From dab20177b626198603176604c0f85ea1d67044ef Mon Sep 17 00:00:00 2001 From: Rafael J. Wysocki Date: Mon, 29 Jun 2020 13:58:28 +0200 Subject: intel_idle: Eliminate redundant static variable The value of the lapic_timer_always_reliable static variable in the intel_idle driver reflects the boot_cpu_has(X86_FEATURE_ARAT) value and so it also reflects the static_cpu_has(X86_FEATURE_ARAT) value. Hence, the lapic_timer_always_reliable check in intel_idle() is redundant and apart from this lapic_timer_always_reliable is only used in two places in which boot_cpu_has(X86_FEATURE_ARAT) can be used directly. Eliminate the lapic_timer_always_reliable variable in accordance with the above observations. No intentional functional impact. Signed-off-by: Rafael J. Wysocki --- drivers/idle/intel_idle.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) (limited to 'drivers') diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index f4495841bf68..fa23a7ea01ac 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -66,8 +66,6 @@ static struct cpuidle_device __percpu *intel_idle_cpuidle_devices; static unsigned long auto_demotion_disable_flags; static bool disable_promotion_to_c1e; -static bool lapic_timer_always_reliable; - struct idle_cpu { struct cpuidle_state *state_table; @@ -142,7 +140,7 @@ static __cpuidle int intel_idle(struct cpuidle_device *dev, if (state->flags & CPUIDLE_FLAG_TLB_FLUSHED) leave_mm(cpu); - if (!static_cpu_has(X86_FEATURE_ARAT) && !lapic_timer_always_reliable) { + if (!static_cpu_has(X86_FEATURE_ARAT)) { /* * Switch over to one-shot tick broadcast if the target C-state * is deeper than C1. @@ -1562,7 +1560,7 @@ static int intel_idle_cpu_online(unsigned int cpu) { struct cpuidle_device *dev; - if (!lapic_timer_always_reliable) + if (!boot_cpu_has(X86_FEATURE_ARAT)) tick_broadcast_enable(); /* @@ -1655,16 +1653,13 @@ static int __init intel_idle_init(void) goto init_driver_fail; } - if (boot_cpu_has(X86_FEATURE_ARAT)) /* Always Reliable APIC Timer */ - lapic_timer_always_reliable = true; - retval = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "idle/intel:online", intel_idle_cpu_online, NULL); if (retval < 0) goto hp_setup_fail; pr_debug("Local APIC timer is reliable in %s\n", - lapic_timer_always_reliable ? "all C-states" : "C1"); + boot_cpu_has(X86_FEATURE_ARAT) ? "all C-states" : "C1"); return 0; -- cgit v1.2.3 From 0317561912d90cf06457c255351549576953704d Mon Sep 17 00:00:00 2001 From: Ulf Hansson Date: Tue, 7 Jul 2020 14:58:00 +0200 Subject: cpuidle: psci: Split into two separate build objects The combined build object for the PSCI cpuidle driver and the PSCI PM domain, is a bit messy. Therefore let's split it up by adding a new Kconfig ARM_PSCI_CPUIDLE_DOMAIN and convert into two separate objects. Reviewed-by: Lina Iyer Reviewed-by: Sudeep Holla Signed-off-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki --- drivers/cpuidle/Kconfig.arm | 10 ++++++++++ drivers/cpuidle/Makefile | 5 ++--- drivers/cpuidle/cpuidle-psci.h | 2 +- 3 files changed, 13 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/cpuidle/Kconfig.arm b/drivers/cpuidle/Kconfig.arm index 51a7e89085c0..0844fadc4be8 100644 --- a/drivers/cpuidle/Kconfig.arm +++ b/drivers/cpuidle/Kconfig.arm @@ -23,6 +23,16 @@ config ARM_PSCI_CPUIDLE It provides an idle driver that is capable of detecting and managing idle states through the PSCI firmware interface. +config ARM_PSCI_CPUIDLE_DOMAIN + bool "PSCI CPU idle Domain" + depends on ARM_PSCI_CPUIDLE + depends on PM_GENERIC_DOMAINS_OF + default y + help + Select this to enable the PSCI based CPUidle driver to use PM domains, + which is needed to support the hierarchical DT based layout of the + idle states. + config ARM_BIG_LITTLE_CPUIDLE bool "Support for ARM big.LITTLE processors" depends on ARCH_VEXPRESS_TC2_PM || ARCH_EXYNOS || COMPILE_TEST diff --git a/drivers/cpuidle/Makefile b/drivers/cpuidle/Makefile index f07800cbb43f..26bbc5e74123 100644 --- a/drivers/cpuidle/Makefile +++ b/drivers/cpuidle/Makefile @@ -21,9 +21,8 @@ obj-$(CONFIG_ARM_U8500_CPUIDLE) += cpuidle-ux500.o obj-$(CONFIG_ARM_AT91_CPUIDLE) += cpuidle-at91.o obj-$(CONFIG_ARM_EXYNOS_CPUIDLE) += cpuidle-exynos.o obj-$(CONFIG_ARM_CPUIDLE) += cpuidle-arm.o -obj-$(CONFIG_ARM_PSCI_CPUIDLE) += cpuidle_psci.o -cpuidle_psci-y := cpuidle-psci.o -cpuidle_psci-$(CONFIG_PM_GENERIC_DOMAINS_OF) += cpuidle-psci-domain.o +obj-$(CONFIG_ARM_PSCI_CPUIDLE) += cpuidle-psci.o +obj-$(CONFIG_ARM_PSCI_CPUIDLE_DOMAIN) += cpuidle-psci-domain.o obj-$(CONFIG_ARM_TEGRA_CPUIDLE) += cpuidle-tegra.o obj-$(CONFIG_ARM_QCOM_SPM_CPUIDLE) += cpuidle-qcom-spm.o diff --git a/drivers/cpuidle/cpuidle-psci.h b/drivers/cpuidle/cpuidle-psci.h index 7299a04dd467..ac8170684d4f 100644 --- a/drivers/cpuidle/cpuidle-psci.h +++ b/drivers/cpuidle/cpuidle-psci.h @@ -8,7 +8,7 @@ struct device_node; void psci_set_domain_state(u32 state); int __init psci_dt_parse_state_node(struct device_node *np, u32 *state); -#ifdef CONFIG_PM_GENERIC_DOMAINS_OF +#ifdef CONFIG_ARM_PSCI_CPUIDLE_DOMAIN struct device __init *psci_dt_attach_cpu(int cpu); #else static inline struct device __init *psci_dt_attach_cpu(int cpu) { return NULL; } -- cgit v1.2.3 From 4b072cd6ace3fcc5fd0688980a74592a7ddee5e5 Mon Sep 17 00:00:00 2001 From: Ulf Hansson Date: Tue, 7 Jul 2020 14:58:01 +0200 Subject: cpuidle: psci: Fail cpuidle registration if set OSI mode failed Currently we allow the cpuidle driver registration to succeed, even if we failed to enable the OSI mode when the hierarchical DT layout is used. This means running in a degraded mode, by using the available idle states per CPU, while also preventing the domain idle states. Moving forward, this behaviour looks quite questionable to maintain, as complexity seems to grow around it, especially when trying to add support for deferred probe, for example. Therefore, let's make the cpuidle driver registration to fail in this situation, thus relying on the default architectural cpuidle backend for WFI to be used. Reviewed-by: Lina Iyer Signed-off-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki --- drivers/cpuidle/cpuidle-psci-domain.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'drivers') diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c index 423f03bbeb74..f07786aad673 100644 --- a/drivers/cpuidle/cpuidle-psci-domain.c +++ b/drivers/cpuidle/cpuidle-psci-domain.c @@ -26,7 +26,6 @@ struct psci_pd_provider { }; static LIST_HEAD(psci_pd_providers); -static bool osi_mode_enabled __initdata; static int psci_pd_power_off(struct generic_pm_domain *pd) { @@ -272,7 +271,6 @@ static int __init psci_idle_init_domains(void) goto remove_pd; } - osi_mode_enabled = true; of_node_put(np); pr_info("Initialized CPU PM domain topology\n"); return pd_count; @@ -293,9 +291,6 @@ struct device __init *psci_dt_attach_cpu(int cpu) { struct device *dev; - if (!osi_mode_enabled) - return NULL; - dev = dev_pm_domain_attach_by_name(get_cpu_device(cpu), "psci"); if (IS_ERR_OR_NULL(dev)) return dev; -- cgit v1.2.3 From 166bf83529c47ddcd4854ff2047f4f980dc492c8 Mon Sep 17 00:00:00 2001 From: Ulf Hansson Date: Tue, 7 Jul 2020 14:58:02 +0200 Subject: cpuidle: psci: Fix error path via converting to a platform driver The current error paths for the cpuidle-psci driver, may leak memory or possibly leave CPU devices attached to their PM domains. These are quite harmless issues, but still deserves to be taken care of. Although, rather than fixing them by keeping track of allocations that needs to be freed, which tends to become a bit messy, let's convert into a platform driver. In this way, it gets easier to fix the memory leaks as we can rely on the devm_* functions. Moreover, converting to a platform driver also enables support for deferred probe, which subsequent changes takes benefit from. Signed-off-by: Ulf Hansson Reviewed-by: Lukasz Luba Signed-off-by: Rafael J. Wysocki --- drivers/cpuidle/cpuidle-psci-domain.c | 10 ++- drivers/cpuidle/cpuidle-psci.c | 141 +++++++++++++++++++--------------- drivers/cpuidle/cpuidle-psci.h | 9 ++- 3 files changed, 95 insertions(+), 65 deletions(-) (limited to 