| /kernel/linux/linux-6.6/Documentation/devicetree/bindings/cpu/ |
| D | idle-states.yaml | 1 # SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
3 ---
4 $id: http://devicetree.org/schemas/cpu/idle-states.yaml#
5 $schema: http://devicetree.org/meta-schemas/core.yaml#
10 - Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
11 - Anup Patel <anup@brainfault.org>
15 1 - Introduction
18 ARM and RISC-V systems contain HW capable of managing power consumption
19 dynamically, where cores can be put in different low-power states (ranging
22 run-time, can be specified through device tree bindings representing the
[all …]
|
| /kernel/linux/linux-5.10/Documentation/devicetree/bindings/arm/ |
| D | idle-states.yaml | 1 # SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
3 ---
4 $id: http://devicetree.org/schemas/arm/idle-states.yaml#
5 $schema: http://devicetree.org/meta-schemas/core.yaml#
10 - Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
14 1 - Introduction
18 where cores can be put in different low-power states (ranging from simple wfi
20 range of dynamic idle states that a processor can enter at run-time, can be
27 - Running
28 - Idle_standby
[all …]
|
| /kernel/linux/linux-5.10/drivers/cpuidle/ |
| D | cpuidle-pseries.c | 1 // SPDX-License-Identifier: GPL-2.0
3 * cpuidle-pseries - idle state cpuidle driver.
78 * were soft-disabled in check_and_cede_processor()
92 * "ibm,get-systems-parameter" RTAS call with the token
98 * table with all the parameters to ibm,get-system-parameters.
99 * CEDE_LATENCY_TOKEN corresponds to the token value for Cede Latency
105 * If the platform supports the cede latency settings information system
109 * a. The first byte is the length “N” of each cede latency setting record minus
112 * b. For each supported cede latency setting a cede latency setting record
115 * -----------------------------
[all …]
|
| D | dt_idle_states.c | 1 // SPDX-License-Identifier: GPL-2.0-only
9 #define pr_fmt(fmt) "DT idle-states: " fmt
33 idle_state->enter = match_id->data; in init_state_node()
39 idle_state->enter_s2idle = match_id->data; in init_state_node()
41 err = of_property_read_u32(state_node, "wakeup-latency-us", in init_state_node()
42 &idle_state->exit_latency); in init_state_node()
46 err = of_property_read_u32(state_node, "entry-latency-us", in init_state_node()
49 pr_debug(" * %pOF missing entry-latency-us property\n", in init_state_node()
51 return -EINVAL; in init_state_node()
54 err = of_property_read_u32(state_node, "exit-latency-us", in init_state_node()
[all …]
|
| /kernel/linux/linux-6.6/drivers/cpuidle/ |
| D | dt_idle_states.c | 1 // SPDX-License-Identifier: GPL-2.0-only
9 #define pr_fmt(fmt) "DT idle-states: " fmt
32 idle_state->enter = match_id->data; in init_state_node()
38 idle_state->enter_s2idle = match_id->data; in init_state_node()
40 err = of_property_read_u32(state_node, "wakeup-latency-us", in init_state_node()
41 &idle_state->exit_latency); in init_state_node()
45 err = of_property_read_u32(state_node, "entry-latency-us", in init_state_node()
48 pr_debug(" * %pOF missing entry-latency-us property\n", in init_state_node()
50 return -EINVAL; in init_state_node()
53 err = of_property_read_u32(state_node, "exit-latency-us", in init_state_node()
[all …]
|
| D | cpuidle-pseries.c | 1 // SPDX-License-Identifier: GPL-2.0
3 * cpuidle-pseries - idle state cpuidle driver.
47 dev->poll_time_limit = false; in snooze_loop()
58 dev->poll_time_limit = true; in snooze_loop()
80 * were soft-disabled in check_and_cede_processor()
94 * "ibm,get-systems-parameter" RTAS call with the token
100 * table with all the parameters to ibm,get-system-parameters.
101 * CEDE_LATENCY_TOKEN corresponds to the token value for Cede Latency
107 * If the platform supports the cede latency settings information system
111 * a. The first byte is the length “N” of each cede latency setting record minus
[all …]
|
| /kernel/linux/linux-6.6/Documentation/tools/rtla/ |
| D | rtla-timerlat-top.rst | 2 rtla-timerlat-top
4 -------------------------------------------
5 Measures the operating system timer latency
6 -------------------------------------------
22 seem with the option **-T**.
