Blame view
mm/vmalloc.c
69.7 KB
1da177e4c
|
1 2 3 4 5 6 7 |
/* * linux/mm/vmalloc.c * * Copyright (C) 1993 Linus Torvalds * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999 * SMP-safe vmalloc/vfree/ioremap, Tigran Aivazian <tigran@veritas.com>, May 2000 * Major rework to support vmap/vunmap, Christoph Hellwig, SGI, August 2002 |
930fc45a4
|
8 |
* Numa awareness, Christoph Lameter, SGI, June 2005 |
1da177e4c
|
9 |
*/ |
db64fe022
|
10 |
#include <linux/vmalloc.h> |
1da177e4c
|
11 12 13 |
#include <linux/mm.h> #include <linux/module.h> #include <linux/highmem.h> |
c3edc4010
|
14 |
#include <linux/sched/signal.h> |
1da177e4c
|
15 16 17 |
#include <linux/slab.h> #include <linux/spinlock.h> #include <linux/interrupt.h> |
5f6a6a9c4
|
18 |
#include <linux/proc_fs.h> |
a10aa5798
|
19 |
#include <linux/seq_file.h> |
3ac7fe5a4
|
20 |
#include <linux/debugobjects.h> |
230169693
|
21 |
#include <linux/kallsyms.h> |
db64fe022
|
22 |
#include <linux/list.h> |
4da56b99d
|
23 |
#include <linux/notifier.h> |
db64fe022
|
24 25 26 |
#include <linux/rbtree.h> #include <linux/radix-tree.h> #include <linux/rcupdate.h> |
f0aa66179
|
27 |
#include <linux/pfn.h> |
89219d37a
|
28 |
#include <linux/kmemleak.h> |
60063497a
|
29 |
#include <linux/atomic.h> |
3b32123d7
|
30 |
#include <linux/compiler.h> |
32fcfd407
|
31 |
#include <linux/llist.h> |
0f616be12
|
32 |
#include <linux/bitops.h> |
3b32123d7
|
33 |
|
7c0f6ba68
|
34 |
#include <linux/uaccess.h> |
1da177e4c
|
35 |
#include <asm/tlbflush.h> |
2dca6999e
|
36 |
#include <asm/shmparam.h> |
1da177e4c
|
37 |
|
dd56b0464
|
38 |
#include "internal.h" |
32fcfd407
|
39 40 41 42 43 44 45 46 47 48 49 |
struct vfree_deferred { struct llist_head list; struct work_struct wq; }; static DEFINE_PER_CPU(struct vfree_deferred, vfree_deferred); static void __vunmap(const void *, int); static void free_work(struct work_struct *w) { struct vfree_deferred *p = container_of(w, struct vfree_deferred, wq); |
894e58c14
|
50 51 52 53 |
struct llist_node *t, *llnode; llist_for_each_safe(llnode, t, llist_del_all(&p->list)) __vunmap((void *)llnode, 1); |
32fcfd407
|
54 |
} |
db64fe022
|
55 |
/*** Page table manipulation functions ***/ |
b221385bc
|
56 |
|
1da177e4c
|
57 58 59 60 61 62 63 64 65 66 |
static void vunmap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end) { pte_t *pte; pte = pte_offset_kernel(pmd, addr); do { pte_t ptent = ptep_get_and_clear(&init_mm, addr, pte); WARN_ON(!pte_none(ptent) && !pte_present(ptent)); } while (pte++, addr += PAGE_SIZE, addr != end); } |
db64fe022
|
67 |
static void vunmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end) |
1da177e4c
|
68 69 70 71 72 73 74 |
{ pmd_t *pmd; unsigned long next; pmd = pmd_offset(pud, addr); do { next = pmd_addr_end(addr, end); |
b9820d8f3
|
75 76 |
if (pmd_clear_huge(pmd)) continue; |
1da177e4c
|
77 78 79 80 81 |
if (pmd_none_or_clear_bad(pmd)) continue; vunmap_pte_range(pmd, addr, next); } while (pmd++, addr = next, addr != end); } |
c2febafc6
|
82 |
static void vunmap_pud_range(p4d_t *p4d, unsigned long addr, unsigned long end) |
1da177e4c
|
83 84 85 |
{ pud_t *pud; unsigned long next; |
c2febafc6
|
86 |
pud = pud_offset(p4d, addr); |
1da177e4c
|
87 88 |
do { next = pud_addr_end(addr, end); |
b9820d8f3
|
89 90 |
if (pud_clear_huge(pud)) continue; |
1da177e4c
|
91 92 93 94 95 |
if (pud_none_or_clear_bad(pud)) continue; vunmap_pmd_range(pud, addr, next); } while (pud++, addr = next, addr != end); } |
c2febafc6
|
96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 |
static void vunmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end) { p4d_t *p4d; unsigned long next; p4d = p4d_offset(pgd, addr); do { next = p4d_addr_end(addr, end); if (p4d_clear_huge(p4d)) continue; if (p4d_none_or_clear_bad(p4d)) continue; vunmap_pud_range(p4d, addr, next); } while (p4d++, addr = next, addr != end); } |
db64fe022
|
111 |
static void vunmap_page_range(unsigned long addr, unsigned long end) |
1da177e4c
|
112 113 114 |
{ pgd_t *pgd; unsigned long next; |
1da177e4c
|
115 116 117 |
BUG_ON(addr >= end); pgd = pgd_offset_k(addr); |
1da177e4c
|
118 119 120 121 |
do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) continue; |
c2febafc6
|
122 |
vunmap_p4d_range(pgd, addr, next); |
1da177e4c
|
123 |
} while (pgd++, addr = next, addr != end); |
1da177e4c
|
124 125 126 |
} static int vmap_pte_range(pmd_t *pmd, unsigned long addr, |
db64fe022
|
127 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr) |
1da177e4c
|
128 129 |
{ pte_t *pte; |
db64fe022
|
130 131 132 133 |
/* * nr is a running index into the array which helps higher level * callers keep track of where we're up to. */ |
872fec16d
|
134 |
pte = pte_alloc_kernel(pmd, addr); |
1da177e4c
|
135 136 137 |
if (!pte) return -ENOMEM; do { |
db64fe022
|
138 139 140 141 142 |
struct page *page = pages[*nr]; if (WARN_ON(!pte_none(*pte))) return -EBUSY; if (WARN_ON(!page)) |
1da177e4c
|
143 144 |
return -ENOMEM; set_pte_at(&init_mm, addr, pte, mk_pte(page, prot)); |
db64fe022
|
145 |
(*nr)++; |
1da177e4c
|
146 147 148 |
} while (pte++, addr += PAGE_SIZE, addr != end); return 0; } |
db64fe022
|
149 150 |
static int vmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end, pgprot_t prot, struct page **pages, int *nr) |
1da177e4c
|
151 152 153 154 155 156 157 158 159 |
{ pmd_t *pmd; unsigned long next; pmd = pmd_alloc(&init_mm, pud, addr); if (!pmd) return -ENOMEM; do { next = pmd_addr_end(addr, end); |
db64fe022
|
160 |
if (vmap_pte_range(pmd, addr, next, prot, pages, nr)) |
1da177e4c
|
161 162 163 164 |
return -ENOMEM; } while (pmd++, addr = next, addr != end); return 0; } |
c2febafc6
|
165 |
static int vmap_pud_range(p4d_t *p4d, unsigned long addr, |
db64fe022
|
166 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr) |
1da177e4c
|
167 168 169 |
{ pud_t *pud; unsigned long next; |
c2febafc6
|
170 |
pud = pud_alloc(&init_mm, p4d, addr); |
1da177e4c
|
171 172 173 174 |
if (!pud) return -ENOMEM; do { next = pud_addr_end(addr, end); |
db64fe022
|
175 |
if (vmap_pmd_range(pud, addr, next, prot, pages, nr)) |
1da177e4c
|
176 177 178 179 |
return -ENOMEM; } while (pud++, addr = next, addr != end); return 0; } |
c2febafc6
|
180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 |
static int vmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end, pgprot_t prot, struct page **pages, int *nr) { p4d_t *p4d; unsigned long next; p4d = p4d_alloc(&init_mm, pgd, addr); if (!p4d) return -ENOMEM; do { next = p4d_addr_end(addr, end); if (vmap_pud_range(p4d, addr, next, prot, pages, nr)) return -ENOMEM; } while (p4d++, addr = next, addr != end); return 0; } |
db64fe022
|
196 197 198 199 200 201 |
/* * Set up page tables in kva (addr, end). The ptes shall have prot "prot", and * will have pfns corresponding to the "pages" array. * * Ie. pte at addr+N*PAGE_SIZE shall point to pfn corresponding to pages[N] */ |
8fc489850
|
202 203 |
static int vmap_page_range_noflush(unsigned long start, unsigned long end, pgprot_t prot, struct page **pages) |
1da177e4c
|
204 205 206 |
{ pgd_t *pgd; unsigned long next; |
2e4e27c7d
|
207 |
unsigned long addr = start; |
db64fe022
|
208 209 |
int err = 0; int nr = 0; |
1da177e4c
|
210 211 212 |
BUG_ON(addr >= end); pgd = pgd_offset_k(addr); |
1da177e4c
|
213 214 |
do { next = pgd_addr_end(addr, end); |
c2febafc6
|
215 |
err = vmap_p4d_range(pgd, addr, next, prot, pages, &nr); |
1da177e4c
|
216 |
if (err) |
bf88c8c83
|
217 |
return err; |
1da177e4c
|
218 |
} while (pgd++, addr = next, addr != end); |
db64fe022
|
219 |
|
db64fe022
|
220 |
return nr; |
1da177e4c
|
221 |
} |
8fc489850
|
222 223 224 225 226 227 228 229 230 |
static int vmap_page_range(unsigned long start, unsigned long end, pgprot_t prot, struct page **pages) { int ret; ret = vmap_page_range_noflush(start, end, prot, pages); flush_cache_vmap(start, end); return ret; } |
81ac3ad90
|
231 |
int is_vmalloc_or_module_addr(const void *x) |
73bdf0a60
|
232 233 |
{ /* |
ab4f2ee13
|
234 |
* ARM, x86-64 and sparc64 put modules in a special place, |
73bdf0a60
|
235 236 237 238 239 240 241 242 243 244 |
* and fall back on vmalloc() if that fails. Others * just put it in the vmalloc space. */ #if defined(CONFIG_MODULES) && defined(MODULES_VADDR) unsigned long addr = (unsigned long)x; if (addr >= MODULES_VADDR && addr < MODULES_END) return 1; #endif return is_vmalloc_addr(x); } |
48667e7a4
|
245 |
/* |
add688fbd
|
246 |
* Walk a vmap address to the struct page it maps. |
48667e7a4
|
247 |
*/ |
add688fbd
|
248 |
struct page *vmalloc_to_page(const void *vmalloc_addr) |
48667e7a4
|
249 250 |
{ unsigned long addr = (unsigned long) vmalloc_addr; |
add688fbd
|
251 |
struct page *page = NULL; |
48667e7a4
|
252 |
pgd_t *pgd = pgd_offset_k(addr); |
c2febafc6
|
253 254 255 256 |
p4d_t *p4d; pud_t *pud; pmd_t *pmd; pte_t *ptep, pte; |
48667e7a4
|
257 |
|
7aa413def
|
258 259 260 261 |
/* * XXX we might need to change this if we add VIRTUAL_BUG_ON for * architectures that do not vmalloc module space */ |
73bdf0a60
|
262 |
VIRTUAL_BUG_ON(!is_vmalloc_or_module_addr(vmalloc_addr)); |
59ea74633
|
263 |
|
c2febafc6
|
264 265 266 267 268 269 |
if (pgd_none(*pgd)) return NULL; p4d = p4d_offset(pgd, addr); if (p4d_none(*p4d)) return NULL; pud = pud_offset(p4d, addr); |
029c54b09
|
270 271 272 273 274 275 276 277 278 279 280 |
/* * Don't dereference bad PUD or PMD (below) entries. This will also * identify huge mappings, which we may encounter on architectures * that define CONFIG_HAVE_ARCH_HUGE_VMAP=y. Such regions will be * identified as vmalloc addresses by is_vmalloc_addr(), but are * not [unambiguously] associated with a struct page, so there is * no correct value to return for them. */ WARN_ON_ONCE(pud_bad(*pud)); if (pud_none(*pud) || pud_bad(*pud)) |
c2febafc6
|
281 282 |
return NULL; pmd = pmd_offset(pud, addr); |
029c54b09
|
283 284 |
WARN_ON_ONCE(pmd_bad(*pmd)); if (pmd_none(*pmd) || pmd_bad(*pmd)) |
c2febafc6
|
285 286 287 288 289 290 291 |
return NULL; ptep = pte_offset_map(pmd, addr); pte = *ptep; if (pte_present(pte)) page = pte_page(pte); pte_unmap(ptep); |
add688fbd
|
292 |
return page; |
48667e7a4
|
293 |
} |
add688fbd
|
294 |
EXPORT_SYMBOL(vmalloc_to_page); |
48667e7a4
|
295 296 |
/* |
add688fbd
|
297 |
* Map a vmalloc()-space virtual address to the physical page frame number. |
48667e7a4
|
298 |
*/ |
add688fbd
|
299 |
unsigned long vmalloc_to_pfn(const void *vmalloc_addr) |
48667e7a4
|
300 |
{ |
add688fbd
|
301 |
return page_to_pfn(vmalloc_to_page(vmalloc_addr)); |
48667e7a4
|
302 |
} |
add688fbd
|
303 |
EXPORT_SYMBOL(vmalloc_to_pfn); |
48667e7a4
|
304 |
|
db64fe022
|
305 306 |
/*** Global kva allocator ***/ |
78c72746f
|
307 |
#define VM_LAZY_FREE 0x02 |
db64fe022
|
308 |
#define VM_VM_AREA 0x04 |
db64fe022
|
309 |
static DEFINE_SPINLOCK(vmap_area_lock); |
f1c4069e1
|
310 311 |
/* Export for kexec only */ LIST_HEAD(vmap_area_list); |
80c4bd7a5
|
312 |
static LLIST_HEAD(vmap_purge_list); |
89699605f
|
313 314 315 316 317 318 319 |
static struct rb_root vmap_area_root = RB_ROOT; /* The vmap cache globals are protected by vmap_area_lock */ static struct rb_node *free_vmap_cache; static unsigned long cached_hole_size; static unsigned long cached_vstart; static unsigned long cached_align; |
ca23e405e
|
320 |
static unsigned long vmap_area_pcpu_hole; |
db64fe022
|
321 322 |
static struct vmap_area *__find_vmap_area(unsigned long addr) |
1da177e4c
|
323 |
{ |
db64fe022
|
324 325 326 327 328 329 330 331 |
struct rb_node *n = vmap_area_root.rb_node; while (n) { struct vmap_area *va; va = rb_entry(n, struct vmap_area, rb_node); if (addr < va->va_start) n = n->rb_left; |
