[PATCH v2] tty: hvc: Fix data abort due to race in hvc_open
Potentially, hvc_open() can be called in parallel when two tasks calls open() on /dev/hvcX. In such a scenario, if the hp->ops->notifier_add() callback in the function fails, where it sets the tty->driver_data to NULL, the parallel hvc_open() can see this NULL and cause a memory abort. Hence, do a NULL check at the beginning, before proceeding ahead. The issue can be easily reproduced by launching two tasks simultaneously that does an open() call on /dev/hvcX. For example: $ cat /dev/hvc0 & cat /dev/hvc0 & Cc: sta...@vger.kernel.org Signed-off-by: Raghavendra Rao Ananta --- drivers/tty/hvc/hvc_console.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/tty/hvc/hvc_console.c b/drivers/tty/hvc/hvc_console.c index 436cc51c92c3..80709f754cc8 100644 --- a/drivers/tty/hvc/hvc_console.c +++ b/drivers/tty/hvc/hvc_console.c @@ -350,6 +350,9 @@ static int hvc_open(struct tty_struct *tty, struct file * filp) unsigned long flags; int rc = 0; + if (!hp) + return -ENODEV; + spin_lock_irqsave(>port.lock, flags); /* Check and then increment for fast path open. */ if (hp->port.count++ > 0) { -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
[PATCH] of: Add of_get_memory_prop()
On some embedded systems, the '/memory' dt-property gets updated by the bootloader (for example, the DDR configuration) and then gets passed onto the kernel. The device drivers may have to read the properties at runtime to make decisions. Hence, add of_get_memory_prop() for the device drivers to query the requested properties. Signed-off-by: Raghavendra Rao Ananta --- drivers/of/fdt.c | 27 +++ include/linux/of_fdt.h | 1 + 2 files changed, 28 insertions(+) diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 223d617ecfe1..925cf2852433 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -79,6 +79,33 @@ void __init of_fdt_limit_memory(int limit) } } +/** + * of_fdt_get_memory_prop - Return the requested property from the /memory node + * + * On match, returns a non-zero positive value which represents the property + * value. Otherwise returns -ENOENT. + */ +int of_fdt_get_memory_prop(const char *pname) +{ + int memory; + int len; + fdt32_t *prop = NULL; + + if (!pname) + return -EINVAL; + + memory = fdt_path_offset(initial_boot_params, "/memory"); + if (memory > 0) + prop = fdt_getprop_w(initial_boot_params, memory, + pname, ); + + if (!prop || len != sizeof(u32)) + return -ENOENT; + + return fdt32_to_cpu(*prop); +} +EXPORT_SYMBOL_GPL(of_fdt_get_memory_prop); + static bool of_fdt_device_is_available(const void *blob, unsigned long node) { const char *status = fdt_getprop(blob, node, "status", NULL); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index acf820e88952..537f29373358 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -38,6 +38,7 @@ extern char __dtb_end[]; /* Other Prototypes */ extern u64 of_flat_dt_translate_address(unsigned long node); extern void of_fdt_limit_memory(int limit); +extern int of_fdt_get_memory_prop(const char *pname); #endif /* CONFIG_OF_FLATTREE */ #ifdef CONFIG_OF_EARLY_FLATTREE -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH 1/1] perf: Add CPU hotplug support for events
On 02/16/2018 12:39 PM, Peter Zijlstra wrote: On Fri, Feb 16, 2018 at 10:06:29AM -0800, Raghavendra Rao Ananta wrote: No this is absolutely disguisting. You can simply keep the events in the dead CPU's context. It's really not that hard. Keeping the events in the dead CPU's context was also an idea that we had. However, detaching that event from the PMU when the CPU is offline would be a pain. Consider the scenario in which an event is about to be destroyed when the CPU is offline (yet still attached to the CPU). During it's destruction, a cross-cpu call is made (from perf_remove_from_context()) to the offlined CPU to detach the event from the CPU's PMU. As the CPU is offline, that would not be possible, and again a separate logic has to be written for cleaning up the events whose CPUs are offlined. That is actually really simple to deal with. The real problems are with semantics, is an event enabled when the CPU is dead? Can you disable/enable an event on a dead CPU. The below patch (_completely_ untested) should do most of it, but needs help with the details. I suspect we want to allow enable/disable on events that are on a dead CPU, and equally I think we want to account the time an enabled event spends on a dead CPU to go towards the 'enabled' bucket. I've gone through your diff, and it gave me a hint of similar texture what we are trying to do (except for maintaining an offline event list). Nevertheless, I tried to test your patch. I created an hw event, and tried to offline the CPU in parallel, and I immediately hit a watchdog soft lockup bug! Tried the same this by first switching off the CPU (without any event created), and I hit into similar issue. I am sure we can fix it, but apart from the "why we are doing hotplug?" question, was was there specifically any issue with our patch? Also, you _still_ don't explain why you care about dead CPUs. I wanted to understand, if we no longer care about hotplugging of CPUs, then why do we still have exported symbols such as cpu_up() and cpu_down()? Moreover, we also have the hotplug interface exposed to users-space as well (through sysfs). As long as these interfaces exist, there's always a potential chance of bringing the CPU up/down. Can you please clear this thing up for me? -- Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH 1/1] perf: Add CPU hotplug support for events