'drivers') diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c index f07786aad673..e48e578aaa7d 100644 --- a/drivers/cpuidle/cpuidle-psci-domain.c +++ b/drivers/cpuidle/cpuidle-psci-domain.c @@ -287,7 +287,7 @@ out: } subsys_initcall(psci_idle_init_domains); -struct device __init *psci_dt_attach_cpu(int cpu) +struct device *psci_dt_attach_cpu(int cpu) { struct device *dev; @@ -301,3 +301,11 @@ struct device __init *psci_dt_attach_cpu(int cpu) return dev; } + +void psci_dt_detach_cpu(struct device *dev) +{ + if (IS_ERR_OR_NULL(dev)) + return; + + dev_pm_domain_detach(dev, false); +} diff --git a/drivers/cpuidle/cpuidle-psci.c b/drivers/cpuidle/cpuidle-psci.c index 3806f911b61c..74463841805f 100644 --- a/drivers/cpuidle/cpuidle-psci.c +++ b/drivers/cpuidle/cpuidle-psci.c @@ -17,9 +17,11 @@ #include #include #include +#include #include #include #include +#include #include @@ -33,7 +35,7 @@ struct psci_cpuidle_data { static DEFINE_PER_CPU_READ_MOSTLY(struct psci_cpuidle_data, psci_cpuidle_data); static DEFINE_PER_CPU(u32, domain_state); -static bool psci_cpuidle_use_cpuhp __initdata; +static bool psci_cpuidle_use_cpuhp; void psci_set_domain_state(u32 state) { @@ -104,7 +106,7 @@ static int psci_idle_cpuhp_down(unsigned int cpu) return 0; } -static void __init psci_idle_init_cpuhp(void) +static void psci_idle_init_cpuhp(void) { int err; @@ -127,30 +129,13 @@ static int psci_enter_idle_state(struct cpuidle_device *dev, return psci_enter_state(idx, state[idx]); } -static struct cpuidle_driver psci_idle_driver __initdata = { - .name = "psci_idle", - .owner = THIS_MODULE, - /* - * PSCI idle states relies on architectural WFI to - * be represented as state index 0. - */ - .states[0] = { - .enter = psci_enter_idle_state, - .exit_latency = 1, - .target_residency = 1, - .power_usage = UINT_MAX, - .name = "WFI", - .desc = "ARM WFI", - } -}; - -static const struct of_device_id psci_idle_state_match[] __initconst = { +static const struct of_device_id psci_idle_state_match[] = { { .compatible = "arm,idle-state", .data = psci_enter_idle_state }, { }, }; -int __init psci_dt_parse_state_node(struct device_node *np, u32 *state) +int psci_dt_parse_state_node(struct device_node *np, u32 *state) { int err = of_property_read_u32(np, "arm,psci-suspend-param", state); @@ -167,9 +152,9 @@ int __init psci_dt_parse_state_node(struct device_node *np, u32 *state) return 0; } -static int __init psci_dt_cpu_init_topology(struct cpuidle_driver *drv, - struct psci_cpuidle_data *data, - unsigned int state_count, int cpu) +static int psci_dt_cpu_init_topology(struct cpuidle_driver *drv, + struct psci_cpuidle_data *data, + unsigned int state_count, int cpu) { /* Currently limit the hierarchical topology to be used in OSI mode. */ if (!psci_has_osi_support()) @@ -190,9 +175,9 @@ static int __init psci_dt_cpu_init_topology(struct cpuidle_driver *drv, return 0; } -static int __init psci_dt_cpu_init_idle(struct cpuidle_driver *drv, - struct device_node *cpu_node, - unsigned int state_count, int cpu) +static int psci_dt_cpu_init_idle(struct device *dev, struct cpuidle_driver *drv, + struct device_node *cpu_node, + unsigned int state_count, int cpu) { int i, ret = 0; u32 *psci_states; @@ -200,7 +185,8 @@ static int __init psci_dt_cpu_init_idle(struct cpuidle_driver *drv, struct psci_cpuidle_data *data = per_cpu_ptr(&psci_cpuidle_data, cpu); state_count++; /* Add WFI state too */ - psci_states = kcalloc(state_count, sizeof(*psci_states), GFP_KERNEL); + psci_states = devm_kcalloc(dev, state_count, sizeof(*psci_states), + GFP_KERNEL); if (!psci_states) return -ENOMEM; @@ -213,32 +199,26 @@ static int __init psci_dt_cpu_init_idle(struct cpuidle_driver *drv, of_node_put(state_node); if (ret) - goto free_mem; + return ret; pr_debug("psci-power-state %#x index %d\n", psci_states[i], i); } - if (i != state_count) { - ret = -ENODEV; - goto free_mem; - } + if (i != state_count) + return -ENODEV; /* Initialize optional data, used for the hierarchical topology. */ ret = psci_dt_cpu_init_topology(drv, data, state_count, cpu); if (ret < 0) - goto free_mem; + return ret; /* Idle states parsed correctly, store them in the per-cpu struct. */ data->psci_states = psci_states; return 0; - -free_mem: - kfree(psci_states); - return ret; } -static __init int psci_cpu_init_idle(struct cpuidle_driver *drv, - unsigned int cpu, unsigned int state_count) +static int psci_cpu_init_idle(struct device *dev, struct cpuidle_driver *drv, + unsigned int cpu, unsigned int state_count) { struct device_node *cpu_node; int ret; @@ -254,14 +234,22 @@ static __init int psci_cpu_init_idle(struct cpuidle_driver *drv, if (!cpu_node) return -ENODEV; - ret = psci_dt_cpu_init_idle(drv, cpu_node, state_count, cpu); + ret = psci_dt_cpu_init_idle(dev, drv, cpu_node, state_count, cpu); of_node_put(cpu_node); return ret; } -static int __init psci_idle_init_cpu(int cpu) +static void psci_cpu_deinit_idle(int cpu) +{ + struct psci_cpuidle_data *data = per_cpu_ptr(&psci_cpuidle_data, cpu); + + psci_dt_detach_cpu(data->dev); + psci_cpuidle_use_cpuhp = false; +} + +static int psci_idle_init_cpu(struct device *dev, int cpu) { struct cpuidle_driver *drv; struct device_node *cpu_node; @@ -284,17 +272,26 @@ static int __init psci_idle_init_cpu(int cpu) if (ret) return ret; - drv = kmemdup(&psci_idle_driver, sizeof(*drv), GFP_KERNEL); + drv = devm_kzalloc(dev, sizeof(*drv), GFP_KERNEL); if (!drv) return -ENOMEM; + drv->name = "psci_idle"; + drv->owner = THIS_MODULE; drv->cpumask = (struct cpumask *)cpumask_of(cpu); /* - * Initialize idle states data, starting at index 1, since - * by default idle state 0 is the quiescent state reached - * by the cpu by executing the wfi instruction. - * + * PSCI idle states relies on architectural WFI to be represented as + * state index 0. + */ + drv->states[0].enter = psci_enter_idle_state; + drv->states[0].exit_latency = 1; + drv->states[0].target_residency = 1; + drv->states[0].power_usage = UINT_MAX; + strcpy(drv->states[0].name, "WFI"); + strcpy(drv->states[0].desc, "ARM WFI"); + + /* * If no DT idle states are detected (ret == 0) let the driver * initialization fail accordingly since there is no reason to * initialize the idle driver if only wfi is supported, the @@ -302,48 +299,45 @@ static int __init psci_idle_init_cpu(int cpu) * on idle entry. */ ret = dt_init_idle_driver(drv, psci_idle_state_match, 1); - if (ret <= 0) { - ret = ret ? : -ENODEV; - goto out_kfree_drv; - } + if (ret <= 0) + return ret ? : -ENODEV; /* * Initialize PSCI idle states. */ - ret = psci_cpu_init_idle(drv, cpu, ret); + ret = psci_cpu_init_idle(dev, drv, cpu, ret); if (ret) { pr_err("CPU %d failed to PSCI idle\n", cpu); - goto out_kfree_drv; + return ret; } ret = cpuidle_register(drv, NULL); if (ret) - goto out_kfree_drv; + goto deinit; cpuidle_cooling_register(drv); return 0; - -out_kfree_drv: - kfree(drv); +deinit: + psci_cpu_deinit_idle(cpu); return ret; } /* - * psci_idle_init - Initializes PSCI cpuidle driver + * psci_idle_probe - Initializes PSCI cpuidle driver * * Initializes PSCI cpuidle driver for all CPUs, if any CPU fails * to register cpuidle driver then rollback to cancel all CPUs * registration. */ -static int __init psci_idle_init(void) +static int psci_cpuidle_probe(struct platform_device *pdev) { int cpu, ret; struct cpuidle_driver *drv; struct cpuidle_device *dev; for_each_possible_cpu(cpu) { - ret = psci_idle_init_cpu(cpu); + ret = psci_idle_init_cpu(&pdev->dev, cpu); if (ret) goto out_fail; } @@ -356,9 +350,34 @@ out_fail: dev = per_cpu(cpuidle_devices, cpu); drv = cpuidle_get_cpu_driver(dev); cpuidle_unregister(drv); - kfree(drv); + psci_cpu_deinit_idle(cpu); } return ret; } + +static struct platform_driver psci_cpuidle_driver = { + .probe = psci_cpuidle_probe, + .driver = { + .name = "psci-cpuidle", + }, +}; + +static int __init psci_idle_init(void) +{ + struct platform_device *pdev; + int ret; + + ret = platform_driver_register(&psci_cpuidle_driver); + if (ret) + return ret; + + pdev = platform_device_register_simple("psci-cpuidle", -1, NULL, 0); + if (IS_ERR(pdev)) { + platform_driver_unregister(&psci_cpuidle_driver); + return