Blame view
mm/mempolicy.c
71.3 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 3 4 |
/* * Simple NUMA memory policy for the Linux kernel. * * Copyright 2003,2004 Andi Kleen, SuSE Labs. |
8bccd85ff [PATCH] Implement... |
5 |
* (C) Copyright 2005 Christoph Lameter, Silicon Graphics, Inc. |
1da177e4c Linux-2.6.12-rc2 |
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
* Subject to the GNU Public License, version 2. * * NUMA policy allows the user to give hints in which node(s) memory should * be allocated. * * Support four policies per VMA and per process: * * The VMA policy has priority over the process policy for a page fault. * * interleave Allocate memory interleaved over a set of nodes, * with normal fallback if it fails. * For VMA based allocations this interleaves based on the * offset into the backing object or offset into the mapping * for anonymous memory. For process policy an process counter * is used. |
8bccd85ff [PATCH] Implement... |
21 |
* |
1da177e4c Linux-2.6.12-rc2 |
22 23 |
* bind Only allocate memory on a specific set of nodes, * no fallback. |
8bccd85ff [PATCH] Implement... |
24 25 26 27 |
* FIXME: memory is allocated starting with the first node * to the last. It would be better if bind would truly restrict * the allocation to memory nodes instead * |
1da177e4c Linux-2.6.12-rc2 |
28 |
* preferred Try a specific node first before normal fallback. |
00ef2d2f8 mm: use NUMA_NO_NODE |
29 |
* As a special case NUMA_NO_NODE here means do the allocation |
1da177e4c Linux-2.6.12-rc2 |
30 31 32 |
* on the local CPU. This is normally identical to default, * but useful to set in a VMA when you have a non default * process policy. |
8bccd85ff [PATCH] Implement... |
33 |
* |
1da177e4c Linux-2.6.12-rc2 |
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 |
* default Allocate on the local node first, or when on a VMA * use the process policy. This is what Linux always did * in a NUMA aware kernel and still does by, ahem, default. * * The process policy is applied for most non interrupt memory allocations * in that process' context. Interrupts ignore the policies and always * try to allocate on the local CPU. The VMA policy is only applied for memory * allocations for a VMA in the VM. * * Currently there are a few corner cases in swapping where the policy * is not applied, but the majority should be handled. When process policy * is used it is not remembered over swap outs/swap ins. * * Only the highest zone in the zone hierarchy gets policied. Allocations * requesting a lower zone just use default policy. This implies that * on systems with highmem kernel lowmem allocation don't get policied. * Same with GFP_DMA allocations. * * For shmfs/tmpfs/hugetlbfs shared memory the policy is shared between * all users and remembered even when nobody has memory mapped. */ /* Notebook: fix mmap readahead to honour policy and enable policy for any page cache object statistics for bigpages global policy for page cache? currently it uses process policy. Requires first item above. handle mremap for shared memory (currently ignored for the policy) grows down? make bind policy root only? It can trigger oom much faster and the kernel is not always grateful with that. |
1da177e4c Linux-2.6.12-rc2 |
66 |
*/ |
b1de0d139 mm: convert some ... |
67 |
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt |
1da177e4c Linux-2.6.12-rc2 |
68 69 70 71 72 73 |
#include <linux/mempolicy.h> #include <linux/mm.h> #include <linux/highmem.h> #include <linux/hugetlb.h> #include <linux/kernel.h> #include <linux/sched.h> |
1da177e4c Linux-2.6.12-rc2 |
74 75 |
#include <linux/nodemask.h> #include <linux/cpuset.h> |
1da177e4c Linux-2.6.12-rc2 |
76 77 |
#include <linux/slab.h> #include <linux/string.h> |
b95f1b31b mm: Map most file... |
78 |
#include <linux/export.h> |
b488893a3 pid namespaces: c... |
79 |
#include <linux/nsproxy.h> |
1da177e4c Linux-2.6.12-rc2 |
80 81 82 |
#include <linux/interrupt.h> #include <linux/init.h> #include <linux/compat.h> |
dc9aa5b9d [PATCH] Swap Migr... |
83 |
#include <linux/swap.h> |
1a75a6c82 [PATCH] Fold numa... |
84 85 |
#include <linux/seq_file.h> #include <linux/proc_fs.h> |
b20a35035 [PATCH] page migr... |
86 |
#include <linux/migrate.h> |
62b61f611 ksm: memory hotre... |
87 |
#include <linux/ksm.h> |
95a402c38 [PATCH] page migr... |
88 |
#include <linux/rmap.h> |
86c3a7645 [PATCH] SELinux: ... |
89 |
#include <linux/security.h> |
dbcb0f19c mm/mempolicy.c: c... |
90 |
#include <linux/syscalls.h> |
095f1fc4e mempolicy: rework... |
91 |
#include <linux/ctype.h> |
6d9c285a6 mm: move inc_zone... |
92 |
#include <linux/mm_inline.h> |
b24f53a0b mm: mempolicy: Ad... |
93 |
#include <linux/mmu_notifier.h> |
b1de0d139 mm: convert some ... |
94 |
#include <linux/printk.h> |
dc9aa5b9d [PATCH] Swap Migr... |
95 |
|
1da177e4c Linux-2.6.12-rc2 |
96 97 |
#include <asm/tlbflush.h> #include <asm/uaccess.h> |
62695a84e vmscan: move isol... |
98 |
#include "internal.h" |
38e35860d [PATCH] mempolici... |
99 |
/* Internal flags */ |
dc9aa5b9d [PATCH] Swap Migr... |
100 |
#define MPOL_MF_DISCONTIG_OK (MPOL_MF_INTERNAL << 0) /* Skip checks for continuous vmas */ |
38e35860d [PATCH] mempolici... |
101 |
#define MPOL_MF_INVERT (MPOL_MF_INTERNAL << 1) /* Invert check for nodemask */ |
dc9aa5b9d [PATCH] Swap Migr... |
102 |
|
fcc234f88 [PATCH] mm: kill ... |
103 104 |
static struct kmem_cache *policy_cache; static struct kmem_cache *sn_cache; |
1da177e4c Linux-2.6.12-rc2 |
105 |
|
1da177e4c Linux-2.6.12-rc2 |
106 107 |
/* Highest zone. An specific allocation for a zone below that is not policied. */ |
6267276f3 [PATCH] optional ... |
108 |
enum zone_type policy_zone = 0; |
1da177e4c Linux-2.6.12-rc2 |
109 |
|
bea904d54 mempolicy: use MP... |
110 111 112 |
/* * run-time system-wide default policy => local allocation */ |
e754d79d3 mm/mempolicy.c: q... |
113 |
static struct mempolicy default_policy = { |
1da177e4c Linux-2.6.12-rc2 |
114 |
.refcnt = ATOMIC_INIT(1), /* never free it */ |
bea904d54 mempolicy: use MP... |
115 |
.mode = MPOL_PREFERRED, |
fc36b8d3d mempolicy: use MP... |
116 |
.flags = MPOL_F_LOCAL, |
1da177e4c Linux-2.6.12-rc2 |
117 |
}; |
5606e3877 mm: numa: Migrate... |
118 |
static struct mempolicy preferred_node_policy[MAX_NUMNODES]; |
74d2c3a05 mempolicy: introd... |
119 |
struct mempolicy *get_task_policy(struct task_struct *p) |
5606e3877 mm: numa: Migrate... |
120 121 |
{ struct mempolicy *pol = p->mempolicy; |
f15ca78e3 mempolicy: change... |
122 |
int node; |
5606e3877 mm: numa: Migrate... |
123 |
|
f15ca78e3 mempolicy: change... |
124 125 |
if (pol) return pol; |
5606e3877 mm: numa: Migrate... |
126 |
|
f15ca78e3 mempolicy: change... |
127 128 129 130 131 132 |
node = numa_node_id(); if (node != NUMA_NO_NODE) { pol = &preferred_node_policy[node]; /* preferred_node_policy is not initialised early in boot */ if (pol->mode) return pol; |
5606e3877 mm: numa: Migrate... |
133 |
} |
f15ca78e3 mempolicy: change... |
134 |
return &default_policy; |
5606e3877 mm: numa: Migrate... |
135 |
} |
37012946d mempolicy: create... |
136 137 |
static const struct mempolicy_operations { int (*create)(struct mempolicy *pol, const nodemask_t *nodes); |
708c1bbc9 mempolicy: restru... |
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 |
/* * If read-side task has no lock to protect task->mempolicy, write-side * task will rebind the task->mempolicy by two step. The first step is * setting all the newly nodes, and the second step is cleaning all the * disallowed nodes. In this way, we can avoid finding no node to alloc * page. * If we have a lock to protect task->mempolicy in read-side, we do * rebind directly. * * step: * MPOL_REBIND_ONCE - do rebind work at once * MPOL_REBIND_STEP1 - set all the newly nodes * MPOL_REBIND_STEP2 - clean all the disallowed nodes */ void (*rebind)(struct mempolicy *pol, const nodemask_t *nodes, enum mpol_rebind_step step); |
37012946d mempolicy: create... |
154 |
} mpol_ops[MPOL_MAX]; |
f5b087b52 mempolicy: add MP... |
155 156 |
static inline int mpol_store_user_nodemask(const struct mempolicy *pol) { |
6d556294d mempolicy: remove... |
157 |
return pol->flags & MPOL_MODE_FLAGS; |
4c50bc011 mempolicy: add MP... |
158 159 160 161 162 163 164 165 |
} static void mpol_relative_nodemask(nodemask_t *ret, const nodemask_t *orig, const nodemask_t *rel) { nodemask_t tmp; nodes_fold(tmp, *orig, nodes_weight(*rel)); nodes_onto(*ret, tmp, *rel); |
f5b087b52 mempolicy: add MP... |
166 |
} |
37012946d mempolicy: create... |
167 168 169 170 171 172 173 174 175 176 177 |
static int mpol_new_interleave(struct mempolicy *pol, const nodemask_t *nodes) { if (nodes_empty(*nodes)) return -EINVAL; pol->v.nodes = *nodes; return 0; } static int mpol_new_preferred(struct mempolicy *pol, const nodemask_t *nodes) { if (!nodes) |
fc36b8d3d mempolicy: use MP... |
178 |
pol->flags |= MPOL_F_LOCAL; /* local allocation */ |
37012946d mempolicy: create... |
179 180 181 182 183 184 185 186 187 |
else if (nodes_empty(*nodes)) return -EINVAL; /* no allowed nodes */ else pol->v.preferred_node = first_node(*nodes); return 0; } static int mpol_new_bind(struct mempolicy *pol, const nodemask_t *nodes) { |
859f7ef14 mm/mempolicy.c: r... |
188 |
if (nodes_empty(*nodes)) |
37012946d mempolicy: create... |
189 190 191 192 |
return -EINVAL; pol->v.nodes = *nodes; return 0; } |
58568d2a8 cpuset,mm: update... |
193 194 195 196 197 198 199 200 201 |
/* * mpol_set_nodemask is called after mpol_new() to set up the nodemask, if * any, for the new policy. mpol_new() has already validated the nodes * parameter with respect to the policy mode and flags. But, we need to * handle an empty nodemask with MPOL_PREFERRED here. * * Must be called holding task's alloc_lock to protect task's mems_allowed * and mempolicy. May also be called holding the mmap_semaphore for write. */ |
4bfc44958 mm: make set_memp... |
202 203 |
static int mpol_set_nodemask(struct mempolicy *pol, const nodemask_t *nodes, struct nodemask_scratch *nsc) |
58568d2a8 cpuset,mm: update... |
204 |
{ |
58568d2a8 cpuset,mm: update... |
205 206 207 208 209 |
int ret; /* if mode is MPOL_DEFAULT, pol is NULL. This is right. */ if (pol == NULL) return 0; |
01f13bd60 mempolicy: use N_... |
210 |
/* Check N_MEMORY */ |
4bfc44958 mm: make set_memp... |
211 |
nodes_and(nsc->mask1, |
01f13bd60 mempolicy: use N_... |
212 |
cpuset_current_mems_allowed, node_states[N_MEMORY]); |
58568d2a8 cpuset,mm: update... |
213 214 215 216 217 218 |
VM_BUG_ON(!nodes); if (pol->mode == MPOL_PREFERRED && nodes_empty(*nodes)) nodes = NULL; /* explicit local allocation */ else { if (pol->flags & MPOL_F_RELATIVE_NODES) |
859f7ef14 mm/mempolicy.c: r... |
219 |
mpol_relative_nodemask(&nsc->mask2, nodes, &nsc->mask1); |
58568d2a8 cpuset,mm: update... |
220 |
else |
4bfc44958 mm: make set_memp... |
221 |
nodes_and(nsc->mask2, *nodes, nsc->mask1); |
58568d2a8 cpuset,mm: update... |
222 223 224 225 226 227 |
if (mpol_store_user_nodemask(pol)) pol->w.user_nodemask = *nodes; else pol->w.cpuset_mems_allowed = cpuset_current_mems_allowed; } |
4bfc44958 mm: make set_memp... |
228 229 230 231 |
if (nodes) ret = mpol_ops[pol->mode].create(pol, &nsc->mask2); else ret = mpol_ops[pol->mode].create(pol, NULL); |
58568d2a8 cpuset,mm: update... |
232 233 234 235 236 237 238 |
return ret; } /* * This function just creates a new policy, does some check and simple * initialization. You must invoke mpol_set_nodemask() to set nodes. */ |
028fec414 mempolicy: suppor... |
239 240 |
static struct mempolicy *mpol_new(unsigned short mode, unsigned short flags, nodemask_t *nodes) |
1da177e4c Linux-2.6.12-rc2 |
241 242 |
{ struct mempolicy *policy; |
028fec414 mempolicy: suppor... |
243 244 |
pr_debug("setting mode %d flags %d nodes[0] %lx ", |
00ef2d2f8 mm: use NUMA_NO_NODE |
245 |
mode, flags, nodes ? nodes_addr(*nodes)[0] : NUMA_NO_NODE); |
140d5a490 numa: mempolicy: ... |
246 |
|
3e1f06456 mempolicy: disall... |
247 248 |
if (mode == MPOL_DEFAULT) { if (nodes && !nodes_empty(*nodes)) |
37012946d mempolicy: create... |
249 |
return ERR_PTR(-EINVAL); |
d3a710337 mm: mempolicy: Ad... |
250 |
return NULL; |
37012946d mempolicy: create... |
251 |
} |
3e1f06456 mempolicy: disall... |
252 253 254 255 256 257 258 259 260 261 262 263 |
VM_BUG_ON(!nodes); /* * MPOL_PREFERRED cannot be used with MPOL_F_STATIC_NODES or * MPOL_F_RELATIVE_NODES if the nodemask is empty (local allocation). * All other modes require a valid pointer to a non-empty nodemask. */ if (mode == MPOL_PREFERRED) { if (nodes_empty(*nodes)) { if (((flags & MPOL_F_STATIC_NODES) || (flags & MPOL_F_RELATIVE_NODES))) return ERR_PTR(-EINVAL); |
3e1f06456 mempolicy: disall... |
264 |
} |
479e2802d mm: mempolicy: Ma... |
265 266 267 268 |
} else if (mode == MPOL_LOCAL) { if (!nodes_empty(*nodes)) return ERR_PTR(-EINVAL); mode = MPOL_PREFERRED; |
3e1f06456 mempolicy: disall... |
269 270 |
} else if (nodes_empty(*nodes)) return ERR_PTR(-EINVAL); |
1da177e4c Linux-2.6.12-rc2 |
271 272 273 274 |
policy = kmem_cache_alloc(policy_cache, GFP_KERNEL); if (!policy) return ERR_PTR(-ENOMEM); atomic_set(&policy->refcnt, 1); |
45c4745af mempolicy: rename... |
275 |
policy->mode = mode; |
3e1f06456 mempolicy: disall... |
276 |
policy->flags = flags; |
37012946d mempolicy: create... |
277 |
|
1da177e4c Linux-2.6.12-rc2 |
278 |
return policy; |
37012946d mempolicy: create... |
279 |
} |
52cd3b074 mempolicy: rework... |
280 281 282 283 284 |
/* Slow path of a mpol destructor. */ void __mpol_put(struct mempolicy *p) { if (!atomic_dec_and_test(&p->refcnt)) return; |
52cd3b074 mempolicy: rework... |
285 286 |
kmem_cache_free(policy_cache, p); } |
708c1bbc9 mempolicy: restru... |
287 288 |
static void mpol_rebind_default(struct mempolicy *pol, const nodemask_t *nodes, enum mpol_rebind_step step) |
37012946d mempolicy: create... |
289 290 |
{ } |
708c1bbc9 mempolicy: restru... |
291 292 293 294 295 296 297 298 |
/* * step: * MPOL_REBIND_ONCE - do rebind work at once * MPOL_REBIND_STEP1 - set all the newly nodes * MPOL_REBIND_STEP2 - clean all the disallowed nodes */ static void mpol_rebind_nodemask(struct mempolicy *pol, const nodemask_t *nodes, enum mpol_rebind_step step) |
37012946d mempolicy: create... |
299 300 301 302 303 304 305 306 |
{ nodemask_t tmp; if (pol->flags & MPOL_F_STATIC_NODES) nodes_and(tmp, pol->w.user_nodemask, *nodes); else if (pol->flags & MPOL_F_RELATIVE_NODES) mpol_relative_nodemask(&tmp, &pol->w.user_nodemask, nodes); else { |
708c1bbc9 mempolicy: restru... |
307 308 309 310 311 312 313 314 315 316 317 318 319 |
/* * if step == 1, we use ->w.cpuset_mems_allowed to cache the * result */ if (step == MPOL_REBIND_ONCE || step == MPOL_REBIND_STEP1) { nodes_remap(tmp, pol->v.nodes, pol->w.cpuset_mems_allowed, *nodes); pol->w.cpuset_mems_allowed = step ? tmp : *nodes; } else if (step == MPOL_REBIND_STEP2) { tmp = pol->w.cpuset_mems_allowed; pol->w.cpuset_mems_allowed = *nodes; } else BUG(); |
37012946d mempolicy: create... |