35 **--aa-only** *us*
38 Print the auto-analysis if the system hits the stop tracing condition. This option
45 In the example below, the timerlat tracer is dispatched in cpus *1-23* in the
46 automatic trace mode, instructing the tracer to stop if a *40 us* latency or
49 # timerlat -a 40 -c 1-23 -q
[all …]
|
| /kernel/linux/linux-6.6/Documentation/trace/ |
| D | ftrace.rst | 2 ftrace - Function Tracer
13 - Written for: 2.6.28-rc2
14 - Updated for: 3.10
15 - Updated for: 4.13 - Copyright 2017 VMware Inc. Steven Rostedt
16 - Converted to rst format - Changbin Du <changbin.du@intel.com>
19 ------------
24 performance issues that take place outside of user-space.
28 There's latency tracing to examine what occurs between interrupts
41 ----------------------
43 See Documentation/trace/ftrace-design.rst for details for arch porters and such.
[all …]
|
| D | timerlat-tracer.rst | 6 find sources of wakeup latencies of real-time threads. Like cyclictest,
8 computes a *wakeup latency* value as the difference between the *current
13 -----
28 # _-----=> irqs-off
29 # / _----=> need-resched
30 # | / _---=> hardirq/softirq
31 # || / _--=> preempt-depth
34 # TASK-PID CPU# |||| TIMESTAMP ID CONTEXT LATENCY
36 <idle>-0 [000] d.h1 54.029328: #1 context irq timer_latency 932 ns
37 <...>-867 [000] .... 54.029339: #1 context thread timer_latency 11700 ns
[all …]
|
| /kernel/linux/linux-5.10/Documentation/trace/ |
| D | ftrace.rst | 2 ftrace - Function Tracer
13 - Written for: 2.6.28-rc2
14 - Updated for: 3.10
15 - Updated for: 4.13 - Copyright 2017 VMware Inc. Steven Rostedt
16 - Converted to rst format - Changbin Du <changbin.du@intel.com>
19 ------------
24 performance issues that take place outside of user-space.
28 There's latency tracing to examine what occurs between interrupts
41 ----------------------
43 See :doc:`ftrace-design` for details for arch porters and such.
[all …]
|
| /kernel/linux/linux-5.10/Documentation/virt/kvm/ |
| D | halt-polling.rst | 1 .. SPDX-License-Identifier: GPL-2.0
7 The KVM halt polling system provides a feature within KVM whereby the latency
11 vcpus of a single vcore have ceded, the host kernel polls for wakeup conditions
14 Polling provides a latency advantage in cases where the guest can be run again
15 very quickly by at least saving us a trip through the scheduler, normally on
16 the order of a few micro-seconds, although performance benefits are workload
17 dependant. In the event that no wakeup source arrives during the polling
20 wakeup periods where the time spent halt polling is minimised and the time
27 The powerpc kvm-hv specific case is implemented in:
39 kvm_vcpu->halt_poll_ns
[all …]
|
| /kernel/linux/linux-6.6/Documentation/virt/kvm/ |
| D | halt-polling.rst | 1 .. SPDX-License-Identifier: GPL-2.0
7 The KVM halt polling system provides a feature within KVM whereby the latency
11 vcpus of a single vcore have ceded, the host kernel polls for wakeup conditions
14 Polling provides a latency advantage in cases where the guest can be run again
15 very quickly by at least saving us a trip through the scheduler, normally on
16 the order of a few micro-seconds, although performance benefits are workload
17 dependent. In the event that no wakeup source arrives during the polling
20 wakeup periods where the time spent halt polling is minimised and the time
27 The powerpc kvm-hv specific case is implemented in:
39 kvm_vcpu->halt_poll_ns
[all …]
|
| /kernel/linux/linux-5.10/drivers/gpu/drm/i915/gt/ |
| D | intel_rc6.c | 2 * SPDX-License-Identifier: MIT
20 * low-voltage mode when idle, using down to 0V while at this stage. This
26 * among each other with the latency required to enter and leave RC6 and
34 * require higher latency to switch to and wake up.
