1 # SPDX-License-Identifier: GPL-2.0-only 1 # SPDX-License-Identifier: GPL-2.0-only
2 2
3 config PREEMPT_NONE_BUILD 3 config PREEMPT_NONE_BUILD
4 bool 4 bool
5 5
6 config PREEMPT_VOLUNTARY_BUILD 6 config PREEMPT_VOLUNTARY_BUILD
7 bool 7 bool
8 8
9 config PREEMPT_BUILD 9 config PREEMPT_BUILD
10 bool 10 bool
11 select PREEMPTION 11 select PREEMPTION
12 select UNINLINE_SPIN_UNLOCK if !ARCH_I 12 select UNINLINE_SPIN_UNLOCK if !ARCH_INLINE_SPIN_UNLOCK
13 13
14 choice 14 choice
15 prompt "Preemption Model" 15 prompt "Preemption Model"
16 default PREEMPT_NONE 16 default PREEMPT_NONE
17 17
18 config PREEMPT_NONE 18 config PREEMPT_NONE
19 bool "No Forced Preemption (Server)" 19 bool "No Forced Preemption (Server)"
20 select PREEMPT_NONE_BUILD if !PREEMPT_ 20 select PREEMPT_NONE_BUILD if !PREEMPT_DYNAMIC
21 help 21 help
22 This is the traditional Linux preemp 22 This is the traditional Linux preemption model, geared towards
23 throughput. It will still provide go 23 throughput. It will still provide good latencies most of the
24 time, but there are no guarantees an 24 time, but there are no guarantees and occasional longer delays
25 are possible. 25 are possible.
26 26
27 Select this option if you are buildi 27 Select this option if you are building a kernel for a server or
28 scientific/computation system, or if 28 scientific/computation system, or if you want to maximize the
29 raw processing power of the kernel, 29 raw processing power of the kernel, irrespective of scheduling
30 latencies. 30 latencies.
31 31
32 config PREEMPT_VOLUNTARY 32 config PREEMPT_VOLUNTARY
33 bool "Voluntary Kernel Preemption (Des 33 bool "Voluntary Kernel Preemption (Desktop)"
34 depends on !ARCH_NO_PREEMPT 34 depends on !ARCH_NO_PREEMPT
35 select PREEMPT_VOLUNTARY_BUILD if !PRE 35 select PREEMPT_VOLUNTARY_BUILD if !PREEMPT_DYNAMIC
36 help 36 help
37 This option reduces the latency of t 37 This option reduces the latency of the kernel by adding more
38 "explicit preemption points" to the 38 "explicit preemption points" to the kernel code. These new
39 preemption points have been selected 39 preemption points have been selected to reduce the maximum
40 latency of rescheduling, providing f 40 latency of rescheduling, providing faster application reactions,
41 at the cost of slightly lower throug 41 at the cost of slightly lower throughput.
42 42
43 This allows reaction to interactive 43 This allows reaction to interactive events by allowing a
44 low priority process to voluntarily 44 low priority process to voluntarily preempt itself even if it
45 is in kernel mode executing a system 45 is in kernel mode executing a system call. This allows
46 applications to run more 'smoothly' 46 applications to run more 'smoothly' even when the system is
47 under load. 47 under load.
48 48
49 Select this if you are building a ke 49 Select this if you are building a kernel for a desktop system.
50 50
51 config PREEMPT 51 config PREEMPT
52 bool "Preemptible Kernel (Low-Latency 52 bool "Preemptible Kernel (Low-Latency Desktop)"
53 depends on !ARCH_NO_PREEMPT 53 depends on !ARCH_NO_PREEMPT
54 select PREEMPT_BUILD 54 select PREEMPT_BUILD
55 help 55 help
56 This option reduces the latency of t 56 This option reduces the latency of the kernel by making
57 all kernel code (that is not executi 57 all kernel code (that is not executing in a critical section)
58 preemptible. This allows reaction t 58 preemptible. This allows reaction to interactive events by
59 permitting a low priority process to 59 permitting a low priority process to be preempted involuntarily
60 even if it is in kernel mode executi 60 even if it is in kernel mode executing a system call and would
61 otherwise not be about to reach a na 61 otherwise not be about to reach a natural preemption point.
62 This allows applications to run more 62 This allows applications to run more 'smoothly' even when the
63 system is under load, at the cost of 63 system is under load, at the cost of slightly lower throughput
64 and a slight runtime overhead to ker 64 and a slight runtime overhead to kernel code.
65 65
66 Select this if you are building a ke 66 Select this if you are building a kernel for a desktop or
67 embedded system with latency require 67 embedded system with latency requirements in the milliseconds
68 range. 68 range.
69 69
70 config PREEMPT_RT 70 config PREEMPT_RT
71 bool "Fully Preemptible Kernel (Real-T 71 bool "Fully Preemptible Kernel (Real-Time)"
72 depends on EXPERT && ARCH_SUPPORTS_RT 72 depends on EXPERT && ARCH_SUPPORTS_RT
73 select PREEMPTION 73 select PREEMPTION
74 help 74 help
75 This option turns the kernel into a 75 This option turns the kernel into a real-time kernel by replacing
76 various locking primitives (spinlock 76 various locking primitives (spinlocks, rwlocks, etc.) with
77 preemptible priority-inheritance awa 77 preemptible priority-inheritance aware variants, enforcing
78 interrupt threading and introducing 78 interrupt threading and introducing mechanisms to break up long
79 non-preemptible sections. This makes 79 non-preemptible sections. This makes the kernel, except for very
80 low level and critical code paths (e 80 low level and critical code paths (entry code, scheduler, low
81 level interrupt handling) fully pree 81 level interrupt handling) fully preemptible and brings most
82 execution contexts under scheduler c 82 execution contexts under scheduler control.
