Blame view
kernel/cgroup/cpuset.c
101 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 3 4 5 6 |
/* * kernel/cpuset.c * * Processor and Memory placement constraints for sets of tasks. * * Copyright (C) 2003 BULL SA. |
029190c51 cpuset sched_load... |
7 |
* Copyright (C) 2004-2007 Silicon Graphics, Inc. |
8793d854e Task Control Grou... |
8 |
* Copyright (C) 2006 Google, Inc |
1da177e4c Linux-2.6.12-rc2 |
9 10 11 |
* * Portions derived from Patrick Mochel's sysfs code. * sysfs is Copyright (c) 2001-3 Patrick Mochel |
1da177e4c Linux-2.6.12-rc2 |
12 |
* |
825a46af5 [PATCH] cpuset me... |
13 |
* 2003-10-10 Written by Simon Derr. |
1da177e4c Linux-2.6.12-rc2 |
14 |
* 2003-10-22 Updates by Stephen Hemminger. |
825a46af5 [PATCH] cpuset me... |
15 |
* 2004 May-July Rework by Paul Jackson. |
8793d854e Task Control Grou... |
16 |
* 2006 Rework by Paul Menage to use generic cgroups |
cf417141c sched, cpuset: re... |
17 18 |
* 2008 Rework of the scheduler domains and CPU hotplug handling * by Max Krasnyansky |
1da177e4c Linux-2.6.12-rc2 |
19 20 21 22 23 |
* * This file is subject to the terms and conditions of the GNU General Public * License. See the file COPYING in the main directory of the Linux * distribution for more details. */ |
1da177e4c Linux-2.6.12-rc2 |
24 25 26 27 28 29 30 31 32 33 34 35 |
#include <linux/cpu.h> #include <linux/cpumask.h> #include <linux/cpuset.h> #include <linux/err.h> #include <linux/errno.h> #include <linux/file.h> #include <linux/fs.h> #include <linux/init.h> #include <linux/interrupt.h> #include <linux/kernel.h> #include <linux/kmod.h> #include <linux/list.h> |
68860ec10 [PATCH] cpusets: ... |
36 |
#include <linux/mempolicy.h> |
1da177e4c Linux-2.6.12-rc2 |
37 |
#include <linux/mm.h> |
f481891fd cpuset: update to... |
38 |
#include <linux/memory.h> |
9984de1a5 kernel: Map most ... |
39 |
#include <linux/export.h> |
1da177e4c Linux-2.6.12-rc2 |
40 |
#include <linux/mount.h> |
a18753747 cpuset: Use fs_co... |
41 |
#include <linux/fs_context.h> |
1da177e4c Linux-2.6.12-rc2 |
42 43 44 |
#include <linux/namei.h> #include <linux/pagemap.h> #include <linux/proc_fs.h> |
6b9c2603c [PATCH] cpuset: u... |
45 |
#include <linux/rcupdate.h> |
1da177e4c Linux-2.6.12-rc2 |
46 |
#include <linux/sched.h> |
f9a25f776 cpusets: Rebuild ... |
47 |
#include <linux/sched/deadline.h> |
6e84f3152 sched/headers: Pr... |
48 |
#include <linux/sched/mm.h> |
f719ff9bc sched/headers: Pr... |
49 |
#include <linux/sched/task.h> |
1da177e4c Linux-2.6.12-rc2 |
50 |
#include <linux/seq_file.h> |
22fb52dd7 [PATCH] SELinux: ... |
51 |
#include <linux/security.h> |
1da177e4c Linux-2.6.12-rc2 |
52 |
#include <linux/slab.h> |
1da177e4c Linux-2.6.12-rc2 |
53 54 55 56 |
#include <linux/spinlock.h> #include <linux/stat.h> #include <linux/string.h> #include <linux/time.h> |
d2b436580 cpuset: Replace a... |
57 |
#include <linux/time64.h> |
1da177e4c Linux-2.6.12-rc2 |
58 59 |
#include <linux/backing-dev.h> #include <linux/sort.h> |
da99ecf11 mm: replace TIF_M... |
60 |
#include <linux/oom.h> |
edb938217 sched/isolation: ... |
61 |
#include <linux/sched/isolation.h> |
7c0f6ba68 Replace <asm/uacc... |
62 |
#include <linux/uaccess.h> |
60063497a atomic: use <linu... |
63 |
#include <linux/atomic.h> |
3d3f26a7b [PATCH] kernel/cp... |
64 |
#include <linux/mutex.h> |
956db3ca0 hotplug cpu: move... |
65 |
#include <linux/cgroup.h> |
e44193d39 cpuset: let hotpl... |
66 |
#include <linux/wait.h> |
1da177e4c Linux-2.6.12-rc2 |
67 |
|
4d1ac6a16 ANDROID: sched/cp... |
68 |
#include <trace/hooks/sched.h> |
89affbf5d cpuset: fix a dea... |
69 |
DEFINE_STATIC_KEY_FALSE(cpusets_pre_enable_key); |
002f29062 cpuset: use stati... |
70 |
DEFINE_STATIC_KEY_FALSE(cpusets_enabled_key); |
202f72d5d [PATCH] cpuset: n... |
71 |
|
3e0d98b9f [PATCH] cpuset: m... |
72 73 74 75 76 |
/* See "Frequency meter" comments, below. */ struct fmeter { int cnt; /* unprocessed events count */ int val; /* most recent output value */ |
d2b436580 cpuset: Replace a... |
77 |
time64_t time; /* clock (secs) when val computed */ |
3e0d98b9f [PATCH] cpuset: m... |
78 79 |
spinlock_t lock; /* guards read or write of above */ }; |
1da177e4c Linux-2.6.12-rc2 |
80 |
struct cpuset { |
8793d854e Task Control Grou... |
81 |
struct cgroup_subsys_state css; |
1da177e4c Linux-2.6.12-rc2 |
82 |
unsigned long flags; /* "unsigned long" so bitops work */ |
e2b9a3d7d cpuset: add cs->e... |
83 |
|
7e88291be cpuset: make cs->... |
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 |
/* * On default hierarchy: * * The user-configured masks can only be changed by writing to * cpuset.cpus and cpuset.mems, and won't be limited by the * parent masks. * * The effective masks is the real masks that apply to the tasks * in the cpuset. They may be changed if the configured masks are * changed or hotplug happens. * * effective_mask == configured_mask & parent's effective_mask, * and if it ends up empty, it will inherit the parent's mask. * * * On legacy hierachy: * * The user-configured masks are always the same with effective masks. */ |
e2b9a3d7d cpuset: add cs->e... |
103 104 |
/* user-configured CPUs and Memory Nodes allow to tasks */ cpumask_var_t cpus_allowed; |
c8dc4422c ANDROID: cpuset: ... |
105 |
cpumask_var_t cpus_requested; |
e2b9a3d7d cpuset: add cs->e... |
106 107 108 109 110 |
nodemask_t mems_allowed; /* effective CPUs and Memory Nodes allow to tasks */ cpumask_var_t effective_cpus; nodemask_t effective_mems; |
1da177e4c Linux-2.6.12-rc2 |
111 |
|
33ad801df cpuset: record ol... |
112 |
/* |
58b748425 cpuset: Define da... |
113 114 115 |
* CPUs allocated to child sub-partitions (default hierarchy only) * - CPUs granted by the parent = effective_cpus U subparts_cpus * - effective_cpus and subparts_cpus are mutually exclusive. |
4b842da27 cpuset: Make CPU ... |
116 117 118 |
* * effective_cpus contains only onlined CPUs, but subparts_cpus * may have offlined ones. |
58b748425 cpuset: Define da... |
119 120 121 122 |
*/ cpumask_var_t subparts_cpus; /* |
33ad801df cpuset: record ol... |
123 124 125 126 127 128 129 130 131 132 |
* This is old Memory Nodes tasks took on. * * - top_cpuset.old_mems_allowed is initialized to mems_allowed. * - A new cpuset's old_mems_allowed is initialized when some * task is moved into it. * - old_mems_allowed is used in cpuset_migrate_mm() when we change * cpuset.mems_allowed and have tasks' nodemask updated, and * then old_mems_allowed is updated to mems_allowed. */ nodemask_t old_mems_allowed; |
3e0d98b9f [PATCH] cpuset: m... |
133 |
struct fmeter fmeter; /* memory_pressure filter */ |
029190c51 cpuset sched_load... |
134 |
|
452477fa6 cpuset: pin down ... |
135 136 137 138 139 |
/* * Tasks are being attached to this cpuset. Used to prevent * zeroing cpus/mems_allowed between ->can_attach() and ->attach(). */ int attach_in_progress; |
029190c51 cpuset sched_load... |
140 141 |
/* partition number for rebuild_sched_domains() */ int pn; |
956db3ca0 hotplug cpu: move... |
142 |
|
1d3504fcf sched, cpuset: cu... |
143 144 |
/* for custom sched domain */ int relax_domain_level; |
58b748425 cpuset: Define da... |
145 146 147 148 149 150 |
/* number of CPUs in subparts_cpus */ int nr_subparts_cpus; /* partition root state */ int partition_root_state; |
4716909cc cpuset: Track cpu... |
151 152 153 154 155 156 157 158 |
/* * Default hierarchy only: * use_parent_ecpus - set if using parent's effective_cpus * child_ecpus_count - # of children with use_parent_ecpus set */ int use_parent_ecpus; int child_ecpus_count; |
58b748425 cpuset: Define da... |
159 160 161 162 163 164 |
}; /* * Partition root states: * * 0 - not a partition root |
3881b8612 cpuset: Add an er... |
165 |
* |
58b748425 cpuset: Define da... |
166 |
* 1 - partition root |
3881b8612 cpuset: Add an er... |
167 168 169 170 171 172 173 |
* * -1 - invalid partition root * None of the cpus in cpus_allowed can be put into the parent's * subparts_cpus. In this case, the cpuset is not a real partition * root anymore. However, the CPU_EXCLUSIVE bit will still be set * and the cpuset can be restored back to a partition root if the * parent cpuset can give more CPUs back to this child cpuset. |
58b748425 cpuset: Define da... |
174 175 176 |
*/ #define PRS_DISABLED 0 #define PRS_ENABLED 1 |
3881b8612 cpuset: Add an er... |
177 |
#define PRS_ERROR -1 |
58b748425 cpuset: Define da... |
178 179 180 181 182 183 184 185 |
/* * Temporary cpumasks for working with partitions that are passed among * functions to avoid memory allocation in inner functions. */ struct tmpmasks { cpumask_var_t addmask, delmask; /* For partition root */ cpumask_var_t new_cpus; /* For update_cpumasks_hier() */ |
1da177e4c Linux-2.6.12-rc2 |
186 |
}; |
a7c6d554a cgroup: add/updat... |
187 |
static inline struct cpuset *css_cs(struct cgroup_subsys_state *css) |
8793d854e Task Control Grou... |
188 |
{ |
a7c6d554a cgroup: add/updat... |
189 |
return css ? container_of(css, struct cpuset, css) : NULL; |
8793d854e Task Control Grou... |
190 191 192 193 194 |
} /* Retrieve the cpuset for a task */ static inline struct cpuset *task_cs(struct task_struct *task) { |
073219e99 cgroup: clean up ... |
195 |
return css_cs(task_css(task, cpuset_cgrp_id)); |
8793d854e Task Control Grou... |
196 |
} |
8793d854e Task Control Grou... |
197 |
|
c9710d801 cpuset: drop "con... |
198 |
static inline struct cpuset *parent_cs(struct cpuset *cs) |
c431069fe cpuset: remove cp... |
199 |
{ |
5c9d535b8 cgroup: remove cs... |
200 |
return css_cs(cs->css.parent); |
c431069fe cpuset: remove cp... |
201 |
} |
1da177e4c Linux-2.6.12-rc2 |
202 203 |
/* bits in struct cpuset flags field */ typedef enum { |
efeb77b2f cpuset: introduce... |
204 |
CS_ONLINE, |
1da177e4c Linux-2.6.12-rc2 |
205 206 |
CS_CPU_EXCLUSIVE, CS_MEM_EXCLUSIVE, |
786083667 Cpuset hardwall f... |
207 |
CS_MEM_HARDWALL, |
45b07ef31 [PATCH] cpusets: ... |
208 |
CS_MEMORY_MIGRATE, |
029190c51 cpuset sched_load... |
209 |
CS_SCHED_LOAD_BALANCE, |
825a46af5 [PATCH] cpuset me... |
210 211 |
CS_SPREAD_PAGE, CS_SPREAD_SLAB, |
1da177e4c Linux-2.6.12-rc2 |
212 213 214 |
} cpuset_flagbits_t; /* convenient tests for these bits */ |
41c25707d cpuset: consider ... |
215 |
static inline bool is_cpuset_online(struct cpuset *cs) |
efeb77b2f cpuset: introduce... |
216 |
{ |
41c25707d cpuset: consider ... |
217 |
return test_bit(CS_ONLINE, &cs->flags) && !css_is_dying(&cs->css); |
efeb77b2f cpuset: introduce... |
218 |
} |
1da177e4c Linux-2.6.12-rc2 |
219 220 |
static inline int is_cpu_exclusive(const struct cpuset *cs) { |
7b5b9ef0e [PATCH] cpuset cl... |
221 |
return test_bit(CS_CPU_EXCLUSIVE, &cs->flags); |
1da177e4c Linux-2.6.12-rc2 |
222 223 224 225 |
} static inline int is_mem_exclusive(const struct cpuset *cs) { |
7b5b9ef0e [PATCH] cpuset cl... |
226 |
return test_bit(CS_MEM_EXCLUSIVE, &cs->flags); |
1da177e4c Linux-2.6.12-rc2 |
227 |
} |
786083667 Cpuset hardwall f... |
228 229 230 231 |
static inline int is_mem_hardwall(const struct cpuset *cs) { return test_bit(CS_MEM_HARDWALL, &cs->flags); } |
029190c51 cpuset sched_load... |
232 233 234 235 |
static inline int is_sched_load_balance(const struct cpuset *cs) { return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); } |
45b07ef31 [PATCH] cpusets: ... |
236 237 |
static inline int is_memory_migrate(const struct cpuset *cs) { |
7b5b9ef0e [PATCH] cpuset cl... |
238 |
return test_bit(CS_MEMORY_MIGRATE, &cs->flags); |
45b07ef31 [PATCH] cpusets: ... |
239 |
} |
825a46af5 [PATCH] cpuset me... |
240 241 242 243 244 245 246 247 248 |
static inline int is_spread_page(const struct cpuset *cs) { return test_bit(CS_SPREAD_PAGE, &cs->flags); } static inline int is_spread_slab(const struct cpuset *cs) { return test_bit(CS_SPREAD_SLAB, &cs->flags); } |
58b748425 cpuset: Define da... |
249 250 |
static inline int is_partition_root(const struct cpuset *cs) { |
3881b8612 cpuset: Add an er... |
251 |
return cs->partition_root_state > 0; |
58b748425 cpuset: Define da... |
252 |
} |
1da177e4c Linux-2.6.12-rc2 |
253 |
static struct cpuset top_cpuset = { |
efeb77b2f cpuset: introduce... |
254 255 |
.flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)), |
58b748425 cpuset: Define da... |
256 |
.partition_root_state = PRS_ENABLED, |
1da177e4c Linux-2.6.12-rc2 |
257 |
}; |
ae8086ce1 cpuset: introduce... |
258 259 260 |
/** * cpuset_for_each_child - traverse online children of a cpuset * @child_cs: loop cursor pointing to the current child |
492eb21b9 cgroup: make hier... |
261 |
* @pos_css: used for iteration |
ae8086ce1 cpuset: introduce... |
262 263 264 265 266 |
* @parent_cs: target cpuset to walk children of * * Walk @child_cs through the online children of @parent_cs. Must be used * with RCU read locked. */ |
492eb21b9 cgroup: make hier... |
267 268 269 |
#define cpuset_for_each_child(child_cs, pos_css, parent_cs) \ css_for_each_child((pos_css), &(parent_cs)->css) \ if (is_cpuset_online(((child_cs) = css_cs((pos_css))))) |
ae8086ce1 cpuset: introduce... |
270 |
|
fc560a26a cpuset: replace c... |
271 272 273 |
/** * cpuset_for_each_descendant_pre - pre-order walk of a cpuset's descendants * @des_cs: loop cursor pointing to the current descendant |
492eb21b9 cgroup: make hier... |
274 |
* @pos_css: used for iteration |
fc560a26a cpuset: replace c... |
275 276 277 |
* @root_cs: target cpuset to walk ancestor of * * Walk @des_cs through the online descendants of @root_cs. Must be used |
492eb21b9 cgroup: make hier... |
278 |
* with RCU read locked. The caller may modify @pos_css by calling |
bd8815a6d cgroup: make css_... |
279 280 |
* css_rightmost_descendant() to skip subtree. @root_cs is included in the * iteration and the first node to be visited. |
fc560a26a cpuset: replace c... |
281 |
*/ |
492eb21b9 cgroup: make hier... |
282 283 284 |
#define cpuset_for_each_descendant_pre(des_cs, pos_css, root_cs) \ css_for_each_descendant_pre((pos_css), &(root_cs)->css) \ if (is_cpuset_online(((des_cs) = css_cs((pos_css))))) |
fc560a26a cpuset: replace c... |
285 |
|
1da177e4c Linux-2.6.12-rc2 |
286 |
/* |
8447a0fee cpuset: convert c... |
287 288 289 290 |
* There are two global locks guarding cpuset structures - cpuset_mutex and * callback_lock. We also require taking task_lock() when dereferencing a * task's cpuset pointer. See "The task_lock() exception", at the end of this * comment. |
5d21cc2db cpuset: replace c... |
291 |
* |
8447a0fee cpuset: convert c... |
292 |
* A task must hold both locks to modify cpusets. If a task holds |
5d21cc2db cpuset: replace c... |
293 |
* cpuset_mutex, then it blocks others wanting that mutex, ensuring that it |
8447a0fee cpuset: convert c... |
294 |
* is the only task able to also acquire callback_lock and be able to |
5d21cc2db cpuset: replace c... |
295 296 297 |
* modify cpusets. It can perform various checks on the cpuset structure * first, knowing nothing will change. It can also allocate memory while * just holding cpuset_mutex. While it is performing these checks, various |
8447a0fee cpuset: convert c... |
298 299 |
* callback routines can briefly acquire callback_lock to query cpusets. * Once it is ready to make the changes, it takes callback_lock, blocking |
5d21cc2db cpuset: replace c... |
300 |
* everyone else. |
053199edf [PATCH] cpusets: ... |
301 302 |
* * Calls to the kernel memory allocator can not be made while holding |
8447a0fee cpuset: convert c... |
303 |
* callback_lock, as that would risk double tripping on callback_lock |
053199edf [PATCH] cpusets: ... |
304 305 306 |
* from one of the callbacks into the cpuset code from within * __alloc_pages(). * |
8447a0fee cpuset: convert c... |
307 |
* If a task is only holding callback_lock, then it has read-only |
053199edf [PATCH] cpusets: ... |
308 309 |
* access to cpusets. * |
58568d2a8 cpuset,mm: update... |
310 311 312 |
* Now, the task_struct fields mems_allowed and mempolicy may be changed * by other task, we use alloc_lock in the task_struct fields to protect * them. |
053199edf [PATCH] cpusets: ... |
313 |
* |
8447a0fee cpuset: convert c... |
314 |
* The cpuset_common_file_read() handlers only hold callback_lock across |
053199edf [PATCH] cpusets: ... |
315 316 317 |
* small pieces of code, such as when reading out possibly multi-word * cpumasks and nodemasks. * |
2df167a30 cgroups: update c... |
318 319 |
* Accessing a task's cpuset should be done in accordance with the * guidelines for accessing subsystem state in kernel/cgroup.c |
1da177e4c Linux-2.6.12-rc2 |
320 |
*/ |
18695a120 Revert "cgroup/cp... |
321 |
static DEFINE_MUTEX(cpuset_mutex); |
8447a0fee cpuset: convert c... |
322 |
static DEFINE_SPINLOCK(callback_lock); |
4247bdc60 [PATCH] cpuset se... |
323 |
|
e93ad19d0 cpuset: make mm m... |
324 |
static struct workqueue_struct *cpuset_migrate_mm_wq; |
cf417141c sched, cpuset: re... |
325 |
/* |
1d3a64fbd ANDROID: cpu/hotp... |
326 327 |
* CPU / memory hotplug is handled asynchronously * for hotplug, synchronously for resume_cpus |
3a5a6d0c2 cpuset: don't nes... |
328 |
*/ |
3a5a6d0c2 cpuset: don't nes... |
329 |
static DECLARE_WORK(cpuset_hotplug_work, cpuset_hotplug_workfn); |
e44193d39 cpuset: let hotpl... |
330 |
static DECLARE_WAIT_QUEUE_HEAD(cpuset_attach_wq); |
3a5a6d0c2 cpuset: don't nes... |
331 |
/* |
0c05b9bdb docs: cgroup-v1: ... |
332 333 334 335 336 337 |
* Cgroup v2 behavior is used on the "cpus" and "mems" control files when * on default hierarchy or when the cpuset_v2_mode flag is set by mounting * the v1 cpuset cgroup filesystem with the "cpuset_v2_mode" mount option. * With v2 behavior, "cpus" and "mems" are always what the users have * requested and won't be changed by hotplug events. Only the effective * cpus or mems will be affected. |
b8d1b8ee9 cpuset: Allow v2 ... |
338 339 340 341 342 343 344 345 |
*/ static inline bool is_in_v2_mode(void) { return cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || (cpuset_cgrp_subsys.root->flags & CGRP_ROOT_CPUSET_V2_MODE); } /* |
300ed6cbb cpuset: convert c... |
346 |
* Return in pmask the portion of a cpusets's cpus_allowed that |
1da177e4c Linux-2.6.12-rc2 |
347 |
* are online. If none are online, walk up the cpuset hierarchy |
28b89b9e6 cpuset: handle ra... |
348 |
* until we find one that does have some online cpus. |
1da177e4c Linux-2.6.12-rc2 |
349 350 |
* * One way or another, we guarantee to return some non-empty subset |
bca99ddbf ANDROID: cgroup/c... |
351 |
* of cpu_active_mask. |
1da177e4c Linux-2.6.12-rc2 |
352 |
* |
8447a0fee cpuset: convert c... |
353 |
* Call with callback_lock or cpuset_mutex held. |
1da177e4c Linux-2.6.12-rc2 |
354 |
*/ |
c9710d801 cpuset: drop "con... |
355 |
static void guarantee_online_cpus(struct cpuset *cs, struct cpumask *pmask) |
1da177e4c Linux-2.6.12-rc2 |
356 |
{ |
bca99ddbf ANDROID: cgroup/c... |
357 |
while (!cpumask_intersects(cs->effective_cpus, cpu_active_mask)) { |
c431069fe cpuset: remove cp... |
358 |
cs = parent_cs(cs); |
28b89b9e6 cpuset: handle ra... |
359 360 361 362 363 |
if (unlikely(!cs)) { /* * The top cpuset doesn't have any online cpu as a * consequence of a race between cpuset_hotplug_work * and cpu hotplug notifier. But we know the top |
7b7b8a2c9 kernel/: fix repe... |
364 |
* cpuset's effective_cpus is on its way to be |
28b89b9e6 cpuset: handle ra... |
365 366 |
* identical to cpu_online_mask. */ |
bca99ddbf ANDROID: cgroup/c... |
367 |
cpumask_copy(pmask, cpu_active_mask); |
28b89b9e6 cpuset: handle ra... |
368 369 370 |
return; } } |
bca99ddbf ANDROID: cgroup/c... |
371 |
cpumask_and(pmask, cs->effective_cpus, cpu_active_mask); |
1da177e4c Linux-2.6.12-rc2 |
372 373 374 375 |
} /* * Return in *pmask the portion of a cpusets's mems_allowed that |
0e1e7c7a7 Memoryless nodes:... |
376 377 |
* are online, with memory. If none are online with memory, walk * up the cpuset hierarchy until we find one that does have some |
40df2deb5 cpuset: cleanup g... |
378 |
* online mems. The top cpuset always has some mems online. |
1da177e4c Linux-2.6.12-rc2 |
379 380 |
* * One way or another, we guarantee to return some non-empty subset |
38d7bee9d cpuset: use N_MEM... |
381 |
* of node_states[N_MEMORY]. |
1da177e4c Linux-2.6.12-rc2 |
382 |
* |
8447a0fee cpuset: convert c... |
383 |
* Call with callback_lock or cpuset_mutex held. |
1da177e4c Linux-2.6.12-rc2 |
384 |
*/ |
c9710d801 cpuset: drop "con... |
385 |
static void guarantee_online_mems(struct cpuset *cs, nodemask_t *pmask) |
1da177e4c Linux-2.6.12-rc2 |
386 |
{ |
ae1c80238 cpuset: apply cs-... |
387 |
while (!nodes_intersects(cs->effective_mems, node_states[N_MEMORY])) |
c431069fe cpuset: remove cp... |
388 |
cs = parent_cs(cs); |
ae1c80238 cpuset: apply cs-... |
389 |
nodes_and(*pmask, cs->effective_mems, node_states[N_MEMORY]); |
1da177e4c Linux-2.6.12-rc2 |
390 |
} |
f3b39d47e cpusets: restruct... |
391 392 393 |
/* * update task's spread flag if cpuset's page/slab spread flag is set * |
8447a0fee cpuset: convert c... |
394 |
* Call with callback_lock or cpuset_mutex held. |
f3b39d47e cpusets: restruct... |
395 396 397 398 399 |
*/ static void cpuset_update_task_spread_flag(struct cpuset *cs, struct task_struct *tsk) { if (is_spread_page(cs)) |
2ad654bc5 cpuset: PF_SPREAD... |
400 |
task_set_spread_page(tsk); |
