]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /* |
2 | * Simple NUMA memory policy for the Linux kernel. | |
3 | * | |
4 | * Copyright 2003,2004 Andi Kleen, SuSE Labs. | |
5 | * Subject to the GNU Public License, version 2. | |
6 | * | |
7 | * NUMA policy allows the user to give hints in which node(s) memory should | |
8 | * be allocated. | |
9 | * | |
10 | * Support four policies per VMA and per process: | |
11 | * | |
12 | * The VMA policy has priority over the process policy for a page fault. | |
13 | * | |
14 | * interleave Allocate memory interleaved over a set of nodes, | |
15 | * with normal fallback if it fails. | |
16 | * For VMA based allocations this interleaves based on the | |
17 | * offset into the backing object or offset into the mapping | |
18 | * for anonymous memory. For process policy an process counter | |
19 | * is used. | |
20 | * bind Only allocate memory on a specific set of nodes, | |
21 | * no fallback. | |
22 | * preferred Try a specific node first before normal fallback. | |
23 | * As a special case node -1 here means do the allocation | |
24 | * on the local CPU. This is normally identical to default, | |
25 | * but useful to set in a VMA when you have a non default | |
26 | * process policy. | |
27 | * default Allocate on the local node first, or when on a VMA | |
28 | * use the process policy. This is what Linux always did | |
29 | * in a NUMA aware kernel and still does by, ahem, default. | |
30 | * | |
31 | * The process policy is applied for most non interrupt memory allocations | |
32 | * in that process' context. Interrupts ignore the policies and always | |
33 | * try to allocate on the local CPU. The VMA policy is only applied for memory | |
34 | * allocations for a VMA in the VM. | |
35 | * | |
36 | * Currently there are a few corner cases in swapping where the policy | |
37 | * is not applied, but the majority should be handled. When process policy | |
38 | * is used it is not remembered over swap outs/swap ins. | |
39 | * | |
40 | * Only the highest zone in the zone hierarchy gets policied. Allocations | |
41 | * requesting a lower zone just use default policy. This implies that | |
42 | * on systems with highmem kernel lowmem allocation don't get policied. | |
43 | * Same with GFP_DMA allocations. | |
44 | * | |
45 | * For shmfs/tmpfs/hugetlbfs shared memory the policy is shared between | |
46 | * all users and remembered even when nobody has memory mapped. | |
47 | */ | |
48 | ||
49 | /* Notebook: | |
50 | fix mmap readahead to honour policy and enable policy for any page cache | |
51 | object | |
52 | statistics for bigpages | |
53 | global policy for page cache? currently it uses process policy. Requires | |
54 | first item above. | |
55 | handle mremap for shared memory (currently ignored for the policy) | |
56 | grows down? | |
57 | make bind policy root only? It can trigger oom much faster and the | |
58 | kernel is not always grateful with that. | |
59 | could replace all the switch()es with a mempolicy_ops structure. | |
60 | */ | |
61 | ||
62 | #include <linux/mempolicy.h> | |
63 | #include <linux/mm.h> | |
64 | #include <linux/highmem.h> | |
65 | #include <linux/hugetlb.h> | |
66 | #include <linux/kernel.h> | |
67 | #include <linux/sched.h> | |
68 | #include <linux/mm.h> | |
69 | #include <linux/nodemask.h> | |
70 | #include <linux/cpuset.h> | |
71 | #include <linux/gfp.h> | |
72 | #include <linux/slab.h> | |
73 | #include <linux/string.h> | |
74 | #include <linux/module.h> | |
75 | #include <linux/interrupt.h> | |
76 | #include <linux/init.h> | |
77 | #include <linux/compat.h> | |
78 | #include <linux/mempolicy.h> | |
79 | #include <asm/tlbflush.h> | |
80 | #include <asm/uaccess.h> | |
81 | ||
82 | static kmem_cache_t *policy_cache; | |
83 | static kmem_cache_t *sn_cache; | |
84 | ||
85 | #define PDprintk(fmt...) | |
86 | ||
87 | /* Highest zone. An specific allocation for a zone below that is not | |
88 | policied. */ | |
89 | static int policy_zone; | |
90 | ||
91 | static struct mempolicy default_policy = { | |
92 | .refcnt = ATOMIC_INIT(1), /* never free it */ | |
93 | .policy = MPOL_DEFAULT, | |
94 | }; | |
95 | ||
96 | /* Check if all specified nodes are online */ | |
97 | static int nodes_online(unsigned long *nodes) | |
98 | { | |
99 | DECLARE_BITMAP(online2, MAX_NUMNODES); | |
100 | ||
101 | bitmap_copy(online2, nodes_addr(node_online_map), MAX_NUMNODES); | |
102 | if (bitmap_empty(online2, MAX_NUMNODES)) | |
103 | set_bit(0, online2); | |
104 | if (!bitmap_subset(nodes, online2, MAX_NUMNODES)) | |
105 | return -EINVAL; | |
106 | return 0; | |
107 | } | |
108 | ||
109 | /* Do sanity checking on a policy */ | |
110 | static int mpol_check_policy(int mode, unsigned long *nodes) | |
111 | { | |
112 | int empty = bitmap_empty(nodes, MAX_NUMNODES); | |
113 | ||
114 | switch (mode) { | |
115 | case MPOL_DEFAULT: | |
116 | if (!empty) | |
117 | return -EINVAL; | |