PTR_ERR(pdev); + } + + return 0; +} device_initcall(psci_idle_init); diff --git a/drivers/cpuidle/cpuidle-psci.h b/drivers/cpuidle/cpuidle-psci.h index ac8170684d4f..d8e925e84c27 100644 --- a/drivers/cpuidle/cpuidle-psci.h +++ b/drivers/cpuidle/cpuidle-psci.h @@ -3,15 +3,18 @@ #ifndef __CPUIDLE_PSCI_H #define __CPUIDLE_PSCI_H +struct device; struct device_node; void psci_set_domain_state(u32 state); -int __init psci_dt_parse_state_node(struct device_node *np, u32 *state); +int psci_dt_parse_state_node(struct device_node *np, u32 *state); #ifdef CONFIG_ARM_PSCI_CPUIDLE_DOMAIN -struct device __init *psci_dt_attach_cpu(int cpu); +struct device *psci_dt_attach_cpu(int cpu); +void psci_dt_detach_cpu(struct device *dev); #else -static inline struct device __init *psci_dt_attach_cpu(int cpu) { return NULL; } +static inline struct device *psci_dt_attach_cpu(int cpu) { return NULL; } +static inline void psci_dt_detach_cpu(struct device *dev) { } #endif #endif /* __CPUIDLE_PSCI_H */ -- cgit v1.2.3 From ee7c34caac3892af08f9c4c91b834d2cb2871eb0 Mon Sep 17 00:00:00 2001 From: Ulf Hansson Date: Tue, 7 Jul 2020 14:58:03 +0200 Subject: cpuidle: psci: Convert PM domain to platform driver To enable support for deferred probing and to allow implementation of the ->sync_state() callback from subsequent changes, let's convert into a platform driver. Reviewed-by: Lina Iyer Signed-off-by: Ulf Hansson Signed-off-by: Rafael J. Wysocki --- drivers/cpuidle/cpuidle-psci-domain.c | 45 ++++++++++++++++++++++------------- 1 file changed, 28 insertions(+), 17 deletions(-) (limited to 'drivers') diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c index e48e578aaa7d..bf527d2bb4b6 100644 --- a/drivers/cpuidle/cpuidle-psci-domain.c +++ b/drivers/cpuidle/cpuidle-psci-domain.c @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include @@ -42,8 +43,8 @@ static int psci_pd_power_off(struct generic_pm_domain *pd) return 0; } -static int __init psci_pd_parse_state_nodes(struct genpd_power_state *states, - int state_count) +static int psci_pd_parse_state_nodes(struct genpd_power_state *states, + int state_count) { int i, ret; u32 psci_state, *psci_state_buf; @@ -72,7 +73,7 @@ free_state: return ret; } -static int __init psci_pd_parse_states(struct device_node *np, +static int psci_pd_parse_states(struct device_node *np, struct genpd_power_state **states, int *state_count) { int ret; @@ -100,7 +101,7 @@ static void psci_pd_free_states(struct genpd_power_state *states, kfree(states); } -static int __init psci_pd_init(struct device_node *np) +static int psci_pd_init(struct device_node *np) { struct generic_pm_domain *pd; struct psci_pd_provider *pd_provider; @@ -167,7 +168,7 @@ out: return ret; } -static void __init psci_pd_remove(void) +static void psci_pd_remove(void) { struct psci_pd_provider *pd_provider, *it; struct generic_pm_domain *genpd; @@ -185,7 +186,7 @@ static void __init psci_pd_remove(void) } } -static int __init psci_pd_init_topology(struct device_node *np, bool add) +static int psci_pd_init_topology(struct device_node *np, bool add) { struct device_node *node; struct of_phandle_args child, parent; @@ -211,24 +212,24 @@ static int __init psci_pd_init_topology(struct device_node *np, bool add) return 0; } -static int __init psci_pd_add_topology(struct device_node *np) +static int psci_pd_add_topology(struct device_node *np) { return psci_pd_init_topology(np, true); } -static void __init psci_pd_remove_topology(struct device_node *np) +static void psci_pd_remove_topology(struct device_node *np) { psci_pd_init_topology(np, false); } -static const struct of_device_id psci_of_match[] __initconst = { +static const struct of_device_id psci_of_match[] = { { .compatible = "arm,psci-1.0" }, {} }; -static int __init psci_idle_init_domains(void) +static int psci_cpuidle_domain_probe(struct platform_device *pdev) { - struct device_node *np = of_find_matching_node(NULL, psci_of_match); + struct device_node *np = pdev->dev.of_node; struct device_node *node; int ret = 