f3b39d47e cpusets: restruct... |
401 |
else |
2ad654bc5 cpuset: PF_SPREAD... |
402 |
task_clear_spread_page(tsk); |
f3b39d47e cpusets: restruct... |
403 |
if (is_spread_slab(cs)) |
2ad654bc5 cpuset: PF_SPREAD... |
404 |
task_set_spread_slab(tsk); |
f3b39d47e cpusets: restruct... |
405 |
else |
2ad654bc5 cpuset: PF_SPREAD... |
406 |
task_clear_spread_slab(tsk); |
f3b39d47e cpusets: restruct... |
407 |
} |
1da177e4c Linux-2.6.12-rc2 |
408 409 410 411 412 |
/* * is_cpuset_subset(p, q) - Is cpuset p a subset of cpuset q? * * One cpuset is a subset of another if all its allowed CPUs and * Memory Nodes are a subset of the other, and its exclusive flags |
5d21cc2db cpuset: replace c... |
413 |
* are only set if the other's are set. Call holding cpuset_mutex. |
1da177e4c Linux-2.6.12-rc2 |
414 415 416 417 |
*/ static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) { |
c8dc4422c ANDROID: cpuset: ... |
418 |
return cpumask_subset(p->cpus_requested, q->cpus_requested) && |
1da177e4c Linux-2.6.12-rc2 |
419 420 421 422 |
nodes_subset(p->mems_allowed, q->mems_allowed) && is_cpu_exclusive(p) <= is_cpu_exclusive(q) && is_mem_exclusive(p) <= is_mem_exclusive(q); } |
645fcc9d2 cpuset: don't all... |
423 |
/** |
bf92370c0 cpuset: Simply al... |
424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 |
* alloc_cpumasks - allocate three cpumasks for cpuset * @cs: the cpuset that have cpumasks to be allocated. * @tmp: the tmpmasks structure pointer * Return: 0 if successful, -ENOMEM otherwise. * * Only one of the two input arguments should be non-NULL. */ static inline int alloc_cpumasks(struct cpuset *cs, struct tmpmasks *tmp) { cpumask_var_t *pmask1, *pmask2, *pmask3; if (cs) { pmask1 = &cs->cpus_allowed; pmask2 = &cs->effective_cpus; pmask3 = &cs->subparts_cpus; } else { pmask1 = &tmp->new_cpus; pmask2 = &tmp->addmask; pmask3 = &tmp->delmask; } if (!zalloc_cpumask_var(pmask1, GFP_KERNEL)) return -ENOMEM; if (!zalloc_cpumask_var(pmask2, GFP_KERNEL)) goto free_one; if (!zalloc_cpumask_var(pmask3, GFP_KERNEL)) goto free_two; |
c8dc4422c ANDROID: cpuset: ... |
453 454 |
if (cs && !zalloc_cpumask_var(&cs->cpus_requested, GFP_KERNEL)) goto free_three; |
bf92370c0 cpuset: Simply al... |
455 |
return 0; |
c8dc4422c ANDROID: cpuset: ... |
456 457 |
free_three: free_cpumask_var(*pmask3); |
bf92370c0 cpuset: Simply al... |
458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 |
free_two: free_cpumask_var(*pmask2); free_one: free_cpumask_var(*pmask1); return -ENOMEM; } /** * free_cpumasks - free cpumasks in a tmpmasks structure * @cs: the cpuset that have cpumasks to be free. * @tmp: the tmpmasks structure pointer */ static inline void free_cpumasks(struct cpuset *cs, struct tmpmasks *tmp) { if (cs) { free_cpumask_var(cs->cpus_allowed); |
c8dc4422c ANDROID: cpuset: ... |
474 |
free_cpumask_var(cs->cpus_requested); |
bf92370c0 cpuset: Simply al... |
475 476 477 478 479 480 481 482 483 484 485 |
free_cpumask_var(cs->effective_cpus); free_cpumask_var(cs->subparts_cpus); } if (tmp) { free_cpumask_var(tmp->new_cpus); free_cpumask_var(tmp->addmask); free_cpumask_var(tmp->delmask); } } /** |
645fcc9d2 cpuset: don't all... |
486 487 488 |
* alloc_trial_cpuset - allocate a trial cpuset * @cs: the cpuset that the trial cpuset duplicates */ |
c9710d801 cpuset: drop "con... |
489 |
static struct cpuset *alloc_trial_cpuset(struct cpuset *cs) |
645fcc9d2 cpuset: don't all... |
490 |
{ |
300ed6cbb cpuset: convert c... |
491 492 493 494 495 |
struct cpuset *trial; trial = kmemdup(cs, sizeof(*cs), GFP_KERNEL); if (!trial) return NULL; |
bf92370c0 cpuset: Simply al... |
496 497 498 499 |
if (alloc_cpumasks(trial, NULL)) { kfree(trial); return NULL; } |
300ed6cbb cpuset: convert c... |
500 |
|
e2b9a3d7d cpuset: add cs->e... |
501 |
cpumask_copy(trial->cpus_allowed, cs->cpus_allowed); |
c8dc4422c ANDROID: cpuset: ... |
502 |
cpumask_copy(trial->cpus_requested, cs->cpus_requested); |
e2b9a3d7d cpuset: add cs->e... |
503 |
cpumask_copy(trial->effective_cpus, cs->effective_cpus); |
300ed6cbb cpuset: convert c... |
504 |
return trial; |
645fcc9d2 cpuset: don't all... |
505 506 507 |
} /** |
bf92370c0 cpuset: Simply al... |
508 509 |
* free_cpuset - free the cpuset * @cs: the cpuset to be freed |
645fcc9d2 cpuset: don't all... |
510 |
*/ |
bf92370c0 cpuset: Simply al... |
511 |
static inline void free_cpuset(struct cpuset *cs) |
645fcc9d2 cpuset: don't all... |
512 |
{ |
bf92370c0 cpuset: Simply al... |
513 514 |
free_cpumasks(cs, NULL); kfree(cs); |
645fcc9d2 cpuset: don't all... |
515 |
} |
1da177e4c Linux-2.6.12-rc2 |
516 517 518 519 520 521 522 |
/* * validate_change() - Used to validate that any proposed cpuset change * follows the structural rules for cpusets. * * If we replaced the flag and mask values of the current cpuset * (cur) with those values in the trial cpuset (trial), would * our various subset and exclusive rules still be valid? Presumes |
5d21cc2db cpuset: replace c... |
523 |
* cpuset_mutex held. |
1da177e4c Linux-2.6.12-rc2 |
524 525 526 527 528 529 530 531 532 533 534 |
* * 'cur' is the address of an actual, in-use cpuset. Operations * such as list traversal that depend on the actual address of the * cpuset in the list must use cur below, not trial. * * 'trial' is the address of bulk structure copy of cur, with * perhaps one or more of the fields cpus_allowed, mems_allowed, * or flags changed to new, trial values. * * Return 0 if valid, -errno if not. */ |
c9710d801 cpuset: drop "con... |
535 |
static int validate_change(struct cpuset *cur, struct cpuset *trial) |
1da177e4c Linux-2.6.12-rc2 |
536 |
{ |
492eb21b9 cgroup: make hier... |
537 |
struct cgroup_subsys_state *css; |
1da177e4c Linux-2.6.12-rc2 |
538 |
struct cpuset *c, *par; |
ae8086ce1 cpuset: introduce... |
539 540 541 |
int ret; rcu_read_lock(); |
1da177e4c Linux-2.6.12-rc2 |
542 543 |
/* Each of our child cpusets must be a subset of us */ |
ae8086ce1 cpuset: introduce... |
544 |
ret = -EBUSY; |
492eb21b9 cgroup: make hier... |
545 |
cpuset_for_each_child(c, css, cur) |
ae8086ce1 cpuset: introduce... |
546 547 |
if (!is_cpuset_subset(c, trial)) goto out; |
1da177e4c Linux-2.6.12-rc2 |
548 549 |
/* Remaining checks don't apply to root cpuset */ |
ae8086ce1 cpuset: introduce... |
550 |
ret = 0; |
696040670 [PATCH] cpuset: m... |
551 |
if (cur == &top_cpuset) |
ae8086ce1 cpuset: introduce... |
552 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
553 |
|
c431069fe cpuset: remove cp... |
554 |
par = parent_cs(cur); |
696040670 [PATCH] cpuset: m... |
555 |
|
7e88291be cpuset: make cs->... |
556 |
/* On legacy hiearchy, we must be a subset of our parent cpuset. */ |
ae8086ce1 cpuset: introduce... |
557 |
ret = -EACCES; |
b8d1b8ee9 cpuset: Allow v2 ... |
558 |
if (!is_in_v2_mode() && !is_cpuset_subset(trial, par)) |
ae8086ce1 cpuset: introduce... |
559 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
560 |
|
2df167a30 cgroups: update c... |
561 562 563 564 |
/* * If either I or some sibling (!= me) is exclusive, we can't * overlap */ |
ae8086ce1 cpuset: introduce... |
565 |
ret = -EINVAL; |
492eb21b9 cgroup: make hier... |
566 |
cpuset_for_each_child(c, css, par) { |
1da177e4c Linux-2.6.12-rc2 |
567 568 |
if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && c != cur && |
c8dc4422c ANDROID: cpuset: ... |
569 |
cpumask_intersects(trial->cpus_requested, c->cpus_requested)) |
ae8086ce1 cpuset: introduce... |
570 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
571 572 573 |
if ((is_mem_exclusive(trial) || is_mem_exclusive(c)) && c != cur && nodes_intersects(trial->mems_allowed, c->mems_allowed)) |
ae8086ce1 cpuset: introduce... |
574 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
575 |
} |
452477fa6 cpuset: pin down ... |
576 577 |
/* * Cpusets with tasks - existing or newly being attached - can't |
1c09b195d cpuset: fix a reg... |
578 |
* be changed to have empty cpus_allowed or mems_allowed. |
452477fa6 cpuset: pin down ... |
579 |
*/ |
ae8086ce1 cpuset: introduce... |
580 |
ret = -ENOSPC; |
27bd4dbb8 cgroup: replace c... |
581 |
if ((cgroup_is_populated(cur->css.cgroup) || cur->attach_in_progress)) { |
1c09b195d cpuset: fix a reg... |
582 583 584 585 586 587 588 |
if (!cpumask_empty(cur->cpus_allowed) && cpumask_empty(trial->cpus_allowed)) goto out; if (!nodes_empty(cur->mems_allowed) && nodes_empty(trial->mems_allowed)) goto out; } |
020958b62 cpusets: decrusti... |
589 |
|
f82f80426 sched/deadline: E... |
590 591 592 593 594 595 596 597 598 |
/* * We can't shrink if we won't have enough room for SCHED_DEADLINE * tasks. */ ret = -EBUSY; if (is_cpu_exclusive(cur) && !cpuset_cpumask_can_shrink(cur->cpus_allowed, trial->cpus_allowed)) goto out; |
ae8086ce1 cpuset: introduce... |
599 600 601 602 |
ret = 0; out: rcu_read_unlock(); return ret; |
1da177e4c Linux-2.6.12-rc2 |
603 |
} |
db7f47cf4 cpusets: allow cp... |
604 |
#ifdef CONFIG_SMP |
85d7b9498 [PATCH] Dynamic s... |
605 |
/* |
cf417141c sched, cpuset: re... |
606 |
* Helper routine for generate_sched_domains(). |
8b5f1c52d cpuset: use effec... |
607 |
* Do cpusets a, b have overlapping effective cpus_allowed masks? |
029190c51 cpuset sched_load... |
608 |
*/ |
029190c51 cpuset sched_load... |
609 610 |
static int cpusets_overlap(struct cpuset *a, struct cpuset *b) { |
8b5f1c52d cpuset: use effec... |
611 |
return cpumask_intersects(a->effective_cpus, b->effective_cpus); |
029190c51 cpuset sched_load... |
612 |
} |
1d3504fcf sched, cpuset: cu... |
613 614 615 |
static void update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c) { |
1d3504fcf sched, cpuset: cu... |
616 617 618 619 |
if (dattr->relax_domain_level < c->relax_domain_level) dattr->relax_domain_level = c->relax_domain_level; return; } |
fc560a26a cpuset: replace c... |
620 621 |
static void update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *root_cs) |
f5393693e cpuset: speed up ... |
622 |
{ |
fc560a26a cpuset: replace c... |
623 |
struct cpuset *cp; |
492eb21b9 cgroup: make hier... |
624 |
struct cgroup_subsys_state *pos_css; |
f5393693e cpuset: speed up ... |
625 |
|
fc560a26a cpuset: replace c... |
626 |
rcu_read_lock(); |
492eb21b9 cgroup: make hier... |
627 |
cpuset_for_each_descendant_pre(cp, pos_css, root_cs) { |
fc560a26a cpuset: replace c... |
628 629 |
/* skip the whole subtree if @cp doesn't have any CPU */ if (cpumask_empty(cp->cpus_allowed)) { |
492eb21b9 cgroup: make hier... |
630 |
pos_css = css_rightmost_descendant(pos_css); |
f5393693e cpuset: speed up ... |
631 |
continue; |
fc560a26a cpuset: replace c... |
632 |
} |
f5393693e cpuset: speed up ... |
633 634 635 |
if (is_sched_load_balance(cp)) update_domain_attr(dattr, cp); |
f5393693e cpuset: speed up ... |
636 |
} |
fc560a26a cpuset: replace c... |
637 |
rcu_read_unlock(); |
f5393693e cpuset: speed up ... |
638 |
} |
be040bea9 cpuset: Make nr_c... |
639 640 641 642 643 644 |
/* Must be called with cpuset_mutex held. */ static inline int nr_cpusets(void) { /* jump label reference count + the top-level cpuset */ return static_key_count(&cpusets_enabled_key.key) + 1; } |
029190c51 cpuset sched_load... |
645 |
/* |
cf417141c sched, cpuset: re... |
646 647 648 649 650 |
* generate_sched_domains() * * This function builds a partial partition of the systems CPUs * A 'partial partition' is a set of non-overlapping subsets whose * union is a subset of that set. |
0a0fca9d8 sched: Rename sch... |
651 |
* The output of this function needs to be passed to kernel/sched/core.c |
cf417141c sched, cpuset: re... |
652 653 654 |
* partition_sched_domains() routine, which will rebuild the scheduler's * load balancing domains (sched domains) as specified by that partial * partition. |
029190c51 cpuset sched_load... |
655 |
* |
da82c92f1 docs: cgroup-v1: ... |
656 |
* See "What is sched_load_balance" in Documentation/admin-guide/cgroup-v1/cpusets.rst |
029190c51 cpuset sched_load... |
657 658 659 660 661 662 663 |
* for a background explanation of this. * * Does not return errors, on the theory that the callers of this * routine would rather not worry about failures to rebuild sched * domains when operating in the severe memory shortage situations * that could cause allocation failures below. * |
5d21cc2db cpuset: replace c... |
664 |
* Must be called with cpuset_mutex held. |
029190c51 cpuset sched_load... |
665 666 |
* * The three key local variables below are: |
b6fbbf31d cgroup/cpuset: Up... |
667 668 669 670 |
* cp - cpuset pointer, used (together with pos_css) to perform a * top-down scan of all cpusets. For our purposes, rebuilding * the schedulers sched domains, we can ignore !is_sched_load_ * balance cpusets. |
029190c51 cpuset sched_load... |
671 672 673 674 675 676 677 678 |
* csa - (for CpuSet Array) Array of pointers to all the cpusets * that need to be load balanced, for convenient iterative * access by the subsequent code that finds the best partition, * i.e the set of domains (subsets) of CPUs such that the * cpus_allowed of every cpuset marked is_sched_load_balance * is a subset of one of these domains, while there are as * many such domains as possible, each as small as possible. * doms - Conversion of 'csa' to an array of cpumasks, for passing to |
0a0fca9d8 sched: Rename sch... |
679 |
* the kernel/sched/core.c routine partition_sched_domains() in a |
029190c51 cpuset sched_load... |
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 |
* convenient format, that can be easily compared to the prior * value to determine what partition elements (sched domains) * were changed (added or removed.) * * Finding the best partition (set of domains): * The triple nested loops below over i, j, k scan over the * load balanced cpusets (using the array of cpuset pointers in * csa[]) looking for pairs of cpusets that have overlapping * cpus_allowed, but which don't have the same 'pn' partition * number and gives them in the same partition number. It keeps * looping on the 'restart' label until it can no longer find * any such pairs. * * The union of the cpus_allowed masks from the set of * all cpusets having the same 'pn' value then form the one * element of the partition (one sched domain) to be passed to * partition_sched_domains(). */ |
acc3f5d7c cpumask: Partitio... |
698 |
static int generate_sched_domains(cpumask_var_t **domains, |
cf417141c sched, cpuset: re... |
699 |
struct sched_domain_attr **attributes) |
029190c51 cpuset sched_load... |
700 |
{ |
b6fbbf31d cgroup/cpuset: Up... |
701 |
struct cpuset *cp; /* top-down scan of cpusets */ |
029190c51 cpuset sched_load... |
702 703 704 |
struct cpuset **csa; /* array of all cpuset ptrs */ int csn; /* how many cpuset ptrs in csa so far */ int i, j, k; /* indices for partition finding loops */ |
acc3f5d7c cpumask: Partitio... |
705 |
cpumask_var_t *doms; /* resulting partition; i.e. sched domains */ |
1d3504fcf sched, cpuset: cu... |
706 |
struct sched_domain_attr *dattr; /* attributes for custom domains */ |
1583715dd sched, cpusets: f... |
707 |
int ndoms = 0; /* number of sched domains in result */ |
6af866af3 cpuset: remove re... |
708 |
int nslot; /* next empty doms[] struct cpumask slot */ |
492eb21b9 cgroup: make hier... |
709 |
struct cgroup_subsys_state *pos_css; |
0ccea8feb cpuset: Make gene... |
710 |
bool root_load_balance = is_sched_load_balance(&top_cpuset); |
029190c51 cpuset sched_load... |
711 |
|
029190c51 cpuset sched_load... |
712 |
doms = NULL; |
1d3504fcf sched, cpuset: cu... |
713 |
dattr = NULL; |
cf417141c sched, cpuset: re... |
714 |
csa = NULL; |
029190c51 cpuset sched_load... |
715 716 |
/* Special case for the 99% of systems with one, full, sched domain */ |
0ccea8feb cpuset: Make gene... |
717 |
if (root_load_balance && !top_cpuset.nr_subparts_cpus) { |
acc3f5d7c cpumask: Partitio... |
718 719 |
ndoms = 1; doms = alloc_sched_domains(ndoms); |
029190c51 cpuset sched_load... |
720 |
if (!doms) |
cf417141c sched, cpuset: re... |
721 |
goto done; |
1d3504fcf sched, cpuset: cu... |
722 723 724 |
dattr = kmalloc(sizeof(struct sched_domain_attr), GFP_KERNEL); if (dattr) { *dattr = SD_ATTR_INIT; |
93a655755 cpuset: fix wrong... |
725 |
update_domain_attr_tree(dattr, &top_cpuset); |
1d3504fcf sched, cpuset: cu... |
726 |
} |
47b8ea718 cpusets, isolcpus... |
727 |
cpumask_and(doms[0], top_cpuset.effective_cpus, |
edb938217 sched/isolation: ... |
728 |
housekeeping_cpumask(HK_FLAG_DOMAIN)); |
cf417141c sched, cpuset: re... |
729 |
|
cf417141c sched, cpuset: re... |
730 |
goto done; |
029190c51 cpuset sched_load... |
731 |
} |
6da2ec560 treewide: kmalloc... |
732 |
csa = kmalloc_array(nr_cpusets(), sizeof(cp), GFP_KERNEL); |
029190c51 cpuset sched_load... |
733 734 735 |
if (!csa) goto done; csn = 0; |
fc560a26a cpuset: replace c... |
736 |
rcu_read_lock(); |
0ccea8feb cpuset: Make gene... |
737 738 |
if (root_load_balance) csa[csn++] = &top_cpuset; |
492eb21b9 cgroup: make hier... |
739 |
cpuset_for_each_descendant_pre(cp, pos_css, &top_cpuset) { |
bd8815a6d cgroup: make css_... |
740 741 |
if (cp == &top_cpuset) continue; |
f5393693e cpuset: speed up ... |
742 |
/* |
fc560a26a cpuset: replace c... |
743 744 745 746 747 748 |
* Continue traversing beyond @cp iff @cp has some CPUs and * isn't load balancing. The former is obvious. The * latter: All child cpusets contain a subset of the * parent's cpus, so just skip them, and then we call * update_domain_attr_tree() to calc relax_domain_level of * the corresponding sched domain. |
0ccea8feb cpuset: Make gene... |
749 750 751 |
* * If root is load-balancing, we can skip @cp if it * is a subset of the root's effective_cpus. |
f5393693e cpuset: speed up ... |
752 |
*/ |
fc560a26a cpuset: replace c... |
753 |
if (!cpumask_empty(cp->cpus_allowed) && |
47b8ea718 cpusets, isolcpus... |
754 |
!(is_sched_load_balance(cp) && |
edb938217 sched/isolation: ... |
755 756 |
cpumask_intersects(cp->cpus_allowed, housekeeping_cpumask(HK_FLAG_DOMAIN)))) |
f5393693e cpuset: speed up ... |
757 |
continue; |
489a5393a cpuset: don't pas... |
758 |
|
0ccea8feb cpuset: Make gene... |
759 760 761 |
if (root_load_balance && cpumask_subset(cp->cpus_allowed, top_cpuset.effective_cpus)) continue; |
cd1cb3350 sched/topology: D... |
762 763 |
if (is_sched_load_balance(cp) && !cpumask_empty(cp->effective_cpus)) |
fc560a26a cpuset: replace c... |
764 |
csa[csn++] = cp; |
0ccea8feb cpuset: Make gene... |
765 766 767 |
/* skip @cp's subtree if not a partition root */ if (!is_partition_root(cp)) pos_css = css_rightmost_descendant(pos_css); |
fc560a26a cpuset: replace c... |
768 769 |
} rcu_read_unlock(); |
029190c51 cpuset sched_load... |
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 |
for (i = 0; i < csn; i++) csa[i]->pn = i; ndoms = csn; restart: /* Find the best partition (set of sched domains) */ for (i = 0; i < csn; i++) { struct cpuset *a = csa[i]; int apn = a->pn; for (j = 0; j < csn; j++) { struct cpuset *b = csa[j]; int bpn = b->pn; if (apn != bpn && cpusets_overlap(a, b)) { for (k = 0; k < csn; k++) { struct cpuset *c = csa[k]; if (c->pn == bpn) c->pn = apn; } ndoms--; /* one less element */ goto restart; } } } |
cf417141c sched, cpuset: re... |
797 798 799 800 |
/* * Now we know how many domains to create. * Convert <csn, csa> to <ndoms, doms> and populate cpu masks. */ |
acc3f5d7c cpumask: Partitio... |
801 |
doms = alloc_sched_domains(ndoms); |
700018e0a cpuset: fix regre... |
802 |
if (!doms) |
cf417141c sched, cpuset: re... |
803 |
goto done; |
cf417141c sched, cpuset: re... |
804 805 806 807 808 |
/* * The rest of the code, including the scheduler, can deal with * dattr==NULL case. No need to abort if alloc fails. */ |
6da2ec560 treewide: kmalloc... |
809 810 |
dattr = kmalloc_array(ndoms, sizeof(struct sched_domain_attr), GFP_KERNEL); |
029190c51 cpuset sched_load... |
811 812 813 |
for (nslot = 0, i = 0; i < csn; i++) { struct cpuset *a = csa[i]; |
6af866af3 cpuset: remove re... |
814 |
struct cpumask *dp; |
029190c51 cpuset sched_load... |
815 |
int apn = a->pn; |
cf417141c sched, cpuset: re... |
816 817 818 819 |
if (apn < 0) { /* Skip completed partitions */ continue; } |
acc3f5d7c cpumask: Partitio... |
820 |
dp = doms[nslot]; |
cf417141c sched, cpuset: re... |
821 822 823 824 |
if (nslot == ndoms) { static int warnings = 10; if (warnings) { |
12d3089c1 kernel/cpuset.c: ... |
825 826 827 |
pr_warn("rebuild_sched_domains confused: nslot %d, ndoms %d, csn %d, i %d, apn %d ", nslot, ndoms, csn, i, apn); |
cf417141c sched, cpuset: re... |
828 |
warnings--; |
029190c51 cpuset sched_load... |
829 |
} |
cf417141c sched, cpuset: re... |
830 831 |
continue; } |
029190c51 cpuset sched_load... |
832 |
|
6af866af3 cpuset: remove re... |
833 |
cpumask_clear(dp); |
cf417141c sched, cpuset: re... |
834 835 836 837 838 839 |
if (dattr) *(dattr + nslot) = SD_ATTR_INIT; for (j = i; j < csn; j++) { struct cpuset *b = csa[j]; if (apn == b->pn) { |
8b5f1c52d cpuset: use effec... |
840 |
cpumask_or(dp, dp, b->effective_cpus); |
edb938217 sched/isolation: ... |
841 |
cpumask_and(dp, dp, housekeeping_cpumask(HK_FLAG_DOMAIN)); |
cf417141c sched, cpuset: re... |
842 843 844 845 846 |
if (dattr) update_domain_attr_tree(dattr + nslot, b); /* Done with this partition */ b->pn = -1; |
