Blame view
mm/vmalloc.c
91.1 KB
457c89965 treewide: Add SPD... |
1 |
// SPDX-License-Identifier: GPL-2.0-only |
1da177e4c Linux-2.6.12-rc2 |
2 |
/* |
1da177e4c Linux-2.6.12-rc2 |
3 4 5 6 |
* 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 [PATCH] vmalloc_node |
7 |
* Numa awareness, Christoph Lameter, SGI, June 2005 |
d758ffe6b mm/vmalloc: updat... |
8 |
* Improving global KVA allocator, Uladzislau Rezki, Sony, May 2019 |
1da177e4c Linux-2.6.12-rc2 |
9 |
*/ |
db64fe022 mm: rewrite vmap ... |
10 |
#include <linux/vmalloc.h> |
1da177e4c Linux-2.6.12-rc2 |
11 12 13 |
#include <linux/mm.h> #include <linux/module.h> #include <linux/highmem.h> |
c3edc4010 sched/headers: Mo... |
14 |
#include <linux/sched/signal.h> |
1da177e4c Linux-2.6.12-rc2 |
15 16 17 |
#include <linux/slab.h> #include <linux/spinlock.h> #include <linux/interrupt.h> |
5f6a6a9c4 proc: move /proc/... |
18 |
#include <linux/proc_fs.h> |
a10aa5798 vmalloc: show vma... |
19 |
#include <linux/seq_file.h> |
868b104d7 mm/vmalloc: Add f... |
20 |
#include <linux/set_memory.h> |
3ac7fe5a4 infrastructure to... |
21 |
#include <linux/debugobjects.h> |
230169693 vmallocinfo: add ... |
22 |
#include <linux/kallsyms.h> |
db64fe022 mm: rewrite vmap ... |
23 |
#include <linux/list.h> |
4da56b99d mm/vmap: Add a no... |
24 |
#include <linux/notifier.h> |
db64fe022 mm: rewrite vmap ... |
25 |
#include <linux/rbtree.h> |
0f14599c6 vmalloc: convert ... |
26 |
#include <linux/xarray.h> |
db64fe022 mm: rewrite vmap ... |
27 |
#include <linux/rcupdate.h> |
f0aa66179 vmalloc: implemen... |
28 |
#include <linux/pfn.h> |
89219d37a kmemleak: Add the... |
29 |
#include <linux/kmemleak.h> |
60063497a atomic: use <linu... |
30 |
#include <linux/atomic.h> |
3b32123d7 mm: use macros fr... |
31 |
#include <linux/compiler.h> |
32fcfd407 make vfree() safe... |
32 |
#include <linux/llist.h> |
0f616be12 mm: change __get_... |
33 |
#include <linux/bitops.h> |
68ad4a330 mm/vmalloc.c: kee... |
34 |
#include <linux/rbtree_augmented.h> |
bdebd6a28 vmalloc: fix rema... |
35 |
#include <linux/overflow.h> |
3b32123d7 mm: use macros fr... |
36 |
|
7c0f6ba68 Replace <asm/uacc... |
37 |
#include <linux/uaccess.h> |
1da177e4c Linux-2.6.12-rc2 |
38 |
#include <asm/tlbflush.h> |
2dca6999e mm, perf_event: M... |
39 |
#include <asm/shmparam.h> |
1da177e4c Linux-2.6.12-rc2 |
40 |
|
dd56b0464 mm: page_alloc: h... |
41 |
#include "internal.h" |
2a681cfa5 mm: move p?d_allo... |
42 |
#include "pgalloc-track.h" |
dd56b0464 mm: page_alloc: h... |
43 |
|
186525bd6 mm, x86/mm: Untan... |
44 45 46 47 48 49 50 |
bool is_vmalloc_addr(const void *x) { unsigned long addr = (unsigned long)x; return addr >= VMALLOC_START && addr < VMALLOC_END; } EXPORT_SYMBOL(is_vmalloc_addr); |
32fcfd407 make vfree() safe... |
51 52 53 54 55 56 57 58 59 60 61 |
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 mm/vmalloc.c: don... |
62 63 64 65 |
struct llist_node *t, *llnode; llist_for_each_safe(llnode, t, llist_del_all(&p->list)) __vunmap((void *)llnode, 1); |
32fcfd407 make vfree() safe... |
66 |
} |
db64fe022 mm: rewrite vmap ... |
67 |
/*** Page table manipulation functions ***/ |
b221385bc [PATCH] mm/: make... |
68 |
|
2ba3e6947 mm/vmalloc: track... |
69 70 |
static void vunmap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
71 72 73 74 75 76 77 78 |
{ 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); |
2ba3e6947 mm/vmalloc: track... |
79 |
*mask |= PGTBL_PTE_MODIFIED; |
1da177e4c Linux-2.6.12-rc2 |
80 |
} |
2ba3e6947 mm/vmalloc: track... |
81 82 |
static void vunmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
83 84 85 |
{ pmd_t *pmd; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
86 |
int cleared; |
1da177e4c Linux-2.6.12-rc2 |
87 88 89 90 |
pmd = pmd_offset(pud, addr); do { next = pmd_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
91 92 93 94 95 96 |
cleared = pmd_clear_huge(pmd); if (cleared || pmd_bad(*pmd)) *mask |= PGTBL_PMD_MODIFIED; if (cleared) |
b9820d8f3 mm: change vunmap... |
97 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
98 99 |
if (pmd_none_or_clear_bad(pmd)) continue; |
2ba3e6947 mm/vmalloc: track... |
100 |
vunmap_pte_range(pmd, addr, next, mask); |
e47110e90 mm/vunmap: add co... |
101 102 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
103 104 |
} while (pmd++, addr = next, addr != end); } |
2ba3e6947 mm/vmalloc: track... |
105 106 |
static void vunmap_pud_range(p4d_t *p4d, unsigned long addr, unsigned long end, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
107 108 109 |
{ pud_t *pud; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
110 |
int cleared; |
1da177e4c Linux-2.6.12-rc2 |
111 |
|
c2febafc6 mm: convert gener... |
112 |
pud = pud_offset(p4d, addr); |
1da177e4c Linux-2.6.12-rc2 |
113 114 |
do { next = pud_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
115 116 117 118 119 120 |
cleared = pud_clear_huge(pud); if (cleared || pud_bad(*pud)) *mask |= PGTBL_PUD_MODIFIED; if (cleared) |
b9820d8f3 mm: change vunmap... |
121 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
122 123 |
if (pud_none_or_clear_bad(pud)) continue; |
2ba3e6947 mm/vmalloc: track... |
124 |
vunmap_pmd_range(pud, addr, next, mask); |
1da177e4c Linux-2.6.12-rc2 |
125 126 |
} while (pud++, addr = next, addr != end); } |
2ba3e6947 mm/vmalloc: track... |
127 128 |
static void vunmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end, pgtbl_mod_mask *mask) |
c2febafc6 mm: convert gener... |
129 130 131 |
{ p4d_t *p4d; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
132 |
int cleared; |
c2febafc6 mm: convert gener... |
133 134 135 136 |
p4d = p4d_offset(pgd, addr); do { next = p4d_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
137 138 139 140 141 142 |
cleared = p4d_clear_huge(p4d); if (cleared || p4d_bad(*p4d)) *mask |= PGTBL_P4D_MODIFIED; if (cleared) |
c2febafc6 mm: convert gener... |
143 144 145 |
continue; if (p4d_none_or_clear_bad(p4d)) continue; |
2ba3e6947 mm/vmalloc: track... |
146 |
vunmap_pud_range(p4d, addr, next, mask); |
c2febafc6 mm: convert gener... |
147 148 |
} while (p4d++, addr = next, addr != end); } |
b521c43f5 mm: remove vmap_p... |
149 150 |
/** * unmap_kernel_range_noflush - unmap kernel VM area |
2ba3e6947 mm/vmalloc: track... |
151 |
* @start: start of the VM area to unmap |
b521c43f5 mm: remove vmap_p... |
152 153 154 155 156 157 158 159 160 161 |
* @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. */ |
2ba3e6947 mm/vmalloc: track... |
162 |
void unmap_kernel_range_noflush(unsigned long start, unsigned long size) |
1da177e4c Linux-2.6.12-rc2 |
163 |
{ |
2ba3e6947 mm/vmalloc: track... |
164 |
unsigned long end = start + size; |
1da177e4c Linux-2.6.12-rc2 |
165 |
unsigned long next; |
b521c43f5 mm: remove vmap_p... |
166 |
pgd_t *pgd; |
2ba3e6947 mm/vmalloc: track... |
167 168 |
unsigned long addr = start; pgtbl_mod_mask mask = 0; |
1da177e4c Linux-2.6.12-rc2 |
169 170 171 |
BUG_ON(addr >= end); pgd = pgd_offset_k(addr); |
1da177e4c Linux-2.6.12-rc2 |
172 173 |
do { next = pgd_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
174 175 |
if (pgd_bad(*pgd)) mask |= PGTBL_PGD_MODIFIED; |
1da177e4c Linux-2.6.12-rc2 |
176 177 |
if (pgd_none_or_clear_bad(pgd)) continue; |
2ba3e6947 mm/vmalloc: track... |
178 |
vunmap_p4d_range(pgd, addr, next, &mask); |
1da177e4c Linux-2.6.12-rc2 |
179 |
} while (pgd++, addr = next, addr != end); |
2ba3e6947 mm/vmalloc: track... |
180 181 182 |
if (mask & ARCH_PAGE_TABLE_SYNC_MASK) arch_sync_kernel_mappings(start, end); |
1da177e4c Linux-2.6.12-rc2 |
183 184 185 |
} static int vmap_pte_range(pmd_t *pmd, unsigned long addr, |
2ba3e6947 mm/vmalloc: track... |
186 187 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
188 189 |
{ pte_t *pte; |
db64fe022 mm: rewrite vmap ... |
190 191 192 193 |
/* * nr is a running index into the array which helps higher level * callers keep track of where we're up to. */ |
2ba3e6947 mm/vmalloc: track... |
194 |
pte = pte_alloc_kernel_track(pmd, addr, mask); |
1da177e4c Linux-2.6.12-rc2 |
195 196 197 |
if (!pte) return -ENOMEM; do { |
db64fe022 mm: rewrite vmap ... |
198 199 200 201 202 |
struct page *page = pages[*nr]; if (WARN_ON(!pte_none(*pte))) return -EBUSY; if (WARN_ON(!page)) |
1da177e4c Linux-2.6.12-rc2 |
203 204 |
return -ENOMEM; set_pte_at(&init_mm, addr, pte, mk_pte(page, prot)); |
db64fe022 mm: rewrite vmap ... |
205 |
(*nr)++; |
1da177e4c Linux-2.6.12-rc2 |
206 |
} while (pte++, addr += PAGE_SIZE, addr != end); |
2ba3e6947 mm/vmalloc: track... |
207 |
*mask |= PGTBL_PTE_MODIFIED; |
1da177e4c Linux-2.6.12-rc2 |
208 209 |
return 0; } |
db64fe022 mm: rewrite vmap ... |
210 |
static int vmap_pmd_range(pud_t *pud, unsigned long addr, |
2ba3e6947 mm/vmalloc: track... |
211 212 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
213 214 215 |
{ pmd_t *pmd; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
216 |
pmd = pmd_alloc_track(&init_mm, pud, addr, mask); |
1da177e4c Linux-2.6.12-rc2 |
217 218 219 220 |
if (!pmd) return -ENOMEM; do { next = pmd_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
221 |
if (vmap_pte_range(pmd, addr, next, prot, pages, nr, mask)) |
1da177e4c Linux-2.6.12-rc2 |
222 223 224 225 |
return -ENOMEM; } while (pmd++, addr = next, addr != end); return 0; } |
c2febafc6 mm: convert gener... |
226 |
static int vmap_pud_range(p4d_t *p4d, unsigned long addr, |
2ba3e6947 mm/vmalloc: track... |
227 228 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr, pgtbl_mod_mask *mask) |
1da177e4c Linux-2.6.12-rc2 |
229 230 231 |
{ pud_t *pud; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
232 |
pud = pud_alloc_track(&init_mm, p4d, addr, mask); |
1da177e4c Linux-2.6.12-rc2 |
233 234 235 236 |
if (!pud) return -ENOMEM; do { next = pud_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
237 |
if (vmap_pmd_range(pud, addr, next, prot, pages, nr, mask)) |
1da177e4c Linux-2.6.12-rc2 |
238 239 240 241 |
return -ENOMEM; } while (pud++, addr = next, addr != end); return 0; } |
c2febafc6 mm: convert gener... |
242 |
static int vmap_p4d_range(pgd_t *pgd, unsigned long addr, |
2ba3e6947 mm/vmalloc: track... |
243 244 |
unsigned long end, pgprot_t prot, struct page **pages, int *nr, pgtbl_mod_mask *mask) |
c2febafc6 mm: convert gener... |