118 | break; | |
119 | case MPOL_BIND: | |
120 | case MPOL_INTERLEAVE: | |
121 | /* Preferred will only use the first bit, but allow | |
122 | more for now. */ | |
123 | if (empty) | |
124 | return -EINVAL; | |
125 | break; | |
126 | } | |
127 | return nodes_online(nodes); | |
128 | } | |
129 | ||
130 | /* Copy a node mask from user space. */ | |
131 | static int get_nodes(unsigned long *nodes, unsigned long __user *nmask, | |
132 | unsigned long maxnode, int mode) | |
133 | { | |
134 | unsigned long k; | |
135 | unsigned long nlongs; | |
136 | unsigned long endmask; | |
137 | ||
138 | --maxnode; | |
139 | bitmap_zero(nodes, MAX_NUMNODES); | |
140 | if (maxnode == 0 || !nmask) | |
141 | return 0; | |
142 | ||
143 | nlongs = BITS_TO_LONGS(maxnode); | |
144 | if ((maxnode % BITS_PER_LONG) == 0) | |
145 | endmask = ~0UL; | |
146 | else | |
147 | endmask = (1UL << (maxnode % BITS_PER_LONG)) - 1; | |
148 | ||
149 | /* When the user specified more nodes than supported just check | |
150 | if the non supported part is all zero. */ | |
151 | if (nlongs > BITS_TO_LONGS(MAX_NUMNODES)) { | |
152 | if (nlongs > PAGE_SIZE/sizeof(long)) | |
153 | return -EINVAL; | |
154 | for (k = BITS_TO_LONGS(MAX_NUMNODES); k < nlongs; k++) { | |
155 | unsigned long t; | |
156 | if (get_user(t, nmask + k)) | |
157 | return -EFAULT; | |
158 | if (k == nlongs - 1) { | |
159 | if (t & endmask) | |
160 | return -EINVAL; | |
161 | } else if (t) | |
162 | return -EINVAL; | |
163 | } | |
164 | nlongs = BITS_TO_LONGS(MAX_NUMNODES); | |
165 | endmask = ~0UL; | |
166 | } | |
167 | ||
168 | if (copy_from_user(nodes, nmask, nlongs*sizeof(unsigned long))) | |
169 | return -EFAULT; | |
170 | nodes[nlongs-1] &= endmask; | |
171 | /* Update current mems_allowed */ | |
172 | cpuset_update_current_mems_allowed(); | |
173 | /* Ignore nodes not set in current->mems_allowed */ | |
174 | cpuset_restrict_to_mems_allowed(nodes); | |
175 | return mpol_check_policy(mode, nodes); | |
176 | } | |
177 | ||
178 | /* Generate a custom zonelist for the BIND policy. */ | |
179 | static struct zonelist *bind_zonelist(unsigned long *nodes) | |
180 | { | |
181 | struct zonelist *zl; | |
182 | int num, max, nd; | |
183 | ||
184 | max = 1 + MAX_NR_ZONES * bitmap_weight(nodes, MAX_NUMNODES); | |
185 | zl = kmalloc(sizeof(void *) * max, GFP_KERNEL); | |
186 | if (!zl) | |
187 | return NULL; | |
188 | num = 0; | |
189 | for (nd = find_first_bit(nodes, MAX_NUMNODES); | |
190 | nd < MAX_NUMNODES; | |
191 | nd = find_next_bit(nodes, MAX_NUMNODES, 1+nd)) { | |
192 | int k; | |
193 | for (k = MAX_NR_ZONES-1; k >= 0; k--) { | |
194 | struct zone *z = &NODE_DATA(nd)->node_zones[k]; | |
195 | if (!z->present_pages) | |
196 | continue; | |
197 | zl->zones[num++] = z; | |
198 | if (k > policy_zone) | |
199 | policy_zone = k; | |
200 | } | |
201 | } | |
202 | BUG_ON(num >= max); | |
203 | zl->zones[num] = NULL; | |
204 | return zl; | |
205 | } | |
206 | ||
207 | /* Create a new policy */ | |
208 | static struct mempolicy *mpol_new(int mode, unsigned long *nodes) | |
209 | { | |
210 | struct mempolicy *policy; | |
211 | ||
212 | PDprintk("setting mode %d nodes[0] %lx\n", mode, nodes[0]); | |
213 | if (mode == MPOL_DEFAULT) | |
214 | return NULL; | |
215 | policy = kmem_cache_alloc(policy_cache, GFP_KERNEL); | |
216 | if (!policy) | |
217 | return ERR_PTR(-ENOMEM); | |
218 | atomic_set(&policy->refcnt, 1); | |
219 | switch (mode) { | |
220 | case MPOL_INTERLEAVE: | |
221 | bitmap_copy(policy->v.nodes, nodes, MAX_NUMNODES); | |
222 | break; | |
223 | case MPOL_PREFERRED: | |
224 | policy->v.preferred_node = find_first_bit(nodes, MAX_NUMNODES); | |
225 | if (policy->v.preferred_node >= MAX_NUMNODES) | |
226 | policy->v.preferred_node = -1; | |
227 | break; | |
228 | case MPOL_BIND: | |
229 | policy->v.zonelist = bind_zonelist(nodes); | |
230 | if (policy->v.zonelist == NULL) { | |
231 | kmem_cache_free(policy_cache, policy); | |
232 | return ERR_PTR(-ENOMEM); | |
233 | } | |
234 | break; | |
235 | } | |
236 | policy->policy = mode; | |
237 | return policy; | |
238 | } | |
239 | ||
240 | /* Ensure all existing pages follow the policy. */ | |
241 | static int | |
242 | verify_pages(struct mm_struct *mm, | |
243 | unsigned long addr, unsigned long end, unsigned long *nodes) | |
244 | { | |
941150a3 HD |
245 | int err = 0; |
246 | ||
247 | spin_lock(&mm->page_table_lock); | |
1da177e4 LT |
248 | while (addr < end) { |
249 | struct page *p; | |
250 | pte_t *pte; | |
251 | pmd_t *pmd; | |
252 | pud_t *pud; | |
253 | pgd_t *pgd; | |
254 | pgd = pgd_offset(mm, addr); | |
255 | if (pgd_none(*pgd)) { | |
256 | unsigned long next = (addr + PGDIR_SIZE) & PGDIR_MASK; | |
257 | if (next > addr) | |
258 | break; | |
259 | addr = next; | |
260 | continue; | |
261 | } | |
262 | pud = pud_offset(pgd, addr); | |
263 | if (pud_none(*pud)) { | |
264 | addr = (addr + PUD_SIZE) & PUD_MASK; | |
265 | continue; | |