320 |
} |
f5b087b52 mempolicy: add MP... |
321 |
|
708c1bbc9 mempolicy: restru... |
322 323 324 325 326 327 328 329 330 |
if (nodes_empty(tmp)) tmp = *nodes; if (step == MPOL_REBIND_STEP1) nodes_or(pol->v.nodes, pol->v.nodes, tmp); else if (step == MPOL_REBIND_ONCE || step == MPOL_REBIND_STEP2) pol->v.nodes = tmp; else BUG(); |
37012946d mempolicy: create... |
331 |
if (!node_isset(current->il_next, tmp)) { |
0edaf86cf include/linux/nod... |
332 |
current->il_next = next_node_in(current->il_next, tmp); |
37012946d mempolicy: create... |
333 334 335 336 337 338 |
if (current->il_next >= MAX_NUMNODES) current->il_next = numa_node_id(); } } static void mpol_rebind_preferred(struct mempolicy *pol, |
708c1bbc9 mempolicy: restru... |
339 340 |
const nodemask_t *nodes, enum mpol_rebind_step step) |
37012946d mempolicy: create... |
341 342 |
{ nodemask_t tmp; |
37012946d mempolicy: create... |
343 344 |
if (pol->flags & MPOL_F_STATIC_NODES) { int node = first_node(pol->w.user_nodemask); |
fc36b8d3d mempolicy: use MP... |
345 |
if (node_isset(node, *nodes)) { |
37012946d mempolicy: create... |
346 |
pol->v.preferred_node = node; |
fc36b8d3d mempolicy: use MP... |
347 348 349 |
pol->flags &= ~MPOL_F_LOCAL; } else pol->flags |= MPOL_F_LOCAL; |
37012946d mempolicy: create... |
350 351 352 |
} else if (pol->flags & MPOL_F_RELATIVE_NODES) { mpol_relative_nodemask(&tmp, &pol->w.user_nodemask, nodes); pol->v.preferred_node = first_node(tmp); |
fc36b8d3d mempolicy: use MP... |
353 |
} else if (!(pol->flags & MPOL_F_LOCAL)) { |
37012946d mempolicy: create... |
354 355 356 357 358 |
pol->v.preferred_node = node_remap(pol->v.preferred_node, pol->w.cpuset_mems_allowed, *nodes); pol->w.cpuset_mems_allowed = *nodes; } |
1da177e4c Linux-2.6.12-rc2 |
359 |
} |
708c1bbc9 mempolicy: restru... |
360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 |
/* * mpol_rebind_policy - Migrate a policy to a different set of nodes * * If read-side task has no lock to protect task->mempolicy, write-side * task will rebind the task->mempolicy by two step. The first step is * setting all the newly nodes, and the second step is cleaning all the * disallowed nodes. In this way, we can avoid finding no node to alloc * page. * If we have a lock to protect task->mempolicy in read-side, we do * rebind directly. * * step: * MPOL_REBIND_ONCE - do rebind work at once * MPOL_REBIND_STEP1 - set all the newly nodes * MPOL_REBIND_STEP2 - clean all the disallowed nodes */ static void mpol_rebind_policy(struct mempolicy *pol, const nodemask_t *newmask, enum mpol_rebind_step step) |
1d0d2680a mempolicy: move r... |
378 |
{ |
1d0d2680a mempolicy: move r... |
379 380 |
if (!pol) return; |
89c522c78 mm/mempolicy.c: u... |
381 |
if (!mpol_store_user_nodemask(pol) && step == MPOL_REBIND_ONCE && |
1d0d2680a mempolicy: move r... |
382 383 |
nodes_equal(pol->w.cpuset_mems_allowed, *newmask)) return; |
708c1bbc9 mempolicy: restru... |
384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 |
if (step == MPOL_REBIND_STEP1 && (pol->flags & MPOL_F_REBINDING)) return; if (step == MPOL_REBIND_STEP2 && !(pol->flags & MPOL_F_REBINDING)) BUG(); if (step == MPOL_REBIND_STEP1) pol->flags |= MPOL_F_REBINDING; else if (step == MPOL_REBIND_STEP2) pol->flags &= ~MPOL_F_REBINDING; else if (step >= MPOL_REBIND_NSTEP) BUG(); mpol_ops[pol->mode].rebind(pol, newmask, step); |
1d0d2680a mempolicy: move r... |
399 400 401 402 403 |
} /* * Wrapper for mpol_rebind_policy() that just requires task * pointer, and updates task mempolicy. |
58568d2a8 cpuset,mm: update... |
404 405 |
* * Called with task's alloc_lock held. |
1d0d2680a mempolicy: move r... |
406 |
*/ |
708c1bbc9 mempolicy: restru... |
407 408 |
void mpol_rebind_task(struct task_struct *tsk, const nodemask_t *new, enum mpol_rebind_step step) |
1d0d2680a mempolicy: move r... |
409 |
{ |
708c1bbc9 mempolicy: restru... |
410 |
mpol_rebind_policy(tsk->mempolicy, new, step); |
1d0d2680a mempolicy: move r... |
411 412 413 414 415 416 417 418 419 420 421 422 423 424 |
} /* * Rebind each vma in mm to new nodemask. * * Call holding a reference to mm. Takes mm->mmap_sem during call. */ void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new) { struct vm_area_struct *vma; down_write(&mm->mmap_sem); for (vma = mm->mmap; vma; vma = vma->vm_next) |
708c1bbc9 mempolicy: restru... |
425 |
mpol_rebind_policy(vma->vm_policy, new, MPOL_REBIND_ONCE); |
1d0d2680a mempolicy: move r... |
426 427 |
up_write(&mm->mmap_sem); } |
37012946d mempolicy: create... |
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 |
static const struct mempolicy_operations mpol_ops[MPOL_MAX] = { [MPOL_DEFAULT] = { .rebind = mpol_rebind_default, }, [MPOL_INTERLEAVE] = { .create = mpol_new_interleave, .rebind = mpol_rebind_nodemask, }, [MPOL_PREFERRED] = { .create = mpol_new_preferred, .rebind = mpol_rebind_preferred, }, [MPOL_BIND] = { .create = mpol_new_bind, .rebind = mpol_rebind_nodemask, }, }; |
fc3012896 [PATCH] Simplify ... |
445 446 |
static void migrate_page_add(struct page *page, struct list_head *pagelist, unsigned long flags); |
1a75a6c82 [PATCH] Fold numa... |
447 |
|
6f4576e36 mempolicy: apply ... |
448 449 450 451 452 453 |
struct queue_pages { struct list_head *pagelist; unsigned long flags; nodemask_t *nmask; struct vm_area_struct *prev; }; |
980949457 mm/mempolicy: ren... |
454 455 456 457 |
/* * Scan through pages checking if pages follow certain conditions, * and move them to the pagelist if they do. */ |
6f4576e36 mempolicy: apply ... |
458 459 |
static int queue_pages_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) |
1da177e4c Linux-2.6.12-rc2 |
460 |
{ |
6f4576e36 mempolicy: apply ... |
461 462 463 464 |
struct vm_area_struct *vma = walk->vma; struct page *page; struct queue_pages *qp = walk->private; unsigned long flags = qp->flags; |
248db92da migrate_pages: tr... |
465 |
int nid, ret; |
91612e0df [PATCH] mbind: ch... |
466 |
pte_t *pte; |
705e87c0c [PATCH] mm: pte_o... |
467 |
spinlock_t *ptl; |
941150a32 [PATCH] mbind: fi... |
468 |
|
248db92da migrate_pages: tr... |
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 |
if (pmd_trans_huge(*pmd)) { ptl = pmd_lock(walk->mm, pmd); if (pmd_trans_huge(*pmd)) { page = pmd_page(*pmd); if (is_huge_zero_page(page)) { spin_unlock(ptl); split_huge_pmd(vma, pmd, addr); } else { get_page(page); spin_unlock(ptl); lock_page(page); ret = split_huge_page(page); unlock_page(page); put_page(page); if (ret) return 0; } } else { spin_unlock(ptl); } } |
91612e0df [PATCH] mbind: ch... |
490 |
|
337d9abf1 mm: thp: check pm... |
491 492 |
if (pmd_trans_unstable(pmd)) return 0; |
248db92da migrate_pages: tr... |
493 |
retry: |
6f4576e36 mempolicy: apply ... |
494 495 |
pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl); for (; addr != end; pte++, addr += PAGE_SIZE) { |
91612e0df [PATCH] mbind: ch... |
496 |
if (!pte_present(*pte)) |
1da177e4c Linux-2.6.12-rc2 |
497 |
continue; |
6aab341e0 mm: re-architect ... |
498 499 |
page = vm_normal_page(vma, addr, *pte); if (!page) |
1da177e4c Linux-2.6.12-rc2 |
500 |
continue; |
053837fce [PATCH] mm: migra... |
501 |
/* |
62b61f611 ksm: memory hotre... |
502 503 |
* vm_normal_page() filters out zero pages, but there might * still be PageReserved pages to skip, perhaps in a VDSO. |
053837fce [PATCH] mm: migra... |
504 |
*/ |
b79bc0a0c ksm: enable KSM p... |
505 |
if (PageReserved(page)) |
f4598c8b3 [PATCH] migration... |
506 |
continue; |
6aab341e0 mm: re-architect ... |
507 |
nid = page_to_nid(page); |
6f4576e36 mempolicy: apply ... |
508 |
if (node_isset(nid, *qp->nmask) == !!(flags & MPOL_MF_INVERT)) |
38e35860d [PATCH] mempolici... |
509 |
continue; |
800d8c63b shmem: add huge p... |
510 |
if (PageTransCompound(page)) { |
248db92da migrate_pages: tr... |
511 512 513 514 515 516 517 518 519 520 521 522 523 524 |
get_page(page); pte_unmap_unlock(pte, ptl); lock_page(page); ret = split_huge_page(page); unlock_page(page); put_page(page); /* Failed to split -- skip. */ if (ret) { pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl); continue; } goto retry; } |
38e35860d [PATCH] mempolici... |
525 |
|
77bf45e78 mempolicy: do not... |
526 |
migrate_page_add(page, qp->pagelist, flags); |
6f4576e36 mempolicy: apply ... |
527 528 529 530 |
} pte_unmap_unlock(pte - 1, ptl); cond_resched(); return 0; |
91612e0df [PATCH] mbind: ch... |
531 |
} |
6f4576e36 mempolicy: apply ... |
532 533 534 |
static int queue_pages_hugetlb(pte_t *pte, unsigned long hmask, unsigned long addr, unsigned long end, struct mm_walk *walk) |
e2d8cf405 migrate: add huge... |
535 536 |
{ #ifdef CONFIG_HUGETLB_PAGE |
6f4576e36 mempolicy: apply ... |
537 538 |
struct queue_pages *qp = walk->private; unsigned long flags = qp->flags; |
e2d8cf405 migrate: add huge... |
539 540 |
int nid; struct page *page; |
cb900f412 mm, hugetlb: conv... |
541 |
spinlock_t *ptl; |
d4c54919e mm: add !pte_pres... |
542 |
pte_t entry; |
e2d8cf405 migrate: add huge... |
543 |
|
6f4576e36 mempolicy: apply ... |
544 545 |
ptl = huge_pte_lock(hstate_vma(walk->vma), walk->mm, pte); entry = huge_ptep_get(pte); |
d4c54919e mm: add !pte_pres... |
546 547 548 |
if (!pte_present(entry)) goto unlock; page = pte_page(entry); |
e2d8cf405 migrate: add huge... |
549 |
nid = page_to_nid(page); |
6f4576e36 mempolicy: apply ... |
550 |
if (node_isset(nid, *qp->nmask) == !!(flags & MPOL_MF_INVERT)) |
e2d8cf405 migrate: add huge... |
551 552 553 554 |
goto unlock; /* With MPOL_MF_MOVE, we migrate only unshared hugepage. */ if (flags & (MPOL_MF_MOVE_ALL) || (flags & MPOL_MF_MOVE && page_mapcount(page) == 1)) |
6f4576e36 mempolicy: apply ... |
555 |
isolate_huge_page(page, qp->pagelist); |
e2d8cf405 migrate: add huge... |
556 |
unlock: |
cb900f412 mm, hugetlb: conv... |
557 |
spin_unlock(ptl); |
e2d8cf405 migrate: add huge... |
558 559 560 |
#else BUG(); #endif |
91612e0df [PATCH] mbind: ch... |
561 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
562 |
} |
5877231f6 mm: Move change_p... |
563 |
#ifdef CONFIG_NUMA_BALANCING |
b24f53a0b mm: mempolicy: Ad... |
564 |
/* |
4b10e7d56 mm: mempolicy: Im... |
565 566 567 568 569 570 571 |
* This is used to mark a range of virtual addresses to be inaccessible. * These are later cleared by a NUMA hinting fault. Depending on these * faults, pages may be migrated for better NUMA placement. * * This is assuming that NUMA faults are handled using PROT_NONE. If * an architecture makes a different choice, it will need further * changes to the core. |
b24f53a0b mm: mempolicy: Ad... |
572 |
*/ |
4b10e7d56 mm: mempolicy: Im... |
573 574 |
unsigned long change_prot_numa(struct vm_area_struct *vma, unsigned long addr, unsigned long end) |
b24f53a0b mm: mempolicy: Ad... |
575 |
{ |
4b10e7d56 mm: mempolicy: Im... |
576 |
int nr_updated; |
b24f53a0b mm: mempolicy: Ad... |
577 |
|
4d9424669 mm: convert p[te|... |
578 |
nr_updated = change_protection(vma, addr, end, PAGE_NONE, 0, 1); |
03c5a6e16 mm: numa: Add pte... |
579 580 |
if (nr_updated) count_vm_numa_events(NUMA_PTE_UPDATES, nr_updated); |
b24f53a0b mm: mempolicy: Ad... |
581 |
|
4b10e7d56 mm: mempolicy: Im... |
582 |
return nr_updated; |
b24f53a0b mm: mempolicy: Ad... |
583 584 585 586 587 588 589 |
} #else static unsigned long change_prot_numa(struct vm_area_struct *vma, unsigned long addr, unsigned long end) { return 0; } |
5877231f6 mm: Move change_p... |
590 |
#endif /* CONFIG_NUMA_BALANCING */ |
b24f53a0b mm: mempolicy: Ad... |
591 |
|
6f4576e36 mempolicy: apply ... |
592 593 594 595 596 597 598 |
static int queue_pages_test_walk(unsigned long start, unsigned long end, struct mm_walk *walk) { struct vm_area_struct *vma = walk->vma; struct queue_pages *qp = walk->private; unsigned long endvma = vma->vm_end; unsigned long flags = qp->flags; |
77bf45e78 mempolicy: do not... |
599 |
if (!vma_migratable(vma)) |
48684a65b mm: pagewalk: fix... |
600 |
return 1; |
6f4576e36 mempolicy: apply ... |
601 602 603 604 605 606 607 608 609 610 611 612 613 |
if (endvma > end) endvma = end; if (vma->vm_start > start) start = vma->vm_start; if (!(flags & MPOL_MF_DISCONTIG_OK)) { if (!vma->vm_next && vma->vm_end < end) return -EFAULT; if (qp->prev && qp->prev->vm_end < vma->vm_start) return -EFAULT; } qp->prev = vma; |
6f4576e36 mempolicy: apply ... |
614 615 |
if (flags & MPOL_MF_LAZY) { /* Similar to task_numa_work, skip inaccessible VMAs */ |
4355c018c mm/mempolicy.c: s... |
616 617 618 |
if (!is_vm_hugetlb_page(vma) && (vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)) && !(vma->vm_flags & VM_MIXEDMAP)) |
6f4576e36 mempolicy: apply ... |
619 620 621 |
change_prot_numa(vma, start, endvma); return 1; } |
77bf45e78 mempolicy: do not... |
622 623 |
/* queue pages from current vma */ if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) |
6f4576e36 mempolicy: apply ... |
624 625 626 |
return 0; return 1; } |
dc9aa5b9d [PATCH] Swap Migr... |
627 |
/* |
980949457 mm/mempolicy: ren... |
628 629 630 631 632 |
* Walk through page tables and collect pages to be migrated. * * If pages found in a given range are on a set of nodes (determined by * @nodes and @flags,) it's isolated and queued to the pagelist which is * passed via @private.) |
dc9aa5b9d [PATCH] Swap Migr... |
633 |
*/ |
d05f0cdcb mm: fix crashes f... |
634 |
static int |
980949457 mm/mempolicy: ren... |
635 |
queue_pages_range(struct mm_struct *mm, unsigned long start, unsigned long end, |
6f4576e36 mempolicy: apply ... |
636 637 |
nodemask_t *nodes, unsigned long flags, struct list_head *pagelist) |
1da177e4c Linux-2.6.12-rc2 |
638 |
{ |
6f4576e36 mempolicy: apply ... |
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 |
struct queue_pages qp = { .pagelist = pagelist, .flags = flags, .nmask = nodes, .prev = NULL, }; struct mm_walk queue_pages_walk = { .hugetlb_entry = queue_pages_hugetlb, .pmd_entry = queue_pages_pte_range, .test_walk = queue_pages_test_walk, .mm = mm, .private = &qp, }; return walk_page_range(start, end, &queue_pages_walk); |
1da177e4c Linux-2.6.12-rc2 |
654 |
} |
869833f2c mempolicy: remove... |
655 656 657 658 659 660 |
/* * Apply policy to a single VMA * This must be called with the mmap_sem held for writing. */ static int vma_replace_policy(struct vm_area_struct *vma, struct mempolicy *pol) |
8d34694c1 revert "mm: mempo... |
661 |
{ |
869833f2c mempolicy: remove... |
662 663 664 |
int err; struct mempolicy *old; struct mempolicy *new; |
8d34694c1 revert "mm: mempo... |
665 666 667 668 669 670 |
pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p ", vma->vm_start, vma->vm_end, vma->vm_pgoff, vma->vm_ops, vma->vm_file, vma->vm_ops ? vma->vm_ops->set_policy : NULL); |
869833f2c mempolicy: remove... |
671 672 673 674 675 |
new = mpol_dup(pol); if (IS_ERR(new)) return PTR_ERR(new); if (vma->vm_ops && vma->vm_ops->set_policy) { |
8d34694c1 revert "mm: mempo... |
676 |
err = vma->vm_ops->set_policy(vma, new); |
869833f2c mempolicy: remove... |
677 678 |
if (err) goto err_out; |