245 246 247 |
{ p4d_t *p4d; unsigned long next; |
2ba3e6947 mm/vmalloc: track... |
248 |
p4d = p4d_alloc_track(&init_mm, pgd, addr, mask); |
c2febafc6 mm: convert gener... |
249 250 251 252 |
if (!p4d) return -ENOMEM; do { next = p4d_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
253 |
if (vmap_pud_range(p4d, addr, next, prot, pages, nr, mask)) |
c2febafc6 mm: convert gener... |
254 255 256 257 |
return -ENOMEM; } while (p4d++, addr = next, addr != end); return 0; } |
b521c43f5 mm: remove vmap_p... |
258 259 260 261 262 263 |
/** * 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 |
db64fe022 mm: rewrite vmap ... |
264 |
* |
b521c43f5 mm: remove vmap_p... |
265 266 267 268 269 270 271 272 273 |
* 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: |
60bb44652 mm: don't return ... |
274 |
* 0 on success, -errno on failure. |
db64fe022 mm: rewrite vmap ... |
275 |
*/ |
b521c43f5 mm: remove vmap_p... |
276 277 |
int map_kernel_range_noflush(unsigned long addr, unsigned long size, pgprot_t prot, struct page **pages) |
1da177e4c Linux-2.6.12-rc2 |
278 |
{ |
2ba3e6947 mm/vmalloc: track... |
279 |
unsigned long start = addr; |
b521c43f5 mm: remove vmap_p... |
280 |
unsigned long end = addr + size; |
1da177e4c Linux-2.6.12-rc2 |
281 |
unsigned long next; |
b521c43f5 mm: remove vmap_p... |
282 |
pgd_t *pgd; |
db64fe022 mm: rewrite vmap ... |
283 284 |
int err = 0; int nr = 0; |
2ba3e6947 mm/vmalloc: track... |
285 |
pgtbl_mod_mask mask = 0; |
1da177e4c Linux-2.6.12-rc2 |
286 287 288 |
BUG_ON(addr >= end); pgd = pgd_offset_k(addr); |
1da177e4c Linux-2.6.12-rc2 |
289 290 |
do { next = pgd_addr_end(addr, end); |
2ba3e6947 mm/vmalloc: track... |
291 292 293 |
if (pgd_bad(*pgd)) mask |= PGTBL_PGD_MODIFIED; err = vmap_p4d_range(pgd, addr, next, prot, pages, &nr, &mask); |
1da177e4c Linux-2.6.12-rc2 |
294 |
if (err) |
bf88c8c83 vmalloc.c: fix do... |
295 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
296 |
} while (pgd++, addr = next, addr != end); |
db64fe022 mm: rewrite vmap ... |
297 |
|
2ba3e6947 mm/vmalloc: track... |
298 299 |
if (mask & ARCH_PAGE_TABLE_SYNC_MASK) arch_sync_kernel_mappings(start, end); |
60bb44652 mm: don't return ... |
300 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
301 |
} |
ed1f324c5 mm: remove map_vm... |
302 303 |
int map_kernel_range(unsigned long start, unsigned long size, pgprot_t prot, struct page **pages) |
8fc489850 vmalloc: add un/m... |
304 305 |
{ int ret; |
a29adb620 mm: rename vmap_p... |
306 307 |
ret = map_kernel_range_noflush(start, size, prot, pages); flush_cache_vmap(start, start + size); |
8fc489850 vmalloc: add un/m... |
308 309 |
return ret; } |
81ac3ad90 kcore: register m... |
310 |
int is_vmalloc_or_module_addr(const void *x) |
73bdf0a60 Introduce is_vmal... |
311 312 |
{ /* |
ab4f2ee13 [ARM] fix naming ... |
313 |
* ARM, x86-64 and sparc64 put modules in a special place, |
73bdf0a60 Introduce is_vmal... |
314 315 316 317 318 319 320 321 322 323 |
* 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 Move vmalloc_to_p... |
324 |
/* |
add688fbd Revert "mm/vmallo... |
325 |
* Walk a vmap address to the struct page it maps. |
48667e7a4 Move vmalloc_to_p... |
326 |
*/ |
add688fbd Revert "mm/vmallo... |
327 |
struct page *vmalloc_to_page(const void *vmalloc_addr) |
48667e7a4 Move vmalloc_to_p... |
328 329 |
{ unsigned long addr = (unsigned long) vmalloc_addr; |
add688fbd Revert "mm/vmallo... |
330 |
struct page *page = NULL; |
48667e7a4 Move vmalloc_to_p... |
331 |
pgd_t *pgd = pgd_offset_k(addr); |
c2febafc6 mm: convert gener... |
332 333 334 335 |
p4d_t *p4d; pud_t *pud; pmd_t *pmd; pte_t *ptep, pte; |
48667e7a4 Move vmalloc_to_p... |
336 |
|
7aa413def x86, MM: virtual ... |
337 338 339 340 |
/* * XXX we might need to change this if we add VIRTUAL_BUG_ON for * architectures that do not vmalloc module space */ |
73bdf0a60 Introduce is_vmal... |
341 |
VIRTUAL_BUG_ON(!is_vmalloc_or_module_addr(vmalloc_addr)); |
59ea74633 MM: virtual addre... |
342 |
|
c2febafc6 mm: convert gener... |
343 344 345 346 347 348 |
if (pgd_none(*pgd)) return NULL; p4d = p4d_offset(pgd, addr); if (p4d_none(*p4d)) return NULL; pud = pud_offset(p4d, addr); |
029c54b09 mm/vmalloc.c: hug... |
349 350 351 352 353 354 355 356 357 358 359 |
/* * 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 mm: convert gener... |
360 361 |
return NULL; pmd = pmd_offset(pud, addr); |
029c54b09 mm/vmalloc.c: hug... |
362 363 |
WARN_ON_ONCE(pmd_bad(*pmd)); if (pmd_none(*pmd) || pmd_bad(*pmd)) |
c2febafc6 mm: convert gener... |
364 365 366 367 368 369 370 |
return NULL; ptep = pte_offset_map(pmd, addr); pte = *ptep; if (pte_present(pte)) page = pte_page(pte); pte_unmap(ptep); |
add688fbd Revert "mm/vmallo... |
371 |
return page; |
48667e7a4 Move vmalloc_to_p... |
372 |
} |
add688fbd Revert "mm/vmallo... |
373 |
EXPORT_SYMBOL(vmalloc_to_page); |
48667e7a4 Move vmalloc_to_p... |
374 375 |
/* |
add688fbd Revert "mm/vmallo... |
376 |
* Map a vmalloc()-space virtual address to the physical page frame number. |
48667e7a4 Move vmalloc_to_p... |
377 |
*/ |
add688fbd Revert "mm/vmallo... |
378 |
unsigned long vmalloc_to_pfn(const void *vmalloc_addr) |
48667e7a4 Move vmalloc_to_p... |
379 |
{ |
add688fbd Revert "mm/vmallo... |
380 |
return page_to_pfn(vmalloc_to_page(vmalloc_addr)); |
48667e7a4 Move vmalloc_to_p... |
381 |
} |
add688fbd Revert "mm/vmallo... |
382 |
EXPORT_SYMBOL(vmalloc_to_pfn); |
48667e7a4 Move vmalloc_to_p... |
383 |
|
db64fe022 mm: rewrite vmap ... |
384 385 |
/*** Global kva allocator ***/ |
bb850f4da mm/vmap: add DEBU... |
386 |
#define DEBUG_AUGMENT_PROPAGATE_CHECK 0 |
a6cf4e0fe mm/vmap: add DEBU... |
387 |
#define DEBUG_AUGMENT_LOWEST_MATCH_CHECK 0 |
bb850f4da mm/vmap: add DEBU... |
388 |
|
db64fe022 mm: rewrite vmap ... |
389 |
|
db64fe022 mm: rewrite vmap ... |
390 |
static DEFINE_SPINLOCK(vmap_area_lock); |
e36176be1 mm/vmalloc: rewor... |
391 |
static DEFINE_SPINLOCK(free_vmap_area_lock); |
f1c4069e1 mm, vmalloc: expo... |
392 393 |
/* Export for kexec only */ LIST_HEAD(vmap_area_list); |
80c4bd7a5 mm/vmalloc: keep ... |
394 |
static LLIST_HEAD(vmap_purge_list); |
89699605f mm: vmap area cache |
395 |
static struct rb_root vmap_area_root = RB_ROOT; |
68ad4a330 mm/vmalloc.c: kee... |
396 |
static bool vmap_initialized __read_mostly; |
89699605f mm: vmap area cache |
397 |
|
68ad4a330 mm/vmalloc.c: kee... |
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 |
/* * This kmem_cache is used for vmap_area objects. Instead of * allocating from slab we reuse an object from this cache to * make things faster. Especially in "no edge" splitting of * free block. */ static struct kmem_cache *vmap_area_cachep; /* * This linked list is used in pair with free_vmap_area_root. * It gives O(1) access to prev/next to perform fast coalescing. */ static LIST_HEAD(free_vmap_area_list); /* * This augment red-black tree represents the free vmap space. * All vmap_area objects in this tree are sorted by va->va_start * address. It is used for allocation and merging when a vmap * object is released. * * Each vmap_area node contains a maximum available free block * of its sub-tree, right or left. Therefore it is possible to * find a lowest match of free area. */ static struct rb_root free_vmap_area_root = RB_ROOT; |
82dd23e84 mm/vmalloc.c: pre... |
423 424 425 426 427 428 |
/* * Preload a CPU with one object for "no edge" split case. The * aim is to get rid of allocations from the atomic context, thus * to use more permissive allocation masks. */ static DEFINE_PER_CPU(struct vmap_area *, ne_fit_preload_node); |
68ad4a330 mm/vmalloc.c: kee... |
429 430 431 432 433 434 435 436 437 438 439 440 441 442 |
static __always_inline unsigned long va_size(struct vmap_area *va) { return (va->va_end - va->va_start); } static __always_inline unsigned long get_subtree_max_size(struct rb_node *node) { struct vmap_area *va; va = rb_entry_safe(node, struct vmap_area, rb_node); return va ? va->subtree_max_size : 0; } |
89699605f mm: vmap area cache |
443 |
|
68ad4a330 mm/vmalloc.c: kee... |
444 445 446 447 448 449 450 451 452 453 |
/* * Gets called when remove the node and rotate. */ static __always_inline unsigned long compute_subtree_max_size(struct vmap_area *va) { return max3(va_size(va), get_subtree_max_size(va->rb_node.rb_left), get_subtree_max_size(va->rb_node.rb_right)); } |
315cc066b augmented rbtree:... |
454 455 |
RB_DECLARE_CALLBACKS_MAX(static, free_vmap_area_rb_augment_cb, struct vmap_area, rb_node, unsigned long, subtree_max_size, va_size) |
68ad4a330 mm/vmalloc.c: kee... |
456 457 458 459 |
static void purge_vmap_area_lazy(void); static BLOCKING_NOTIFIER_HEAD(vmap_notify_list); static unsigned long lazy_max_pages(void); |
db64fe022 mm: rewrite vmap ... |
460 |
|
97105f0ab mm: vmalloc: show... |
461 462 463 464 465 466 |
static atomic_long_t nr_vmalloc_pages; unsigned long vmalloc_nr_pages(void) { return atomic_long_read(&nr_vmalloc_pages); } |
e46a1a5e3 ANDROID: mm: Expo... |
467 |
EXPORT_SYMBOL_GPL(vmalloc_nr_pages); |
97105f0ab mm: vmalloc: show... |
468 |
|
db64fe022 mm: rewrite vmap ... |
469 |
static struct vmap_area *__find_vmap_area(unsigned long addr) |
1da177e4c Linux-2.6.12-rc2 |
470 |
{ |
db64fe022 mm: rewrite vmap ... |
471 472 473 474 475 476 477 478 |
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 vmalloc: make fin... |
479 |
else if (addr >= va->va_end) |
db64fe022 mm: rewrite vmap ... |
480 481 482 483 484 485 486 |
n = n->rb_right; else return va; } return NULL; } |
68ad4a330 mm/vmalloc.c: kee... |
487 488 489 |
/* * This function returns back addresses of parent node * and its left or right link for further processing. |
9c801f61d mm/vmalloc.c: rem... |
490 491 492 493 |
* * Otherwise NULL is returned. In that case all further * steps regarding inserting of conflicting overlap range * have to be declined and actually considered as a bug. |
68ad4a330 mm/vmalloc.c: kee... |
494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 |
*/ static __always_inline struct rb_node ** find_va_links(struct vmap_area *va, struct rb_root *root, struct rb_node *from, struct rb_node **parent) { struct vmap_area *tmp_va; struct rb_node **link; if (root) { link = &root->rb_node; if (unlikely(!*link)) { *parent = NULL; return link; } } else { link = &from; } |