266 | } | |
267 | pmd = pmd_offset(pud, addr); | |
268 | if (pmd_none(*pmd)) { | |
269 | addr = (addr + PMD_SIZE) & PMD_MASK; | |
270 | continue; | |
271 | } | |
272 | p = NULL; | |
273 | pte = pte_offset_map(pmd, addr); | |
941150a3 HD |
274 | if (pte_present(*pte)) { |
275 | unsigned long pfn = pte_pfn(*pte); | |
276 | if (pfn_valid(pfn)) | |
277 | p = pfn_to_page(pfn); | |
278 | } | |
1da177e4 LT |
279 | pte_unmap(pte); |
280 | if (p) { | |
281 | unsigned nid = page_to_nid(p); | |
941150a3 HD |
282 | if (!test_bit(nid, nodes)) { |
283 | err = -EIO; | |
284 | break; | |
285 | } | |
1da177e4 LT |
286 | } |
287 | addr += PAGE_SIZE; | |
288 | } | |
941150a3 HD |
289 | spin_unlock(&mm->page_table_lock); |
290 | return err; | |
1da177e4 LT |
291 | } |
292 | ||
293 | /* Step 1: check the range */ | |
294 | static struct vm_area_struct * | |
295 | check_range(struct mm_struct *mm, unsigned long start, unsigned long end, | |
296 | unsigned long *nodes, unsigned long flags) | |
297 | { | |
298 | int err; | |
299 | struct vm_area_struct *first, *vma, *prev; | |
300 | ||
301 | first = find_vma(mm, start); | |
302 | if (!first) | |
303 | return ERR_PTR(-EFAULT); | |
304 | prev = NULL; | |
305 | for (vma = first; vma && vma->vm_start < end; vma = vma->vm_next) { | |
306 | if (!vma->vm_next && vma->vm_end < end) | |
307 | return ERR_PTR(-EFAULT); | |
308 | if (prev && prev->vm_end < vma->vm_start) | |
309 | return ERR_PTR(-EFAULT); | |
310 | if ((flags & MPOL_MF_STRICT) && !is_vm_hugetlb_page(vma)) { | |
311 | err = verify_pages(vma->vm_mm, | |
312 | vma->vm_start, vma->vm_end, nodes); | |
313 | if (err) { | |
314 | first = ERR_PTR(err); | |
315 | break; | |
316 | } | |
317 | } | |
318 | prev = vma; | |
319 | } | |
320 | return first; | |
321 | } | |
322 | ||
323 | /* Apply policy to a single VMA */ | |
324 | static int policy_vma(struct vm_area_struct *vma, struct mempolicy *new) | |
325 | { | |
326 | int err = 0; | |
327 | struct mempolicy *old = vma->vm_policy; | |
328 | ||
329 | PDprintk("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n", | |
330 | vma->vm_start, vma->vm_end, vma->vm_pgoff, | |
331 | vma->vm_ops, vma->vm_file, | |
332 | vma->vm_ops ? vma->vm_ops->set_policy : NULL); | |
333 | ||
334 | if (vma->vm_ops && vma->vm_ops->set_policy) | |
335 | err = vma->vm_ops->set_policy(vma, new); | |
336 | if (!err) { | |
337 | mpol_get(new); | |
338 | vma->vm_policy = new; | |
339 | mpol_free(old); | |
340 | } | |
341 | return err; | |
342 | } | |
343 | ||
344 | /* Step 2: apply policy to a range and do splits. */ | |
345 | static int mbind_range(struct vm_area_struct *vma, unsigned long start, | |
346 | unsigned long end, struct mempolicy *new) | |
347 | { | |
348 | struct vm_area_struct *next; | |
349 | int err; | |
350 | ||
351 | err = 0; | |
352 | for (; vma && vma->vm_start < end; vma = next) { | |
353 | next = vma->vm_next; | |
354 | if (vma->vm_start < start) | |
355 | err = split_vma(vma->vm_mm, vma, start, 1); | |
356 | if (!err && vma->vm_end > end) | |
357 | err = split_vma(vma->vm_mm, vma, end, 0); | |
358 | if (!err) | |
359 | err = policy_vma(vma, new); | |
360 | if (err) | |
361 | break; | |
362 | } | |
363 | return err; | |
364 | } | |
365 | ||
366 | /* Change policy for a memory range */ | |
367 | asmlinkage long sys_mbind(unsigned long start, unsigned long len, | |
368 | unsigned long mode, | |
369 | unsigned long __user *nmask, unsigned long maxnode, | |
370 | unsigned flags) | |
371 | { | |
372 | struct vm_area_struct *vma; | |
373 | struct mm_struct *mm = current->mm; | |
374 | struct mempolicy *new; | |
375 | unsigned long end; | |
376 | DECLARE_BITMAP(nodes, MAX_NUMNODES); | |
377 | int err; | |
378 | ||
379 | if ((flags & ~(unsigned long)(MPOL_MF_STRICT)) || mode > MPOL_MAX) | |
380 | return -EINVAL; | |
381 | if (start & ~PAGE_MASK) | |
382 | return -EINVAL; | |
383 | if (mode == MPOL_DEFAULT) | |
384 | flags &= ~MPOL_MF_STRICT; | |
385 | len = (len + PAGE_SIZE - 1) & PAGE_MASK; | |
386 | end = start + len; | |
387 | if (end < start) | |
388 | return -EINVAL; | |
389 | if (end == start) | |
390 | return 0; | |
391 | ||
392 | err = get_nodes(nodes, nmask, maxnode, mode); | |
393 | if (err) | |
394 | return err; | |
395 | ||
396 | new = mpol_new(mode, nodes); | |
397 | if (IS_ERR(new)) | |
398 | return PTR_ERR(new); | |
399 | ||
400 | PDprintk("mbind %lx-%lx mode:%ld nodes:%lx\n",start,start+len, | |
401 | mode,nodes[0]); | |
402 | ||
403 | down_write(&mm->mmap_sem); | |
404 | vma = check_range(mm, start, end, nodes, flags); | |
405 | err = PTR_ERR(vma); | |
406 | if (!IS_ERR(vma)) | |
407 | err = mbind_range(vma, start, end, new); | |
408 | up_write(&mm->mmap_sem); | |
409 | mpol_free(new); | |
410 | return err; | |
411 | } | |
412 | ||
413 | /* Set the process memory policy */ | |
414 | asmlinkage long sys_set_mempolicy(int mode, unsigned long __user *nmask, | |
415 | unsigned long maxnode) | |