029190c51 cpuset sched_load... |
847 |
} |
029190c51 cpuset sched_load... |
848 |
} |
cf417141c sched, cpuset: re... |
849 |
nslot++; |
029190c51 cpuset sched_load... |
850 851 |
} BUG_ON(nslot != ndoms); |
cf417141c sched, cpuset: re... |
852 853 |
done: kfree(csa); |
700018e0a cpuset: fix regre... |
854 855 856 857 858 859 |
/* * Fallback to the default domain if kmalloc() failed. * See comments in partition_sched_domains(). */ if (doms == NULL) ndoms = 1; |
cf417141c sched, cpuset: re... |
860 861 862 863 |
*domains = doms; *attributes = dattr; return ndoms; } |
f9a25f776 cpusets: Rebuild ... |
864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 |
static void update_tasks_root_domain(struct cpuset *cs) { struct css_task_iter it; struct task_struct *task; css_task_iter_start(&cs->css, 0, &it); while ((task = css_task_iter_next(&it))) dl_add_task_root_domain(task); css_task_iter_end(&it); } static void rebuild_root_domains(void) { struct cpuset *cs = NULL; struct cgroup_subsys_state *pos_css; |
18695a120 Revert "cgroup/cp... |
881 |
lockdep_assert_held(&cpuset_mutex); |
f9a25f776 cpusets: Rebuild ... |
882 883 |
lockdep_assert_cpus_held(); lockdep_assert_held(&sched_domains_mutex); |
f9a25f776 cpusets: Rebuild ... |
884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 |
rcu_read_lock(); /* * Clear default root domain DL accounting, it will be computed again * if a task belongs to it. */ dl_clear_root_domain(&def_root_domain); cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { if (cpumask_empty(cs->effective_cpus)) { pos_css = css_rightmost_descendant(pos_css); continue; } css_get(&cs->css); rcu_read_unlock(); update_tasks_root_domain(cs); rcu_read_lock(); css_put(&cs->css); } rcu_read_unlock(); } static void partition_and_rebuild_sched_domains(int ndoms_new, cpumask_var_t doms_new[], struct sched_domain_attr *dattr_new) { mutex_lock(&sched_domains_mutex); partition_sched_domains_locked(ndoms_new, doms_new, dattr_new); rebuild_root_domains(); mutex_unlock(&sched_domains_mutex); } |
cf417141c sched, cpuset: re... |
920 921 922 |
/* * Rebuild scheduler domains. * |
699140ba8 cpuset: drop asyn... |
923 924 925 926 927 |
* If the flag 'sched_load_balance' of any cpuset with non-empty * 'cpus' changes, or if the 'cpus' allowed changes in any cpuset * which has that flag enabled, or if any cpuset with a non-empty * 'cpus' is removed, then call this routine to rebuild the * scheduler's dynamic sched domains. |
cf417141c sched, cpuset: re... |
928 |
* |
5d21cc2db cpuset: replace c... |
929 |
* Call with cpuset_mutex held. Takes get_online_cpus(). |
cf417141c sched, cpuset: re... |
930 |
*/ |
699140ba8 cpuset: drop asyn... |
931 |
static void rebuild_sched_domains_locked(void) |
cf417141c sched, cpuset: re... |
932 |
{ |
771b663fa cpuset: fix race ... |
933 |
struct cgroup_subsys_state *pos_css; |
cf417141c sched, cpuset: re... |
934 |
struct sched_domain_attr *attr; |
acc3f5d7c cpumask: Partitio... |
935 |
cpumask_var_t *doms; |
771b663fa cpuset: fix race ... |
936 |
struct cpuset *cs; |
cf417141c sched, cpuset: re... |
937 |
int ndoms; |
18695a120 Revert "cgroup/cp... |
938 |
lockdep_assert_held(&cpuset_mutex); |
cf417141c sched, cpuset: re... |
939 |
|
5b16c2a49 cpuset: fix cpu h... |
940 |
/* |
771b663fa cpuset: fix race ... |
941 |
* If we have raced with CPU hotplug, return early to avoid |
5b16c2a49 cpuset: fix cpu h... |
942 |
* passing doms with offlined cpu to partition_sched_domains(). |
771b663fa cpuset: fix race ... |
943 944 945 946 947 |
* Anyways, cpuset_hotplug_workfn() will rebuild sched domains. * * With no CPUs in any subpartitions, top_cpuset's effective CPUs * should be the same as the active CPUs, so checking only top_cpuset * is enough to detect racing CPU offlines. |
5b16c2a49 cpuset: fix cpu h... |
948 |
*/ |
0ccea8feb cpuset: Make gene... |
949 950 |
if (!top_cpuset.nr_subparts_cpus && !cpumask_equal(top_cpuset.effective_cpus, cpu_active_mask)) |
d74b27d63 cgroup/cpuset: Ch... |
951 |
return; |
0ccea8feb cpuset: Make gene... |
952 |
|
771b663fa cpuset: fix race ... |
953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 |
/* * With subpartition CPUs, however, the effective CPUs of a partition * root should be only a subset of the active CPUs. Since a CPU in any * partition root could be offlined, all must be checked. */ if (top_cpuset.nr_subparts_cpus) { rcu_read_lock(); cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { if (!is_partition_root(cs)) { pos_css = css_rightmost_descendant(pos_css); continue; } if (!cpumask_subset(cs->effective_cpus, cpu_active_mask)) { rcu_read_unlock(); return; } } rcu_read_unlock(); } |
5b16c2a49 cpuset: fix cpu h... |
973 |
|
cf417141c sched, cpuset: re... |
974 |
/* Generate domain masks and attrs */ |
cf417141c sched, cpuset: re... |
975 |
ndoms = generate_sched_domains(&doms, &attr); |
cf417141c sched, cpuset: re... |
976 977 |
/* Have scheduler rebuild the domains */ |
f9a25f776 cpusets: Rebuild ... |
978 |
partition_and_rebuild_sched_domains(ndoms, doms, attr); |
cf417141c sched, cpuset: re... |
979 |
} |
db7f47cf4 cpusets: allow cp... |
980 |
#else /* !CONFIG_SMP */ |
699140ba8 cpuset: drop asyn... |
981 |
static void rebuild_sched_domains_locked(void) |
db7f47cf4 cpusets: allow cp... |
982 983 |
{ } |
db7f47cf4 cpusets: allow cp... |
984 |
#endif /* CONFIG_SMP */ |
029190c51 cpuset sched_load... |
985 |
|
cf417141c sched, cpuset: re... |
986 987 |
void rebuild_sched_domains(void) { |
d74b27d63 cgroup/cpuset: Ch... |
988 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
989 |
mutex_lock(&cpuset_mutex); |
699140ba8 cpuset: drop asyn... |
990 |
rebuild_sched_domains_locked(); |
18695a120 Revert "cgroup/cp... |
991 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
992 |
put_online_cpus(); |
029190c51 cpuset sched_load... |
993 |
} |
4d1ac6a16 ANDROID: sched/cp... |
994 995 996 997 998 999 1000 1001 1002 1003 1004 |
static int update_cpus_allowed(struct cpuset *cs, struct task_struct *p, const struct cpumask *new_mask) { int ret = -EINVAL; trace_android_rvh_update_cpus_allowed(p, cs->cpus_requested, new_mask, &ret); if (!ret) return ret; return set_cpus_allowed_ptr(p, new_mask); } |
58f4790b7 cpusets: update_c... |
1005 |
/** |
0b2f630a2 cpusets: restruct... |
1006 1007 |
* update_tasks_cpumask - Update the cpumasks of tasks in the cpuset. * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed |
0b2f630a2 cpusets: restruct... |
1008 |
* |
d66393e54 cpuset: use css_t... |
1009 1010 1011 |
* Iterate through each task of @cs updating its cpus_allowed to the * effective cpuset's. As this function is called with cpuset_mutex held, * cpuset membership stays stable. |
0b2f630a2 cpusets: restruct... |
1012 |
*/ |
d66393e54 cpuset: use css_t... |
1013 |
static void update_tasks_cpumask(struct cpuset *cs) |
0b2f630a2 cpusets: restruct... |
1014 |
{ |
d66393e54 cpuset: use css_t... |
1015 1016 |
struct css_task_iter it; struct task_struct *task; |
bc2fb7ed0 cgroup: add @flag... |
1017 |
css_task_iter_start(&cs->css, 0, &it); |
d66393e54 cpuset: use css_t... |
1018 |
while ((task = css_task_iter_next(&it))) |
4d1ac6a16 ANDROID: sched/cp... |
1019 |
update_cpus_allowed(cs, task, cs->effective_cpus); |
d66393e54 cpuset: use css_t... |
1020 |
css_task_iter_end(&it); |
0b2f630a2 cpusets: restruct... |
1021 |
} |
ee8dde0cd cpuset: Add new v... |
1022 1023 1024 1025 1026 1027 1028 |
/** * compute_effective_cpumask - Compute the effective cpumask of the cpuset * @new_cpus: the temp variable for the new effective_cpus mask * @cs: the cpuset the need to recompute the new effective_cpus mask * @parent: the parent cpuset * * If the parent has subpartition CPUs, include them in the list of |
4b842da27 cpuset: Make CPU ... |
1029 1030 1031 |
* allowable CPUs in computing the new effective_cpus mask. Since offlined * CPUs are not removed from subparts_cpus, we have to use cpu_active_mask * to mask those out. |
ee8dde0cd cpuset: Add new v... |
1032 1033 1034 1035 1036 1037 1038 |
*/ static void compute_effective_cpumask(struct cpumask *new_cpus, struct cpuset *cs, struct cpuset *parent) { if (parent->nr_subparts_cpus) { cpumask_or(new_cpus, parent->effective_cpus, parent->subparts_cpus); |
c8dc4422c ANDROID: cpuset: ... |
1039 |
cpumask_and(new_cpus, new_cpus, cs->cpus_requested); |
4b842da27 cpuset: Make CPU ... |
1040 |
cpumask_and(new_cpus, new_cpus, cpu_active_mask); |
ee8dde0cd cpuset: Add new v... |
1041 |
} else { |
c8dc4422c ANDROID: cpuset: ... |
1042 |
cpumask_and(new_cpus, cs->cpus_requested, parent_cs(cs)->effective_cpus); |
ee8dde0cd cpuset: Add new v... |
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 |
} } /* * Commands for update_parent_subparts_cpumask */ enum subparts_cmd { partcmd_enable, /* Enable partition root */ partcmd_disable, /* Disable partition root */ partcmd_update, /* Update parent's subparts_cpus */ }; /** * update_parent_subparts_cpumask - update subparts_cpus mask of parent cpuset * @cpuset: The cpuset that requests change in partition root state * @cmd: Partition root state change command * @newmask: Optional new cpumask for partcmd_update * @tmp: Temporary addmask and delmask * Return: 0, 1 or an error code * * For partcmd_enable, the cpuset is being transformed from a non-partition * root to a partition root. The cpus_allowed mask of the given cpuset will * be put into parent's subparts_cpus and taken away from parent's * effective_cpus. The function will return 0 if all the CPUs listed in * cpus_allowed can be granted or an error code will be returned. * * For partcmd_disable, the cpuset is being transofrmed from a partition * root back to a non-partition root. any CPUs in cpus_allowed that are in * parent's subparts_cpus will be taken away from that cpumask and put back * into parent's effective_cpus. 0 should always be returned. * * For partcmd_update, if the optional newmask is specified, the cpu * list is to be changed from cpus_allowed to newmask. Otherwise, |
3881b8612 cpuset: Add an er... |
1076 1077 1078 1079 1080 1081 |
* cpus_allowed is assumed to remain the same. The cpuset should either * be a partition root or an invalid partition root. The partition root * state may change if newmask is NULL and none of the requested CPUs can * be granted by the parent. The function will return 1 if changes to * parent's subparts_cpus and effective_cpus happen or 0 otherwise. * Error code should only be returned when newmask is non-NULL. |
ee8dde0cd cpuset: Add new v... |
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 |
* * The partcmd_enable and partcmd_disable commands are used by * update_prstate(). The partcmd_update command is used by * update_cpumasks_hier() with newmask NULL and update_cpumask() with * newmask set. * * The checking is more strict when enabling partition root than the * other two commands. * * Because of the implicit cpu exclusive nature of a partition root, * cpumask changes that violates the cpu exclusivity rule will not be * permitted when checked by validate_change(). The validate_change() * function will also prevent any changes to the cpu list if it is not * a superset of children's cpu lists. */ static int update_parent_subparts_cpumask(struct cpuset *cpuset, int cmd, struct cpumask *newmask, struct tmpmasks *tmp) { struct cpuset *parent = parent_cs(cpuset); int adding; /* Moving cpus from effective_cpus to subparts_cpus */ int deleting; /* Moving cpus from subparts_cpus to effective_cpus */ |
3881b8612 cpuset: Add an er... |
1104 |
bool part_error = false; /* Partition error? */ |
ee8dde0cd cpuset: Add new v... |
1105 |
|
18695a120 Revert "cgroup/cp... |
1106 |
lockdep_assert_held(&cpuset_mutex); |
ee8dde0cd cpuset: Add new v... |
1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 |
/* * The parent must be a partition root. * The new cpumask, if present, or the current cpus_allowed must * not be empty. */ if (!is_partition_root(parent) || (newmask && cpumask_empty(newmask)) || (!newmask && cpumask_empty(cpuset->cpus_allowed))) return -EINVAL; /* * Enabling/disabling partition root is not allowed if there are * online children. */ if ((cmd != partcmd_update) && css_has_online_children(&cpuset->css)) return -EBUSY; /* * Enabling partition root is not allowed if not all the CPUs * can be granted from parent's effective_cpus or at least one * CPU will be left after that. */ if ((cmd == partcmd_enable) && (!cpumask_subset(cpuset->cpus_allowed, parent->effective_cpus) || cpumask_equal(cpuset->cpus_allowed, parent->effective_cpus))) return -EINVAL; /* * A cpumask update cannot make parent's effective_cpus become empty. */ adding = deleting = false; if (cmd == partcmd_enable) { cpumask_copy(tmp->addmask, cpuset->cpus_allowed); adding = true; } else if (cmd == partcmd_disable) { deleting = cpumask_and(tmp->delmask, cpuset->cpus_allowed, parent->subparts_cpus); } else if (newmask) { /* * partcmd_update with newmask: * * delmask = cpus_allowed & ~newmask & parent->subparts_cpus * addmask = newmask & parent->effective_cpus * & ~parent->subparts_cpus */ cpumask_andnot(tmp->delmask, cpuset->cpus_allowed, newmask); deleting = cpumask_and(tmp->delmask, tmp->delmask, parent->subparts_cpus); cpumask_and(tmp->addmask, newmask, parent->effective_cpus); adding = cpumask_andnot(tmp->addmask, tmp->addmask, parent->subparts_cpus); /* * Return error if the new effective_cpus could become empty. */ |
4b842da27 cpuset: Make CPU ... |
1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 |
if (adding && cpumask_equal(parent->effective_cpus, tmp->addmask)) { if (!deleting) return -EINVAL; /* * As some of the CPUs in subparts_cpus might have * been offlined, we need to compute the real delmask * to confirm that. */ if (!cpumask_and(tmp->addmask, tmp->delmask, cpu_active_mask)) return -EINVAL; cpumask_copy(tmp->addmask, parent->effective_cpus); } |
ee8dde0cd cpuset: Add new v... |
1177 1178 1179 1180 1181 1182 1183 |
} else { /* * partcmd_update w/o newmask: * * addmask = cpus_allowed & parent->effectiveb_cpus * * Note that parent's subparts_cpus may have been |
3881b8612 cpuset: Add an er... |
1184 1185 |
* pre-shrunk in case there is a change in the cpu list. * So no deletion is needed. |
ee8dde0cd cpuset: Add new v... |
1186 1187 1188 |
*/ adding = cpumask_and(tmp->addmask, cpuset->cpus_allowed, parent->effective_cpus); |
3881b8612 cpuset: Add an er... |
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 |
part_error = cpumask_equal(tmp->addmask, parent->effective_cpus); } if (cmd == partcmd_update) { int prev_prs = cpuset->partition_root_state; /* * Check for possible transition between PRS_ENABLED * and PRS_ERROR. */ switch (cpuset->partition_root_state) { case PRS_ENABLED: if (part_error) cpuset->partition_root_state = PRS_ERROR; break; case PRS_ERROR: if (!part_error) cpuset->partition_root_state = PRS_ENABLED; break; } /* * Set part_error if previously in invalid state. */ part_error = (prev_prs == PRS_ERROR); } if (!part_error && (cpuset->partition_root_state == PRS_ERROR)) return 0; /* Nothing need to be done */ if (cpuset->partition_root_state == PRS_ERROR) { /* * Remove all its cpus from parent's subparts_cpus. */ adding = false; deleting = cpumask_and(tmp->delmask, cpuset->cpus_allowed, parent->subparts_cpus); |
ee8dde0cd cpuset: Add new v... |
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 |
} if (!adding && !deleting) return 0; /* * Change the parent's subparts_cpus. * Newly added CPUs will be removed from effective_cpus and * newly deleted ones will be added back to effective_cpus. */ spin_lock_irq(&callback_lock); if (adding) { cpumask_or(parent->subparts_cpus, parent->subparts_cpus, tmp->addmask); cpumask_andnot(parent->effective_cpus, parent->effective_cpus, tmp->addmask); } if (deleting) { cpumask_andnot(parent->subparts_cpus, parent->subparts_cpus, tmp->delmask); |
4b842da27 cpuset: Make CPU ... |
1246 1247 1248 1249 |
/* * Some of the CPUs in subparts_cpus might have been offlined. */ cpumask_and(tmp->delmask, tmp->delmask, cpu_active_mask); |
ee8dde0cd cpuset: Add new v... |
1250 1251 1252 1253 1254 1255 1256 1257 1258 |
cpumask_or(parent->effective_cpus, parent->effective_cpus, tmp->delmask); } parent->nr_subparts_cpus = cpumask_weight(parent->subparts_cpus); spin_unlock_irq(&callback_lock); return cmd == partcmd_update; } |
5c5cc6232 cpuset: allow to ... |
1259 |
/* |
734d45130 cpuset: update cs... |
1260 |
* update_cpumasks_hier - Update effective cpumasks and tasks in the subtree |
ee8dde0cd cpuset: Add new v... |
1261 1262 |
* @cs: the cpuset to consider * @tmp: temp variables for calculating effective_cpus & partition setup |
734d45130 cpuset: update cs... |
1263 1264 1265 |
* * When congifured cpumask is changed, the effective cpumasks of this cpuset * and all its descendants need to be updated. |
5c5cc6232 cpuset: allow to ... |
1266 |
* |
734d45130 cpuset: update cs... |
1267 |
* On legacy hierachy, effective_cpus will be the same with cpu_allowed. |
5c5cc6232 cpuset: allow to ... |
1268 1269 1270 |
* * Called with cpuset_mutex held */ |
ee8dde0cd cpuset: Add new v... |
1271 |
static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp) |
5c5cc6232 cpuset: allow to ... |
1272 1273 |
{ struct cpuset *cp; |
492eb21b9 cgroup: make hier... |
1274 |
struct cgroup_subsys_state *pos_css; |
8b5f1c52d cpuset: use effec... |
1275 |
bool need_rebuild_sched_domains = false; |
5c5cc6232 cpuset: allow to ... |
1276 1277 |
rcu_read_lock(); |
734d45130 cpuset: update cs... |
1278 1279 |
cpuset_for_each_descendant_pre(cp, pos_css, cs) { struct cpuset *parent = parent_cs(cp); |
ee8dde0cd cpuset: Add new v... |
1280 |
compute_effective_cpumask(tmp->new_cpus, cp, parent); |
734d45130 cpuset: update cs... |
1281 |
|
554b0d1c8 cpuset: inherit a... |
1282 1283 1284 1285 |
/* * If it becomes empty, inherit the effective mask of the * parent, which is guaranteed to have some CPUs. */ |
4716909cc cpuset: Track cpu... |
1286 |
if (is_in_v2_mode() && cpumask_empty(tmp->new_cpus)) { |
ee8dde0cd cpuset: Add new v... |
1287 |
cpumask_copy(tmp->new_cpus, parent->effective_cpus); |
4716909cc cpuset: Track cpu... |
1288 1289 1290 1291 1292 1293 1294 1295 1296 |
if (!cp->use_parent_ecpus) { cp->use_parent_ecpus = true; parent->child_ecpus_count++; } } else if (cp->use_parent_ecpus) { cp->use_parent_ecpus = false; WARN_ON_ONCE(!parent->child_ecpus_count); parent->child_ecpus_count--; } |
554b0d1c8 cpuset: inherit a... |
1297 |
|
ee8dde0cd cpuset: Add new v... |
1298 1299 1300 1301 |
/* * Skip the whole subtree if the cpumask remains the same * and has no partition root state. */ |
3881b8612 cpuset: Add an er... |
1302 |
if (!cp->partition_root_state && |
ee8dde0cd cpuset: Add new v... |
1303 |
cpumask_equal(tmp->new_cpus, cp->effective_cpus)) { |
734d45130 cpuset: update cs... |
1304 1305 |
pos_css = css_rightmost_descendant(pos_css); continue; |
5c5cc6232 cpuset: allow to ... |
1306 |
} |
734d45130 cpuset: update cs... |
1307 |
|
ee8dde0cd cpuset: Add new v... |
1308 1309 1310 1311 1312 1313 |
/* * update_parent_subparts_cpumask() should have been called * for cs already in update_cpumask(). We should also call * update_tasks_cpumask() again for tasks in the parent * cpuset if the parent's subparts_cpus changes. */ |
3881b8612 cpuset: Add an er... |
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 |
if ((cp != cs) && cp->partition_root_state) { switch (parent->partition_root_state) { case PRS_DISABLED: /* * If parent is not a partition root or an * invalid partition root, clear the state * state and the CS_CPU_EXCLUSIVE flag. */ WARN_ON_ONCE(cp->partition_root_state != PRS_ERROR); cp->partition_root_state = 0; /* * clear_bit() is an atomic operation and * readers aren't interested in the state * of CS_CPU_EXCLUSIVE anyway. So we can * just update the flag without holding * the callback_lock. */ clear_bit(CS_CPU_EXCLUSIVE, &cp->flags); break; case PRS_ENABLED: if (update_parent_subparts_cpumask(cp, partcmd_update, NULL, tmp)) update_tasks_cpumask(parent); break; case PRS_ERROR: /* * When parent is invalid, it has to be too. */ cp->partition_root_state = PRS_ERROR; if (cp->nr_subparts_cpus) { cp->nr_subparts_cpus = 0; cpumask_clear(cp->subparts_cpus); } break; } |
ee8dde0cd cpuset: Add new v... |
1352 |
} |
ec903c0c8 cgroup: rename cs... |
1353 |
if (!css_tryget_online(&cp->css)) |
5c5cc6232 cpuset: allow to ... |
1354 1355 |
continue; rcu_read_unlock(); |
8447a0fee cpuset: convert c... |
1356 |
spin_lock_irq(&callback_lock); |
ee8dde0cd cpuset: Add new v... |
1357 1358 |
cpumask_copy(cp->effective_cpus, tmp->new_cpus); |
3881b8612 cpuset: Add an er... |
1359 1360 1361 1362 1363 |
if (cp->nr_subparts_cpus && (cp->partition_root_state != PRS_ENABLED)) { cp->nr_subparts_cpus = 0; cpumask_clear(cp->subparts_cpus); } else if (cp->nr_subparts_cpus) { |
ee8dde0cd cpuset: Add new v... |
1364 1365 1366 |
/* * Make sure that effective_cpus & subparts_cpus * are mutually exclusive. |
3881b8612 cpuset: Add an er... |
1367 1368 1369 1370 1371 |
* * In the unlikely event that effective_cpus * becomes empty. we clear cp->nr_subparts_cpus and * let its child partition roots to compete for * CPUs again. |