db64fe022 mm: rewrite vmap ... |
512 |
|
68ad4a330 mm/vmalloc.c: kee... |
513 514 515 516 517 518 519 |
/* * Go to the bottom of the tree. When we hit the last point * we end up with parent rb_node and correct direction, i name * it link, where the new va->rb_node will be attached to. */ do { tmp_va = rb_entry(*link, struct vmap_area, rb_node); |
db64fe022 mm: rewrite vmap ... |
520 |
|
68ad4a330 mm/vmalloc.c: kee... |
521 522 523 524 525 526 527 528 529 530 531 |
/* * During the traversal we also do some sanity check. * Trigger the BUG() if there are sides(left/right) * or full overlaps. */ if (va->va_start < tmp_va->va_end && va->va_end <= tmp_va->va_start) link = &(*link)->rb_left; else if (va->va_end > tmp_va->va_start && va->va_start >= tmp_va->va_end) link = &(*link)->rb_right; |
9c801f61d mm/vmalloc.c: rem... |
532 533 534 535 536 537 538 |
else { WARN(1, "vmalloc bug: 0x%lx-0x%lx overlaps with 0x%lx-0x%lx ", va->va_start, va->va_end, tmp_va->va_start, tmp_va->va_end); return NULL; } |
68ad4a330 mm/vmalloc.c: kee... |
539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 |
} while (*link); *parent = &tmp_va->rb_node; return link; } static __always_inline struct list_head * get_va_next_sibling(struct rb_node *parent, struct rb_node **link) { struct list_head *list; if (unlikely(!parent)) /* * The red-black tree where we try to find VA neighbors * before merging or inserting is empty, i.e. it means * there is no free vmap space. Normally it does not * happen but we handle this case anyway. */ return NULL; list = &rb_entry(parent, struct vmap_area, rb_node)->list; return (&parent->rb_right == link ? list->next : list); } static __always_inline void link_va(struct vmap_area *va, struct rb_root *root, struct rb_node *parent, struct rb_node **link, struct list_head *head) { /* * VA is still not in the list, but we can * identify its future previous list_head node. */ if (likely(parent)) { head = &rb_entry(parent, struct vmap_area, rb_node)->list; if (&parent->rb_right != link) head = head->prev; |
db64fe022 mm: rewrite vmap ... |
575 |
} |
68ad4a330 mm/vmalloc.c: kee... |
576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 |
/* Insert to the rb-tree */ rb_link_node(&va->rb_node, parent, link); if (root == &free_vmap_area_root) { /* * Some explanation here. Just perform simple insertion * to the tree. We do not set va->subtree_max_size to * its current size before calling rb_insert_augmented(). * It is because of we populate the tree from the bottom * to parent levels when the node _is_ in the tree. * * Therefore we set subtree_max_size to zero after insertion, * to let __augment_tree_propagate_from() puts everything to * the correct order later on. */ rb_insert_augmented(&va->rb_node, root, &free_vmap_area_rb_augment_cb); va->subtree_max_size = 0; } else { rb_insert_color(&va->rb_node, root); } |
db64fe022 mm: rewrite vmap ... |
596 |
|
68ad4a330 mm/vmalloc.c: kee... |
597 598 |
/* Address-sort this list */ list_add(&va->list, head); |
db64fe022 mm: rewrite vmap ... |
599 |
} |
68ad4a330 mm/vmalloc.c: kee... |
600 601 602 |
static __always_inline void unlink_va(struct vmap_area *va, struct rb_root *root) { |
460e42d19 mm/vmalloc.c: swi... |
603 604 |
if (WARN_ON(RB_EMPTY_NODE(&va->rb_node))) return; |
db64fe022 mm: rewrite vmap ... |
605 |
|
460e42d19 mm/vmalloc.c: swi... |
606 607 608 609 610 611 612 613 |
if (root == &free_vmap_area_root) rb_erase_augmented(&va->rb_node, root, &free_vmap_area_rb_augment_cb); else rb_erase(&va->rb_node, root); list_del(&va->list); RB_CLEAR_NODE(&va->rb_node); |
68ad4a330 mm/vmalloc.c: kee... |
614 |
} |
bb850f4da mm/vmap: add DEBU... |
615 616 |
#if DEBUG_AUGMENT_PROPAGATE_CHECK static void |
da27c9ed1 mm/vmalloc: simpl... |
617 |
augment_tree_propagate_check(void) |
bb850f4da mm/vmap: add DEBU... |
618 619 |
{ struct vmap_area *va; |
da27c9ed1 mm/vmalloc: simpl... |
620 |
unsigned long computed_size; |
bb850f4da mm/vmap: add DEBU... |
621 |
|
da27c9ed1 mm/vmalloc: simpl... |
622 623 624 625 626 627 |
list_for_each_entry(va, &free_vmap_area_list, list) { computed_size = compute_subtree_max_size(va); if (computed_size != va->subtree_max_size) pr_emerg("tree is corrupted: %lu, %lu ", va_size(va), va->subtree_max_size); |
bb850f4da mm/vmap: add DEBU... |
628 |
} |
bb850f4da mm/vmap: add DEBU... |
629 630 |
} #endif |
68ad4a330 mm/vmalloc.c: kee... |
631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 |
/* * This function populates subtree_max_size from bottom to upper * levels starting from VA point. The propagation must be done * when VA size is modified by changing its va_start/va_end. Or * in case of newly inserting of VA to the tree. * * It means that __augment_tree_propagate_from() must be called: * - After VA has been inserted to the tree(free path); * - After VA has been shrunk(allocation path); * - After VA has been increased(merging path). * * Please note that, it does not mean that upper parent nodes * and their subtree_max_size are recalculated all the time up * to the root node. * * 4--8 * /\ * / \ * / \ * 2--2 8--8 * * For example if we modify the node 4, shrinking it to 2, then * no any modification is required. If we shrink the node 2 to 1 * its subtree_max_size is updated only, and set to 1. If we shrink * the node 8 to 6, then its subtree_max_size is set to 6 and parent * node becomes 4--6. */ static __always_inline void augment_tree_propagate_from(struct vmap_area *va) { |
15ae144f7 mm/vmalloc: switc... |
661 662 663 664 665 666 |
/* * Populate the tree from bottom towards the root until * the calculated maximum available size of checked node * is equal to its current one. */ free_vmap_area_rb_augment_cb_propagate(&va->rb_node, NULL); |
bb850f4da mm/vmap: add DEBU... |
667 668 |
#if DEBUG_AUGMENT_PROPAGATE_CHECK |
da27c9ed1 mm/vmalloc: simpl... |
669 |
augment_tree_propagate_check(); |
bb850f4da mm/vmap: add DEBU... |
670 |
#endif |
68ad4a330 mm/vmalloc.c: kee... |
671 672 673 674 675 676 677 678 679 680 |
} static void insert_vmap_area(struct vmap_area *va, struct rb_root *root, struct list_head *head) { struct rb_node **link; struct rb_node *parent; link = find_va_links(va, root, NULL, &parent); |
9c801f61d mm/vmalloc.c: rem... |
681 682 |
if (link) link_va(va, root, parent, link, head); |
68ad4a330 mm/vmalloc.c: kee... |
683 684 685 686 687 688 689 690 691 692 693 694 695 696 |
} static void insert_vmap_area_augment(struct vmap_area *va, struct rb_node *from, struct rb_root *root, struct list_head *head) { struct rb_node **link; struct rb_node *parent; if (from) link = find_va_links(va, NULL, from, &parent); else link = find_va_links(va, root, NULL, &parent); |
9c801f61d mm/vmalloc.c: rem... |
697 698 699 700 |
if (link) { link_va(va, root, parent, link, head); augment_tree_propagate_from(va); } |
68ad4a330 mm/vmalloc.c: kee... |
701 702 703 704 705 706 707 |
} /* * Merge de-allocated chunk of VA memory with previous * and next free blocks. If coalesce is not done a new * free area is inserted. If VA has been merged, it is * freed. |
9c801f61d mm/vmalloc.c: rem... |
708 709 710 711 712 |
* * Please note, it can return NULL in case of overlap * ranges, followed by WARN() report. Despite it is a * buggy behaviour, a system can be alive and keep * ongoing. |
68ad4a330 mm/vmalloc.c: kee... |
713 |
*/ |
3c5c3cfb9 kasan: support ba... |
714 |
static __always_inline struct vmap_area * |
68ad4a330 mm/vmalloc.c: kee... |
715 716 717 718 719 720 721 722 723 724 725 726 727 728 |
merge_or_add_vmap_area(struct vmap_area *va, struct rb_root *root, struct list_head *head) { struct vmap_area *sibling; struct list_head *next; struct rb_node **link; struct rb_node *parent; bool merged = false; /* * Find a place in the tree where VA potentially will be * inserted, unless it is merged with its sibling/siblings. */ link = find_va_links(va, root, NULL, &parent); |
9c801f61d mm/vmalloc.c: rem... |
729 730 |
if (!link) return NULL; |
68ad4a330 mm/vmalloc.c: kee... |
731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 |
/* * Get next node of VA to check if merging can be done. */ next = get_va_next_sibling(parent, link); if (unlikely(next == NULL)) goto insert; /* * start end * | | * |<------VA------>|<-----Next----->| * | | * start end */ if (next != head) { sibling = list_entry(next, struct vmap_area, list); if (sibling->va_start == va->va_end) { sibling->va_start = va->va_start; |
68ad4a330 mm/vmalloc.c: kee... |
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 |
/* Free vmap_area object. */ kmem_cache_free(vmap_area_cachep, va); /* Point to the new merged area. */ va = sibling; merged = true; } } /* * start end * | | * |<-----Prev----->|<------VA------>| * | | * start end */ if (next->prev != head) { sibling = list_entry(next->prev, struct vmap_area, list); if (sibling->va_end == va->va_start) { |
5dd786409 mm/vmalloc: simpl... |
769 770 771 772 773 774 775 |
/* * If both neighbors are coalesced, it is important * to unlink the "next" node first, followed by merging * with "previous" one. Otherwise the tree might not be * fully populated if a sibling's augmented value is * "normalized" because of rotation operations. */ |
54f63d9d8 mm/vmalloc.c: get... |
776 777 |
if (merged) unlink_va(va, root); |
68ad4a330 mm/vmalloc.c: kee... |
778 |
|
5dd786409 mm/vmalloc: simpl... |
779 |
sibling->va_end = va->va_end; |
68ad4a330 mm/vmalloc.c: kee... |
780 781 |
/* Free vmap_area object. */ kmem_cache_free(vmap_area_cachep, va); |
3c5c3cfb9 kasan: support ba... |
782 783 784 785 |
/* Point to the new merged area. */ va = sibling; merged = true; |
68ad4a330 mm/vmalloc.c: kee... |
786 787 788 789 |
} } insert: |
5dd786409 mm/vmalloc: simpl... |
790 |
if (!merged) |
68ad4a330 mm/vmalloc.c: kee... |
791 |
link_va(va, root, parent, link, head); |
3c5c3cfb9 kasan: support ba... |
792 |
|
5dd786409 mm/vmalloc: simpl... |
793 794 795 796 |
/* * Last step is to check and update the tree. */ augment_tree_propagate_from(va); |
3c5c3cfb9 kasan: support ba... |
797 |
return va; |
68ad4a330 mm/vmalloc.c: kee... |
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 |
