~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

TOMOYO Linux Cross Reference
Linux/Documentation/scheduler/sched-rt-group.rst

Version: ~ [ linux-6.12-rc7 ] ~ [ linux-6.11.7 ] ~ [ linux-6.10.14 ] ~ [ linux-6.9.12 ] ~ [ linux-6.8.12 ] ~ [ linux-6.7.12 ] ~ [ linux-6.6.60 ] ~ [ linux-6.5.13 ] ~ [ linux-6.4.16 ] ~ [ linux-6.3.13 ] ~ [ linux-6.2.16 ] ~ [ linux-6.1.116 ] ~ [ linux-6.0.19 ] ~ [ linux-5.19.17 ] ~ [ linux-5.18.19 ] ~ [ linux-5.17.15 ] ~ [ linux-5.16.20 ] ~ [ linux-5.15.171 ] ~ [ linux-5.14.21 ] ~ [ linux-5.13.19 ] ~ [ linux-5.12.19 ] ~ [ linux-5.11.22 ] ~ [ linux-5.10.229 ] ~ [ linux-5.9.16 ] ~ [ linux-5.8.18 ] ~ [ linux-5.7.19 ] ~ [ linux-5.6.19 ] ~ [ linux-5.5.19 ] ~ [ linux-5.4.285 ] ~ [ linux-5.3.18 ] ~ [ linux-5.2.21 ] ~ [ linux-5.1.21 ] ~ [ linux-5.0.21 ] ~ [ linux-4.20.17 ] ~ [ linux-4.19.323 ] ~ [ linux-4.18.20 ] ~ [ linux-4.17.19 ] ~ [ linux-4.16.18 ] ~ [ linux-4.15.18 ] ~ [ linux-4.14.336 ] ~ [ linux-4.13.16 ] ~ [ linux-4.12.14 ] ~ [ linux-4.11.12 ] ~ [ linux-4.10.17 ] ~ [ linux-4.9.337 ] ~ [ linux-4.4.302 ] ~ [ linux-3.10.108 ] ~ [ linux-2.6.32.71 ] ~ [ linux-2.6.0 ] ~ [ linux-2.4.37.11 ] ~ [ unix-v6-master ] ~ [ ccs-tools-1.8.12 ] ~ [ policy-sample ] ~
Architecture: ~ [ i386 ] ~ [ alpha ] ~ [ m68k ] ~ [ mips ] ~ [ ppc ] ~ [ sparc ] ~ [ sparc64 ] ~

Diff markup

Differences between /Documentation/scheduler/sched-rt-group.rst (Architecture sparc64) and /Documentation/scheduler/sched-rt-group.rst (Architecture sparc)