ee8dde0cd cpuset: Add new v... |
1372 1373 1374 |
*/ cpumask_andnot(cp->effective_cpus, cp->effective_cpus, cp->subparts_cpus); |
3881b8612 cpuset: Add an er... |
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 |
if (cpumask_empty(cp->effective_cpus)) { cpumask_copy(cp->effective_cpus, tmp->new_cpus); cpumask_clear(cp->subparts_cpus); cp->nr_subparts_cpus = 0; } else if (!cpumask_subset(cp->subparts_cpus, tmp->new_cpus)) { cpumask_andnot(cp->subparts_cpus, cp->subparts_cpus, tmp->new_cpus); cp->nr_subparts_cpus = cpumask_weight(cp->subparts_cpus); } |
ee8dde0cd cpuset: Add new v... |
1386 |
} |
8447a0fee cpuset: convert c... |
1387 |
spin_unlock_irq(&callback_lock); |
734d45130 cpuset: update cs... |
1388 |
|
b8d1b8ee9 cpuset: Allow v2 ... |
1389 |
WARN_ON(!is_in_v2_mode() && |
734d45130 cpuset: update cs... |
1390 |
!cpumask_equal(cp->cpus_allowed, cp->effective_cpus)); |
d66393e54 cpuset: use css_t... |
1391 |
update_tasks_cpumask(cp); |
5c5cc6232 cpuset: allow to ... |
1392 |
|
8b5f1c52d cpuset: use effec... |
1393 |
/* |
0ccea8feb cpuset: Make gene... |
1394 1395 1396 1397 |
* On legacy hierarchy, if the effective cpumask of any non- * empty cpuset is changed, we need to rebuild sched domains. * On default hierarchy, the cpuset needs to be a partition * root as well. |
8b5f1c52d cpuset: use effec... |
1398 1399 |
*/ if (!cpumask_empty(cp->cpus_allowed) && |
0ccea8feb cpuset: Make gene... |
1400 1401 1402 |
is_sched_load_balance(cp) && (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || is_partition_root(cp))) |
8b5f1c52d cpuset: use effec... |
1403 |
need_rebuild_sched_domains = true; |
5c5cc6232 cpuset: allow to ... |
1404 1405 1406 1407 |
rcu_read_lock(); css_put(&cp->css); } rcu_read_unlock(); |
8b5f1c52d cpuset: use effec... |
1408 1409 1410 |
if (need_rebuild_sched_domains) rebuild_sched_domains_locked(); |
5c5cc6232 cpuset: allow to ... |
1411 |
} |
0b2f630a2 cpusets: restruct... |
1412 |
/** |
4716909cc cpuset: Track cpu... |
1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 |
* update_sibling_cpumasks - Update siblings cpumasks * @parent: Parent cpuset * @cs: Current cpuset * @tmp: Temp variables */ static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs, struct tmpmasks *tmp) { struct cpuset *sibling; struct cgroup_subsys_state *pos_css; /* * Check all its siblings and call update_cpumasks_hier() * if their use_parent_ecpus flag is set in order for them * to use the right effective_cpus value. */ rcu_read_lock(); cpuset_for_each_child(sibling, pos_css, parent) { if (sibling == cs) continue; if (!sibling->use_parent_ecpus) continue; update_cpumasks_hier(sibling, tmp); } rcu_read_unlock(); } /** |
58f4790b7 cpusets: update_c... |
1442 1443 |
* update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it * @cs: the cpuset to consider |
fc34ac1dc kernel/cpuset.c: ... |
1444 |
* @trialcs: trial cpuset |
58f4790b7 cpusets: update_c... |
1445 1446 |
* @buf: buffer of cpu numbers written to this cpuset */ |
645fcc9d2 cpuset: don't all... |
1447 1448 |
static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, const char *buf) |
1da177e4c Linux-2.6.12-rc2 |
1449 |
{ |
58f4790b7 cpusets: update_c... |
1450 |
int retval; |
ee8dde0cd cpuset: Add new v... |
1451 |
struct tmpmasks tmp; |
1da177e4c Linux-2.6.12-rc2 |
1452 |
|
5f054e31c documentation: re... |
1453 |
/* top_cpuset.cpus_allowed tracks cpu_online_mask; it's read-only */ |
4c4d50f7b [PATCH] cpuset: t... |
1454 1455 |
if (cs == &top_cpuset) return -EACCES; |
6f7f02e78 cpusets: allow em... |
1456 |
/* |
54a1b2c15 ANDROID: kernel: ... |
1457 |
* An empty cpus_requested is ok only if the cpuset has no tasks. |
020958b62 cpusets: decrusti... |
1458 1459 1460 |
* Since cpulist_parse() fails on an empty mask, we special case * that parsing. The validate_change() call ensures that cpusets * with tasks have cpus. |
6f7f02e78 cpusets: allow em... |
1461 |
*/ |
020958b62 cpusets: decrusti... |
1462 |
if (!*buf) { |
54a1b2c15 ANDROID: kernel: ... |
1463 |
cpumask_clear(trialcs->cpus_requested); |
6f7f02e78 cpusets: allow em... |
1464 |
} else { |
c8dc4422c ANDROID: cpuset: ... |
1465 |
retval = cpulist_parse(buf, trialcs->cpus_requested); |
6f7f02e78 cpusets: allow em... |
1466 1467 |
if (retval < 0) return retval; |
54a1b2c15 ANDROID: kernel: ... |
1468 |
} |
37340746a cpusets: fix bug ... |
1469 |
|
54a1b2c15 ANDROID: kernel: ... |
1470 1471 |
if (!cpumask_subset(trialcs->cpus_requested, cpu_present_mask)) return -EINVAL; |
c8dc4422c ANDROID: cpuset: ... |
1472 |
|
54a1b2c15 ANDROID: kernel: ... |
1473 |
cpumask_and(trialcs->cpus_allowed, trialcs->cpus_requested, cpu_active_mask); |
029190c51 cpuset sched_load... |
1474 |
|
8707d8b8c Fix cpusets updat... |
1475 |
/* Nothing to do if the cpus didn't change */ |
c8dc4422c ANDROID: cpuset: ... |
1476 |
if (cpumask_equal(cs->cpus_requested, trialcs->cpus_requested)) |
8707d8b8c Fix cpusets updat... |
1477 |
return 0; |
58f4790b7 cpusets: update_c... |
1478 |
|
a73456f37 cpuset: re-struct... |
1479 1480 1481 |
retval = validate_change(cs, trialcs); if (retval < 0) return retval; |
ee8dde0cd cpuset: Add new v... |
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 |
#ifdef CONFIG_CPUMASK_OFFSTACK /* * Use the cpumasks in trialcs for tmpmasks when they are pointers * to allocated cpumasks. */ tmp.addmask = trialcs->subparts_cpus; tmp.delmask = trialcs->effective_cpus; tmp.new_cpus = trialcs->cpus_allowed; #endif if (cs->partition_root_state) { /* Cpumask of a partition root cannot be empty */ if (cpumask_empty(trialcs->cpus_allowed)) return -EINVAL; if (update_parent_subparts_cpumask(cs, partcmd_update, trialcs->cpus_allowed, &tmp) < 0) return -EINVAL; } |
8447a0fee cpuset: convert c... |
1500 |
spin_lock_irq(&callback_lock); |
300ed6cbb cpuset: convert c... |
1501 |
cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed); |
c8dc4422c ANDROID: cpuset: ... |
1502 |
cpumask_copy(cs->cpus_requested, trialcs->cpus_requested); |
ee8dde0cd cpuset: Add new v... |
1503 1504 1505 1506 1507 1508 1509 1510 1511 |
/* * Make sure that subparts_cpus is a subset of cpus_allowed. */ if (cs->nr_subparts_cpus) { cpumask_andnot(cs->subparts_cpus, cs->subparts_cpus, cs->cpus_allowed); cs->nr_subparts_cpus = cpumask_weight(cs->subparts_cpus); } |
8447a0fee cpuset: convert c... |
1512 |
spin_unlock_irq(&callback_lock); |
029190c51 cpuset sched_load... |
1513 |
|
ee8dde0cd cpuset: Add new v... |
1514 |
update_cpumasks_hier(cs, &tmp); |
4716909cc cpuset: Track cpu... |
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 |
if (cs->partition_root_state) { struct cpuset *parent = parent_cs(cs); /* * For partition root, update the cpumasks of sibling * cpusets if they use parent's effective_cpus. */ if (parent->child_ecpus_count) update_sibling_cpumasks(parent, cs, &tmp); } |
85d7b9498 [PATCH] Dynamic s... |
1526 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1527 |
} |
053199edf [PATCH] cpusets: ... |
1528 |
/* |
e93ad19d0 cpuset: make mm m... |
1529 1530 1531 1532 1533 |
* Migrate memory region from one set of nodes to another. This is * performed asynchronously as it can be called from process migration path * holding locks involved in process management. All mm migrations are * performed in the queued order and can be waited for by flushing * cpuset_migrate_mm_wq. |
e4e364e86 [PATCH] cpuset: m... |
1534 |
*/ |
e93ad19d0 cpuset: make mm m... |
1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 |
struct cpuset_migrate_mm_work { struct work_struct work; struct mm_struct *mm; nodemask_t from; nodemask_t to; }; static void cpuset_migrate_mm_workfn(struct work_struct *work) { struct cpuset_migrate_mm_work *mwork = container_of(work, struct cpuset_migrate_mm_work, work); /* on a wq worker, no need to worry about %current's mems_allowed */ do_migrate_pages(mwork->mm, &mwork->from, &mwork->to, MPOL_MF_MOVE_ALL); mmput(mwork->mm); kfree(mwork); } |
e4e364e86 [PATCH] cpuset: m... |
1552 1553 1554 |
static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, const nodemask_t *to) { |
e93ad19d0 cpuset: make mm m... |
1555 |
struct cpuset_migrate_mm_work *mwork; |
e4e364e86 [PATCH] cpuset: m... |
1556 |
|
e93ad19d0 cpuset: make mm m... |
1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 |
mwork = kzalloc(sizeof(*mwork), GFP_KERNEL); if (mwork) { mwork->mm = mm; mwork->from = *from; mwork->to = *to; INIT_WORK(&mwork->work, cpuset_migrate_mm_workfn); queue_work(cpuset_migrate_mm_wq, &mwork->work); } else { mmput(mm); } } |
e4e364e86 [PATCH] cpuset: m... |
1568 |
|
5cf1cacb4 cgroup, cpuset: r... |
1569 |
static void cpuset_post_attach(void) |
e93ad19d0 cpuset: make mm m... |
1570 1571 |
{ flush_workqueue(cpuset_migrate_mm_wq); |
e4e364e86 [PATCH] cpuset: m... |
1572 |
} |
3b6766fe6 cpuset: rewrite u... |
1573 |
/* |
58568d2a8 cpuset,mm: update... |
1574 1575 1576 1577 |
* cpuset_change_task_nodemask - change task's mems_allowed and mempolicy * @tsk: the task to change * @newmems: new nodes that the task will be set * |
5f155f27c mm, cpuset: alway... |
1578 1579 1580 1581 |
* We use the mems_allowed_seq seqlock to safely update both tsk->mems_allowed * and rebind an eventual tasks' mempolicy. If the task is allocating in * parallel, it might temporarily see an empty intersection, which results in * a seqlock check and retry before OOM or allocation failure. |
58568d2a8 cpuset,mm: update... |
1582 1583 1584 1585 |
*/ static void cpuset_change_task_nodemask(struct task_struct *tsk, nodemask_t *newmems) { |
c0ff7453b cpuset,mm: fix no... |
1586 |
task_lock(tsk); |
c0ff7453b cpuset,mm: fix no... |
1587 |
|
5f155f27c mm, cpuset: alway... |
1588 1589 |
local_irq_disable(); write_seqcount_begin(&tsk->mems_allowed_seq); |
c0ff7453b cpuset,mm: fix no... |
1590 |
|
cc9a6c877 cpuset: mm: reduc... |
1591 |
nodes_or(tsk->mems_allowed, tsk->mems_allowed, *newmems); |
213980c0f mm, mempolicy: si... |
1592 |
mpol_rebind_task(tsk, newmems); |
58568d2a8 cpuset,mm: update... |
1593 |
tsk->mems_allowed = *newmems; |
cc9a6c877 cpuset: mm: reduc... |
1594 |
|
5f155f27c mm, cpuset: alway... |
1595 1596 |
write_seqcount_end(&tsk->mems_allowed_seq); local_irq_enable(); |
cc9a6c877 cpuset: mm: reduc... |
1597 |
|
c0ff7453b cpuset,mm: fix no... |
1598 |
task_unlock(tsk); |
58568d2a8 cpuset,mm: update... |
1599 |
} |
8793d854e Task Control Grou... |
1600 |
static void *cpuset_being_rebound; |
0b2f630a2 cpusets: restruct... |
1601 1602 1603 |
/** * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed |
0b2f630a2 cpusets: restruct... |
1604 |
* |
d66393e54 cpuset: use css_t... |
1605 1606 1607 |
* Iterate through each task of @cs updating its mems_allowed to the * effective cpuset's. As this function is called with cpuset_mutex held, * cpuset membership stays stable. |
0b2f630a2 cpusets: restruct... |
1608 |
*/ |
d66393e54 cpuset: use css_t... |
1609 |
static void update_tasks_nodemask(struct cpuset *cs) |
1da177e4c Linux-2.6.12-rc2 |
1610 |
{ |
33ad801df cpuset: record ol... |
1611 |
static nodemask_t newmems; /* protected by cpuset_mutex */ |
d66393e54 cpuset: use css_t... |
1612 1613 |
struct css_task_iter it; struct task_struct *task; |
59dac16fb [PATCH] cpuset: u... |
1614 |
|
846a16bf0 mempolicy: rename... |
1615 |
cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ |
4225399a6 [PATCH] cpuset: r... |
1616 |
|
ae1c80238 cpuset: apply cs-... |
1617 |
guarantee_online_mems(cs, &newmems); |
33ad801df cpuset: record ol... |
1618 |
|
4225399a6 [PATCH] cpuset: r... |
1619 |
/* |
c1e8d7c6a mmap locking API:... |
1620 |
* The mpol_rebind_mm() call takes mmap_lock, which we couldn't |
3b6766fe6 cpuset: rewrite u... |
1621 1622 1623 |
* take while holding tasklist_lock. Forks can happen - the * mpol_dup() cpuset_being_rebound check will catch such forks, * and rebind their vma mempolicies too. Because we still hold |
5d21cc2db cpuset: replace c... |
1624 |
* the global cpuset_mutex, we know that no other rebind effort |
3b6766fe6 cpuset: rewrite u... |
1625 |
* will be contending for the global variable cpuset_being_rebound. |
4225399a6 [PATCH] cpuset: r... |
1626 |
* It's ok if we rebind the same mm twice; mpol_rebind_mm() |
04c19fa6f [PATCH] cpuset: m... |
1627 |
* is idempotent. Also migrate pages in each mm to new nodes. |
4225399a6 [PATCH] cpuset: r... |
1628 |
*/ |
bc2fb7ed0 cgroup: add @flag... |
1629 |
css_task_iter_start(&cs->css, 0, &it); |
d66393e54 cpuset: use css_t... |
1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 |
while ((task = css_task_iter_next(&it))) { struct mm_struct *mm; bool migrate; cpuset_change_task_nodemask(task, &newmems); mm = get_task_mm(task); if (!mm) continue; migrate = is_memory_migrate(cs); mpol_rebind_mm(mm, &cs->mems_allowed); if (migrate) cpuset_migrate_mm(mm, &cs->old_mems_allowed, &newmems); |
e93ad19d0 cpuset: make mm m... |
1645 1646 |
else mmput(mm); |
d66393e54 cpuset: use css_t... |
1647 1648 |
} css_task_iter_end(&it); |
4225399a6 [PATCH] cpuset: r... |
1649 |
|
33ad801df cpuset: record ol... |
1650 1651 1652 1653 1654 |
/* * All the tasks' nodemasks have been updated, update * cs->old_mems_allowed. */ cs->old_mems_allowed = newmems; |
2df167a30 cgroups: update c... |
1655 |
/* We're done rebinding vmas to this cpuset's new mems_allowed. */ |
8793d854e Task Control Grou... |
1656 |
cpuset_being_rebound = NULL; |
1da177e4c Linux-2.6.12-rc2 |
1657 |
} |
0b2f630a2 cpusets: restruct... |
1658 |
/* |
734d45130 cpuset: update cs... |
1659 1660 1661 |
* update_nodemasks_hier - Update effective nodemasks and tasks in the subtree * @cs: the cpuset to consider * @new_mems: a temp variable for calculating new effective_mems |
5c5cc6232 cpuset: allow to ... |
1662 |
* |
734d45130 cpuset: update cs... |
1663 1664 |
* When configured nodemask is changed, the effective nodemasks of this cpuset * and all its descendants need to be updated. |
5c5cc6232 cpuset: allow to ... |
1665 |
* |
734d45130 cpuset: update cs... |
1666 |
* On legacy hiearchy, effective_mems will be the same with mems_allowed. |
5c5cc6232 cpuset: allow to ... |
1667 1668 1669 |
* * Called with cpuset_mutex held */ |
734d45130 cpuset: update cs... |
1670 |
static void update_nodemasks_hier(struct cpuset *cs, nodemask_t *new_mems) |
5c5cc6232 cpuset: allow to ... |
1671 1672 |
{ struct cpuset *cp; |
492eb21b9 cgroup: make hier... |
1673 |
struct cgroup_subsys_state *pos_css; |
5c5cc6232 cpuset: allow to ... |
1674 1675 |
rcu_read_lock(); |
734d45130 cpuset: update cs... |
1676 1677 1678 1679 |
cpuset_for_each_descendant_pre(cp, pos_css, cs) { struct cpuset *parent = parent_cs(cp); nodes_and(*new_mems, cp->mems_allowed, parent->effective_mems); |
554b0d1c8 cpuset: inherit a... |
1680 1681 1682 1683 |
/* * If it becomes empty, inherit the effective mask of the * parent, which is guaranteed to have some MEMs. */ |
b8d1b8ee9 cpuset: Allow v2 ... |
1684 |
if (is_in_v2_mode() && nodes_empty(*new_mems)) |
554b0d1c8 cpuset: inherit a... |
1685 |
*new_mems = parent->effective_mems; |
734d45130 cpuset: update cs... |
1686 1687 1688 1689 |
/* Skip the whole subtree if the nodemask remains the same. */ if (nodes_equal(*new_mems, cp->effective_mems)) { pos_css = css_rightmost_descendant(pos_css); continue; |
5c5cc6232 cpuset: allow to ... |
1690 |
} |
734d45130 cpuset: update cs... |
1691 |
|
ec903c0c8 cgroup: rename cs... |
1692 |
if (!css_tryget_online(&cp->css)) |
5c5cc6232 cpuset: allow to ... |
1693 1694 |
continue; rcu_read_unlock(); |
8447a0fee cpuset: convert c... |
1695 |
spin_lock_irq(&callback_lock); |
734d45130 cpuset: update cs... |
1696 |
cp->effective_mems = *new_mems; |
8447a0fee cpuset: convert c... |
1697 |
spin_unlock_irq(&callback_lock); |
734d45130 cpuset: update cs... |
1698 |
|
b8d1b8ee9 cpuset: Allow v2 ... |
1699 |
WARN_ON(!is_in_v2_mode() && |
a13812683 cpuset: fix the W... |
1700 |
!nodes_equal(cp->mems_allowed, cp->effective_mems)); |
734d45130 cpuset: update cs... |
1701 |
|
d66393e54 cpuset: use css_t... |
1702 |
update_tasks_nodemask(cp); |
5c5cc6232 cpuset: allow to ... |
1703 1704 1705 1706 1707 1708 1709 1710 |
rcu_read_lock(); css_put(&cp->css); } rcu_read_unlock(); } /* |
0b2f630a2 cpusets: restruct... |
1711 1712 |
* Handle user request to change the 'mems' memory placement * of a cpuset. Needs to validate the request, update the |
58568d2a8 cpuset,mm: update... |
1713 1714 1715 1716 |
* cpusets mems_allowed, and for each task in the cpuset, * update mems_allowed and rebind task's mempolicy and any vma * mempolicies and if the cpuset is marked 'memory_migrate', * migrate the tasks pages to the new memory. |
0b2f630a2 cpusets: restruct... |
1717 |
* |
8447a0fee cpuset: convert c... |
1718 |
* Call with cpuset_mutex held. May take callback_lock during call. |
0b2f630a2 cpusets: restruct... |
1719 |
* Will take tasklist_lock, scan tasklist for tasks in cpuset cs, |
c1e8d7c6a mmap locking API:... |
1720 |
* lock each such tasks mm->mmap_lock, scan its vma's and rebind |
0b2f630a2 cpusets: restruct... |
1721 1722 |
* their mempolicies to the cpusets new mems_allowed. */ |
645fcc9d2 cpuset: don't all... |
1723 1724 |
static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, const char *buf) |
0b2f630a2 cpusets: restruct... |
1725 |
{ |
0b2f630a2 cpusets: restruct... |
1726 1727 1728 |
int retval; /* |
38d7bee9d cpuset: use N_MEM... |
1729 |
* top_cpuset.mems_allowed tracks node_stats[N_MEMORY]; |
0b2f630a2 cpusets: restruct... |
1730 1731 |
* it's read-only */ |
53feb2976 cpuset: alloc nod... |
1732 1733 1734 1735 |
if (cs == &top_cpuset) { retval = -EACCES; goto done; } |
0b2f630a2 cpusets: restruct... |
1736 |
|
0b2f630a2 cpusets: restruct... |
1737 1738 1739 1740 1741 1742 1743 |
/* * An empty mems_allowed is ok iff there are no tasks in the cpuset. * Since nodelist_parse() fails on an empty mask, we special case * that parsing. The validate_change() call ensures that cpusets * with tasks have memory. */ if (!*buf) { |
645fcc9d2 cpuset: don't all... |
1744 |
nodes_clear(trialcs->mems_allowed); |
0b2f630a2 cpusets: restruct... |
1745 |
} else { |
645fcc9d2 cpuset: don't all... |
1746 |
retval = nodelist_parse(buf, trialcs->mems_allowed); |
0b2f630a2 cpusets: restruct... |
1747 1748 |
if (retval < 0) goto done; |
645fcc9d2 cpuset: don't all... |
1749 |
if (!nodes_subset(trialcs->mems_allowed, |
5d8ba82c3 cpuset: allow wri... |
1750 1751 |
top_cpuset.mems_allowed)) { retval = -EINVAL; |
53feb2976 cpuset: alloc nod... |
1752 1753 |
goto done; } |
0b2f630a2 cpusets: restruct... |
1754 |
} |
33ad801df cpuset: record ol... |
1755 1756 |
if (nodes_equal(cs->mems_allowed, trialcs->mems_allowed)) { |
0b2f630a2 cpusets: restruct... |
1757 1758 1759 |
retval = 0; /* Too easy - nothing to do */ goto done; } |
645fcc9d2 cpuset: don't all... |
1760 |
retval = validate_change(cs, trialcs); |
0b2f630a2 cpusets: restruct... |
1761 1762 |
if (retval < 0) goto done; |
8447a0fee cpuset: convert c... |
1763 |
spin_lock_irq(&callback_lock); |
645fcc9d2 cpuset: don't all... |
1764 |
cs->mems_allowed = trialcs->mems_allowed; |
8447a0fee cpuset: convert c... |
1765 |
spin_unlock_irq(&callback_lock); |
0b2f630a2 cpusets: restruct... |
1766 |
|
734d45130 cpuset: update cs... |
1767 |
/* use trialcs->mems_allowed as a temp variable */ |
24ee3cf89 cpuset: use trial... |
1768 |
update_nodemasks_hier(cs, &trialcs->mems_allowed); |
0b2f630a2 cpusets: restruct... |
1769 1770 1771 |
done: return retval; } |
77ef80c65 kernel/cpuset: cu... |
1772 |
bool current_cpuset_is_being_rebound(void) |
8793d854e Task Control Grou... |
1773 |
{ |
77ef80c65 kernel/cpuset: cu... |
1774 |
bool ret; |
391acf970 cpuset,mempolicy:... |
1775 1776 1777 1778 1779 1780 |
rcu_read_lock(); ret = task_cs(current) == cpuset_being_rebound; rcu_read_unlock(); return ret; |
8793d854e Task Control Grou... |
1781 |
} |
5be7a4792 Fix cpuset sched_... |
1782 |
static int update_relax_domain_level(struct cpuset *cs, s64 val) |
1d3504fcf sched, cpuset: cu... |
1783 |
{ |
db7f47cf4 cpusets: allow cp... |
1784 |
#ifdef CONFIG_SMP |
60495e776 sched: Dynamic sc... |
1785 |
if (val < -1 || val >= sched_domain_level_max) |
30e0e1781 cpuset: limit the... |
1786 |
return -EINVAL; |
db7f47cf4 cpusets: allow cp... |
1787 |
#endif |
1d3504fcf sched, cpuset: cu... |
1788 1789 1790 |
if (val != cs->relax_domain_level) { cs->relax_domain_level = val; |
300ed6cbb cpuset: convert c... |
1791 1792 |
if (!cpumask_empty(cs->cpus_allowed) && is_sched_load_balance(cs)) |
699140ba8 cpuset: drop asyn... |
1793 |
rebuild_sched_domains_locked(); |
1d3504fcf sched, cpuset: cu... |
1794 1795 1796 1797 |
} return 0; } |
72ec70299 cgroup: make task... |
1798 |
/** |
950592f7b cpusets: update t... |
1799 1800 |
* update_tasks_flags - update the spread flags of tasks in the cpuset. * @cs: the cpuset in which each task's spread flags needs to be changed |
950592f7b cpusets: update t... |
1801 |
* |
d66393e54 cpuset: use css_t... |
1802 1803 1804 |
* Iterate through each task of @cs updating its spread flags. As this * function is called with cpuset_mutex held, cpuset membership stays * stable. |
950592f7b cpusets: update t... |
1805 |
*/ |
d66393e54 cpuset: use css_t... |
1806 |
static void update_tasks_flags(struct cpuset *cs) |