} static __always_inline bool is_within_this_va(struct vmap_area *va, unsigned long size, unsigned long align, unsigned long vstart) { unsigned long nva_start_addr; if (va->va_start > vstart) nva_start_addr = ALIGN(va->va_start, align); else nva_start_addr = ALIGN(vstart, align); /* Can be overflowed due to big size or alignment. */ if (nva_start_addr + size < nva_start_addr || nva_start_addr < vstart) return false; return (nva_start_addr + size <= va->va_end); } /* * Find the first free block(lowest start address) in the tree, * that will accomplish the request corresponding to passing * parameters. */ static __always_inline struct vmap_area * find_vmap_lowest_match(unsigned long size, unsigned long align, unsigned long vstart) { struct vmap_area *va; struct rb_node *node; unsigned long length; /* Start from the root. */ node = free_vmap_area_root.rb_node; /* Adjust the search size for alignment overhead. */ length = size + align - 1; while (node) { va = rb_entry(node, struct vmap_area, rb_node); if (get_subtree_max_size(node->rb_left) >= length && vstart < va->va_start) { node = node->rb_left; } else { if (is_within_this_va(va, size, align, vstart)) return va; /* * Does not make sense to go deeper towards the right * sub-tree if it does not have a free block that is * equal or bigger to the requested search length. */ if (get_subtree_max_size(node->rb_right) >= length) { node = node->rb_right; continue; } /* |
3806b0414 mm/vmalloc.c: fix... |
859 |
* OK. We roll back and find the first right sub-tree, |
68ad4a330 mm/vmalloc.c: kee... |
860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 |
* that will satisfy the search criteria. It can happen * only once due to "vstart" restriction. */ while ((node = rb_parent(node))) { va = rb_entry(node, struct vmap_area, rb_node); if (is_within_this_va(va, size, align, vstart)) return va; if (get_subtree_max_size(node->rb_right) >= length && vstart <= va->va_start) { node = node->rb_right; break; } } } } return NULL; } |
a6cf4e0fe mm/vmap: add DEBU... |
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 |
#if DEBUG_AUGMENT_LOWEST_MATCH_CHECK #include <linux/random.h> static struct vmap_area * find_vmap_lowest_linear_match(unsigned long size, unsigned long align, unsigned long vstart) { struct vmap_area *va; list_for_each_entry(va, &free_vmap_area_list, list) { if (!is_within_this_va(va, size, align, vstart)) continue; return va; } return NULL; } static void find_vmap_lowest_match_check(unsigned long size) { struct vmap_area *va_1, *va_2; unsigned long vstart; unsigned int rnd; get_random_bytes(&rnd, sizeof(rnd)); vstart = VMALLOC_START + rnd; va_1 = find_vmap_lowest_match(size, 1, vstart); va_2 = find_vmap_lowest_linear_match(size, 1, vstart); if (va_1 != va_2) pr_emerg("not lowest: t: 0x%p, l: 0x%p, v: 0x%lx ", va_1, va_2, vstart); } #endif |
68ad4a330 mm/vmalloc.c: kee... |
917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 |
enum fit_type { NOTHING_FIT = 0, FL_FIT_TYPE = 1, /* full fit */ LE_FIT_TYPE = 2, /* left edge fit */ RE_FIT_TYPE = 3, /* right edge fit */ NE_FIT_TYPE = 4 /* no edge fit */ }; static __always_inline enum fit_type classify_va_fit_type(struct vmap_area *va, unsigned long nva_start_addr, unsigned long size) { enum fit_type type; /* Check if it is within VA. */ if (nva_start_addr < va->va_start || nva_start_addr + size > va->va_end) return NOTHING_FIT; /* Now classify. */ if (va->va_start == nva_start_addr) { if (va->va_end == nva_start_addr + size) type = FL_FIT_TYPE; else type = LE_FIT_TYPE; } else if (va->va_end == nva_start_addr + size) { type = RE_FIT_TYPE; } else { type = NE_FIT_TYPE; } return type; } static __always_inline int adjust_va_to_fit_type(struct vmap_area *va, unsigned long nva_start_addr, unsigned long size, enum fit_type type) { |
2c9292336 mm/vmalloc.c: avo... |
956 |
struct vmap_area *lva = NULL; |
68ad4a330 mm/vmalloc.c: kee... |
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 |
if (type == FL_FIT_TYPE) { /* * No need to split VA, it fully fits. * * | | * V NVA V * |---------------| */ unlink_va(va, &free_vmap_area_root); kmem_cache_free(vmap_area_cachep, va); } else if (type == LE_FIT_TYPE) { /* * Split left edge of fit VA. * * | | * V NVA V R * |-------|-------| */ va->va_start += size; } else if (type == RE_FIT_TYPE) { /* * Split right edge of fit VA. * * | | * L V NVA V * |-------|-------| */ va->va_end = nva_start_addr; } else if (type == NE_FIT_TYPE) { /* * Split no edge of fit VA. * * | | * L V NVA V R * |---|-------|---| */ |
82dd23e84 mm/vmalloc.c: pre... |
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 |
lva = __this_cpu_xchg(ne_fit_preload_node, NULL); if (unlikely(!lva)) { /* * For percpu allocator we do not do any pre-allocation * and leave it as it is. The reason is it most likely * never ends up with NE_FIT_TYPE splitting. In case of * percpu allocations offsets and sizes are aligned to * fixed align request, i.e. RE_FIT_TYPE and FL_FIT_TYPE * are its main fitting cases. * * There are a few exceptions though, as an example it is * a first allocation (early boot up) when we have "one" * big free space that has to be split. |
060650a2a mm/vmalloc: add m... |
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 |
* * Also we can hit this path in case of regular "vmap" * allocations, if "this" current CPU was not preloaded. * See the comment in alloc_vmap_area() why. If so, then * GFP_NOWAIT is used instead to get an extra object for * split purpose. That is rare and most time does not * occur. * * What happens if an allocation gets failed. Basically, * an "overflow" path is triggered to purge lazily freed * areas to free some memory, then, the "retry" path is * triggered to repeat one more time. See more details * in alloc_vmap_area() function. |
82dd23e84 mm/vmalloc.c: pre... |
1020 1021 1022 1023 1024 |
*/ lva = kmem_cache_alloc(vmap_area_cachep, GFP_NOWAIT); if (!lva) return -1; } |
68ad4a330 mm/vmalloc.c: kee... |
1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 |
/* * Build the remainder. */ lva->va_start = va->va_start; lva->va_end = nva_start_addr; /* * Shrink this VA to remaining size. */ va->va_start = nva_start_addr + size; } else { return -1; } if (type != FL_FIT_TYPE) { augment_tree_propagate_from(va); |
2c9292336 mm/vmalloc.c: avo... |
1042 |
if (lva) /* type == NE_FIT_TYPE */ |
68ad4a330 mm/vmalloc.c: kee... |
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 |
insert_vmap_area_augment(lva, &va->rb_node, &free_vmap_area_root, &free_vmap_area_list); } return 0; } /* * Returns a start address of the newly allocated area, if success. * Otherwise a vend is returned that indicates failure. */ static __always_inline unsigned long __alloc_vmap_area(unsigned long size, unsigned long align, |
cacca6baf mm/vmalloc.c: rem... |
1056 |
unsigned long vstart, unsigned long vend) |
68ad4a330 mm/vmalloc.c: kee... |
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 |
{ unsigned long nva_start_addr; struct vmap_area *va; enum fit_type type; int ret; va = find_vmap_lowest_match(size, align, vstart); if (unlikely(!va)) return vend; if (va->va_start > vstart) nva_start_addr = ALIGN(va->va_start, align); else nva_start_addr = ALIGN(vstart, align); /* Check the "vend" restriction. */ if (nva_start_addr + size > vend) return vend; /* Classify what we have found. */ type = classify_va_fit_type(va, nva_start_addr, size); if (WARN_ON_ONCE(type == NOTHING_FIT)) return vend; /* Update the free vmap_area. */ ret = adjust_va_to_fit_type(va, nva_start_addr, size, type); if (ret) return vend; |
a6cf4e0fe mm/vmap: add DEBU... |
1085 1086 1087 |
#if DEBUG_AUGMENT_LOWEST_MATCH_CHECK find_vmap_lowest_match_check(size); #endif |
68ad4a330 mm/vmalloc.c: kee... |
1088 1089 |
return nva_start_addr; } |
4da56b99d mm/vmap: Add a no... |
1090 |
|
db64fe022 mm: rewrite vmap ... |
1091 |
/* |
d98c9e83b kasan: fix crashe... |
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 |
* Free a region of KVA allocated by alloc_vmap_area */ static void free_vmap_area(struct vmap_area *va) { /* * Remove from the busy tree/list. */ spin_lock(&vmap_area_lock); unlink_va(va, &vmap_area_root); spin_unlock(&vmap_area_lock); /* * Insert/Merge it back to the free tree/list. */ spin_lock(&free_vmap_area_lock); merge_or_add_vmap_area(va, &free_vmap_area_root, &free_vmap_area_list); spin_unlock(&free_vmap_area_lock); } /* |
db64fe022 mm: rewrite vmap ... |
1112 1113 1114 1115 1116 1117 1118 1119 |
* 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) { |
82dd23e84 mm/vmalloc.c: pre... |
1120 |
struct vmap_area *va, *pva; |
1da177e4c Linux-2.6.12-rc2 |
1121 |
unsigned long addr; |
db64fe022 mm: rewrite vmap ... |
1122 |
int purged = 0; |
d98c9e83b kasan: fix crashe... |
1123 |
int ret; |
db64fe022 mm: rewrite vmap ... |
1124 |
|
7766970cc mm: vmap fix over... |
1125 |
BUG_ON(!size); |
891c49abf mm/vmalloc: use o... |
1126 |
BUG_ON(offset_in_page(size)); |
89699605f mm: vmap area cache |
1127 |
BUG_ON(!is_power_of_2(align)); |
db64fe022 mm: rewrite vmap ... |
1128 |
|
68ad4a330 mm/vmalloc.c: kee... |
1129 1130 |
if (unlikely(!vmap_initialized)) return ERR_PTR(-EBUSY); |
5803ed292 mm: mark all call... |
1131 |
might_sleep(); |
f07116d77 mm/vmalloc: respe... |
1132 |
gfp_mask = gfp_mask & GFP_RECLAIM_MASK; |
4da56b99d mm/vmap: Add a no... |
1133 |
|
f07116d77 mm/vmalloc: respe... |
1134 |
va = kmem_cache_alloc_node(vmap_area_cachep, gfp_mask, node); |
db64fe022 mm: rewrite vmap ... |
1135 1136 |
if (unlikely(!va)) return ERR_PTR(-ENOMEM); |
7f88f88f8 mm: kmemleak: avo... |
1137 1138 1139 1140 |
/* * Only scan the relevant parts containing pointers to other objects * to avoid false negatives. */ |
f07116d77 mm/vmalloc: respe... |
1141 |
kmemleak_scan_area(&va->rb_node, SIZE_MAX, gfp_mask); |
7f88f88f8 mm: kmemleak: avo... |
1142 |
|
db64fe022 mm: rewrite vmap ... |
1143 |
retry: |
82dd23e84 mm/vmalloc.c: pre... |
1144 |
/* |
81f1ba586 mm/vmalloc: remov... |
1145 1146 1147 1148 1149 1150 |
* Preload this CPU with one extra vmap_area object. It is used * when fit type of free area is NE_FIT_TYPE. Please note, it * does not guarantee that an allocation occurs on a CPU that * is preloaded, instead we minimize the case when it is not. * It can happen because of cpu migration, because there is a * race until the below spinlock is taken. |