  1 ==========================                          1 ==========================
  2 Real-Time group scheduling                          2 Real-Time group scheduling
  3 ==========================                          3 ==========================
  4                                                     4 
  5 .. CONTENTS                                         5 .. CONTENTS
  6                                                     6 
  7    0. WARNING                                       7    0. WARNING
  8    1. Overview                                      8    1. Overview
  9      1.1 The problem                                9      1.1 The problem
 10      1.2 The solution                              10      1.2 The solution
 11    2. The interface                                11    2. The interface
 12      2.1 System-wide settings                      12      2.1 System-wide settings
 13      2.2 Default behaviour                         13      2.2 Default behaviour
 14      2.3 Basis for grouping tasks                  14      2.3 Basis for grouping tasks
 15    3. Future plans                                 15    3. Future plans
 16                                                    16 
 17                                                    17 
 18 0. WARNING                                         18 0. WARNING
 19 ==========                                         19 ==========
 20                                                    20 
 21  Fiddling with these settings can result in an     21  Fiddling with these settings can result in an unstable system, the knobs are
 22  root only and assumes root knows what he is d     22  root only and assumes root knows what he is doing.
 23                                                    23 
 24 Most notable:                                      24 Most notable:
 25                                                    25 
 26  * very small values in sched_rt_period_us can     26  * very small values in sched_rt_period_us can result in an unstable
 27    system when the period is smaller than eith     27    system when the period is smaller than either the available hrtimer
 28    resolution, or the time it takes to handle      28    resolution, or the time it takes to handle the budget refresh itself.
 29                                                    29 
 30  * very small values in sched_rt_runtime_us ca     30  * very small values in sched_rt_runtime_us can result in an unstable
 31    system when the runtime is so small the sys     31    system when the runtime is so small the system has difficulty making
 32    forward progress (NOTE: the migration threa     32    forward progress (NOTE: the migration thread and kstopmachine both
 33    are real-time processes).                       33    are real-time processes).
 34                                                    34 
 35 1. Overview                                        35 1. Overview
 36 ===========                                        36 ===========
 37                                                    37 
 38                                                    38 
 39 1.1 The problem                                    39 1.1 The problem
 40 ---------------                                    40 ---------------
 41                                                    41 
 42 Real-time scheduling is all about determinism,     42 Real-time scheduling is all about determinism, a group has to be able to rely on
 43 the amount of bandwidth (eg. CPU time) being c     43 the amount of bandwidth (eg. CPU time) being constant. In order to schedule
 44 multiple groups of real-time tasks, each group     44 multiple groups of real-time tasks, each group must be assigned a fixed portion
 45 of the CPU time available.  Without a minimum      45 of the CPU time available.  Without a minimum guarantee a real-time group can
 46 obviously fall short. A fuzzy upper limit is o     46 obviously fall short. A fuzzy upper limit is of no use since it cannot be
 47 relied upon. Which leaves us with just the sin     47 relied upon. Which leaves us with just the single fixed portion.
 48                                                    48 
 49 1.2 The solution                                   49 1.2 The solution
 50 ----------------                                   50 ----------------
 51                                                    51 
 52 CPU time is divided by means of specifying how     52 CPU time is divided by means of specifying how much time can be spent running
 53 in a given period. We allocate this "run time"     53 in a given period. We allocate this "run time" for each real-time group which
 54 the other real-time groups will not be permitt     54 the other real-time groups will not be permitted to use.
 55                                                    55 
 56 Any time not allocated to a real-time group wi     56 Any time not allocated to a real-time group will be used to run normal priority
 57 tasks (SCHED_OTHER). Any allocated run time no     57 tasks (SCHED_OTHER). Any allocated run time not used will also be picked up by
 58 SCHED_OTHER.                                       58 SCHED_OTHER.
 59                                                    59 
 60 Let's consider an example: a frame fixed real-     60 Let's consider an example: a frame fixed real-time renderer must deliver 25
 61 frames a second, which yields a period of 0.04     61 frames a second, which yields a period of 0.04s per frame. Now say it will also
 62 have to play some music and respond to input,      62 have to play some music and respond to input, leaving it with around 80% CPU
 63 time dedicated for the graphics. We can then g     63 time dedicated for the graphics. We can then give this group a run time of 0.8
 64 * 0.04s = 0.032s.                                  64 * 0.04s = 0.032s.
 65                                                    65 
 66 This way the graphics group will have a 0.04s      66 This way the graphics group will have a 0.04s period with a 0.032s run time