8d34694c1 revert "mm: mempo... |
679 |
} |
869833f2c mempolicy: remove... |
680 681 682 683 684 685 686 687 |
old = vma->vm_policy; vma->vm_policy = new; /* protected by mmap_sem */ mpol_put(old); return 0; err_out: mpol_put(new); |
8d34694c1 revert "mm: mempo... |
688 689 |
return err; } |
1da177e4c Linux-2.6.12-rc2 |
690 |
/* Step 2: apply policy to a range and do splits. */ |
9d8cebd4b mm: fix mbind vma... |
691 692 |
static int mbind_range(struct mm_struct *mm, unsigned long start, unsigned long end, struct mempolicy *new_pol) |
1da177e4c Linux-2.6.12-rc2 |
693 694 |
{ struct vm_area_struct *next; |
9d8cebd4b mm: fix mbind vma... |
695 696 697 |
struct vm_area_struct *prev; struct vm_area_struct *vma; int err = 0; |
e26a51148 mm/mempolicy.c: r... |
698 |
pgoff_t pgoff; |
9d8cebd4b mm: fix mbind vma... |
699 700 |
unsigned long vmstart; unsigned long vmend; |
1da177e4c Linux-2.6.12-rc2 |
701 |
|
097d59106 vm: avoid using f... |
702 |
vma = find_vma(mm, start); |
9d8cebd4b mm: fix mbind vma... |
703 704 |
if (!vma || vma->vm_start > start) return -EFAULT; |
097d59106 vm: avoid using f... |
705 |
prev = vma->vm_prev; |
e26a51148 mm/mempolicy.c: r... |
706 707 |
if (start > vma->vm_start) prev = vma; |
9d8cebd4b mm: fix mbind vma... |
708 |
for (; vma && vma->vm_start < end; prev = vma, vma = next) { |
1da177e4c Linux-2.6.12-rc2 |
709 |
next = vma->vm_next; |
9d8cebd4b mm: fix mbind vma... |
710 711 |
vmstart = max(start, vma->vm_start); vmend = min(end, vma->vm_end); |
e26a51148 mm/mempolicy.c: r... |
712 713 714 715 716 |
if (mpol_equal(vma_policy(vma), new_pol)) continue; pgoff = vma->vm_pgoff + ((vmstart - vma->vm_start) >> PAGE_SHIFT); |
9d8cebd4b mm: fix mbind vma... |
717 |
prev = vma_merge(mm, prev, vmstart, vmend, vma->vm_flags, |
19a809afe userfaultfd: teac... |
718 719 |
vma->anon_vma, vma->vm_file, pgoff, new_pol, vma->vm_userfaultfd_ctx); |
9d8cebd4b mm: fix mbind vma... |
720 721 722 |
if (prev) { vma = prev; next = vma->vm_next; |
3964acd0d mm: mempolicy: fi... |
723 724 725 726 |
if (mpol_equal(vma_policy(vma), new_pol)) continue; /* vma_merge() joined vma && vma->next, case 8 */ goto replace; |
9d8cebd4b mm: fix mbind vma... |
727 728 729 730 731 732 733 734 735 736 737 |
} if (vma->vm_start != vmstart) { err = split_vma(vma->vm_mm, vma, vmstart, 1); if (err) goto out; } if (vma->vm_end != vmend) { err = split_vma(vma->vm_mm, vma, vmend, 0); if (err) goto out; } |
3964acd0d mm: mempolicy: fi... |
738 |
replace: |
869833f2c mempolicy: remove... |
739 |
err = vma_replace_policy(vma, new_pol); |
8d34694c1 revert "mm: mempo... |
740 741 |
if (err) goto out; |
1da177e4c Linux-2.6.12-rc2 |
742 |
} |
9d8cebd4b mm: fix mbind vma... |
743 744 |
out: |
1da177e4c Linux-2.6.12-rc2 |
745 746 |
return err; } |
1da177e4c Linux-2.6.12-rc2 |
747 |
/* Set the process memory policy */ |
028fec414 mempolicy: suppor... |
748 749 |
static long do_set_mempolicy(unsigned short mode, unsigned short flags, nodemask_t *nodes) |
1da177e4c Linux-2.6.12-rc2 |
750 |
{ |
58568d2a8 cpuset,mm: update... |
751 |
struct mempolicy *new, *old; |
4bfc44958 mm: make set_memp... |
752 |
NODEMASK_SCRATCH(scratch); |
58568d2a8 cpuset,mm: update... |
753 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
754 |
|
4bfc44958 mm: make set_memp... |
755 756 |
if (!scratch) return -ENOMEM; |
f4e53d910 mempolicy: write ... |
757 |
|
4bfc44958 mm: make set_memp... |
758 759 760 761 762 |
new = mpol_new(mode, flags, nodes); if (IS_ERR(new)) { ret = PTR_ERR(new); goto out; } |
2c7c3a7d0 mempolicy: kill d... |
763 |
|
58568d2a8 cpuset,mm: update... |
764 |
task_lock(current); |
4bfc44958 mm: make set_memp... |
765 |
ret = mpol_set_nodemask(new, nodes, scratch); |
58568d2a8 cpuset,mm: update... |
766 767 |
if (ret) { task_unlock(current); |
58568d2a8 cpuset,mm: update... |
768 |
mpol_put(new); |
4bfc44958 mm: make set_memp... |
769 |
goto out; |
58568d2a8 cpuset,mm: update... |
770 771 |
} old = current->mempolicy; |
1da177e4c Linux-2.6.12-rc2 |
772 |
current->mempolicy = new; |
45c4745af mempolicy: rename... |
773 |
if (new && new->mode == MPOL_INTERLEAVE && |
f5b087b52 mempolicy: add MP... |
774 |
nodes_weight(new->v.nodes)) |
dfcd3c0dc [PATCH] Convert m... |
775 |
current->il_next = first_node(new->v.nodes); |
58568d2a8 cpuset,mm: update... |
776 |
task_unlock(current); |
58568d2a8 cpuset,mm: update... |
777 |
mpol_put(old); |
4bfc44958 mm: make set_memp... |
778 779 780 781 |
ret = 0; out: NODEMASK_SCRATCH_FREE(scratch); return ret; |
1da177e4c Linux-2.6.12-rc2 |
782 |
} |
bea904d54 mempolicy: use MP... |
783 784 |
/* * Return nodemask for policy for get_mempolicy() query |
58568d2a8 cpuset,mm: update... |
785 786 |
* * Called with task's alloc_lock held |
bea904d54 mempolicy: use MP... |
787 788 |
*/ static void get_policy_nodemask(struct mempolicy *p, nodemask_t *nodes) |
1da177e4c Linux-2.6.12-rc2 |
789 |
{ |
dfcd3c0dc [PATCH] Convert m... |
790 |
nodes_clear(*nodes); |
bea904d54 mempolicy: use MP... |
791 792 |
if (p == &default_policy) return; |
45c4745af mempolicy: rename... |
793 |
switch (p->mode) { |
19770b326 mm: filter based ... |
794 795 |
case MPOL_BIND: /* Fall through */ |
1da177e4c Linux-2.6.12-rc2 |
796 |
case MPOL_INTERLEAVE: |
dfcd3c0dc [PATCH] Convert m... |
797 |
*nodes = p->v.nodes; |
1da177e4c Linux-2.6.12-rc2 |
798 799 |
break; case MPOL_PREFERRED: |
fc36b8d3d mempolicy: use MP... |
800 |
if (!(p->flags & MPOL_F_LOCAL)) |
dfcd3c0dc [PATCH] Convert m... |
801 |
node_set(p->v.preferred_node, *nodes); |
53f2556b6 mempolicy: mPOL_P... |
802 |
/* else return empty node mask for local allocation */ |
1da177e4c Linux-2.6.12-rc2 |
803 804 805 806 807 |
break; default: BUG(); } } |
d4edcf0d5 mm/gup: Switch al... |
808 |
static int lookup_node(unsigned long addr) |
1da177e4c Linux-2.6.12-rc2 |
809 810 811 |
{ struct page *p; int err; |
768ae309a mm: replace get_u... |
812 |
err = get_user_pages(addr & PAGE_MASK, 1, 0, &p, NULL); |
1da177e4c Linux-2.6.12-rc2 |
813 814 815 816 817 818 |
if (err >= 0) { err = page_to_nid(p); put_page(p); } return err; } |
1da177e4c Linux-2.6.12-rc2 |
819 |
/* Retrieve NUMA policy */ |
dbcb0f19c mm/mempolicy.c: c... |
820 821 |
static long do_get_mempolicy(int *policy, nodemask_t *nmask, unsigned long addr, unsigned long flags) |
1da177e4c Linux-2.6.12-rc2 |
822 |
{ |
8bccd85ff [PATCH] Implement... |
823 |
int err; |
1da177e4c Linux-2.6.12-rc2 |
824 825 826 |
struct mm_struct *mm = current->mm; struct vm_area_struct *vma = NULL; struct mempolicy *pol = current->mempolicy; |
754af6f5a Mem Policy: add M... |
827 828 |
if (flags & ~(unsigned long)(MPOL_F_NODE|MPOL_F_ADDR|MPOL_F_MEMS_ALLOWED)) |
1da177e4c Linux-2.6.12-rc2 |
829 |
return -EINVAL; |
754af6f5a Mem Policy: add M... |
830 831 832 833 834 |
if (flags & MPOL_F_MEMS_ALLOWED) { if (flags & (MPOL_F_NODE|MPOL_F_ADDR)) return -EINVAL; *policy = 0; /* just so it's initialized */ |
58568d2a8 cpuset,mm: update... |
835 |
task_lock(current); |
754af6f5a Mem Policy: add M... |
836 |
*nmask = cpuset_current_mems_allowed; |
58568d2a8 cpuset,mm: update... |
837 |
task_unlock(current); |
754af6f5a Mem Policy: add M... |
838 839 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
840 |
if (flags & MPOL_F_ADDR) { |
bea904d54 mempolicy: use MP... |
841 842 843 844 845 |
/* * Do NOT fall back to task policy if the * vma/shared policy at addr is NULL. We * want to return MPOL_DEFAULT in this case. */ |
1da177e4c Linux-2.6.12-rc2 |
846 847 848 849 850 851 852 853 854 855 856 857 858 859 |
down_read(&mm->mmap_sem); vma = find_vma_intersection(mm, addr, addr+1); if (!vma) { up_read(&mm->mmap_sem); return -EFAULT; } if (vma->vm_ops && vma->vm_ops->get_policy) pol = vma->vm_ops->get_policy(vma, addr); else pol = vma->vm_policy; } else if (addr) return -EINVAL; if (!pol) |
bea904d54 mempolicy: use MP... |
860 |
pol = &default_policy; /* indicates default behavior */ |
1da177e4c Linux-2.6.12-rc2 |
861 862 863 |
if (flags & MPOL_F_NODE) { if (flags & MPOL_F_ADDR) { |
d4edcf0d5 mm/gup: Switch al... |
864 |
err = lookup_node(addr); |
1da177e4c Linux-2.6.12-rc2 |
865 866 |
if (err < 0) goto out; |
8bccd85ff [PATCH] Implement... |
867 |
*policy = err; |
1da177e4c Linux-2.6.12-rc2 |
868 |
} else if (pol == current->mempolicy && |
45c4745af mempolicy: rename... |
869 |
pol->mode == MPOL_INTERLEAVE) { |
8bccd85ff [PATCH] Implement... |
870 |
*policy = current->il_next; |
1da177e4c Linux-2.6.12-rc2 |
871 872 873 874 |
} else { err = -EINVAL; goto out; } |
bea904d54 mempolicy: use MP... |
875 876 877 |
} else { *policy = pol == &default_policy ? MPOL_DEFAULT : pol->mode; |
d79df630f mempolicy: mask o... |
878 879 880 881 882 |
/* * Internal mempolicy flags must be masked off before exposing * the policy to userspace. */ *policy |= (pol->flags & MPOL_MODE_FLAGS); |
bea904d54 mempolicy: use MP... |
883 |
} |
1da177e4c Linux-2.6.12-rc2 |
884 885 886 887 888 |
if (vma) { up_read(¤t->mm->mmap_sem); vma = NULL; } |
1da177e4c Linux-2.6.12-rc2 |
889 |
err = 0; |
58568d2a8 cpuset,mm: update... |
890 |
if (nmask) { |
c6b6ef8bb mempolicy: fix ge... |
891 892 893 894 895 896 897 |
if (mpol_store_user_nodemask(pol)) { *nmask = pol->w.user_nodemask; } else { task_lock(current); get_policy_nodemask(pol, nmask); task_unlock(current); } |
58568d2a8 cpuset,mm: update... |
898 |
} |
1da177e4c Linux-2.6.12-rc2 |
899 900 |
out: |
52cd3b074 mempolicy: rework... |
901 |
mpol_cond_put(pol); |
1da177e4c Linux-2.6.12-rc2 |
902 903 904 905 |
if (vma) up_read(¤t->mm->mmap_sem); return err; } |
b20a35035 [PATCH] page migr... |
906 |
#ifdef CONFIG_MIGRATION |
8bccd85ff [PATCH] Implement... |
907 |
/* |
6ce3c4c0f [PATCH] Move page... |
908 909 |
* page migration */ |
fc3012896 [PATCH] Simplify ... |
910 911 |
static void migrate_page_add(struct page *page, struct list_head *pagelist, unsigned long flags) |
6ce3c4c0f [PATCH] Move page... |
912 913 |
{ /* |
fc3012896 [PATCH] Simplify ... |
914 |
* Avoid migrating a page that is shared with others. |
6ce3c4c0f [PATCH] Move page... |
915 |
*/ |
62695a84e vmscan: move isol... |
916 917 918 |
if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1) { if (!isolate_lru_page(page)) { list_add_tail(&page->lru, pagelist); |
599d0c954 mm, vmscan: move ... |
919 |
inc_node_page_state(page, NR_ISOLATED_ANON + |
6d9c285a6 mm: move inc_zone... |
920 |
page_is_file_cache(page)); |
62695a84e vmscan: move isol... |
921 922 |
} } |
7e2ab150d [PATCH] Direct Mi... |
923 |
} |
6ce3c4c0f [PATCH] Move page... |
924 |
|
742755a1d [PATCH] page migr... |
925 |
static struct page *new_node_page(struct page *page, unsigned long node, int **x) |
95a402c38 [PATCH] page migr... |
926 |
{ |
e2d8cf405 migrate: add huge... |
927 928 929 930 |
if (PageHuge(page)) return alloc_huge_page_node(page_hstate(compound_head(page)), node); else |
96db800f5 mm: rename alloc_... |
931 |
return __alloc_pages_node(node, GFP_HIGHUSER_MOVABLE | |
b360edb43 mm, mempolicy: mi... |
932 |
__GFP_THISNODE, 0); |
95a402c38 [PATCH] page migr... |
933 |
} |
7e2ab150d [PATCH] Direct Mi... |
934 935 936 937 |
/* * Migrate pages from one node to a target node. * Returns error or the number of pages not migrated. */ |
dbcb0f19c mm/mempolicy.c: c... |
938 939 |
static int migrate_to_node(struct mm_struct *mm, int source, int dest, int flags) |
7e2ab150d [PATCH] Direct Mi... |
940 941 942 943 944 945 946 |
{ nodemask_t nmask; LIST_HEAD(pagelist); int err = 0; nodes_clear(nmask); node_set(source, nmask); |
6ce3c4c0f [PATCH] Move page... |
947 |
|
082708072 mm: revert 0def08... |
948 949 950 951 952 953 |
/* * This does not "check" the range but isolates all pages that * need migration. Between passing in the full user address * space range and MPOL_MF_DISCONTIG_OK, this call can not fail. */ VM_BUG_ON(!(flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))); |
980949457 mm/mempolicy: ren... |
954 |
queue_pages_range(mm, mm->mmap->vm_start, mm->task_size, &nmask, |
7e2ab150d [PATCH] Direct Mi... |
955 |
flags | MPOL_MF_DISCONTIG_OK, &pagelist); |
cf608ac19 mm: compaction: f... |
956 |
if (!list_empty(&pagelist)) { |
68711a746 mm, migration: ad... |
957 |
err = migrate_pages(&pagelist, new_node_page, NULL, dest, |
9c620e2bc mm: remove offlin... |
958 |
MIGRATE_SYNC, MR_SYSCALL); |
cf608ac19 mm: compaction: f... |
959 |
if (err) |
e2d8cf405 migrate: add huge... |
960 |
putback_movable_pages(&pagelist); |
cf608ac19 mm: compaction: f... |
961 |
} |
95a402c38 [PATCH] page migr... |
962 |
|
7e2ab150d [PATCH] Direct Mi... |
963 |
return err; |
6ce3c4c0f [PATCH] Move page... |
964 965 966 |
} /* |
7e2ab150d [PATCH] Direct Mi... |
967 968 |
* Move pages between the two nodesets so as to preserve the physical * layout as much as possible. |
39743889a [PATCH] Swap Migr... |
969 970 971 |
* * Returns the number of page that could not be moved. */ |
0ce72d4f7 mm: do_migrate_pa... |
972 973 |
int do_migrate_pages(struct mm_struct *mm, const nodemask_t *from, const nodemask_t *to, int flags) |
39743889a [PATCH] Swap Migr... |
974 |
{ |
7e2ab150d [PATCH] Direct Mi... |
975 |
int busy = 0; |
0aedadf91 mm: move migrate_... |
976 |
int err; |
7e2ab150d [PATCH] Direct Mi... |
977 |
nodemask_t tmp; |
39743889a [PATCH] Swap Migr... |
978 |
|
0aedadf91 mm: move migrate_... |
979 980 981 |
err = migrate_prep(); if (err) return err; |
53f2556b6 mempolicy: mPOL_P... |
982 |
down_read(&mm->mmap_sem); |
39743889a [PATCH] Swap Migr... |
983 |
|
da0aa1389 mm/mempolicy.c: f... |
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 |
/* * Find a 'source' bit set in 'tmp' whose corresponding 'dest' * bit in 'to' is not also set in 'tmp'. Clear the found 'source' * bit in 'tmp', and return that <source, dest> pair for migration. * The pair of nodemasks 'to' and 'from' define the map. * * If no pair of bits is found that way, fallback to picking some * pair of 'source' and 'dest' bits that are not the same. If the * 'source' and 'dest' bits are the same, this represents a node * that will be migrating to itself, so no pages need move. * * If no bits are left in 'tmp', or if all remaining bits left * in 'tmp' correspond to the same bit in 'to', return false * (nothing left to migrate). * * This lets us pick a pair of nodes to migrate between, such that * if possible the dest node is not already occupied by some other * source node, minimizing the risk of overloading the memory on a * node that would happen if we migrated incoming memory to a node * before migrating outgoing memory source that same node. * * A single scan of tmp is sufficient. As we go, we remember the * most recent <s, d> pair that moved (s != d). If we find a pair * that not only moved, but what's better, moved to an empty slot * (d is not set in tmp), then we break out then, with that pair. |
ae0e47f02 Remove one to man... |
1009 |
* Otherwise when we finish scanning from_tmp, we at least have the |
da0aa1389 mm/mempolicy.c: f... |
1010 1011 1012 1013 |