416 | { | |
417 | int err; | |
418 | struct mempolicy *new; | |
419 | DECLARE_BITMAP(nodes, MAX_NUMNODES); | |
420 | ||
421 | if (mode > MPOL_MAX) | |
422 | return -EINVAL; | |
423 | err = get_nodes(nodes, nmask, maxnode, mode); | |
424 | if (err) | |
425 | return err; | |
426 | new = mpol_new(mode, nodes); | |
427 | if (IS_ERR(new)) | |
428 | return PTR_ERR(new); | |
429 | mpol_free(current->mempolicy); | |
430 | current->mempolicy = new; | |
431 | if (new && new->policy == MPOL_INTERLEAVE) | |
432 | current->il_next = find_first_bit(new->v.nodes, MAX_NUMNODES); | |
433 | return 0; | |
434 | } | |
435 | ||
436 | /* Fill a zone bitmap for a policy */ | |
437 | static void get_zonemask(struct mempolicy *p, unsigned long *nodes) | |
438 | { | |
439 | int i; | |
440 | ||
441 | bitmap_zero(nodes, MAX_NUMNODES); | |
442 | switch (p->policy) { | |
443 | case MPOL_BIND: | |
444 | for (i = 0; p->v.zonelist->zones[i]; i++) | |
445 | __set_bit(p->v.zonelist->zones[i]->zone_pgdat->node_id, nodes); | |
446 | break; | |
447 | case MPOL_DEFAULT: | |
448 | break; | |
449 | case MPOL_INTERLEAVE: | |
450 | bitmap_copy(nodes, p->v.nodes, MAX_NUMNODES); | |
451 | break; | |
452 | case MPOL_PREFERRED: | |
453 | /* or use current node instead of online map? */ | |
454 | if (p->v.preferred_node < 0) | |
455 | bitmap_copy(nodes, nodes_addr(node_online_map), MAX_NUMNODES); | |
456 | else | |
457 | __set_bit(p->v.preferred_node, nodes); | |
458 | break; | |
459 | default: | |
460 | BUG(); | |
461 | } | |
462 | } | |
463 | ||
464 | static int lookup_node(struct mm_struct *mm, unsigned long addr) | |
465 | { | |
466 | struct page *p; | |
467 | int err; | |
468 | ||
469 | err = get_user_pages(current, mm, addr & PAGE_MASK, 1, 0, 0, &p, NULL); | |
470 | if (err >= 0) { | |
471 | err = page_to_nid(p); | |
472 | put_page(p); | |
473 | } | |
474 | return err; | |
475 | } | |
476 | ||
477 | /* Copy a kernel node mask to user space */ | |
478 | static int copy_nodes_to_user(unsigned long __user *mask, unsigned long maxnode, | |
479 | void *nodes, unsigned nbytes) | |
480 | { | |
481 | unsigned long copy = ALIGN(maxnode-1, 64) / 8; | |
482 | ||
483 | if (copy > nbytes) { | |
484 | if (copy > PAGE_SIZE) | |
485 | return -EINVAL; | |
486 | if (clear_user((char __user *)mask + nbytes, copy - nbytes)) | |
487 | return -EFAULT; | |
488 | copy = nbytes; | |
489 | } | |
490 | return copy_to_user(mask, nodes, copy) ? -EFAULT : 0; | |
491 | } | |
492 | ||
493 | /* Retrieve NUMA policy */ | |
494 | asmlinkage long sys_get_mempolicy(int __user *policy, | |
495 | unsigned long __user *nmask, | |
496 | unsigned long maxnode, | |
497 | unsigned long addr, unsigned long flags) | |
498 | { | |
499 | int err, pval; | |
500 | struct mm_struct *mm = current->mm; | |
501 | struct vm_area_struct *vma = NULL; | |
502 | struct mempolicy *pol = current->mempolicy; | |
503 | ||
504 | if (flags & ~(unsigned long)(MPOL_F_NODE|MPOL_F_ADDR)) | |
505 | return -EINVAL; | |
506 | if (nmask != NULL && maxnode < MAX_NUMNODES) | |
507 | return -EINVAL; | |
508 | if (flags & MPOL_F_ADDR) { | |
509 | down_read(&mm->mmap_sem); | |
510 | vma = find_vma_intersection(mm, addr, addr+1); | |
511 | if (!vma) { | |
512 | up_read(&mm->mmap_sem); | |
513 | return -EFAULT; | |
514 | } | |
515 | if (vma->vm_ops && vma->vm_ops->get_policy) | |
516 | pol = vma->vm_ops->get_policy(vma, addr); | |
517 | else | |
518 | pol = vma->vm_policy; | |
519 | } else if (addr) | |
520 | return -EINVAL; | |
521 | ||
522 | if (!pol) | |
523 | pol = &default_policy; | |
524 | ||
525 | if (flags & MPOL_F_NODE) { | |
526 | if (flags & MPOL_F_ADDR) { | |
527 | err = lookup_node(mm, addr); | |
528 | if (err < 0) | |
529 | goto out; | |
530 | pval = err; | |
531 | } else if (pol == current->mempolicy && | |
532 | pol->policy == MPOL_INTERLEAVE) { | |
533 | pval = current->il_next; | |
534 | } else { | |
535 | err = -EINVAL; | |
536 | goto out; | |
537 | } | |
538 | } else | |
539 | pval = pol->policy; | |
540 | ||
541 | if (vma) { | |
542 | up_read(¤t->mm->mmap_sem); | |
543 | vma = NULL; | |
544 | } | |
545 | ||
546 | if (policy && put_user(pval, policy)) | |
547 | return -EFAULT; | |
548 | ||
549 | err = 0; | |
550 | if (nmask) { | |
551 | DECLARE_BITMAP(nodes, MAX_NUMNODES); | |
552 | get_zonemask(pol, nodes); | |
553 | err = copy_nodes_to_user(nmask, maxnode, nodes, sizeof(nodes)); | |
554 | } | |
555 | ||
556 | out: | |
557 | if (vma) | |
558 | up_read(¤t->mm->mmap_sem); | |
559 | return err; | |
560 | } | |
561 | ||
562 | #ifdef CONFIG_COMPAT | |
563 | ||
564 | asmlinkage long compat_sys_get_mempolicy(int __user *policy, | |
565 | compat_ulong_t __user *nmask, | |
566 | compat_ulong_t maxnode, | |
567 | compat_ulong_t addr, compat_ulong_t flags) | |
568 | { | |
569 | long err; | |
570 | unsigned long __user *nm = NULL; | |
571 | unsigned long nr_bits, alloc_size; | |
572 | DECLARE_BITMAP(bm, MAX_NUMNODES); | |
573 | ||