0, pd_count = 0; @@ -237,7 +238,7 @@ static int __init psci_idle_init_domains(void) /* Currently limit the hierarchical topology to be used in OSI mode. */ if (!psci_has_osi_support()) - goto out; + return 0; /* * Parse child nodes for the "#power-domain-cells" property and @@ -256,7 +257,7 @@ static int __init psci_idle_init_domains(void) /* Bail out if not using the hierarchical CPU topology. */ if (!pd_count) - goto out; + return 0; /* Link genpd masters/subdomains to model the CPU topology. */ ret = psci_pd_add_topology(np); @@ -271,9 +272,8 @@ static int __init psci_idle_init_domains(void) goto remove_pd; } - of_node_put(np); pr_info("Initialized CPU PM domain topology\n"); - return pd_count; + return 0; put_node: of_node_put(node); @@ -281,10 +281,21 @@ remove_pd: if (pd_count) psci_pd_remove(); pr_err("failed to create CPU PM domains ret=%d\n", ret); -out: - of_node_put(np); return ret; } + +static struct platform_driver psci_cpuidle_domain_driver = { + .probe = psci_cpuidle_domain_probe, + .driver = { + .name = "psci-cpuidle-domain", + .of_match_table = psci_of_match, + }, +}; + +static int __init psci_idle_init_domains(void) +{ + return platform_driver_register(&psci_cpuidle_domain_driver); +} subsys_initcall(psci_idle_init_domains); struct device *psci_dt_attach_cpu(int cpu) -- cgit v1.2.3 From 81f94ddfeceddb3b8ffa6cfe5ddd3b896711d9ad Mon Sep 17 00:00:00 2001 From: Ulf Hansson Date: Tue, 7 Jul 2020 14:58:04 +0200 Subject: cpuidle: psci: Prevent domain idlestates until consumers are ready Depending on the SoC/platform, additional devices may be part of the PSCI PM domain topology. This is the case with 'qcom,rpmh-rsc' device, for example, even if this is not yet visible in the corresponding DTS-files. Without going into too much details, a device like the 'qcom,rpmh-rsc' may have HW constraints that needs to be obeyed to, before a domain idlestate can be picked. Therefore, let's implement the ->sync_state() callback to receive a notification when all consumers of the PSCI PM domain providers have been attached/probed to it. In this way, we can make sure all constraints from all relevant devices, are taken into account before allowing a domain idlestate to be picked. Acked-by: Saravana Kannan Signed-off-by: Ulf Hansson Reviewed-by: Lukasz Luba Signed-off-by: Rafael J. Wysocki --- drivers/cpuidle/cpuidle-psci-domain.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'drivers') diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c index bf527d2bb4b6..b6e9649ab0da 100644 --- a/drivers/cpuidle/cpuidle-psci-domain.c +++ b/drivers/cpuidle/cpuidle-psci-domain.c @@ -27,6 +27,7 @@ struct psci_pd_provider { }; static LIST_HEAD(psci_pd_providers); +static bool psci_pd_allow_domain_state; static int psci_pd_power_off(struct generic_pm_domain *pd) { @@ -36,6 +37,9 @@ static int psci_pd_power_off(struct generic_pm_domain *pd) if (!state->data) return 0; + if (!psci_pd_allow_domain_state) + return -EBUSY; + /* OSI mode is enabled, set the corresponding domain state. */ pd_state = state->data; psci_set_domain_state(*pd_state); @@ -222,6 +226,15 @@ static void psci_pd_remove_topology(struct device_node *np) psci_pd_init_topology(np, false); } +static void psci_cpuidle_domain_sync_state(struct device *dev) +{ + /* + * All devices have now been attached/probed to the PM domain topology, + * hence it's fine to allow domain states to be picked. + */ + psci_pd_allow_domain_state = true; +} + static const struct of_device_id psci_of_match[] = { { .compatible = "arm,psci-1.0" }, {} @@ -289,6 +302,7 @@ static struct platform_driver psci_cpuidle_domain_driver = { .driver = { .name = "psci-cpuidle-domain", .of_match_table = psci_of_match, + .sync_state = psci_cpuidle_domain_sync_state, }, }; -- cgit v1.2.3 From efe9711214e6138a5a2a46ca4068bfce50c03444 Mon Sep 17 00:00:00 2001 From: Neal Liu Date: Mon, 27 Jul 2020 11:25:46 +0800 Subject: cpuidle: change enter_s2idle() prototype Control Flow Integrity(CFI) is a security mechanism that disallows changes to the original control flow graph