cef2ac3f6
|
332 |
else if (addr >= va->va_end) |
db64fe022
|
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 |
n = n->rb_right; else return va; } return NULL; } static void __insert_vmap_area(struct vmap_area *va) { struct rb_node **p = &vmap_area_root.rb_node; struct rb_node *parent = NULL; struct rb_node *tmp; while (*p) { |
170168d0a
|
348 |
struct vmap_area *tmp_va; |
db64fe022
|
349 350 |
parent = *p; |
170168d0a
|
351 352 |
tmp_va = rb_entry(parent, struct vmap_area, rb_node); if (va->va_start < tmp_va->va_end) |
db64fe022
|
353 |
p = &(*p)->rb_left; |
170168d0a
|
354 |
else if (va->va_end > tmp_va->va_start) |
db64fe022
|
355 356 357 358 359 360 361 |
p = &(*p)->rb_right; else BUG(); } rb_link_node(&va->rb_node, parent, p); rb_insert_color(&va->rb_node, &vmap_area_root); |
4341fa454
|
362 |
/* address-sort this list */ |
db64fe022
|
363 364 365 366 367 368 369 370 371 372 |
tmp = rb_prev(&va->rb_node); if (tmp) { struct vmap_area *prev; prev = rb_entry(tmp, struct vmap_area, rb_node); list_add_rcu(&va->list, &prev->list); } else list_add_rcu(&va->list, &vmap_area_list); } static void purge_vmap_area_lazy(void); |
4da56b99d
|
373 |
static BLOCKING_NOTIFIER_HEAD(vmap_notify_list); |
db64fe022
|
374 375 376 377 378 379 380 381 382 383 384 |
/* * Allocate a region of KVA of the specified size and alignment, within the * vstart and vend. */ static struct vmap_area *alloc_vmap_area(unsigned long size, unsigned long align, unsigned long vstart, unsigned long vend, int node, gfp_t gfp_mask) { struct vmap_area *va; struct rb_node *n; |
1da177e4c
|
385 |
unsigned long addr; |
db64fe022
|
386 |
int purged = 0; |
89699605f
|
387 |
struct vmap_area *first; |
db64fe022
|
388 |
|
7766970cc
|
389 |
BUG_ON(!size); |
891c49abf
|
390 |
BUG_ON(offset_in_page(size)); |
89699605f
|
391 |
BUG_ON(!is_power_of_2(align)); |
db64fe022
|
392 |
|
5803ed292
|
393 |
might_sleep(); |
4da56b99d
|
394 |
|
db64fe022
|
395 396 397 398 |
va = kmalloc_node(sizeof(struct vmap_area), gfp_mask & GFP_RECLAIM_MASK, node); if (unlikely(!va)) return ERR_PTR(-ENOMEM); |
7f88f88f8
|
399 400 401 402 403 |
/* * Only scan the relevant parts containing pointers to other objects * to avoid false negatives. */ kmemleak_scan_area(&va->rb_node, SIZE_MAX, gfp_mask & GFP_RECLAIM_MASK); |
db64fe022
|
404 405 |
retry: spin_lock(&vmap_area_lock); |
89699605f
|
406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 |
/* * Invalidate cache if we have more permissive parameters. * cached_hole_size notes the largest hole noticed _below_ * the vmap_area cached in free_vmap_cache: if size fits * into that hole, we want to scan from vstart to reuse * the hole instead of allocating above free_vmap_cache. * Note that __free_vmap_area may update free_vmap_cache * without updating cached_hole_size or cached_align. */ if (!free_vmap_cache || size < cached_hole_size || vstart < cached_vstart || align < cached_align) { nocache: cached_hole_size = 0; free_vmap_cache = NULL; } /* record if we encounter less permissive parameters */ cached_vstart = vstart; cached_align = align; /* find starting point for our search */ if (free_vmap_cache) { first = rb_entry(free_vmap_cache, struct vmap_area, rb_node); |
248ac0e19
|
430 |
addr = ALIGN(first->va_end, align); |
89699605f
|
431 432 |
if (addr < vstart) goto nocache; |
bcb615a81
|
433 |
if (addr + size < addr) |
89699605f
|
434 435 436 437 |
goto overflow; } else { addr = ALIGN(vstart, align); |
bcb615a81
|
438 |
if (addr + size < addr) |
89699605f
|
439 440 441 442 443 444 |
goto overflow; n = vmap_area_root.rb_node; first = NULL; while (n) { |
db64fe022
|
445 446 447 |
struct vmap_area *tmp; tmp = rb_entry(n, struct vmap_area, rb_node); if (tmp->va_end >= addr) { |
db64fe022
|
448 |
first = tmp; |
89699605f
|
449 450 451 452 |
if (tmp->va_start <= addr) break; n = n->rb_left; } else |
db64fe022
|
453 |
n = n->rb_right; |
89699605f
|
454 |
} |
db64fe022
|
455 456 457 |
if (!first) goto found; |
db64fe022
|
458 |
} |
89699605f
|
459 460 |
/* from the starting point, walk areas until a suitable hole is found */ |
248ac0e19
|
461 |
while (addr + size > first->va_start && addr + size <= vend) { |
89699605f
|
462 463 |
if (addr + cached_hole_size < first->va_start) cached_hole_size = first->va_start - addr; |
248ac0e19
|
464 |
addr = ALIGN(first->va_end, align); |
bcb615a81
|
465 |
if (addr + size < addr) |
89699605f
|
466 |
goto overflow; |
92ca922f0
|
467 |
if (list_is_last(&first->list, &vmap_area_list)) |
89699605f
|
468 |
goto found; |
92ca922f0
|
469 |
|
6219c2a2e
|
470 |
first = list_next_entry(first, list); |
db64fe022
|
471 |
} |
89699605f
|
472 473 474 |
found: if (addr + size > vend) goto overflow; |
db64fe022
|
475 476 477 478 479 |
va->va_start = addr; va->va_end = addr + size; va->flags = 0; __insert_vmap_area(va); |
89699605f
|
480 |
free_vmap_cache = &va->rb_node; |
db64fe022
|
481 |
spin_unlock(&vmap_area_lock); |
61e165578
|
482 |
BUG_ON(!IS_ALIGNED(va->va_start, align)); |
89699605f
|
483 484 |
BUG_ON(va->va_start < vstart); BUG_ON(va->va_end > vend); |
db64fe022
|
485 |
return va; |
89699605f
|
486 487 488 489 490 491 492 493 |
overflow: spin_unlock(&vmap_area_lock); if (!purged) { purge_vmap_area_lazy(); purged = 1; goto retry; } |
4da56b99d
|
494 495 496 497 498 499 500 501 502 |
if (gfpflags_allow_blocking(gfp_mask)) { unsigned long freed = 0; blocking_notifier_call_chain(&vmap_notify_list, 0, &freed); if (freed > 0) { purged = 0; goto retry; } } |
03497d761
|
503 |
if (!(gfp_mask & __GFP_NOWARN) && printk_ratelimit()) |
756a025f0
|
504 505 506 |
pr_warn("vmap allocation for size %lu failed: use vmalloc=<size> to increase size ", size); |
89699605f
|
507 508 |
kfree(va); return ERR_PTR(-EBUSY); |
db64fe022
|
509 |
} |
4da56b99d
|
510 511 512 513 514 515 516 517 518 519 520 |
int register_vmap_purge_notifier(struct notifier_block *nb) { return blocking_notifier_chain_register(&vmap_notify_list, nb); } EXPORT_SYMBOL_GPL(register_vmap_purge_notifier); int unregister_vmap_purge_notifier(struct notifier_block *nb) { return blocking_notifier_chain_unregister(&vmap_notify_list, nb); } EXPORT_SYMBOL_GPL(unregister_vmap_purge_notifier); |
db64fe022
|
521 522 523 |
static void __free_vmap_area(struct vmap_area *va) { BUG_ON(RB_EMPTY_NODE(&va->rb_node)); |
89699605f
|
524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 |
if (free_vmap_cache) { if (va->va_end < cached_vstart) { free_vmap_cache = NULL; } else { struct vmap_area *cache; cache = rb_entry(free_vmap_cache, struct vmap_area, rb_node); if (va->va_start <= cache->va_start) { free_vmap_cache = rb_prev(&va->rb_node); /* * We don't try to update cached_hole_size or * cached_align, but it won't go very wrong. */ } } } |
db64fe022
|
540 541 542 |
rb_erase(&va->rb_node, &vmap_area_root); RB_CLEAR_NODE(&va->rb_node); list_del_rcu(&va->list); |
ca23e405e
|
543 544 545 546 547 548 549 550 |
/* * Track the highest possible candidate for pcpu area * allocation. Areas outside of vmalloc area can be returned * here too, consider only end addresses which fall inside * vmalloc area proper. */ if (va->va_end > VMALLOC_START && va->va_end <= VMALLOC_END) vmap_area_pcpu_hole = max(vmap_area_pcpu_hole, va->va_end); |
14769de93
|
551 |
kfree_rcu(va, rcu_head); |
db64fe022
|
552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 |
} /* * Free a region of KVA allocated by alloc_vmap_area */ static void free_vmap_area(struct vmap_area *va) { spin_lock(&vmap_area_lock); __free_vmap_area(va); spin_unlock(&vmap_area_lock); } /* * Clear the pagetable entries of a given vmap_area */ static void unmap_vmap_area(struct vmap_area *va) { vunmap_page_range(va->va_start, va->va_end); } /* * lazy_max_pages is the maximum amount of virtual address space we gather up * before attempting to purge with a TLB flush. * * There is a tradeoff here: a larger number will cover more kernel page tables * and take slightly longer to purge, but it will linearly reduce the number of * global TLB flushes that must be performed. It would seem natural to scale * this number up linearly with the number of CPUs (because vmapping activity * could also scale linearly with the number of CPUs), however it is likely * that in practice, workloads might be constrained in other ways that mean * vmap activity will not scale linearly with CPUs. Also, I want to be * conservative and not introduce a big latency on huge systems, so go with * a less aggressive log scale. It will still be an improvement over the old * code, and it will be simple to change the scale factor if we find that it * becomes a problem on bigger systems. */ static unsigned long lazy_max_pages(void) { unsigned int log; log = fls(num_online_cpus()); return log * (32UL * 1024 * 1024 / PAGE_SIZE); } static atomic_t vmap_lazy_nr = ATOMIC_INIT(0); |
0574ecd14
|
598 599 600 601 602 |
/* * Serialize vmap purging. There is no actual criticial section protected * by this look, but we want to avoid concurrent calls for performance * reasons and to make the pcpu_get_vm_areas more deterministic. */ |
f9e099776
|
603 |
static DEFINE_MUTEX(vmap_purge_lock); |
0574ecd14
|
604 |
|
02b709df8
|
605 606 |
/* for per-CPU blocks */ static void purge_fragmented_blocks_allcpus(void); |
db64fe022
|
607 |
/* |
3ee48b6af
|
608 609 610 611 612 613 614 615 616 |
* called before a call to iounmap() if the caller wants vm_area_struct's * immediately freed. */ void set_iounmap_nonlazy(void) { atomic_set(&vmap_lazy_nr, lazy_max_pages()+1); } /* |
db64fe022
|
617 |
* Purges all lazily-freed vmap areas. |
db64fe022
|
618 |
*/ |
0574ecd14
|
619 |
static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end) |
db64fe022
|
620 |
{ |
80c4bd7a5
|
621 |
struct llist_node *valist; |
db64fe022
|
622 |
struct vmap_area *va; |
cbb766766
|
623 |
struct vmap_area *n_va; |
763b218dd
|
624 |
bool do_free = false; |
db64fe022
|
625 |
|
0574ecd14
|
626 |
lockdep_assert_held(&vmap_purge_lock); |
02b709df8
|
627 |
|
80c4bd7a5
|
628 629 |
valist = llist_del_all(&vmap_purge_list); llist_for_each_entry(va, valist, purge_list) { |
0574ecd14
|
630 631 632 633 |
if (va->va_start < start) start = va->va_start; if (va->va_end > end) end = va->va_end; |
763b218dd
|
634 |
do_free = true; |
db64fe022
|
635 |
} |
db64fe022
|
636 |
|
763b218dd
|
637 |
if (!do_free) |
0574ecd14
|
638 |
return false; |
db64fe022
|
639 |
|
0574ecd14
|
640 |
flush_tlb_kernel_range(start, end); |
db64fe022
|
641 |
|
0574ecd14
|
642 |
spin_lock(&vmap_area_lock); |
763b218dd
|
643 644 |
llist_for_each_entry_safe(va, n_va, valist, purge_list) { int nr = (va->va_end - va->va_start) >> PAGE_SHIFT; |
0574ecd14
|
645 |
__free_vmap_area(va); |
763b218dd
|
646 647 648 |
atomic_sub(nr, &vmap_lazy_nr); cond_resched_lock(&vmap_area_lock); } |
0574ecd14
|
649 650 |
spin_unlock(&vmap_area_lock); return true; |
db64fe022
|
651 652 653 |
} /* |
496850e5f
|
654 655 656 657 658 |
* Kick off a purge of the outstanding lazy areas. Don't bother if somebody * is already purging. */ static void try_purge_vmap_area_lazy(void) { |
f9e099776
|
659 |
if (mutex_trylock(&vmap_purge_lock)) { |
0574ecd14
|
660 |
__purge_vmap_area_lazy(ULONG_MAX, 0); |
f9e099776
|
661 |
mutex_unlock(&vmap_purge_lock); |
0574ecd14
|
662 |
} |
496850e5f
|
663 664 665 |
} /* |
db64fe022
|
666 667 668 669 |
* Kick off a purge of the outstanding lazy areas. */ static void purge_vmap_area_lazy(void) { |
f9e099776
|
670 |
mutex_lock(&vmap_purge_lock); |
0574ecd14
|
671 672 |
purge_fragmented_blocks_allcpus(); __purge_vmap_area_lazy(ULONG_MAX, 0); |
f9e099776
|
673 |
mutex_unlock(&vmap_purge_lock); |