82dd23e84 mm/vmalloc.c: pre... |
1151 1152 1153 |
* * The preload is done in non-atomic context, thus it allows us * to use more permissive allocation masks to be more stable under |
81f1ba586 mm/vmalloc: remov... |
1154 1155 |
* low memory condition and high memory pressure. In rare case, * if not preloaded, GFP_NOWAIT is used. |
82dd23e84 mm/vmalloc.c: pre... |
1156 |
* |
81f1ba586 mm/vmalloc: remov... |
1157 |
* Set "pva" to NULL here, because of "retry" path. |
82dd23e84 mm/vmalloc.c: pre... |
1158 |
*/ |
81f1ba586 mm/vmalloc: remov... |
1159 |
pva = NULL; |
82dd23e84 mm/vmalloc.c: pre... |
1160 |
|
81f1ba586 mm/vmalloc: remov... |
1161 1162 1163 1164 1165 1166 |
if (!this_cpu_read(ne_fit_preload_node)) /* * Even if it fails we do not really care about that. * Just proceed as it is. If needed "overflow" path * will refill the cache we allocate from. */ |
f07116d77 mm/vmalloc: respe... |
1167 |
pva = kmem_cache_alloc_node(vmap_area_cachep, gfp_mask, node); |
82dd23e84 mm/vmalloc.c: pre... |
1168 |
|
e36176be1 mm/vmalloc: rewor... |
1169 |
spin_lock(&free_vmap_area_lock); |
81f1ba586 mm/vmalloc: remov... |
1170 1171 1172 |
if (pva && __this_cpu_cmpxchg(ne_fit_preload_node, NULL, pva)) kmem_cache_free(vmap_area_cachep, pva); |
89699605f mm: vmap area cache |
1173 |
|
afd07389d mm/vmalloc.c: fix... |
1174 |
/* |
68ad4a330 mm/vmalloc.c: kee... |
1175 1176 |
* If an allocation fails, the "vend" address is * returned. Therefore trigger the overflow path. |
afd07389d mm/vmalloc.c: fix... |
1177 |
*/ |
cacca6baf mm/vmalloc.c: rem... |
1178 |
addr = __alloc_vmap_area(size, align, vstart, vend); |
e36176be1 mm/vmalloc: rewor... |
1179 |
spin_unlock(&free_vmap_area_lock); |
68ad4a330 mm/vmalloc.c: kee... |
1180 |
if (unlikely(addr == vend)) |
89699605f mm: vmap area cache |
1181 |
goto overflow; |
db64fe022 mm: rewrite vmap ... |
1182 1183 1184 |
va->va_start = addr; va->va_end = addr + size; |
688fcbfc0 mm/vmalloc: modif... |
1185 |
va->vm = NULL; |
68ad4a330 mm/vmalloc.c: kee... |
1186 |
|
d98c9e83b kasan: fix crashe... |
1187 |
|
e36176be1 mm/vmalloc: rewor... |
1188 1189 |
spin_lock(&vmap_area_lock); insert_vmap_area(va, &vmap_area_root, &vmap_area_list); |
db64fe022 mm: rewrite vmap ... |
1190 |
spin_unlock(&vmap_area_lock); |
61e165578 mm/vmalloc.c: use... |
1191 |
BUG_ON(!IS_ALIGNED(va->va_start, align)); |
89699605f mm: vmap area cache |
1192 1193 |
BUG_ON(va->va_start < vstart); BUG_ON(va->va_end > vend); |
d98c9e83b kasan: fix crashe... |
1194 1195 1196 1197 1198 |
ret = kasan_populate_vmalloc(addr, size); if (ret) { free_vmap_area(va); return ERR_PTR(ret); } |
db64fe022 mm: rewrite vmap ... |
1199 |
return va; |
89699605f mm: vmap area cache |
1200 1201 |
overflow: |
89699605f mm: vmap area cache |
1202 1203 1204 1205 1206 |
if (!purged) { purge_vmap_area_lazy(); purged = 1; goto retry; } |
4da56b99d mm/vmap: Add a no... |
1207 1208 1209 1210 1211 1212 1213 1214 1215 |
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 mm: Silence vmap(... |
1216 |
if (!(gfp_mask & __GFP_NOWARN) && printk_ratelimit()) |
756a025f0 mm: coalesce spli... |
1217 1218 1219 |
pr_warn("vmap allocation for size %lu failed: use vmalloc=<size> to increase size ", size); |
68ad4a330 mm/vmalloc.c: kee... |
1220 1221 |
kmem_cache_free(vmap_area_cachep, va); |
89699605f mm: vmap area cache |
1222 |
return ERR_PTR(-EBUSY); |
db64fe022 mm: rewrite vmap ... |
1223 |
} |
4da56b99d mm/vmap: Add a no... |
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 |
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); |
e36176be1 mm/vmalloc: rewor... |
1235 |
/* |
db64fe022 mm: rewrite vmap ... |
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 |
* 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); } |
4d36e6f80 mm/vmalloc.c: con... |
1259 |
static atomic_long_t vmap_lazy_nr = ATOMIC_LONG_INIT(0); |
db64fe022 mm: rewrite vmap ... |
1260 |
|
0574ecd14 mm: refactor __pu... |
1261 1262 1263 1264 1265 |
/* * 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 mm: turn vmap_pur... |
1266 |
static DEFINE_MUTEX(vmap_purge_lock); |
0574ecd14 mm: refactor __pu... |
1267 |
|
02b709df8 mm: purge fragmen... |
1268 1269 |
/* for per-CPU blocks */ static void purge_fragmented_blocks_allcpus(void); |
db64fe022 mm: rewrite vmap ... |
1270 |
/* |
3ee48b6af mm, x86: Saving v... |
1271 1272 1273 1274 1275 |
* called before a call to iounmap() if the caller wants vm_area_struct's * immediately freed. */ void set_iounmap_nonlazy(void) { |
4d36e6f80 mm/vmalloc.c: con... |
1276 |
atomic_long_set(&vmap_lazy_nr, lazy_max_pages()+1); |
3ee48b6af mm, x86: Saving v... |
1277 1278 1279 |
} /* |
db64fe022 mm: rewrite vmap ... |
1280 |
* Purges all lazily-freed vmap areas. |
db64fe022 mm: rewrite vmap ... |
1281 |
*/ |
0574ecd14 mm: refactor __pu... |
1282 |
static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end) |
db64fe022 mm: rewrite vmap ... |
1283 |
{ |
4d36e6f80 mm/vmalloc.c: con... |
1284 |
unsigned long resched_threshold; |
80c4bd7a5 mm/vmalloc: keep ... |
1285 |
struct llist_node *valist; |
db64fe022 mm: rewrite vmap ... |
1286 |
struct vmap_area *va; |
cbb766766 mm: fix lazy vmap... |
1287 |
struct vmap_area *n_va; |
db64fe022 mm: rewrite vmap ... |
1288 |
|
0574ecd14 mm: refactor __pu... |
1289 |
lockdep_assert_held(&vmap_purge_lock); |
02b709df8 mm: purge fragmen... |
1290 |
|
80c4bd7a5 mm/vmalloc: keep ... |
1291 |
valist = llist_del_all(&vmap_purge_list); |
68571be99 mm/vmalloc.c: add... |
1292 1293 1294 1295 1296 1297 1298 |
if (unlikely(valist == NULL)) return false; /* * TODO: to calculate a flush range without looping. * The list can be up to lazy_max_pages() elements. */ |
80c4bd7a5 mm/vmalloc: keep ... |
1299 |
llist_for_each_entry(va, valist, purge_list) { |
0574ecd14 mm: refactor __pu... |
1300 1301 1302 1303 |
if (va->va_start < start) start = va->va_start; if (va->va_end > end) end = va->va_end; |
db64fe022 mm: rewrite vmap ... |
1304 |
} |
db64fe022 mm: rewrite vmap ... |
1305 |
|
0574ecd14 mm: refactor __pu... |
1306 |
flush_tlb_kernel_range(start, end); |
4d36e6f80 mm/vmalloc.c: con... |
1307 |
resched_threshold = lazy_max_pages() << 1; |
db64fe022 mm: rewrite vmap ... |
1308 |
|
e36176be1 mm/vmalloc: rewor... |
1309 |
spin_lock(&free_vmap_area_lock); |
763b218dd mm: add preempt p... |
1310 |
llist_for_each_entry_safe(va, n_va, valist, purge_list) { |
4d36e6f80 mm/vmalloc.c: con... |
1311 |
unsigned long nr = (va->va_end - va->va_start) >> PAGE_SHIFT; |
3c5c3cfb9 kasan: support ba... |
1312 1313 |
unsigned long orig_start = va->va_start; unsigned long orig_end = va->va_end; |
763b218dd mm: add preempt p... |
1314 |
|
dd3b8353b mm/vmalloc: do no... |
1315 1316 1317 1318 1319 |
/* * Finally insert or merge lazily-freed area. It is * detached and there is no need to "unlink" it from * anything. */ |
3c5c3cfb9 kasan: support ba... |
1320 1321 |
va = merge_or_add_vmap_area(va, &free_vmap_area_root, &free_vmap_area_list); |
9c801f61d mm/vmalloc.c: rem... |
1322 1323 |
if (!va) continue; |
3c5c3cfb9 kasan: support ba... |
1324 1325 1326 |
if (is_vmalloc_or_module_addr((void *)orig_start)) kasan_release_vmalloc(orig_start, orig_end, va->va_start, va->va_end); |
dd3b8353b mm/vmalloc: do no... |
1327 |
|
4d36e6f80 mm/vmalloc.c: con... |
1328 |
atomic_long_sub(nr, &vmap_lazy_nr); |
68571be99 mm/vmalloc.c: add... |
1329 |
|
4d36e6f80 mm/vmalloc.c: con... |
1330 |
if (atomic_long_read(&vmap_lazy_nr) < resched_threshold) |
e36176be1 mm/vmalloc: rewor... |
1331 |
cond_resched_lock(&free_vmap_area_lock); |
763b218dd mm: add preempt p... |
1332 |
} |
e36176be1 mm/vmalloc: rewor... |
1333 |
spin_unlock(&free_vmap_area_lock); |
0574ecd14 mm: refactor __pu... |
1334 |
return true; |
db64fe022 mm: rewrite vmap ... |
1335 1336 1337 |
} /* |
496850e5f mm: vmalloc failu... |
1338 1339 1340 1341 1342 |
* 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 mm: turn vmap_pur... |
1343 |
if (mutex_trylock(&vmap_purge_lock)) { |
0574ecd14 mm: refactor __pu... |
1344 |
__purge_vmap_area_lazy(ULONG_MAX, 0); |
f9e099776 mm: turn vmap_pur... |
1345 |
mutex_unlock(&vmap_purge_lock); |
0574ecd14 mm: refactor __pu... |
1346 |
} |
496850e5f mm: vmalloc failu... |
1347 1348 1349 |
} /* |
db64fe022 mm: rewrite vmap ... |
1350 1351 1352 1353 |
* Kick off a purge of the outstanding lazy areas. */ static void purge_vmap_area_lazy(void) { |
f9e099776 mm: turn vmap_pur... |
1354 |
mutex_lock(&vmap_purge_lock); |
0574ecd14 mm: refactor __pu... |
1355 1356 |
purge_fragmented_blocks_allcpus(); __purge_vmap_area_lazy(ULONG_MAX, 0); |
f9e099776 mm: turn vmap_pur... |
1357 |
mutex_unlock(&vmap_purge_lock); |
db64fe022 mm: rewrite vmap ... |
1358 1359 1360 |
} /* |
64141da58 vmalloc: eagerly ... |
1361 1362 1363 |
* 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 mm: rewrite vmap ... |
1364 |
*/ |
64141da58 vmalloc: eagerly ... |
1365 |
static void free_vmap_area_noflush(struct vmap_area *va) |
db64fe022 mm: rewrite vmap ... |
1366 |
{ |
4d36e6f80 mm/vmalloc.c: con... |
1367 |
unsigned long nr_lazy; |
80c4bd7a5 mm/vmalloc: keep ... |
1368 |
|
dd3b8353b mm/vmalloc: do no... |
1369 1370 1371 |
spin_lock(&vmap_area_lock); unlink_va(va, &vmap_area_root); spin_unlock(&vmap_area_lock); |
4d36e6f80 mm/vmalloc.c: con... |
1372 1373 |
nr_lazy = atomic_long_add_return((va->va_end - va->va_start) >> PAGE_SHIFT, &vmap_lazy_nr); |
80c4bd7a5 mm/vmalloc: keep ... |
1374 1375 1376 1377 1378 |
/* 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 mm: vmalloc failu... |
1379 |
try_purge_vmap_area_lazy(); |
db64fe022 mm: rewrite vmap ... |
1380 |
} |
b29acbdcf mm: vmalloc fix l... |
1381 1382 1383 1384 1385 1386 |
/* * 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); |
855e57a11 mm: remove unmap_... |
1387 |
unmap_kernel_range_noflush(va->va_start, va->va_end - va->va_start); |
8e57f8acb mm, debug_pageall... |
1388 |
if (debug_pagealloc_enabled_static()) |
82a2e924f mm: vmalloc: clea... |
1389 |
flush_tlb_kernel_range(va->va_start, va->va_end); |
c8eef01e2 mm: remove free_u... |
1390 |
free_vmap_area_noflush(va); |
b29acbdcf mm: vmalloc fix l... |
1391 |
} |
db64fe022 mm: rewrite vmap ... |