574 | nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); | |
575 | alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; | |
576 | ||
577 | if (nmask) | |
578 | nm = compat_alloc_user_space(alloc_size); | |
579 | ||
580 | err = sys_get_mempolicy(policy, nm, nr_bits+1, addr, flags); | |
581 | ||
582 | if (!err && nmask) { | |
583 | err = copy_from_user(bm, nm, alloc_size); | |
584 | /* ensure entire bitmap is zeroed */ | |
585 | err |= clear_user(nmask, ALIGN(maxnode-1, 8) / 8); | |
586 | err |= compat_put_bitmap(nmask, bm, nr_bits); | |
587 | } | |
588 | ||
589 | return err; | |
590 | } | |
591 | ||
592 | asmlinkage long compat_sys_set_mempolicy(int mode, compat_ulong_t __user *nmask, | |
593 | compat_ulong_t maxnode) | |
594 | { | |
595 | long err = 0; | |
596 | unsigned long __user *nm = NULL; | |
597 | unsigned long nr_bits, alloc_size; | |
598 | DECLARE_BITMAP(bm, MAX_NUMNODES); | |
599 | ||
600 | nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); | |
601 | alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; | |
602 | ||
603 | if (nmask) { | |
604 | err = compat_get_bitmap(bm, nmask, nr_bits); | |
605 | nm = compat_alloc_user_space(alloc_size); | |
606 | err |= copy_to_user(nm, bm, alloc_size); | |
607 | } | |
608 | ||
609 | if (err) | |
610 | return -EFAULT; | |
611 | ||
612 | return sys_set_mempolicy(mode, nm, nr_bits+1); | |
613 | } | |
614 | ||
615 | asmlinkage long compat_sys_mbind(compat_ulong_t start, compat_ulong_t len, | |
616 | compat_ulong_t mode, compat_ulong_t __user *nmask, | |
617 | compat_ulong_t maxnode, compat_ulong_t flags) | |
618 | { | |
619 | long err = 0; | |
620 | unsigned long __user *nm = NULL; | |
621 | unsigned long nr_bits, alloc_size; | |
622 | DECLARE_BITMAP(bm, MAX_NUMNODES); | |
623 | ||
624 | nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES); | |
625 | alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8; | |
626 | ||
627 | if (nmask) { | |
628 | err = compat_get_bitmap(bm, nmask, nr_bits); | |
629 | nm = compat_alloc_user_space(alloc_size); | |
630 | err |= copy_to_user(nm, bm, alloc_size); | |
631 | } | |
632 | ||
633 | if (err) | |
634 | return -EFAULT; | |
635 | ||
636 | return sys_mbind(start, len, mode, nm, nr_bits+1, flags); | |
637 | } | |
638 | ||
639 | #endif | |
640 | ||
641 | /* Return effective policy for a VMA */ | |
642 | static struct mempolicy * | |
643 | get_vma_policy(struct vm_area_struct *vma, unsigned long addr) | |
644 | { | |
645 | struct mempolicy *pol = current->mempolicy; | |
646 | ||
647 | if (vma) { | |
648 | if (vma->vm_ops && vma->vm_ops->get_policy) | |
649 | pol = vma->vm_ops->get_policy(vma, addr); | |
650 | else if (vma->vm_policy && | |
651 | vma->vm_policy->policy != MPOL_DEFAULT) | |
652 | pol = vma->vm_policy; | |
653 | } | |
654 | if (!pol) | |
655 | pol = &default_policy; | |
656 | return pol; | |
657 | } | |
658 | ||
659 | /* Return a zonelist representing a mempolicy */ | |
660 | static struct zonelist *zonelist_policy(unsigned int __nocast gfp, struct mempolicy *policy) | |
661 | { | |
662 | int nd; | |
663 | ||
664 | switch (policy->policy) { | |
665 | case MPOL_PREFERRED: | |
666 | nd = policy->v.preferred_node; | |
667 | if (nd < 0) | |
668 | nd = numa_node_id(); | |
669 | break; | |
670 | case MPOL_BIND: | |
671 | /* Lower zones don't get a policy applied */ | |
672 | /* Careful: current->mems_allowed might have moved */ | |
01424961 | 673 | if ((gfp & GFP_ZONEMASK) >= policy_zone) |
1da177e4 LT |
674 | if (cpuset_zonelist_valid_mems_allowed(policy->v.zonelist)) |
675 | return policy->v.zonelist; | |
676 | /*FALL THROUGH*/ | |
677 | case MPOL_INTERLEAVE: /* should not happen */ | |
678 | case MPOL_DEFAULT: | |
679 | nd = numa_node_id(); | |
680 | break; | |
681 | default: | |
682 | nd = 0; | |
683 | BUG(); | |
684 | } | |
685 | return NODE_DATA(nd)->node_zonelists + (gfp & GFP_ZONEMASK); | |
686 | } | |
687 | ||
688 | /* Do dynamic interleaving for a process */ | |
689 | static unsigned interleave_nodes(struct mempolicy *policy) | |
690 | { | |
691 | unsigned nid, next; | |
692 | struct task_struct *me = current; | |
693 | ||
694 | nid = me->il_next; | |
695 | BUG_ON(nid >= MAX_NUMNODES); | |
696 | next = find_next_bit(policy->v.nodes, MAX_NUMNODES, 1+nid); | |
697 | if (next >= MAX_NUMNODES) | |
698 | next = find_first_bit(policy->v.nodes, MAX_NUMNODES); | |
699 | me->il_next = next; | |
700 | return nid; | |
701 | } | |
702 | ||
703 | /* Do static interleaving for a VMA with known offset. */ | |
704 | static unsigned offset_il_node(struct mempolicy *pol, | |
705 | struct vm_area_struct *vma, unsigned long off) | |
706 | { | |
707 | unsigned nnodes = bitmap_weight(pol->v.nodes, MAX_NUMNODES); | |
708 | unsigned target = (unsigned)off % nnodes; | |
709 | int c; | |
710 | int nid = -1; | |
711 | ||
712 | c = 0; | |
713 | do { | |
714 | nid = find_next_bit(pol->v.nodes, MAX_NUMNODES, nid+1); | |
715 | c++; | |