On 02/16/2018 12:39 PM, Peter Zijlstra wrote: On Fri, Feb 16, 2018 at 10:06:29AM -0800, Raghavendra Rao Ananta wrote: No this is absolutely disguisting. You can simply keep the events in the dead CPU's context. It's really not that hard. Keeping the events in the dead CPU's context was also an idea that we had. However, detaching that event from the PMU when the CPU is offline would be a pain. Consider the scenario in which an event is about to be destroyed when the CPU is offline (yet still attached to the CPU). During it's destruction, a cross-cpu call is made (from perf_remove_from_context()) to the offlined CPU to detach the event from the CPU's PMU. As the CPU is offline, that would not be possible, and again a separate logic has to be written for cleaning up the events whose CPUs are offlined. That is actually really simple to deal with. The real problems are with semantics, is an event enabled when the CPU is dead? Can you disable/enable an event on a dead CPU. The below patch (_completely_ untested) should do most of it, but needs help with the details. I suspect we want to allow enable/disable on events that are on a dead CPU, and equally I think we want to account the time an enabled event spends on a dead CPU to go towards the 'enabled' bucket. I've gone through your diff, and it gave me a hint of similar texture what we are trying to do (except for maintaining an offline event list). Nevertheless, I tried to test your patch. I created an hw event, and tried to offline the CPU in parallel, and I immediately hit a watchdog soft lockup bug! Tried the same this by first switching off the CPU (without any event created), and I hit into similar issue. I am sure we can fix it, but apart from the "why we are doing hotplug?" question, was was there specifically any issue with our patch? Also, you _still_ don't explain why you care about dead CPUs. I wanted to understand, if we no longer care about hotplugging of CPUs, then why do we still have exported symbols such as cpu_up() and cpu_down()? Moreover, we also have the hotplug interface exposed to users-space as well (through sysfs). As long as these interfaces exist, there's always a potential chance of bringing the CPU up/down. Can you please clear this thing up for me? -- Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH 1/1] perf: Add CPU hotplug support for events
On 02/16/2018 12:21 AM, Peter Zijlstra wrote: On Thu, Feb 15, 2018 at 03:01:41PM -0800, Raghavendra Rao Ananta wrote: Perf framework doesn't allow prevserving CPU events across CPU hotplugs. The events are scheduled out as and when the CPU walks offline. Moreover, the framework also doesn't allow the clients to create events on an offline CPU. As a result, the clients have to keep on monitoring the CPU state until it comes back online. Therefore, introducing the perf framework to support creation and preserving of (CPU) events for offline CPUs. Through this, the CPU's online state would be transparent to the client and it not have to worry about monitoring the CPU's state. Success would be returned to the client even while creating the event on an offline CPU. If during the lifetime of the event the CPU walks offline, the event would be preserved and would continue to count as soon as (and if) the CPU comes back online. Signed-off-by: Raghavendra Rao Ananta <rana...@codeaurora.org> --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_head sb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; No this is absolutely disguisting. You can simply keep the events in the dead CPU's context. It's really not that hard. Keeping the events in the dead CPU's context was also an idea that we had. However, detaching that event from the PMU when the CPU is offline would be a pain. Consider the scenario in which an event is about to be destroyed when the CPU is offline (yet still attached to the CPU). During it's destruction, a cross-cpu call is made (from perf_remove_from_context()) to the offlined CPU to detach the event from the CPU's PMU. As the CPU is offline, that would not be possible, and again a separate logic has to be written for cleaning up the events whose CPUs are offlined. Hence, I thought it would be a cleaner way to maintain the events. Also, you _still_ don't explain why you care about dead CPUs. It's just not only about dead CPUs. It's the fact that the CPUs can come and go online. The embedded world, specifically Android mobile SoCs, rely on CPU hotplugs to manage power and thermal constraints. These hotplugs can happen at a very rapid pace. Adjacently, they also rely on many perf event counters for its management. Therefore, there is a need to preserve these events across hotplugs. In such a scenario, a perf client (kernel or user-space) can create events even when the CPU is offline. If the CPU comes online during the lifetime of the event, the registered event can start counting spontaneously. As an extension to this, the events' count can also be preserved across CPU hotplugs. This takes the burden off of the clients to monitor the state of the CPU. -- Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH 1/1] perf: Add CPU hotplug support for events
On 02/16/2018 12:21 AM, Peter Zijlstra wrote: On Thu, Feb 15, 2018 at 03:01:41PM -0800, Raghavendra Rao Ananta wrote: Perf framework doesn't allow prevserving CPU events across CPU hotplugs. The events are scheduled out as and when the CPU walks offline. Moreover, the framework also doesn't allow the clients to create events on an offline CPU. As a result, the clients have to keep on monitoring the CPU state until it comes back online. Therefore, introducing the perf framework to support creation and preserving of (CPU) events for offline CPUs. Through this, the CPU's online state would be transparent to the client and it not have to worry about monitoring the CPU's state. Success would be returned to the client even while creating the event on an offline CPU. If during the lifetime of the event the CPU walks offline, the event would be preserved and would continue to count as soon as (and if) the CPU comes back online. Signed-off-by: Raghavendra Rao Ananta --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_head sb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; No this is absolutely disguisting. You can simply keep the events in the dead CPU's context. It's really not that hard. Keeping the events in the dead CPU's context was also an idea that we had. However, detaching that event from the PMU when the CPU is offline would be a pain. Consider the scenario in which an event is about to be destroyed when the CPU is offline (yet still attached to the CPU). During it's destruction, a cross-cpu call is made (from perf_remove_from_context()) to the offlined CPU to detach the event from the CPU's