44 return rc6_to_gt(rc)->uncore; in rc6_to_uncore()
49 return rc6_to_gt(rc)->i915; in rc6_to_i915()
60 struct intel_uncore *uncore = gt->uncore; in gen11_rc6_enable()
73 set(uncore, RING_MAX_IDLE(engine->mmio_base), 10); in gen11_rc6_enable()
84 * Bspec's guidance is to use 25us (really 25 * 1280ns) here. What we in gen11_rc6_enable()
86 * it takes us to service a CS interrupt and submit a new ELSP - that in gen11_rc6_enable()
[all …]
|
| /kernel/linux/linux-6.6/drivers/gpu/drm/i915/gt/ |
| D | intel_rc6.c | 1 // SPDX-License-Identifier: MIT
24 * low-voltage mode when idle, using down to 0V while at this stage. This
30 * among each other with the latency required to enter and leave RC6 and
38 * require higher latency to switch to and wake up.
48 return rc6_to_gt(rc)->uncore; in rc6_to_uncore()
53 return rc6_to_gt(rc)->i915; in rc6_to_i915()
59 struct intel_uncore *uncore = gt->uncore; in gen11_rc6_enable()
68 if (!intel_uc_uses_guc_rc(>->uc)) { in gen11_rc6_enable()
76 intel_uncore_write_fw(uncore, RING_MAX_IDLE(engine->mmio_base), 10); in gen11_rc6_enable()
88 * Bspec's guidance is to use 25us (really 25 * 1280ns) here. What we in gen11_rc6_enable()
[all …]
|
| /kernel/linux/linux-5.10/drivers/cpuidle/governors/ |
| D | menu.c | 1 // SPDX-License-Identifier: GPL-2.0-only
3 * menu.c - the menu idle governor
5 * Copyright (C) 2006-2007 Adam Belay <abelay@novell.com>
36 * 3) Latency tolerance (from pmqos infrastructure)
40 * -----------------------
43 * provides us this duration in the "target_residency" field. So all that we
68 * Repeatable-interval-detector
69 * ----------------------------
79 * ---------------------------
88 * This rule-of-thumb is implemented using a performance-multiplier:
[all …]
|
| /kernel/linux/linux-6.6/arch/arm64/boot/dts/sprd/ |
| D | sc9860.dtsi | 6 * SPDX-License-Identifier: (GPL-2.0+ OR MIT)
9 #include <dt-bindings/interrupt-controller/arm-gic.h>
10 #include <dt-bindings/input/input.h>
11 #include <dt-bindings/gpio/gpio.h>
16 #address-cells = <2>;
17 #size-cells = <0>;
19 cpu-map {
53 compatible = "arm,cortex-a53";
55 enable-method = "psci";
56 cpu-idle-states = <&CORE_PD &CLUSTER_PD>;
[all …]
|
| /kernel/linux/linux-5.10/arch/arm64/boot/dts/sprd/ |
| D | sc9860.dtsi | 6 * SPDX-License-Identifier: (GPL-2.0+ OR MIT)
9 #include <dt-bindings/interrupt-controller/arm-gic.h>
10 #include <dt-bindings/input/input.h>
11 #include <dt-bindings/gpio/gpio.h>
16 #address-cells = <2>;
17 #size-cells = <0>;
19 cpu-map {
53 compatible = "arm,cortex-a53";
55 enable-method = "psci";
56 cpu-idle-states = <&CORE_PD &CLUSTER_PD>;
[all …]
|
| /kernel/linux/linux-6.6/drivers/cpuidle/governors/ |
| D | menu.c | 1 // SPDX-License-Identifier: GPL-2.0-only
3 * menu.c - the menu idle governor
5 * Copyright (C) 2006-2007 Adam Belay <abelay@novell.com>
38 * 3) Latency tolerance (from pmqos infrastructure)
42 * -----------------------
45 * provides us this duration in the "target_residency" field. So all that we
70 * Repeatable-interval-detector
71 * ----------------------------
81 * ---------------------------
90 * This rule-of-thumb is implemented using a performance-multiplier:
[all …]
|
| /kernel/linux/linux-6.6/tools/testing/selftests/bpf/benchs/ |
| D | bench_local_storage_rcu_tasks_trace.c | 1 // SPDX-License-Identifier: GPL-2.0
72 fprintf(stderr, "benchmark doesn't support multi-producer!\n"); in validate()
94 return -1; in kthread_pid_ticks()
169 if (!bpf_program__attach(ctx.skel->progs.get_local)) { in local_storage_tasks_trace_setup()
174 if (!bpf_program__attach(ctx.skel->progs.pregp_step)) { in local_storage_tasks_trace_setup()
179 if (!bpf_program__attach(ctx.skel->progs.postgp)) { in local_storage_tasks_trace_setup()
193 res->gp_ct = atomic_swap(&ctx.skel->bss->gp_hits, 0); in measure()
194 res->gp_ns = atomic_swap(&ctx.skel->bss->gp_times, 0); in measure()
196 res->stime = ticks - ctx.prev_kthread_stime; in measure()
209 if (ctx.skel->bss->unexpected) { in report_progress()
[all …]
|
| /kernel/linux/linux-5.10/Documentation/driver-api/thermal/ |
| D | cpu-idle-cooling.rst | 1 .. SPDX-License-Identifier: GPL-2.0
8 ----------
26 budget lower than the requested one and under-utilize the CPU, thus
27 losing performance. In other words, one OPP under-utilizes the CPU
33 ----------
58 ---------------
67 latencies as the CPUs will have to wakeup from a deep sleep state.
