest_subcstate - 1);
+	}
+
+	mwait_play_dead(eax);
+}
+
 /*
  * Kick all "offline" CPUs out of mwait on kexec(). See comment in
  * mwait_play_dead().
@@ -1343,9 +1383,9 @@ void native_play_dead(void)
 	play_dead_common();
 	tboot_shutdown(TB_SHUTDOWN_WFS);
 
-	/* Below returns only on error. */
-	cpuidle_play_dead();
-	hlt_play_dead();
+	mwait_play_dead_cpuid_hint();
+	if (cpuidle_play_dead())
+		hlt_play_dead();
 }
 
 #else /* ... !CONFIG_HOTPLUG_CPU */

drivers/cpufreq/Kconfig

@@ -217,6 +217,18 @@ config CPUFREQ_DT
 
 	  If in doubt, say N.
 
+config CPUFREQ_DT_RUST
+	tristate "Rust based Generic DT based cpufreq driver"
+	depends on HAVE_CLK && OF && RUST
+	select CPUFREQ_DT_PLATDEV
+	select PM_OPP
+	help
+	  This adds a Rust based generic DT based cpufreq driver for frequency
+	  management.  It supports both uniprocessor (UP) and symmetric
+	  multiprocessor (SMP) systems.
+
+	  If in doubt, say N.
+
 config CPUFREQ_VIRT
 	tristate "Virtual cpufreq driver"
 	depends on GENERIC_ARCH_TOPOLOGY

drivers/cpufreq/Makefile

@@ -15,6 +15,7 @@ obj-$(CONFIG_CPU_FREQ_GOV_COMMON)		+= cpufreq_governor.o
 obj-$(CONFIG_CPU_FREQ_GOV_ATTR_SET)	+= cpufreq_governor_attr_set.o
 
 obj-$(CONFIG_CPUFREQ_DT)		+= cpufreq-dt.o
+obj-$(CONFIG_CPUFREQ_DT_RUST)		+= rcpufreq_dt.o
 obj-$(CONFIG_CPUFREQ_DT_PLATDEV)	+= cpufreq-dt-platdev.o
 obj-$(CONFIG_CPUFREQ_VIRT)		+= virtual-cpufreq.o
 

drivers/cpufreq/acpi-cpufreq.c

@@ -660,7 +660,7 @@ static u64 get_max_boost_ratio(unsigned int cpu, u64 *nominal_freq)
 	nominal_perf = perf_caps.nominal_perf;
 
 	if (nominal_freq)
-		*nominal_freq = perf_caps.nominal_freq;
+		*nominal_freq = perf_caps.nominal_freq * 1000;
 
 	if (!highest_perf || !nominal_perf) {
 		pr_debug("CPU%d: highest or nominal performance missing\n", cpu);