PMU. As the CPU is offline, that would not be possible, and again a separate logic has to be written for cleaning up the events whose CPUs are offlined. Hence, I thought it would be a cleaner way to maintain the events. Also, you _still_ don't explain why you care about dead CPUs. It's just not only about dead CPUs. It's the fact that the CPUs can come and go online. The embedded world, specifically Android mobile SoCs, rely on CPU hotplugs to manage power and thermal constraints. These hotplugs can happen at a very rapid pace. Adjacently, they also rely on many perf event counters for its management. Therefore, there is a need to preserve these events across hotplugs. In such a scenario, a perf client (kernel or user-space) can create events even when the CPU is offline. If the CPU comes online during the lifetime of the event, the registered event can start counting spontaneously. As an extension to this, the events' count can also be preserved across CPU hotplugs. This takes the burden off of the clients to monitor the state of the CPU. -- Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
[PATCH 1/1] perf: Add CPU hotplug support for events
Perf framework doesn't allow prevserving CPU events across CPU hotplugs. The events are scheduled out as and when the CPU walks offline. Moreover, the framework also doesn't allow the clients to create events on an offline CPU. As a result, the clients have to keep on monitoring the CPU state until it comes back online. Therefore, introducing the perf framework to support creation and preserving of (CPU) events for offline CPUs. Through this, the CPU's online state would be transparent to the client and it not have to worry about monitoring the CPU's state. Success would be returned to the client even while creating the event on an offline CPU. If during the lifetime of the event the CPU walks offline, the event would be preserved and would continue to count as soon as (and if) the CPU comes back online. Signed-off-by: Raghavendra Rao Ananta <rana...@codeaurora.org> --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_headsb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; #endif /* CONFIG_PERF_EVENTS */ }; diff --git a/kernel/events/core.c b/kernel/events/core.c index 96db9ae..5d0a155 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2329,6 +2329,30 @@ static int __perf_install_in_context(void *info) return ret; } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static DEFINE_PER_CPU(struct list_head, dormant_event_list); +static DEFINE_PER_CPU(spinlock_t, dormant_event_list_lock); + +static void perf_prepare_install_in_context(struct perf_event *event) +{ + int cpu = event->cpu; + bool prepare_hp_sched = !READ_ONCE(event->ctx->task); + + if (!prepare_hp_sched) + return; + + spin_lock(_cpu(dormant_event_list_lock, cpu)); + if (event->state == PERF_EVENT_STATE_DORMANT) + goto out; + + event->state = PERF_EVENT_STATE_DORMANT; + list_add_tail(>dormant_entry, + _cpu(dormant_event_list, cpu)); +out: + spin_unlock(_cpu(dormant_event_list_lock, cpu)); +} +#endif + /* * Attach a performance event to a context. * @@ -2353,6 +2377,15 @@ static int __perf_install_in_context(void *info) smp_store_release(>ctx, ctx); if (!task) { + struct perf_cpu_context *cpuctx = + container_of(ctx, struct perf_cpu_context, ctx); + +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + if (!cpuctx->online) { + perf_prepare_install_in_context(event); + return; + } +#endif cpu_function_call(cpu, __perf_install_in_context, event); return; } @@ -2421,6 +2454,43 @@ static int __perf_install_in_context(void *info) raw_spin_unlock_irq(>lock); } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static void perf_deferred_install_in_context(int cpu) +{ + struct perf_event *event, *tmp; + struct perf_event_context *ctx; + + /* This function is called twice while coming online. Once for +* CPUHP_PERF_PREPARE and the other for CPUHP_AP_PERF_ONLINE. +* Only during the CPUHP_AP_PERF_ONLINE state, we can confirm +* that CPU PMU is ready and can be installed to. +*/ + if (!cpu_online(cpu)) + return; + + spin_lock(_cpu(dormant_event_list_lock, cpu)); + list_for_each_entry_safe(event, tmp, + _cpu(dormant_event_list, cpu), dormant_entry) { + if (cpu != event->cpu) + continue; + + list_del(>dormant_entry); + event->state = PERF_EVENT_STATE_INACTIVE; + spin_unlock(_cpu(dormant_event_list_lock, cpu)); + + ctx = event->ctx; + perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); + + mutex_lock(>mutex); + perf_install_in_context(ctx, event, cpu); + mutex_unlock(>mutex); + + spin_lock(_cpu(dormant_event_l
[PATCH 1/1] perf: Add CPU hotplug support for events
Perf framework doesn't allow prevserving CPU events across CPU hotplugs. The events are scheduled out as and when the CPU walks offline. Moreover, the framework also doesn't allow the clients to create events on an offline CPU. As a result, the clients have to keep on monitoring the CPU state until it comes back online. Therefore, introducing the perf framework to support creation and preserving of (CPU) events for offline CPUs. Through this, the CPU's online state would be transparent to the client and it not have to worry about monitoring the CPU's state. Success would be returned to the client even while creating the event on an offline CPU. If during the lifetime of the event the CPU walks offline, the event would be preserved and would continue to count as soon as (and if) the CPU comes back online. Signed-off-by: Raghavendra Rao Ananta --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_headsb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; #endif /* CONFIG_PERF_EVENTS */ }; diff --git a/kernel/events/core.c b/kernel/events/core.c index 96db9ae..5d0a155 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2329,6 +2329,30 @@ static int __perf_install_in_context(void *info) return ret; } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static DEFINE_PER_CPU(struct list_head, dormant_event_list); +static DEFINE_PER_CPU(spinlock_t, dormant_event_list_lock); + +static void perf_prepare_install_in_context(struct perf_event *event) +{ + int cpu = event->cpu; + bool prepare_hp_sched = !READ_ONCE(event->ctx->task); + + if (!prepare_hp_sched) + return; + + spin_lock(_cpu(dormant_event_list_lock, cpu)); + if (event->state == PERF_EVENT_STATE_DORMANT) + goto out; + + event->state = PERF_EVENT_STATE_DORMANT; + list_add_tail(>dormant_entry, + _cpu(dormant_event_list, cpu)); +out: + spin_unlock(_cpu(dormant_event_list_lock, cpu)); +} +#endif + /* * Attach a performance event to a context. * @@ -2353,6 +2377,15 @@ static int __perf_install_in_context(void *info) smp_store_release(>ctx, ctx); if (!task) { + struct perf_cpu_context *cpuctx = + container_of(ctx, struct perf_cpu_context, ctx); + +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + if (!cpuctx->online) { + perf_prepare_install_in_context(event); + return; + } +#endif cpu_function_call(cpu, __perf_install_in_context, event); return; } @@ -2421,6 +2454,43 @@ static int __perf_install_in_context(void *info) raw_spin_unlock_irq(>lock); } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static void perf_deferred_install_in_context(int cpu) +{ + struct perf_event *event, *tmp; + struct perf_event_context *ctx; + + /* This function is called twice while coming online. Once for +* CPUHP_PERF_PREPARE and the other for CPUHP_AP_PERF_ONLINE. +* Only during the CPUHP_AP_PERF_ONLINE state, we can confirm +* that CPU PMU is ready and can be installed to. +*/ + if (!cpu_online(cpu)) + return; + + spin_lock(_cpu(dormant_event_list_lock, cpu)); + list_for_each_entry_safe(event, tmp, + _cpu(dormant_event_list, cpu), dormant_entry) { + if (cpu != event->cpu) + continue; + + list_del(>dormant_entry); + event->state = PERF_EVENT_STATE_INACTIVE; + spin_unlock(_cpu(dormant_event_list_lock, cpu)); + + ctx = event->ctx; + perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); + + mutex_lock(>mutex); + perf_install_in_context(ctx, event, cpu); + mutex_unlock(>mutex); + + spin_lock(_cpu(dormant_event_list_lock, cpu)); + } + spin_unlock(_c
[PATCH 0/1] perf: Add CPU hotplug support for events
The embedded world, specifically Android mobile SoCs, rely on CPU hotplugs to manage power and thermal constraints. These hotplugs can happen at a very rapid pace. Adjacently, they also relies on many perf event counters for its management. Therefore, there is a need to preserve these events across hotplugs. In such a scenario, a perf client (kernel or user-space) can create events even when the CPU is offline. If the CPU comes online during the lifetime of the event, the registered event can start counting spontaneously. As an extension to this, the events' count can also be preserved across CPU hotplugs. This takes the burden off of the clients to monitor the state of the CPU. The tests were conducted on x86_64 PC. Sample results: /* CPU-1 is offline: Event created when CPU is offline */ # ./perf stat -C 1 -e cycles -v -I 1000 Using CPUID GenuineIntel-6-5E cycles: 0 0 0 # time counts unit events 1.000143727cycles cycles: 0 0 0 2.000309394cycles cycles: 0 0 0 3.000466412cycles cycles: 0 0 0 4.000599885cycles cycles: 0 0 0 5.000750095cycles cycles: 0 0 0 6.000901789cycles cycles: 0 0 0 7.001035315cycles cycles: 0 0 0 8.001195270cycles /* CPU-1 made online: Event started counting */ cycles: 21046256 719464572 719464572 9.001379820 21,046,256 cycles cycles: 19537953 1000222653 1000222653 10.001634052 19,537,953 cycles cycles: 27480135 195468 195468 11.001764556 27,480,135 cycles cycles: 27723233 1000108008 1000108008 12.001909310 27,723,233 cycles cycles: 19137349 1000133462 1000133462 13.002079041 19,137,349 cycles cycles: 48913285 1000101268 1000101268 14.002218928 48,913,285 cycles cycles: 27259199 1000128972 1000128972 15.002388700 27,259,199 cycles cycles: 19249055 195789 195789 16.002514765 19,249,055 cycles cycles: 27530051 1000246860 1000246860 17.002801610 27,530,051 cycles cycles: 34348072 1000272874 1000272874 18.003110100 34,348,072 cycles cycles: 11526457 107984264 107984264 19.003435811 11,526,457 cycles /* CPU-1 made offline */ cycles: 0 0 0 20.003582803cycles cycles: 0 0 0 21.003896484cycles cycles: 0 0 0 22.004212989cycles cycles: 0 0 0 23.004346689cycles cycles: 0 0 0 24.004668259cycles cycles: 0 0 0 25.004983504cycles cycles: 0 0 0 # time counts unit events 26.005315741cycles /* CPU-1 made online: Event preserved across hotplug */ cycles: 27210082 933493459 933493459 27.005652287 27,210,082 cycles cycles: 41950431 1000112865 1000112865 28.005805475 41,950,431 cycles cycles: 35075124 1000141146 1000141146 29.005974101 35,075,124 cycles cycles: 45240055 1000132743 1000132743 30.006140008 45,240,055 cycles cycles: 43426180 177828 177828 31.006253035 43,426,180 cycles cycles: 34593167 1000315835 1000315835 32.006605393 34,593,167 cycles cycles: 105078270 1000136171 1000136171 33.006773971105,078,270 cycles Raghavendra Rao Ananta (1): perf: Add CPU hotplug support for events include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
[PATCH 0/1] perf: Add CPU hotplug support for events
The embedded world, specifically Android mobile SoCs, rely on CPU hotplugs to manage power and thermal constraints. These hotplugs can happen at a very rapid pace. Adjacently, they also relies on many perf event counters for its management. Therefore, there is a need to preserve these events across hotplugs. In such a scenario, a perf client (kernel or user-space) can create events even when the CPU is offline. If the CPU comes online during the lifetime of the event, the registered event can start counting spontaneously. As an extension to this, the events' count can also be preserved across CPU hotplugs. This takes the burden off of the clients to monitor the state of the CPU. The tests were conducted on x86_64 PC. Sample results: /* CPU-1 is offline: Event created when CPU is offline */ # ./perf stat -C 1 -e cycles -v -I 1000 Using CPUID GenuineIntel-6-5E cycles: 0 0 0 # time counts unit events 1.000143727cycles cycles: 0 0 0 2.000309394cycles cycles: 0 0 0 3.000466412cycles cycles: 0 0 0 4.000599885cycles cycles: 0 0 0 5.000750095cycles cycles: 0 0 0 6.000901789cycles cycles: 0 0 0 7.001035315cycles cycles: 0 0 0 8.001195270cycles /* CPU-1 made online: Event started counting */ cycles: 21046256 719464572 719464572 9.001379820 21,046,256 cycles