83 83
84 Select this if you are building a ke 84 Select this if you are building a kernel for systems which
85 require real-time guarantees. 85 require real-time guarantees.
86 86
87 endchoice 87 endchoice
88 88
89 config PREEMPT_COUNT 89 config PREEMPT_COUNT
90 bool 90 bool
91 91
92 config PREEMPTION 92 config PREEMPTION
93 bool 93 bool
94 select PREEMPT_COUNT 94 select PREEMPT_COUNT
95 95
96 config PREEMPT_DYNAMIC 96 config PREEMPT_DYNAMIC
97 bool "Preemption behaviour defined on 97 bool "Preemption behaviour defined on boot"
98 depends on HAVE_PREEMPT_DYNAMIC && !PR 98 depends on HAVE_PREEMPT_DYNAMIC && !PREEMPT_RT
99 select JUMP_LABEL if HAVE_PREEMPT_DYNA 99 select JUMP_LABEL if HAVE_PREEMPT_DYNAMIC_KEY
100 select PREEMPT_BUILD 100 select PREEMPT_BUILD
101 default y if HAVE_PREEMPT_DYNAMIC_CALL 101 default y if HAVE_PREEMPT_DYNAMIC_CALL
102 help 102 help
103 This option allows to define the pre 103 This option allows to define the preemption model on the kernel
104 command line parameter and thus over 104 command line parameter and thus override the default preemption
105 model defined during compile time. 105 model defined during compile time.
106 106
107 The feature is primarily interesting 107 The feature is primarily interesting for Linux distributions which
108 provide a pre-built kernel binary to 108 provide a pre-built kernel binary to reduce the number of kernel
109 flavors they offer while still offer 109 flavors they offer while still offering different usecases.
110 110
111 The runtime overhead is negligible w 111 The runtime overhead is negligible with HAVE_STATIC_CALL_INLINE enabled
112 but if runtime patching is not avail 112 but if runtime patching is not available for the specific architecture
113 then the potential overhead should b 113 then the potential overhead should be considered.
114 114
115 Interesting if you want the same pre 115 Interesting if you want the same pre-built kernel should be used for
116 both Server and Desktop workloads. 116 both Server and Desktop workloads.
117 117
118 config SCHED_CORE 118 config SCHED_CORE
119 bool "Core Scheduling for SMT" 119 bool "Core Scheduling for SMT"
120 depends on SCHED_SMT 120 depends on SCHED_SMT
121 help 121 help
122 This option permits Core Scheduling, 122 This option permits Core Scheduling, a means of coordinated task
123 selection across SMT siblings. When 123 selection across SMT siblings. When enabled -- see
124 prctl(PR_SCHED_CORE) -- task selecti 124 prctl(PR_SCHED_CORE) -- task selection ensures that all SMT siblings
125 will execute a task from the same 'c 125 will execute a task from the same 'core group', forcing idle when no
126 matching task is found. 126 matching task is found.
127 127
128 Use of this feature includes: 128 Use of this feature includes:
129 - mitigation of some (not all) SMT 129 - mitigation of some (not all) SMT side channels;
130 - limiting SMT interference to impr 130 - limiting SMT interference to improve determinism and/or performance.
131 131
132 SCHED_CORE is default disabled. When 132 SCHED_CORE is default disabled. When it is enabled and unused,
133 which is the likely usage by Linux d 133 which is the likely usage by Linux distributions, there should
134 be no measurable impact on performan 134 be no measurable impact on performance.
135 135
136 config SCHED_CLASS_EXT 136 config SCHED_CLASS_EXT
137 bool "Extensible Scheduling Class" 137 bool "Extensible Scheduling Class"
138 depends on BPF_SYSCALL && BPF_JIT && D 138 depends on BPF_SYSCALL && BPF_JIT && DEBUG_INFO_BTF
139 select STACKTRACE if STACKTRACE_SUPPOR 139 select STACKTRACE if STACKTRACE_SUPPORT
140 help 140 help
141 This option enables a new scheduler 141 This option enables a new scheduler class sched_ext (SCX), which
142 allows scheduling policies to be imp 142 allows scheduling policies to be implemented as BPF programs to
143 achieve the following: 143 achieve the following:
144 144
145 - Ease of experimentation and explor 145 - Ease of experimentation and exploration: Enabling rapid
146 iteration of new scheduling polici 146 iteration of new scheduling policies.
147 - Customization: Building applicatio 147 - Customization: Building application-specific schedulers which
148 implement policies that are not ap 148 implement policies that are not applicable to general-purpose
149 schedulers. 149 schedulers.
150 - Rapid scheduler deployments: Non-d 150 - Rapid scheduler deployments: Non-disruptive swap outs of
151 scheduling policies in production 151 scheduling policies in production environments.
152 152
153 sched_ext leverages BPF struct_ops f 153 sched_ext leverages BPF struct_ops feature to define a structure
154 which exports function callbacks and 154 which exports function callbacks and flags to BPF programs that
155 wish to implement scheduling policie 155 wish to implement scheduling policies. The struct_ops structure
156 exported by sched_ext is struct sche 156 exported by sched_ext is struct sched_ext_ops, and is conceptually
157 similar to struct sched_class. 157 similar to struct sched_class.
158 158
159 For more information: 159 For more information:
160 Documentation/scheduler/sched-ext. 160 Documentation/scheduler/sched-ext.rst
161 https://github.com/sched-ext/scx 161 https://github.com/sched-ext/scx
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.