drivers/cpufreq/rcpufreq_dt.rs

@@ -0,0 +1,226 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! Rust based implementation of the cpufreq-dt driver.
+
+use kernel::{
+    c_str,
+    clk::Clk,
+    cpu, cpufreq,
+    cpumask::CpumaskVar,
+    device::{Core, Device},
+    error::code::*,
+    fmt,
+    macros::vtable,
+    module_platform_driver, of, opp, platform,
+    prelude::*,
+    str::CString,
+    sync::Arc,
+};
+
+/// Finds exact supply name from the OF node.
+fn find_supply_name_exact(dev: &Device, name: &str) -> Option<CString> {
+    let prop_name = CString::try_from_fmt(fmt!("{}-supply", name)).ok()?;
+    dev.property_present(&prop_name)
+        .then(|| CString::try_from_fmt(fmt!("{name}")).ok())
+        .flatten()
+}
+
+/// Finds supply name for the CPU from DT.
+fn find_supply_names(dev: &Device, cpu: u32) -> Option<KVec<CString>> {
+    // Try "cpu0" for older DTs, fallback to "cpu".
+    let name = (cpu == 0)
+        .then(|| find_supply_name_exact(dev, "cpu0"))
+        .flatten()
+        .or_else(|| find_supply_name_exact(dev, "cpu"))?;
+
+    let mut list = KVec::with_capacity(1, GFP_KERNEL).ok()?;
+    list.push(name, GFP_KERNEL).ok()?;
+
+    Some(list)
+}
+
+/// Represents the cpufreq dt device.
+struct CPUFreqDTDevice {
+    opp_table: opp::Table,
+    freq_table: opp::FreqTable,
+    _mask: CpumaskVar,
+    _token: Option<opp::ConfigToken>,
+    _clk: Clk,
+}
+
+#[derive(Default)]
+struct CPUFreqDTDriver;
+
+#[vtable]
+impl opp::ConfigOps for CPUFreqDTDriver {}
+
+#[vtable]
+impl cpufreq::Driver for CPUFreqDTDriver {
+    const NAME: &'static CStr = c_str!("cpufreq-dt");
+    const FLAGS: u16 = cpufreq::flags::NEED_INITIAL_FREQ_CHECK | cpufreq::flags::IS_COOLING_DEV;
+    const BOOST_ENABLED: bool = true;
+
+    type PData = Arc<CPUFreqDTDevice>;
+
+    fn init(policy: &mut cpufreq::Policy) -> Result<Self::PData> {
+        let cpu = policy.cpu();
+        // SAFETY: The CPU device is only used during init; it won't get hot-unplugged. The cpufreq
+        // core  registers with CPU notifiers and the cpufreq core/driver won't use the CPU device,
+        // once the CPU is hot-unplugged.
+        let dev = unsafe { cpu::from_cpu(cpu)? };
+        let mut mask = CpumaskVar::new_zero(GFP_KERNEL)?;
+
+        mask.set(cpu);
+
+        let token = find_supply_names(dev, cpu)
+            .map(|names| {
+                opp::Config::<Self>::new()
+                    .set_regulator_names(names)?
+                    .set(dev)
+            })
+            .transpose()?;
+
+        // Get OPP-sharing information from "operating-points-v2" bindings.
+        let fallback = match opp::Table::of_sharing_cpus(dev, &mut mask) {
+            Ok(()) => false,
+            Err(e) if e == ENOENT => {
+                // "operating-points-v2" not supported. If the platform hasn't
+                // set sharing CPUs, fallback to all CPUs share the `Policy`
+                // for backward compatibility.
+                opp::Table::sharing_cpus(dev, &mut mask).is_err()
+            }
+            Err(e) => return Err(e),
+        };
+
+        // Initialize OPP tables for all policy cpus.
+        //
+        // For platforms not using "operating-points-v2" bindings, we do this
+        // before updating policy cpus. Otherwise, we will end up creating
+        // duplicate OPPs for the CPUs.
+        //
+        // OPPs might be populated at runtime, don't fail for error here unless
+        // it is -EPROBE_DEFER.
+        let mut opp_table = match opp::Table::from_of_cpumask(dev, &mut mask) {
+            Ok(table) => table,
+            Err(e) => {
+                if e == EPROBE_DEFER {
+                    return Err(e);
+                }
+
+                // The table is added dynamically ?
+                opp::Table::from_dev(dev)?
+            }
+        };
+
+        // The OPP table must be initialized, statically or dynamically, by this point.
+        opp_table.opp_count()?;
+
+        // Set sharing cpus for fallback scenario.
+        if fallback {
+            mask.setall();
+            opp_table.set_sharing_cpus(&mut mask)?;
+        }
+
+        let mut transition_latency = opp_table.max_transition_latency_ns() as u32;
+        if transition_latency == 0 {
+            transition_latency = cpufreq::ETERNAL_LATENCY_NS;
+        }
+
+        policy
+            .set_dvfs_possible_from_any_cpu(true)
+            .set_suspend_freq(opp_table.suspend_freq())
+            .set_transition_latency_ns(transition_latency);
+
+        let freq_table = opp_table.cpufreq_table()?;
+        // SAFETY: The `freq_table` is not dropped while it is getting used by the C code.
+        unsafe { policy.set_freq_table(&freq_table) };
+
+        // SAFETY: The returned `clk` is not dropped while it is getting used by the C code.
+        let clk = unsafe { policy.set_clk(dev, None)? };
+
+        mask.copy(policy.cpus());
+
+        Ok(Arc::new(
+            CPUFreqDTDevice {
+                opp_table,
+                freq_table,
+                _mask: mask,
+                _token: token,
+                _clk: clk,
+            },
+            GFP_KERNEL,
+        )?)
+    }
+
+    fn exit(_policy: &mut cpufreq::Policy, _data: Option<Self::PData>) -> Result {
+        Ok(())
+    }
+
+    fn online(_policy: &mut cpufreq::Policy) -> Result {
+        // We did light-weight tear down earlier, nothing to do here.
+        Ok(())
+    }
+
+    fn offline(_policy: &mut cpufreq::Policy) -> Result {
+        // Preserve policy->data and don't free resources on light-weight
+        // tear down.
+        Ok(())
+    }
+
+    fn suspend(policy: &mut cpufreq::Policy) -> Result {
+        policy.generic_suspend()
+    }
+
+    fn verify(data: &mut cpufreq::PolicyData) -> Result {
+        data.generic_verify()
+    }
+
+    fn target_index(policy: &mut cpufreq::Policy, index: cpufreq::TableIndex) -> Result {
+        let Some(data) = policy.data::<Self::PData>() else {
+            return Err(ENOENT);
+        };
+
+        let freq = data.freq_table.freq(index)?;
+        data.opp_table.set_rate(freq)
+    }
+
+    fn get(policy: &mut cpufreq::Policy) -> Result<u32> {
+        policy.generic_get()
+    }
+
+    fn set_boost(_policy: &mut cpufreq::Policy, _state: i32) -> Result {
+        Ok(())
+    }
+
+    fn register_em(policy: &mut cpufreq::Policy) {
+        policy.register_em_opp()
+    }
+}
+
+kernel::of_device_table!(
+    OF_TABLE,
+    MODULE_OF_TABLE,
+    <CPUFreqDTDriver as platform::Driver>::IdInfo,
+    [(of::DeviceId::new(c_str!("operating-points-v2")), ())]
+);
+
+impl platform::Driver for CPUFreqDTDriver {
+    type IdInfo = ();
+    const OF_ID_TABLE: Option<of::IdTable<Self::IdInfo>> = Some(&OF_TABLE);
+
+    fn probe(
+        pdev: &platform::Device<Core>,
+        _id_info: Option<&Self::IdInfo>,
+    ) -> Result<Pin<KBox<Self>>> {
+        cpufreq::Registration::<CPUFreqDTDriver>::new_foreign_owned(pdev.as_ref())?;
+        Ok(KBox::new(Self {}, GFP_KERNEL)?.into())
+    }
+}
+
+module_platform_driver! {
+    type: CPUFreqDTDriver,
+    name: "cpufreq-dt",
+    author: "Viresh Kumar <[email protected]>",
+    description: "Generic CPUFreq DT driver",
+    license: "GPL v2",
+}