cycles: 19537953 1000222653 1000222653 10.001634052 19,537,953 cycles cycles: 27480135 195468 195468 11.001764556 27,480,135 cycles cycles: 27723233 1000108008 1000108008 12.001909310 27,723,233 cycles cycles: 19137349 1000133462 1000133462 13.002079041 19,137,349 cycles cycles: 48913285 1000101268 1000101268 14.002218928 48,913,285 cycles cycles: 27259199 1000128972 1000128972 15.002388700 27,259,199 cycles cycles: 19249055 195789 195789 16.002514765 19,249,055 cycles cycles: 27530051 1000246860 1000246860 17.002801610 27,530,051 cycles cycles: 34348072 1000272874 1000272874 18.003110100 34,348,072 cycles cycles: 11526457 107984264 107984264 19.003435811 11,526,457 cycles /* CPU-1 made offline */ cycles: 0 0 0 20.003582803cycles cycles: 0 0 0 21.003896484cycles cycles: 0 0 0 22.004212989cycles cycles: 0 0 0 23.004346689cycles cycles: 0 0 0 24.004668259cycles cycles: 0 0 0 25.004983504cycles cycles: 0 0 0 # time counts unit events 26.005315741cycles /* CPU-1 made online: Event preserved across hotplug */ cycles: 27210082 933493459 933493459 27.005652287 27,210,082 cycles cycles: 41950431 1000112865 1000112865 28.005805475 41,950,431 cycles cycles: 35075124 1000141146 1000141146 29.005974101 35,075,124 cycles cycles: 45240055 1000132743 1000132743 30.006140008 45,240,055 cycles cycles: 43426180 177828 177828 31.006253035 43,426,180 cycles cycles: 34593167 1000315835 1000315835 32.006605393 34,593,167 cycles cycles: 105078270 1000136171 1000136171 33.006773971105,078,270 cycles Raghavendra Rao Ananta (1): perf: Add CPU hotplug support for events include/linux/perf_event.h | 7 +++ kernel/events/core.c | 123 + 2 files changed, 97 insertions(+), 33 deletions(-) -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
On 02/13/2018 08:08 AM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 02:22:30PM -0800, Raghavendra Rao Ananta wrote: On 02/12/2018 01:21 PM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 10:04:42PM +0100, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 09:42:05AM -0800, Raghavendra Rao Ananta wrote: Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. nope, we don't check on that Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? hum, I wonder what's wrong.. will check I think the user space needs to enable the event once the cpu gets online.. which we dont do and your app does..? maybe we could add perf_event_attr::enable_on_online ;-) I'll check what we can do in user space, I guess we can monitor the cpu state and enable event accordingly jirka Yes, probably that's the reason. In order for an event to get scheduled-in, it expects the event to be at least in PERF_EVENT_STATE_INACTIVE state. If you notice, in my patch, when the cpu wakes up, we are initializing the state of the event (perf_event__state_init()) and then trying to schedule-in. Since the event was created with a disabled state, it seems that the same this is followed and the state gets initialized to PERF_EVENT_STATE_OFF. Unfortunately, events in this state could not be scheduled. One way for things to get working is, instead of calling perf_event__state_init() before the event is scheduled-in (when the cpu wakes up), we can do something like: perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); could you add check in ioctl call that set the inactive state on the dormant event.. that would start it once the cpu is online.. as requested I am a little confused. When you say "check", do you mean a new ioctl command? So the flow (from the user-space perspective) would go something like this? 1. // CPU offline 2. perf_event_open(); // event started as disabled --> added to dormant list in the kernel 3. ioctl(SET_INACTIVE); // change the state of the event to inactive 4. // CPU woken up! 5. // schedule the (inactive) event by traversing the dormant list Is this what you were trying to mention, or am I missing something? Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
On 02/13/2018 08:08 AM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 02:22:30PM -0800, Raghavendra Rao Ananta wrote: On 02/12/2018 01:21 PM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 10:04:42PM +0100, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 09:42:05AM -0800, Raghavendra Rao Ananta wrote: Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. nope, we don't check on that Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? hum, I wonder what's wrong.. will check I think the user space needs to enable the event once the cpu gets online.. which we dont do and your app does..? maybe we could add perf_event_attr::enable_on_online ;-) I'll check what we can do in user space, I guess we can monitor the cpu state and enable event accordingly jirka Yes, probably that's the reason. In order for an event to get scheduled-in, it expects the event to be at least in PERF_EVENT_STATE_INACTIVE state. If you notice, in my patch, when the cpu wakes up, we are initializing the state of the event (perf_event__state_init()) and then trying to schedule-in. Since the event was created with a disabled state, it seems that the same this is followed and the state gets initialized to PERF_EVENT_STATE_OFF. Unfortunately, events in this state could not be scheduled. One way for things to get working is, instead of calling perf_event__state_init() before the event is scheduled-in (when the cpu wakes up), we can do something like: perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); could you add check in ioctl call that set the inactive state on the dormant event.. that would start it once the cpu is online.. as requested I am a little confused. When you say "check", do you mean a new ioctl command? So the flow (from the user-space perspective) would go something like this? 1. // CPU offline 2. perf_event_open(); // event started as disabled --> added to dormant list in the kernel 3. ioctl(SET_INACTIVE); // change the state of the event to inactive 4. // CPU woken up! 5. // schedule the (inactive) event by traversing the dormant list Is this what you were trying to mention, or am I missing something? Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