1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 |
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 mm: rewrite vmap ... |
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 |
/*** 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 mm: fix wrong vma... |
1425 1426 1427 1428 |
#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 mm: rewrite vmap ... |
1429 1430 1431 1432 1433 1434 |
#define VMAP_BLOCK_SIZE (VMAP_BBMAP_BITS * PAGE_SIZE) struct vmap_block_queue { spinlock_t lock; struct list_head free; |
db64fe022 mm: rewrite vmap ... |
1435 1436 1437 1438 1439 |
}; struct vmap_block { spinlock_t lock; struct vmap_area *va; |
db64fe022 mm: rewrite vmap ... |
1440 |
unsigned long free, dirty; |
7d61bfe8f mm/vmalloc: get r... |
1441 |
unsigned long dirty_min, dirty_max; /*< dirty range */ |
de5604231 mm: percpu-vmap f... |
1442 1443 |
struct list_head free_list; struct rcu_head rcu_head; |
02b709df8 mm: purge fragmen... |
1444 |
struct list_head purge; |
db64fe022 mm: rewrite vmap ... |
1445 1446 1447 1448 1449 1450 |
}; /* Queue of free and dirty vmap blocks, for allocation and flushing purposes */ static DEFINE_PER_CPU(struct vmap_block_queue, vmap_block_queue); /* |
0f14599c6 vmalloc: convert ... |
1451 |
* XArray of vmap blocks, indexed by address, to quickly find a vmap block |
db64fe022 mm: rewrite vmap ... |
1452 1453 1454 |
* 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. */ |
0f14599c6 vmalloc: convert ... |
1455 |
static DEFINE_XARRAY(vmap_blocks); |
db64fe022 mm: rewrite vmap ... |
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 |
/* * 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 mm/vmalloc: occup... |
1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 |
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 * |
a862f68a8 docs/core-api/mm:... |
1485 |
* Return: virtual address in a newly allocated block or ERR_PTR(-errno) |
cf725ce27 mm/vmalloc: occup... |
1486 1487 |
*/ static void *new_vmap_block(unsigned int order, gfp_t gfp_mask) |
db64fe022 mm: rewrite vmap ... |
1488 1489 1490 1491 1492 1493 |
{ struct vmap_block_queue *vbq; struct vmap_block *vb; struct vmap_area *va; unsigned long vb_idx; int node, err; |
cf725ce27 mm/vmalloc: occup... |
1494 |
void *vaddr; |
db64fe022 mm: rewrite vmap ... |
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 |
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 vmalloc: remove r... |
1506 |
if (IS_ERR(va)) { |
db64fe022 mm: rewrite vmap ... |
1507 |
kfree(vb); |
e7d863407 mm: use ERR_CAST |
1508 |
return ERR_CAST(va); |
db64fe022 mm: rewrite vmap ... |
1509 |
} |
cf725ce27 mm/vmalloc: occup... |
1510 |
vaddr = vmap_block_vaddr(va->va_start, 0); |
db64fe022 mm: rewrite vmap ... |
1511 1512 |
spin_lock_init(&vb->lock); vb->va = va; |
cf725ce27 mm/vmalloc: occup... |
1513 1514 1515 |
/* At least something should be left free */ BUG_ON(VMAP_BBMAP_BITS <= (1UL << order)); vb->free = VMAP_BBMAP_BITS - (1UL << order); |
db64fe022 mm: rewrite vmap ... |
1516 |
vb->dirty = 0; |
7d61bfe8f mm/vmalloc: get r... |
1517 1518 |
vb->dirty_min = VMAP_BBMAP_BITS; vb->dirty_max = 0; |
db64fe022 mm: rewrite vmap ... |
1519 |
INIT_LIST_HEAD(&vb->free_list); |
db64fe022 mm: rewrite vmap ... |
1520 1521 |
vb_idx = addr_to_vb_idx(va->va_start); |
0f14599c6 vmalloc: convert ... |
1522 1523 1524 1525 1526 1527 |
err = xa_insert(&vmap_blocks, vb_idx, vb, gfp_mask); if (err) { kfree(vb); free_vmap_area(va); return ERR_PTR(err); } |
db64fe022 mm: rewrite vmap ... |
1528 1529 |
vbq = &get_cpu_var(vmap_block_queue); |
db64fe022 mm: rewrite vmap ... |
1530 |
spin_lock(&vbq->lock); |
68ac546f2 mm/vmalloc: fix p... |
1531 |
list_add_tail_rcu(&vb->free_list, &vbq->free); |
db64fe022 mm: rewrite vmap ... |
1532 |
spin_unlock(&vbq->lock); |
3f04ba859 vmalloc: fix use ... |
1533 |
put_cpu_var(vmap_block_queue); |
db64fe022 mm: rewrite vmap ... |
1534 |
|
cf725ce27 mm/vmalloc: occup... |
1535 |
return vaddr; |
db64fe022 mm: rewrite vmap ... |
1536 |
} |
db64fe022 mm: rewrite vmap ... |
1537 1538 1539 |
static void free_vmap_block(struct vmap_block *vb) { struct vmap_block *tmp; |
db64fe022 mm: rewrite vmap ... |
1540 |
|
0f14599c6 vmalloc: convert ... |
1541 |
tmp = xa_erase(&vmap_blocks, addr_to_vb_idx(vb->va->va_start)); |
db64fe022 mm: rewrite vmap ... |
1542 |
BUG_ON(tmp != vb); |
64141da58 vmalloc: eagerly ... |
1543 |
free_vmap_area_noflush(vb->va); |
22a3c7d18 vmalloc,rcu: Conv... |
1544 |
kfree_rcu(vb, rcu_head); |
db64fe022 mm: rewrite vmap ... |
1545 |
} |
02b709df8 mm: purge fragmen... |
1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 |
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 mm/vmalloc: get r... |
1563 1564 |
vb->dirty_min = 0; vb->dirty_max = VMAP_BBMAP_BITS; |
02b709df8 mm: purge fragmen... |
1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 |
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 mm: purge fragmen... |
1580 1581 1582 1583 1584 1585 1586 |
static void purge_fragmented_blocks_allcpus(void) { int cpu; for_each_possible_cpu(cpu) purge_fragmented_blocks(cpu); } |
db64fe022 mm: rewrite vmap ... |
1587 1588 1589 1590 |
static void *vb_alloc(unsigned long size, gfp_t gfp_mask) { struct vmap_block_queue *vbq; struct vmap_block *vb; |
cf725ce27 mm/vmalloc: occup... |
1591 |
void *vaddr = NULL; |
db64fe022 mm: rewrite vmap ... |
1592 |
unsigned int order; |
891c49abf mm/vmalloc: use o... |
1593 |
BUG_ON(offset_in_page(size)); |
db64fe022 mm: rewrite vmap ... |
1594 |
BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC); |
aa91c4d89 mm: make vb_alloc... |
1595 1596 1597 1598 1599 1600 1601 1602 |
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 mm: rewrite vmap ... |
1603 |
order = get_order(size); |
db64fe022 mm: rewrite vmap ... |
1604 1605 1606 |
rcu_read_lock(); vbq = &get_cpu_var(vmap_block_queue); list_for_each_entry_rcu(vb, &vbq->free, free_list) { |
cf725ce27 mm/vmalloc: occup... |
1607 |
unsigned long pages_off; |
db64fe022 mm: rewrite vmap ... |
1608 1609 |
spin_lock(&vb->lock); |
cf725ce27 mm/vmalloc: occup... |
1610 1611 1612 1613 |
if (vb->free < (1UL << order)) { spin_unlock(&vb->lock); continue; } |
02b709df8 mm: purge fragmen... |
1614 |
|
cf725ce27 mm/vmalloc: occup... |
1615 1616 |
pages_off = VMAP_BBMAP_BITS - vb->free; vaddr = vmap_block_vaddr(vb->va->va_start, pages_off); |
02b709df8 mm: purge fragmen... |
1617 1618 1619 1620 1621 1622 |
vb->free -= 1UL << order; if (vb->free == 0) { spin_lock(&vbq->lock); list_del_rcu(&vb->free_list); spin_unlock(&vbq->lock); } |
cf725ce27 mm/vmalloc: occup... |
1623 |
|
02b709df8 mm: purge fragmen... |
1624 1625 |
spin_unlock(&vb->lock); break; |
db64fe022 mm: rewrite vmap ... |
1626 |
} |
02b709df8 mm: purge fragmen... |
1627 |
|
3f04ba859 vmalloc: fix use ... |
1628 |
put_cpu_var(vmap_block_queue); |
db64fe022 mm: rewrite vmap ... |
1629 |
rcu_read_unlock(); |
cf725ce27 mm/vmalloc: occup... |
1630 1631 1632 |
/* Allocate new block if nothing was found */ if (!vaddr) vaddr = new_vmap_block(order, gfp_mask); |
db64fe022 mm: rewrite vmap ... |
1633 |
|
cf725ce27 mm/vmalloc: occup... |
1634 |
return vaddr; |
db64fe022 mm: rewrite vmap ... |
1635 |
} |
78a0e8c48 mm: pass addr as ... |
1636 |
static void vb_free(unsigned long addr, unsigned long size) |
db64fe022 mm: rewrite vmap ... |
1637 1638 |
{ unsigned long offset; |
db64fe022 mm: rewrite vmap ... |
1639 1640 |
unsigned int order; struct vmap_block *vb; |
891c49abf mm/vmalloc: use o... |
1641 |
BUG_ON(offset_in_page(size)); |
db64fe022 mm: rewrite vmap ... |
1642 |
BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC); |
b29acbdcf mm: vmalloc fix l... |
1643 |
|
78a0e8c48 mm: pass addr as ... |
1644 |
flush_cache_vunmap(addr, addr + size); |
b29acbdcf mm: vmalloc fix l... |
1645 |
|
db64fe022 mm: rewrite vmap ... |
1646 |
order = get_order(size); |
78a0e8c48 mm: pass addr as ... |
1647 |
offset = (addr & (VMAP_BLOCK_SIZE - 1)) >> PAGE_SHIFT; |
0f14599c6 vmalloc: convert ... |
1648 |
vb = xa_load(&vmap_blocks, addr_to_vb_idx(addr)); |
db64fe022 mm: rewrite vmap ... |
1649 |
|
b521c43f5 mm: remove vmap_p... |
1650 |
unmap_kernel_range_noflush(addr, size); |
64141da58 vmalloc: eagerly ... |
1651 |
|
8e57f8acb mm, debug_pageall... |
1652 |
if (debug_pagealloc_enabled_static()) |
78a0e8c48 mm: pass addr as ... |
1653 |
flush_tlb_kernel_range(addr, addr + size); |
82a2e924f mm: vmalloc: clea... |
1654 |
|
db64fe022 mm: rewrite vmap ... |
1655 |
spin_lock(&vb->lock); |
7d61bfe8f mm/vmalloc: get r... |
1656 1657 1658 1659 |
/* Expand dirty range */ vb->dirty_min = min(vb->dirty_min, offset); vb->dirty_max = max(vb->dirty_max, offset + (1UL << order)); |
d086817dc vmap: remove need... |
1660 |
|
db64fe022 mm: rewrite vmap ... |
1661 1662 |
vb->dirty += 1UL << order; if (vb->dirty == VMAP_BBMAP_BITS) { |
de5604231 mm: percpu-vmap f... |
1663 |
BUG_ON(vb->free); |
db64fe022 mm: rewrite vmap ... |
1664 1665 1666 1667 1668 |
spin_unlock(&vb->lock); free_vmap_block(vb); } else spin_unlock(&vb->lock); } |
868b104d7 mm/vmalloc: Add f... |
1669 |
static void _vm_unmap_aliases(unsigned long start, unsigned long end, int flush) |
db64fe022 mm: rewrite vmap ... |
1670 |
{ |
db64fe022 mm: rewrite vmap ... |
1671 |
int cpu; |
db64fe022 mm: rewrite vmap ... |
1672 |
|
9b4633340 vmap: cope with v... |
1673 1674 |
if (unlikely(!vmap_initialized)) return; |
5803ed292 mm: mark all call... |
1675 |
might_sleep(); |
db64fe022 mm: rewrite vmap ... |
1676 1677 1678 1679 1680 1681 |
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 mm: rewrite vmap ... |
1682 |
spin_lock(&vb->lock); |
7d61bfe8f mm/vmalloc: get r... |
1683 1684 |
if (vb->dirty) { unsigned long va_start = vb->va->va_start; |
db64fe022 mm: rewrite vmap ... |
1685 |
unsigned long s, e; |
b136be5e0 mm, vmalloc: use ... |
1686 |
|
7d61bfe8f mm/vmalloc: get r... |
1687 1688 |
s = va_start + (vb->dirty_min << PAGE_SHIFT); e = va_start + (vb->dirty_max << PAGE_SHIFT); |
db64fe022 mm: rewrite vmap ... |
1689 |
|
7d61bfe8f mm/vmalloc: get r... |
1690 1691 |
start = min(s, start); end = max(e, end); |
db64fe022 mm: rewrite vmap ... |
1692 |
|
7d61bfe8f mm/vmalloc: get r... |
1693 |
flush = 1; |
db64fe022 mm: rewrite vmap ... |
1694 1695 1696 1697 1698 |