* most recent <s, d> pair that moved. If we get all the way through * the scan of tmp without finding any node that moved, much less * moved to an empty node, then there is nothing left worth migrating. */ |
d49847113 [PATCH] SwapMig: ... |
1014 |
|
0ce72d4f7 mm: do_migrate_pa... |
1015 |
tmp = *from; |
7e2ab150d [PATCH] Direct Mi... |
1016 1017 |
while (!nodes_empty(tmp)) { int s,d; |
b76ac7e73 mm/mempolicy: use... |
1018 |
int source = NUMA_NO_NODE; |
7e2ab150d [PATCH] Direct Mi... |
1019 1020 1021 |
int dest = 0; for_each_node_mask(s, tmp) { |
4a5b18cc1 mm: do_migrate_pa... |
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 |
/* * do_migrate_pages() tries to maintain the relative * node relationship of the pages established between * threads and memory areas. * * However if the number of source nodes is not equal to * the number of destination nodes we can not preserve * this node relative relationship. In that case, skip * copying memory from a node that is in the destination * mask. * * Example: [2,3,4] -> [3,4,5] moves everything. * [0-7] - > [3,4,5] moves only 0,1,2,6,7. */ |
0ce72d4f7 mm: do_migrate_pa... |
1037 1038 |
if ((nodes_weight(*from) != nodes_weight(*to)) && (node_isset(s, *to))) |
4a5b18cc1 mm: do_migrate_pa... |
1039 |
continue; |
0ce72d4f7 mm: do_migrate_pa... |
1040 |
d = node_remap(s, *from, *to); |
7e2ab150d [PATCH] Direct Mi... |
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 |
if (s == d) continue; source = s; /* Node moved. Memorize */ dest = d; /* dest not in remaining from nodes? */ if (!node_isset(dest, tmp)) break; } |
b76ac7e73 mm/mempolicy: use... |
1051 |
if (source == NUMA_NO_NODE) |
7e2ab150d [PATCH] Direct Mi... |
1052 1053 1054 1055 1056 1057 1058 1059 |
break; node_clear(source, tmp); err = migrate_to_node(mm, source, dest, flags); if (err > 0) busy += err; if (err < 0) break; |
39743889a [PATCH] Swap Migr... |
1060 1061 |
} up_read(&mm->mmap_sem); |
7e2ab150d [PATCH] Direct Mi... |
1062 1063 1064 |
if (err < 0) return err; return busy; |
b20a35035 [PATCH] page migr... |
1065 1066 |
} |
3ad33b243 Migration: find c... |
1067 1068 |
/* * Allocate a new page for page migration based on vma policy. |
d05f0cdcb mm: fix crashes f... |
1069 |
* Start by assuming the page is mapped by the same vma as contains @start. |
3ad33b243 Migration: find c... |
1070 1071 1072 1073 |
* Search forward from there, if not. N.B., this assumes that the * list of pages handed to migrate_pages()--which is how we get here-- * is in virtual address order. */ |
d05f0cdcb mm: fix crashes f... |
1074 |
static struct page *new_page(struct page *page, unsigned long start, int **x) |
95a402c38 [PATCH] page migr... |
1075 |
{ |
d05f0cdcb mm: fix crashes f... |
1076 |
struct vm_area_struct *vma; |
3ad33b243 Migration: find c... |
1077 |
unsigned long uninitialized_var(address); |
95a402c38 [PATCH] page migr... |
1078 |
|
d05f0cdcb mm: fix crashes f... |
1079 |
vma = find_vma(current->mm, start); |
3ad33b243 Migration: find c... |
1080 1081 1082 1083 1084 1085 |
while (vma) { address = page_address_in_vma(page, vma); if (address != -EFAULT) break; vma = vma->vm_next; } |
11c731e81 mm/mempolicy: fix... |
1086 1087 |
if (PageHuge(page)) { |
cc81717ed mm: new_vma_page(... |
1088 1089 |
BUG_ON(!vma); return alloc_huge_page_noerr(vma, address, 1); |
11c731e81 mm/mempolicy: fix... |
1090 |
} |
0bf598d86 mbind: add BUG_ON... |
1091 |
/* |
11c731e81 mm/mempolicy: fix... |
1092 |
* if !vma, alloc_page_vma() will use task or system default policy |
0bf598d86 mbind: add BUG_ON... |
1093 |
*/ |
3ad33b243 Migration: find c... |
1094 |
return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address); |
95a402c38 [PATCH] page migr... |
1095 |
} |
b20a35035 [PATCH] page migr... |
1096 1097 1098 1099 1100 |
#else static void migrate_page_add(struct page *page, struct list_head *pagelist, unsigned long flags) { |
39743889a [PATCH] Swap Migr... |
1101 |
} |
0ce72d4f7 mm: do_migrate_pa... |
1102 1103 |
int do_migrate_pages(struct mm_struct *mm, const nodemask_t *from, const nodemask_t *to, int flags) |
b20a35035 [PATCH] page migr... |
1104 1105 1106 |
{ return -ENOSYS; } |
95a402c38 [PATCH] page migr... |
1107 |
|
d05f0cdcb mm: fix crashes f... |
1108 |
static struct page *new_page(struct page *page, unsigned long start, int **x) |
95a402c38 [PATCH] page migr... |
1109 1110 1111 |
{ return NULL; } |
b20a35035 [PATCH] page migr... |
1112 |
#endif |
dbcb0f19c mm/mempolicy.c: c... |
1113 |
static long do_mbind(unsigned long start, unsigned long len, |
028fec414 mempolicy: suppor... |
1114 1115 |
unsigned short mode, unsigned short mode_flags, nodemask_t *nmask, unsigned long flags) |
6ce3c4c0f [PATCH] Move page... |
1116 |
{ |
6ce3c4c0f [PATCH] Move page... |
1117 1118 1119 1120 1121 |
struct mm_struct *mm = current->mm; struct mempolicy *new; unsigned long end; int err; LIST_HEAD(pagelist); |
b24f53a0b mm: mempolicy: Ad... |
1122 |
if (flags & ~(unsigned long)MPOL_MF_VALID) |
6ce3c4c0f [PATCH] Move page... |
1123 |
return -EINVAL; |
74c002410 [PATCH] Consisten... |
1124 |
if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE)) |
6ce3c4c0f [PATCH] Move page... |
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 |
return -EPERM; if (start & ~PAGE_MASK) return -EINVAL; if (mode == MPOL_DEFAULT) flags &= ~MPOL_MF_STRICT; len = (len + PAGE_SIZE - 1) & PAGE_MASK; end = start + len; if (end < start) return -EINVAL; if (end == start) return 0; |
028fec414 mempolicy: suppor... |
1140 |
new = mpol_new(mode, mode_flags, nmask); |
6ce3c4c0f [PATCH] Move page... |
1141 1142 |
if (IS_ERR(new)) return PTR_ERR(new); |
b24f53a0b mm: mempolicy: Ad... |
1143 1144 |
if (flags & MPOL_MF_LAZY) new->flags |= MPOL_F_MOF; |
6ce3c4c0f [PATCH] Move page... |
1145 1146 1147 1148 1149 1150 |
/* * If we are using the default policy then operation * on discontinuous address spaces is okay after all */ if (!new) flags |= MPOL_MF_DISCONTIG_OK; |
028fec414 mempolicy: suppor... |
1151 1152 1153 |
pr_debug("mbind %lx-%lx mode:%d flags:%d nodes:%lx ", start, start + len, mode, mode_flags, |
00ef2d2f8 mm: use NUMA_NO_NODE |
1154 |
nmask ? nodes_addr(*nmask)[0] : NUMA_NO_NODE); |
6ce3c4c0f [PATCH] Move page... |
1155 |
|
0aedadf91 mm: move migrate_... |
1156 1157 1158 1159 |
if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) { err = migrate_prep(); if (err) |
b05ca7385 do_mbind(): fix m... |
1160 |
goto mpol_out; |
0aedadf91 mm: move migrate_... |
1161 |
} |
4bfc44958 mm: make set_memp... |
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 |
{ NODEMASK_SCRATCH(scratch); if (scratch) { down_write(&mm->mmap_sem); task_lock(current); err = mpol_set_nodemask(new, nmask, scratch); task_unlock(current); if (err) up_write(&mm->mmap_sem); } else err = -ENOMEM; NODEMASK_SCRATCH_FREE(scratch); } |
b05ca7385 do_mbind(): fix m... |
1175 1176 |
if (err) goto mpol_out; |
d05f0cdcb mm: fix crashes f... |
1177 |
err = queue_pages_range(mm, start, end, nmask, |
6ce3c4c0f [PATCH] Move page... |
1178 |
flags | MPOL_MF_INVERT, &pagelist); |
d05f0cdcb mm: fix crashes f... |
1179 |
if (!err) |
9d8cebd4b mm: fix mbind vma... |
1180 |
err = mbind_range(mm, start, end, new); |
7e2ab150d [PATCH] Direct Mi... |
1181 |
|
b24f53a0b mm: mempolicy: Ad... |
1182 1183 |
if (!err) { int nr_failed = 0; |
cf608ac19 mm: compaction: f... |
1184 |
if (!list_empty(&pagelist)) { |
b24f53a0b mm: mempolicy: Ad... |
1185 |
WARN_ON_ONCE(flags & MPOL_MF_LAZY); |
d05f0cdcb mm: fix crashes f... |
1186 1187 |
nr_failed = migrate_pages(&pagelist, new_page, NULL, start, MIGRATE_SYNC, MR_MEMPOLICY_MBIND); |
cf608ac19 mm: compaction: f... |
1188 |
if (nr_failed) |
74060e4d7 mm: mbind: add hu... |
1189 |
putback_movable_pages(&pagelist); |
cf608ac19 mm: compaction: f... |
1190 |
} |
6ce3c4c0f [PATCH] Move page... |
1191 |
|
b24f53a0b mm: mempolicy: Ad... |
1192 |
if (nr_failed && (flags & MPOL_MF_STRICT)) |
6ce3c4c0f [PATCH] Move page... |
1193 |
err = -EIO; |
ab8a3e14e mbind(): fix leak... |
1194 |
} else |
b0e5fd735 mm/mempolicy: cor... |
1195 |
putback_movable_pages(&pagelist); |
b20a35035 [PATCH] page migr... |
1196 |
|
6ce3c4c0f [PATCH] Move page... |
1197 |
up_write(&mm->mmap_sem); |
b05ca7385 do_mbind(): fix m... |
1198 |
mpol_out: |
f0be3d32b mempolicy: rename... |
1199 |
mpol_put(new); |
6ce3c4c0f [PATCH] Move page... |
1200 1201 |
return err; } |
39743889a [PATCH] Swap Migr... |
1202 |
/* |
8bccd85ff [PATCH] Implement... |
1203 1204 1205 1206 |
* User space interface with variable sized bitmaps for nodelists. */ /* Copy a node mask from user space. */ |
39743889a [PATCH] Swap Migr... |
1207 |
static int get_nodes(nodemask_t *nodes, const unsigned long __user *nmask, |
8bccd85ff [PATCH] Implement... |
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 |
unsigned long maxnode) { unsigned long k; unsigned long nlongs; unsigned long endmask; --maxnode; nodes_clear(*nodes); if (maxnode == 0 || !nmask) return 0; |
a9c930bac [PATCH] Fix units... |
1218 |
if (maxnode > PAGE_SIZE*BITS_PER_BYTE) |
636f13c17 [PATCH] sys_mbind... |
1219 |
return -EINVAL; |
8bccd85ff [PATCH] Implement... |
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 |
nlongs = BITS_TO_LONGS(maxnode); if ((maxnode % BITS_PER_LONG) == 0) endmask = ~0UL; else endmask = (1UL << (maxnode % BITS_PER_LONG)) - 1; /* When the user specified more nodes than supported just check if the non supported part is all zero. */ if (nlongs > BITS_TO_LONGS(MAX_NUMNODES)) { if (nlongs > PAGE_SIZE/sizeof(long)) return -EINVAL; for (k = BITS_TO_LONGS(MAX_NUMNODES); k < nlongs; k++) { unsigned long t; if (get_user(t, nmask + k)) return -EFAULT; if (k == nlongs - 1) { if (t & endmask) return -EINVAL; } else if (t) return -EINVAL; } nlongs = BITS_TO_LONGS(MAX_NUMNODES); endmask = ~0UL; } if (copy_from_user(nodes_addr(*nodes), nmask, nlongs*sizeof(unsigned long))) return -EFAULT; nodes_addr(*nodes)[nlongs-1] &= endmask; return 0; } /* Copy a kernel node mask to user space */ static int copy_nodes_to_user(unsigned long __user *mask, unsigned long maxnode, nodemask_t *nodes) { unsigned long copy = ALIGN(maxnode-1, 64) / 8; const int nbytes = BITS_TO_LONGS(MAX_NUMNODES) * sizeof(long); if (copy > nbytes) { if (copy > PAGE_SIZE) return -EINVAL; if (clear_user((char __user *)mask + nbytes, copy - nbytes)) return -EFAULT; copy = nbytes; } return copy_to_user(mask, nodes_addr(*nodes), copy) ? -EFAULT : 0; } |
938bb9f5e [CVE-2009-0029] S... |
1268 |
SYSCALL_DEFINE6(mbind, unsigned long, start, unsigned long, len, |
f7f28ca98 mm: constify nmas... |
1269 |
unsigned long, mode, const unsigned long __user *, nmask, |
938bb9f5e [CVE-2009-0029] S... |
1270 |
unsigned long, maxnode, unsigned, flags) |
8bccd85ff [PATCH] Implement... |
1271 1272 1273 |
{ nodemask_t nodes; int err; |
028fec414 mempolicy: suppor... |
1274 |
unsigned short mode_flags; |
8bccd85ff [PATCH] Implement... |
1275 |
|
028fec414 mempolicy: suppor... |
1276 1277 |
mode_flags = mode & MPOL_MODE_FLAGS; mode &= ~MPOL_MODE_FLAGS; |
a3b51e014 mempolicy: conver... |
1278 1279 |
if (mode >= MPOL_MAX) return -EINVAL; |
4c50bc011 mempolicy: add MP... |
1280 1281 1282 |
if ((mode_flags & MPOL_F_STATIC_NODES) && (mode_flags & MPOL_F_RELATIVE_NODES)) return -EINVAL; |
8bccd85ff [PATCH] Implement... |
1283 1284 1285 |
err = get_nodes(&nodes, nmask, maxnode); if (err) return err; |
028fec414 mempolicy: suppor... |
1286 |
return do_mbind(start, len, mode, mode_flags, &nodes, flags); |
8bccd85ff [PATCH] Implement... |
1287 1288 1289 |
} /* Set the process memory policy */ |
23c8902d4 mm: constify nmas... |
1290 |
SYSCALL_DEFINE3(set_mempolicy, int, mode, const unsigned long __user *, nmask, |
938bb9f5e [CVE-2009-0029] S... |
1291 |
unsigned long, maxnode) |
8bccd85ff [PATCH] Implement... |
1292 1293 1294 |
{ int err; nodemask_t nodes; |
028fec414 mempolicy: suppor... |
1295 |
unsigned short flags; |
8bccd85ff [PATCH] Implement... |
1296 |
|
028fec414 mempolicy: suppor... |
1297 1298 1299 |
flags = mode & MPOL_MODE_FLAGS; mode &= ~MPOL_MODE_FLAGS; if ((unsigned int)mode >= MPOL_MAX) |
8bccd85ff [PATCH] Implement... |
1300 |
return -EINVAL; |
4c50bc011 mempolicy: add MP... |
1301 1302 |
if ((flags & MPOL_F_STATIC_NODES) && (flags & MPOL_F_RELATIVE_NODES)) return -EINVAL; |
8bccd85ff [PATCH] Implement... |
1303 1304 1305 |
err = get_nodes(&nodes, nmask, maxnode); if (err) return err; |
028fec414 mempolicy: suppor... |
1306 |
return do_set_mempolicy(mode, flags, &nodes); |
8bccd85ff [PATCH] Implement... |
1307 |
} |
938bb9f5e [CVE-2009-0029] S... |
1308 1309 1310 |
SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode, const unsigned long __user *, old_nodes, const unsigned long __user *, new_nodes) |
39743889a [PATCH] Swap Migr... |
1311 |
{ |
c69e8d9c0 CRED: Use RCU to ... |
1312 |
const struct cred *cred = current_cred(), *tcred; |
596d7cfa2 mempolicy: reduce... |
1313 |
struct mm_struct *mm = NULL; |
39743889a [PATCH] Swap Migr... |
1314 |
struct task_struct *task; |
39743889a [PATCH] Swap Migr... |
1315 1316 |
nodemask_t task_nodes; int err; |
596d7cfa2 mempolicy: reduce... |
1317 1318 1319 1320 1321 1322 |
nodemask_t *old; nodemask_t *new; NODEMASK_SCRATCH(scratch); if (!scratch) return -ENOMEM; |
39743889a [PATCH] Swap Migr... |
1323 |
|
596d7cfa2 mempolicy: reduce... |
1324 1325 1326 1327 |
old = &scratch->mask1; new = &scratch->mask2; err = get_nodes(old, old_nodes, maxnode); |
39743889a [PATCH] Swap Migr... |
1328 |
if (err) |
596d7cfa2 mempolicy: reduce... |
1329 |
goto out; |
39743889a [PATCH] Swap Migr... |
1330 |
|
596d7cfa2 mempolicy: reduce... |
1331 |
err = get_nodes(new, new_nodes, maxnode); |
39743889a [PATCH] Swap Migr... |
1332 |
if (err) |
596d7cfa2 mempolicy: reduce... |
1333 |
goto out; |
39743889a [PATCH] Swap Migr... |
1334 1335 |
/* Find the mm_struct */ |
55cfaa3cb mm/mempolicy.c: a... |
1336 |
rcu_read_lock(); |
228ebcbe6 Uninline find_tas... |
1337 |
task = pid ? find_task_by_vpid(pid) : current; |
39743889a [PATCH] Swap Migr... |
1338 |
if (!task) { |
55cfaa3cb mm/mempolicy.c: a... |
1339 |
rcu_read_unlock(); |
596d7cfa2 mempolicy: reduce... |
1340 1341 |
err = -ESRCH; goto out; |
39743889a [PATCH] Swap Migr... |
1342 |
} |
3268c63ed mm: fix move/migr... |
1343 |
get_task_struct(task); |
39743889a [PATCH] Swap Migr... |
1344 |
|
596d7cfa2 mempolicy: reduce... |
1345 |
err = -EINVAL; |
39743889a [PATCH] Swap Migr... |
1346 1347 1348 1349 |
/* * Check if this process has the right to modify the specified * process. The right exists if the process has administrative |
7f927fcc2 [PATCH] Typo fixes |
1350 |
* capabilities, superuser privileges or the same |
39743889a [PATCH] Swap Migr... |
1351 1352 |
* userid as the target process. */ |
c69e8d9c0 CRED: Use RCU to ... |
1353 |
tcred = __task_cred(task); |
b38a86eb1 userns: Convert t... |
1354 1355 |
if (!uid_eq(cred->euid, tcred->suid) && !uid_eq(cred->euid, tcred->uid) && !uid_eq(cred->uid, tcred->suid) && !uid_eq(cred->uid, tcred->uid) && |
74c002410 [PATCH] Consisten... |
1356 |