716 | } while (c <= target); | |
717 | BUG_ON(nid >= MAX_NUMNODES); | |
718 | BUG_ON(!test_bit(nid, pol->v.nodes)); | |
719 | return nid; | |
720 | } | |
721 | ||
722 | /* Allocate a page in interleaved policy. | |
723 | Own path because it needs to do special accounting. */ | |
724 | static struct page *alloc_page_interleave(unsigned int __nocast gfp, unsigned order, unsigned nid) | |
725 | { | |
726 | struct zonelist *zl; | |
727 | struct page *page; | |
728 | ||
729 | BUG_ON(!node_online(nid)); | |
730 | zl = NODE_DATA(nid)->node_zonelists + (gfp & GFP_ZONEMASK); | |
731 | page = __alloc_pages(gfp, order, zl); | |
732 | if (page && page_zone(page) == zl->zones[0]) { | |
e7c8d5c9 | 733 | zone_pcp(zl->zones[0],get_cpu())->interleave_hit++; |
1da177e4 LT |
734 | put_cpu(); |
735 | } | |
736 | return page; | |
737 | } | |
738 | ||
739 | /** | |
740 | * alloc_page_vma - Allocate a page for a VMA. | |
741 | * | |
742 | * @gfp: | |
743 | * %GFP_USER user allocation. | |
744 | * %GFP_KERNEL kernel allocations, | |
745 | * %GFP_HIGHMEM highmem/user allocations, | |
746 | * %GFP_FS allocation should not call back into a file system. | |
747 | * %GFP_ATOMIC don't sleep. | |
748 | * | |
749 | * @vma: Pointer to VMA or NULL if not available. | |
750 | * @addr: Virtual Address of the allocation. Must be inside the VMA. | |
751 | * | |
752 | * This function allocates a page from the kernel page pool and applies | |
753 | * a NUMA policy associated with the VMA or the current process. | |
754 | * When VMA is not NULL caller must hold down_read on the mmap_sem of the | |
755 | * mm_struct of the VMA to prevent it from going away. Should be used for | |
756 | * all allocations for pages that will be mapped into | |
757 | * user space. Returns NULL when no page can be allocated. | |
758 | * | |
759 | * Should be called with the mm_sem of the vma hold. | |
760 | */ | |
761 | struct page * | |
762 | alloc_page_vma(unsigned int __nocast gfp, struct vm_area_struct *vma, unsigned long addr) | |
763 | { | |
764 | struct mempolicy *pol = get_vma_policy(vma, addr); | |
765 | ||
766 | cpuset_update_current_mems_allowed(); | |
767 | ||
768 | if (unlikely(pol->policy == MPOL_INTERLEAVE)) { | |
769 | unsigned nid; | |
770 | if (vma) { | |
771 | unsigned long off; | |
772 | BUG_ON(addr >= vma->vm_end); | |
773 | BUG_ON(addr < vma->vm_start); | |
774 | off = vma->vm_pgoff; | |
775 | off += (addr - vma->vm_start) >> PAGE_SHIFT; | |
776 | nid = offset_il_node(pol, vma, off); | |
777 | } else { | |
778 | /* fall back to process interleaving */ | |
779 | nid = interleave_nodes(pol); | |
780 | } | |
781 | return alloc_page_interleave(gfp, 0, nid); | |
782 | } | |
783 | return __alloc_pages(gfp, 0, zonelist_policy(gfp, pol)); | |
784 | } | |
785 | ||
786 | /** | |
787 | * alloc_pages_current - Allocate pages. | |
788 | * | |
789 | * @gfp: | |
790 | * %GFP_USER user allocation, | |
791 | * %GFP_KERNEL kernel allocation, | |
792 | * %GFP_HIGHMEM highmem allocation, | |
793 | * %GFP_FS don't call back into a file system. | |
794 | * %GFP_ATOMIC don't sleep. | |
795 | * @order: Power of two of allocation size in pages. 0 is a single page. | |
796 | * | |
797 | * Allocate a page from the kernel page pool. When not in | |
798 | * interrupt context and apply the current process NUMA policy. | |
799 | * Returns NULL when no page can be allocated. | |
800 | * | |
801 | * Don't call cpuset_update_current_mems_allowed() unless | |
802 | * 1) it's ok to take cpuset_sem (can WAIT), and | |
803 | * 2) allocating for current task (not interrupt). | |
804 | */ | |
805 | struct page *alloc_pages_current(unsigned int __nocast gfp, unsigned order) | |
806 | { | |
807 | struct mempolicy *pol = current->mempolicy; | |
808 | ||
809 | if ((gfp & __GFP_WAIT) && !in_interrupt()) | |
810 | cpuset_update_current_mems_allowed(); | |
811 | if (!pol || in_interrupt()) | |
812 | pol = &default_policy; | |
813 | if (pol->policy == MPOL_INTERLEAVE) | |
814 | return alloc_page_interleave(gfp, order, interleave_nodes(pol)); | |
815 | return __alloc_pages(gfp, order, zonelist_policy(gfp, pol)); | |
816 | } | |
817 | EXPORT_SYMBOL(alloc_pages_current); | |
818 | ||
819 | /* Slow path of a mempolicy copy */ | |
820 | struct mempolicy *__mpol_copy(struct mempolicy *old) | |
821 | { | |
822 | struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL); | |
823 | ||
824 | if (!new) | |
825 | return ERR_PTR(-ENOMEM); | |
826 | *new = *old; | |
827 | atomic_set(&new->refcnt, 1); | |
828 | if (new->policy == MPOL_BIND) { | |
829 | int sz = ksize(old->v.zonelist); | |
830 | new->v.zonelist = kmalloc(sz, SLAB_KERNEL); | |
831 | if (!new->v.zonelist) { | |
832 | kmem_cache_free(policy_cache, new); | |
833 | return ERR_PTR(-ENOMEM); | |
834 | } | |
835 | memcpy(new->v.zonelist, old->v.zonelist, sz); | |
836 | } | |
837 | return new; | |
838 | } | |
839 | ||
840 | /* Slow path of a mempolicy comparison */ | |