} spin_unlock(&vb->lock); } rcu_read_unlock(); } |
f9e099776 mm: turn vmap_pur... |
1699 |
mutex_lock(&vmap_purge_lock); |
0574ecd14 mm: refactor __pu... |
1700 1701 1702 |
purge_fragmented_blocks_allcpus(); if (!__purge_vmap_area_lazy(start, end) && flush) flush_tlb_kernel_range(start, end); |
f9e099776 mm: turn vmap_pur... |
1703 |
mutex_unlock(&vmap_purge_lock); |
db64fe022 mm: rewrite vmap ... |
1704 |
} |
868b104d7 mm/vmalloc: Add f... |
1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 |
/** * 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 flush = 0; _vm_unmap_aliases(start, end, flush); } |
db64fe022 mm: rewrite vmap ... |
1726 1727 1728 1729 1730 1731 1732 1733 1734 |
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 mm: fix overflow ... |
1735 |
unsigned long size = (unsigned long)count << PAGE_SHIFT; |
db64fe022 mm: rewrite vmap ... |
1736 |
unsigned long addr = (unsigned long)mem; |
9c3acf604 mm: remove free_u... |
1737 |
struct vmap_area *va; |
db64fe022 mm: rewrite vmap ... |
1738 |
|
5803ed292 mm: mark all call... |
1739 |
might_sleep(); |
db64fe022 mm: rewrite vmap ... |
1740 1741 1742 |
BUG_ON(!addr); BUG_ON(addr < VMALLOC_START); BUG_ON(addr > VMALLOC_END); |
a1c0b1a07 mm/vmalloc: use P... |
1743 |
BUG_ON(!PAGE_ALIGNED(addr)); |
db64fe022 mm: rewrite vmap ... |
1744 |
|
d98c9e83b kasan: fix crashe... |
1745 |
kasan_poison_vmalloc(mem, size); |
9c3acf604 mm: remove free_u... |
1746 |
if (likely(count <= VMAP_MAX_ALLOC)) { |
05e3ff950 mm: vmalloc: pass... |
1747 |
debug_check_no_locks_freed(mem, size); |
78a0e8c48 mm: pass addr as ... |
1748 |
vb_free(addr, size); |
9c3acf604 mm: remove free_u... |
1749 1750 1751 1752 1753 |
return; } va = find_vmap_area(addr); BUG_ON(!va); |
05e3ff950 mm: vmalloc: pass... |
1754 1755 |
debug_check_no_locks_freed((void *)va->va_start, (va->va_end - va->va_start)); |
9c3acf604 mm: remove free_u... |
1756 |
free_unmap_vmap_area(va); |
db64fe022 mm: rewrite vmap ... |
1757 1758 1759 1760 1761 1762 1763 1764 |
} 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 |
e99c97ade mm: fix kernel-do... |
1765 |
* |
364376383 mm/vmalloc.c: enh... |
1766 1767 1768 1769 1770 1771 |
* 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 mm: fix kernel-do... |
1772 |
* Returns: a pointer to the address that has been mapped, or %NULL on failure |
db64fe022 mm: rewrite vmap ... |
1773 |
*/ |
d4efd79a8 mm: remove the pr... |
1774 |
void *vm_map_ram(struct page **pages, unsigned int count, int node) |
db64fe022 mm: rewrite vmap ... |
1775 |
{ |
65ee03c4b mm: fix overflow ... |
1776 |
unsigned long size = (unsigned long)count << PAGE_SHIFT; |
db64fe022 mm: rewrite vmap ... |
1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 |
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; } |
d98c9e83b kasan: fix crashe... |
1795 1796 |
kasan_unpoison_vmalloc(mem, size); |
d4efd79a8 mm: remove the pr... |
1797 |
if (map_kernel_range(addr, size, PAGE_KERNEL, pages) < 0) { |
db64fe022 mm: rewrite vmap ... |
1798 1799 1800 1801 1802 1803 |
vm_unmap_ram(mem, count); return NULL; } return mem; } EXPORT_SYMBOL(vm_map_ram); |
4341fa454 mm, vmalloc: remo... |
1804 |
static struct vm_struct *vmlist __initdata; |
92eac1681 docs/mm: vmalloc:... |
1805 |
|
f0aa66179 vmalloc: implemen... |
1806 |
/** |
be9b7335e mm: add vm_area_a... |
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 |
* 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 vmalloc: implemen... |
1833 1834 |
* vm_area_register_early - register vmap area early during boot * @vm: vm_struct to register |
c0c0a2937 vmalloc: add @ali... |
1835 |
* @align: requested alignment |
f0aa66179 vmalloc: implemen... |
1836 1837 1838 1839 1840 1841 1842 1843 |
* * 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 vmalloc: add @ali... |
1844 |
void __init vm_area_register_early(struct vm_struct *vm, size_t align) |
f0aa66179 vmalloc: implemen... |
1845 1846 |
{ static size_t vm_init_off __initdata; |
c0c0a2937 vmalloc: add @ali... |
1847 1848 1849 1850 |
unsigned long addr; addr = ALIGN(VMALLOC_START + vm_init_off, align); vm_init_off = PFN_ALIGN(addr + vm->size) - VMALLOC_START; |
f0aa66179 vmalloc: implemen... |
1851 |
|
c0c0a2937 vmalloc: add @ali... |
1852 |
vm->addr = (void *)addr; |
f0aa66179 vmalloc: implemen... |
1853 |
|
be9b7335e mm: add vm_area_a... |
1854 |
vm_area_add_early(vm); |
f0aa66179 vmalloc: implemen... |
1855 |
} |
68ad4a330 mm/vmalloc.c: kee... |
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 |
static void vmap_init_free_space(void) { unsigned long vmap_start = 1; const unsigned long vmap_end = ULONG_MAX; struct vmap_area *busy, *free; /* * B F B B B F * -|-----|.....|-----|-----|-----|.....|- * | The KVA space | * |<--------------------------------->| */ list_for_each_entry(busy, &vmap_area_list, list) { if (busy->va_start - vmap_start > 0) { free = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT); if (!WARN_ON_ONCE(!free)) { free->va_start = vmap_start; free->va_end = busy->va_start; insert_vmap_area_augment(free, NULL, &free_vmap_area_root, &free_vmap_area_list); } } vmap_start = busy->va_end; } if (vmap_end - vmap_start > 0) { free = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT); if (!WARN_ON_ONCE(!free)) { free->va_start = vmap_start; free->va_end = vmap_end; insert_vmap_area_augment(free, NULL, &free_vmap_area_root, &free_vmap_area_list); } } } |
db64fe022 mm: rewrite vmap ... |
1896 1897 |
void __init vmalloc_init(void) { |
822c18f2e alpha: fix vmallo... |
1898 1899 |
struct vmap_area *va; struct vm_struct *tmp; |
db64fe022 mm: rewrite vmap ... |
1900 |
int i; |
68ad4a330 mm/vmalloc.c: kee... |
1901 1902 1903 1904 |
/* * Create the cache for vmap_area objects. */ vmap_area_cachep = KMEM_CACHE(vmap_area, SLAB_PANIC); |
db64fe022 mm: rewrite vmap ... |
1905 1906 |
for_each_possible_cpu(i) { struct vmap_block_queue *vbq; |
32fcfd407 make vfree() safe... |
1907 |
struct vfree_deferred *p; |
db64fe022 mm: rewrite vmap ... |
1908 1909 1910 1911 |
vbq = &per_cpu(vmap_block_queue, i); spin_lock_init(&vbq->lock); INIT_LIST_HEAD(&vbq->free); |
32fcfd407 make vfree() safe... |
1912 1913 1914 |
p = &per_cpu(vfree_deferred, i); init_llist_head(&p->list); INIT_WORK(&p->wq, free_work); |
db64fe022 mm: rewrite vmap ... |
1915 |
} |
9b4633340 vmap: cope with v... |
1916 |
|
822c18f2e alpha: fix vmallo... |
1917 1918 |
/* Import existing vmlist entries. */ for (tmp = vmlist; tmp; tmp = tmp->next) { |
68ad4a330 mm/vmalloc.c: kee... |
1919 1920 1921 |
va = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT); if (WARN_ON_ONCE(!va)) continue; |
822c18f2e alpha: fix vmallo... |
1922 1923 |
va->va_start = (unsigned long)tmp->addr; va->va_end = va->va_start + tmp->size; |
dbda591d9 mm: fix faulty in... |
1924 |
va->vm = tmp; |
68ad4a330 mm/vmalloc.c: kee... |
1925 |
insert_vmap_area(va, &vmap_area_root, &vmap_area_list); |
822c18f2e alpha: fix vmallo... |
1926 |
} |
ca23e405e vmalloc: implemen... |
1927 |
|
68ad4a330 mm/vmalloc.c: kee... |
1928 1929 1930 1931 |
/* * Now we can initialize a free vmap space. */ vmap_init_free_space(); |
9b4633340 vmap: cope with v... |
1932 |
vmap_initialized = true; |
db64fe022 mm: rewrite vmap ... |
1933 |
} |
8fc489850 vmalloc: add un/m... |
1934 |
/** |
8fc489850 vmalloc: add un/m... |
1935 1936 1937 1938 1939 1940 1941 |
* 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 mm: rewrite vmap ... |
1942 1943 1944 |
void unmap_kernel_range(unsigned long addr, unsigned long size) { unsigned long end = addr + size; |
f6fcba701 vmalloc: call flu... |
1945 1946 |
flush_cache_vunmap(addr, end); |
b521c43f5 mm: remove vmap_p... |
1947 |
unmap_kernel_range_noflush(addr, size); |
db64fe022 mm: rewrite vmap ... |
1948 1949 |
flush_tlb_kernel_range(addr, end); } |
e36176be1 mm/vmalloc: rewor... |
1950 1951 |
static inline void setup_vmalloc_vm_locked(struct vm_struct *vm, struct vmap_area *va, unsigned long flags, const void *caller) |
cf88c7900 vmalloc: separate... |
1952 |
{ |
cf88c7900 vmalloc: separate... |
1953 1954 1955 1956 |
vm->flags = flags; vm->addr = (void *)va->va_start; vm->size = va->va_end - va->va_start; vm->caller = caller; |
db1aecafe mm/vmalloc.c: cha... |
1957 |
va->vm = vm; |
e36176be1 mm/vmalloc: rewor... |
1958 1959 1960 1961 1962 1963 1964 |
} static void setup_vmalloc_vm(struct vm_struct *vm, struct vmap_area *va, unsigned long flags, const void *caller) { spin_lock(&vmap_area_lock); setup_vmalloc_vm_locked(vm, va, flags, caller); |
c69480ade mm, vmalloc: prot... |
1965 |
spin_unlock(&vmap_area_lock); |
f5252e009 mm: avoid null po... |
1966 |
} |
cf88c7900 vmalloc: separate... |
1967 |
|
20fc02b47 mm/vmalloc.c: ren... |
1968 |
static void clear_vm_uninitialized_flag(struct vm_struct *vm) |
f5252e009 mm: avoid null po... |
1969 |
{ |
d4033afdf mm, vmalloc: iter... |
1970 |
/* |
20fc02b47 mm/vmalloc.c: ren... |
1971 |
* Before removing VM_UNINITIALIZED, |
d4033afdf mm, vmalloc: iter... |
1972 1973 1974 1975 |
* we should make sure that vm has proper values. * Pair with smp_rmb() in show_numa_info(). */ smp_wmb(); |
20fc02b47 mm/vmalloc.c: ren... |
1976 |
vm->flags &= ~VM_UNINITIALIZED; |
cf88c7900 vmalloc: separate... |
1977 |
} |
db64fe022 mm: rewrite vmap ... |
1978 |
static struct vm_struct *__get_vm_area_node(unsigned long size, |
2dca6999e mm, perf_event: M... |
1979 |
unsigned long align, unsigned long flags, unsigned long start, |
5e6cafc83 mm: vmalloc: use ... |
1980 |
unsigned long end, int node, gfp_t gfp_mask, const void *caller) |
db64fe022 mm: rewrite vmap ... |
1981 |
{ |
0006526d7 mm/vmalloc.c: rem... |
1982 |
struct vmap_area *va; |
db64fe022 mm: rewrite vmap ... |
1983 |
struct vm_struct *area; |
d98c9e83b kasan: fix crashe... |
1984 |
unsigned long requested_size = size; |
1da177e4c Linux-2.6.12-rc2 |
1985 |
|
52fd24ca1 [PATCH] __vmalloc... |
1986 |
BUG_ON(in_interrupt()); |
1da177e4c Linux-2.6.12-rc2 |
1987 |
size = PAGE_ALIGN(size); |
31be83095 [PATCH] Fix stran... |
1988 1989 |
if (unlikely(!size)) return NULL; |
1da177e4c Linux-2.6.12-rc2 |
1990 |
|
252e5c6e2 mm/vmalloc.c: fix... |
1991 1992 1993 |
if (flags & VM_IOREMAP) align = 1ul << clamp_t(int, get_count_order_long(size), PAGE_SHIFT, IOREMAP_MAX_ORDER); |
cf88c7900 vmalloc: separate... |