of a compiled binary, making it significantly harder to perform such attacks. init_state_node() assign same function callback to different function pointer declarations. static int init_state_node(struct cpuidle_state *idle_state, const struct of_device_id *matches, struct device_node *state_node) { ... idle_state->enter = match_id->data; ... idle_state->enter_s2idle = match_id->data; } Function declarations: struct cpuidle_state { ... int (*enter) (struct cpuidle_device *dev, struct cpuidle_driver *drv, int index); void (*enter_s2idle) (struct cpuidle_device *dev, struct cpuidle_driver *drv, int index); }; In this case, either enter() or enter_s2idle() would cause CFI check failed since they use same callee. Align function prototype of enter() since it needs return value for some use cases. The return value of enter_s2idle() is no need currently. Signed-off-by: Neal Liu Reviewed-by: Sami Tolvanen Signed-off-by: Rafael J. Wysocki --- drivers/acpi/processor_idle.c | 6 ++++-- drivers/cpuidle/cpuidle-tegra.c | 8 +++++--- drivers/idle/intel_idle.c | 6 ++++-- include/linux/cpuidle.h | 9 ++++++--- 4 files changed, 19 insertions(+), 10 deletions(-) (limited to 'drivers') diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 75534c5b5433..6ffb6c99c2de 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -655,8 +655,8 @@ static int acpi_idle_enter(struct cpuidle_device *dev, return index; } -static void acpi_idle_enter_s2idle(struct cpuidle_device *dev, - struct cpuidle_driver *drv, int index) +static int acpi_idle_enter_s2idle(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int index) { struct acpi_processor_cx *cx = per_cpu(acpi_cstate[index], dev->cpu); @@ -674,6 +674,8 @@ static void acpi_idle_enter_s2idle(struct cpuidle_device *dev, } } acpi_idle_do_entry(cx); + + return 0; } static int acpi_processor_setup_cpuidle_cx(struct acpi_processor *pr, diff --git a/drivers/cpuidle/cpuidle-tegra.c b/drivers/cpuidle/cpuidle-tegra.c index 150045849d78..a12fb141875a 100644 --- a/drivers/cpuidle/cpuidle-tegra.c +++ b/drivers/cpuidle/cpuidle-tegra.c @@ -253,11 +253,13 @@ static int tegra_cpuidle_enter(struct cpuidle_device *dev, return err ? -1 : index; } -static void tegra114_enter_s2idle(struct cpuidle_device *dev, - struct cpuidle_driver *drv, - int index) +static int tegra114_enter_s2idle(struct cpuidle_device *dev, + struct cpuidle_driver *drv, + int index) { tegra_cpuidle_enter(dev, drv, index); + + return 0; } /* diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index fa23a7ea01ac..3f86f36dab2b 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -173,13 +173,15 @@ static __cpuidle int intel_idle(struct cpuidle_device *dev, * Invoked as a suspend-to-idle callback routine with frozen user space, frozen * scheduler tick and suspended scheduler clock on the target CPU. */ -static __cpuidle void intel_idle_s2idle(struct cpuidle_device *dev, - struct cpuidle_driver *drv, int index) +static __cpuidle int intel_idle_s2idle(struct cpuidle_device *dev, + struct cpuidle_driver *drv, int index) { unsigned long eax = flg2MWAIT(drv->states[index].flags); unsigned long ecx = 1; /* break on interrupt flag */ mwait_idle_with_hints(eax, ecx); + + return 0; } /* diff --git a/include/linux/cpuidle.h b/include/linux/cpuidle.h index ec2ef63771f0..b65909ae4e20 100644 --- a/include/linux/cpuidle.h +++ b/include/linux/cpuidle.h @@ -65,10 +65,13 @@ struct cpuidle_state { * CPUs execute ->enter_s2idle with the local tick or entire timekeeping * suspended, so it must not re-enable interrupts at any point (even * temporarily) or attempt to change states of clock event devices. + * + * This callback may point to the same function as ->enter if all of + * the above requirements are met by it. */ - void (*enter_s2idle) (struct cpuidle_device *dev, - struct cpuidle_driver *drv, - int index); + int (*enter_s2idle)(struct cpuidle_device *dev, + struct cpuidle_driver *drv, + int index); }; /* Idle State Flags */ -- cgit v1.2.3 From a472ad2bcea479ba068880125d7273fc95c14b70 Mon Sep 17 00:00:00 2001 