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