On 02/12/2018 01:21 PM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 10:04:42PM +0100, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 09:42:05AM -0800, Raghavendra Rao Ananta wrote: Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. nope, we don't check on that Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? hum, I wonder what's wrong.. will check I think the user space needs to enable the event once the cpu gets online.. which we dont do and your app does..? maybe we could add perf_event_attr::enable_on_online ;-) I'll check what we can do in user space, I guess we can monitor the cpu state and enable event accordingly jirka Yes, probably that's the reason. In order for an event to get scheduled-in, it expects the event to be at least in PERF_EVENT_STATE_INACTIVE state. If you notice, in my patch, when the cpu wakes up, we are initializing the state of the event (perf_event__state_init()) and then trying to schedule-in. Since the event was created with a disabled state, it seems that the same this is followed and the state gets initialized to PERF_EVENT_STATE_OFF. Unfortunately, events in this state could not be scheduled. One way for things to get working is, instead of calling perf_event__state_init() before the event is scheduled-in (when the cpu wakes up), we can do something like: perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); I made this change and ran the same test as yours, and I see things working out for us: # ./perf stat -C 1 -e sched:sched_switch -v -I 1000 failed to read counter sched:sched_switch # time counts unit events 1.000115547sched:sched_switch failed to read counter sched:sched_switch 2.000265492sched:sched_switch failed to read counter sched:sched_switch 3.000379462sched:sched_switch failed to read counter sched:sched_switch 4.000523872sched:sched_switch failed to read counter sched:sched_switch 5.000614808sched:sched_switch /* CPU bought ONLINE here */ sched:sched_switch: 541 284808940 284808940 6.000767761541 sched:sched_switch sched:sched_switch: 180 1000119686 1000119686 7.000907234180 sched:sched_switch sched:sched_switch: 248 1000129929 1000129929 8.001026518248 sched:sched_switch sched:sched_switch: 253 1000173050 1000173050 9.001203689253 sched:sched_switch sched:sched_switch: 620 1000113378 1000113378 10.001323334620 sched:sched_switch sched:sched_switch: 366 1000121839 1000121839 11.001448354366 sched:sched_switch sched:sched_switch: 327 1000147664 1000147664 12.001591432327 sched:sched_switch ^Csched:sched_switch: 272 488810681 488810681 12.490414290272 sched:sched_switch sched:sched_switch: 6 75893 75893 Yes, so as you mentioned adding something like perf_event_attr::enable_on_online gives us a control as to put the event in INACTIVE state. -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
On 02/12/2018 01:21 PM, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 10:04:42PM +0100, Jiri Olsa wrote: On Mon, Feb 12, 2018 at 09:42:05AM -0800, Raghavendra Rao Ananta wrote: Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. nope, we don't check on that Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? hum, I wonder what's wrong.. will check I think the user space needs to enable the event once the cpu gets online.. which we dont do and your app does..? maybe we could add perf_event_attr::enable_on_online ;-) I'll check what we can do in user space, I guess we can monitor the cpu state and enable event accordingly jirka Yes, probably that's the reason. In order for an event to get scheduled-in, it expects the event to be at least in PERF_EVENT_STATE_INACTIVE state. If you notice, in my patch, when the cpu wakes up, we are initializing the state of the event (perf_event__state_init()) and then trying to schedule-in. Since the event was created with a disabled state, it seems that the same this is followed and the state gets initialized to PERF_EVENT_STATE_OFF. Unfortunately, events in this state could not be scheduled. One way for things to get working is, instead of calling perf_event__state_init() before the event is scheduled-in (when the cpu wakes up), we can do something like: perf_event_set_state(event, PERF_EVENT_STATE_INACTIVE); I made this change and ran the same test as yours, and I see things working out for us: # ./perf stat -C 1 -e sched:sched_switch -v -I 1000 failed to read counter sched:sched_switch # time counts unit events 1.000115547sched:sched_switch failed to read counter sched:sched_switch 2.000265492sched:sched_switch failed to read counter sched:sched_switch 3.000379462sched:sched_switch failed to read counter sched:sched_switch 4.000523872sched:sched_switch failed to read counter sched:sched_switch 5.000614808sched:sched_switch /* CPU bought ONLINE here */ sched:sched_switch: 541 284808940 284808940 6.000767761541 sched:sched_switch sched:sched_switch: 180 1000119686 1000119686 7.000907234180 sched:sched_switch sched:sched_switch: 248 1000129929 1000129929 8.001026518248 sched:sched_switch sched:sched_switch: 253 1000173050 1000173050 9.001203689253 sched:sched_switch sched:sched_switch: 620 1000113378 1000113378 10.001323334620 sched:sched_switch sched:sched_switch: 366 1000121839 1000121839 11.001448354366 sched:sched_switch sched:sched_switch: 327 1000147664 1000147664 12.001591432327 sched:sched_switch ^Csched:sched_switch: 272 488810681 488810681 12.490414290272 sched:sched_switch sched:sched_switch: 6 75893 75893 Yes, so as you mentioned adding something like perf_event_attr::enable_on_online gives us a control as to put the event in INACTIVE state. -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? Find an inline comment. On 02/12/2018 01:43 AM, Jiri Olsa wrote: On Fri, Feb 09, 2018 at 03:07:00PM -0800, Raghavendra Rao Ananta wrote: Perf framework doesn't allow creation of hardware events if the requested CPU is offline. However, creation of an event is achievable if the event is attached to the PMU as soon as the CPU is online again. So, introducing a feature that could allow to create events even when the CPU is offline and return a success to the caller. If, during the time of event creation, the CPU is found offline, the event is moved to a new state (PERF_EVENT_STATE_DORMANT). As and when the CPU is know to be woken up (through hotplug notifiers), all the dormant events would be attached to the PMU (by perf_install_in_context()). If during the life time of the event, the CPU hasn't come online, the dormant event would just be freed. Signed-off-by: Raghavendra Rao Ananta <rana...