950592f7b cpusets: update t... |
1807 |
{ |
d66393e54 cpuset: use css_t... |
1808 1809 |
struct css_task_iter it; struct task_struct *task; |
bc2fb7ed0 cgroup: add @flag... |
1810 |
css_task_iter_start(&cs->css, 0, &it); |
d66393e54 cpuset: use css_t... |
1811 1812 1813 |
while ((task = css_task_iter_next(&it))) cpuset_update_task_spread_flag(cs, task); css_task_iter_end(&it); |
950592f7b cpusets: update t... |
1814 1815 1816 |
} /* |
1da177e4c Linux-2.6.12-rc2 |
1817 |
* update_flag - read a 0 or a 1 in a file and update associated flag |
786083667 Cpuset hardwall f... |
1818 1819 1820 |
* bit: the bit to update (see cpuset_flagbits_t) * cs: the cpuset to update * turning_on: whether the flag is being set or cleared |
053199edf [PATCH] cpusets: ... |
1821 |
* |
5d21cc2db cpuset: replace c... |
1822 |
* Call with cpuset_mutex held. |
1da177e4c Linux-2.6.12-rc2 |
1823 |
*/ |
700fe1ab9 CGroup API files:... |
1824 1825 |
static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, int turning_on) |
1da177e4c Linux-2.6.12-rc2 |
1826 |
{ |
645fcc9d2 cpuset: don't all... |
1827 |
struct cpuset *trialcs; |
40b6a7623 cpuset.c: remove ... |
1828 |
int balance_flag_changed; |
950592f7b cpusets: update t... |
1829 |
int spread_flag_changed; |
950592f7b cpusets: update t... |
1830 |
int err; |
1da177e4c Linux-2.6.12-rc2 |
1831 |
|
645fcc9d2 cpuset: don't all... |
1832 1833 1834 |
trialcs = alloc_trial_cpuset(cs); if (!trialcs) return -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
1835 |
if (turning_on) |
645fcc9d2 cpuset: don't all... |
1836 |
set_bit(bit, &trialcs->flags); |
1da177e4c Linux-2.6.12-rc2 |
1837 |
else |
645fcc9d2 cpuset: don't all... |
1838 |
clear_bit(bit, &trialcs->flags); |
1da177e4c Linux-2.6.12-rc2 |
1839 |
|
645fcc9d2 cpuset: don't all... |
1840 |
err = validate_change(cs, trialcs); |
85d7b9498 [PATCH] Dynamic s... |
1841 |
if (err < 0) |
645fcc9d2 cpuset: don't all... |
1842 |
goto out; |
029190c51 cpuset sched_load... |
1843 |
|
029190c51 cpuset sched_load... |
1844 |
balance_flag_changed = (is_sched_load_balance(cs) != |
645fcc9d2 cpuset: don't all... |
1845 |
is_sched_load_balance(trialcs)); |
029190c51 cpuset sched_load... |
1846 |
|
950592f7b cpusets: update t... |
1847 1848 |
spread_flag_changed = ((is_spread_slab(cs) != is_spread_slab(trialcs)) || (is_spread_page(cs) != is_spread_page(trialcs))); |
8447a0fee cpuset: convert c... |
1849 |
spin_lock_irq(&callback_lock); |
645fcc9d2 cpuset: don't all... |
1850 |
cs->flags = trialcs->flags; |
8447a0fee cpuset: convert c... |
1851 |
spin_unlock_irq(&callback_lock); |
85d7b9498 [PATCH] Dynamic s... |
1852 |
|
300ed6cbb cpuset: convert c... |
1853 |
if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) |
699140ba8 cpuset: drop asyn... |
1854 |
rebuild_sched_domains_locked(); |
029190c51 cpuset sched_load... |
1855 |
|
950592f7b cpusets: update t... |
1856 |
if (spread_flag_changed) |
d66393e54 cpuset: use css_t... |
1857 |
update_tasks_flags(cs); |
645fcc9d2 cpuset: don't all... |
1858 |
out: |
bf92370c0 cpuset: Simply al... |
1859 |
free_cpuset(trialcs); |
645fcc9d2 cpuset: don't all... |
1860 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
1861 |
} |
053199edf [PATCH] cpusets: ... |
1862 |
/* |
ee8dde0cd cpuset: Add new v... |
1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 |
* update_prstate - update partititon_root_state * cs: the cpuset to update * val: 0 - disabled, 1 - enabled * * Call with cpuset_mutex held. */ static int update_prstate(struct cpuset *cs, int val) { int err; struct cpuset *parent = parent_cs(cs); struct tmpmasks tmp; if ((val != 0) && (val != 1)) return -EINVAL; if (val == cs->partition_root_state) return 0; /* |
3881b8612 cpuset: Add an er... |
1881 |
* Cannot force a partial or invalid partition root to a full |
ee8dde0cd cpuset: Add new v... |
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 |
* partition root. */ if (val && cs->partition_root_state) return -EINVAL; if (alloc_cpumasks(NULL, &tmp)) return -ENOMEM; err = -EINVAL; if (!cs->partition_root_state) { /* * Turning on partition root requires setting the * CS_CPU_EXCLUSIVE bit implicitly as well and cpus_allowed * cannot be NULL. */ if (cpumask_empty(cs->cpus_allowed)) goto out; err = update_flag(CS_CPU_EXCLUSIVE, cs, 1); if (err) goto out; err = update_parent_subparts_cpumask(cs, partcmd_enable, NULL, &tmp); if (err) { update_flag(CS_CPU_EXCLUSIVE, cs, 0); goto out; } cs->partition_root_state = PRS_ENABLED; } else { |
3881b8612 cpuset: Add an er... |
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 |
/* * Turning off partition root will clear the * CS_CPU_EXCLUSIVE bit. */ if (cs->partition_root_state == PRS_ERROR) { cs->partition_root_state = 0; update_flag(CS_CPU_EXCLUSIVE, cs, 0); err = 0; goto out; } |
ee8dde0cd cpuset: Add new v... |
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 |
err = update_parent_subparts_cpumask(cs, partcmd_disable, NULL, &tmp); if (err) goto out; cs->partition_root_state = 0; /* Turning off CS_CPU_EXCLUSIVE will not return error */ update_flag(CS_CPU_EXCLUSIVE, cs, 0); } /* * Update cpumask of parent's tasks except when it is the top * cpuset as some system daemons cannot be mapped to other CPUs. */ if (parent != &top_cpuset) update_tasks_cpumask(parent); |
4716909cc cpuset: Track cpu... |
1939 1940 |
if (parent->child_ecpus_count) update_sibling_cpumasks(parent, cs, &tmp); |
ee8dde0cd cpuset: Add new v... |
1941 1942 1943 |
rebuild_sched_domains_locked(); out: free_cpumasks(NULL, &tmp); |
645fcc9d2 cpuset: don't all... |
1944 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
1945 |
} |
053199edf [PATCH] cpusets: ... |
1946 |
/* |
80f7228b5 typo fixes: occur... |
1947 |
* Frequency meter - How fast is some event occurring? |
3e0d98b9f [PATCH] cpuset: m... |
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 |
* * These routines manage a digitally filtered, constant time based, * event frequency meter. There are four routines: * fmeter_init() - initialize a frequency meter. * fmeter_markevent() - called each time the event happens. * fmeter_getrate() - returns the recent rate of such events. * fmeter_update() - internal routine used to update fmeter. * * A common data structure is passed to each of these routines, * which is used to keep track of the state required to manage the * frequency meter and its digital filter. * * The filter works on the number of events marked per unit time. * The filter is single-pole low-pass recursive (IIR). The time unit * is 1 second. Arithmetic is done using 32-bit integers scaled to * simulate 3 decimal digits of precision (multiplied by 1000). * * With an FM_COEF of 933, and a time base of 1 second, the filter * has a half-life of 10 seconds, meaning that if the events quit * happening, then the rate returned from the fmeter_getrate() * will be cut in half each 10 seconds, until it converges to zero. * * It is not worth doing a real infinitely recursive filter. If more * than FM_MAXTICKS ticks have elapsed since the last filter event, * just compute FM_MAXTICKS ticks worth, by which point the level * will be stable. * * Limit the count of unprocessed events to FM_MAXCNT, so as to avoid * arithmetic overflow in the fmeter_update() routine. * * Given the simple 32 bit integer arithmetic used, this meter works * best for reporting rates between one per millisecond (msec) and * one per 32 (approx) seconds. At constant rates faster than one * per msec it maxes out at values just under 1,000,000. At constant * rates between one per msec, and one per second it will stabilize * to a value N*1000, where N is the rate of events per second. * At constant rates between one per second and one per 32 seconds, * it will be choppy, moving up on the seconds that have an event, * and then decaying until the next event. At rates slower than * about one in 32 seconds, it decays all the way back to zero between * each event. */ #define FM_COEF 933 /* coefficient for half-life of 10 secs */ |
d2b436580 cpuset: Replace a... |
1992 |
#define FM_MAXTICKS ((u32)99) /* useless computing more ticks than this */ |
3e0d98b9f [PATCH] cpuset: m... |
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 |
#define FM_MAXCNT 1000000 /* limit cnt to avoid overflow */ #define FM_SCALE 1000 /* faux fixed point scale */ /* Initialize a frequency meter */ static void fmeter_init(struct fmeter *fmp) { fmp->cnt = 0; fmp->val = 0; fmp->time = 0; spin_lock_init(&fmp->lock); } /* Internal meter update - process cnt events and update value */ static void fmeter_update(struct fmeter *fmp) { |
d2b436580 cpuset: Replace a... |
2008 2009 2010 2011 2012 |
time64_t now; u32 ticks; now = ktime_get_seconds(); ticks = now - fmp->time; |
3e0d98b9f [PATCH] cpuset: m... |
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 |
if (ticks == 0) return; ticks = min(FM_MAXTICKS, ticks); while (ticks-- > 0) fmp->val = (FM_COEF * fmp->val) / FM_SCALE; fmp->time = now; fmp->val += ((FM_SCALE - FM_COEF) * fmp->cnt) / FM_SCALE; fmp->cnt = 0; } /* Process any previous ticks, then bump cnt by one (times scale). */ static void fmeter_markevent(struct fmeter *fmp) { spin_lock(&fmp->lock); fmeter_update(fmp); fmp->cnt = min(FM_MAXCNT, fmp->cnt + FM_SCALE); spin_unlock(&fmp->lock); } /* Process any previous ticks, then return current value. */ static int fmeter_getrate(struct fmeter *fmp) { int val; spin_lock(&fmp->lock); fmeter_update(fmp); val = fmp->val; spin_unlock(&fmp->lock); return val; } |
57fce0a68 cpuset: don't use... |
2046 |
static struct cpuset *cpuset_attach_old_cs; |
5d21cc2db cpuset: replace c... |
2047 |
/* Called by cgroups to determine if a cpuset is usable; cpuset_mutex held */ |
1f7dd3e5a cgroup: fix handl... |
2048 |
static int cpuset_can_attach(struct cgroup_taskset *tset) |
f780bdb7c cgroups: add per-... |
2049 |
{ |
1f7dd3e5a cgroup: fix handl... |
2050 2051 |
struct cgroup_subsys_state *css; struct cpuset *cs; |
bb9d97b6d cgroup: don't use... |
2052 2053 |
struct task_struct *task; int ret; |
1da177e4c Linux-2.6.12-rc2 |
2054 |
|
57fce0a68 cpuset: don't use... |
2055 |
/* used later by cpuset_attach() */ |
1f7dd3e5a cgroup: fix handl... |
2056 2057 |
cpuset_attach_old_cs = task_cs(cgroup_taskset_first(tset, &css)); cs = css_cs(css); |
57fce0a68 cpuset: don't use... |
2058 |
|
18695a120 Revert "cgroup/cp... |
2059 |
mutex_lock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2060 |
|
aa6ec29be cgroup: remove sa... |
2061 |
/* allow moving tasks into an empty cpuset if on default hierarchy */ |
5d21cc2db cpuset: replace c... |
2062 |
ret = -ENOSPC; |
b8d1b8ee9 cpuset: Allow v2 ... |
2063 |
if (!is_in_v2_mode() && |
88fa523bf cpuset: allow to ... |
2064 |
(cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))) |
5d21cc2db cpuset: replace c... |
2065 |
goto out_unlock; |
9985b0bab sched: prevent bo... |
2066 |
|
1f7dd3e5a cgroup: fix handl... |
2067 |
cgroup_taskset_for_each(task, css, tset) { |
7f51412a4 sched/deadline: F... |
2068 2069 |
ret = task_can_attach(task, cs->cpus_allowed); if (ret) |
5d21cc2db cpuset: replace c... |
2070 2071 2072 2073 |
goto out_unlock; ret = security_task_setscheduler(task); if (ret) goto out_unlock; |
bb9d97b6d cgroup: don't use... |
2074 |
} |
f780bdb7c cgroups: add per-... |
2075 |
|
452477fa6 cpuset: pin down ... |
2076 2077 2078 2079 2080 |
/* * Mark attach is in progress. This makes validate_change() fail * changes which zero cpus/mems_allowed. */ cs->attach_in_progress++; |
5d21cc2db cpuset: replace c... |
2081 2082 |
ret = 0; out_unlock: |
18695a120 Revert "cgroup/cp... |
2083 |
mutex_unlock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2084 |
return ret; |
8793d854e Task Control Grou... |
2085 |
} |
f780bdb7c cgroups: add per-... |
2086 |
|
1f7dd3e5a cgroup: fix handl... |
2087 |
static void cpuset_cancel_attach(struct cgroup_taskset *tset) |
452477fa6 cpuset: pin down ... |
2088 |
{ |
1f7dd3e5a cgroup: fix handl... |
2089 |
struct cgroup_subsys_state *css; |
1f7dd3e5a cgroup: fix handl... |
2090 2091 |
cgroup_taskset_first(tset, &css); |
1f7dd3e5a cgroup: fix handl... |
2092 |
|
18695a120 Revert "cgroup/cp... |
2093 |
mutex_lock(&cpuset_mutex); |
eb95419b0 cgroup: pass arou... |
2094 |
css_cs(css)->attach_in_progress--; |
18695a120 Revert "cgroup/cp... |
2095 |
mutex_unlock(&cpuset_mutex); |
8793d854e Task Control Grou... |
2096 |
} |
1da177e4c Linux-2.6.12-rc2 |
2097 |
|
4e4c9a140 cpuset: cleanup c... |
2098 |
/* |
5d21cc2db cpuset: replace c... |
2099 |
* Protected by cpuset_mutex. cpus_attach is used only by cpuset_attach() |
4e4c9a140 cpuset: cleanup c... |
2100 2101 2102 2103 |
* but we can't allocate it dynamically there. Define it global and * allocate from cpuset_init(). */ static cpumask_var_t cpus_attach; |
1f7dd3e5a cgroup: fix handl... |
2104 |
static void cpuset_attach(struct cgroup_taskset *tset) |
8793d854e Task Control Grou... |
2105 |
{ |
67bd2c598 cpuset: remove un... |
2106 |
/* static buf protected by cpuset_mutex */ |
4e4c9a140 cpuset: cleanup c... |
2107 |
static nodemask_t cpuset_attach_nodemask_to; |
bb9d97b6d cgroup: don't use... |
2108 |
struct task_struct *task; |
4530eddb5 cgroup, memcg, cp... |
2109 |
struct task_struct *leader; |
1f7dd3e5a cgroup: fix handl... |
2110 2111 |
struct cgroup_subsys_state *css; struct cpuset *cs; |
57fce0a68 cpuset: don't use... |
2112 |
struct cpuset *oldcs = cpuset_attach_old_cs; |
22fb52dd7 [PATCH] SELinux: ... |
2113 |
|
1f7dd3e5a cgroup: fix handl... |
2114 2115 |
cgroup_taskset_first(tset, &css); cs = css_cs(css); |
18695a120 Revert "cgroup/cp... |
2116 |
mutex_lock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2117 |
|
4e4c9a140 cpuset: cleanup c... |
2118 2119 2120 2121 |
/* prepare for attach */ if (cs == &top_cpuset) cpumask_copy(cpus_attach, cpu_possible_mask); else |
ae1c80238 cpuset: apply cs-... |
2122 |
guarantee_online_cpus(cs, cpus_attach); |
4e4c9a140 cpuset: cleanup c... |
2123 |
|
ae1c80238 cpuset: apply cs-... |
2124 |
guarantee_online_mems(cs, &cpuset_attach_nodemask_to); |
4e4c9a140 cpuset: cleanup c... |
2125 |
|
1f7dd3e5a cgroup: fix handl... |
2126 |
cgroup_taskset_for_each(task, css, tset) { |
bb9d97b6d cgroup: don't use... |
2127 2128 2129 2130 |
/* * can_attach beforehand should guarantee that this doesn't * fail. TODO: have a better way to handle failure here */ |
4d1ac6a16 ANDROID: sched/cp... |
2131 |
WARN_ON_ONCE(update_cpus_allowed(cs, task, cpus_attach)); |
bb9d97b6d cgroup: don't use... |
2132 2133 2134 2135 |
cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to); cpuset_update_task_spread_flag(cs, task); } |
22fb52dd7 [PATCH] SELinux: ... |
2136 |
|
f780bdb7c cgroups: add per-... |
2137 |
/* |
4530eddb5 cgroup, memcg, cp... |
2138 2139 |
* Change mm for all threadgroup leaders. This is expensive and may * sleep and should be moved outside migration path proper. |
f780bdb7c cgroups: add per-... |
2140 |
*/ |
ae1c80238 cpuset: apply cs-... |
2141 |
cpuset_attach_nodemask_to = cs->effective_mems; |
1f7dd3e5a cgroup: fix handl... |
2142 |
cgroup_taskset_for_each_leader(leader, css, tset) { |
3df9ca0a2 cpuset: migrate m... |
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 |
struct mm_struct *mm = get_task_mm(leader); if (mm) { mpol_rebind_mm(mm, &cpuset_attach_nodemask_to); /* * old_mems_allowed is the same with mems_allowed * here, except if this task is being moved * automatically due to hotplug. In that case * @mems_allowed has been updated and is empty, so * @old_mems_allowed is the right nodesets that we * migrate mm from. */ |
e93ad19d0 cpuset: make mm m... |
2156 |
if (is_memory_migrate(cs)) |
3df9ca0a2 cpuset: migrate m... |
2157 2158 |
cpuset_migrate_mm(mm, &oldcs->old_mems_allowed, &cpuset_attach_nodemask_to); |
e93ad19d0 cpuset: make mm m... |
2159 2160 |
else mmput(mm); |
f047cecf2 cpuset: fix to mi... |
2161 |
} |
4225399a6 [PATCH] cpuset: r... |
2162 |
} |
452477fa6 cpuset: pin down ... |
2163 |
|
33ad801df cpuset: record ol... |
2164 |
cs->old_mems_allowed = cpuset_attach_nodemask_to; |
02bb58637 cpuset: schedule ... |
2165 |
|
452477fa6 cpuset: pin down ... |
2166 |
cs->attach_in_progress--; |
e44193d39 cpuset: let hotpl... |
2167 2168 |
if (!cs->attach_in_progress) wake_up(&cpuset_attach_wq); |
5d21cc2db cpuset: replace c... |
2169 |
|
18695a120 Revert "cgroup/cp... |
2170 |
mutex_unlock(&cpuset_mutex); |
1da177e4c Linux-2.6.12-rc2 |
2171 2172 2173 2174 2175 |
} /* The various types of files and directories in a cpuset file system */ typedef enum { |
45b07ef31 [PATCH] cpusets: ... |
2176 |
FILE_MEMORY_MIGRATE, |
1da177e4c Linux-2.6.12-rc2 |
2177 2178 |
FILE_CPULIST, FILE_MEMLIST, |
afd1a8b3e cpuset: export ef... |
2179 2180 |
FILE_EFFECTIVE_CPULIST, FILE_EFFECTIVE_MEMLIST, |
5cf8114d6 cpuset: Expose cp... |
2181 |
FILE_SUBPARTS_CPULIST, |
1da177e4c Linux-2.6.12-rc2 |
2182 2183 |
FILE_CPU_EXCLUSIVE, FILE_MEM_EXCLUSIVE, |
786083667 Cpuset hardwall f... |
2184 |
FILE_MEM_HARDWALL, |
029190c51 cpuset sched_load... |
2185 |
FILE_SCHED_LOAD_BALANCE, |
ee8dde0cd cpuset: Add new v... |
2186 |
FILE_PARTITION_ROOT, |
1d3504fcf sched, cpuset: cu... |
2187 |
FILE_SCHED_RELAX_DOMAIN_LEVEL, |
3e0d98b9f [PATCH] cpuset: m... |
2188 2189 |
FILE_MEMORY_PRESSURE_ENABLED, FILE_MEMORY_PRESSURE, |
825a46af5 [PATCH] cpuset me... |
2190 2191 |
FILE_SPREAD_PAGE, FILE_SPREAD_SLAB, |
1da177e4c Linux-2.6.12-rc2 |
2192 |
} cpuset_filetype_t; |
182446d08 cgroup: pass arou... |
2193 2194 |
static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft, u64 val) |
700fe1ab9 CGroup API files:... |
2195 |
{ |
182446d08 cgroup: pass arou... |
2196 |
struct cpuset *cs = css_cs(css); |
700fe1ab9 CGroup API files:... |
2197 |
cpuset_filetype_t type = cft->private; |
a903f0865 cpuset: fix the r... |
2198 |
int retval = 0; |
700fe1ab9 CGroup API files:... |
2199 |
|
d74b27d63 cgroup/cpuset: Ch... |
2200 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2201 |
mutex_lock(&cpuset_mutex); |
a903f0865 cpuset: fix the r... |
2202 2203 |
if (!is_cpuset_online(cs)) { retval = -ENODEV; |
5d21cc2db cpuset: replace c... |
2204 |
goto out_unlock; |
a903f0865 cpuset: fix the r... |
2205 |
} |
700fe1ab9 CGroup API files:... |
2206 2207 |
switch (type) { |
1da177e4c Linux-2.6.12-rc2 |
2208 |
case FILE_CPU_EXCLUSIVE: |
700fe1ab9 CGroup API files:... |
2209 |
retval = update_flag(CS_CPU_EXCLUSIVE, cs, val); |
1da177e4c Linux-2.6.12-rc2 |
2210 2211 |
break; case FILE_MEM_EXCLUSIVE: |
700fe1ab9 CGroup API files:... |
2212 |
retval = update_flag(CS_MEM_EXCLUSIVE, cs, val); |
1da177e4c Linux-2.6.12-rc2 |
2213 |
break; |
786083667 Cpuset hardwall f... |
2214 2215 2216 |
case FILE_MEM_HARDWALL: retval = update_flag(CS_MEM_HARDWALL, cs, val); break; |
029190c51 cpuset sched_load... |
2217 |
case FILE_SCHED_LOAD_BALANCE: |
700fe1ab9 CGroup API files:... |
2218 |
retval = update_flag(CS_SCHED_LOAD_BALANCE, cs, val); |
1d3504fcf sched, cpuset: cu... |
2219 |
break; |
45b07ef31 [PATCH] cpusets: ... |
2220 |
case FILE_MEMORY_MIGRATE: |
700fe1ab9 CGroup API files:... |
2221 |
retval = update_flag(CS_MEMORY_MIGRATE, cs, val); |
45b07ef31 [PATCH] cpusets: ... |
2222 |
break; |
3e0d98b9f [PATCH] cpuset: m... |
2223 |
case FILE_MEMORY_PRESSURE_ENABLED: |
700fe1ab9 CGroup API files:... |
2224 |
cpuset_memory_pressure_enabled = !!val; |
3e0d98b9f [PATCH] cpuset: m... |
2225 |
break; |
825a46af5 [PATCH] cpuset me... |
2226 |
case FILE_SPREAD_PAGE: |
700fe1ab9 CGroup API files:... |
2227 |
retval = update_flag(CS_SPREAD_PAGE, cs, val); |
825a46af5 [PATCH] cpuset me... |
2228 2229 |
break; case FILE_SPREAD_SLAB: |
700fe1ab9 CGroup API files:... |
2230 |
retval = update_flag(CS_SPREAD_SLAB, cs, val); |
825a46af5 [PATCH] cpuset me... |
2231 |
break; |
1da177e4c Linux-2.6.12-rc2 |
2232 2233 |
default: retval = -EINVAL; |
700fe1ab9 CGroup API files:... |
2234 |
break; |
1da177e4c Linux-2.6.12-rc2 |
2235 |
} |
5d21cc2db cpuset: replace c... |
2236 |
out_unlock: |
18695a120 Revert "cgroup/cp... |
2237 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2238 |
put_online_cpus(); |
1da177e4c Linux-2.6.12-rc2 |
2239 2240 |
return retval; } |
182446d08 cgroup: pass arou... |
2241 2242 |
static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft, s64 val) |
5be7a4792 Fix cpuset sched_... |
2243 |
{ |
182446d08 cgroup: pass arou... |
2244 |
struct cpuset *cs = css_cs(css); |
5be7a4792 Fix cpuset sched_... |
2245 |
cpuset_filetype_t type = cft->private; |
5d21cc2db cpuset: replace c... |
2246 |
int retval = -ENODEV; |
5be7a4792 Fix cpuset sched_... |
2247 |
|
d74b27d63 cgroup/cpuset: Ch... |
2248 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2249 |
mutex_lock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2250 2251 |
if (!is_cpuset_online(cs)) goto out_unlock; |
e37123953 cgroup files: rem... |
2252 |
|
5be7a4792 Fix cpuset sched_... |
2253 2254 2255 2256 2257 2258 2259 2260 |
switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: retval = update_relax_domain_level(cs, val); break; default: retval = -EINVAL; break; } |
5d21cc2db cpuset: replace c... |
2261 |
out_unlock: |
18695a120 Revert "cgroup/cp... |
2262 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2263 |