db64fe022
|
674 675 676 |
} /* |
64141da58
|
677 678 679 |
* Free a vmap area, caller ensuring that the area has been unmapped * and flush_cache_vunmap had been called for the correct range * previously. |
db64fe022
|
680 |
*/ |
64141da58
|
681 |
static void free_vmap_area_noflush(struct vmap_area *va) |
db64fe022
|
682 |
{ |
80c4bd7a5
|
683 684 685 686 687 688 689 690 691 |
int nr_lazy; nr_lazy = atomic_add_return((va->va_end - va->va_start) >> PAGE_SHIFT, &vmap_lazy_nr); /* After this point, we may free va at any time */ llist_add(&va->purge_list, &vmap_purge_list); if (unlikely(nr_lazy > lazy_max_pages())) |
496850e5f
|
692 |
try_purge_vmap_area_lazy(); |
db64fe022
|
693 |
} |
b29acbdcf
|
694 695 696 697 698 699 |
/* * Free and unmap a vmap area */ static void free_unmap_vmap_area(struct vmap_area *va) { flush_cache_vunmap(va->va_start, va->va_end); |
c8eef01e2
|
700 |
unmap_vmap_area(va); |
82a2e924f
|
701 702 |
if (debug_pagealloc_enabled()) flush_tlb_kernel_range(va->va_start, va->va_end); |
c8eef01e2
|
703 |
free_vmap_area_noflush(va); |
b29acbdcf
|
704 |
} |
db64fe022
|
705 706 707 708 709 710 711 712 713 714 |
static struct vmap_area *find_vmap_area(unsigned long addr) { struct vmap_area *va; spin_lock(&vmap_area_lock); va = __find_vmap_area(addr); spin_unlock(&vmap_area_lock); return va; } |
db64fe022
|
715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 |
/*** Per cpu kva allocator ***/ /* * vmap space is limited especially on 32 bit architectures. Ensure there is * room for at least 16 percpu vmap blocks per CPU. */ /* * If we had a constant VMALLOC_START and VMALLOC_END, we'd like to be able * to #define VMALLOC_SPACE (VMALLOC_END-VMALLOC_START). Guess * instead (we just need a rough idea) */ #if BITS_PER_LONG == 32 #define VMALLOC_SPACE (128UL*1024*1024) #else #define VMALLOC_SPACE (128UL*1024*1024*1024) #endif #define VMALLOC_PAGES (VMALLOC_SPACE / PAGE_SIZE) #define VMAP_MAX_ALLOC BITS_PER_LONG /* 256K with 4K pages */ #define VMAP_BBMAP_BITS_MAX 1024 /* 4MB with 4K pages */ #define VMAP_BBMAP_BITS_MIN (VMAP_MAX_ALLOC*2) #define VMAP_MIN(x, y) ((x) < (y) ? (x) : (y)) /* can't use min() */ #define VMAP_MAX(x, y) ((x) > (y) ? (x) : (y)) /* can't use max() */ |
f982f9151
|
738 739 740 741 |
#define VMAP_BBMAP_BITS \ VMAP_MIN(VMAP_BBMAP_BITS_MAX, \ VMAP_MAX(VMAP_BBMAP_BITS_MIN, \ VMALLOC_PAGES / roundup_pow_of_two(NR_CPUS) / 16)) |
db64fe022
|
742 743 |
#define VMAP_BLOCK_SIZE (VMAP_BBMAP_BITS * PAGE_SIZE) |
9b4633340
|
744 |
static bool vmap_initialized __read_mostly = false; |
db64fe022
|
745 746 747 |
struct vmap_block_queue { spinlock_t lock; struct list_head free; |
db64fe022
|
748 749 750 751 752 |
}; struct vmap_block { spinlock_t lock; struct vmap_area *va; |
db64fe022
|
753 |
unsigned long free, dirty; |
7d61bfe8f
|
754 |
unsigned long dirty_min, dirty_max; /*< dirty range */ |
de5604231
|
755 756 |
struct list_head free_list; struct rcu_head rcu_head; |
02b709df8
|
757 |
struct list_head purge; |
db64fe022
|
758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 |
}; /* Queue of free and dirty vmap blocks, for allocation and flushing purposes */ static DEFINE_PER_CPU(struct vmap_block_queue, vmap_block_queue); /* * Radix tree of vmap blocks, indexed by address, to quickly find a vmap block * in the free path. Could get rid of this if we change the API to return a * "cookie" from alloc, to be passed to free. But no big deal yet. */ static DEFINE_SPINLOCK(vmap_block_tree_lock); static RADIX_TREE(vmap_block_tree, GFP_ATOMIC); /* * We should probably have a fallback mechanism to allocate virtual memory * out of partially filled vmap blocks. However vmap block sizing should be * fairly reasonable according to the vmalloc size, so it shouldn't be a * big problem. */ static unsigned long addr_to_vb_idx(unsigned long addr) { addr -= VMALLOC_START & ~(VMAP_BLOCK_SIZE-1); addr /= VMAP_BLOCK_SIZE; return addr; } |
cf725ce27
|
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 |
static void *vmap_block_vaddr(unsigned long va_start, unsigned long pages_off) { unsigned long addr; addr = va_start + (pages_off << PAGE_SHIFT); BUG_ON(addr_to_vb_idx(addr) != addr_to_vb_idx(va_start)); return (void *)addr; } /** * new_vmap_block - allocates new vmap_block and occupies 2^order pages in this * block. Of course pages number can't exceed VMAP_BBMAP_BITS * @order: how many 2^order pages should be occupied in newly allocated block * @gfp_mask: flags for the page level allocator * * Returns: virtual address in a newly allocated block or ERR_PTR(-errno) */ static void *new_vmap_block(unsigned int order, gfp_t gfp_mask) |
db64fe022
|
802 803 804 805 806 807 |
{ struct vmap_block_queue *vbq; struct vmap_block *vb; struct vmap_area *va; unsigned long vb_idx; int node, err; |
cf725ce27
|
808 |
void *vaddr; |
db64fe022
|
809 810 811 812 813 814 815 816 817 818 819 |
node = numa_node_id(); vb = kmalloc_node(sizeof(struct vmap_block), gfp_mask & GFP_RECLAIM_MASK, node); if (unlikely(!vb)) return ERR_PTR(-ENOMEM); va = alloc_vmap_area(VMAP_BLOCK_SIZE, VMAP_BLOCK_SIZE, VMALLOC_START, VMALLOC_END, node, gfp_mask); |
ddf9c6d47
|
820 |
if (IS_ERR(va)) { |
db64fe022
|
821 |
kfree(vb); |
e7d863407
|
822 |
return ERR_CAST(va); |
db64fe022
|
823 824 825 826 827 828 829 830 |
} err = radix_tree_preload(gfp_mask); if (unlikely(err)) { kfree(vb); free_vmap_area(va); return ERR_PTR(err); } |
cf725ce27
|
831 |
vaddr = vmap_block_vaddr(va->va_start, 0); |
db64fe022
|
832 833 |
spin_lock_init(&vb->lock); vb->va = va; |
cf725ce27
|
834 835 836 |
/* At least something should be left free */ BUG_ON(VMAP_BBMAP_BITS <= (1UL << order)); vb->free = VMAP_BBMAP_BITS - (1UL << order); |
db64fe022
|
837 |
vb->dirty = 0; |
7d61bfe8f
|
838 839 |
vb->dirty_min = VMAP_BBMAP_BITS; vb->dirty_max = 0; |
db64fe022
|
840 |
INIT_LIST_HEAD(&vb->free_list); |
db64fe022
|
841 842 843 844 845 846 847 848 849 |
vb_idx = addr_to_vb_idx(va->va_start); spin_lock(&vmap_block_tree_lock); err = radix_tree_insert(&vmap_block_tree, vb_idx, vb); spin_unlock(&vmap_block_tree_lock); BUG_ON(err); radix_tree_preload_end(); vbq = &get_cpu_var(vmap_block_queue); |
db64fe022
|
850 |
spin_lock(&vbq->lock); |
68ac546f2
|
851 |
list_add_tail_rcu(&vb->free_list, &vbq->free); |
db64fe022
|
852 |
spin_unlock(&vbq->lock); |
3f04ba859
|
853 |
put_cpu_var(vmap_block_queue); |
db64fe022
|
854 |
|
cf725ce27
|
855 |
return vaddr; |
db64fe022
|
856 |
} |
db64fe022
|
857 858 859 860 |
static void free_vmap_block(struct vmap_block *vb) { struct vmap_block *tmp; unsigned long vb_idx; |
db64fe022
|
861 862 863 864 865 |
vb_idx = addr_to_vb_idx(vb->va->va_start); spin_lock(&vmap_block_tree_lock); tmp = radix_tree_delete(&vmap_block_tree, vb_idx); spin_unlock(&vmap_block_tree_lock); BUG_ON(tmp != vb); |
64141da58
|
866 |
free_vmap_area_noflush(vb->va); |
22a3c7d18
|
867 |
kfree_rcu(vb, rcu_head); |
db64fe022
|
868 |
} |
02b709df8
|
869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 |
static void purge_fragmented_blocks(int cpu) { LIST_HEAD(purge); struct vmap_block *vb; struct vmap_block *n_vb; struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu); rcu_read_lock(); list_for_each_entry_rcu(vb, &vbq->free, free_list) { if (!(vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS)) continue; spin_lock(&vb->lock); if (vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS) { vb->free = 0; /* prevent further allocs after releasing lock */ vb->dirty = VMAP_BBMAP_BITS; /* prevent purging it again */ |
7d61bfe8f
|
886 887 |
vb->dirty_min = 0; vb->dirty_max = VMAP_BBMAP_BITS; |
02b709df8
|
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 |
spin_lock(&vbq->lock); list_del_rcu(&vb->free_list); spin_unlock(&vbq->lock); spin_unlock(&vb->lock); list_add_tail(&vb->purge, &purge); } else spin_unlock(&vb->lock); } rcu_read_unlock(); list_for_each_entry_safe(vb, n_vb, &purge, purge) { list_del(&vb->purge); free_vmap_block(vb); } } |
02b709df8
|
903 904 905 906 907 908 909 |
static void purge_fragmented_blocks_allcpus(void) { int cpu; for_each_possible_cpu(cpu) purge_fragmented_blocks(cpu); } |
db64fe022
|
910 911 912 913 |
static void *vb_alloc(unsigned long size, gfp_t gfp_mask) { struct vmap_block_queue *vbq; struct vmap_block *vb; |
cf725ce27
|
914 |
void *vaddr = NULL; |
db64fe022
|
915 |
unsigned int order; |
891c49abf
|
916 |
BUG_ON(offset_in_page(size)); |
db64fe022
|
917 |
BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC); |
aa91c4d89
|
918 919 920 921 922 923 924 925 |
if (WARN_ON(size == 0)) { /* * Allocating 0 bytes isn't what caller wants since * get_order(0) returns funny result. Just warn and terminate * early. */ return NULL; } |
db64fe022
|
926 |
order = get_order(size); |
db64fe022
|
927 928 929 |
rcu_read_lock(); vbq = &get_cpu_var(vmap_block_queue); list_for_each_entry_rcu(vb, &vbq->free, free_list) { |
cf725ce27
|
930 |
unsigned long pages_off; |
db64fe022
|
931 932 |
spin_lock(&vb->lock); |
cf725ce27
|
933 934 935 936 |
if (vb->free < (1UL << order)) { spin_unlock(&vb->lock); continue; } |
02b709df8
|
937 |
|
cf725ce27
|
938 939 |
pages_off = VMAP_BBMAP_BITS - vb->free; vaddr = vmap_block_vaddr(vb->va->va_start, pages_off); |
02b709df8
|
940 941 942 943 944 945 |
vb->free -= 1UL << order; if (vb->free == 0) { spin_lock(&vbq->lock); list_del_rcu(&vb->free_list); spin_unlock(&vbq->lock); } |
cf725ce27
|
946 |
|
02b709df8
|
947 948 |
spin_unlock(&vb->lock); break; |
db64fe022
|
949 |
} |
02b709df8
|
950 |
|
3f04ba859
|
951 |
put_cpu_var(vmap_block_queue); |
db64fe022
|
952 |
rcu_read_unlock(); |
cf725ce27
|
953 954 955 |
/* Allocate new block if nothing was found */ if (!vaddr) vaddr = new_vmap_block(order, gfp_mask); |
db64fe022
|
956 |
|
cf725ce27
|
957 |
return vaddr; |
db64fe022
|
958 959 960 961 962 963 964 965 |
} static void vb_free(const void *addr, unsigned long size) { unsigned long offset; unsigned long vb_idx; unsigned int order; struct vmap_block *vb; |
891c49abf
|
966 |
BUG_ON(offset_in_page(size)); |
db64fe022
|
967 |
BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC); |
b29acbdcf
|
968 969 |
flush_cache_vunmap((unsigned long)addr, (unsigned long)addr + size); |
db64fe022
|
970 971 972 |
order = get_order(size); offset = (unsigned long)addr & (VMAP_BLOCK_SIZE - 1); |
7d61bfe8f
|
973 |
offset >>= PAGE_SHIFT; |
db64fe022
|
974 975 976 977 978 979 |
vb_idx = addr_to_vb_idx((unsigned long)addr); rcu_read_lock(); vb = radix_tree_lookup(&vmap_block_tree, vb_idx); rcu_read_unlock(); BUG_ON(!vb); |
64141da58
|
980 |
vunmap_page_range((unsigned long)addr, (unsigned long)addr + size); |
82a2e924f
|
981 982 983 |
if (debug_pagealloc_enabled()) flush_tlb_kernel_range((unsigned long)addr, (unsigned long)addr + size); |
db64fe022
|
984 |
spin_lock(&vb->lock); |
7d61bfe8f
|
985 986 987 988 |
/* Expand dirty range */ vb->dirty_min = min(vb->dirty_min, offset); vb->dirty_max = max(vb->dirty_max, offset + (1UL << order)); |
d086817dc
|
989 |
|
db64fe022
|
990 991 |
vb->dirty += 1UL << order; if (vb->dirty == VMAP_BBMAP_BITS) { |
de5604231
|
992 |
BUG_ON(vb->free); |
db64fe022
|
993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 |
spin_unlock(&vb->lock); free_vmap_block(vb); } else spin_unlock(&vb->lock); } /** * vm_unmap_aliases - unmap outstanding lazy aliases in the vmap layer * * The vmap/vmalloc layer lazily flushes kernel virtual mappings primarily * to amortize TLB flushing overheads. What this means is that any page you * have now, may, in a former life, have been mapped into kernel virtual * address by the vmap layer and so there might be some CPUs with TLB entries * still referencing that page (additional to the regular 1:1 kernel mapping). * * vm_unmap_aliases flushes all such lazy mappings. After it returns, we can * be sure that none of the pages we have control over will have any aliases * from the vmap layer. */ void vm_unmap_aliases(void) { unsigned long start = ULONG_MAX, end = 0; int cpu; int flush = 0; |