 67 limit. Now if the audio thread needs to refill     67 limit. Now if the audio thread needs to refill the DMA buffer every 0.005s, but
 68 needs only about 3% CPU time to do so, it can      68 needs only about 3% CPU time to do so, it can do with a 0.03 * 0.005s =
 69 0.00015s. So this group can be scheduled with      69 0.00015s. So this group can be scheduled with a period of 0.005s and a run time
 70 of 0.00015s.                                       70 of 0.00015s.
 71                                                    71 
 72 The remaining CPU time will be used for user i     72 The remaining CPU time will be used for user input and other tasks. Because
 73 real-time tasks have explicitly allocated the      73 real-time tasks have explicitly allocated the CPU time they need to perform
 74 their tasks, buffer underruns in the graphics      74 their tasks, buffer underruns in the graphics or audio can be eliminated.
 75                                                    75 
 76 NOTE: the above example is not fully implement     76 NOTE: the above example is not fully implemented yet. We still
 77 lack an EDF scheduler to make non-uniform peri     77 lack an EDF scheduler to make non-uniform periods usable.
 78                                                    78 
 79                                                    79 
 80 2. The Interface                                   80 2. The Interface
 81 ================                                   81 ================
 82                                                    82 
 83                                                    83 
 84 2.1 System wide settings                           84 2.1 System wide settings
 85 ------------------------                           85 ------------------------
 86                                                    86 
 87 The system wide settings are configured under      87 The system wide settings are configured under the /proc virtual file system:
 88                                                    88 
 89 /proc/sys/kernel/sched_rt_period_us:               89 /proc/sys/kernel/sched_rt_period_us:
 90   The scheduling period that is equivalent to      90   The scheduling period that is equivalent to 100% CPU bandwidth.
 91                                                    91 
 92 /proc/sys/kernel/sched_rt_runtime_us:              92 /proc/sys/kernel/sched_rt_runtime_us:
 93   A global limit on how much time real-time sc     93   A global limit on how much time real-time scheduling may use. This is always
 94   less or equal to the period_us, as it denote     94   less or equal to the period_us, as it denotes the time allocated from the
 95   period_us for the real-time tasks. Even with     95   period_us for the real-time tasks. Even without CONFIG_RT_GROUP_SCHED enabled,
 96   this will limit time reserved to real-time p     96   this will limit time reserved to real-time processes. With
 97   CONFIG_RT_GROUP_SCHED=y it signifies the tot     97   CONFIG_RT_GROUP_SCHED=y it signifies the total bandwidth available to all
 98   real-time groups.                                98   real-time groups.
 99                                                    99 
100   * Time is specified in us because the interf    100   * Time is specified in us because the interface is s32. This gives an
101     operating range from 1us to about 35 minut    101     operating range from 1us to about 35 minutes.
102   * sched_rt_period_us takes values from 1 to     102   * sched_rt_period_us takes values from 1 to INT_MAX.
103   * sched_rt_runtime_us takes values from -1 t    103   * sched_rt_runtime_us takes values from -1 to sched_rt_period_us.
104   * A run time of -1 specifies runtime == peri    104   * A run time of -1 specifies runtime == period, ie. no limit.
105                                                   105 
106                                                   106 
107 2.2 Default behaviour                             107 2.2 Default behaviour
108 ---------------------                             108 ---------------------
109                                                   109 
110 The default values for sched_rt_period_us (100    110 The default values for sched_rt_period_us (1000000 or 1s) and
111 sched_rt_runtime_us (950000 or 0.95s).  This g    111 sched_rt_runtime_us (950000 or 0.95s).  This gives 0.05s to be used by
112 SCHED_OTHER (non-RT tasks). These defaults wer    112 SCHED_OTHER (non-RT tasks). These defaults were chosen so that a run-away
113 real-time tasks will not lock up the machine b    113 real-time tasks will not lock up the machine but leave a little time to recover
114 it.  By setting runtime to -1 you'd get the ol    114 it.  By setting runtime to -1 you'd get the old behaviour back.
115                                                   115 
116 By default all bandwidth is assigned to the ro    116 By default all bandwidth is assigned to the root group and new groups get the
117 period from /proc/sys/kernel/sched_rt_period_u    117 period from /proc/sys/kernel/sched_rt_period_us and a run time of 0. If you
118 want to assign bandwidth to another group, red    118 want to assign bandwidth to another group, reduce the root group's bandwidth
119 and assign some or all of the difference to an    119 and assign some or all of the difference to another group.
120                                                   120 
121 Real-time group scheduling means you have to a    121 Real-time group scheduling means you have to assign a portion of total CPU
122 bandwidth to the group before it will accept r    122 bandwidth to the group before it will accept real-time tasks. Therefore you will
123 not be able to run real-time tasks as any user    123 not be able to run real-time tasks as any user other than root until you have
124 done that, even if the user has the rights to     124 done that, even if the user has the rights to run processes with real-time
125 priority!                                         125 priority!
126                                                   126 