put_online_cpus(); |
5be7a4792 Fix cpuset sched_... |
2264 2265 |
return retval; } |
1da177e4c Linux-2.6.12-rc2 |
2266 |
/* |
e37123953 cgroup files: rem... |
2267 2268 |
* Common handling for a write to a "cpus" or "mems" file. */ |
451af504d cgroup: replace c... |
2269 2270 |
static ssize_t cpuset_write_resmask(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
e37123953 cgroup files: rem... |
2271 |
{ |
451af504d cgroup: replace c... |
2272 |
struct cpuset *cs = css_cs(of_css(of)); |
645fcc9d2 cpuset: don't all... |
2273 |
struct cpuset *trialcs; |
5d21cc2db cpuset: replace c... |
2274 |
int retval = -ENODEV; |
e37123953 cgroup files: rem... |
2275 |
|
451af504d cgroup: replace c... |
2276 |
buf = strstrip(buf); |
3a5a6d0c2 cpuset: don't nes... |
2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 |
/* * CPU or memory hotunplug may leave @cs w/o any execution * resources, in which case the hotplug code asynchronously updates * configuration and transfers all tasks to the nearest ancestor * which can execute. * * As writes to "cpus" or "mems" may restore @cs's execution * resources, wait for the previously scheduled operations before * proceeding, so that we don't end up keep removing tasks added * after execution capability is restored. |
76bb5ab8f cpuset: break ker... |
2287 2288 2289 2290 2291 2292 2293 2294 |
* * cpuset_hotplug_work calls back into cgroup core via * cgroup_transfer_tasks() and waiting for it from a cgroupfs * operation like this one can lead to a deadlock through kernfs * active_ref protection. Let's break the protection. Losing the * protection is okay as we check whether @cs is online after * grabbing cpuset_mutex anyway. This only happens on the legacy * hierarchies. |
3a5a6d0c2 cpuset: don't nes... |
2295 |
*/ |
76bb5ab8f cpuset: break ker... |
2296 2297 |
css_get(&cs->css); kernfs_break_active_protection(of->kn); |
3a5a6d0c2 cpuset: don't nes... |
2298 |
flush_work(&cpuset_hotplug_work); |
d74b27d63 cgroup/cpuset: Ch... |
2299 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2300 |
mutex_lock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2301 2302 |
if (!is_cpuset_online(cs)) goto out_unlock; |
e37123953 cgroup files: rem... |
2303 |
|
645fcc9d2 cpuset: don't all... |
2304 |
trialcs = alloc_trial_cpuset(cs); |
b75f38d65 cpuset: add a mis... |
2305 2306 |
if (!trialcs) { retval = -ENOMEM; |
5d21cc2db cpuset: replace c... |
2307 |
goto out_unlock; |
b75f38d65 cpuset: add a mis... |
2308 |
} |
645fcc9d2 cpuset: don't all... |
2309 |
|
451af504d cgroup: replace c... |
2310 |
switch (of_cft(of)->private) { |
e37123953 cgroup files: rem... |
2311 |
case FILE_CPULIST: |
645fcc9d2 cpuset: don't all... |
2312 |
retval = update_cpumask(cs, trialcs, buf); |
e37123953 cgroup files: rem... |
2313 2314 |
break; case FILE_MEMLIST: |
645fcc9d2 cpuset: don't all... |
2315 |
retval = update_nodemask(cs, trialcs, buf); |
e37123953 cgroup files: rem... |
2316 2317 2318 2319 2320 |
break; default: retval = -EINVAL; break; } |
645fcc9d2 cpuset: don't all... |
2321 |
|
bf92370c0 cpuset: Simply al... |
2322 |
free_cpuset(trialcs); |
5d21cc2db cpuset: replace c... |
2323 |
out_unlock: |
18695a120 Revert "cgroup/cp... |
2324 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2325 |
put_online_cpus(); |
76bb5ab8f cpuset: break ker... |
2326 2327 |
kernfs_unbreak_active_protection(of->kn); css_put(&cs->css); |
e93ad19d0 cpuset: make mm m... |
2328 |
flush_workqueue(cpuset_migrate_mm_wq); |
451af504d cgroup: replace c... |
2329 |
return retval ?: nbytes; |
e37123953 cgroup files: rem... |
2330 2331 2332 |
} /* |
1da177e4c Linux-2.6.12-rc2 |
2333 2334 2335 2336 2337 2338 |
* These ascii lists should be read in a single call, by using a user * buffer large enough to hold the entire map. If read in smaller * chunks, there is no guarantee of atomicity. Since the display format * used, list of ranges of sequential numbers, is variable length, * and since these maps can change value dynamically, one could read * gibberish by doing partial reads while a list was changing. |
1da177e4c Linux-2.6.12-rc2 |
2339 |
*/ |
2da8ca822 cgroup: replace c... |
2340 |
static int cpuset_common_seq_show(struct seq_file *sf, void *v) |
1da177e4c Linux-2.6.12-rc2 |
2341 |
{ |
2da8ca822 cgroup: replace c... |
2342 2343 |
struct cpuset *cs = css_cs(seq_css(sf)); cpuset_filetype_t type = seq_cft(sf)->private; |
51ffe4117 cpuset: convert a... |
2344 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
2345 |
|
8447a0fee cpuset: convert c... |
2346 |
spin_lock_irq(&callback_lock); |
1da177e4c Linux-2.6.12-rc2 |
2347 2348 2349 |
switch (type) { case FILE_CPULIST: |
c8dc4422c ANDROID: cpuset: ... |
2350 2351 |
seq_printf(sf, "%*pbl ", cpumask_pr_args(cs->cpus_requested)); |
1da177e4c Linux-2.6.12-rc2 |
2352 2353 |
break; case FILE_MEMLIST: |
e8e6d97c9 cpuset: use %*pb[... |
2354 2355 |
seq_printf(sf, "%*pbl ", nodemask_pr_args(&cs->mems_allowed)); |
1da177e4c Linux-2.6.12-rc2 |
2356 |
break; |
afd1a8b3e cpuset: export ef... |
2357 |
case FILE_EFFECTIVE_CPULIST: |
e8e6d97c9 cpuset: use %*pb[... |
2358 2359 |
seq_printf(sf, "%*pbl ", cpumask_pr_args(cs->effective_cpus)); |
afd1a8b3e cpuset: export ef... |
2360 2361 |
break; case FILE_EFFECTIVE_MEMLIST: |
e8e6d97c9 cpuset: use %*pb[... |
2362 2363 |
seq_printf(sf, "%*pbl ", nodemask_pr_args(&cs->effective_mems)); |
afd1a8b3e cpuset: export ef... |
2364 |
break; |
5cf8114d6 cpuset: Expose cp... |
2365 2366 2367 2368 |
case FILE_SUBPARTS_CPULIST: seq_printf(sf, "%*pbl ", cpumask_pr_args(cs->subparts_cpus)); break; |
1da177e4c Linux-2.6.12-rc2 |
2369 |
default: |
51ffe4117 cpuset: convert a... |
2370 |
ret = -EINVAL; |
1da177e4c Linux-2.6.12-rc2 |
2371 |
} |
1da177e4c Linux-2.6.12-rc2 |
2372 |
|
8447a0fee cpuset: convert c... |
2373 |
spin_unlock_irq(&callback_lock); |
51ffe4117 cpuset: convert a... |
2374 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
2375 |
} |
182446d08 cgroup: pass arou... |
2376 |
static u64 cpuset_read_u64(struct cgroup_subsys_state *css, struct cftype *cft) |
700fe1ab9 CGroup API files:... |
2377 |
{ |
182446d08 cgroup: pass arou... |
2378 |
struct cpuset *cs = css_cs(css); |
700fe1ab9 CGroup API files:... |
2379 2380 2381 2382 2383 2384 |
cpuset_filetype_t type = cft->private; switch (type) { case FILE_CPU_EXCLUSIVE: return is_cpu_exclusive(cs); case FILE_MEM_EXCLUSIVE: return is_mem_exclusive(cs); |
786083667 Cpuset hardwall f... |
2385 2386 |
case FILE_MEM_HARDWALL: return is_mem_hardwall(cs); |
700fe1ab9 CGroup API files:... |
2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 |
case FILE_SCHED_LOAD_BALANCE: return is_sched_load_balance(cs); case FILE_MEMORY_MIGRATE: return is_memory_migrate(cs); case FILE_MEMORY_PRESSURE_ENABLED: return cpuset_memory_pressure_enabled; case FILE_MEMORY_PRESSURE: return fmeter_getrate(&cs->fmeter); case FILE_SPREAD_PAGE: return is_spread_page(cs); case FILE_SPREAD_SLAB: return is_spread_slab(cs); default: BUG(); } |
cf417141c sched, cpuset: re... |
2402 2403 2404 |
/* Unreachable but makes gcc happy */ return 0; |
700fe1ab9 CGroup API files:... |
2405 |
} |
1da177e4c Linux-2.6.12-rc2 |
2406 |
|
182446d08 cgroup: pass arou... |
2407 |
static s64 cpuset_read_s64(struct cgroup_subsys_state *css, struct cftype *cft) |
5be7a4792 Fix cpuset sched_... |
2408 |
{ |
182446d08 cgroup: pass arou... |
2409 |
struct cpuset *cs = css_cs(css); |
5be7a4792 Fix cpuset sched_... |
2410 2411 2412 2413 2414 2415 2416 |
cpuset_filetype_t type = cft->private; switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: return cs->relax_domain_level; default: BUG(); } |
cf417141c sched, cpuset: re... |
2417 2418 2419 |
/* Unrechable but makes gcc happy */ return 0; |
5be7a4792 Fix cpuset sched_... |
2420 |
} |
bb5b553c3 cpuset: Use descr... |
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 |
static int sched_partition_show(struct seq_file *seq, void *v) { struct cpuset *cs = css_cs(seq_css(seq)); switch (cs->partition_root_state) { case PRS_ENABLED: seq_puts(seq, "root "); break; case PRS_DISABLED: seq_puts(seq, "member "); break; case PRS_ERROR: seq_puts(seq, "root invalid "); break; } return 0; } static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct cpuset *cs = css_cs(of_css(of)); int val; int retval = -ENODEV; buf = strstrip(buf); /* |
b1e3aeb11 cpuset: Minor cgr... |
2452 |
* Convert "root" to ENABLED, and convert "member" to DISABLED. |
bb5b553c3 cpuset: Use descr... |
2453 |
*/ |
b1e3aeb11 cpuset: Minor cgr... |
2454 |
if (!strcmp(buf, "root")) |
bb5b553c3 cpuset: Use descr... |
2455 |
val = PRS_ENABLED; |
b1e3aeb11 cpuset: Minor cgr... |
2456 |
else if (!strcmp(buf, "member")) |
bb5b553c3 cpuset: Use descr... |
2457 2458 2459 2460 2461 |
val = PRS_DISABLED; else return -EINVAL; css_get(&cs->css); |
d74b27d63 cgroup/cpuset: Ch... |
2462 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2463 |
mutex_lock(&cpuset_mutex); |
bb5b553c3 cpuset: Use descr... |
2464 2465 2466 2467 2468 |
if (!is_cpuset_online(cs)) goto out_unlock; retval = update_prstate(cs, val); out_unlock: |
18695a120 Revert "cgroup/cp... |
2469 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2470 |
put_online_cpus(); |
bb5b553c3 cpuset: Use descr... |
2471 2472 2473 |
css_put(&cs->css); return retval ?: nbytes; } |
1da177e4c Linux-2.6.12-rc2 |
2474 2475 2476 2477 |
/* * for the common functions, 'private' gives the type of file */ |
4ec22e9c5 cpuset: Enable cp... |
2478 |
static struct cftype legacy_files[] = { |
addf2c739 Cpuset hardwall f... |
2479 2480 |
{ .name = "cpus", |
2da8ca822 cgroup: replace c... |
2481 |
.seq_show = cpuset_common_seq_show, |
451af504d cgroup: replace c... |
2482 |
.write = cpuset_write_resmask, |
e37123953 cgroup files: rem... |
2483 |
.max_write_len = (100U + 6 * NR_CPUS), |
addf2c739 Cpuset hardwall f... |
2484 2485 2486 2487 2488 |
.private = FILE_CPULIST, }, { .name = "mems", |
2da8ca822 cgroup: replace c... |
2489 |
.seq_show = cpuset_common_seq_show, |
451af504d cgroup: replace c... |
2490 |
.write = cpuset_write_resmask, |
e37123953 cgroup files: rem... |
2491 |
.max_write_len = (100U + 6 * MAX_NUMNODES), |
addf2c739 Cpuset hardwall f... |
2492 2493 2494 2495 |
.private = FILE_MEMLIST, }, { |
afd1a8b3e cpuset: export ef... |
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 |
.name = "effective_cpus", .seq_show = cpuset_common_seq_show, .private = FILE_EFFECTIVE_CPULIST, }, { .name = "effective_mems", .seq_show = cpuset_common_seq_show, .private = FILE_EFFECTIVE_MEMLIST, }, { |
addf2c739 Cpuset hardwall f... |
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 |
.name = "cpu_exclusive", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_CPU_EXCLUSIVE, }, { .name = "mem_exclusive", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_MEM_EXCLUSIVE, }, { |
786083667 Cpuset hardwall f... |
2522 2523 2524 2525 2526 2527 2528 |
.name = "mem_hardwall", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_MEM_HARDWALL, }, { |
addf2c739 Cpuset hardwall f... |
2529 2530 2531 2532 2533 2534 2535 2536 |
.name = "sched_load_balance", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_SCHED_LOAD_BALANCE, }, { .name = "sched_relax_domain_level", |
5be7a4792 Fix cpuset sched_... |
2537 2538 |
.read_s64 = cpuset_read_s64, .write_s64 = cpuset_write_s64, |
addf2c739 Cpuset hardwall f... |
2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 |
.private = FILE_SCHED_RELAX_DOMAIN_LEVEL, }, { .name = "memory_migrate", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_MEMORY_MIGRATE, }, { .name = "memory_pressure", .read_u64 = cpuset_read_u64, |
1c08c22c8 cpuset: Fix incor... |
2552 |
.private = FILE_MEMORY_PRESSURE, |
addf2c739 Cpuset hardwall f... |
2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 |
}, { .name = "memory_spread_page", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_SPREAD_PAGE, }, { .name = "memory_spread_slab", .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_SPREAD_SLAB, }, |
3e0d98b9f [PATCH] cpuset: m... |
2568 |
|
4baf6e332 cgroup: convert a... |
2569 2570 2571 2572 2573 2574 2575 |
{ .name = "memory_pressure_enabled", .flags = CFTYPE_ONLY_ON_ROOT, .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_MEMORY_PRESSURE_ENABLED, }, |
1da177e4c Linux-2.6.12-rc2 |
2576 |
|
4baf6e332 cgroup: convert a... |
2577 2578 |
{ } /* terminate */ }; |
1da177e4c Linux-2.6.12-rc2 |
2579 2580 |
/* |
4ec22e9c5 cpuset: Enable cp... |
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 |
* This is currently a minimal set for the default hierarchy. It can be * expanded later on by migrating more features and control files from v1. */ static struct cftype dfl_files[] = { { .name = "cpus", .seq_show = cpuset_common_seq_show, .write = cpuset_write_resmask, .max_write_len = (100U + 6 * NR_CPUS), .private = FILE_CPULIST, .flags = CFTYPE_NOT_ON_ROOT, }, { .name = "mems", .seq_show = cpuset_common_seq_show, .write = cpuset_write_resmask, .max_write_len = (100U + 6 * MAX_NUMNODES), .private = FILE_MEMLIST, .flags = CFTYPE_NOT_ON_ROOT, }, { .name = "cpus.effective", .seq_show = cpuset_common_seq_show, .private = FILE_EFFECTIVE_CPULIST, |
4ec22e9c5 cpuset: Enable cp... |
2607 2608 2609 2610 2611 2612 |
}, { .name = "mems.effective", .seq_show = cpuset_common_seq_show, .private = FILE_EFFECTIVE_MEMLIST, |
4ec22e9c5 cpuset: Enable cp... |
2613 |
}, |
ee8dde0cd cpuset: Add new v... |
2614 |
{ |
b1e3aeb11 cpuset: Minor cgr... |
2615 |
.name = "cpus.partition", |
bb5b553c3 cpuset: Use descr... |
2616 2617 |
.seq_show = sched_partition_show, .write = sched_partition_write, |
ee8dde0cd cpuset: Add new v... |
2618 2619 2620 |
.private = FILE_PARTITION_ROOT, .flags = CFTYPE_NOT_ON_ROOT, }, |
5cf8114d6 cpuset: Expose cp... |
2621 2622 2623 2624 2625 2626 |
{ .name = "cpus.subpartitions", .seq_show = cpuset_common_seq_show, .private = FILE_SUBPARTS_CPULIST, .flags = CFTYPE_DEBUG, }, |
4ec22e9c5 cpuset: Enable cp... |
2627 2628 2629 2630 2631 |
{ } /* terminate */ }; /* |
92fb97487 cgroup: rename ->... |
2632 |
* cpuset_css_alloc - allocate a cpuset css |
c9e5fe66f cpuset: rename @c... |
2633 |
* cgrp: control group that the new cpuset will be part of |
1da177e4c Linux-2.6.12-rc2 |
2634 |
*/ |
eb95419b0 cgroup: pass arou... |
2635 2636 |
static struct cgroup_subsys_state * cpuset_css_alloc(struct cgroup_subsys_state *parent_css) |
1da177e4c Linux-2.6.12-rc2 |
2637 |
{ |
c8f699bb5 cpuset: introduce... |
2638 |
struct cpuset *cs; |
1da177e4c Linux-2.6.12-rc2 |
2639 |
|
eb95419b0 cgroup: pass arou... |
2640 |
if (!parent_css) |
8793d854e Task Control Grou... |
2641 |
return &top_cpuset.css; |
033fa1c5f cgroup, cpuset: r... |
2642 |
|
c8f699bb5 cpuset: introduce... |
2643 |
cs = kzalloc(sizeof(*cs), GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
2644 |
if (!cs) |
8793d854e Task Control Grou... |
2645 |
return ERR_PTR(-ENOMEM); |
bf92370c0 cpuset: Simply al... |
2646 2647 2648 2649 2650 |
if (alloc_cpumasks(cs, NULL)) { kfree(cs); return ERR_PTR(-ENOMEM); } |
1da177e4c Linux-2.6.12-rc2 |
2651 |
|
029190c51 cpuset sched_load... |
2652 |
set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); |
f9a86fcbb cpuset: modify cp... |
2653 |
nodes_clear(cs->mems_allowed); |
e2b9a3d7d cpuset: add cs->e... |
2654 |
nodes_clear(cs->effective_mems); |
3e0d98b9f [PATCH] cpuset: m... |
2655 |
fmeter_init(&cs->fmeter); |
1d3504fcf sched, cpuset: cu... |
2656 |
cs->relax_domain_level = -1; |
1da177e4c Linux-2.6.12-rc2 |
2657 |
|
c8f699bb5 cpuset: introduce... |
2658 2659 |
return &cs->css; } |
eb95419b0 cgroup: pass arou... |
2660 |
static int cpuset_css_online(struct cgroup_subsys_state *css) |
c8f699bb5 cpuset: introduce... |
2661 |
{ |
eb95419b0 cgroup: pass arou... |
2662 |
struct cpuset *cs = css_cs(css); |
c431069fe cpuset: remove cp... |
2663 |
struct cpuset *parent = parent_cs(cs); |
ae8086ce1 cpuset: introduce... |
2664 |
struct cpuset *tmp_cs; |
492eb21b9 cgroup: make hier... |
2665 |
struct cgroup_subsys_state *pos_css; |
c8f699bb5 cpuset: introduce... |
2666 2667 2668 |
if (!parent) return 0; |
d74b27d63 cgroup/cpuset: Ch... |
2669 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2670 |
mutex_lock(&cpuset_mutex); |
5d21cc2db cpuset: replace c... |
2671 |
|
efeb77b2f cpuset: introduce... |
2672 |
set_bit(CS_ONLINE, &cs->flags); |
c8f699bb5 cpuset: introduce... |
2673 2674 2675 2676 |
if (is_spread_page(parent)) set_bit(CS_SPREAD_PAGE, &cs->flags); if (is_spread_slab(parent)) set_bit(CS_SPREAD_SLAB, &cs->flags); |
1da177e4c Linux-2.6.12-rc2 |
2677 |
|
664eeddee mm: page_alloc: u... |
2678 |
cpuset_inc(); |
033fa1c5f cgroup, cpuset: r... |
2679 |
|
8447a0fee cpuset: convert c... |
2680 |
spin_lock_irq(&callback_lock); |
b8d1b8ee9 cpuset: Allow v2 ... |
2681 |
if (is_in_v2_mode()) { |
e2b9a3d7d cpuset: add cs->e... |
2682 2683 |
cpumask_copy(cs->effective_cpus, parent->effective_cpus); cs->effective_mems = parent->effective_mems; |
4716909cc cpuset: Track cpu... |
2684 2685 |
cs->use_parent_ecpus = true; parent->child_ecpus_count++; |
e2b9a3d7d cpuset: add cs->e... |
2686 |
} |
8447a0fee cpuset: convert c... |
2687 |
spin_unlock_irq(&callback_lock); |
e2b9a3d7d cpuset: add cs->e... |
2688 |
|
eb95419b0 cgroup: pass arou... |
2689 |
if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags)) |
5d21cc2db cpuset: replace c... |
2690 |
goto out_unlock; |
033fa1c5f cgroup, cpuset: r... |
2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 |
/* * Clone @parent's configuration if CGRP_CPUSET_CLONE_CHILDREN is * set. This flag handling is implemented in cgroup core for * histrical reasons - the flag may be specified during mount. * * Currently, if any sibling cpusets have exclusive cpus or mem, we * refuse to clone the configuration - thereby refusing the task to * be entered, and as a result refusing the sys_unshare() or * clone() which initiated it. If this becomes a problem for some * users who wish to allow that scenario, then this could be * changed to grant parent->cpus_allowed-sibling_cpus_exclusive * (and likewise for mems) to the new cgroup. */ |
ae8086ce1 cpuset: introduce... |
2705 |
rcu_read_lock(); |
492eb21b9 cgroup: make hier... |
2706 |
cpuset_for_each_child(tmp_cs, pos_css, parent) { |
ae8086ce1 cpuset: introduce... |
2707 2708 |
if (is_mem_exclusive(tmp_cs) || is_cpu_exclusive(tmp_cs)) { rcu_read_unlock(); |
5d21cc2db cpuset: replace c... |
2709 |
goto out_unlock; |
ae8086ce1 cpuset: introduce... |
2710 |
} |
033fa1c5f cgroup, cpuset: r... |
2711 |
} |
ae8086ce1 cpuset: introduce... |
2712 |
rcu_read_unlock(); |
033fa1c5f cgroup, cpuset: r... |
2713 |
|
8447a0fee cpuset: convert c... |
2714 |
spin_lock_irq(&callback_lock); |
033fa1c5f cgroup, cpuset: r... |
2715 |
cs->mems_allowed = parent->mems_allowed; |
790317e1b cpuset: initializ... |
2716 |
cs->effective_mems = parent->mems_allowed; |
033fa1c5f cgroup, cpuset: r... |
2717 |
cpumask_copy(cs->cpus_allowed, parent->cpus_allowed); |
c8dc4422c ANDROID: cpuset: ... |
2718 |
cpumask_copy(cs->cpus_requested, parent->cpus_requested); |
790317e1b cpuset: initializ... |
2719 |
cpumask_copy(cs->effective_cpus, parent->cpus_allowed); |
cea74465e cpuset: lock vs u... |
2720 |
spin_unlock_irq(&callback_lock); |
5d21cc2db cpuset: replace c... |
2721 |
out_unlock: |
18695a120 Revert "cgroup/cp... |
2722 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2723 |
put_online_cpus(); |
c8f699bb5 cpuset: introduce... |
2724 2725 |
return 0; } |
0b9e6965a cpuset: relocate ... |
2726 2727 2728 |
/* * If the cpuset being removed has its flag 'sched_load_balance' * enabled, then simulate turning sched_load_balance off, which |
ee8dde0cd cpuset: Add new v... |
2729 2730 2731 2732 2733 2734 |
* will call rebuild_sched_domains_locked(). That is not needed * in the default hierarchy where only changes in partition * will cause repartitioning. * * If the cpuset has the 'sched.partition' flag enabled, simulate * turning 'sched.partition" off. |
0b9e6965a cpuset: relocate ... |
2735 |
*/ |
eb95419b0 cgroup: pass arou... |
2736 |
static void cpuset_css_offline(struct cgroup_subsys_state *css) |
c8f699bb5 cpuset: introduce... |
2737 |
{ |
eb95419b0 cgroup: pass arou... |
2738 |
struct cpuset *cs = css_cs(css); |
c8f699bb5 cpuset: introduce... |
2739 |
|
d74b27d63 cgroup/cpuset: Ch... |
2740 |
get_online_cpus(); |
18695a120 Revert "cgroup/cp... |
2741 |
mutex_lock(&cpuset_mutex); |
c8f699bb5 cpuset: introduce... |
2742 |
|
ee8dde0cd cpuset: Add new v... |
2743 2744 2745 2746 2747 |
if (is_partition_root(cs)) update_prstate(cs, 0); if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) && is_sched_load_balance(cs)) |
c8f699bb5 cpuset: introduce... |
2748 |
update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); |
4716909cc cpuset: Track cpu... |
2749 2750 2751 2752 2753 2754 |
if (cs->use_parent_ecpus) { struct cpuset *parent = parent_cs(cs); cs->use_parent_ecpus = false; parent->child_ecpus_count--; } |
664eeddee mm: page_alloc: u... |
2755 |
cpuset_dec(); |
efeb77b2f cpuset: introduce... |
2756 |
clear_bit(CS_ONLINE, &cs->flags); |
c8f699bb5 cpuset: introduce... |
2757 |
|
18695a120 Revert "cgroup/cp... |
2758 |
mutex_unlock(&cpuset_mutex); |
d74b27d63 cgroup/cpuset: Ch... |
2759 |