841 | int __mpol_equal(struct mempolicy *a, struct mempolicy *b) | |
842 | { | |
843 | if (!a || !b) | |
844 | return 0; | |
845 | if (a->policy != b->policy) | |
846 | return 0; | |
847 | switch (a->policy) { | |
848 | case MPOL_DEFAULT: | |
849 | return 1; | |
850 | case MPOL_INTERLEAVE: | |
851 | return bitmap_equal(a->v.nodes, b->v.nodes, MAX_NUMNODES); | |
852 | case MPOL_PREFERRED: | |
853 | return a->v.preferred_node == b->v.preferred_node; | |
854 | case MPOL_BIND: { | |
855 | int i; | |
856 | for (i = 0; a->v.zonelist->zones[i]; i++) | |
857 | if (a->v.zonelist->zones[i] != b->v.zonelist->zones[i]) | |
858 | return 0; | |
859 | return b->v.zonelist->zones[i] == NULL; | |
860 | } | |
861 | default: | |
862 | BUG(); | |
863 | return 0; | |
864 | } | |
865 | } | |
866 | ||
867 | /* Slow path of a mpol destructor. */ | |
868 | void __mpol_free(struct mempolicy *p) | |
869 | { | |
870 | if (!atomic_dec_and_test(&p->refcnt)) | |
871 | return; | |
872 | if (p->policy == MPOL_BIND) | |
873 | kfree(p->v.zonelist); | |
874 | p->policy = MPOL_DEFAULT; | |
875 | kmem_cache_free(policy_cache, p); | |
876 | } | |
877 | ||
878 | /* | |
879 | * Hugetlb policy. Same as above, just works with node numbers instead of | |
880 | * zonelists. | |
881 | */ | |
882 | ||
883 | /* Find first node suitable for an allocation */ | |
884 | int mpol_first_node(struct vm_area_struct *vma, unsigned long addr) | |
885 | { | |
886 | struct mempolicy *pol = get_vma_policy(vma, addr); | |
887 | ||
888 | switch (pol->policy) { | |
889 | case MPOL_DEFAULT: | |
890 | return numa_node_id(); | |
891 | case MPOL_BIND: | |
892 | return pol->v.zonelist->zones[0]->zone_pgdat->node_id; | |
893 | case MPOL_INTERLEAVE: | |
894 | return interleave_nodes(pol); | |
895 | case MPOL_PREFERRED: | |
896 | return pol->v.preferred_node >= 0 ? | |
897 | pol->v.preferred_node : numa_node_id(); | |
898 | } | |
899 | BUG(); | |
900 | return 0; | |
901 | } | |
902 | ||
903 | /* Find secondary valid nodes for an allocation */ | |
904 | int mpol_node_valid(int nid, struct vm_area_struct *vma, unsigned long addr) | |
905 | { | |
906 | struct mempolicy *pol = get_vma_policy(vma, addr); | |
907 | ||
908 | switch (pol->policy) { | |
909 | case MPOL_PREFERRED: | |
910 | case MPOL_DEFAULT: | |
911 | case MPOL_INTERLEAVE: | |
912 | return 1; | |
913 | case MPOL_BIND: { | |
914 | struct zone **z; | |
915 | for (z = pol->v.zonelist->zones; *z; z++) | |
916 | if ((*z)->zone_pgdat->node_id == nid) | |
917 | return 1; | |
918 | return 0; | |
919 | } | |
920 | default: | |
921 | BUG(); | |
922 | return 0; | |
923 | } | |
924 | } | |
925 | ||
926 | /* | |
927 | * Shared memory backing store policy support. | |
928 | * | |
929 | * Remember policies even when nobody has shared memory mapped. | |
930 | * The policies are kept in Red-Black tree linked from the inode. | |
931 | * They are protected by the sp->lock spinlock, which should be held | |
932 | * for any accesses to the tree. | |
933 | */ | |
934 | ||
935 | /* lookup first element intersecting start-end */ | |
936 | /* Caller holds sp->lock */ | |
937 | static struct sp_node * | |
938 | sp_lookup(struct shared_policy *sp, unsigned long start, unsigned long end) | |
939 | { | |
940 | struct rb_node *n = sp->root.rb_node; | |
941 | ||
942 | while (n) { | |
943 | struct sp_node *p = rb_entry(n, struct sp_node, nd); | |
944 | ||
945 | if (start >= p->end) | |
946 | n = n->rb_right; | |
947 | else if (end <= p->start) | |
948 | n = n->rb_left; | |
949 | else | |
950 | break; | |
951 | } | |
952 | if (!n) | |
953 | return NULL; | |
954 | for (;;) { | |
955 | struct sp_node *w = NULL; | |
956 | struct rb_node *prev = rb_prev(n); | |
957 | if (!prev) | |
958 | break; | |
959 | w = rb_entry(prev, struct sp_node, nd); | |
960 | if (w->end <= start) | |
961 | break; | |
962 | n = prev; | |
963 | } | |
964 | return rb_entry(n, struct sp_node, nd); | |
965 | } | |
966 | ||
967 | /* Insert a new shared policy into the list. */ | |
968 | /* Caller holds sp->lock */ | |
969 | static void sp_insert(struct shared_policy *sp, struct sp_node *new) | |
970 | { | |
971 | struct rb_node **p = &sp->root.rb_node; | |
972 | struct rb_node *parent = NULL; | |
973 | struct sp_node *nd; | |
974 | ||
975 | while (*p) { | |
976 | parent = *p; | |
977 | nd = rb_entry(parent, struct sp_node, nd); | |
978 | if (new->start < nd->start) | |
979 | p = &(*p)->rb_left; | |
980 | else if (new->end > nd->end) | |
981 | p = &(*p)->rb_right; | |
982 | else | |
983 | BUG(); | |
984 | } | |
985 | rb_link_node(&new->nd, parent, p); | |
986 | rb_insert_color(&new->nd, &sp->root); | |
987 | PDprintk("inserting %lx-%lx: %d\n", new->start, new->end, | |
988 | new->policy ? new->policy->policy : 0); | |
989 | } | |
990 | ||
991 | /* Find shared policy intersecting idx */ | |
992 | struct mempolicy * | |
993 | mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx) | |