!capable(CAP_SYS_NICE)) { |
c69e8d9c0 CRED: Use RCU to ... |
1357 |
rcu_read_unlock(); |
39743889a [PATCH] Swap Migr... |
1358 |
err = -EPERM; |
3268c63ed mm: fix move/migr... |
1359 |
goto out_put; |
39743889a [PATCH] Swap Migr... |
1360 |
} |
c69e8d9c0 CRED: Use RCU to ... |
1361 |
rcu_read_unlock(); |
39743889a [PATCH] Swap Migr... |
1362 1363 1364 |
task_nodes = cpuset_mems_allowed(task); /* Is the user allowed to access the target nodes? */ |
596d7cfa2 mempolicy: reduce... |
1365 |
if (!nodes_subset(*new, task_nodes) && !capable(CAP_SYS_NICE)) { |
39743889a [PATCH] Swap Migr... |
1366 |
err = -EPERM; |
3268c63ed mm: fix move/migr... |
1367 |
goto out_put; |
39743889a [PATCH] Swap Migr... |
1368 |
} |
01f13bd60 mempolicy: use N_... |
1369 |
if (!nodes_subset(*new, node_states[N_MEMORY])) { |
3b42d28b2 Page migration: D... |
1370 |
err = -EINVAL; |
3268c63ed mm: fix move/migr... |
1371 |
goto out_put; |
3b42d28b2 Page migration: D... |
1372 |
} |
86c3a7645 [PATCH] SELinux: ... |
1373 1374 |
err = security_task_movememory(task); if (err) |
3268c63ed mm: fix move/migr... |
1375 |
goto out_put; |
86c3a7645 [PATCH] SELinux: ... |
1376 |
|
3268c63ed mm: fix move/migr... |
1377 1378 |
mm = get_task_mm(task); put_task_struct(task); |
f2a9ef880 mm: fix NULL ptr ... |
1379 1380 |
if (!mm) { |
3268c63ed mm: fix move/migr... |
1381 |
err = -EINVAL; |
f2a9ef880 mm: fix NULL ptr ... |
1382 1383 1384 1385 1386 |
goto out; } err = do_migrate_pages(mm, old, new, capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE); |
3268c63ed mm: fix move/migr... |
1387 1388 1389 |
mmput(mm); out: |
596d7cfa2 mempolicy: reduce... |
1390 |
NODEMASK_SCRATCH_FREE(scratch); |
39743889a [PATCH] Swap Migr... |
1391 |
return err; |
3268c63ed mm: fix move/migr... |
1392 1393 1394 1395 |
out_put: put_task_struct(task); goto out; |
39743889a [PATCH] Swap Migr... |
1396 |
} |
8bccd85ff [PATCH] Implement... |
1397 |
/* Retrieve NUMA policy */ |
938bb9f5e [CVE-2009-0029] S... |
1398 1399 1400 |
SYSCALL_DEFINE5(get_mempolicy, int __user *, policy, unsigned long __user *, nmask, unsigned long, maxnode, unsigned long, addr, unsigned long, flags) |
8bccd85ff [PATCH] Implement... |
1401 |
{ |
dbcb0f19c mm/mempolicy.c: c... |
1402 1403 |
int err; int uninitialized_var(pval); |
8bccd85ff [PATCH] Implement... |
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 |
nodemask_t nodes; if (nmask != NULL && maxnode < MAX_NUMNODES) return -EINVAL; err = do_get_mempolicy(&pval, &nodes, addr, flags); if (err) return err; if (policy && put_user(pval, policy)) return -EFAULT; if (nmask) err = copy_nodes_to_user(nmask, maxnode, &nodes); return err; } |
1da177e4c Linux-2.6.12-rc2 |
1422 |
#ifdef CONFIG_COMPAT |
c93e0f6c8 mm/compat: conver... |
1423 1424 1425 1426 |
COMPAT_SYSCALL_DEFINE5(get_mempolicy, int __user *, policy, compat_ulong_t __user *, nmask, compat_ulong_t, maxnode, compat_ulong_t, addr, compat_ulong_t, flags) |
1da177e4c Linux-2.6.12-rc2 |
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 |
{ long err; unsigned long __user *nm = NULL; unsigned long nr_bits, alloc_size; DECLARE_BITMAP(bm, MAX_NUMNODES); nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; if (nmask) nm = compat_alloc_user_space(alloc_size); err = sys_get_mempolicy(policy, nm, nr_bits+1, addr, flags); if (!err && nmask) { |
2bbff6c76 mm/mempolicy.c: m... |
1442 1443 1444 |
unsigned long copy_size; copy_size = min_t(unsigned long, sizeof(bm), alloc_size); err = copy_from_user(bm, nm, copy_size); |
1da177e4c Linux-2.6.12-rc2 |
1445 1446 1447 1448 1449 1450 1451 |
/* ensure entire bitmap is zeroed */ err |= clear_user(nmask, ALIGN(maxnode-1, 8) / 8); err |= compat_put_bitmap(nmask, bm, nr_bits); } return err; } |
c93e0f6c8 mm/compat: conver... |
1452 1453 |
COMPAT_SYSCALL_DEFINE3(set_mempolicy, int, mode, compat_ulong_t __user *, nmask, compat_ulong_t, maxnode) |
1da177e4c Linux-2.6.12-rc2 |
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 |
{ long err = 0; unsigned long __user *nm = NULL; unsigned long nr_bits, alloc_size; DECLARE_BITMAP(bm, MAX_NUMNODES); nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; if (nmask) { err = compat_get_bitmap(bm, nmask, nr_bits); nm = compat_alloc_user_space(alloc_size); err |= copy_to_user(nm, bm, alloc_size); } if (err) return -EFAULT; return sys_set_mempolicy(mode, nm, nr_bits+1); } |
c93e0f6c8 mm/compat: conver... |
1474 1475 1476 |
COMPAT_SYSCALL_DEFINE6(mbind, compat_ulong_t, start, compat_ulong_t, len, compat_ulong_t, mode, compat_ulong_t __user *, nmask, compat_ulong_t, maxnode, compat_ulong_t, flags) |
1da177e4c Linux-2.6.12-rc2 |
1477 1478 1479 1480 |
{ long err = 0; unsigned long __user *nm = NULL; unsigned long nr_bits, alloc_size; |
dfcd3c0dc [PATCH] Convert m... |
1481 |
nodemask_t bm; |
1da177e4c Linux-2.6.12-rc2 |
1482 1483 1484 1485 1486 |
nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; if (nmask) { |
dfcd3c0dc [PATCH] Convert m... |
1487 |
err = compat_get_bitmap(nodes_addr(bm), nmask, nr_bits); |
1da177e4c Linux-2.6.12-rc2 |
1488 |
nm = compat_alloc_user_space(alloc_size); |
dfcd3c0dc [PATCH] Convert m... |
1489 |
err |= copy_to_user(nm, nodes_addr(bm), alloc_size); |
1da177e4c Linux-2.6.12-rc2 |
1490 1491 1492 1493 1494 1495 1496 1497 1498 |
} if (err) return -EFAULT; return sys_mbind(start, len, mode, nm, nr_bits+1, flags); } #endif |
74d2c3a05 mempolicy: introd... |
1499 1500 |
struct mempolicy *__get_vma_policy(struct vm_area_struct *vma, unsigned long addr) |
1da177e4c Linux-2.6.12-rc2 |
1501 |
{ |
8d90274b3 mempolicy: saniti... |
1502 |
struct mempolicy *pol = NULL; |
1da177e4c Linux-2.6.12-rc2 |
1503 1504 |
if (vma) { |
480eccf9a Fix NUMA Memory P... |
1505 |
if (vma->vm_ops && vma->vm_ops->get_policy) { |
8d90274b3 mempolicy: saniti... |
1506 |
pol = vma->vm_ops->get_policy(vma, addr); |
00442ad04 mempolicy: fix a ... |
1507 |
} else if (vma->vm_policy) { |
1da177e4c Linux-2.6.12-rc2 |
1508 |
pol = vma->vm_policy; |
00442ad04 mempolicy: fix a ... |
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 |
/* * shmem_alloc_page() passes MPOL_F_SHARED policy with * a pseudo vma whose vma->vm_ops=NULL. Take a reference * count on these policies which will be dropped by * mpol_cond_put() later */ if (mpol_needs_cond_ref(pol)) mpol_get(pol); } |
1da177e4c Linux-2.6.12-rc2 |
1519 |
} |
f15ca78e3 mempolicy: change... |
1520 |
|
74d2c3a05 mempolicy: introd... |
1521 1522 1523 1524 |
return pol; } /* |
dd6eecb91 mempolicy: unexpo... |
1525 |
* get_vma_policy(@vma, @addr) |
74d2c3a05 mempolicy: introd... |
1526 1527 1528 1529 |
* @vma: virtual memory area whose policy is sought * @addr: address in @vma for shared policy lookup * * Returns effective policy for a VMA at specified address. |
dd6eecb91 mempolicy: unexpo... |
1530 |
* Falls back to current->mempolicy or system default policy, as necessary. |
74d2c3a05 mempolicy: introd... |
1531 1532 1533 1534 1535 |
* Shared policies [those marked as MPOL_F_SHARED] require an extra reference * count--added by the get_policy() vm_op, as appropriate--to protect against * freeing by another task. It is the caller's responsibility to free the * extra reference for shared policies. */ |
dd6eecb91 mempolicy: unexpo... |
1536 1537 |
static struct mempolicy *get_vma_policy(struct vm_area_struct *vma, unsigned long addr) |
74d2c3a05 mempolicy: introd... |
1538 1539 |
{ struct mempolicy *pol = __get_vma_policy(vma, addr); |
8d90274b3 mempolicy: saniti... |
1540 |
if (!pol) |
dd6eecb91 mempolicy: unexpo... |
1541 |
pol = get_task_policy(current); |
8d90274b3 mempolicy: saniti... |
1542 |
|
1da177e4c Linux-2.6.12-rc2 |
1543 1544 |
return pol; } |
6b6482bbf mempolicy: remove... |
1545 |
bool vma_policy_mof(struct vm_area_struct *vma) |
fc3147245 mm: numa: Limit N... |
1546 |
{ |
6b6482bbf mempolicy: remove... |
1547 |
struct mempolicy *pol; |
fc3147245 mm: numa: Limit N... |
1548 |
|
6b6482bbf mempolicy: remove... |
1549 1550 |
if (vma->vm_ops && vma->vm_ops->get_policy) { bool ret = false; |
fc3147245 mm: numa: Limit N... |
1551 |
|
6b6482bbf mempolicy: remove... |
1552 1553 1554 1555 |
pol = vma->vm_ops->get_policy(vma, vma->vm_start); if (pol && (pol->flags & MPOL_F_MOF)) ret = true; mpol_cond_put(pol); |
8d90274b3 mempolicy: saniti... |
1556 |
|
6b6482bbf mempolicy: remove... |
1557 |
return ret; |
fc3147245 mm: numa: Limit N... |
1558 |
} |
6b6482bbf mempolicy: remove... |
1559 |
pol = vma->vm_policy; |
8d90274b3 mempolicy: saniti... |
1560 |
if (!pol) |
6b6482bbf mempolicy: remove... |
1561 |
pol = get_task_policy(current); |
8d90274b3 mempolicy: saniti... |
1562 |
|
fc3147245 mm: numa: Limit N... |
1563 1564 |
return pol->flags & MPOL_F_MOF; } |
d3eb1570a mempolicy: fix is... |
1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 |
static int apply_policy_zone(struct mempolicy *policy, enum zone_type zone) { enum zone_type dynamic_policy_zone = policy_zone; BUG_ON(dynamic_policy_zone == ZONE_MOVABLE); /* * if policy->v.nodes has movable memory only, * we apply policy when gfp_zone(gfp) = ZONE_MOVABLE only. * * policy->v.nodes is intersect with node_states[N_MEMORY]. * so if the following test faile, it implies * policy->v.nodes has movable memory only. */ if (!nodes_intersects(policy->v.nodes, node_states[N_HIGH_MEMORY])) dynamic_policy_zone = ZONE_MOVABLE; return zone >= dynamic_policy_zone; } |
52cd3b074 mempolicy: rework... |
1584 1585 1586 1587 1588 |
/* * Return a nodemask representing a mempolicy for filtering nodes for * page allocation */ static nodemask_t *policy_nodemask(gfp_t gfp, struct mempolicy *policy) |
19770b326 mm: filter based ... |
1589 1590 |
{ /* Lower zones don't get a nodemask applied for MPOL_BIND */ |
45c4745af mempolicy: rename... |
1591 |
if (unlikely(policy->mode == MPOL_BIND) && |
d3eb1570a mempolicy: fix is... |
1592 |
apply_policy_zone(policy, gfp_zone(gfp)) && |
19770b326 mm: filter based ... |
1593 1594 1595 1596 1597 |
cpuset_nodemask_valid_mems_allowed(&policy->v.nodes)) return &policy->v.nodes; return NULL; } |
52cd3b074 mempolicy: rework... |
1598 |
/* Return a zonelist indicated by gfp for node representing a mempolicy */ |
2f5f9486f mm: change alloc_... |
1599 1600 |
static struct zonelist *policy_zonelist(gfp_t gfp, struct mempolicy *policy, int nd) |
1da177e4c Linux-2.6.12-rc2 |
1601 |
{ |
45c4745af mempolicy: rename... |
1602 |
switch (policy->mode) { |
1da177e4c Linux-2.6.12-rc2 |
1603 |
case MPOL_PREFERRED: |
fc36b8d3d mempolicy: use MP... |
1604 1605 |
if (!(policy->flags & MPOL_F_LOCAL)) nd = policy->v.preferred_node; |
1da177e4c Linux-2.6.12-rc2 |
1606 1607 |
break; case MPOL_BIND: |
19770b326 mm: filter based ... |
1608 |
/* |
52cd3b074 mempolicy: rework... |
1609 1610 |
* Normally, MPOL_BIND allocations are node-local within the * allowed nodemask. However, if __GFP_THISNODE is set and the |
6eb27e1fd mempolicy: remove... |
1611 |
* current node isn't part of the mask, we use the zonelist for |
52cd3b074 mempolicy: rework... |
1612 |
* the first node in the mask instead. |
19770b326 mm: filter based ... |
1613 |
*/ |
19770b326 mm: filter based ... |
1614 1615 1616 1617 |
if (unlikely(gfp & __GFP_THISNODE) && unlikely(!node_isset(nd, policy->v.nodes))) nd = first_node(policy->v.nodes); break; |
1da177e4c Linux-2.6.12-rc2 |
1618 |
default: |
1da177e4c Linux-2.6.12-rc2 |
1619 1620 |
BUG(); } |
0e88460da mm: introduce nod... |
1621 |
return node_zonelist(nd, gfp); |
1da177e4c Linux-2.6.12-rc2 |
1622 1623 1624 1625 1626 1627 1628 1629 1630 |
} /* Do dynamic interleaving for a process */ static unsigned interleave_nodes(struct mempolicy *policy) { unsigned nid, next; struct task_struct *me = current; nid = me->il_next; |
0edaf86cf include/linux/nod... |
1631 |
next = next_node_in(nid, policy->v.nodes); |
f5b087b52 mempolicy: add MP... |
1632 1633 |
if (next < MAX_NUMNODES) me->il_next = next; |
1da177e4c Linux-2.6.12-rc2 |
1634 1635 |
return nid; } |
dc85da15d [PATCH] NUMA poli... |
1636 1637 1638 1639 |
/* * Depending on the memory policy provide a node from which to allocate the * next slab entry. */ |
2a389610a mm, mempolicy: re... |
1640 |
unsigned int mempolicy_slab_node(void) |
dc85da15d [PATCH] NUMA poli... |
1641 |
{ |
e7b691b08 slab/mempolicy: a... |
1642 |
struct mempolicy *policy; |
2a389610a mm, mempolicy: re... |
1643 |
int node = numa_mem_id(); |
e7b691b08 slab/mempolicy: a... |
1644 1645 |
if (in_interrupt()) |
2a389610a mm, mempolicy: re... |
1646 |
return node; |
e7b691b08 slab/mempolicy: a... |
1647 1648 |
policy = current->mempolicy; |
fc36b8d3d mempolicy: use MP... |
1649 |
if (!policy || policy->flags & MPOL_F_LOCAL) |
2a389610a mm, mempolicy: re... |
1650 |
return node; |
bea904d54 mempolicy: use MP... |
1651 1652 1653 |
switch (policy->mode) { case MPOL_PREFERRED: |
fc36b8d3d mempolicy: use MP... |
1654 1655 1656 1657 |
/* * handled MPOL_F_LOCAL above */ return policy->v.preferred_node; |
765c4507a [PATCH] GFP_THISN... |
1658 |
|
dc85da15d [PATCH] NUMA poli... |
1659 1660 |
case MPOL_INTERLEAVE: return interleave_nodes(policy); |
dd1a239f6 mm: have zonelist... |
1661 |
case MPOL_BIND: { |
c33d6c06f mm, page_alloc: a... |
1662 |
struct zoneref *z; |
dc85da15d [PATCH] NUMA poli... |
1663 1664 1665 1666 |
/* * Follow bind policy behavior and start allocation at the * first node. */ |
19770b326 mm: filter based ... |
1667 |
struct zonelist *zonelist; |
19770b326 mm: filter based ... |
1668 |
enum zone_type highest_zoneidx = gfp_zone(GFP_KERNEL); |
c9634cf01 mm: use zonelist ... |
1669 |
zonelist = &NODE_DATA(node)->node_zonelists[ZONELIST_FALLBACK]; |
c33d6c06f mm, page_alloc: a... |
1670 1671 1672 |
z = first_zones_zonelist(zonelist, highest_zoneidx, &policy->v.nodes); return z->zone ? z->zone->node : node; |
dd1a239f6 mm: have zonelist... |
1673 |
} |
dc85da15d [PATCH] NUMA poli... |
1674 |
|
dc85da15d [PATCH] NUMA poli... |
1675 |
default: |
bea904d54 mempolicy: use MP... |
1676 |
BUG(); |
dc85da15d [PATCH] NUMA poli... |
1677 1678 |
} } |
fee83b3ab mm/mempolicy.c:of... |
1679 1680 1681 1682 1683 |
/* * Do static interleaving for a VMA with known offset @n. Returns the n'th * node in pol->v.nodes (starting from n=0), wrapping around if n exceeds the * number of present nodes. */ |
1da177e4c Linux-2.6.12-rc2 |
1684 |
static unsigned offset_il_node(struct mempolicy *pol, |
fee83b3ab mm/mempolicy.c:of... |
1685 |
struct vm_area_struct *vma, unsigned long n) |
1da177e4c Linux-2.6.12-rc2 |
1686 |
{ |
dfcd3c0dc [PATCH] Convert m... |
1687 |
unsigned nnodes = nodes_weight(pol->v.nodes); |
f5b087b52 mempolicy: add MP... |
1688 |
unsigned target; |
fee83b3ab mm/mempolicy.c:of... |
1689 1690 |
int i; int nid; |
1da177e4c Linux-2.6.12-rc2 |
1691 |
|
f5b087b52 mempolicy: add MP... |
1692 1693 |
if (!nnodes) return numa_node_id(); |
fee83b3ab mm/mempolicy.c:of... |
1694 1695 1696 |
target = (unsigned int)n % nnodes; nid = first_node(pol->v.nodes); for (i = 0; i < target; i++) |
dfcd3c0dc [PATCH] Convert m... |
1697 |
nid = next_node(nid, pol->v.nodes); |