70 performance penalty and a fixed latency. Mitigation can be increased
78 |------- -------
81 <------>
[all …]
|
| /kernel/linux/linux-6.6/Documentation/driver-api/thermal/ |
| D | cpu-idle-cooling.rst | 1 .. SPDX-License-Identifier: GPL-2.0
8 ----------
26 budget lower than the requested one and under-utilize the CPU, thus
27 losing performance. In other words, one OPP under-utilizes the CPU
33 ----------
58 ---------------
67 latencies as the CPUs will have to wakeup from a deep sleep state.
70 performance penalty and a fixed latency. Mitigation can be increased
78 |------- -------
81 <------>
[all …]
|
| /kernel/linux/linux-6.6/arch/arm64/boot/dts/qcom/ |
| D | sdx75.dtsi | 1 // SPDX-License-Identifier: BSD-3-Clause
9 #include <dt-bindings/clock/qcom,rpmh.h>
10 #include <dt-bindings/clock/qcom,sdx75-gcc.h>
11 #include <dt-bindings/interrupt-controller/arm-gic.h>
12 #include <dt-bindings/power/qcom,rpmhpd.h>
13 #include <dt-bindings/power/qcom-rpmpd.h>
14 #include <dt-bindings/soc/qcom,rpmh-rsc.h>
17 #address-cells = <2>;
18 #size-cells = <2>;
19 interrupt-parent = <&intc>;
[all …]
|
| /kernel/linux/linux-6.6/drivers/base/power/ |
| D | domain_governor.c | 1 // SPDX-License-Identifier: GPL-2.0
3 * drivers/base/power/domain_governor.c - Governors for device PM domains.
21 if (dev->power.subsys_data && dev->power.subsys_data->domain_data) { in dev_update_qos_constraint()
22 struct gpd_timing_data *td = dev_gpd_data(dev)->td; in dev_update_qos_constraint()
25 * Only take suspend-time QoS constraints of devices into in dev_update_qos_constraint()
31 constraint_ns = td ? td->effective_constraint_ns : in dev_update_qos_constraint()
51 * default_suspend_ok - Default PM domain governor routine to suspend devices.
56 struct gpd_timing_data *td = dev_gpd_data(dev)->td; in default_suspend_ok()
62 spin_lock_irqsave(&dev->power.lock, flags); in default_suspend_ok()
64 if (!td->constraint_changed) { in default_suspend_ok()
[all …]
|
| /kernel/linux/linux-6.6/Documentation/accounting/ |
| D | psi.rst | 4 PSI - Pressure Stall Information
11 latency spikes, throughput losses, and run the risk of OOM kills.
14 either play it safe and under-utilize their hardware resources, or
23 scarcity aids users in sizing workloads to hardware--or provisioning
38 respective file in /proc/pressure/ -- cpu, memory, and io.
48 The "full" line indicates the share of time in which all non-idle
63 (in us) is tracked and exported as well, to allow detection of latency
75 generate a wakeup event.
83 <some|full> <stall amount in us> <time window in us>
116 Notifications to the userspace are rate-limited to one per tracking window.
[all …]
|
| /kernel/linux/linux-5.10/Documentation/accounting/ |
| D | psi.rst | 4 PSI - Pressure Stall Information
11 latency spikes, throughput losses, and run the risk of OOM kills.
14 either play it safe and under-utilize their hardware resources, or
23 scarcity aids users in sizing workloads to hardware--or provisioning
38 respective file in /proc/pressure/ -- cpu, memory, and io.
52 The "full" line indicates the share of time in which all non-idle
64 (in us) is tracked and exported as well, to allow detection of latency
76 generate a wakeup event.
84 <some|full> <stall amount in us> <time window in us>
113 Notifications to the userspace are rate-limited to one per tracking window.
[all …]
|