@codeaurora.org> hum, I tried and for some reason I'm getting zero counts/times when the cpu 1 is set back online (in 9.second) [root@ibm-x3650m4-02 perf]# ./perf stat -C 1 -e sched:sched_switch -v -I 1000 failed to read counter sched:sched_switch # time counts unit events 1.000921624sched:sched_switch failed to read counter sched:sched_switch 2.001725364sched:sched_switch failed to read counter sched:sched_switch 3.002685350sched:sched_switch failed to read counter sched:sched_switch 4.003463851sched:sched_switch failed to read counter sched:sched_switch 5.004651601sched:sched_switch failed to read counter sched:sched_switch 6.005338294sched:sched_switch failed to read counter sched:sched_switch 7.006351155sched:sched_switch failed to read counter sched:sched_switch 8.007239698sched:sched_switch sched:sched_switch: 0 0 0 9.008665621sched:sched_switch sched:sched_switch: 0 0 0 10.009570492sched:sched_switch sched:sched_switch: 0 0 0 11.010811591sched:sched_switch sched:sched_switch: 0 0 0 12.011614182sched:sched_switch sched:sched_switch: 0 0 0 13.012299851sched:sched_switch looks like the dormant event wasn't scheduled in properly also while at it, could we also handle cpu going offline case, so the event would survive until it's back online Sure, I have a plan for that one as well, but wondering how this goes first. jirka Thank you. Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
Re: [PATCH] perf: Add support for creating offline events
Hi Jiri, Thank you for the response. Does perf tool has its own check to see if the CPU was offline during the lifetime of an event? If so, it might ignore these type of events. Initially, I tested the same using perf tool and found similar results. Then I debugged further and found that the perf core was actually sending data to the userspace (copy_to_user()) and the corresponding count for the data. Hence, I tested this further by writing my own userspace application, and I was able to read the count through this, even when the CPU was made offline and back online. Do you think we also have to modify the perf tool accordingly? Find an inline comment. On 02/12/2018 01:43 AM, Jiri Olsa wrote: On Fri, Feb 09, 2018 at 03:07:00PM -0800, Raghavendra Rao Ananta wrote: Perf framework doesn't allow creation of hardware events if the requested CPU is offline. However, creation of an event is achievable if the event is attached to the PMU as soon as the CPU is online again. So, introducing a feature that could allow to create events even when the CPU is offline and return a success to the caller. If, during the time of event creation, the CPU is found offline, the event is moved to a new state (PERF_EVENT_STATE_DORMANT). As and when the CPU is know to be woken up (through hotplug notifiers), all the dormant events would be attached to the PMU (by perf_install_in_context()). If during the life time of the event, the CPU hasn't come online, the dormant event would just be freed. Signed-off-by: Raghavendra Rao Ananta hum, I tried and for some reason I'm getting zero counts/times when the cpu 1 is set back online (in 9.second) [root@ibm-x3650m4-02 perf]# ./perf stat -C 1 -e sched:sched_switch -v -I 1000 failed to read counter sched:sched_switch # time counts unit events 1.000921624sched:sched_switch failed to read counter sched:sched_switch 2.001725364sched:sched_switch failed to read counter sched:sched_switch 3.002685350sched:sched_switch failed to read counter sched:sched_switch 4.003463851sched:sched_switch failed to read counter sched:sched_switch 5.004651601sched:sched_switch failed to read counter sched:sched_switch 6.005338294sched:sched_switch failed to read counter sched:sched_switch 7.006351155sched:sched_switch failed to read counter sched:sched_switch 8.007239698sched:sched_switch sched:sched_switch: 0 0 0 9.008665621sched:sched_switch sched:sched_switch: 0 0 0 10.009570492sched:sched_switch sched:sched_switch: 0 0 0 11.010811591sched:sched_switch sched:sched_switch: 0 0 0 12.011614182sched:sched_switch sched:sched_switch: 0 0 0 13.012299851sched:sched_switch looks like the dormant event wasn't scheduled in properly also while at it, could we also handle cpu going offline case, so the event would survive until it's back online Sure, I have a plan for that one as well, but wondering how this goes first. jirka Thank you. Raghavendra -- Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project
[PATCH] perf: Add support for creating offline events
Perf framework doesn't allow creation of hardware events if the requested CPU is offline. However, creation of an event is achievable if the event is attached to the PMU as soon as the CPU is online again. So, introducing a feature that could allow to create events even when the CPU is offline and return a success to the caller. If, during the time of event creation, the CPU is found offline, the event is moved to a new state (PERF_EVENT_STATE_DORMANT). As and when the CPU is know to be woken up (through hotplug notifiers), all the dormant events would be attached to the PMU (by perf_install_in_context()). If during the life time of the event, the CPU hasn't come online, the dormant event would just be freed. Signed-off-by: Raghavendra Rao Ananta <rana...