9b4633340
|
1017 1018 |
if (unlikely(!vmap_initialized)) return; |
5803ed292
|
1019 |
might_sleep(); |
db64fe022
|
1020 1021 1022 1023 1024 1025 |
for_each_possible_cpu(cpu) { struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu); struct vmap_block *vb; rcu_read_lock(); list_for_each_entry_rcu(vb, &vbq->free, free_list) { |
db64fe022
|
1026 |
spin_lock(&vb->lock); |
7d61bfe8f
|
1027 1028 |
if (vb->dirty) { unsigned long va_start = vb->va->va_start; |
db64fe022
|
1029 |
unsigned long s, e; |
b136be5e0
|
1030 |
|
7d61bfe8f
|
1031 1032 |
s = va_start + (vb->dirty_min << PAGE_SHIFT); e = va_start + (vb->dirty_max << PAGE_SHIFT); |
db64fe022
|
1033 |
|
7d61bfe8f
|
1034 1035 |
start = min(s, start); end = max(e, end); |
db64fe022
|
1036 |
|
7d61bfe8f
|
1037 |
flush = 1; |
db64fe022
|
1038 1039 1040 1041 1042 |
} spin_unlock(&vb->lock); } rcu_read_unlock(); } |
f9e099776
|
1043 |
mutex_lock(&vmap_purge_lock); |
0574ecd14
|
1044 1045 1046 |
purge_fragmented_blocks_allcpus(); if (!__purge_vmap_area_lazy(start, end) && flush) flush_tlb_kernel_range(start, end); |
f9e099776
|
1047 |
mutex_unlock(&vmap_purge_lock); |
db64fe022
|
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 |
} EXPORT_SYMBOL_GPL(vm_unmap_aliases); /** * vm_unmap_ram - unmap linear kernel address space set up by vm_map_ram * @mem: the pointer returned by vm_map_ram * @count: the count passed to that vm_map_ram call (cannot unmap partial) */ void vm_unmap_ram(const void *mem, unsigned int count) { |
65ee03c4b
|
1058 |
unsigned long size = (unsigned long)count << PAGE_SHIFT; |
db64fe022
|
1059 |
unsigned long addr = (unsigned long)mem; |
9c3acf604
|
1060 |
struct vmap_area *va; |
db64fe022
|
1061 |
|
5803ed292
|
1062 |
might_sleep(); |
db64fe022
|
1063 1064 1065 |
BUG_ON(!addr); BUG_ON(addr < VMALLOC_START); BUG_ON(addr > VMALLOC_END); |
a1c0b1a07
|
1066 |
BUG_ON(!PAGE_ALIGNED(addr)); |
db64fe022
|
1067 |
|
9c3acf604
|
1068 |
if (likely(count <= VMAP_MAX_ALLOC)) { |
05e3ff950
|
1069 |
debug_check_no_locks_freed(mem, size); |
db64fe022
|
1070 |
vb_free(mem, size); |
9c3acf604
|
1071 1072 1073 1074 1075 |
return; } va = find_vmap_area(addr); BUG_ON(!va); |
05e3ff950
|
1076 1077 |
debug_check_no_locks_freed((void *)va->va_start, (va->va_end - va->va_start)); |
9c3acf604
|
1078 |
free_unmap_vmap_area(va); |
db64fe022
|
1079 1080 1081 1082 1083 1084 1085 1086 1087 |
} EXPORT_SYMBOL(vm_unmap_ram); /** * vm_map_ram - map pages linearly into kernel virtual address (vmalloc space) * @pages: an array of pointers to the pages to be mapped * @count: number of pages * @node: prefer to allocate data structures on this node * @prot: memory protection to use. PAGE_KERNEL for regular RAM |
e99c97ade
|
1088 |
* |
364376383
|
1089 1090 1091 1092 1093 1094 |
* If you use this function for less than VMAP_MAX_ALLOC pages, it could be * faster than vmap so it's good. But if you mix long-life and short-life * objects with vm_map_ram(), it could consume lots of address space through * fragmentation (especially on a 32bit machine). You could see failures in * the end. Please use this function for short-lived objects. * |
e99c97ade
|
1095 |
* Returns: a pointer to the address that has been mapped, or %NULL on failure |
db64fe022
|
1096 1097 1098 |
*/ void *vm_map_ram(struct page **pages, unsigned int count, int node, pgprot_t prot) { |
65ee03c4b
|
1099 |
unsigned long size = (unsigned long)count << PAGE_SHIFT; |
db64fe022
|
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 |
unsigned long addr; void *mem; if (likely(count <= VMAP_MAX_ALLOC)) { mem = vb_alloc(size, GFP_KERNEL); if (IS_ERR(mem)) return NULL; addr = (unsigned long)mem; } else { struct vmap_area *va; va = alloc_vmap_area(size, PAGE_SIZE, VMALLOC_START, VMALLOC_END, node, GFP_KERNEL); if (IS_ERR(va)) return NULL; addr = va->va_start; mem = (void *)addr; } if (vmap_page_range(addr, addr + size, prot, pages) < 0) { vm_unmap_ram(mem, count); return NULL; } return mem; } EXPORT_SYMBOL(vm_map_ram); |
4341fa454
|
1125 |
static struct vm_struct *vmlist __initdata; |
f0aa66179
|
1126 |
/** |
be9b7335e
|
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 |
* vm_area_add_early - add vmap area early during boot * @vm: vm_struct to add * * This function is used to add fixed kernel vm area to vmlist before * vmalloc_init() is called. @vm->addr, @vm->size, and @vm->flags * should contain proper values and the other fields should be zero. * * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING. */ void __init vm_area_add_early(struct vm_struct *vm) { struct vm_struct *tmp, **p; BUG_ON(vmap_initialized); for (p = &vmlist; (tmp = *p) != NULL; p = &tmp->next) { if (tmp->addr >= vm->addr) { BUG_ON(tmp->addr < vm->addr + vm->size); break; } else BUG_ON(tmp->addr + tmp->size > vm->addr); } vm->next = *p; *p = vm; } /** |
f0aa66179
|
1153 1154 |
* vm_area_register_early - register vmap area early during boot * @vm: vm_struct to register |
c0c0a2937
|
1155 |
* @align: requested alignment |
f0aa66179
|
1156 1157 1158 1159 1160 1161 1162 1163 |
* * This function is used to register kernel vm area before * vmalloc_init() is called. @vm->size and @vm->flags should contain * proper values on entry and other fields should be zero. On return, * vm->addr contains the allocated address. * * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING. */ |
c0c0a2937
|
1164 |
void __init vm_area_register_early(struct vm_struct *vm, size_t align) |
f0aa66179
|
1165 1166 |
{ static size_t vm_init_off __initdata; |
c0c0a2937
|
1167 1168 1169 1170 |
unsigned long addr; addr = ALIGN(VMALLOC_START + vm_init_off, align); vm_init_off = PFN_ALIGN(addr + vm->size) - VMALLOC_START; |
f0aa66179
|
1171 |
|
c0c0a2937
|
1172 |
vm->addr = (void *)addr; |
f0aa66179
|
1173 |
|
be9b7335e
|
1174 |
vm_area_add_early(vm); |
f0aa66179
|
1175 |
} |
db64fe022
|
1176 1177 |
void __init vmalloc_init(void) { |
822c18f2e
|
1178 1179 |
struct vmap_area *va; struct vm_struct *tmp; |
db64fe022
|
1180 1181 1182 1183 |
int i; for_each_possible_cpu(i) { struct vmap_block_queue *vbq; |
32fcfd407
|
1184 |
struct vfree_deferred *p; |
db64fe022
|
1185 1186 1187 1188 |
vbq = &per_cpu(vmap_block_queue, i); spin_lock_init(&vbq->lock); INIT_LIST_HEAD(&vbq->free); |
32fcfd407
|
1189 1190 1191 |
p = &per_cpu(vfree_deferred, i); init_llist_head(&p->list); INIT_WORK(&p->wq, free_work); |
db64fe022
|
1192 |
} |
9b4633340
|
1193 |
|
822c18f2e
|
1194 1195 |
/* Import existing vmlist entries. */ for (tmp = vmlist; tmp; tmp = tmp->next) { |
43ebdac42
|
1196 |
va = kzalloc(sizeof(struct vmap_area), GFP_NOWAIT); |
dbda591d9
|
1197 |
va->flags = VM_VM_AREA; |
822c18f2e
|
1198 1199 |
va->va_start = (unsigned long)tmp->addr; va->va_end = va->va_start + tmp->size; |
dbda591d9
|
1200 |
va->vm = tmp; |
822c18f2e
|
1201 1202 |
__insert_vmap_area(va); } |
ca23e405e
|
1203 1204 |
vmap_area_pcpu_hole = VMALLOC_END; |
9b4633340
|
1205 |
vmap_initialized = true; |
db64fe022
|
1206 |
} |
8fc489850
|
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 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 |
/** * map_kernel_range_noflush - map kernel VM area with the specified pages * @addr: start of the VM area to map * @size: size of the VM area to map * @prot: page protection flags to use * @pages: pages to map * * Map PFN_UP(@size) pages at @addr. The VM area @addr and @size * specify should have been allocated using get_vm_area() and its * friends. * * NOTE: * This function does NOT do any cache flushing. The caller is * responsible for calling flush_cache_vmap() on to-be-mapped areas * before calling this function. * * RETURNS: * The number of pages mapped on success, -errno on failure. */ int map_kernel_range_noflush(unsigned long addr, unsigned long size, pgprot_t prot, struct page **pages) { return vmap_page_range_noflush(addr, addr + size, prot, pages); } /** * unmap_kernel_range_noflush - unmap kernel VM area * @addr: start of the VM area to unmap * @size: size of the VM area to unmap * * Unmap PFN_UP(@size) pages at @addr. The VM area @addr and @size * specify should have been allocated using get_vm_area() and its * friends. * * NOTE: * This function does NOT do any cache flushing. The caller is * responsible for calling flush_cache_vunmap() on to-be-mapped areas * before calling this function and flush_tlb_kernel_range() after. */ void unmap_kernel_range_noflush(unsigned long addr, unsigned long size) { vunmap_page_range(addr, addr + size); } |
81e88fdc4
|
1250 |
EXPORT_SYMBOL_GPL(unmap_kernel_range_noflush); |
8fc489850
|
1251 1252 1253 1254 1255 1256 1257 1258 1259 |
/** * unmap_kernel_range - unmap kernel VM area and flush cache and TLB * @addr: start of the VM area to unmap * @size: size of the VM area to unmap * * Similar to unmap_kernel_range_noflush() but flushes vcache before * the unmapping and tlb after. */ |
db64fe022
|
1260 1261 1262 |
void unmap_kernel_range(unsigned long addr, unsigned long size) { unsigned long end = addr + size; |
f6fcba701
|
1263 1264 |
flush_cache_vunmap(addr, end); |
db64fe022
|
1265 1266 1267 |
vunmap_page_range(addr, end); flush_tlb_kernel_range(addr, end); } |
93ef6d6ca
|
1268 |
EXPORT_SYMBOL_GPL(unmap_kernel_range); |
db64fe022
|
1269 |
|
f6f8ed473
|
1270 |
int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page **pages) |
db64fe022
|
1271 1272 |
{ unsigned long addr = (unsigned long)area->addr; |
762216ab4
|
1273 |
unsigned long end = addr + get_vm_area_size(area); |
db64fe022
|
1274 |
int err; |
f6f8ed473
|
1275 |
err = vmap_page_range(addr, end, prot, pages); |
db64fe022
|
1276 |
|
f6f8ed473
|
1277 |
return err > 0 ? 0 : err; |
db64fe022
|
1278 1279 |
} EXPORT_SYMBOL_GPL(map_vm_area); |
f5252e009
|
1280 |
static void setup_vmalloc_vm(struct vm_struct *vm, struct vmap_area *va, |
5e6cafc83
|
1281 |
unsigned long flags, const void *caller) |
cf88c7900
|
1282 |
{ |
c69480ade
|
1283 |
spin_lock(&vmap_area_lock); |
cf88c7900
|
1284 1285 1286 1287 |
vm->flags = flags; vm->addr = (void *)va->va_start; vm->size = va->va_end - va->va_start; vm->caller = caller; |
db1aecafe
|
1288 |
va->vm = vm; |
cf88c7900
|
1289 |
va->flags |= VM_VM_AREA; |
c69480ade
|
1290 |
spin_unlock(&vmap_area_lock); |
f5252e009
|
1291 |
} |
cf88c7900
|
1292 |
|
20fc02b47
|
1293 |
static void clear_vm_uninitialized_flag(struct vm_struct *vm) |
f5252e009
|
1294 |
{ |
d4033afdf
|
1295 |
/* |
20fc02b47
|
1296 |
* Before removing VM_UNINITIALIZED, |
d4033afdf
|
1297 1298 1299 1300 |
* we should make sure that vm has proper values. * Pair with smp_rmb() in show_numa_info(). */ smp_wmb(); |
20fc02b47
|
1301 |
vm->flags &= ~VM_UNINITIALIZED; |
cf88c7900
|
1302 |
} |
db64fe022
|
1303 |
static struct vm_struct *__get_vm_area_node(unsigned long size, |
2dca6999e
|
1304 |
unsigned long align, unsigned long flags, unsigned long start, |
5e6cafc83
|
1305 |
unsigned long end, int node, gfp_t gfp_mask, const void *caller) |
db64fe022
|
1306 |
{ |
0006526d7
|
1307 |
struct vmap_area *va; |
db64fe022
|
1308 |
struct vm_struct *area; |
1da177e4c
|
1309 |
|
52fd24ca1
|
1310 |
BUG_ON(in_interrupt()); |
1da177e4c
|
1311 |
size = PAGE_ALIGN(size); |
31be83095
|
1312 1313 |
if (unlikely(!size)) return NULL; |
1da177e4c
|
1314 |
|
252e5c6e2
|
1315 1316 1317 |