put_online_cpus(); |
1da177e4c Linux-2.6.12-rc2 |
2760 |
} |
eb95419b0 cgroup: pass arou... |
2761 |
static void cpuset_css_free(struct cgroup_subsys_state *css) |
1da177e4c Linux-2.6.12-rc2 |
2762 |
{ |
eb95419b0 cgroup: pass arou... |
2763 |
struct cpuset *cs = css_cs(css); |
1da177e4c Linux-2.6.12-rc2 |
2764 |
|
bf92370c0 cpuset: Simply al... |
2765 |
free_cpuset(cs); |
1da177e4c Linux-2.6.12-rc2 |
2766 |
} |
39bd0d15e cpuset: initializ... |
2767 2768 |
static void cpuset_bind(struct cgroup_subsys_state *root_css) { |
18695a120 Revert "cgroup/cp... |
2769 |
mutex_lock(&cpuset_mutex); |
8447a0fee cpuset: convert c... |
2770 |
spin_lock_irq(&callback_lock); |
39bd0d15e cpuset: initializ... |
2771 |
|
b8d1b8ee9 cpuset: Allow v2 ... |
2772 |
if (is_in_v2_mode()) { |
39bd0d15e cpuset: initializ... |
2773 2774 2775 2776 2777 2778 2779 |
cpumask_copy(top_cpuset.cpus_allowed, cpu_possible_mask); top_cpuset.mems_allowed = node_possible_map; } else { cpumask_copy(top_cpuset.cpus_allowed, top_cpuset.effective_cpus); top_cpuset.mems_allowed = top_cpuset.effective_mems; } |
8447a0fee cpuset: convert c... |
2780 |
spin_unlock_irq(&callback_lock); |
18695a120 Revert "cgroup/cp... |
2781 |
mutex_unlock(&cpuset_mutex); |
39bd0d15e cpuset: initializ... |
2782 |
} |
06f4e9489 cpuset: make sure... |
2783 2784 2785 2786 2787 |
/* * Make sure the new task conform to the current state of its parent, * which could have been changed by cpuset just after it inherits the * state from the parent and before it sits on the cgroup's task list. */ |
8a15b8174 cpuset: fix non s... |
2788 |
static void cpuset_fork(struct task_struct *task) |
06f4e9489 cpuset: make sure... |
2789 2790 2791 |
{ if (task_css_is_root(task, cpuset_cgrp_id)) return; |
3bd370625 sched/core: Provi... |
2792 |
set_cpus_allowed_ptr(task, current->cpus_ptr); |
06f4e9489 cpuset: make sure... |
2793 2794 |
task->mems_allowed = current->mems_allowed; } |
073219e99 cgroup: clean up ... |
2795 |
struct cgroup_subsys cpuset_cgrp_subsys = { |
39bd0d15e cpuset: initializ... |
2796 2797 2798 2799 2800 2801 2802 |
.css_alloc = cpuset_css_alloc, .css_online = cpuset_css_online, .css_offline = cpuset_css_offline, .css_free = cpuset_css_free, .can_attach = cpuset_can_attach, .cancel_attach = cpuset_cancel_attach, .attach = cpuset_attach, |
5cf1cacb4 cgroup, cpuset: r... |
2803 |
.post_attach = cpuset_post_attach, |
39bd0d15e cpuset: initializ... |
2804 |
.bind = cpuset_bind, |
06f4e9489 cpuset: make sure... |
2805 |
.fork = cpuset_fork, |
4ec22e9c5 cpuset: Enable cp... |
2806 2807 |
.legacy_cftypes = legacy_files, .dfl_cftypes = dfl_files, |
b38e42e96 cgroup: convert c... |
2808 |
.early_init = true, |
4ec22e9c5 cpuset: Enable cp... |
2809 |
.threaded = true, |
8793d854e Task Control Grou... |
2810 |
}; |
1da177e4c Linux-2.6.12-rc2 |
2811 2812 2813 |
/** * cpuset_init - initialize cpusets at system boot * |
d5f68d330 cpuset: move moun... |
2814 |
* Description: Initialize top_cpuset |
1da177e4c Linux-2.6.12-rc2 |
2815 2816 2817 2818 |
**/ int __init cpuset_init(void) { |
75fa8e5d3 cgroup: switch to... |
2819 2820 |
BUG_ON(!alloc_cpumask_var(&top_cpuset.cpus_allowed, GFP_KERNEL)); BUG_ON(!alloc_cpumask_var(&top_cpuset.effective_cpus, GFP_KERNEL)); |
bf92370c0 cpuset: Simply al... |
2821 |
BUG_ON(!zalloc_cpumask_var(&top_cpuset.subparts_cpus, GFP_KERNEL)); |
c8dc4422c ANDROID: cpuset: ... |
2822 |
BUG_ON(!alloc_cpumask_var(&top_cpuset.cpus_requested, GFP_KERNEL)); |
58568d2a8 cpuset,mm: update... |
2823 |
|
300ed6cbb cpuset: convert c... |
2824 |
cpumask_setall(top_cpuset.cpus_allowed); |
c8dc4422c ANDROID: cpuset: ... |
2825 |
cpumask_setall(top_cpuset.cpus_requested); |
f9a86fcbb cpuset: modify cp... |
2826 |
nodes_setall(top_cpuset.mems_allowed); |
e2b9a3d7d cpuset: add cs->e... |
2827 2828 |
cpumask_setall(top_cpuset.effective_cpus); nodes_setall(top_cpuset.effective_mems); |
1da177e4c Linux-2.6.12-rc2 |
2829 |
|
3e0d98b9f [PATCH] cpuset: m... |
2830 |
fmeter_init(&top_cpuset.fmeter); |
029190c51 cpuset sched_load... |
2831 |
set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags); |
1d3504fcf sched, cpuset: cu... |
2832 |
top_cpuset.relax_domain_level = -1; |
1da177e4c Linux-2.6.12-rc2 |
2833 |
|
75fa8e5d3 cgroup: switch to... |
2834 |
BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL)); |
2341d1b65 cpuset: convert c... |
2835 |
|
8793d854e Task Control Grou... |
2836 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
2837 |
} |
b1aac8bb8 [PATCH] cpuset: h... |
2838 |
/* |
cf417141c sched, cpuset: re... |
2839 |
* If CPU and/or memory hotplug handlers, below, unplug any CPUs |
b1aac8bb8 [PATCH] cpuset: h... |
2840 2841 |
* or memory nodes, we need to walk over the cpuset hierarchy, * removing that CPU or node from all cpusets. If this removes the |
956db3ca0 hotplug cpu: move... |
2842 2843 |
* last CPU or node from a cpuset, then move the tasks in the empty * cpuset to its next-highest non-empty parent. |
b1aac8bb8 [PATCH] cpuset: h... |
2844 |
*/ |
956db3ca0 hotplug cpu: move... |
2845 2846 2847 |
static void remove_tasks_in_empty_cpuset(struct cpuset *cs) { struct cpuset *parent; |
c8d9c90c7 hotplug cpu: move... |
2848 |
/* |
956db3ca0 hotplug cpu: move... |
2849 2850 2851 |
* Find its next-highest non-empty parent, (top cpuset * has online cpus, so can't be empty). */ |
c431069fe cpuset: remove cp... |
2852 |
parent = parent_cs(cs); |
300ed6cbb cpuset: convert c... |
2853 |
while (cpumask_empty(parent->cpus_allowed) || |
b45012955 hotplug cpu move ... |
2854 |
nodes_empty(parent->mems_allowed)) |
c431069fe cpuset: remove cp... |
2855 |
parent = parent_cs(parent); |
956db3ca0 hotplug cpu: move... |
2856 |
|
8cc993452 cgroup, cpuset: r... |
2857 |
if (cgroup_transfer_tasks(parent->css.cgroup, cs->css.cgroup)) { |
12d3089c1 kernel/cpuset.c: ... |
2858 |
pr_err("cpuset: failed to transfer tasks out of empty cpuset "); |
e61734c55 cgroup: remove cg... |
2859 2860 2861 |
pr_cont_cgroup_name(cs->css.cgroup); pr_cont(" "); |
8cc993452 cgroup, cpuset: r... |
2862 |
} |
956db3ca0 hotplug cpu: move... |
2863 |
} |
be4c9dd7a cpuset: enable on... |
2864 2865 2866 2867 |
static void hotplug_update_tasks_legacy(struct cpuset *cs, struct cpumask *new_cpus, nodemask_t *new_mems, bool cpus_updated, bool mems_updated) |
390a36aad cpuset: refactor ... |
2868 2869 |
{ bool is_empty; |
8447a0fee cpuset: convert c... |
2870 |
spin_lock_irq(&callback_lock); |
be4c9dd7a cpuset: enable on... |
2871 2872 2873 2874 |
cpumask_copy(cs->cpus_allowed, new_cpus); cpumask_copy(cs->effective_cpus, new_cpus); cs->mems_allowed = *new_mems; cs->effective_mems = *new_mems; |
8447a0fee cpuset: convert c... |
2875 |
spin_unlock_irq(&callback_lock); |
390a36aad cpuset: refactor ... |
2876 2877 2878 2879 2880 |
/* * Don't call update_tasks_cpumask() if the cpuset becomes empty, * as the tasks will be migratecd to an ancestor. */ |
be4c9dd7a cpuset: enable on... |
2881 |
if (cpus_updated && !cpumask_empty(cs->cpus_allowed)) |
390a36aad cpuset: refactor ... |
2882 |
update_tasks_cpumask(cs); |
be4c9dd7a cpuset: enable on... |
2883 |
if (mems_updated && !nodes_empty(cs->mems_allowed)) |
390a36aad cpuset: refactor ... |
2884 2885 2886 2887 |
update_tasks_nodemask(cs); is_empty = cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed); |
18695a120 Revert "cgroup/cp... |
2888 |
mutex_unlock(&cpuset_mutex); |
390a36aad cpuset: refactor ... |
2889 2890 2891 2892 2893 2894 2895 2896 |
/* * Move tasks to the nearest ancestor with execution resources, * This is full cgroup operation which will also call back into * cpuset. Should be done outside any lock. */ if (is_empty) remove_tasks_in_empty_cpuset(cs); |
18695a120 Revert "cgroup/cp... |
2897 |
mutex_lock(&cpuset_mutex); |
390a36aad cpuset: refactor ... |
2898 |
} |
be4c9dd7a cpuset: enable on... |
2899 2900 2901 2902 |
static void hotplug_update_tasks(struct cpuset *cs, struct cpumask *new_cpus, nodemask_t *new_mems, bool cpus_updated, bool mems_updated) |
390a36aad cpuset: refactor ... |
2903 |
{ |
be4c9dd7a cpuset: enable on... |
2904 2905 2906 2907 |
if (cpumask_empty(new_cpus)) cpumask_copy(new_cpus, parent_cs(cs)->effective_cpus); if (nodes_empty(*new_mems)) *new_mems = parent_cs(cs)->effective_mems; |
8447a0fee cpuset: convert c... |
2908 |
spin_lock_irq(&callback_lock); |
be4c9dd7a cpuset: enable on... |
2909 2910 |
cpumask_copy(cs->effective_cpus, new_cpus); cs->effective_mems = *new_mems; |
8447a0fee cpuset: convert c... |
2911 |
spin_unlock_irq(&callback_lock); |
390a36aad cpuset: refactor ... |
2912 |
|
be4c9dd7a cpuset: enable on... |
2913 |
if (cpus_updated) |
390a36aad cpuset: refactor ... |
2914 |
update_tasks_cpumask(cs); |
be4c9dd7a cpuset: enable on... |
2915 |
if (mems_updated) |
390a36aad cpuset: refactor ... |
2916 2917 |
update_tasks_nodemask(cs); } |
4b842da27 cpuset: Make CPU ... |
2918 2919 2920 2921 2922 2923 |
static bool force_rebuild; void cpuset_force_rebuild(void) { force_rebuild = true; } |
deb7aa308 cpuset: reorganiz... |
2924 |
/** |
388afd854 cpuset: remove as... |
2925 |
* cpuset_hotplug_update_tasks - update tasks in a cpuset for hotunplug |
deb7aa308 cpuset: reorganiz... |
2926 |
* @cs: cpuset in interest |
4b842da27 cpuset: Make CPU ... |
2927 |
* @tmp: the tmpmasks structure pointer |
956db3ca0 hotplug cpu: move... |
2928 |
* |
deb7aa308 cpuset: reorganiz... |
2929 2930 2931 |
* Compare @cs's cpu and mem masks against top_cpuset and if some have gone * offline, update @cs accordingly. If @cs ends up with no CPU or memory, * all its tasks are moved to the nearest ancestor with both resources. |
80d1fa646 cpusets, hotplug:... |
2932 |
*/ |
4b842da27 cpuset: Make CPU ... |
2933 |
static void cpuset_hotplug_update_tasks(struct cpuset *cs, struct tmpmasks *tmp) |
80d1fa646 cpusets, hotplug:... |
2934 |
{ |
be4c9dd7a cpuset: enable on... |
2935 2936 2937 2938 |
static cpumask_t new_cpus; static nodemask_t new_mems; bool cpus_updated; bool mems_updated; |
4b842da27 cpuset: Make CPU ... |
2939 |
struct cpuset *parent; |
e44193d39 cpuset: let hotpl... |
2940 2941 |
retry: wait_event(cpuset_attach_wq, cs->attach_in_progress == 0); |
80d1fa646 cpusets, hotplug:... |
2942 |
|
18695a120 Revert "cgroup/cp... |
2943 |
mutex_lock(&cpuset_mutex); |
7ddf96b02 cpusets, hotplug:... |
2944 |
|
e44193d39 cpuset: let hotpl... |
2945 2946 2947 2948 2949 |
/* * We have raced with task attaching. We wait until attaching * is finished, so we won't attach a task to an empty cpuset. */ if (cs->attach_in_progress) { |
18695a120 Revert "cgroup/cp... |
2950 |
mutex_unlock(&cpuset_mutex); |
e44193d39 cpuset: let hotpl... |
2951 2952 |
goto retry; } |
4b842da27 cpuset: Make CPU ... |
2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 |
parent = parent_cs(cs); compute_effective_cpumask(&new_cpus, cs, parent); nodes_and(new_mems, cs->mems_allowed, parent->effective_mems); if (cs->nr_subparts_cpus) /* * Make sure that CPUs allocated to child partitions * do not show up in effective_cpus. */ cpumask_andnot(&new_cpus, &new_cpus, cs->subparts_cpus); if (!tmp || !cs->partition_root_state) goto update_tasks; |
80d1fa646 cpusets, hotplug:... |
2966 |
|
4b842da27 cpuset: Make CPU ... |
2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 |
/* * In the unlikely event that a partition root has empty * effective_cpus or its parent becomes erroneous, we have to * transition it to the erroneous state. */ if (is_partition_root(cs) && (cpumask_empty(&new_cpus) || (parent->partition_root_state == PRS_ERROR))) { if (cs->nr_subparts_cpus) { cs->nr_subparts_cpus = 0; cpumask_clear(cs->subparts_cpus); compute_effective_cpumask(&new_cpus, cs, parent); } |
80d1fa646 cpusets, hotplug:... |
2979 |
|
4b842da27 cpuset: Make CPU ... |
2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 |
/* * If the effective_cpus is empty because the child * partitions take away all the CPUs, we can keep * the current partition and let the child partitions * fight for available CPUs. */ if ((parent->partition_root_state == PRS_ERROR) || cpumask_empty(&new_cpus)) { update_parent_subparts_cpumask(cs, partcmd_disable, NULL, tmp); cs->partition_root_state = PRS_ERROR; } cpuset_force_rebuild(); } /* * On the other hand, an erroneous partition root may be transitioned * back to a regular one or a partition root with no CPU allocated * from the parent may change to erroneous. */ if (is_partition_root(parent) && ((cs->partition_root_state == PRS_ERROR) || !cpumask_intersects(&new_cpus, parent->subparts_cpus)) && update_parent_subparts_cpumask(cs, partcmd_update, NULL, tmp)) cpuset_force_rebuild(); update_tasks: |
be4c9dd7a cpuset: enable on... |
3007 3008 |
cpus_updated = !cpumask_equal(&new_cpus, cs->effective_cpus); mems_updated = !nodes_equal(new_mems, cs->effective_mems); |
deb7aa308 cpuset: reorganiz... |
3009 |
|
b8d1b8ee9 cpuset: Allow v2 ... |
3010 |
if (is_in_v2_mode()) |
be4c9dd7a cpuset: enable on... |
3011 3012 |
hotplug_update_tasks(cs, &new_cpus, &new_mems, cpus_updated, mems_updated); |
390a36aad cpuset: refactor ... |
3013 |
else |
be4c9dd7a cpuset: enable on... |
3014 3015 |
hotplug_update_tasks_legacy(cs, &new_cpus, &new_mems, cpus_updated, mems_updated); |
8d0339487 cpuset: make CPU ... |
3016 |
|
18695a120 Revert "cgroup/cp... |
3017 |
mutex_unlock(&cpuset_mutex); |
b1aac8bb8 [PATCH] cpuset: h... |
3018 |
} |
deb7aa308 cpuset: reorganiz... |
3019 |
/** |
3a5a6d0c2 cpuset: don't nes... |
3020 |
* cpuset_hotplug_workfn - handle CPU/memory hotunplug for a cpuset |
956db3ca0 hotplug cpu: move... |
3021 |
* |
deb7aa308 cpuset: reorganiz... |
3022 3023 3024 3025 3026 |
* This function is called after either CPU or memory configuration has * changed and updates cpuset accordingly. The top_cpuset is always * synchronized to cpu_active_mask and N_MEMORY, which is necessary in * order to make cpusets transparent (of no affect) on systems that are * actively using CPU hotplug but making no active use of cpusets. |
956db3ca0 hotplug cpu: move... |
3027 |
* |
deb7aa308 cpuset: reorganiz... |
3028 |
* Non-root cpusets are only affected by offlining. If any CPUs or memory |
388afd854 cpuset: remove as... |
3029 3030 |
* nodes have been taken down, cpuset_hotplug_update_tasks() is invoked on * all descendants. |
956db3ca0 hotplug cpu: move... |
3031 |
* |
deb7aa308 cpuset: reorganiz... |
3032 3033 |
* Note that CPU offlining during suspend is ignored. We don't modify * cpusets across suspend/resume cycles at all. |
956db3ca0 hotplug cpu: move... |
3034 |
*/ |
1d3a64fbd ANDROID: cpu/hotp... |
3035 |
void cpuset_hotplug_workfn(struct work_struct *work) |
b1aac8bb8 [PATCH] cpuset: h... |
3036 |
{ |
5c5cc6232 cpuset: allow to ... |
3037 3038 |
static cpumask_t new_cpus; static nodemask_t new_mems; |
deb7aa308 cpuset: reorganiz... |
3039 |
bool cpus_updated, mems_updated; |
b8d1b8ee9 cpuset: Allow v2 ... |
3040 |
bool on_dfl = is_in_v2_mode(); |
4b842da27 cpuset: Make CPU ... |
3041 3042 3043 3044 |
struct tmpmasks tmp, *ptmp = NULL; if (on_dfl && !alloc_cpumasks(NULL, &tmp)) ptmp = &tmp; |
b1aac8bb8 [PATCH] cpuset: h... |
3045 |
|
18695a120 Revert "cgroup/cp... |
3046 |
mutex_lock(&cpuset_mutex); |
956db3ca0 hotplug cpu: move... |
3047 |
|
deb7aa308 cpuset: reorganiz... |
3048 3049 3050 |
/* fetch the available cpus/mems and find out which changed how */ cpumask_copy(&new_cpus, cpu_active_mask); new_mems = node_states[N_MEMORY]; |
7ddf96b02 cpusets, hotplug:... |
3051 |
|
4b842da27 cpuset: Make CPU ... |
3052 3053 3054 3055 3056 |
/* * If subparts_cpus is populated, it is likely that the check below * will produce a false positive on cpus_updated when the cpu list * isn't changed. It is extra work, but it is better to be safe. */ |
7e88291be cpuset: make cs->... |
3057 3058 |
cpus_updated = !cpumask_equal(top_cpuset.effective_cpus, &new_cpus); mems_updated = !nodes_equal(top_cpuset.effective_mems, new_mems); |
7ddf96b02 cpusets, hotplug:... |
3059 |
|
deb7aa308 cpuset: reorganiz... |
3060 3061 |
/* synchronize cpus_allowed to cpu_active_mask */ if (cpus_updated) { |
8447a0fee cpuset: convert c... |
3062 |
spin_lock_irq(&callback_lock); |
7e88291be cpuset: make cs->... |
3063 3064 |
if (!on_dfl) cpumask_copy(top_cpuset.cpus_allowed, &new_cpus); |
4b842da27 cpuset: Make CPU ... |
3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 |
/* * Make sure that CPUs allocated to child partitions * do not show up in effective_cpus. If no CPU is left, * we clear the subparts_cpus & let the child partitions * fight for the CPUs again. */ if (top_cpuset.nr_subparts_cpus) { if (cpumask_subset(&new_cpus, top_cpuset.subparts_cpus)) { top_cpuset.nr_subparts_cpus = 0; cpumask_clear(top_cpuset.subparts_cpus); } else { cpumask_andnot(&new_cpus, &new_cpus, top_cpuset.subparts_cpus); } } |
1344ab9c2 cpuset: update cp... |
3081 |
cpumask_copy(top_cpuset.effective_cpus, &new_cpus); |
8447a0fee cpuset: convert c... |
3082 |
spin_unlock_irq(&callback_lock); |
deb7aa308 cpuset: reorganiz... |
3083 3084 |
/* we don't mess with cpumasks of tasks in top_cpuset */ } |
b45012955 hotplug cpu move ... |
3085 |
|
deb7aa308 cpuset: reorganiz... |
3086 3087 |
/* synchronize mems_allowed to N_MEMORY */ if (mems_updated) { |
8447a0fee cpuset: convert c... |
3088 |
spin_lock_irq(&callback_lock); |
7e88291be cpuset: make cs->... |
3089 3090 |
if (!on_dfl) top_cpuset.mems_allowed = new_mems; |
1344ab9c2 cpuset: update cp... |
3091 |
top_cpuset.effective_mems = new_mems; |
8447a0fee cpuset: convert c... |
3092 |
spin_unlock_irq(&callback_lock); |
d66393e54 cpuset: use css_t... |
3093 |
update_tasks_nodemask(&top_cpuset); |
deb7aa308 cpuset: reorganiz... |
3094 |
} |
b45012955 hotplug cpu move ... |
3095 |
|
18695a120 Revert "cgroup/cp... |
3096 |
mutex_unlock(&cpuset_mutex); |
388afd854 cpuset: remove as... |
3097 |
|
5c5cc6232 cpuset: allow to ... |
3098 3099 |
/* if cpus or mems changed, we need to propagate to descendants */ if (cpus_updated || mems_updated) { |
deb7aa308 cpuset: reorganiz... |
3100 |
struct cpuset *cs; |
492eb21b9 cgroup: make hier... |
3101 |
struct cgroup_subsys_state *pos_css; |
f9b4fb8da cpusets: update t... |
3102 |
|
fc560a26a cpuset: replace c... |
3103 |
rcu_read_lock(); |
492eb21b9 cgroup: make hier... |
3104 |
cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { |
ec903c0c8 cgroup: rename cs... |
3105 |
if (cs == &top_cpuset || !css_tryget_online(&cs->css)) |
388afd854 cpuset: remove as... |
3106 3107 |
continue; rcu_read_unlock(); |
7ddf96b02 cpusets, hotplug:... |
3108 |
|
4b842da27 cpuset: Make CPU ... |
3109 |
cpuset_hotplug_update_tasks(cs, ptmp); |
b45012955 hotplug cpu move ... |
3110 |
|
388afd854 cpuset: remove as... |
3111 3112 3113 3114 3115 |
rcu_read_lock(); css_put(&cs->css); } rcu_read_unlock(); } |
8d0339487 cpuset: make CPU ... |
3116 |
|
deb7aa308 cpuset: reorganiz... |
3117 |
/* rebuild sched domains if cpus_allowed has changed */ |
50e766323 sched/cpuset/pm: ... |
3118 3119 |
if (cpus_updated || force_rebuild) { force_rebuild = false; |
e0e80a02e cpuset: use rebui... |
3120 |
rebuild_sched_domains(); |
50e766323 sched/cpuset/pm: ... |
3121 |
} |
4b842da27 cpuset: Make CPU ... |
3122 3123 |
free_cpumasks(NULL, ptmp); |
b1aac8bb8 [PATCH] cpuset: h... |
3124 |
} |
30e03acda cpuset: Remove cp... |
3125 |
void cpuset_update_active_cpus(void) |
4c4d50f7b [PATCH] cpuset: t... |
3126 |
{ |
3a5a6d0c2 cpuset: don't nes... |
3127 3128 3129 3130 |
/* * We're inside cpu hotplug critical region which usually nests * inside cgroup synchronization. Bounce actual hotplug processing * to a work item to avoid reverse locking order. |
3a5a6d0c2 cpuset: don't nes... |
3131 |
*/ |
3a5a6d0c2 cpuset: don't nes... |
3132 |
schedule_work(&cpuset_hotplug_work); |
4c4d50f7b [PATCH] cpuset: t... |
3133 |
} |
4c4d50f7b [PATCH] cpuset: t... |
3134 |
|
50e766323 sched/cpuset/pm: ... |
3135 3136 3137 3138 |
void cpuset_wait_for_hotplug(void) { flush_work(&cpuset_hotplug_work); } |
38837fc75 [PATCH] cpuset: t... |
3139 |
/* |
38d7bee9d cpuset: use N_MEM... |
3140 3141 |
* Keep top_cpuset.mems_allowed tracking node_states[N_MEMORY]. * Call this routine anytime after node_states[N_MEMORY] changes. |
a1cd2b13f cpusets: Remove/u... |
3142 |
* See cpuset_update_active_cpus() for CPU hotplug handling. |
38837fc75 [PATCH] cpuset: t... |
3143 |
*/ |
f481891fd cpuset: update to... |
3144 3145 |
static int cpuset_track_online_nodes(struct notifier_block *self, unsigned long action, void *arg) |
38837fc75 [PATCH] cpuset: t... |
3146 |
{ |
3a5a6d0c2 cpuset: don't nes... |
3147 |
schedule_work(&cpuset_hotplug_work); |
f481891fd cpuset: update to... |
3148 |
return NOTIFY_OK; |
38837fc75 [PATCH] cpuset: t... |
3149 |
} |
d8f10cb3d kernel/cpuset.c: ... |
3150 3151 3152 3153 3154 |
static struct notifier_block cpuset_track_online_nodes_nb = { .notifier_call = cpuset_track_online_nodes, .priority = 10, /* ??! */ }; |
38837fc75 [PATCH] cpuset: t... |
3155 |
|
1da177e4c Linux-2.6.12-rc2 |
3156 3157 3158 3159 |
/** * cpuset_init_smp - initialize cpus_allowed * * Description: Finish top cpuset after cpu, node maps are initialized |
d8f10cb3d kernel/cpuset.c: ... |
3160 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
3161 3162 |