From: Chen Yu Date: Fri, 10 Jul 2020 12:12:01 +0800 Subject: intel_idle: Customize IceLake server support On ICX platform, the C1E auto-promotion is enabled by default. As a result, the CPU might fall into C1E more offen than previous platforms. Besides, the C1E is not exposed to sysfs on ICX, which is inconsistent with previous server platforms. So disable C1E auto-promotion and expose C1E as a separate idle state, so the C1E and C6 can be disabled via sysfs when necessary. Beside C1 and C1E, the exit latency of C6 was measured by a dedicated tool. However the exit latency(41us) exposed by _CST is much smaller than the one we measured(128us). This is probably due to the _CST uses the exit latency when woken up from PC0+C6, rather than PC6+C6 when C6 was measured. Choose the latter as we need the longest latency in theory. Reported-by: kernel test robot Tested-by: Artem Bityutskiy Acked-by: Artem Bityutskiy Reviewed-by: Zhang Rui Signed-off-by: Chen Yu Signed-off-by: Rafael J. Wysocki --- drivers/idle/intel_idle.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) (limited to 'drivers') diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index 3f86f36dab2b..fd0fa9e7900b 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -752,6 +752,35 @@ static struct cpuidle_state skx_cstates[] __initdata = { .enter = NULL } }; +static struct cpuidle_state icx_cstates[] __initdata = { + { + .name = "C1", + .desc = "MWAIT 0x00", + .flags = MWAIT2flg(0x00), + .exit_latency = 1, + .target_residency = 1, + .enter = &intel_idle, + .enter_s2idle = intel_idle_s2idle, }, + { + .name = "C1E", + .desc = "MWAIT 0x01", + .flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE, + .exit_latency = 4, + .target_residency = 4, + .enter = &intel_idle, + .enter_s2idle = intel_idle_s2idle, }, + { + .name = "C6", + .desc = "MWAIT 0x20", + .flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED, + .exit_latency = 128, + .target_residency = 384, + .enter = &intel_idle, + .enter_s2idle = intel_idle_s2idle, }, + { + .enter = NULL } +}; + static struct cpuidle_state atom_cstates[] __initdata = { { .name = "C1E", @@ -1056,6 +1085,12 @@ static const struct idle_cpu idle_cpu_skx __initconst = { .use_acpi = true, }; +static const struct idle_cpu idle_cpu_icx __initconst = { + .state_table = icx_cstates, + .disable_promotion_to_c1e = true, + .use_acpi = true, +}; + static const struct idle_cpu idle_cpu_avn __initconst = { .state_table = avn_cstates, .disable_promotion_to_c1e = true, @@ -1110,6 +1145,7 @@ static const struct x86_cpu_id intel_idle_ids[] __initconst = { X86_MATCH_INTEL_FAM6_MODEL(KABYLAKE_L, &idle_cpu_skl), X86_MATCH_INTEL_FAM6_MODEL(KABYLAKE, &idle_cpu_skl), X86_MATCH_INTEL_FAM6_MODEL(SKYLAKE_X, &idle_cpu_skx), + X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_X, &idle_cpu_icx), X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNL, &idle_cpu_knl), X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNM, &idle_cpu_knl), X86_MATCH_INTEL_FAM6_MODEL(ATOM_GOLDMONT, &idle_cpu_bxt), -- cgit v1.2.3 From 0a398945d6f3e54f9f88ba2a17908674c659bfa4 Mon Sep 17 00:00:00 2001 From: Neal Liu Date: Fri, 31 Jul 2020 11:39:04 +0800 Subject: cpuidle: ACPI: fix 'return' with no value build warning Add return value to fix return-type build warning introduced by commit efe9711214e6 ("cpuidle: change enter_s2idle() prototype"). Fixes: efe9711214e6 ("cpuidle: change enter_s2idle() prototype") Signed-off-by: Neal Liu [ rjw: Subject & changelog edits, make acpi_idle_enter_s2idle() return 0 in all cases ] Signed-off-by: Rafael J. Wysocki --- drivers/acpi/processor_idle.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 6ffb6c99c2de..0c65a09aec1b 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -664,11 +664,11 @@ static int acpi_idle_enter_s2idle(struct cpuidle_device *dev, struct acpi_processor *pr = __this_cpu_read(processors); if (unlikely(!pr)) - return; + return 0; if (pr->flags.bm_check) { acpi_idle_enter_bm(pr, cx, false); - return; + return 0; } else { ACPI_FLUSH_CPU_CACHE(); } -- cgit v1.2.3