1da177e4c Linux-2.6.12-rc2 |
1698 1699 |
return nid; } |
5da7ca860 [PATCH] Add NUMA ... |
1700 1701 1702 1703 1704 1705 |
/* Determine a node number for interleave */ static inline unsigned interleave_nid(struct mempolicy *pol, struct vm_area_struct *vma, unsigned long addr, int shift) { if (vma) { unsigned long off; |
3b98b087f [PATCH] fix NUMA ... |
1706 1707 1708 1709 1710 1711 1712 1713 1714 |
/* * for small pages, there is no difference between * shift and PAGE_SHIFT, so the bit-shift is safe. * for huge pages, since vm_pgoff is in units of small * pages, we need to shift off the always 0 bits to get * a useful offset. */ BUG_ON(shift < PAGE_SHIFT); off = vma->vm_pgoff >> (shift - PAGE_SHIFT); |
5da7ca860 [PATCH] Add NUMA ... |
1715 1716 1717 1718 1719 |
off += (addr - vma->vm_start) >> shift; return offset_il_node(pol, vma, off); } else return interleave_nodes(pol); } |
00ac59adf [PATCH] x86_64: F... |
1720 |
#ifdef CONFIG_HUGETLBFS |
480eccf9a Fix NUMA Memory P... |
1721 1722 |
/* * huge_zonelist(@vma, @addr, @gfp_flags, @mpol) |
b46e14acb mm/mempolicy.c: p... |
1723 1724 1725 1726 1727 |
* @vma: virtual memory area whose policy is sought * @addr: address in @vma for shared policy lookup and interleave policy * @gfp_flags: for requested zone * @mpol: pointer to mempolicy pointer for reference counted mempolicy * @nodemask: pointer to nodemask pointer for MPOL_BIND nodemask |
480eccf9a Fix NUMA Memory P... |
1728 |
* |
52cd3b074 mempolicy: rework... |
1729 1730 1731 1732 |
* Returns a zonelist suitable for a huge page allocation and a pointer * to the struct mempolicy for conditional unref after allocation. * If the effective policy is 'BIND, returns a pointer to the mempolicy's * @nodemask for filtering the zonelist. |
c0ff7453b cpuset,mm: fix no... |
1733 |
* |
d26914d11 mm: optimize put_... |
1734 |
* Must be protected by read_mems_allowed_begin() |
480eccf9a Fix NUMA Memory P... |
1735 |
*/ |
396faf030 Allow huge page a... |
1736 |
struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr, |
19770b326 mm: filter based ... |
1737 1738 |
gfp_t gfp_flags, struct mempolicy **mpol, nodemask_t **nodemask) |
5da7ca860 [PATCH] Add NUMA ... |
1739 |
{ |
480eccf9a Fix NUMA Memory P... |
1740 |
struct zonelist *zl; |
5da7ca860 [PATCH] Add NUMA ... |
1741 |
|
dd6eecb91 mempolicy: unexpo... |
1742 |
*mpol = get_vma_policy(vma, addr); |
19770b326 mm: filter based ... |
1743 |
*nodemask = NULL; /* assume !MPOL_BIND */ |
5da7ca860 [PATCH] Add NUMA ... |
1744 |
|
52cd3b074 mempolicy: rework... |
1745 1746 |
if (unlikely((*mpol)->mode == MPOL_INTERLEAVE)) { zl = node_zonelist(interleave_nid(*mpol, vma, addr, |
a55164389 hugetlb: modular ... |
1747 |
huge_page_shift(hstate_vma(vma))), gfp_flags); |
52cd3b074 mempolicy: rework... |
1748 |
} else { |
2f5f9486f mm: change alloc_... |
1749 |
zl = policy_zonelist(gfp_flags, *mpol, numa_node_id()); |
52cd3b074 mempolicy: rework... |
1750 1751 |
if ((*mpol)->mode == MPOL_BIND) *nodemask = &(*mpol)->v.nodes; |
480eccf9a Fix NUMA Memory P... |
1752 1753 |
} return zl; |
5da7ca860 [PATCH] Add NUMA ... |
1754 |
} |
06808b082 hugetlb: derive h... |
1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 |
/* * init_nodemask_of_mempolicy * * If the current task's mempolicy is "default" [NULL], return 'false' * to indicate default policy. Otherwise, extract the policy nodemask * for 'bind' or 'interleave' policy into the argument nodemask, or * initialize the argument nodemask to contain the single node for * 'preferred' or 'local' policy and return 'true' to indicate presence * of non-default mempolicy. * * We don't bother with reference counting the mempolicy [mpol_get/put] * because the current task is examining it's own mempolicy and a task's * mempolicy is only ever changed by the task itself. * * N.B., it is the caller's responsibility to free a returned nodemask. */ bool init_nodemask_of_mempolicy(nodemask_t *mask) { struct mempolicy *mempolicy; int nid; if (!(mask && current->mempolicy)) return false; |
c0ff7453b cpuset,mm: fix no... |
1779 |
task_lock(current); |
06808b082 hugetlb: derive h... |
1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 |
mempolicy = current->mempolicy; switch (mempolicy->mode) { case MPOL_PREFERRED: if (mempolicy->flags & MPOL_F_LOCAL) nid = numa_node_id(); else nid = mempolicy->v.preferred_node; init_nodemask_of_node(mask, nid); break; case MPOL_BIND: /* Fall through */ case MPOL_INTERLEAVE: *mask = mempolicy->v.nodes; break; default: BUG(); } |
c0ff7453b cpuset,mm: fix no... |
1799 |
task_unlock(current); |
06808b082 hugetlb: derive h... |
1800 1801 1802 |
return true; } |
00ac59adf [PATCH] x86_64: F... |
1803 |
#endif |
5da7ca860 [PATCH] Add NUMA ... |
1804 |
|
6f48d0ebd oom: select task ... |
1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 |
/* * mempolicy_nodemask_intersects * * If tsk's mempolicy is "default" [NULL], return 'true' to indicate default * policy. Otherwise, check for intersection between mask and the policy * nodemask for 'bind' or 'interleave' policy. For 'perferred' or 'local' * policy, always return true since it may allocate elsewhere on fallback. * * Takes task_lock(tsk) to prevent freeing of its mempolicy. */ bool mempolicy_nodemask_intersects(struct task_struct *tsk, const nodemask_t *mask) { struct mempolicy *mempolicy; bool ret = true; if (!mask) return ret; task_lock(tsk); mempolicy = tsk->mempolicy; if (!mempolicy) goto out; switch (mempolicy->mode) { case MPOL_PREFERRED: /* * MPOL_PREFERRED and MPOL_F_LOCAL are only preferred nodes to * allocate from, they may fallback to other nodes when oom. * Thus, it's possible for tsk to have allocated memory from * nodes in mask. */ break; case MPOL_BIND: case MPOL_INTERLEAVE: ret = nodes_intersects(mempolicy->v.nodes, *mask); break; default: BUG(); } out: task_unlock(tsk); return ret; } |
1da177e4c Linux-2.6.12-rc2 |
1848 1849 |
/* Allocate a page in interleaved policy. Own path because it needs to do special accounting. */ |
662f3a0b9 [PATCH] Remove ne... |
1850 1851 |
static struct page *alloc_page_interleave(gfp_t gfp, unsigned order, unsigned nid) |
1da177e4c Linux-2.6.12-rc2 |
1852 1853 1854 |
{ struct zonelist *zl; struct page *page; |
0e88460da mm: introduce nod... |
1855 |
zl = node_zonelist(nid, gfp); |
1da177e4c Linux-2.6.12-rc2 |
1856 |
page = __alloc_pages(gfp, order, zl); |
dd1a239f6 mm: have zonelist... |
1857 |
if (page && page_zone(page) == zonelist_zone(&zl->_zonerefs[0])) |
ca889e6c4 [PATCH] Use Zoned... |
1858 |
inc_zone_page_state(page, NUMA_INTERLEAVE_HIT); |
1da177e4c Linux-2.6.12-rc2 |
1859 1860 1861 1862 |
return page; } /** |
0bbbc0b33 thp: add numa awa... |
1863 |
* alloc_pages_vma - Allocate a page for a VMA. |
1da177e4c Linux-2.6.12-rc2 |
1864 1865 1866 1867 1868 1869 1870 1871 |
* * @gfp: * %GFP_USER user allocation. * %GFP_KERNEL kernel allocations, * %GFP_HIGHMEM highmem/user allocations, * %GFP_FS allocation should not call back into a file system. * %GFP_ATOMIC don't sleep. * |
0bbbc0b33 thp: add numa awa... |
1872 |
* @order:Order of the GFP allocation. |
1da177e4c Linux-2.6.12-rc2 |
1873 1874 |
* @vma: Pointer to VMA or NULL if not available. * @addr: Virtual Address of the allocation. Must be inside the VMA. |
be97a41b2 mm/mempolicy.c: m... |
1875 1876 |
* @node: Which node to prefer for allocation (modulo policy). * @hugepage: for hugepages try only the preferred node if possible |
1da177e4c Linux-2.6.12-rc2 |
1877 1878 1879 1880 1881 |
* * This function allocates a page from the kernel page pool and applies * a NUMA policy associated with the VMA or the current process. * When VMA is not NULL caller must hold down_read on the mmap_sem of the * mm_struct of the VMA to prevent it from going away. Should be used for |
be97a41b2 mm/mempolicy.c: m... |
1882 1883 |
* all allocations for pages that will be mapped into user space. Returns * NULL when no page can be allocated. |
1da177e4c Linux-2.6.12-rc2 |
1884 1885 |
*/ struct page * |
0bbbc0b33 thp: add numa awa... |
1886 |
alloc_pages_vma(gfp_t gfp, int order, struct vm_area_struct *vma, |
be97a41b2 mm/mempolicy.c: m... |
1887 |
unsigned long addr, int node, bool hugepage) |
1da177e4c Linux-2.6.12-rc2 |
1888 |
{ |
cc9a6c877 cpuset: mm: reduc... |
1889 |
struct mempolicy *pol; |
c0ff7453b cpuset,mm: fix no... |
1890 |
struct page *page; |
cc9a6c877 cpuset: mm: reduc... |
1891 |
unsigned int cpuset_mems_cookie; |
be97a41b2 mm/mempolicy.c: m... |
1892 1893 |
struct zonelist *zl; nodemask_t *nmask; |
cc9a6c877 cpuset: mm: reduc... |
1894 1895 |
retry_cpuset: |
dd6eecb91 mempolicy: unexpo... |
1896 |
pol = get_vma_policy(vma, addr); |
d26914d11 mm: optimize put_... |
1897 |
cpuset_mems_cookie = read_mems_allowed_begin(); |
1da177e4c Linux-2.6.12-rc2 |
1898 |
|
0867a57c4 mm, thp: respect ... |
1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 |
if (pol->mode == MPOL_INTERLEAVE) { unsigned nid; nid = interleave_nid(pol, vma, addr, PAGE_SHIFT + order); mpol_cond_put(pol); page = alloc_page_interleave(gfp, order, nid); goto out; } if (unlikely(IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && hugepage)) { int hpage_node = node; |
be97a41b2 mm/mempolicy.c: m... |
1910 1911 |
/* * For hugepage allocation and non-interleave policy which |
0867a57c4 mm, thp: respect ... |
1912 1913 1914 1915 |
* allows the current node (or other explicitly preferred * node) we only try to allocate from the current/preferred * node and don't fall back to other nodes, as the cost of * remote accesses would likely offset THP benefits. |
be97a41b2 mm/mempolicy.c: m... |
1916 1917 1918 1919 |
* * If the policy is interleave, or does not allow the current * node in its nodemask, we allocate the standard way. */ |
0867a57c4 mm, thp: respect ... |
1920 1921 1922 |
if (pol->mode == MPOL_PREFERRED && !(pol->flags & MPOL_F_LOCAL)) hpage_node = pol->v.preferred_node; |
be97a41b2 mm/mempolicy.c: m... |
1923 |
nmask = policy_nodemask(gfp, pol); |
0867a57c4 mm, thp: respect ... |
1924 |
if (!nmask || node_isset(hpage_node, *nmask)) { |
be97a41b2 mm/mempolicy.c: m... |
1925 |
mpol_cond_put(pol); |
96db800f5 mm: rename alloc_... |
1926 |
page = __alloc_pages_node(hpage_node, |
5265047ac mm, thp: really l... |
1927 |
gfp | __GFP_THISNODE, order); |
be97a41b2 mm/mempolicy.c: m... |
1928 1929 1930 |
goto out; } } |
be97a41b2 mm/mempolicy.c: m... |
1931 1932 |
nmask = policy_nodemask(gfp, pol); zl = policy_zonelist(gfp, pol, node); |
be97a41b2 mm/mempolicy.c: m... |
1933 |
page = __alloc_pages_nodemask(gfp, order, zl, nmask); |
9b1a1ae9b mm/mempolicy.c: d... |
1934 |
mpol_cond_put(pol); |
be97a41b2 mm/mempolicy.c: m... |
1935 |
out: |
d26914d11 mm: optimize put_... |
1936 |
if (unlikely(!page && read_mems_allowed_retry(cpuset_mems_cookie))) |
cc9a6c877 cpuset: mm: reduc... |
1937 |
goto retry_cpuset; |
c0ff7453b cpuset,mm: fix no... |
1938 |
return page; |
1da177e4c Linux-2.6.12-rc2 |
1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 |
} /** * alloc_pages_current - Allocate pages. * * @gfp: * %GFP_USER user allocation, * %GFP_KERNEL kernel allocation, * %GFP_HIGHMEM highmem allocation, * %GFP_FS don't call back into a file system. * %GFP_ATOMIC don't sleep. * @order: Power of two of allocation size in pages. 0 is a single page. * * Allocate a page from the kernel page pool. When not in * interrupt context and apply the current process NUMA policy. * Returns NULL when no page can be allocated. * |
cf2a473c4 [PATCH] cpuset: c... |
1956 |
* Don't call cpuset_update_task_memory_state() unless |
1da177e4c Linux-2.6.12-rc2 |
1957 1958 1959 |
* 1) it's ok to take cpuset_sem (can WAIT), and * 2) allocating for current task (not interrupt). */ |
dd0fc66fb [PATCH] gfp flags... |
1960 |
struct page *alloc_pages_current(gfp_t gfp, unsigned order) |
1da177e4c Linux-2.6.12-rc2 |
1961 |
{ |
8d90274b3 mempolicy: saniti... |
1962 |
struct mempolicy *pol = &default_policy; |
c0ff7453b cpuset,mm: fix no... |
1963 |
struct page *page; |
cc9a6c877 cpuset: mm: reduc... |
1964 |
unsigned int cpuset_mems_cookie; |
1da177e4c Linux-2.6.12-rc2 |
1965 |
|
8d90274b3 mempolicy: saniti... |
1966 1967 |
if (!in_interrupt() && !(gfp & __GFP_THISNODE)) pol = get_task_policy(current); |
52cd3b074 mempolicy: rework... |
1968 |
|
cc9a6c877 cpuset: mm: reduc... |
1969 |
retry_cpuset: |
d26914d11 mm: optimize put_... |
1970 |
cpuset_mems_cookie = read_mems_allowed_begin(); |
cc9a6c877 cpuset: mm: reduc... |
1971 |
|
52cd3b074 mempolicy: rework... |
1972 1973 1974 1975 |
/* * No reference counting needed for current->mempolicy * nor system default_policy */ |
45c4745af mempolicy: rename... |
1976 |
if (pol->mode == MPOL_INTERLEAVE) |
c0ff7453b cpuset,mm: fix no... |
1977 1978 1979 |
page = alloc_page_interleave(gfp, order, interleave_nodes(pol)); else page = __alloc_pages_nodemask(gfp, order, |
5c4b4be3b mm: use correct n... |
1980 1981 |
policy_zonelist(gfp, pol, numa_node_id()), policy_nodemask(gfp, pol)); |
cc9a6c877 cpuset: mm: reduc... |
1982 |
|
d26914d11 mm: optimize put_... |
1983 |
if (unlikely(!page && read_mems_allowed_retry(cpuset_mems_cookie))) |
cc9a6c877 cpuset: mm: reduc... |
1984 |
goto retry_cpuset; |
c0ff7453b cpuset,mm: fix no... |
1985 |
return page; |
1da177e4c Linux-2.6.12-rc2 |
1986 1987 |
} EXPORT_SYMBOL(alloc_pages_current); |
ef0855d33 mm: mempolicy: tu... |
1988 1989 1990 1991 1992 1993 1994 1995 1996 |
int vma_dup_policy(struct vm_area_struct *src, struct vm_area_struct *dst) { struct mempolicy *pol = mpol_dup(vma_policy(src)); if (IS_ERR(pol)) return PTR_ERR(pol); dst->vm_policy = pol; return 0; } |
4225399a6 [PATCH] cpuset: r... |
1997 |
/* |
846a16bf0 mempolicy: rename... |
1998 |
* If mpol_dup() sees current->cpuset == cpuset_being_rebound, then it |
4225399a6 [PATCH] cpuset: r... |
1999 2000 2001 2002 |
* rebinds the mempolicy its copying by calling mpol_rebind_policy() * with the mems_allowed returned by cpuset_mems_allowed(). This * keeps mempolicies cpuset relative after its cpuset moves. See * further kernel/cpuset.c update_nodemask(). |
708c1bbc9 mempolicy: restru... |
2003 2004 2005 |
* * current's mempolicy may be rebinded by the other task(the task that changes * cpuset's mems), so we needn't do rebind work for current task. |
4225399a6 [PATCH] cpuset: r... |
2006 |
*/ |
4225399a6 [PATCH] cpuset: r... |
2007 |
|
846a16bf0 mempolicy: rename... |
2008 2009 |
/* Slow path of a mempolicy duplicate */ struct mempolicy *__mpol_dup(struct mempolicy *old) |
1da177e4c Linux-2.6.12-rc2 |
2010 2011 2012 2013 2014 |
{ struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL); if (!new) return ERR_PTR(-ENOMEM); |
708c1bbc9 mempolicy: restru... |
2015 2016 2017 2018 2019 2020 2021 2022 |
/* task's mempolicy is protected by alloc_lock */ if (old == current->mempolicy) { task_lock(current); *new = *old; task_unlock(current); } else *new = *old; |
4225399a6 [PATCH] cpuset: r... |
2023 2024 |
if (current_cpuset_is_being_rebound()) { nodemask_t mems = cpuset_mems_allowed(current); |