@codeaurora.org> --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 109 - 2 files changed, 84 insertions(+), 32 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_headsb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; #endif /* CONFIG_PERF_EVENTS */ }; diff --git a/kernel/events/core.c b/kernel/events/core.c index f0549e7..66dbccb 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2329,6 +2329,19 @@ static int __perf_install_in_context(void *info) return ret; } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static LIST_HEAD(dormant_event_list); +static DEFINE_SPINLOCK(dormant_event_list_lock); + +static void perf_prepare_install_in_context(struct perf_event *event) +{ + spin_lock(_event_list_lock); + event->state = PERF_EVENT_STATE_DORMANT; + list_add_tail(>dormant_entry, _event_list); + spin_unlock(_event_list_lock); +} +#endif + /* * Attach a performance event to a context. * @@ -2353,6 +2366,15 @@ static int __perf_install_in_context(void *info) smp_store_release(>ctx, ctx); if (!task) { +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + struct perf_cpu_context *cpuctx = + container_of(ctx, struct perf_cpu_context, ctx); + + if (!cpuctx->online) { + perf_prepare_install_in_context(event); + return; + } +#endif cpu_function_call(cpu, __perf_install_in_context, event); return; } @@ -2421,6 +2443,43 @@ static int __perf_install_in_context(void *info) raw_spin_unlock_irq(>lock); } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static void perf_deferred_install_in_context(int cpu) +{ + struct perf_event *event, *tmp; + struct perf_event_context *ctx; + + /* This function is called twice while coming online. Once for +* CPUHP_PERF_PREPARE and the other for CPUHP_AP_PERF_ONLINE. +* Only during the CPUHP_AP_PERF_ONLINE state, we can confirm +* that CPU PMU is ready and can be installed to. +*/ + if (!cpu_online(cpu)) + return; + + spin_lock(_event_list_lock); + list_for_each_entry_safe(event, tmp, _event_list, + dormant_entry) { + if (cpu != event->cpu) + continue; + + list_del(>dormant_entry); + event->state = PERF_EVENT_STATE_INACTIVE; + spin_unlock(_event_list_lock); + + ctx = event->ctx; + perf_event__state_init(event); + + mutex_lock(>mutex); + perf_install_in_context(ctx, event, cpu); + mutex_unlock(>mutex); + + spin_lock(_event_list_lock); + } + spin_unlock(_event_list_lock); +} +#endif + /* * Cross CPU call to enable a performance event */ @@ -4202,6 +4261,13 @@ int perf_event_release_kernel(struct perf_event *event) struct perf_event *child, *tmp; LIST_HEAD(free_list); +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + spin_lock(_event_list_lock); + if (event->state == PERF_EVENT_STATE_DORMANT) + list_del(>dormant_entry); + spin_unlock(_event_list_lock); +#endif + /*
[PATCH] perf: Add support for creating offline events
Perf framework doesn't allow creation of hardware events if the requested CPU is offline. However, creation of an event is achievable if the event is attached to the PMU as soon as the CPU is online again. So, introducing a feature that could allow to create events even when the CPU is offline and return a success to the caller. If, during the time of event creation, the CPU is found offline, the event is moved to a new state (PERF_EVENT_STATE_DORMANT). As and when the CPU is know to be woken up (through hotplug notifiers), all the dormant events would be attached to the PMU (by perf_install_in_context()). If during the life time of the event, the CPU hasn't come online, the dormant event would just be freed. Signed-off-by: Raghavendra Rao Ananta --- include/linux/perf_event.h | 7 +++ kernel/events/core.c | 109 - 2 files changed, 84 insertions(+), 32 deletions(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7546822..bc07f16 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -489,6 +489,7 @@ struct perf_addr_filters_head { * enum perf_event_state - the states of a event */ enum perf_event_state { + PERF_EVENT_STATE_DORMANT= -5, PERF_EVENT_STATE_DEAD = -4, PERF_EVENT_STATE_EXIT = -3, PERF_EVENT_STATE_ERROR = -2, @@ -687,6 +688,12 @@ struct perf_event { #endif struct list_headsb_list; + + /* Entry into the list that holds the events whose CPUs +* are offline. These events will be removed from the +* list and installed once the CPU wakes up. +*/ + struct list_headdormant_entry; #endif /* CONFIG_PERF_EVENTS */ }; diff --git a/kernel/events/core.c b/kernel/events/core.c index f0549e7..66dbccb 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2329,6 +2329,19 @@ static int __perf_install_in_context(void *info) return ret; } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static LIST_HEAD(dormant_event_list); +static DEFINE_SPINLOCK(dormant_event_list_lock); + +static void perf_prepare_install_in_context(struct perf_event *event) +{ + spin_lock(_event_list_lock); + event->state = PERF_EVENT_STATE_DORMANT; + list_add_tail(>dormant_entry, _event_list); + spin_unlock(_event_list_lock); +} +#endif + /* * Attach a performance event to a context. * @@ -2353,6 +2366,15 @@ static int __perf_install_in_context(void *info) smp_store_release(>ctx, ctx); if (!task) { +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + struct perf_cpu_context *cpuctx = + container_of(ctx, struct perf_cpu_context, ctx); + + if (!cpuctx->online) { + perf_prepare_install_in_context(event); + return; + } +#endif cpu_function_call(cpu, __perf_install_in_context, event); return; } @@ -2421,6 +2443,43 @@ static int __perf_install_in_context(void *info) raw_spin_unlock_irq(>lock); } +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE +static void perf_deferred_install_in_context(int cpu) +{ + struct perf_event *event, *tmp; + struct perf_event_context *ctx; + + /* This function is called twice while coming online. Once for +* CPUHP_PERF_PREPARE and the other for CPUHP_AP_PERF_ONLINE. +* Only during the CPUHP_AP_PERF_ONLINE state, we can confirm +* that CPU PMU is ready and can be installed to. +*/ + if (!cpu_online(cpu)) + return; + + spin_lock(_event_list_lock); + list_for_each_entry_safe(event, tmp, _event_list, + dormant_entry) { + if (cpu != event->cpu) + continue; + + list_del(>dormant_entry); + event->state = PERF_EVENT_STATE_INACTIVE; + spin_unlock(_event_list_lock); + + ctx = event->ctx; + perf_event__state_init(event); + + mutex_lock(>mutex); + perf_install_in_context(ctx, event, cpu); + mutex_unlock(>mutex); + + spin_lock(_event_list_lock); + } + spin_unlock(_event_list_lock); +} +#endif + /* * Cross CPU call to enable a performance event */ @@ -4202,6 +4261,13 @@ int perf_event_release_kernel(struct perf_event *event) struct perf_event *child, *tmp; LIST_HEAD(free_list); +#if defined CONFIG_HOTPLUG_CPU || defined CONFIG_KEXEC_CORE + spin_lock(_event_list_lock); + if (event->state == PERF_EVENT_STATE_DORMANT) + list_del(>dormant_entry); + spin_unlock(_event_list_lock); +#endif + /* * If we got here