if (flags & VM_IOREMAP) align = 1ul << clamp_t(int, get_count_order_long(size), PAGE_SHIFT, IOREMAP_MAX_ORDER); |
cf88c7900
|
1318 |
area = kzalloc_node(sizeof(*area), gfp_mask & GFP_RECLAIM_MASK, node); |
1da177e4c
|
1319 1320 |
if (unlikely(!area)) return NULL; |
71394fe50
|
1321 1322 |
if (!(flags & VM_NO_GUARD)) size += PAGE_SIZE; |
1da177e4c
|
1323 |
|
db64fe022
|
1324 1325 1326 1327 |
va = alloc_vmap_area(size, align, start, end, node, gfp_mask); if (IS_ERR(va)) { kfree(area); return NULL; |
1da177e4c
|
1328 |
} |
1da177e4c
|
1329 |
|
d82b1d857
|
1330 |
setup_vmalloc_vm(area, va, flags, caller); |
f5252e009
|
1331 |
|
1da177e4c
|
1332 |
return area; |
1da177e4c
|
1333 |
} |
930fc45a4
|
1334 1335 1336 |
struct vm_struct *__get_vm_area(unsigned long size, unsigned long flags, unsigned long start, unsigned long end) { |
00ef2d2f8
|
1337 1338 |
return __get_vm_area_node(size, 1, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); |
930fc45a4
|
1339 |
} |
5992b6dac
|
1340 |
EXPORT_SYMBOL_GPL(__get_vm_area); |
930fc45a4
|
1341 |
|
c29686129
|
1342 1343 |
struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags, unsigned long start, unsigned long end, |
5e6cafc83
|
1344 |
const void *caller) |
c29686129
|
1345 |
{ |
00ef2d2f8
|
1346 1347 |
return __get_vm_area_node(size, 1, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, caller); |
c29686129
|
1348 |
} |
1da177e4c
|
1349 |
/** |
183ff22bb
|
1350 |
* get_vm_area - reserve a contiguous kernel virtual area |
1da177e4c
|
1351 1352 1353 1354 1355 1356 1357 1358 1359 |
* @size: size of the area * @flags: %VM_IOREMAP for I/O mappings or VM_ALLOC * * Search an area of @size in the kernel virtual mapping area, * and reserved it for out purposes. Returns the area descriptor * on success or %NULL on failure. */ struct vm_struct *get_vm_area(unsigned long size, unsigned long flags) { |
2dca6999e
|
1360 |
return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END, |
00ef2d2f8
|
1361 1362 |
NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); |
230169693
|
1363 1364 1365 |
} struct vm_struct *get_vm_area_caller(unsigned long size, unsigned long flags, |
5e6cafc83
|
1366 |
const void *caller) |
230169693
|
1367 |
{ |
2dca6999e
|
1368 |
return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END, |
00ef2d2f8
|
1369 |
NUMA_NO_NODE, GFP_KERNEL, caller); |
1da177e4c
|
1370 |
} |
e9da6e990
|
1371 1372 1373 1374 1375 1376 1377 1378 1379 |
/** * find_vm_area - find a continuous kernel virtual area * @addr: base address * * Search for the kernel VM area starting at @addr, and return it. * It is up to the caller to do all required locking to keep the returned * pointer valid. */ struct vm_struct *find_vm_area(const void *addr) |
833423143
|
1380 |
{ |
db64fe022
|
1381 |
struct vmap_area *va; |
833423143
|
1382 |
|
db64fe022
|
1383 1384 |
va = find_vmap_area((unsigned long)addr); if (va && va->flags & VM_VM_AREA) |
db1aecafe
|
1385 |
return va->vm; |
1da177e4c
|
1386 |
|
1da177e4c
|
1387 |
return NULL; |
1da177e4c
|
1388 |
} |
7856dfeb2
|
1389 |
/** |
183ff22bb
|
1390 |
* remove_vm_area - find and remove a continuous kernel virtual area |
7856dfeb2
|
1391 1392 1393 1394 1395 1396 |
* @addr: base address * * Search for the kernel VM area starting at @addr, and remove it. * This function returns the found VM area, but using it is NOT safe * on SMP machines, except for its size or flags. */ |
b3bdda02a
|
1397 |
struct vm_struct *remove_vm_area(const void *addr) |
7856dfeb2
|
1398 |
{ |
db64fe022
|
1399 |
struct vmap_area *va; |
5803ed292
|
1400 |
might_sleep(); |
db64fe022
|
1401 1402 |
va = find_vmap_area((unsigned long)addr); if (va && va->flags & VM_VM_AREA) { |
db1aecafe
|
1403 |
struct vm_struct *vm = va->vm; |
f5252e009
|
1404 |
|
c69480ade
|
1405 1406 1407 |
spin_lock(&vmap_area_lock); va->vm = NULL; va->flags &= ~VM_VM_AREA; |
78c72746f
|
1408 |
va->flags |= VM_LAZY_FREE; |
c69480ade
|
1409 |
spin_unlock(&vmap_area_lock); |
a5af5aa8b
|
1410 |
kasan_free_shadow(vm); |
dd32c2799
|
1411 |
free_unmap_vmap_area(va); |
dd32c2799
|
1412 |
|
db64fe022
|
1413 1414 1415 |
return vm; } return NULL; |
7856dfeb2
|
1416 |
} |
b3bdda02a
|
1417 |
static void __vunmap(const void *addr, int deallocate_pages) |
1da177e4c
|
1418 1419 1420 1421 1422 |
{ struct vm_struct *area; if (!addr) return; |
e69e9d4ae
|
1423 1424 |
if (WARN(!PAGE_ALIGNED(addr), "Trying to vfree() bad address (%p) ", |
ab15d9b4c
|
1425 |
addr)) |
1da177e4c
|
1426 |
return; |
1da177e4c
|
1427 |
|
f3c01d2f3
|
1428 |
area = find_vmap_area((unsigned long)addr)->vm; |
1da177e4c
|
1429 |
if (unlikely(!area)) { |
4c8573e25
|
1430 1431 |
WARN(1, KERN_ERR "Trying to vfree() nonexistent vm area (%p) ", |
1da177e4c
|
1432 |
addr); |
1da177e4c
|
1433 1434 |
return; } |
05e3ff950
|
1435 1436 |
debug_check_no_locks_freed(area->addr, get_vm_area_size(area)); debug_check_no_obj_freed(area->addr, get_vm_area_size(area)); |
9a11b49a8
|
1437 |
|
f3c01d2f3
|
1438 |
remove_vm_area(addr); |
1da177e4c
|
1439 1440 1441 1442 |
if (deallocate_pages) { int i; for (i = 0; i < area->nr_pages; i++) { |
bf53d6f8f
|
1443 1444 1445 |
struct page *page = area->pages[i]; BUG_ON(!page); |
4949148ad
|
1446 |
__free_pages(page, 0); |
1da177e4c
|
1447 |
} |
244d63ee3
|
1448 |
kvfree(area->pages); |
1da177e4c
|
1449 1450 1451 1452 1453 |
} kfree(area); return; } |
bf22e37a6
|
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 |
static inline void __vfree_deferred(const void *addr) { /* * Use raw_cpu_ptr() because this can be called from preemptible * context. Preemption is absolutely fine here, because the llist_add() * implementation is lockless, so it works even if we are adding to * nother cpu's list. schedule_work() should be fine with this too. */ struct vfree_deferred *p = raw_cpu_ptr(&vfree_deferred); if (llist_add((struct llist_node *)addr, &p->list)) schedule_work(&p->wq); } /** * vfree_atomic - release memory allocated by vmalloc() * @addr: memory base address * * This one is just like vfree() but can be called in any atomic context * except NMIs. */ void vfree_atomic(const void *addr) { BUG_ON(in_nmi()); kmemleak_free(addr); if (!addr) return; __vfree_deferred(addr); } |
1da177e4c
|
1486 1487 |
/** * vfree - release memory allocated by vmalloc() |
1da177e4c
|
1488 1489 |
* @addr: memory base address * |
183ff22bb
|
1490 |
* Free the virtually continuous memory area starting at @addr, as |
80e93effc
|
1491 1492 |
* obtained from vmalloc(), vmalloc_32() or __vmalloc(). If @addr is * NULL, no operation is performed. |
1da177e4c
|
1493 |
* |
32fcfd407
|
1494 1495 1496 |
* Must not be called in NMI context (strictly speaking, only if we don't * have CONFIG_ARCH_HAVE_NMI_SAFE_CMPXCHG, but making the calling * conventions for vfree() arch-depenedent would be a really bad idea) |
c9fcee513
|
1497 |
* |
0e056eb55
|
1498 |
* NOTE: assumes that the object at @addr has a size >= sizeof(llist_node) |
1da177e4c
|
1499 |
*/ |
b3bdda02a
|
1500 |
void vfree(const void *addr) |
1da177e4c
|
1501 |
{ |
32fcfd407
|
1502 |
BUG_ON(in_nmi()); |
89219d37a
|
1503 1504 |
kmemleak_free(addr); |
32fcfd407
|
1505 1506 |
if (!addr) return; |
bf22e37a6
|
1507 1508 1509 |
if (unlikely(in_interrupt())) __vfree_deferred(addr); else |
32fcfd407
|
1510 |
__vunmap(addr, 1); |
1da177e4c
|
1511 |
} |
1da177e4c
|
1512 1513 1514 1515 |
EXPORT_SYMBOL(vfree); /** * vunmap - release virtual mapping obtained by vmap() |
1da177e4c
|
1516 1517 1518 1519 1520 |
* @addr: memory base address * * Free the virtually contiguous memory area starting at @addr, * which was created from the page array passed to vmap(). * |
80e93effc
|
1521 |
* Must not be called in interrupt context. |
1da177e4c
|
1522 |
*/ |
b3bdda02a
|
1523 |
void vunmap(const void *addr) |
1da177e4c
|
1524 1525 |
{ BUG_ON(in_interrupt()); |
34754b69a
|
1526 |
might_sleep(); |
32fcfd407
|
1527 1528 |
if (addr) __vunmap(addr, 0); |
1da177e4c
|
1529 |
} |
1da177e4c
|
1530 1531 1532 1533 |
EXPORT_SYMBOL(vunmap); /** * vmap - map an array of pages into virtually contiguous space |
1da177e4c
|
1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 |
* @pages: array of page pointers * @count: number of pages to map * @flags: vm_area->flags * @prot: page protection for the mapping * * Maps @count pages from @pages into contiguous kernel virtual * space. */ void *vmap(struct page **pages, unsigned int count, unsigned long flags, pgprot_t prot) { struct vm_struct *area; |
65ee03c4b
|
1546 |
unsigned long size; /* In bytes */ |
1da177e4c
|
1547 |
|
34754b69a
|
1548 |
might_sleep(); |
4481374ce
|
1549 |
if (count > totalram_pages) |
1da177e4c
|
1550 |
return NULL; |
65ee03c4b
|
1551 1552 |
size = (unsigned long)count << PAGE_SHIFT; area = get_vm_area_caller(size, flags, __builtin_return_address(0)); |
1da177e4c
|
1553 1554 |
if (!area) return NULL; |
230169693
|
1555 |
|
f6f8ed473
|
1556 |
if (map_vm_area(area, prot, pages)) { |
1da177e4c
|
1557 1558 1559 1560 1561 1562 |
vunmap(area->addr); return NULL; } return area->addr; } |
1da177e4c
|
1563 |
EXPORT_SYMBOL(vmap); |
8594a21cf
|
1564 1565 1566 |
static void *__vmalloc_node(unsigned long size, unsigned long align, gfp_t gfp_mask, pgprot_t prot, int node, const void *caller); |
e31d9eb5c
|
1567 |
static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, |
3722e13cf
|
1568 |
pgprot_t prot, int node) |
1da177e4c
|
1569 1570 1571 |
{ struct page **pages; unsigned int nr_pages, array_size, i; |
930f036b4
|
1572 |
const gfp_t nested_gfp = (gfp_mask & GFP_RECLAIM_MASK) | __GFP_ZERO; |
704b862f9
|
1573 1574 1575 1576 |
const gfp_t alloc_mask = gfp_mask | __GFP_NOWARN; const gfp_t highmem_mask = (gfp_mask & (GFP_DMA | GFP_DMA32)) ? 0 : __GFP_HIGHMEM; |
1da177e4c
|
1577 |
|
762216ab4
|
1578 |
nr_pages = get_vm_area_size(area) >> PAGE_SHIFT; |
1da177e4c
|
1579 1580 1581 1582 |
array_size = (nr_pages * sizeof(struct page *)); area->nr_pages = nr_pages; /* Please note that the recursion is strictly bounded. */ |
8757d5fa6
|
1583 |
if (array_size > PAGE_SIZE) { |
704b862f9
|
1584 |
pages = __vmalloc_node(array_size, 1, nested_gfp|highmem_mask, |
3722e13cf
|
1585 |
PAGE_KERNEL, node, area->caller); |
286e1ea3a
|
1586 |
} else { |
976d6dfbb
|
1587 |
pages = kmalloc_node(array_size, nested_gfp, node); |
286e1ea3a
|
1588 |
} |
1da177e4c
|
1589 1590 1591 1592 1593 1594 |
area->pages = pages; if (!area->pages) { remove_vm_area(area->addr); kfree(area); return NULL; } |
1da177e4c
|
1595 1596 |
for (i = 0; i < area->nr_pages; i++) { |
bf53d6f8f
|
1597 |
struct page *page; |
4b90951c0
|
1598 |
if (node == NUMA_NO_NODE) |
704b862f9
|
1599 |
page = alloc_page(alloc_mask|highmem_mask); |
930fc45a4
|
1600 |
else |
704b862f9
|
1601 |
page = alloc_pages_node(node, alloc_mask|highmem_mask, 0); |
bf53d6f8f
|
1602 1603 |
if (unlikely(!page)) { |
1da177e4c
|
1604 1605 1606 1607 |
/* Successfully allocated i pages, free them in __vunmap() */ area->nr_pages = i; goto fail; } |
bf53d6f8f
|
1608 |
area->pages[i] = page; |
704b862f9
|
1609 |
if (gfpflags_allow_blocking(gfp_mask|highmem_mask)) |
660654f90
|
1610 |
cond_resched(); |
1da177e4c
|
1611 |
} |
f6f8ed473
|
1612 |
if (map_vm_area(area, prot, pages)) |
1da177e4c
|
1613 1614 1615 1616 |
goto fail; return area->addr; fail: |
a8e99259e
|
1617 |
warn_alloc(gfp_mask, NULL, |
7877cdcc3
|
1618 |
"vmalloc: allocation failure, allocated %ld of %ld bytes", |
22943ab11
|
1619 |
(area->nr_pages*PAGE_SIZE), area->size); |
1da177e4c
|
1620 1621 1622 1623 1624 |
vfree(area->addr); return NULL; } /** |
d0a21265d
|
1625 |
* __vmalloc_node_range - allocate virtually contiguous memory |
1da177e4c
|
1626 |
* @size: allocation size |