708c1bbc9 mempolicy: restru... |
2025 2026 2027 2028 |
if (new->flags & MPOL_F_REBINDING) mpol_rebind_policy(new, &mems, MPOL_REBIND_STEP2); else mpol_rebind_policy(new, &mems, MPOL_REBIND_ONCE); |
4225399a6 [PATCH] cpuset: r... |
2029 |
} |
1da177e4c Linux-2.6.12-rc2 |
2030 |
atomic_set(&new->refcnt, 1); |
1da177e4c Linux-2.6.12-rc2 |
2031 2032 2033 2034 |
return new; } /* Slow path of a mempolicy comparison */ |
fcfb4dcc9 mm/mempolicy.c: m... |
2035 |
bool __mpol_equal(struct mempolicy *a, struct mempolicy *b) |
1da177e4c Linux-2.6.12-rc2 |
2036 2037 |
{ if (!a || !b) |
fcfb4dcc9 mm/mempolicy.c: m... |
2038 |
return false; |
45c4745af mempolicy: rename... |
2039 |
if (a->mode != b->mode) |
fcfb4dcc9 mm/mempolicy.c: m... |
2040 |
return false; |
198005025 mempolicy: remove... |
2041 |
if (a->flags != b->flags) |
fcfb4dcc9 mm/mempolicy.c: m... |
2042 |
return false; |
198005025 mempolicy: remove... |
2043 2044 |
if (mpol_store_user_nodemask(a)) if (!nodes_equal(a->w.user_nodemask, b->w.user_nodemask)) |
fcfb4dcc9 mm/mempolicy.c: m... |
2045 |
return false; |
198005025 mempolicy: remove... |
2046 |
|
45c4745af mempolicy: rename... |
2047 |
switch (a->mode) { |
19770b326 mm: filter based ... |
2048 2049 |
case MPOL_BIND: /* Fall through */ |
1da177e4c Linux-2.6.12-rc2 |
2050 |
case MPOL_INTERLEAVE: |
fcfb4dcc9 mm/mempolicy.c: m... |
2051 |
return !!nodes_equal(a->v.nodes, b->v.nodes); |
1da177e4c Linux-2.6.12-rc2 |
2052 |
case MPOL_PREFERRED: |
757196618 mempolicy: remove... |
2053 |
return a->v.preferred_node == b->v.preferred_node; |
1da177e4c Linux-2.6.12-rc2 |
2054 2055 |
default: BUG(); |
fcfb4dcc9 mm/mempolicy.c: m... |
2056 |
return false; |
1da177e4c Linux-2.6.12-rc2 |
2057 2058 |
} } |
1da177e4c Linux-2.6.12-rc2 |
2059 |
/* |
1da177e4c Linux-2.6.12-rc2 |
2060 2061 2062 2063 |
* Shared memory backing store policy support. * * Remember policies even when nobody has shared memory mapped. * The policies are kept in Red-Black tree linked from the inode. |
4a8c7bb59 mm/mempolicy.c: c... |
2064 |
* They are protected by the sp->lock rwlock, which should be held |
1da177e4c Linux-2.6.12-rc2 |
2065 2066 |
* for any accesses to the tree. */ |
4a8c7bb59 mm/mempolicy.c: c... |
2067 2068 2069 2070 |
/* * lookup first element intersecting start-end. Caller holds sp->lock for * reading or for writing */ |
1da177e4c Linux-2.6.12-rc2 |
2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 |
static struct sp_node * sp_lookup(struct shared_policy *sp, unsigned long start, unsigned long end) { struct rb_node *n = sp->root.rb_node; while (n) { struct sp_node *p = rb_entry(n, struct sp_node, nd); if (start >= p->end) n = n->rb_right; else if (end <= p->start) n = n->rb_left; else break; } if (!n) return NULL; for (;;) { struct sp_node *w = NULL; struct rb_node *prev = rb_prev(n); if (!prev) break; w = rb_entry(prev, struct sp_node, nd); if (w->end <= start) break; n = prev; } return rb_entry(n, struct sp_node, nd); } |
4a8c7bb59 mm/mempolicy.c: c... |
2100 2101 2102 2103 |
/* * Insert a new shared policy into the list. Caller holds sp->lock for * writing. */ |
1da177e4c Linux-2.6.12-rc2 |
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 |
static void sp_insert(struct shared_policy *sp, struct sp_node *new) { struct rb_node **p = &sp->root.rb_node; struct rb_node *parent = NULL; struct sp_node *nd; while (*p) { parent = *p; nd = rb_entry(parent, struct sp_node, nd); if (new->start < nd->start) p = &(*p)->rb_left; else if (new->end > nd->end) p = &(*p)->rb_right; else BUG(); } rb_link_node(&new->nd, parent, p); rb_insert_color(&new->nd, &sp->root); |
140d5a490 numa: mempolicy: ... |
2122 2123 |
pr_debug("inserting %lx-%lx: %d ", new->start, new->end, |
45c4745af mempolicy: rename... |
2124 |
new->policy ? new->policy->mode : 0); |
1da177e4c Linux-2.6.12-rc2 |
2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 |
} /* Find shared policy intersecting idx */ struct mempolicy * mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx) { struct mempolicy *pol = NULL; struct sp_node *sn; if (!sp->root.rb_node) return NULL; |
4a8c7bb59 mm/mempolicy.c: c... |
2136 |
read_lock(&sp->lock); |
1da177e4c Linux-2.6.12-rc2 |
2137 2138 2139 2140 2141 |
sn = sp_lookup(sp, idx, idx+1); if (sn) { mpol_get(sn->policy); pol = sn->policy; } |
4a8c7bb59 mm/mempolicy.c: c... |
2142 |
read_unlock(&sp->lock); |
1da177e4c Linux-2.6.12-rc2 |
2143 2144 |
return pol; } |
63f74ca21 mempolicy: fix re... |
2145 2146 2147 2148 2149 |
static void sp_free(struct sp_node *n) { mpol_put(n->policy); kmem_cache_free(sn_cache, n); } |
771fb4d80 mm: mempolicy: Ch... |
2150 2151 2152 |
/** * mpol_misplaced - check whether current page node is valid in policy * |
b46e14acb mm/mempolicy.c: p... |
2153 2154 2155 |
* @page: page to be checked * @vma: vm area where page mapped * @addr: virtual address where page mapped |
771fb4d80 mm: mempolicy: Ch... |
2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 |
* * Lookup current policy node id for vma,addr and "compare to" page's * node id. * * Returns: * -1 - not misplaced, page is in the right node * node - node id where the page should be * * Policy determination "mimics" alloc_page_vma(). * Called from fault path where we know the vma and faulting address. */ int mpol_misplaced(struct page *page, struct vm_area_struct *vma, unsigned long addr) { struct mempolicy *pol; |
c33d6c06f mm, page_alloc: a... |
2170 |
struct zoneref *z; |
771fb4d80 mm: mempolicy: Ch... |
2171 2172 |
int curnid = page_to_nid(page); unsigned long pgoff; |
90572890d mm: numa: Change ... |
2173 2174 |
int thiscpu = raw_smp_processor_id(); int thisnid = cpu_to_node(thiscpu); |
771fb4d80 mm: mempolicy: Ch... |
2175 2176 2177 2178 |
int polnid = -1; int ret = -1; BUG_ON(!vma); |
dd6eecb91 mempolicy: unexpo... |
2179 |
pol = get_vma_policy(vma, addr); |
771fb4d80 mm: mempolicy: Ch... |
2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 |
if (!(pol->flags & MPOL_F_MOF)) goto out; switch (pol->mode) { case MPOL_INTERLEAVE: BUG_ON(addr >= vma->vm_end); BUG_ON(addr < vma->vm_start); pgoff = vma->vm_pgoff; pgoff += (addr - vma->vm_start) >> PAGE_SHIFT; polnid = offset_il_node(pol, vma, pgoff); break; case MPOL_PREFERRED: if (pol->flags & MPOL_F_LOCAL) polnid = numa_node_id(); else polnid = pol->v.preferred_node; break; case MPOL_BIND: |
c33d6c06f mm, page_alloc: a... |
2201 |
|
771fb4d80 mm: mempolicy: Ch... |
2202 2203 2204 2205 2206 2207 2208 2209 |
/* * allows binding to multiple nodes. * use current page if in policy nodemask, * else select nearest allowed node, if any. * If no allowed nodes, use current [!misplaced]. */ if (node_isset(curnid, pol->v.nodes)) goto out; |
c33d6c06f mm, page_alloc: a... |
2210 |
z = first_zones_zonelist( |
771fb4d80 mm: mempolicy: Ch... |
2211 2212 |
node_zonelist(numa_node_id(), GFP_HIGHUSER), gfp_zone(GFP_HIGHUSER), |
c33d6c06f mm, page_alloc: a... |
2213 2214 |
&pol->v.nodes); polnid = z->zone->node; |
771fb4d80 mm: mempolicy: Ch... |
2215 2216 2217 2218 2219 |
break; default: BUG(); } |
5606e3877 mm: numa: Migrate... |
2220 2221 |
/* Migrate the page towards the node whose CPU is referencing it */ |
e42c8ff29 mm: numa: Use a t... |
2222 |
if (pol->flags & MPOL_F_MORON) { |
90572890d mm: numa: Change ... |
2223 |
polnid = thisnid; |
5606e3877 mm: numa: Migrate... |
2224 |
|
10f390427 sched/numa, mm: U... |
2225 |
if (!should_numa_migrate_memory(current, page, curnid, thiscpu)) |
de1c9ce6f sched/numa: Skip ... |
2226 |
goto out; |
e42c8ff29 mm: numa: Use a t... |
2227 |
} |
771fb4d80 mm: mempolicy: Ch... |
2228 2229 2230 2231 2232 2233 2234 |
if (curnid != polnid) ret = polnid; out: mpol_cond_put(pol); return ret; } |
c11600e4f mm, mempolicy: ta... |
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 |
/* * Drop the (possibly final) reference to task->mempolicy. It needs to be * dropped after task->mempolicy is set to NULL so that any allocation done as * part of its kmem_cache_free(), such as by KASAN, doesn't reference a freed * policy. */ void mpol_put_task_policy(struct task_struct *task) { struct mempolicy *pol; task_lock(task); pol = task->mempolicy; task->mempolicy = NULL; task_unlock(task); mpol_put(pol); } |
1da177e4c Linux-2.6.12-rc2 |
2251 2252 |
static void sp_delete(struct shared_policy *sp, struct sp_node *n) { |
140d5a490 numa: mempolicy: ... |
2253 2254 |
pr_debug("deleting %lx-l%lx ", n->start, n->end); |
1da177e4c Linux-2.6.12-rc2 |
2255 |
rb_erase(&n->nd, &sp->root); |
63f74ca21 mempolicy: fix re... |
2256 |
sp_free(n); |
1da177e4c Linux-2.6.12-rc2 |
2257 |
} |
42288fe36 mm: mempolicy: Co... |
2258 2259 2260 2261 2262 2263 2264 |
static void sp_node_init(struct sp_node *node, unsigned long start, unsigned long end, struct mempolicy *pol) { node->start = start; node->end = end; node->policy = pol; } |
dbcb0f19c mm/mempolicy.c: c... |
2265 2266 |
static struct sp_node *sp_alloc(unsigned long start, unsigned long end, struct mempolicy *pol) |
1da177e4c Linux-2.6.12-rc2 |
2267 |
{ |
869833f2c mempolicy: remove... |
2268 2269 |
struct sp_node *n; struct mempolicy *newpol; |
1da177e4c Linux-2.6.12-rc2 |
2270 |
|
869833f2c mempolicy: remove... |
2271 |
n = kmem_cache_alloc(sn_cache, GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
2272 2273 |
if (!n) return NULL; |
869833f2c mempolicy: remove... |
2274 2275 2276 2277 2278 2279 2280 |
newpol = mpol_dup(pol); if (IS_ERR(newpol)) { kmem_cache_free(sn_cache, n); return NULL; } newpol->flags |= MPOL_F_SHARED; |
42288fe36 mm: mempolicy: Co... |
2281 |
sp_node_init(n, start, end, newpol); |
869833f2c mempolicy: remove... |
2282 |
|
1da177e4c Linux-2.6.12-rc2 |
2283 2284 2285 2286 2287 2288 2289 |
return n; } /* Replace a policy range. */ static int shared_policy_replace(struct shared_policy *sp, unsigned long start, unsigned long end, struct sp_node *new) { |
b22d127a3 mempolicy: fix a ... |
2290 |
struct sp_node *n; |
42288fe36 mm: mempolicy: Co... |
2291 2292 |
struct sp_node *n_new = NULL; struct mempolicy *mpol_new = NULL; |
b22d127a3 mempolicy: fix a ... |
2293 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
2294 |
|
42288fe36 mm: mempolicy: Co... |
2295 |
restart: |
4a8c7bb59 mm/mempolicy.c: c... |
2296 |
write_lock(&sp->lock); |
1da177e4c Linux-2.6.12-rc2 |
2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 |
n = sp_lookup(sp, start, end); /* Take care of old policies in the same range. */ while (n && n->start < end) { struct rb_node *next = rb_next(&n->nd); if (n->start >= start) { if (n->end <= end) sp_delete(sp, n); else n->start = end; } else { /* Old policy spanning whole new range. */ if (n->end > end) { |
42288fe36 mm: mempolicy: Co... |
2309 2310 2311 2312 2313 |
if (!n_new) goto alloc_new; *mpol_new = *n->policy; atomic_set(&mpol_new->refcnt, 1); |
7880639c3 mm/mempolicy.c: f... |
2314 |
sp_node_init(n_new, end, n->end, mpol_new); |
1da177e4c Linux-2.6.12-rc2 |
2315 |
n->end = start; |
5ca395751 mm/mempolicy.c: f... |
2316 |
sp_insert(sp, n_new); |
42288fe36 mm: mempolicy: Co... |
2317 2318 |
n_new = NULL; mpol_new = NULL; |
1da177e4c Linux-2.6.12-rc2 |
2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 |
break; } else n->end = start; } if (!next) break; n = rb_entry(next, struct sp_node, nd); } if (new) sp_insert(sp, new); |
4a8c7bb59 mm/mempolicy.c: c... |
2329 |
write_unlock(&sp->lock); |
42288fe36 mm: mempolicy: Co... |
2330 2331 2332 2333 2334 2335 2336 |
ret = 0; err_out: if (mpol_new) mpol_put(mpol_new); if (n_new) kmem_cache_free(sn_cache, n_new); |
b22d127a3 mempolicy: fix a ... |
2337 |
return ret; |
42288fe36 mm: mempolicy: Co... |
2338 2339 |
alloc_new: |
4a8c7bb59 mm/mempolicy.c: c... |
2340 |
write_unlock(&sp->lock); |
42288fe36 mm: mempolicy: Co... |
2341 2342 2343 2344 2345 2346 2347 2348 |
ret = -ENOMEM; n_new = kmem_cache_alloc(sn_cache, GFP_KERNEL); if (!n_new) goto err_out; mpol_new = kmem_cache_alloc(policy_cache, GFP_KERNEL); if (!mpol_new) goto err_out; goto restart; |
1da177e4c Linux-2.6.12-rc2 |
2349 |
} |
71fe804b6 mempolicy: use st... |
2350 2351 2352 2353 2354 2355 2356 2357 |
/** * mpol_shared_policy_init - initialize shared policy for inode * @sp: pointer to inode shared policy * @mpol: struct mempolicy to install * * Install non-NULL @mpol in inode's shared policy rb-tree. * On entry, the current task has a reference on a non-NULL @mpol. * This must be released on exit. |
4bfc44958 mm: make set_memp... |
2358 |
* This is called at get_inode() calls and we can use GFP_KERNEL. |
71fe804b6 mempolicy: use st... |
2359 2360 2361 |
*/ void mpol_shared_policy_init(struct shared_policy *sp, struct mempolicy *mpol) { |
58568d2a8 cpuset,mm: update... |
2362 |
int ret; |
71fe804b6 mempolicy: use st... |
2363 |
sp->root = RB_ROOT; /* empty tree == default mempolicy */ |
4a8c7bb59 mm/mempolicy.c: c... |
2364 |
rwlock_init(&sp->lock); |
71fe804b6 mempolicy: use st... |
2365 2366 2367 2368 |
if (mpol) { struct vm_area_struct pvma; struct mempolicy *new; |
4bfc44958 mm: make set_memp... |
2369 |
NODEMASK_SCRATCH(scratch); |
71fe804b6 mempolicy: use st... |
2370 |
|
4bfc44958 mm: make set_memp... |
2371 |
if (!scratch) |
5c0c16549 mempolicy: fix da... |
2372 |
goto put_mpol; |
71fe804b6 mempolicy: use st... |
2373 2374 |
/* contextualize the tmpfs mount point mempolicy */ new = mpol_new(mpol->mode, mpol->flags, &mpol->w.user_nodemask); |
15d77835a mempolicy: factor... |
2375 |
if (IS_ERR(new)) |
0cae3457b mempolicy: ERR_PT... |
2376 |
goto free_scratch; /* no valid nodemask intersection */ |
58568d2a8 cpuset,mm: update... |
2377 2378 |
task_lock(current); |
4bfc44958 mm: make set_memp... |
2379 |
ret = mpol_set_nodemask(new, &mpol->w.user_nodemask, scratch); |
58568d2a8 cpuset,mm: update... |
2380 |
task_unlock(current); |
15d77835a mempolicy: factor... |
2381 |
if (ret) |
5c0c16549 mempolicy: fix da... |
2382 |
goto put_new; |
71fe804b6 mempolicy: use st... |
2383 2384 2385 2386 2387 |
/* Create pseudo-vma that contains just the policy */ memset(&pvma, 0, sizeof(struct vm_area_struct)); pvma.vm_end = TASK_SIZE; /* policy covers entire file */ mpol_set_shared_policy(sp, &pvma, new); /* adds ref */ |
15d77835a mempolicy: factor... |
2388 |
|
5c0c16549 mempolicy: fix da... |
2389 |
put_new: |
71fe804b6 mempolicy: use st... |
2390 |
mpol_put(new); /* drop initial ref */ |
0cae3457b mempolicy: ERR_PT... |
2391 |
free_scratch: |
4bfc44958 mm: make set_memp... |
2392 |
NODEMASK_SCRATCH_FREE(scratch); |
5c0c16549 mempolicy: fix da... |
2393 2394 |
put_mpol: mpol_put(mpol); /* drop our incoming ref on sb mpol */ |
7339ff830 [PATCH] Add tmpfs... |
2395 2396 |
} } |
1da177e4c Linux-2.6.12-rc2 |
2397 2398 2399 2400 2401 2402 |
int mpol_set_shared_policy(struct shared_policy *info, struct vm_area_struct *vma, struct mempolicy *npol) { int err; struct sp_node *new = NULL; unsigned long sz = vma_pages(vma); |
028fec414 mempolicy: suppor... |
2403 2404 |
pr_debug("set_shared_policy %lx sz %lu %d %d %lx ", |