994 | { | |
995 | struct mempolicy *pol = NULL; | |
996 | struct sp_node *sn; | |
997 | ||
998 | if (!sp->root.rb_node) | |
999 | return NULL; | |
1000 | spin_lock(&sp->lock); | |
1001 | sn = sp_lookup(sp, idx, idx+1); | |
1002 | if (sn) { | |
1003 | mpol_get(sn->policy); | |
1004 | pol = sn->policy; | |
1005 | } | |
1006 | spin_unlock(&sp->lock); | |
1007 | return pol; | |
1008 | } | |
1009 | ||
1010 | static void sp_delete(struct shared_policy *sp, struct sp_node *n) | |
1011 | { | |
1012 | PDprintk("deleting %lx-l%x\n", n->start, n->end); | |
1013 | rb_erase(&n->nd, &sp->root); | |
1014 | mpol_free(n->policy); | |
1015 | kmem_cache_free(sn_cache, n); | |
1016 | } | |
1017 | ||
1018 | struct sp_node * | |
1019 | sp_alloc(unsigned long start, unsigned long end, struct mempolicy *pol) | |
1020 | { | |
1021 | struct sp_node *n = kmem_cache_alloc(sn_cache, GFP_KERNEL); | |
1022 | ||
1023 | if (!n) | |
1024 | return NULL; | |
1025 | n->start = start; | |
1026 | n->end = end; | |
1027 | mpol_get(pol); | |
1028 | n->policy = pol; | |
1029 | return n; | |
1030 | } | |
1031 | ||
1032 | /* Replace a policy range. */ | |
1033 | static int shared_policy_replace(struct shared_policy *sp, unsigned long start, | |
1034 | unsigned long end, struct sp_node *new) | |
1035 | { | |
1036 | struct sp_node *n, *new2 = NULL; | |
1037 | ||
1038 | restart: | |
1039 | spin_lock(&sp->lock); | |
1040 | n = sp_lookup(sp, start, end); | |
1041 | /* Take care of old policies in the same range. */ | |
1042 | while (n && n->start < end) { | |
1043 | struct rb_node *next = rb_next(&n->nd); | |
1044 | if (n->start >= start) { | |
1045 | if (n->end <= end) | |
1046 | sp_delete(sp, n); | |
1047 | else | |
1048 | n->start = end; | |
1049 | } else { | |
1050 | /* Old policy spanning whole new range. */ | |
1051 | if (n->end > end) { | |
1052 | if (!new2) { | |
1053 | spin_unlock(&sp->lock); | |
1054 | new2 = sp_alloc(end, n->end, n->policy); | |
1055 | if (!new2) | |
1056 | return -ENOMEM; | |
1057 | goto restart; | |
1058 | } | |
1059 | n->end = start; | |
1060 | sp_insert(sp, new2); | |
1061 | new2 = NULL; | |
1062 | break; | |
1063 | } else | |
1064 | n->end = start; | |
1065 | } | |
1066 | if (!next) | |
1067 | break; | |
1068 | n = rb_entry(next, struct sp_node, nd); | |
1069 | } | |
1070 | if (new) | |
1071 | sp_insert(sp, new); | |
1072 | spin_unlock(&sp->lock); | |
1073 | if (new2) { | |
1074 | mpol_free(new2->policy); | |
1075 | kmem_cache_free(sn_cache, new2); | |
1076 | } | |
1077 | return 0; | |
1078 | } | |
1079 | ||
1080 | int mpol_set_shared_policy(struct shared_policy *info, | |
1081 | struct vm_area_struct *vma, struct mempolicy *npol) | |
1082 | { | |
1083 | int err; | |
1084 | struct sp_node *new = NULL; | |
1085 | unsigned long sz = vma_pages(vma); | |
1086 | ||
1087 | PDprintk("set_shared_policy %lx sz %lu %d %lx\n", | |
1088 | vma->vm_pgoff, | |
1089 | sz, npol? npol->policy : -1, | |
1090 | npol ? npol->v.nodes[0] : -1); | |
1091 | ||
1092 | if (npol) { | |
1093 | new = sp_alloc(vma->vm_pgoff, vma->vm_pgoff + sz, npol); | |
1094 | if (!new) | |
1095 | return -ENOMEM; | |
1096 | } | |
1097 | err = shared_policy_replace(info, vma->vm_pgoff, vma->vm_pgoff+sz, new); | |
1098 | if (err && new) | |
1099 | kmem_cache_free(sn_cache, new); | |
1100 | return err; | |
1101 | } | |
1102 | ||
1103 | /* Free a backing policy store on inode delete. */ | |
1104 | void mpol_free_shared_policy(struct shared_policy *p) | |
1105 | { | |
1106 | struct sp_node *n; | |
1107 | struct rb_node *next; | |
1108 | ||
1109 | if (!p->root.rb_node) | |
1110 | return; | |
1111 | spin_lock(&p->lock); | |
1112 | next = rb_first(&p->root); | |
1113 | while (next) { | |
1114 | n = rb_entry(next, struct sp_node, nd); | |
1115 | next = rb_next(&n->nd); | |
1116 | mpol_free(n->policy); | |
1117 | kmem_cache_free(sn_cache, n); | |
1118 | } | |
1119 | spin_unlock(&p->lock); | |
1120 | p->root = RB_ROOT; | |
1121 | } | |
1122 | ||
1123 | /* assumes fs == KERNEL_DS */ | |
1124 | void __init numa_policy_init(void) | |
1125 | { | |
1126 | policy_cache = kmem_cache_create("numa_policy", | |
1127 | sizeof(struct mempolicy), | |
1128 | 0, SLAB_PANIC, NULL, NULL); | |
1129 | ||
1130 | sn_cache = kmem_cache_create("shared_policy_node", | |
1131 | sizeof(struct sp_node), | |
1132 | 0, SLAB_PANIC, NULL, NULL); | |
1133 | ||
1134 | /* Set interleaving policy for system init. This way not all | |
1135 | the data structures allocated at system boot end up in node zero. */ | |
1136 | ||
1137 | if (sys_set_mempolicy(MPOL_INTERLEAVE, nodes_addr(node_online_map), | |
1138 | MAX_NUMNODES) < 0) | |
1139 | printk("numa_policy_init: interleaving failed\n"); | |
1140 | } | |
1141 | ||
1142 | /* Reset policy of current process to default. | |
1143 | * Assumes fs == KERNEL_DS */ | |
1144 | void numa_default_policy(void) | |
1145 | { | |
1146 | sys_set_mempolicy(MPOL_DEFAULT, NULL, 0); | |
1147 | } |