2dca6999e
|
1627 |
* @align: desired alignment |
d0a21265d
|
1628 1629 |
* @start: vm area range start * @end: vm area range end |
1da177e4c
|
1630 1631 |
* @gfp_mask: flags for the page level allocator * @prot: protection mask for the allocated pages |
cb9e3c292
|
1632 |
* @vm_flags: additional vm area flags (e.g. %VM_NO_GUARD) |
00ef2d2f8
|
1633 |
* @node: node to use for allocation or NUMA_NO_NODE |
c85d194bf
|
1634 |
* @caller: caller's return address |
1da177e4c
|
1635 1636 1637 1638 1639 |
* * Allocate enough pages to cover @size from the page level * allocator with @gfp_mask flags. Map them into contiguous * kernel virtual space, using a pagetable protection of @prot. */ |
d0a21265d
|
1640 1641 |
void *__vmalloc_node_range(unsigned long size, unsigned long align, unsigned long start, unsigned long end, gfp_t gfp_mask, |
cb9e3c292
|
1642 1643 |
pgprot_t prot, unsigned long vm_flags, int node, const void *caller) |
1da177e4c
|
1644 1645 |
{ struct vm_struct *area; |
89219d37a
|
1646 1647 |
void *addr; unsigned long real_size = size; |
1da177e4c
|
1648 1649 |
size = PAGE_ALIGN(size); |
4481374ce
|
1650 |
if (!size || (size >> PAGE_SHIFT) > totalram_pages) |
de7d2b567
|
1651 |
goto fail; |
1da177e4c
|
1652 |
|
cb9e3c292
|
1653 1654 |
area = __get_vm_area_node(size, align, VM_ALLOC | VM_UNINITIALIZED | vm_flags, start, end, node, gfp_mask, caller); |
1da177e4c
|
1655 |
if (!area) |
de7d2b567
|
1656 |
goto fail; |
1da177e4c
|
1657 |
|
3722e13cf
|
1658 |
addr = __vmalloc_area_node(area, gfp_mask, prot, node); |
1368edf06
|
1659 |
if (!addr) |
b82225f3f
|
1660 |
return NULL; |
89219d37a
|
1661 1662 |
/* |
20fc02b47
|
1663 1664 |
* In this function, newly allocated vm_struct has VM_UNINITIALIZED * flag. It means that vm_struct is not fully initialized. |
4341fa454
|
1665 |
* Now, it is fully initialized, so remove this flag here. |
f5252e009
|
1666 |
*/ |
20fc02b47
|
1667 |
clear_vm_uninitialized_flag(area); |
f5252e009
|
1668 |
|
94f4a1618
|
1669 |
kmemleak_vmalloc(area, size, gfp_mask); |
89219d37a
|
1670 1671 |
return addr; |
de7d2b567
|
1672 1673 |
fail: |
a8e99259e
|
1674 |
warn_alloc(gfp_mask, NULL, |
7877cdcc3
|
1675 |
"vmalloc: allocation failure: %lu bytes", real_size); |
de7d2b567
|
1676 |
return NULL; |
1da177e4c
|
1677 |
} |
d0a21265d
|
1678 1679 1680 1681 1682 1683 |
/** * __vmalloc_node - allocate virtually contiguous memory * @size: allocation size * @align: desired alignment * @gfp_mask: flags for the page level allocator * @prot: protection mask for the allocated pages |
00ef2d2f8
|
1684 |
* @node: node to use for allocation or NUMA_NO_NODE |
d0a21265d
|
1685 1686 1687 1688 1689 |
* @caller: caller's return address * * Allocate enough pages to cover @size from the page level * allocator with @gfp_mask flags. Map them into contiguous * kernel virtual space, using a pagetable protection of @prot. |
a7c3e901a
|
1690 |
* |
dcda9b047
|
1691 |
* Reclaim modifiers in @gfp_mask - __GFP_NORETRY, __GFP_RETRY_MAYFAIL |
a7c3e901a
|
1692 1693 1694 1695 1696 |
* and __GFP_NOFAIL are not supported * * Any use of gfp flags outside of GFP_KERNEL should be consulted * with mm people. * |
d0a21265d
|
1697 |
*/ |
8594a21cf
|
1698 |
static void *__vmalloc_node(unsigned long size, unsigned long align, |
d0a21265d
|
1699 |
gfp_t gfp_mask, pgprot_t prot, |
5e6cafc83
|
1700 |
int node, const void *caller) |
d0a21265d
|
1701 1702 |
{ return __vmalloc_node_range(size, align, VMALLOC_START, VMALLOC_END, |
cb9e3c292
|
1703 |
gfp_mask, prot, 0, node, caller); |
d0a21265d
|
1704 |
} |
930fc45a4
|
1705 1706 |
void *__vmalloc(unsigned long size, gfp_t gfp_mask, pgprot_t prot) { |
00ef2d2f8
|
1707 |
return __vmalloc_node(size, 1, gfp_mask, prot, NUMA_NO_NODE, |
230169693
|
1708 |
__builtin_return_address(0)); |
930fc45a4
|
1709 |
} |
1da177e4c
|
1710 |
EXPORT_SYMBOL(__vmalloc); |
8594a21cf
|
1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 |
static inline void *__vmalloc_node_flags(unsigned long size, int node, gfp_t flags) { return __vmalloc_node(size, 1, flags, PAGE_KERNEL, node, __builtin_return_address(0)); } void *__vmalloc_node_flags_caller(unsigned long size, int node, gfp_t flags, void *caller) { return __vmalloc_node(size, 1, flags, PAGE_KERNEL, node, caller); } |
1da177e4c
|
1724 1725 |
/** * vmalloc - allocate virtually contiguous memory |
1da177e4c
|
1726 |
* @size: allocation size |
1da177e4c
|
1727 1728 1729 |
* Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. * |
c1c8897f8
|
1730 |
* For tight control over page level allocator and protection flags |
1da177e4c
|
1731 1732 1733 1734 |
* use __vmalloc() instead. */ void *vmalloc(unsigned long size) { |
00ef2d2f8
|
1735 |
return __vmalloc_node_flags(size, NUMA_NO_NODE, |
19809c2da
|
1736 |
GFP_KERNEL); |
1da177e4c
|
1737 |
} |
1da177e4c
|
1738 |
EXPORT_SYMBOL(vmalloc); |
930fc45a4
|
1739 |
/** |
e1ca7788d
|
1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 |
* vzalloc - allocate virtually contiguous memory with zero fill * @size: allocation size * Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. * The memory allocated is set to zero. * * For tight control over page level allocator and protection flags * use __vmalloc() instead. */ void *vzalloc(unsigned long size) { |
00ef2d2f8
|
1751 |
return __vmalloc_node_flags(size, NUMA_NO_NODE, |
19809c2da
|
1752 |
GFP_KERNEL | __GFP_ZERO); |
e1ca7788d
|
1753 1754 1755 1756 |
} EXPORT_SYMBOL(vzalloc); /** |
ead04089b
|
1757 1758 |
* vmalloc_user - allocate zeroed virtually contiguous memory for userspace * @size: allocation size |
833423143
|
1759 |
* |
ead04089b
|
1760 1761 |
* The resulting memory area is zeroed so it can be mapped to userspace * without leaking data. |
833423143
|
1762 1763 1764 1765 1766 |
*/ void *vmalloc_user(unsigned long size) { struct vm_struct *area; void *ret; |
2dca6999e
|
1767 |
ret = __vmalloc_node(size, SHMLBA, |
19809c2da
|
1768 |
GFP_KERNEL | __GFP_ZERO, |
00ef2d2f8
|
1769 1770 |
PAGE_KERNEL, NUMA_NO_NODE, __builtin_return_address(0)); |
2b4ac44e7
|
1771 |
if (ret) { |
db64fe022
|
1772 |
area = find_vm_area(ret); |
2b4ac44e7
|
1773 |
area->flags |= VM_USERMAP; |
2b4ac44e7
|
1774 |
} |
833423143
|
1775 1776 1777 1778 1779 |
return ret; } EXPORT_SYMBOL(vmalloc_user); /** |
930fc45a4
|
1780 |
* vmalloc_node - allocate memory on a specific node |
930fc45a4
|
1781 |
* @size: allocation size |
d44e0780b
|
1782 |
* @node: numa node |
930fc45a4
|
1783 1784 1785 1786 |
* * Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. * |
c1c8897f8
|
1787 |
* For tight control over page level allocator and protection flags |
930fc45a4
|
1788 1789 1790 1791 |
* use __vmalloc() instead. */ void *vmalloc_node(unsigned long size, int node) { |
19809c2da
|
1792 |
return __vmalloc_node(size, 1, GFP_KERNEL, PAGE_KERNEL, |
230169693
|
1793 |
node, __builtin_return_address(0)); |
930fc45a4
|
1794 1795 |
} EXPORT_SYMBOL(vmalloc_node); |
e1ca7788d
|
1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 |
/** * vzalloc_node - allocate memory on a specific node with zero fill * @size: allocation size * @node: numa node * * Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. * The memory allocated is set to zero. * * For tight control over page level allocator and protection flags * use __vmalloc_node() instead. */ void *vzalloc_node(unsigned long size, int node) { return __vmalloc_node_flags(size, node, |
19809c2da
|
1811 |
GFP_KERNEL | __GFP_ZERO); |
e1ca7788d
|
1812 1813 |
} EXPORT_SYMBOL(vzalloc_node); |
4dc3b16ba
|
1814 1815 1816 |
#ifndef PAGE_KERNEL_EXEC # define PAGE_KERNEL_EXEC PAGE_KERNEL #endif |
1da177e4c
|
1817 1818 |
/** * vmalloc_exec - allocate virtually contiguous, executable memory |
1da177e4c
|
1819 1820 1821 1822 1823 1824 |
* @size: allocation size * * Kernel-internal function to allocate enough pages to cover @size * the page level allocator and map them into contiguous and * executable kernel virtual space. * |
c1c8897f8
|
1825 |
* For tight control over page level allocator and protection flags |
1da177e4c
|
1826 1827 |
* use __vmalloc() instead. */ |
1da177e4c
|
1828 1829 |
void *vmalloc_exec(unsigned long size) { |
19809c2da
|
1830 |
return __vmalloc_node(size, 1, GFP_KERNEL, PAGE_KERNEL_EXEC, |
00ef2d2f8
|
1831 |
NUMA_NO_NODE, __builtin_return_address(0)); |
1da177e4c
|
1832 |
} |
0d08e0d3a
|
1833 |
#if defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA32) |
698d0831b
|
1834 |
#define GFP_VMALLOC32 (GFP_DMA32 | GFP_KERNEL) |
0d08e0d3a
|
1835 |
#elif defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA) |
698d0831b
|
1836 |
#define GFP_VMALLOC32 (GFP_DMA | GFP_KERNEL) |
0d08e0d3a
|
1837 |
#else |
698d0831b
|
1838 1839 1840 1841 1842 |
/* * 64b systems should always have either DMA or DMA32 zones. For others * GFP_DMA32 should do the right thing and use the normal zone. */ #define GFP_VMALLOC32 GFP_DMA32 | GFP_KERNEL |
0d08e0d3a
|
1843 |
#endif |
1da177e4c
|
1844 1845 |
/** * vmalloc_32 - allocate virtually contiguous memory (32bit addressable) |
1da177e4c
|
1846 1847 1848 1849 1850 1851 1852 |
* @size: allocation size * * Allocate enough 32bit PA addressable pages to cover @size from the * page level allocator and map them into contiguous kernel virtual space. */ void *vmalloc_32(unsigned long size) { |
2dca6999e
|
1853 |
return __vmalloc_node(size, 1, GFP_VMALLOC32, PAGE_KERNEL, |
00ef2d2f8
|
1854 |
NUMA_NO_NODE, __builtin_return_address(0)); |
1da177e4c
|
1855 |
} |
1da177e4c
|
1856 |
EXPORT_SYMBOL(vmalloc_32); |
833423143
|
1857 |
/** |
ead04089b
|
1858 |
* vmalloc_32_user - allocate zeroed virtually contiguous 32bit memory |
833423143
|
1859 |
* @size: allocation size |
ead04089b
|
1860 1861 1862 |
* * The resulting memory area is 32bit addressable and zeroed so it can be * mapped to userspace without leaking data. |
833423143
|
1863 1864 1865 1866 1867 |
*/ void *vmalloc_32_user(unsigned long size) { struct vm_struct *area; void *ret; |
2dca6999e
|
1868 |
ret = __vmalloc_node(size, 1, GFP_VMALLOC32 | __GFP_ZERO, PAGE_KERNEL, |
00ef2d2f8
|
1869 |
NUMA_NO_NODE, __builtin_return_address(0)); |
2b4ac44e7
|
1870 |
if (ret) { |
db64fe022
|
1871 |
area = find_vm_area(ret); |
2b4ac44e7
|
1872 |
area->flags |= VM_USERMAP; |
2b4ac44e7
|
1873 |
} |
833423143
|
1874 1875 1876 |
return ret; } EXPORT_SYMBOL(vmalloc_32_user); |
d0107eb07
|
1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 |
/* * small helper routine , copy contents to buf from addr. * If the page is not present, fill zero. */ static int aligned_vread(char *buf, char *addr, unsigned long count) { struct page *p; int copied = 0; while (count) { unsigned long offset, length; |
891c49abf
|
1889 |
offset = offset_in_page(addr); |
d0107eb07
|
1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 |
length = PAGE_SIZE - offset; if (length > count) length = count; p = vmalloc_to_page(addr); /* * To do safe access to this _mapped_ area, we need * lock. But adding lock here means that we need to add * overhead of vmalloc()/vfree() calles for this _debug_ * interface, rarely used. Instead of that, we'll use * kmap() and get small overhead in this access function. */ if (p) { /* * we can expect USER0 is not used (see vread/vwrite's * function description) */ |
9b04c5fec
|
1906 |
void *map = kmap_atomic(p); |
d0107eb07
|
1907 |
memcpy(buf, map + offset, length); |
9b04c5fec
|
1908 |
kunmap_atomic(map); |
d0107eb07
|
1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 |
} else memset(buf, 0, length); addr += length; buf += length; copied += length; count -= length; } return copied; } static int aligned_vwrite(char *buf, char *addr, unsigned long count) { struct page *p; int copied = 0; while (count) { unsigned long offset, length; |