127                                                   127 
128 2.3 Basis for grouping tasks                      128 2.3 Basis for grouping tasks
129 ----------------------------                      129 ----------------------------
130                                                   130 
131 Enabling CONFIG_RT_GROUP_SCHED lets you explic    131 Enabling CONFIG_RT_GROUP_SCHED lets you explicitly allocate real
132 CPU bandwidth to task groups.                     132 CPU bandwidth to task groups.
133                                                   133 
134 This uses the cgroup virtual file system and "    134 This uses the cgroup virtual file system and "<cgroup>/cpu.rt_runtime_us"
135 to control the CPU time reserved for each cont    135 to control the CPU time reserved for each control group.
136                                                   136 
137 For more information on working with control g    137 For more information on working with control groups, you should read
138 Documentation/admin-guide/cgroup-v1/cgroups.rs    138 Documentation/admin-guide/cgroup-v1/cgroups.rst as well.
139                                                   139 
140 Group settings are checked against the followi    140 Group settings are checked against the following limits in order to keep the
141 configuration schedulable:                        141 configuration schedulable:
142                                                   142 
143    \Sum_{i} runtime_{i} / global_period <= glo    143    \Sum_{i} runtime_{i} / global_period <= global_runtime / global_period
144                                                   144 
145 For now, this can be simplified to just the fo    145 For now, this can be simplified to just the following (but see Future plans):
146                                                   146 
147    \Sum_{i} runtime_{i} <= global_runtime         147    \Sum_{i} runtime_{i} <= global_runtime
148                                                   148 
149                                                   149 
150 3. Future plans                                   150 3. Future plans
151 ===============                                   151 ===============
152                                                   152 
153 There is work in progress to make the scheduli    153 There is work in progress to make the scheduling period for each group
154 ("<cgroup>/cpu.rt_period_us") configurable as     154 ("<cgroup>/cpu.rt_period_us") configurable as well.
155                                                   155 
156 The constraint on the period is that a subgrou    156 The constraint on the period is that a subgroup must have a smaller or
157 equal period to its parent. But realistically     157 equal period to its parent. But realistically its not very useful _yet_
158 as its prone to starvation without deadline sc    158 as its prone to starvation without deadline scheduling.
159                                                   159 
160 Consider two sibling groups A and B; both have    160 Consider two sibling groups A and B; both have 50% bandwidth, but A's
161 period is twice the length of B's.                161 period is twice the length of B's.
162                                                   162 
163 * group A: period=100000us, runtime=50000us       163 * group A: period=100000us, runtime=50000us
164                                                   164 
165         - this runs for 0.05s once every 0.1s     165         - this runs for 0.05s once every 0.1s
166                                                   166 
167 * group B: period= 50000us, runtime=25000us       167 * group B: period= 50000us, runtime=25000us
168                                                   168 
169         - this runs for 0.025s twice every 0.1    169         - this runs for 0.025s twice every 0.1s (or once every 0.05 sec).
170                                                   170 
171 This means that currently a while (1) loop in     171 This means that currently a while (1) loop in A will run for the full period of
172 B and can starve B's tasks (assuming they are     172 B and can starve B's tasks (assuming they are of lower priority) for a whole
173 period.                                           173 period.
174                                                   174 
175 The next project will be SCHED_EDF (Earliest D    175 The next project will be SCHED_EDF (Earliest Deadline First scheduling) to bring
176 full deadline scheduling to the linux kernel.     176 full deadline scheduling to the linux kernel. Deadline scheduling the above
177 groups and treating end of the period as a dea    177 groups and treating end of the period as a deadline will ensure that they both
178 get their allocated time.                         178 get their allocated time.
179                                                   179 
180 Implementing SCHED_EDF might take a while to c    180 Implementing SCHED_EDF might take a while to complete. Priority Inheritance is
181 the biggest challenge as the current linux PI     181 the biggest challenge as the current linux PI infrastructure is geared towards
182 the limited static priority levels 0-99. With     182 the limited static priority levels 0-99. With deadline scheduling you need to
183 do deadline inheritance (since priority is inv    183 do deadline inheritance (since priority is inversely proportional to the
184 deadline delta (deadline - now)).                 184 deadline delta (deadline - now)).
185                                                   185 
186 This means the whole PI machinery will have to    186 This means the whole PI machinery will have to be reworked - and that is one of
187 the most complex pieces of code we have.          187 the most complex pieces of code we have.
                                                      

~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

kernel.org | git.kernel.org | LWN.net | Project Home | SVN repository | Mail admin

Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.

sflogo.php