1da177e4c Linux-2.6.12-rc2 |
2405 |
vma->vm_pgoff, |
45c4745af mempolicy: rename... |
2406 |
sz, npol ? npol->mode : -1, |
028fec414 mempolicy: suppor... |
2407 |
npol ? npol->flags : -1, |
00ef2d2f8 mm: use NUMA_NO_NODE |
2408 |
npol ? nodes_addr(npol->v.nodes)[0] : NUMA_NO_NODE); |
1da177e4c Linux-2.6.12-rc2 |
2409 2410 2411 2412 2413 2414 2415 2416 |
if (npol) { new = sp_alloc(vma->vm_pgoff, vma->vm_pgoff + sz, npol); if (!new) return -ENOMEM; } err = shared_policy_replace(info, vma->vm_pgoff, vma->vm_pgoff+sz, new); if (err && new) |
63f74ca21 mempolicy: fix re... |
2417 |
sp_free(new); |
1da177e4c Linux-2.6.12-rc2 |
2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 |
return err; } /* Free a backing policy store on inode delete. */ void mpol_free_shared_policy(struct shared_policy *p) { struct sp_node *n; struct rb_node *next; if (!p->root.rb_node) return; |
4a8c7bb59 mm/mempolicy.c: c... |
2429 |
write_lock(&p->lock); |
1da177e4c Linux-2.6.12-rc2 |
2430 2431 2432 2433 |
next = rb_first(&p->root); while (next) { n = rb_entry(next, struct sp_node, nd); next = rb_next(&n->nd); |
63f74ca21 mempolicy: fix re... |
2434 |
sp_delete(p, n); |
1da177e4c Linux-2.6.12-rc2 |
2435 |
} |
4a8c7bb59 mm/mempolicy.c: c... |
2436 |
write_unlock(&p->lock); |
1da177e4c Linux-2.6.12-rc2 |
2437 |
} |
1a687c2e9 mm: sched: numa: ... |
2438 |
#ifdef CONFIG_NUMA_BALANCING |
c297663c0 mm: numa: initial... |
2439 |
static int __initdata numabalancing_override; |
1a687c2e9 mm: sched: numa: ... |
2440 2441 2442 2443 2444 2445 2446 |
static void __init check_numabalancing_enable(void) { bool numabalancing_default = false; if (IS_ENABLED(CONFIG_NUMA_BALANCING_DEFAULT_ENABLED)) numabalancing_default = true; |
c297663c0 mm: numa: initial... |
2447 2448 2449 |
/* Parsed by setup_numabalancing. override == 1 enables, -1 disables */ if (numabalancing_override) set_numabalancing_state(numabalancing_override == 1); |
b0dc2b9bb mm, numa: really ... |
2450 |
if (num_online_nodes() > 1 && !numabalancing_override) { |
756a025f0 mm: coalesce spli... |
2451 2452 |
pr_info("%s automatic NUMA balancing. Configure with numa_balancing= or the kernel.numa_balancing sysctl ", |
c297663c0 mm: numa: initial... |
2453 |
numabalancing_default ? "Enabling" : "Disabling"); |
1a687c2e9 mm: sched: numa: ... |
2454 2455 2456 2457 2458 2459 2460 2461 2462 |
set_numabalancing_state(numabalancing_default); } } static int __init setup_numabalancing(char *str) { int ret = 0; if (!str) goto out; |
1a687c2e9 mm: sched: numa: ... |
2463 2464 |
if (!strcmp(str, "enable")) { |
c297663c0 mm: numa: initial... |
2465 |
numabalancing_override = 1; |
1a687c2e9 mm: sched: numa: ... |
2466 2467 |
ret = 1; } else if (!strcmp(str, "disable")) { |
c297663c0 mm: numa: initial... |
2468 |
numabalancing_override = -1; |
1a687c2e9 mm: sched: numa: ... |
2469 2470 2471 2472 |
ret = 1; } out: if (!ret) |
4a404bea9 mm/mempolicy.c: c... |
2473 2474 |
pr_warn("Unable to parse numa_balancing= "); |
1a687c2e9 mm: sched: numa: ... |
2475 2476 2477 2478 2479 2480 2481 2482 2483 |
return ret; } __setup("numa_balancing=", setup_numabalancing); #else static inline void __init check_numabalancing_enable(void) { } #endif /* CONFIG_NUMA_BALANCING */ |
1da177e4c Linux-2.6.12-rc2 |
2484 2485 2486 |
/* assumes fs == KERNEL_DS */ void __init numa_policy_init(void) { |
b71636e29 numa: mempolicy: ... |
2487 2488 2489 |
nodemask_t interleave_nodes; unsigned long largest = 0; int nid, prefer = 0; |
1da177e4c Linux-2.6.12-rc2 |
2490 2491 |
policy_cache = kmem_cache_create("numa_policy", sizeof(struct mempolicy), |
20c2df83d mm: Remove slab d... |
2492 |
0, SLAB_PANIC, NULL); |
1da177e4c Linux-2.6.12-rc2 |
2493 2494 2495 |
sn_cache = kmem_cache_create("shared_policy_node", sizeof(struct sp_node), |
20c2df83d mm: Remove slab d... |
2496 |
0, SLAB_PANIC, NULL); |
1da177e4c Linux-2.6.12-rc2 |
2497 |
|
5606e3877 mm: numa: Migrate... |
2498 2499 2500 2501 2502 2503 2504 2505 |
for_each_node(nid) { preferred_node_policy[nid] = (struct mempolicy) { .refcnt = ATOMIC_INIT(1), .mode = MPOL_PREFERRED, .flags = MPOL_F_MOF | MPOL_F_MORON, .v = { .preferred_node = nid, }, }; } |
b71636e29 numa: mempolicy: ... |
2506 2507 2508 2509 2510 2511 |
/* * Set interleaving policy for system init. Interleaving is only * enabled across suitably sized nodes (default is >= 16MB), or * fall back to the largest node if they're all smaller. */ nodes_clear(interleave_nodes); |
01f13bd60 mempolicy: use N_... |
2512 |
for_each_node_state(nid, N_MEMORY) { |
b71636e29 numa: mempolicy: ... |
2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 |
unsigned long total_pages = node_present_pages(nid); /* Preserve the largest node */ if (largest < total_pages) { largest = total_pages; prefer = nid; } /* Interleave this node? */ if ((total_pages << PAGE_SHIFT) >= (16 << 20)) node_set(nid, interleave_nodes); } /* All too small, use the largest */ if (unlikely(nodes_empty(interleave_nodes))) node_set(prefer, interleave_nodes); |
1da177e4c Linux-2.6.12-rc2 |
2529 |
|
028fec414 mempolicy: suppor... |
2530 |
if (do_set_mempolicy(MPOL_INTERLEAVE, 0, &interleave_nodes)) |
b1de0d139 mm: convert some ... |
2531 2532 |
pr_err("%s: interleaving failed ", __func__); |
1a687c2e9 mm: sched: numa: ... |
2533 2534 |
check_numabalancing_enable(); |
1da177e4c Linux-2.6.12-rc2 |
2535 |
} |
8bccd85ff [PATCH] Implement... |
2536 |
/* Reset policy of current process to default */ |
1da177e4c Linux-2.6.12-rc2 |
2537 2538 |
void numa_default_policy(void) { |
028fec414 mempolicy: suppor... |
2539 |
do_set_mempolicy(MPOL_DEFAULT, 0, NULL); |
1da177e4c Linux-2.6.12-rc2 |
2540 |
} |
68860ec10 [PATCH] cpusets: ... |
2541 |
|
4225399a6 [PATCH] cpuset: r... |
2542 |
/* |
095f1fc4e mempolicy: rework... |
2543 2544 2545 2546 |
* Parse and format mempolicy from/to strings */ /* |
f2a07f40d tmpfs mempolicy: ... |
2547 |
* "local" is implemented internally by MPOL_PREFERRED with MPOL_F_LOCAL flag. |
1a75a6c82 [PATCH] Fold numa... |
2548 |
*/ |
345ace9c7 mempolicy: rename... |
2549 2550 2551 2552 2553 2554 |
static const char * const policy_modes[] = { [MPOL_DEFAULT] = "default", [MPOL_PREFERRED] = "prefer", [MPOL_BIND] = "bind", [MPOL_INTERLEAVE] = "interleave", |
d3a710337 mm: mempolicy: Ad... |
2555 |
[MPOL_LOCAL] = "local", |
345ace9c7 mempolicy: rename... |
2556 |
}; |
1a75a6c82 [PATCH] Fold numa... |
2557 |
|
095f1fc4e mempolicy: rework... |
2558 2559 2560 |
#ifdef CONFIG_TMPFS /** |
f2a07f40d tmpfs mempolicy: ... |
2561 |
* mpol_parse_str - parse string to mempolicy, for tmpfs mpol mount option. |
095f1fc4e mempolicy: rework... |
2562 |
* @str: string containing mempolicy to parse |
71fe804b6 mempolicy: use st... |
2563 |
* @mpol: pointer to struct mempolicy pointer, returned on success. |
095f1fc4e mempolicy: rework... |
2564 2565 2566 2567 |
* * Format of input: * <mode>[=<flags>][:<nodelist>] * |
71fe804b6 mempolicy: use st... |
2568 |
* On success, returns 0, else 1 |
095f1fc4e mempolicy: rework... |
2569 |
*/ |
a7a88b237 mempolicy: remove... |
2570 |
int mpol_parse_str(char *str, struct mempolicy **mpol) |
095f1fc4e mempolicy: rework... |
2571 |
{ |
71fe804b6 mempolicy: use st... |
2572 |
struct mempolicy *new = NULL; |
b4652e842 mempolicy: lose u... |
2573 |
unsigned short mode; |
f2a07f40d tmpfs mempolicy: ... |
2574 |
unsigned short mode_flags; |
71fe804b6 mempolicy: use st... |
2575 |
nodemask_t nodes; |
095f1fc4e mempolicy: rework... |
2576 2577 |
char *nodelist = strchr(str, ':'); char *flags = strchr(str, '='); |
095f1fc4e mempolicy: rework... |
2578 2579 2580 2581 2582 |
int err = 1; if (nodelist) { /* NUL-terminate mode or flags string */ *nodelist++ = '\0'; |
71fe804b6 mempolicy: use st... |
2583 |
if (nodelist_parse(nodelist, nodes)) |
095f1fc4e mempolicy: rework... |
2584 |
goto out; |
01f13bd60 mempolicy: use N_... |
2585 |
if (!nodes_subset(nodes, node_states[N_MEMORY])) |
095f1fc4e mempolicy: rework... |
2586 |
goto out; |
71fe804b6 mempolicy: use st... |
2587 2588 |
} else nodes_clear(nodes); |
095f1fc4e mempolicy: rework... |
2589 2590 |
if (flags) *flags++ = '\0'; /* terminate mode string */ |
479e2802d mm: mempolicy: Ma... |
2591 |
for (mode = 0; mode < MPOL_MAX; mode++) { |
345ace9c7 mempolicy: rename... |
2592 |
if (!strcmp(str, policy_modes[mode])) { |
095f1fc4e mempolicy: rework... |
2593 2594 2595 |
break; } } |
a720094de mm: mempolicy: Hi... |
2596 |
if (mode >= MPOL_MAX) |
095f1fc4e mempolicy: rework... |
2597 |
goto out; |
71fe804b6 mempolicy: use st... |
2598 |
switch (mode) { |
095f1fc4e mempolicy: rework... |
2599 |
case MPOL_PREFERRED: |
71fe804b6 mempolicy: use st... |
2600 2601 2602 |
/* * Insist on a nodelist of one node only */ |
095f1fc4e mempolicy: rework... |
2603 2604 2605 2606 |
if (nodelist) { char *rest = nodelist; while (isdigit(*rest)) rest++; |
926f2ae04 tmpfs: cleanup mp... |
2607 2608 |
if (*rest) goto out; |
095f1fc4e mempolicy: rework... |
2609 2610 |
} break; |
095f1fc4e mempolicy: rework... |
2611 2612 2613 2614 2615 |
case MPOL_INTERLEAVE: /* * Default to online nodes with memory if no nodelist */ if (!nodelist) |
01f13bd60 mempolicy: use N_... |
2616 |
nodes = node_states[N_MEMORY]; |
3f226aa1c mempolicy: suppor... |
2617 |
break; |
71fe804b6 mempolicy: use st... |
2618 |
case MPOL_LOCAL: |
3f226aa1c mempolicy: suppor... |
2619 |
/* |
71fe804b6 mempolicy: use st... |
2620 |
* Don't allow a nodelist; mpol_new() checks flags |
3f226aa1c mempolicy: suppor... |
2621 |
*/ |
71fe804b6 mempolicy: use st... |
2622 |
if (nodelist) |
3f226aa1c mempolicy: suppor... |
2623 |
goto out; |
71fe804b6 mempolicy: use st... |
2624 |
mode = MPOL_PREFERRED; |
3f226aa1c mempolicy: suppor... |
2625 |
break; |
413b43dea tmpfs: fix oops o... |
2626 2627 2628 2629 2630 2631 2632 |
case MPOL_DEFAULT: /* * Insist on a empty nodelist */ if (!nodelist) err = 0; goto out; |
d69b2e63e tmpfs: mpol=bind:... |
2633 2634 2635 2636 2637 2638 |
case MPOL_BIND: /* * Insist on a nodelist */ if (!nodelist) goto out; |
095f1fc4e mempolicy: rework... |
2639 |
} |
71fe804b6 mempolicy: use st... |
2640 |
mode_flags = 0; |
095f1fc4e mempolicy: rework... |
2641 2642 2643 2644 2645 2646 |
if (flags) { /* * Currently, we only support two mutually exclusive * mode flags. */ if (!strcmp(flags, "static")) |
71fe804b6 mempolicy: use st... |
2647 |
mode_flags |= MPOL_F_STATIC_NODES; |
095f1fc4e mempolicy: rework... |
2648 |
else if (!strcmp(flags, "relative")) |
71fe804b6 mempolicy: use st... |
2649 |
mode_flags |= MPOL_F_RELATIVE_NODES; |
095f1fc4e mempolicy: rework... |
2650 |
else |
926f2ae04 tmpfs: cleanup mp... |
2651 |
goto out; |
095f1fc4e mempolicy: rework... |
2652 |
} |
71fe804b6 mempolicy: use st... |
2653 2654 2655 |
new = mpol_new(mode, mode_flags, &nodes); if (IS_ERR(new)) |
926f2ae04 tmpfs: cleanup mp... |
2656 |
goto out; |
f2a07f40d tmpfs mempolicy: ... |
2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 |
/* * Save nodes for mpol_to_str() to show the tmpfs mount options * for /proc/mounts, /proc/pid/mounts and /proc/pid/mountinfo. */ if (mode != MPOL_PREFERRED) new->v.nodes = nodes; else if (nodelist) new->v.preferred_node = first_node(nodes); else new->flags |= MPOL_F_LOCAL; /* * Save nodes for contextualization: this will be used to "clone" * the mempolicy in a specific context [cpuset] at a later time. */ new->w.user_nodemask = nodes; |
926f2ae04 tmpfs: cleanup mp... |
2673 |
err = 0; |
71fe804b6 mempolicy: use st... |
2674 |
|
095f1fc4e mempolicy: rework... |
2675 2676 2677 2678 2679 2680 |
out: /* Restore string for error message */ if (nodelist) *--nodelist = ':'; if (flags) *--flags = '='; |
71fe804b6 mempolicy: use st... |
2681 2682 |
if (!err) *mpol = new; |
095f1fc4e mempolicy: rework... |
2683 2684 2685 |
return err; } #endif /* CONFIG_TMPFS */ |
71fe804b6 mempolicy: use st... |
2686 2687 2688 2689 2690 |
/** * mpol_to_str - format a mempolicy structure for printing * @buffer: to contain formatted mempolicy string * @maxlen: length of @buffer * @pol: pointer to mempolicy to be formatted |
71fe804b6 mempolicy: use st... |
2691 |
* |
948927ee9 mm, mempolicy: ma... |
2692 2693 2694 |
* Convert @pol into a string. If @buffer is too short, truncate the string. * Recommend a @maxlen of at least 32 for the longest mode, "interleave", the * longest flag, "relative", and to display at least a few node ids. |
1a75a6c82 [PATCH] Fold numa... |
2695 |
*/ |
948927ee9 mm, mempolicy: ma... |
2696 |
void mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol) |
1a75a6c82 [PATCH] Fold numa... |
2697 2698 |
{ char *p = buffer; |
948927ee9 mm, mempolicy: ma... |
2699 2700 2701 |
nodemask_t nodes = NODE_MASK_NONE; unsigned short mode = MPOL_DEFAULT; unsigned short flags = 0; |
2291990ab mempolicy: clean-... |
2702 |
|
8790c71a1 mm/mempolicy.c: f... |
2703 |
if (pol && pol != &default_policy && !(pol->flags & MPOL_F_MORON)) { |
bea904d54 mempolicy: use MP... |
2704 |
mode = pol->mode; |
948927ee9 mm, mempolicy: ma... |
2705 2706 |
flags = pol->flags; } |
bea904d54 mempolicy: use MP... |
2707 |
|
1a75a6c82 [PATCH] Fold numa... |
2708 2709 |
switch (mode) { case MPOL_DEFAULT: |
1a75a6c82 [PATCH] Fold numa... |
2710 |
break; |
1a75a6c82 [PATCH] Fold numa... |
2711 |
case MPOL_PREFERRED: |
fc36b8d3d mempolicy: use MP... |
2712 |
if (flags & MPOL_F_LOCAL) |
f2a07f40d tmpfs mempolicy: ... |
2713 |
mode = MPOL_LOCAL; |
53f2556b6 mempolicy: mPOL_P... |
2714 |
else |
fc36b8d3d mempolicy: use MP... |
2715 |
node_set(pol->v.preferred_node, nodes); |
1a75a6c82 [PATCH] Fold numa... |
2716 |
break; |
1a75a6c82 [PATCH] Fold numa... |
2717 |
case MPOL_BIND: |
1a75a6c82 [PATCH] Fold numa... |
2718 |
case MPOL_INTERLEAVE: |
f2a07f40d tmpfs mempolicy: ... |
2719 |
nodes = pol->v.nodes; |
1a75a6c82 [PATCH] Fold numa... |
2720 |
break; |
1a75a6c82 [PATCH] Fold numa... |
2721 |
default: |
948927ee9 mm, mempolicy: ma... |
2722 2723 2724 |
WARN_ON_ONCE(1); snprintf(p, maxlen, "unknown"); return; |
1a75a6c82 [PATCH] Fold numa... |
2725 |
} |
b7a9f420e mm, mempolicy: si... |
2726 |
p += snprintf(p, maxlen, "%s", policy_modes[mode]); |
1a75a6c82 [PATCH] Fold numa... |
2727 |
|
fc36b8d3d mempolicy: use MP... |
2728 |
if (flags & MPOL_MODE_FLAGS) { |
948927ee9 mm, mempolicy: ma... |
2729 |
p += snprintf(p, buffer + maxlen - p, "="); |
f5b087b52 mempolicy: add MP... |
2730 |
|
2291990ab mempolicy: clean-... |
2731 2732 2733 |
/* * Currently, the only defined flags are mutually exclusive */ |
f5b087b52 mempolicy: add MP... |
2734 |
if (flags & MPOL_F_STATIC_NODES) |
2291990ab mempolicy: clean-... |
2735 2736 2737 |
p += snprintf(p, buffer + maxlen - p, "static"); else if (flags & MPOL_F_RELATIVE_NODES) p += snprintf(p, buffer + maxlen - p, "relative"); |
f5b087b52 mempolicy: add MP... |
2738 |
} |
9e763e0f4 mm: use %*pb[l] t... |
2739 2740 2741 |
if (!nodes_empty(nodes)) p += scnprintf(p, buffer + maxlen - p, ":%*pbl", nodemask_pr_args(&nodes)); |
1a75a6c82 [PATCH] Fold numa... |
2742 |
} |