891c49abf
|
1927 |
offset = offset_in_page(addr); |
d0107eb07
|
1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 |
length = PAGE_SIZE - offset; if (length > count) length = count; p = vmalloc_to_page(addr); /* * To do safe access to this _mapped_ area, we need * lock. But adding lock here means that we need to add * overhead of vmalloc()/vfree() calles for this _debug_ * interface, rarely used. Instead of that, we'll use * kmap() and get small overhead in this access function. */ if (p) { /* * we can expect USER0 is not used (see vread/vwrite's * function description) */ |
9b04c5fec
|
1944 |
void *map = kmap_atomic(p); |
d0107eb07
|
1945 |
memcpy(map + offset, buf, length); |
9b04c5fec
|
1946 |
kunmap_atomic(map); |
d0107eb07
|
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 |
} addr += length; buf += length; copied += length; count -= length; } return copied; } /** * vread() - read vmalloc area in a safe way. * @buf: buffer for reading data * @addr: vm address. * @count: number of bytes to be read. * * Returns # of bytes which addr and buf should be increased. * (same number to @count). Returns 0 if [addr...addr+count) doesn't * includes any intersect with alive vmalloc area. * * This function checks that addr is a valid vmalloc'ed area, and * copy data from that area to a given buffer. If the given memory range * of [addr...addr+count) includes some valid address, data is copied to * proper area of @buf. If there are memory holes, they'll be zero-filled. * IOREMAP area is treated as memory hole and no copy is done. * * If [addr...addr+count) doesn't includes any intersects with alive |
a8e5202d0
|
1973 |
* vm_struct area, returns 0. @buf should be kernel's buffer. |
d0107eb07
|
1974 1975 1976 1977 1978 1979 1980 |
* * Note: In usual ops, vread() is never necessary because the caller * should know vmalloc() area is valid and can use memcpy(). * This is for routines which have to access vmalloc area without * any informaion, as /dev/kmem. * */ |
1da177e4c
|
1981 1982 |
long vread(char *buf, char *addr, unsigned long count) { |
e81ce85f9
|
1983 1984 |
struct vmap_area *va; struct vm_struct *vm; |
1da177e4c
|
1985 |
char *vaddr, *buf_start = buf; |
d0107eb07
|
1986 |
unsigned long buflen = count; |
1da177e4c
|
1987 1988 1989 1990 1991 |
unsigned long n; /* Don't allow overflow */ if ((unsigned long) addr + count < count) count = -(unsigned long) addr; |
e81ce85f9
|
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 |
spin_lock(&vmap_area_lock); list_for_each_entry(va, &vmap_area_list, list) { if (!count) break; if (!(va->flags & VM_VM_AREA)) continue; vm = va->vm; vaddr = (char *) vm->addr; |
762216ab4
|
2002 |
if (addr >= vaddr + get_vm_area_size(vm)) |
1da177e4c
|
2003 2004 2005 2006 2007 2008 2009 2010 2011 |
continue; while (addr < vaddr) { if (count == 0) goto finished; *buf = '\0'; buf++; addr++; count--; } |
762216ab4
|
2012 |
n = vaddr + get_vm_area_size(vm) - addr; |
d0107eb07
|
2013 2014 |
if (n > count) n = count; |
e81ce85f9
|
2015 |
if (!(vm->flags & VM_IOREMAP)) |
d0107eb07
|
2016 2017 2018 2019 2020 2021 |
aligned_vread(buf, addr, n); else /* IOREMAP area is treated as memory hole */ memset(buf, 0, n); buf += n; addr += n; count -= n; |
1da177e4c
|
2022 2023 |
} finished: |
e81ce85f9
|
2024 |
spin_unlock(&vmap_area_lock); |
d0107eb07
|
2025 2026 2027 2028 2029 2030 2031 2032 |
if (buf == buf_start) return 0; /* zero-fill memory holes */ if (buf != buf_start + buflen) memset(buf, 0, buflen - (buf - buf_start)); return buflen; |
1da177e4c
|
2033 |
} |
d0107eb07
|
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 |
/** * vwrite() - write vmalloc area in a safe way. * @buf: buffer for source data * @addr: vm address. * @count: number of bytes to be read. * * Returns # of bytes which addr and buf should be incresed. * (same number to @count). * If [addr...addr+count) doesn't includes any intersect with valid * vmalloc area, returns 0. * * This function checks that addr is a valid vmalloc'ed area, and * copy data from a buffer to the given addr. If specified range of * [addr...addr+count) includes some valid address, data is copied from * proper area of @buf. If there are memory holes, no copy to hole. * IOREMAP area is treated as memory hole and no copy is done. * * If [addr...addr+count) doesn't includes any intersects with alive |
a8e5202d0
|
2052 |
* vm_struct area, returns 0. @buf should be kernel's buffer. |
d0107eb07
|
2053 2054 2055 2056 2057 |
* * Note: In usual ops, vwrite() is never necessary because the caller * should know vmalloc() area is valid and can use memcpy(). * This is for routines which have to access vmalloc area without * any informaion, as /dev/kmem. |
d0107eb07
|
2058 |
*/ |
1da177e4c
|
2059 2060 |
long vwrite(char *buf, char *addr, unsigned long count) { |
e81ce85f9
|
2061 2062 |
struct vmap_area *va; struct vm_struct *vm; |
d0107eb07
|
2063 2064 2065 |
char *vaddr; unsigned long n, buflen; int copied = 0; |
1da177e4c
|
2066 2067 2068 2069 |
/* Don't allow overflow */ if ((unsigned long) addr + count < count) count = -(unsigned long) addr; |
d0107eb07
|
2070 |
buflen = count; |
1da177e4c
|
2071 |
|
e81ce85f9
|
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 |
spin_lock(&vmap_area_lock); list_for_each_entry(va, &vmap_area_list, list) { if (!count) break; if (!(va->flags & VM_VM_AREA)) continue; vm = va->vm; vaddr = (char *) vm->addr; |
762216ab4
|
2082 |
if (addr >= vaddr + get_vm_area_size(vm)) |
1da177e4c
|
2083 2084 2085 2086 2087 2088 2089 2090 |
continue; while (addr < vaddr) { if (count == 0) goto finished; buf++; addr++; count--; } |
762216ab4
|
2091 |
n = vaddr + get_vm_area_size(vm) - addr; |
d0107eb07
|
2092 2093 |
if (n > count) n = count; |
e81ce85f9
|
2094 |
if (!(vm->flags & VM_IOREMAP)) { |
d0107eb07
|
2095 2096 2097 2098 2099 2100 |
aligned_vwrite(buf, addr, n); copied++; } buf += n; addr += n; count -= n; |
1da177e4c
|
2101 2102 |
} finished: |
e81ce85f9
|
2103 |
spin_unlock(&vmap_area_lock); |
d0107eb07
|
2104 2105 2106 |
if (!copied) return 0; return buflen; |
1da177e4c
|
2107 |
} |
833423143
|
2108 2109 |
/** |
e69e9d4ae
|
2110 2111 2112 2113 2114 |
* remap_vmalloc_range_partial - map vmalloc pages to userspace * @vma: vma to cover * @uaddr: target user address to start at * @kaddr: virtual address of vmalloc kernel memory * @size: size of map area |
7682486b3
|
2115 2116 |
* * Returns: 0 for success, -Exxx on failure |
833423143
|
2117 |
* |
e69e9d4ae
|
2118 2119 2120 2121 |
* This function checks that @kaddr is a valid vmalloc'ed area, * and that it is big enough to cover the range starting at * @uaddr in @vma. Will return failure if that criteria isn't * met. |
833423143
|
2122 |
* |
72fd4a35a
|
2123 |
* Similar to remap_pfn_range() (see mm/memory.c) |
833423143
|
2124 |
*/ |
e69e9d4ae
|
2125 2126 |
int remap_vmalloc_range_partial(struct vm_area_struct *vma, unsigned long uaddr, void *kaddr, unsigned long size) |
833423143
|
2127 2128 |
{ struct vm_struct *area; |
833423143
|
2129 |
|
e69e9d4ae
|
2130 2131 2132 |
size = PAGE_ALIGN(size); if (!PAGE_ALIGNED(uaddr) || !PAGE_ALIGNED(kaddr)) |
833423143
|
2133 |
return -EINVAL; |
e69e9d4ae
|
2134 |
area = find_vm_area(kaddr); |
833423143
|
2135 |
if (!area) |
db64fe022
|
2136 |
return -EINVAL; |
833423143
|
2137 2138 |
if (!(area->flags & VM_USERMAP)) |
db64fe022
|
2139 |
return -EINVAL; |
833423143
|
2140 |
|
e69e9d4ae
|
2141 |
if (kaddr + size > area->addr + area->size) |
db64fe022
|
2142 |
return -EINVAL; |
833423143
|
2143 |
|
833423143
|
2144 |
do { |
e69e9d4ae
|
2145 |
struct page *page = vmalloc_to_page(kaddr); |
db64fe022
|
2146 |
int ret; |
833423143
|
2147 2148 2149 2150 2151 |
ret = vm_insert_page(vma, uaddr, page); if (ret) return ret; uaddr += PAGE_SIZE; |
e69e9d4ae
|
2152 2153 2154 |
kaddr += PAGE_SIZE; size -= PAGE_SIZE; } while (size > 0); |
833423143
|
2155 |
|
314e51b98
|
2156 |
vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP; |
833423143
|
2157 |
|
db64fe022
|
2158 |
return 0; |
833423143
|
2159 |
} |
e69e9d4ae
|
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 |
EXPORT_SYMBOL(remap_vmalloc_range_partial); /** * remap_vmalloc_range - map vmalloc pages to userspace * @vma: vma to cover (map full range of vma) * @addr: vmalloc memory * @pgoff: number of pages into addr before first page to map * * Returns: 0 for success, -Exxx on failure * * This function checks that addr is a valid vmalloc'ed area, and * that it is big enough to cover the vma. Will return failure if * that criteria isn't met. * * Similar to remap_pfn_range() (see mm/memory.c) */ int remap_vmalloc_range(struct vm_area_struct *vma, void *addr, unsigned long pgoff) { return remap_vmalloc_range_partial(vma, vma->vm_start, addr + (pgoff << PAGE_SHIFT), vma->vm_end - vma->vm_start); } |
833423143
|
2183 |
EXPORT_SYMBOL(remap_vmalloc_range); |
1eeb66a1b
|
2184 2185 2186 2187 |
/* * Implement a stub for vmalloc_sync_all() if the architecture chose not to * have one. */ |
3b32123d7
|
2188 |
void __weak vmalloc_sync_all(void) |
1eeb66a1b
|
2189 2190 |
{ } |
5f4352fbf
|
2191 |
|
2f569afd9
|
2192 |
static int f(pte_t *pte, pgtable_t table, unsigned long addr, void *data) |
5f4352fbf
|
2193 |
{ |
cd12909cb
|
2194 2195 2196 2197 2198 2199 |
pte_t ***p = data; if (p) { *(*p) = pte; (*p)++; } |
5f4352fbf
|
2200 2201 2202 2203 2204 2205 |
return 0; } /** * alloc_vm_area - allocate a range of kernel address space * @size: size of the area |
cd12909cb
|
2206 |
* @ptes: returns the PTEs for the address space |
7682486b3
|
2207 2208 |
* * Returns: NULL on failure, vm_struct on success |
5f4352fbf
|
2209 2210 2211 |
* * This function reserves a range of kernel address space, and * allocates pagetables to map that range. No actual mappings |
cd12909cb
|
2212 2213 2214 2215 |
* are created. * * If @ptes is non-NULL, pointers to the PTEs (in init_mm) * allocated for the VM area are returned. |
5f4352fbf
|
2216 |
*/ |
cd12909cb
|
2217 |
struct vm_struct *alloc_vm_area(size_t size, pte_t **ptes) |
5f4352fbf
|
2218 2219 |
{ struct vm_struct *area; |
230169693
|
2220 2221 |
area = get_vm_area_caller(size, VM_IOREMAP, __builtin_return_address(0)); |
5f4352fbf
|
2222 2223 2224 2225 2226 2227 2228 2229 |
if (area == NULL) return NULL; /* * This ensures that page tables are constructed for this region * of kernel virtual address space and mapped into init_mm. */ if (apply_to_page_range(&init_mm, (unsigned long)area->addr, |
cd12909cb
|
2230 |
size, f, ptes ? &ptes : NULL)) { |
5f4352fbf
|
2231 2232 2233 |
free_vm_area(area); return NULL; } |
5f4352fbf
|
2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 |
return area; } EXPORT_SYMBOL_GPL(alloc_vm_area); void free_vm_area(struct vm_struct *area) { struct vm_struct *ret; ret = remove_vm_area(area->addr); BUG_ON(ret != area); kfree(area); } EXPORT_SYMBOL_GPL(free_vm_area); |
a10aa5798
|
2246 |
|
4f8b02b4e
|
2247 |
#ifdef CONFIG_SMP |
ca23e405e
|
2248 2249 |
static struct vmap_area *node_to_va(struct rb_node *n) { |
4583e7731
|
2250 |
return rb_entry_safe(n, struct vmap_area, rb_node); |
ca23e405e
|
2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 |