void __init cpuset_init_smp(void) { |
6ad4c1888 sched: Fix balanc... |
3163 |
cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); |
38d7bee9d cpuset: use N_MEM... |
3164 |
top_cpuset.mems_allowed = node_states[N_MEMORY]; |
33ad801df cpuset: record ol... |
3165 |
top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; |
4c4d50f7b [PATCH] cpuset: t... |
3166 |
|
e2b9a3d7d cpuset: add cs->e... |
3167 3168 |
cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); top_cpuset.effective_mems = node_states[N_MEMORY]; |
d8f10cb3d kernel/cpuset.c: ... |
3169 |
register_hotmemory_notifier(&cpuset_track_online_nodes_nb); |
e93ad19d0 cpuset: make mm m... |
3170 3171 3172 |
cpuset_migrate_mm_wq = alloc_ordered_workqueue("cpuset_migrate_mm", 0); BUG_ON(!cpuset_migrate_mm_wq); |
1da177e4c Linux-2.6.12-rc2 |
3173 3174 3175 |
} /** |
1da177e4c Linux-2.6.12-rc2 |
3176 3177 |
* cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset. * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed. |
6af866af3 cpuset: remove re... |
3178 |
* @pmask: pointer to struct cpumask variable to receive cpus_allowed set. |
1da177e4c Linux-2.6.12-rc2 |
3179 |
* |
300ed6cbb cpuset: convert c... |
3180 |
* Description: Returns the cpumask_var_t cpus_allowed of the cpuset |
1da177e4c Linux-2.6.12-rc2 |
3181 |
* attached to the specified @tsk. Guaranteed to return some non-empty |
5f054e31c documentation: re... |
3182 |
* subset of cpu_online_mask, even if this means going outside the |
1da177e4c Linux-2.6.12-rc2 |
3183 3184 |
* tasks cpuset. **/ |
6af866af3 cpuset: remove re... |
3185 |
void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) |
1da177e4c Linux-2.6.12-rc2 |
3186 |
{ |
8447a0fee cpuset: convert c... |
3187 3188 3189 |
unsigned long flags; spin_lock_irqsave(&callback_lock, flags); |
b8dadcb58 cpuset: use rcu_r... |
3190 |
rcu_read_lock(); |
ae1c80238 cpuset: apply cs-... |
3191 |
guarantee_online_cpus(task_cs(tsk), pmask); |
b8dadcb58 cpuset: use rcu_r... |
3192 |
rcu_read_unlock(); |
8447a0fee cpuset: convert c... |
3193 |
spin_unlock_irqrestore(&callback_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
3194 |
} |
d477f8c20 cpuset: restore s... |
3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 |
/** * cpuset_cpus_allowed_fallback - final fallback before complete catastrophe. * @tsk: pointer to task_struct with which the scheduler is struggling * * Description: In the case that the scheduler cannot find an allowed cpu in * tsk->cpus_allowed, we fall back to task_cs(tsk)->cpus_allowed. In legacy * mode however, this value is the same as task_cs(tsk)->effective_cpus, * which will not contain a sane cpumask during cases such as cpu hotplugging. * This is the absolute last resort for the scheduler and it is only used if * _every_ other avenue has been traveled. **/ |
2baab4e90 sched: Fix select... |
3206 |
void cpuset_cpus_allowed_fallback(struct task_struct *tsk) |
9084bb824 sched: Make selec... |
3207 |
{ |
9084bb824 sched: Make selec... |
3208 |
rcu_read_lock(); |
d477f8c20 cpuset: restore s... |
3209 3210 |
do_set_cpus_allowed(tsk, is_in_v2_mode() ? task_cs(tsk)->cpus_allowed : cpu_possible_mask); |
9084bb824 sched: Make selec... |
3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 |
rcu_read_unlock(); /* * We own tsk->cpus_allowed, nobody can change it under us. * * But we used cs && cs->cpus_allowed lockless and thus can * race with cgroup_attach_task() or update_cpumask() and get * the wrong tsk->cpus_allowed. However, both cases imply the * subsequent cpuset_change_cpumask()->set_cpus_allowed_ptr() * which takes task_rq_lock(). * * If we are called after it dropped the lock we must see all * changes in tsk_cs()->cpus_allowed. Otherwise we can temporary * set any mask even if it is not right from task_cs() pov, * the pending set_cpus_allowed_ptr() will fix things. |
2baab4e90 sched: Fix select... |
3226 3227 3228 |
* * select_fallback_rq() will fix things ups and set cpu_possible_mask * if required. |
9084bb824 sched: Make selec... |
3229 |
*/ |
9084bb824 sched: Make selec... |
3230 |
} |
8f4ab07f4 kernel/cpuset.c: ... |
3231 |
void __init cpuset_init_current_mems_allowed(void) |
1da177e4c Linux-2.6.12-rc2 |
3232 |
{ |
f9a86fcbb cpuset: modify cp... |
3233 |
nodes_setall(current->mems_allowed); |
1da177e4c Linux-2.6.12-rc2 |
3234 |
} |
d9fd8a6d4 [PATCH] kernel/cp... |
3235 |
/** |
909d75a3b [PATCH] cpuset: i... |
3236 3237 3238 3239 3240 |
* cpuset_mems_allowed - return mems_allowed mask from a tasks cpuset. * @tsk: pointer to task_struct from which to obtain cpuset->mems_allowed. * * Description: Returns the nodemask_t mems_allowed of the cpuset * attached to the specified @tsk. Guaranteed to return some non-empty |
38d7bee9d cpuset: use N_MEM... |
3241 |
* subset of node_states[N_MEMORY], even if this means going outside the |
909d75a3b [PATCH] cpuset: i... |
3242 3243 3244 3245 3246 3247 |
* tasks cpuset. **/ nodemask_t cpuset_mems_allowed(struct task_struct *tsk) { nodemask_t mask; |
8447a0fee cpuset: convert c... |
3248 |
unsigned long flags; |
909d75a3b [PATCH] cpuset: i... |
3249 |
|
8447a0fee cpuset: convert c... |
3250 |
spin_lock_irqsave(&callback_lock, flags); |
b8dadcb58 cpuset: use rcu_r... |
3251 |
rcu_read_lock(); |
ae1c80238 cpuset: apply cs-... |
3252 |
guarantee_online_mems(task_cs(tsk), &mask); |
b8dadcb58 cpuset: use rcu_r... |
3253 |
rcu_read_unlock(); |
8447a0fee cpuset: convert c... |
3254 |
spin_unlock_irqrestore(&callback_lock, flags); |
909d75a3b [PATCH] cpuset: i... |
3255 3256 3257 3258 3259 |
return mask; } /** |
19770b326 mm: filter based ... |
3260 3261 |
* cpuset_nodemask_valid_mems_allowed - check nodemask vs. curremt mems_allowed * @nodemask: the nodemask to be checked |
d9fd8a6d4 [PATCH] kernel/cp... |
3262 |
* |
19770b326 mm: filter based ... |
3263 |
* Are any of the nodes in the nodemask allowed in current->mems_allowed? |
1da177e4c Linux-2.6.12-rc2 |
3264 |
*/ |
19770b326 mm: filter based ... |
3265 |
int cpuset_nodemask_valid_mems_allowed(nodemask_t *nodemask) |
1da177e4c Linux-2.6.12-rc2 |
3266 |
{ |
19770b326 mm: filter based ... |
3267 |
return nodes_intersects(*nodemask, current->mems_allowed); |
1da177e4c Linux-2.6.12-rc2 |
3268 |
} |
9bf2229f8 [PATCH] cpusets: ... |
3269 |
/* |
786083667 Cpuset hardwall f... |
3270 3271 |
* nearest_hardwall_ancestor() - Returns the nearest mem_exclusive or * mem_hardwall ancestor to the specified cpuset. Call holding |
8447a0fee cpuset: convert c... |
3272 |
* callback_lock. If no ancestor is mem_exclusive or mem_hardwall |
786083667 Cpuset hardwall f... |
3273 |
* (an unusual configuration), then returns the root cpuset. |
9bf2229f8 [PATCH] cpusets: ... |
3274 |
*/ |
c9710d801 cpuset: drop "con... |
3275 |
static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) |
9bf2229f8 [PATCH] cpusets: ... |
3276 |
{ |
c431069fe cpuset: remove cp... |
3277 3278 |
while (!(is_mem_exclusive(cs) || is_mem_hardwall(cs)) && parent_cs(cs)) cs = parent_cs(cs); |
9bf2229f8 [PATCH] cpusets: ... |
3279 3280 |
return cs; } |
d9fd8a6d4 [PATCH] kernel/cp... |
3281 |
/** |
344736f29 cpuset: simplify ... |
3282 |
* cpuset_node_allowed - Can we allocate on a memory node? |
a1bc5a4ee cpusets: replace ... |
3283 |
* @node: is this an allowed node? |
02a0e53d8 [PATCH] cpuset: r... |
3284 |
* @gfp_mask: memory allocation flags |
d9fd8a6d4 [PATCH] kernel/cp... |
3285 |
* |
6e276d2a5 kernel, cpuset: r... |
3286 3287 3288 |
* If we're in interrupt, yes, we can always allocate. If @node is set in * current's mems_allowed, yes. If it's not a __GFP_HARDWALL request and this * node is set in the nearest hardwalled cpuset ancestor to current's cpuset, |
da99ecf11 mm: replace TIF_M... |
3289 |
* yes. If current has access to memory reserves as an oom victim, yes. |
9bf2229f8 [PATCH] cpusets: ... |
3290 3291 3292 |
* Otherwise, no. * * GFP_USER allocations are marked with the __GFP_HARDWALL bit, |
c596d9f32 cpusets: allow TI... |
3293 |
* and do not allow allocations outside the current tasks cpuset |
da99ecf11 mm: replace TIF_M... |
3294 |
* unless the task has been OOM killed. |
9bf2229f8 [PATCH] cpusets: ... |
3295 |
* GFP_KERNEL allocations are not so marked, so can escape to the |
786083667 Cpuset hardwall f... |
3296 |
* nearest enclosing hardwalled ancestor cpuset. |
9bf2229f8 [PATCH] cpusets: ... |
3297 |
* |
8447a0fee cpuset: convert c... |
3298 |
* Scanning up parent cpusets requires callback_lock. The |
02a0e53d8 [PATCH] cpuset: r... |
3299 3300 3301 3302 |
* __alloc_pages() routine only calls here with __GFP_HARDWALL bit * _not_ set if it's a GFP_KERNEL allocation, and all nodes in the * current tasks mems_allowed came up empty on the first pass over * the zonelist. So only GFP_KERNEL allocations, if all nodes in the |
8447a0fee cpuset: convert c... |
3303 |
* cpuset are short of memory, might require taking the callback_lock. |
9bf2229f8 [PATCH] cpusets: ... |
3304 |
* |
36be57ffe [PATCH] cpuset: u... |
3305 |
* The first call here from mm/page_alloc:get_page_from_freelist() |
02a0e53d8 [PATCH] cpuset: r... |
3306 3307 3308 |
* has __GFP_HARDWALL set in gfp_mask, enforcing hardwall cpusets, * so no allocation on a node outside the cpuset is allowed (unless * in interrupt, of course). |
36be57ffe [PATCH] cpuset: u... |
3309 3310 3311 3312 3313 3314 |
* * The second pass through get_page_from_freelist() doesn't even call * here for GFP_ATOMIC calls. For those calls, the __alloc_pages() * variable 'wait' is not set, and the bit ALLOC_CPUSET is not set * in alloc_flags. That logic and the checks below have the combined * affect that: |
9bf2229f8 [PATCH] cpusets: ... |
3315 3316 |
* in_interrupt - any node ok (current task context irrelevant) * GFP_ATOMIC - any node ok |
da99ecf11 mm: replace TIF_M... |
3317 |
* tsk_is_oom_victim - any node ok |
786083667 Cpuset hardwall f... |
3318 |
* GFP_KERNEL - any node in enclosing hardwalled cpuset ok |
9bf2229f8 [PATCH] cpusets: ... |
3319 |
* GFP_USER - only nodes in current tasks mems allowed ok. |
02a0e53d8 [PATCH] cpuset: r... |
3320 |
*/ |
002f29062 cpuset: use stati... |
3321 |
bool __cpuset_node_allowed(int node, gfp_t gfp_mask) |
1da177e4c Linux-2.6.12-rc2 |
3322 |
{ |
c9710d801 cpuset: drop "con... |
3323 |
struct cpuset *cs; /* current cpuset ancestors */ |
29afd49b7 [PATCH] cpuset: r... |
3324 |
int allowed; /* is allocation in zone z allowed? */ |
8447a0fee cpuset: convert c... |
3325 |
unsigned long flags; |
9bf2229f8 [PATCH] cpusets: ... |
3326 |
|
6e276d2a5 kernel, cpuset: r... |
3327 |
if (in_interrupt()) |
002f29062 cpuset: use stati... |
3328 |
return true; |
9bf2229f8 [PATCH] cpusets: ... |
3329 |
if (node_isset(node, current->mems_allowed)) |
002f29062 cpuset: use stati... |
3330 |
return true; |
c596d9f32 cpusets: allow TI... |
3331 3332 3333 3334 |
/* * Allow tasks that have access to memory reserves because they have * been OOM killed to get memory anywhere. */ |
da99ecf11 mm: replace TIF_M... |
3335 |
if (unlikely(tsk_is_oom_victim(current))) |
002f29062 cpuset: use stati... |
3336 |
return true; |
9bf2229f8 [PATCH] cpusets: ... |
3337 |
if (gfp_mask & __GFP_HARDWALL) /* If hardwall request, stop here */ |
002f29062 cpuset: use stati... |
3338 |
return false; |
9bf2229f8 [PATCH] cpusets: ... |
3339 |
|
5563e7707 [PATCH] cpuset: f... |
3340 |
if (current->flags & PF_EXITING) /* Let dying task have memory */ |
002f29062 cpuset: use stati... |
3341 |
return true; |
5563e7707 [PATCH] cpuset: f... |
3342 |
|
9bf2229f8 [PATCH] cpusets: ... |
3343 |
/* Not hardwall and node outside mems_allowed: scan up cpusets */ |
8447a0fee cpuset: convert c... |
3344 |
spin_lock_irqsave(&callback_lock, flags); |
053199edf [PATCH] cpusets: ... |
3345 |
|
b8dadcb58 cpuset: use rcu_r... |
3346 |
rcu_read_lock(); |
786083667 Cpuset hardwall f... |
3347 |
cs = nearest_hardwall_ancestor(task_cs(current)); |
99afb0fd5 cpuset: fix a rac... |
3348 |
allowed = node_isset(node, cs->mems_allowed); |
b8dadcb58 cpuset: use rcu_r... |
3349 |
rcu_read_unlock(); |
053199edf [PATCH] cpusets: ... |
3350 |
|
8447a0fee cpuset: convert c... |
3351 |
spin_unlock_irqrestore(&callback_lock, flags); |
9bf2229f8 [PATCH] cpusets: ... |
3352 |
return allowed; |
1da177e4c Linux-2.6.12-rc2 |
3353 |
} |
ef08e3b49 [PATCH] cpusets: ... |
3354 |
/** |
6adef3ebe cpusets: new roun... |
3355 3356 |
* cpuset_mem_spread_node() - On which node to begin search for a file page * cpuset_slab_spread_node() - On which node to begin search for a slab page |
825a46af5 [PATCH] cpuset me... |
3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 |
* * If a task is marked PF_SPREAD_PAGE or PF_SPREAD_SLAB (as for * tasks in a cpuset with is_spread_page or is_spread_slab set), * and if the memory allocation used cpuset_mem_spread_node() * to determine on which node to start looking, as it will for * certain page cache or slab cache pages such as used for file * system buffers and inode caches, then instead of starting on the * local node to look for a free page, rather spread the starting * node around the tasks mems_allowed nodes. * * We don't have to worry about the returned node being offline * because "it can't happen", and even if it did, it would be ok. * * The routines calling guarantee_online_mems() are careful to * only set nodes in task->mems_allowed that are online. So it * should not be possible for the following code to return an * offline node. But if it did, that would be ok, as this routine * is not returning the node where the allocation must be, only * the node where the search should start. The zonelist passed to * __alloc_pages() will include all nodes. If the slab allocator * is passed an offline node, it will fall back to the local node. * See kmem_cache_alloc_node(). */ |
6adef3ebe cpusets: new roun... |
3380 |
static int cpuset_spread_node(int *rotor) |
825a46af5 [PATCH] cpuset me... |
3381 |
{ |
0edaf86cf include/linux/nod... |
3382 |
return *rotor = next_node_in(*rotor, current->mems_allowed); |
825a46af5 [PATCH] cpuset me... |
3383 |
} |
6adef3ebe cpusets: new roun... |
3384 3385 3386 |
int cpuset_mem_spread_node(void) { |
778d3b0ff cpusets: randomiz... |
3387 3388 3389 |
if (current->cpuset_mem_spread_rotor == NUMA_NO_NODE) current->cpuset_mem_spread_rotor = node_random(¤t->mems_allowed); |
6adef3ebe cpusets: new roun... |
3390 3391 3392 3393 3394 |
return cpuset_spread_node(¤t->cpuset_mem_spread_rotor); } int cpuset_slab_spread_node(void) { |
778d3b0ff cpusets: randomiz... |
3395 3396 3397 |
if (current->cpuset_slab_spread_rotor == NUMA_NO_NODE) current->cpuset_slab_spread_rotor = node_random(¤t->mems_allowed); |
6adef3ebe cpusets: new roun... |
3398 3399 |
return cpuset_spread_node(¤t->cpuset_slab_spread_rotor); } |
825a46af5 [PATCH] cpuset me... |
3400 3401 3402 |
EXPORT_SYMBOL_GPL(cpuset_mem_spread_node); /** |
bbe373f2c oom: compare cpus... |
3403 3404 3405 3406 3407 3408 3409 3410 |
* cpuset_mems_allowed_intersects - Does @tsk1's mems_allowed intersect @tsk2's? * @tsk1: pointer to task_struct of some task. * @tsk2: pointer to task_struct of some other task. * * Description: Return true if @tsk1's mems_allowed intersects the * mems_allowed of @tsk2. Used by the OOM killer to determine if * one of the task's memory usage might impact the memory available * to the other. |
ef08e3b49 [PATCH] cpusets: ... |
3411 |
**/ |
bbe373f2c oom: compare cpus... |
3412 3413 |
int cpuset_mems_allowed_intersects(const struct task_struct *tsk1, const struct task_struct *tsk2) |
ef08e3b49 [PATCH] cpusets: ... |
3414 |
{ |
bbe373f2c oom: compare cpus... |
3415 |
return nodes_intersects(tsk1->mems_allowed, tsk2->mems_allowed); |
ef08e3b49 [PATCH] cpusets: ... |
3416 |
} |
75aa19941 oom: print trigge... |
3417 |
/** |
da39da3a5 mm, oom: remove t... |
3418 |
* cpuset_print_current_mems_allowed - prints current's cpuset and mems_allowed |
75aa19941 oom: print trigge... |
3419 |
* |
da39da3a5 mm, oom: remove t... |
3420 |
* Description: Prints current's name, cpuset name, and cached copy of its |
b8dadcb58 cpuset: use rcu_r... |
3421 |
* mems_allowed to the kernel log. |
75aa19941 oom: print trigge... |
3422 |
*/ |
da39da3a5 mm, oom: remove t... |
3423 |
void cpuset_print_current_mems_allowed(void) |
75aa19941 oom: print trigge... |
3424 |
{ |
b8dadcb58 cpuset: use rcu_r... |
3425 |
struct cgroup *cgrp; |
75aa19941 oom: print trigge... |
3426 |
|
b8dadcb58 cpuset: use rcu_r... |
3427 |
rcu_read_lock(); |
63f43f55c cpuset: fix cpuse... |
3428 |
|
da39da3a5 mm, oom: remove t... |
3429 |
cgrp = task_cs(current)->css.cgroup; |
ef8444ea0 mm, oom: reorgani... |
3430 |
pr_cont(",cpuset="); |
e61734c55 cgroup: remove cg... |
3431 |
pr_cont_cgroup_name(cgrp); |
ef8444ea0 mm, oom: reorgani... |
3432 |
pr_cont(",mems_allowed=%*pbl", |
da39da3a5 mm, oom: remove t... |
3433 |
nodemask_pr_args(¤t->mems_allowed)); |
f440d98f8 cpuset: use cgrou... |
3434 |
|
cfb5966be cpuset: fix RCU l... |
3435 |
rcu_read_unlock(); |
75aa19941 oom: print trigge... |
3436 |
} |
1da177e4c Linux-2.6.12-rc2 |
3437 |
/* |
3e0d98b9f [PATCH] cpuset: m... |
3438 3439 3440 3441 |
* Collection of memory_pressure is suppressed unless * this flag is enabled by writing "1" to the special * cpuset file 'memory_pressure_enabled' in the root cpuset. */ |
c5b2aff89 [PATCH] cpuset: m... |
3442 |
int cpuset_memory_pressure_enabled __read_mostly; |
3e0d98b9f [PATCH] cpuset: m... |
3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 |
/** * cpuset_memory_pressure_bump - keep stats of per-cpuset reclaims. * * Keep a running average of the rate of synchronous (direct) * page reclaim efforts initiated by tasks in each cpuset. * * This represents the rate at which some task in the cpuset * ran low on memory on all nodes it was allowed to use, and * had to enter the kernels page reclaim code in an effort to * create more free memory by tossing clean pages or swapping * or writing dirty pages. * * Display to user space in the per-cpuset read-only file * "memory_pressure". Value displayed is an integer * representing the recent rate of entry into the synchronous * (direct) page reclaim by any task attached to the cpuset. **/ void __cpuset_memory_pressure_bump(void) { |
b8dadcb58 cpuset: use rcu_r... |
3464 |
rcu_read_lock(); |
8793d854e Task Control Grou... |
3465 |
fmeter_markevent(&task_cs(current)->fmeter); |
b8dadcb58 cpuset: use rcu_r... |
3466 |
rcu_read_unlock(); |
3e0d98b9f [PATCH] cpuset: m... |
3467 |
} |
8793d854e Task Control Grou... |
3468 |
#ifdef CONFIG_PROC_PID_CPUSET |
3e0d98b9f [PATCH] cpuset: m... |
3469 |
/* |
1da177e4c Linux-2.6.12-rc2 |
3470 3471 3472 |
* proc_cpuset_show() * - Print tasks cpuset path into seq_file. * - Used for /proc/<pid>/cpuset. |
053199edf [PATCH] cpusets: ... |
3473 3474 |
* - No need to task_lock(tsk) on this tsk->cpuset reference, as it * doesn't really matter if tsk->cpuset changes after we read it, |
5d21cc2db cpuset: replace c... |
3475 |
* and we take cpuset_mutex, keeping cpuset_attach() from changing it |
2df167a30 cgroups: update c... |
3476 |
* anyway. |
1da177e4c Linux-2.6.12-rc2 |
3477 |
*/ |
52de4779f cpuset: simplify ... |
3478 3479 |
int proc_cpuset_show(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *tsk) |
1da177e4c Linux-2.6.12-rc2 |
3480 |
{ |
4c737b41d cgroup: make cgro... |
3481 |
char *buf; |
8793d854e Task Control Grou... |
3482 |
struct cgroup_subsys_state *css; |
99f895518 [PATCH] proc: don... |
3483 |
int retval; |
1da177e4c Linux-2.6.12-rc2 |
3484 |
|
99f895518 [PATCH] proc: don... |
3485 |
retval = -ENOMEM; |
e61734c55 cgroup: remove cg... |
3486 |
buf = kmalloc(PATH_MAX, GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
3487 |
if (!buf) |
99f895518 [PATCH] proc: don... |
3488 |
goto out; |
a79a908fd cgroup: introduce... |
3489 |
css = task_get_css(tsk, cpuset_cgrp_id); |
4c737b41d cgroup: make cgro... |
3490 3491 |
retval = cgroup_path_ns(css->cgroup, buf, PATH_MAX, current->nsproxy->cgroup_ns); |
a79a908fd cgroup: introduce... |
3492 |
css_put(css); |
4c737b41d cgroup: make cgro... |
3493 |
if (retval >= PATH_MAX) |
679a5e3f1 cpuset: fix error... |
3494 3495 |
retval = -ENAMETOOLONG; if (retval < 0) |
52de4779f cpuset: simplify ... |
3496 |
goto out_free; |
4c737b41d cgroup: make cgro... |
3497 |
seq_puts(m, buf); |
1da177e4c Linux-2.6.12-rc2 |
3498 3499 |
seq_putc(m, ' '); |
e61734c55 cgroup: remove cg... |
3500 |
retval = 0; |
99f895518 [PATCH] proc: don... |
3501 |
out_free: |
1da177e4c Linux-2.6.12-rc2 |
3502 |
kfree(buf); |
99f895518 [PATCH] proc: don... |
3503 |
out: |
1da177e4c Linux-2.6.12-rc2 |
3504 3505 |
return retval; } |
8793d854e Task Control Grou... |
3506 |
#endif /* CONFIG_PROC_PID_CPUSET */ |
1da177e4c Linux-2.6.12-rc2 |
3507 |
|
d01d48278 sched: Always sho... |
3508 |
/* Display task mems_allowed in /proc/<pid>/status file. */ |
df5f8314c proc: seqfile con... |
3509 3510 |
void cpuset_task_status_allowed(struct seq_file *m, struct task_struct *task) { |
e8e6d97c9 cpuset: use %*pb[... |
3511 3512 3513 3514 3515 3516 |
seq_printf(m, "Mems_allowed:\t%*pb ", nodemask_pr_args(&task->mems_allowed)); seq_printf(m, "Mems_allowed_list:\t%*pbl ", nodemask_pr_args(&task->mems_allowed)); |
1da177e4c Linux-2.6.12-rc2 |
3517 |
} |