} /** * pvm_find_next_prev - find the next and prev vmap_area surrounding @end * @end: target address * @pnext: out arg for the next vmap_area * @pprev: out arg for the previous vmap_area * * Returns: %true if either or both of next and prev are found, * %false if no vmap_area exists * * Find vmap_areas end addresses of which enclose @end. ie. if not * NULL, *pnext->va_end > @end and *pprev->va_end <= @end. */ static bool pvm_find_next_prev(unsigned long end, struct vmap_area **pnext, struct vmap_area **pprev) { struct rb_node *n = vmap_area_root.rb_node; struct vmap_area *va = NULL; while (n) { va = rb_entry(n, struct vmap_area, rb_node); if (end < va->va_end) n = n->rb_left; else if (end > va->va_end) n = n->rb_right; else break; } if (!va) return false; if (va->va_end > end) { *pnext = va; *pprev = node_to_va(rb_prev(&(*pnext)->rb_node)); } else { *pprev = va; *pnext = node_to_va(rb_next(&(*pprev)->rb_node)); } return true; } /** * pvm_determine_end - find the highest aligned address between two vmap_areas * @pnext: in/out arg for the next vmap_area * @pprev: in/out arg for the previous vmap_area * @align: alignment * * Returns: determined end address * * Find the highest aligned address between *@pnext and *@pprev below * VMALLOC_END. *@pnext and *@pprev are adjusted so that the aligned * down address is between the end addresses of the two vmap_areas. * * Please note that the address returned by this function may fall * inside *@pnext vmap_area. The caller is responsible for checking * that. */ static unsigned long pvm_determine_end(struct vmap_area **pnext, struct vmap_area **pprev, unsigned long align) { const unsigned long vmalloc_end = VMALLOC_END & ~(align - 1); unsigned long addr; if (*pnext) addr = min((*pnext)->va_start & ~(align - 1), vmalloc_end); else addr = vmalloc_end; while (*pprev && (*pprev)->va_end > addr) { *pnext = *pprev; *pprev = node_to_va(rb_prev(&(*pnext)->rb_node)); } return addr; } /** * pcpu_get_vm_areas - allocate vmalloc areas for percpu allocator * @offsets: array containing offset of each area * @sizes: array containing size of each area * @nr_vms: the number of areas to allocate * @align: alignment, all entries in @offsets and @sizes must be aligned to this |
ca23e405e
|
2337 2338 2339 2340 2341 2342 |
* * Returns: kmalloc'd vm_struct pointer array pointing to allocated * vm_structs on success, %NULL on failure * * Percpu allocator wants to use congruent vm areas so that it can * maintain the offsets among percpu areas. This function allocates |
ec3f64fc9
|
2343 2344 2345 2346 |
* congruent vmalloc areas for it with GFP_KERNEL. These areas tend to * be scattered pretty far, distance between two areas easily going up * to gigabytes. To avoid interacting with regular vmallocs, these * areas are allocated from top. |
ca23e405e
|
2347 2348 2349 2350 2351 2352 |
* * Despite its complicated look, this allocator is rather simple. It * does everything top-down and scans areas from the end looking for * matching slot. While scanning, if any of the areas overlaps with * existing vmap_area, the base address is pulled down to fit the * area. Scanning is repeated till all the areas fit and then all |
c568da282
|
2353 |
* necessary data structures are inserted and the result is returned. |
ca23e405e
|
2354 2355 2356 |
*/ struct vm_struct **pcpu_get_vm_areas(const unsigned long *offsets, const size_t *sizes, int nr_vms, |
ec3f64fc9
|
2357 |
size_t align) |
ca23e405e
|
2358 2359 2360 2361 2362 2363 2364 2365 |
{ const unsigned long vmalloc_start = ALIGN(VMALLOC_START, align); const unsigned long vmalloc_end = VMALLOC_END & ~(align - 1); struct vmap_area **vas, *prev, *next; struct vm_struct **vms; int area, area2, last_area, term_area; unsigned long base, start, end, last_end; bool purged = false; |
ca23e405e
|
2366 |
/* verify parameters and allocate data structures */ |
891c49abf
|
2367 |
BUG_ON(offset_in_page(align) || !is_power_of_2(align)); |
ca23e405e
|
2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 |
for (last_area = 0, area = 0; area < nr_vms; area++) { start = offsets[area]; end = start + sizes[area]; /* is everything aligned properly? */ BUG_ON(!IS_ALIGNED(offsets[area], align)); BUG_ON(!IS_ALIGNED(sizes[area], align)); /* detect the area with the highest address */ if (start > offsets[last_area]) last_area = area; |
c568da282
|
2379 |
for (area2 = area + 1; area2 < nr_vms; area2++) { |
ca23e405e
|
2380 2381 |
unsigned long start2 = offsets[area2]; unsigned long end2 = start2 + sizes[area2]; |
c568da282
|
2382 |
BUG_ON(start2 < end && start < end2); |
ca23e405e
|
2383 2384 2385 2386 2387 2388 2389 2390 |
} } last_end = offsets[last_area] + sizes[last_area]; if (vmalloc_end - vmalloc_start < last_end) { WARN_ON(true); return NULL; } |
4d67d8605
|
2391 2392 |
vms = kcalloc(nr_vms, sizeof(vms[0]), GFP_KERNEL); vas = kcalloc(nr_vms, sizeof(vas[0]), GFP_KERNEL); |
ca23e405e
|
2393 |
if (!vas || !vms) |
f1db7afd9
|
2394 |
goto err_free2; |
ca23e405e
|
2395 2396 |
for (area = 0; area < nr_vms; area++) { |
ec3f64fc9
|
2397 2398 |
vas[area] = kzalloc(sizeof(struct vmap_area), GFP_KERNEL); vms[area] = kzalloc(sizeof(struct vm_struct), GFP_KERNEL); |
ca23e405e
|
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 |
if (!vas[area] || !vms[area]) goto err_free; } retry: spin_lock(&vmap_area_lock); /* start scanning - we scan from the top, begin with the last area */ area = term_area = last_area; start = offsets[area]; end = start + sizes[area]; if (!pvm_find_next_prev(vmap_area_pcpu_hole, &next, &prev)) { base = vmalloc_end - last_end; goto found; } base = pvm_determine_end(&next, &prev, align) - end; while (true) { BUG_ON(next && next->va_end <= base + end); BUG_ON(prev && prev->va_end > base + end); /* * base might have underflowed, add last_end before * comparing. */ if (base + last_end < vmalloc_start + last_end) { spin_unlock(&vmap_area_lock); if (!purged) { purge_vmap_area_lazy(); purged = true; goto retry; } goto err_free; } /* * If next overlaps, move base downwards so that it's * right below next and then recheck. */ if (next && next->va_start < base + end) { base = pvm_determine_end(&next, &prev, align) - end; term_area = area; continue; } /* * If prev overlaps, shift down next and prev and move * base so that it's right below new next and then * recheck. */ if (prev && prev->va_end > base + start) { next = prev; prev = node_to_va(rb_prev(&next->rb_node)); base = pvm_determine_end(&next, &prev, align) - end; term_area = area; continue; } /* * This area fits, move on to the previous one. If * the previous one is the terminal one, we're done. */ area = (area + nr_vms - 1) % nr_vms; if (area == term_area) break; start = offsets[area]; end = start + sizes[area]; pvm_find_next_prev(base + end, &next, &prev); } found: /* we've found a fitting base, insert all va's */ for (area = 0; area < nr_vms; area++) { struct vmap_area *va = vas[area]; va->va_start = base + offsets[area]; va->va_end = va->va_start + sizes[area]; __insert_vmap_area(va); } vmap_area_pcpu_hole = base + offsets[last_area]; spin_unlock(&vmap_area_lock); /* insert all vm's */ for (area = 0; area < nr_vms; area++) |
3645cb4a4
|
2484 2485 |
setup_vmalloc_vm(vms[area], vas[area], VM_ALLOC, pcpu_get_vm_areas); |
ca23e405e
|
2486 2487 2488 2489 2490 2491 |
kfree(vas); return vms; err_free: for (area = 0; area < nr_vms; area++) { |
f1db7afd9
|
2492 2493 |
kfree(vas[area]); kfree(vms[area]); |
ca23e405e
|
2494 |
} |
f1db7afd9
|
2495 |
err_free2: |
ca23e405e
|
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 |
kfree(vas); kfree(vms); return NULL; } /** * pcpu_free_vm_areas - free vmalloc areas for percpu allocator * @vms: vm_struct pointer array returned by pcpu_get_vm_areas() * @nr_vms: the number of allocated areas * * Free vm_structs and the array allocated by pcpu_get_vm_areas(). */ void pcpu_free_vm_areas(struct vm_struct **vms, int nr_vms) { int i; for (i = 0; i < nr_vms; i++) free_vm_area(vms[i]); kfree(vms); } |
4f8b02b4e
|
2516 |
#endif /* CONFIG_SMP */ |
a10aa5798
|
2517 2518 2519 |
#ifdef CONFIG_PROC_FS static void *s_start(struct seq_file *m, loff_t *pos) |
d4033afdf
|
2520 |
__acquires(&vmap_area_lock) |
a10aa5798
|
2521 |
{ |
d4033afdf
|
2522 |
spin_lock(&vmap_area_lock); |
3f5000693
|
2523 |
return seq_list_start(&vmap_area_list, *pos); |
a10aa5798
|
2524 2525 2526 2527 |
} static void *s_next(struct seq_file *m, void *p, loff_t *pos) { |
3f5000693
|
2528 |
return seq_list_next(p, &vmap_area_list, pos); |
a10aa5798
|
2529 2530 2531 |
} static void s_stop(struct seq_file *m, void *p) |
d4033afdf
|
2532 |
__releases(&vmap_area_lock) |
a10aa5798
|
2533 |
{ |
d4033afdf
|
2534 |
spin_unlock(&vmap_area_lock); |
a10aa5798
|
2535 |
} |
a47a126ad
|
2536 2537 |
static void show_numa_info(struct seq_file *m, struct vm_struct *v) { |
e5adfffc8
|
2538 |
if (IS_ENABLED(CONFIG_NUMA)) { |
a47a126ad
|
2539 2540 2541 2542 |
unsigned int nr, *counters = m->private; if (!counters) return; |
af12346cd
|
2543 2544 |
if (v->flags & VM_UNINITIALIZED) return; |
7e5b528b4
|
2545 2546 |
/* Pair with smp_wmb() in clear_vm_uninitialized_flag() */ smp_rmb(); |
af12346cd
|
2547 |
|
a47a126ad
|
2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 |
memset(counters, 0, nr_node_ids * sizeof(unsigned int)); for (nr = 0; nr < v->nr_pages; nr++) counters[page_to_nid(v->pages[nr])]++; for_each_node_state(nr, N_HIGH_MEMORY) if (counters[nr]) seq_printf(m, " N%u=%u", nr, counters[nr]); } } |
a10aa5798
|
2558 2559 |
static int s_show(struct seq_file *m, void *p) { |
3f5000693
|
2560 |
struct vmap_area *va; |
d4033afdf
|
2561 |
struct vm_struct *v; |
3f5000693
|
2562 |
va = list_entry(p, struct vmap_area, list); |
c2ce8c142
|
2563 2564 2565 2566 |
/* * s_show can encounter race with remove_vm_area, !VM_VM_AREA on * behalf of vmap area is being tear down or vm_map_ram allocation. */ |
78c72746f
|
2567 2568 2569 2570 2571 2572 |
if (!(va->flags & VM_VM_AREA)) { seq_printf(m, "0x%pK-0x%pK %7ld %s ", (void *)va->va_start, (void *)va->va_end, va->va_end - va->va_start, va->flags & VM_LAZY_FREE ? "unpurged vm_area" : "vm_map_ram"); |
d4033afdf
|
2573 |
return 0; |
78c72746f
|
2574 |
} |
d4033afdf
|
2575 2576 |
v = va->vm; |
a10aa5798
|
2577 |
|
45ec16908
|
2578 |
seq_printf(m, "0x%pK-0x%pK %7ld", |
a10aa5798
|
2579 |
v->addr, v->addr + v->size, v->size); |
62c70bce8
|
2580 2581 |
if (v->caller) seq_printf(m, " %pS", v->caller); |
230169693
|
2582 |
|
a10aa5798
|
2583 2584 2585 2586 |
if (v->nr_pages) seq_printf(m, " pages=%d", v->nr_pages); if (v->phys_addr) |
199eaa05a
|
2587 |
seq_printf(m, " phys=%pa", &v->phys_addr); |
a10aa5798
|
2588 2589 |
if (v->flags & VM_IOREMAP) |
f4527c908
|
2590 |
seq_puts(m, " ioremap"); |
a10aa5798
|
2591 2592 |
if (v->flags & VM_ALLOC) |
f4527c908
|
2593 |
seq_puts(m, " vmalloc"); |
a10aa5798
|
2594 2595 |
if (v->flags & VM_MAP) |
f4527c908
|
2596 |
seq_puts(m, " vmap"); |
a10aa5798
|
2597 2598 |
if (v->flags & VM_USERMAP) |
f4527c908
|
2599 |
seq_puts(m, " user"); |
a10aa5798
|
2600 |
|
244d63ee3
|
2601 |
if (is_vmalloc_addr(v->pages)) |
f4527c908
|
2602 |
seq_puts(m, " vpages"); |
a10aa5798
|
2603 |
|
a47a126ad
|
2604 |
show_numa_info(m, v); |
a10aa5798
|
2605 2606 2607 2608 |
seq_putc(m, ' '); return 0; } |
5f6a6a9c4
|
2609 |
static const struct seq_operations vmalloc_op = { |
a10aa5798
|
2610 2611 2612 2613 2614 |
.start = s_start, .next = s_next, .stop = s_stop, .show = s_show, }; |
5f6a6a9c4
|
2615 |
|
5f6a6a9c4
|
2616 2617 |
static int __init proc_vmalloc_init(void) { |
fddda2b7b
|
2618 |
if (IS_ENABLED(CONFIG_NUMA)) |
44414d82c
|
2619 2620 2621 |
proc_create_seq_private("vmallocinfo", S_IRUSR, NULL, &vmalloc_op, nr_node_ids * sizeof(unsigned int), NULL); |
fddda2b7b
|
2622 2623 |
else proc_create_seq("vmallocinfo", S_IRUSR, NULL, &vmalloc_op); |
5f6a6a9c4
|
2624 2625 2626 |
return 0; } module_init(proc_vmalloc_init); |
db3808c1b
|
2627 |
|
a10aa5798
|
2628 |
#endif |