1994 |
area = kzalloc_node(sizeof(*area), gfp_mask & GFP_RECLAIM_MASK, node); |
1da177e4c Linux-2.6.12-rc2 |
1995 1996 |
if (unlikely(!area)) return NULL; |
71394fe50 mm: vmalloc: add ... |
1997 1998 |
if (!(flags & VM_NO_GUARD)) size += PAGE_SIZE; |
1da177e4c Linux-2.6.12-rc2 |
1999 |
|
db64fe022 mm: rewrite vmap ... |
2000 2001 2002 2003 |
va = alloc_vmap_area(size, align, start, end, node, gfp_mask); if (IS_ERR(va)) { kfree(area); return NULL; |
1da177e4c Linux-2.6.12-rc2 |
2004 |
} |
1da177e4c Linux-2.6.12-rc2 |
2005 |
|
d98c9e83b kasan: fix crashe... |
2006 |
kasan_unpoison_vmalloc((void *)va->va_start, requested_size); |
f5252e009 mm: avoid null po... |
2007 |
|
d98c9e83b kasan: fix crashe... |
2008 |
setup_vmalloc_vm(area, va, flags, caller); |
3c5c3cfb9 kasan: support ba... |
2009 |
|
1da177e4c Linux-2.6.12-rc2 |
2010 |
return area; |
1da177e4c Linux-2.6.12-rc2 |
2011 |
} |
c29686129 vmalloc: add __ge... |
2012 2013 |
struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags, unsigned long start, unsigned long end, |
5e6cafc83 mm: vmalloc: use ... |
2014 |
const void *caller) |
c29686129 vmalloc: add __ge... |
2015 |
{ |
00ef2d2f8 mm: use NUMA_NO_NODE |
2016 2017 |
return __get_vm_area_node(size, 1, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, caller); |
c29686129 vmalloc: add __ge... |
2018 |
} |
1ac80b081 mm: vmalloc: Expo... |
2019 |
EXPORT_SYMBOL_GPL(__get_vm_area_caller); |
c29686129 vmalloc: add __ge... |
2020 |
|
1da177e4c Linux-2.6.12-rc2 |
2021 |
/** |
92eac1681 docs/mm: vmalloc:... |
2022 2023 2024 |
* get_vm_area - reserve a contiguous kernel virtual area * @size: size of the area * @flags: %VM_IOREMAP for I/O mappings or VM_ALLOC |
1da177e4c Linux-2.6.12-rc2 |
2025 |
* |
92eac1681 docs/mm: vmalloc:... |
2026 2027 2028 |
* 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. |
a862f68a8 docs/core-api/mm:... |
2029 2030 |
* * Return: the area descriptor on success or %NULL on failure. |
1da177e4c Linux-2.6.12-rc2 |
2031 2032 2033 |
*/ struct vm_struct *get_vm_area(unsigned long size, unsigned long flags) { |
2dca6999e mm, perf_event: M... |
2034 |
return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END, |
00ef2d2f8 mm: use NUMA_NO_NODE |
2035 2036 |
NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); |
230169693 vmallocinfo: add ... |
2037 2038 2039 |
} struct vm_struct *get_vm_area_caller(unsigned long size, unsigned long flags, |
5e6cafc83 mm: vmalloc: use ... |
2040 |
const void *caller) |
230169693 vmallocinfo: add ... |
2041 |
{ |
2dca6999e mm, perf_event: M... |
2042 |
return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END, |
00ef2d2f8 mm: use NUMA_NO_NODE |
2043 |
NUMA_NO_NODE, GFP_KERNEL, caller); |
1da177e4c Linux-2.6.12-rc2 |
2044 |
} |
e9da6e990 ARM: dma-mapping:... |
2045 |
/** |
92eac1681 docs/mm: vmalloc:... |
2046 2047 |
* find_vm_area - find a continuous kernel virtual area * @addr: base address |
e9da6e990 ARM: dma-mapping:... |
2048 |
* |
92eac1681 docs/mm: vmalloc:... |
2049 2050 2051 |
* 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. |
a862f68a8 docs/core-api/mm:... |
2052 |
* |
74640617e mm/vmalloc.c: fix... |
2053 |
* Return: the area descriptor on success or %NULL on failure. |
e9da6e990 ARM: dma-mapping:... |
2054 2055 |
*/ struct vm_struct *find_vm_area(const void *addr) |
833423143 [PATCH] mm: intro... |
2056 |
{ |
db64fe022 mm: rewrite vmap ... |
2057 |
struct vmap_area *va; |
833423143 [PATCH] mm: intro... |
2058 |
|
db64fe022 mm: rewrite vmap ... |
2059 |
va = find_vmap_area((unsigned long)addr); |
688fcbfc0 mm/vmalloc: modif... |
2060 2061 |
if (!va) return NULL; |
1da177e4c Linux-2.6.12-rc2 |
2062 |
|
688fcbfc0 mm/vmalloc: modif... |
2063 |
return va->vm; |
1da177e4c Linux-2.6.12-rc2 |
2064 |
} |
7856dfeb2 [PATCH] x86_64: F... |
2065 |
/** |
92eac1681 docs/mm: vmalloc:... |
2066 2067 |
* remove_vm_area - find and remove a continuous kernel virtual area * @addr: base address |
7856dfeb2 [PATCH] x86_64: F... |
2068 |
* |
92eac1681 docs/mm: vmalloc:... |
2069 2070 2071 |
* 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. |
a862f68a8 docs/core-api/mm:... |
2072 |
* |
74640617e mm/vmalloc.c: fix... |
2073 |
* Return: the area descriptor on success or %NULL on failure. |
7856dfeb2 [PATCH] x86_64: F... |
2074 |
*/ |
b3bdda02a vmalloc: add cons... |
2075 |
struct vm_struct *remove_vm_area(const void *addr) |
7856dfeb2 [PATCH] x86_64: F... |
2076 |
{ |
db64fe022 mm: rewrite vmap ... |
2077 |
struct vmap_area *va; |
5803ed292 mm: mark all call... |
2078 |
might_sleep(); |
dd3b8353b mm/vmalloc: do no... |
2079 2080 |
spin_lock(&vmap_area_lock); va = __find_vmap_area((unsigned long)addr); |
688fcbfc0 mm/vmalloc: modif... |
2081 |
if (va && va->vm) { |
db1aecafe mm/vmalloc.c: cha... |
2082 |
struct vm_struct *vm = va->vm; |
f5252e009 mm: avoid null po... |
2083 |
|
c69480ade mm, vmalloc: prot... |
2084 |
va->vm = NULL; |
c69480ade mm, vmalloc: prot... |
2085 |
spin_unlock(&vmap_area_lock); |
a5af5aa8b kasan, module, vm... |
2086 |
kasan_free_shadow(vm); |
dd32c2799 vmalloc: unmap vm... |
2087 |
free_unmap_vmap_area(va); |
dd32c2799 vmalloc: unmap vm... |
2088 |
|
db64fe022 mm: rewrite vmap ... |
2089 2090 |
return vm; } |
dd3b8353b mm/vmalloc: do no... |
2091 2092 |
spin_unlock(&vmap_area_lock); |
db64fe022 mm: rewrite vmap ... |
2093 |
return NULL; |
7856dfeb2 [PATCH] x86_64: F... |
2094 |
} |
868b104d7 mm/vmalloc: Add f... |
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 |
static inline void set_area_direct_map(const struct vm_struct *area, int (*set_direct_map)(struct page *page)) { int i; for (i = 0; i < area->nr_pages; i++) if (page_address(area->pages[i])) set_direct_map(area->pages[i]); } /* Handle removing and resetting vm mappings related to the vm_struct. */ static void vm_remove_mappings(struct vm_struct *area, int deallocate_pages) { |
868b104d7 mm/vmalloc: Add f... |
2108 2109 |
unsigned long start = ULONG_MAX, end = 0; int flush_reset = area->flags & VM_FLUSH_RESET_PERMS; |
31e67340c mm/vmalloc: Avoid... |
2110 |
int flush_dmap = 0; |
868b104d7 mm/vmalloc: Add f... |
2111 |
int i; |
868b104d7 mm/vmalloc: Add f... |
2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 |
remove_vm_area(area->addr); /* If this is not VM_FLUSH_RESET_PERMS memory, no need for the below. */ if (!flush_reset) return; /* * If not deallocating pages, just do the flush of the VM area and * return. */ if (!deallocate_pages) { vm_unmap_aliases(); return; } /* * If execution gets here, flush the vm mapping and reset the direct * map. Find the start and end range of the direct mappings to make sure * the vm_unmap_aliases() flush includes the direct map. */ for (i = 0; i < area->nr_pages; i++) { |
8e41f8726 mm/vmalloc: Fix c... |
2133 2134 |
unsigned long addr = (unsigned long)page_address(area->pages[i]); if (addr) { |
868b104d7 mm/vmalloc: Add f... |
2135 |
start = min(addr, start); |
8e41f8726 mm/vmalloc: Fix c... |
2136 |
end = max(addr + PAGE_SIZE, end); |
31e67340c mm/vmalloc: Avoid... |
2137 |
flush_dmap = 1; |
868b104d7 mm/vmalloc: Add f... |
2138 2139 2140 2141 2142 2143 2144 2145 2146 |
} } /* * Set direct map to something invalid so that it won't be cached if * there are any accesses after the TLB flush, then flush the TLB and * reset the direct map permissions to the default. */ set_area_direct_map(area, set_direct_map_invalid_noflush); |
31e67340c mm/vmalloc: Avoid... |
2147 |
_vm_unmap_aliases(start, end, flush_dmap); |
868b104d7 mm/vmalloc: Add f... |
2148 2149 |
set_area_direct_map(area, set_direct_map_default_noflush); } |
b3bdda02a vmalloc: add cons... |
2150 |
static void __vunmap(const void *addr, int deallocate_pages) |
1da177e4c Linux-2.6.12-rc2 |
2151 2152 2153 2154 2155 |
{ struct vm_struct *area; if (!addr) return; |
e69e9d4ae vmalloc: introduc... |
2156 2157 |
if (WARN(!PAGE_ALIGNED(addr), "Trying to vfree() bad address (%p) ", |
ab15d9b4c mm/vmalloc.c: unb... |
2158 |
addr)) |
1da177e4c Linux-2.6.12-rc2 |
2159 |
return; |
1da177e4c Linux-2.6.12-rc2 |
2160 |
|
6ade20327 mm/vmalloc.c: don... |
2161 |
area = find_vm_area(addr); |
1da177e4c Linux-2.6.12-rc2 |
2162 |
if (unlikely(!area)) { |
4c8573e25 Use WARN() in mm/... |
2163 2164 |
WARN(1, KERN_ERR "Trying to vfree() nonexistent vm area (%p) ", |
1da177e4c Linux-2.6.12-rc2 |
2165 |
addr); |
1da177e4c Linux-2.6.12-rc2 |
2166 2167 |
return; } |
05e3ff950 mm: vmalloc: pass... |
2168 2169 |
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 [PATCH] lockdep: ... |
2170 |
|
23713b480 mm/vmalloc.c: fix... |
2171 |
kasan_poison_vmalloc(area->addr, get_vm_area_size(area)); |
3c5c3cfb9 kasan: support ba... |
2172 |
|
868b104d7 mm/vmalloc: Add f... |
2173 |
vm_remove_mappings(area, deallocate_pages); |
1da177e4c Linux-2.6.12-rc2 |
2174 2175 2176 2177 |
if (deallocate_pages) { int i; for (i = 0; i < area->nr_pages; i++) { |
bf53d6f8f vmalloc: clean up... |
2178 2179 2180 |
struct page *page = area->pages[i]; BUG_ON(!page); |
4949148ad mm: charge/unchar... |
2181 |
__free_pages(page, 0); |
1da177e4c Linux-2.6.12-rc2 |
2182 |
} |
97105f0ab mm: vmalloc: show... |
2183 |
atomic_long_sub(area->nr_pages, &nr_vmalloc_pages); |
1da177e4c Linux-2.6.12-rc2 |
2184 |
|
244d63ee3 mm, vmalloc: remo... |
2185 |
kvfree(area->pages); |
1da177e4c Linux-2.6.12-rc2 |
2186 2187 2188 2189 2190 |
} kfree(area); return; } |
bf22e37a6 mm: add vfree_ato... |
2191 2192 2193 2194 2195 2196 2197 |
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 |
73221d888 mm/vmalloc: fix a... |
2198 |
* another cpu's list. schedule_work() should be fine with this too. |
bf22e37a6 mm: add vfree_ato... |
2199 2200 2201 2202 2203 2204 2205 2206 |
*/ struct vfree_deferred *p = raw_cpu_ptr(&vfree_deferred); if (llist_add((struct llist_node *)addr, &p->list)) schedule_work(&p->wq); } /** |
92eac1681 docs/mm: vmalloc:... |
2207 2208 |
* vfree_atomic - release memory allocated by vmalloc() * @addr: memory base address |
bf22e37a6 mm: add vfree_ato... |
2209 |
* |
92eac1681 docs/mm: vmalloc:... |
2210 2211 |
* This one is just like vfree() but can be called in any atomic context * except NMIs. |
bf22e37a6 mm: add vfree_ato... |
2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 |
*/ void vfree_atomic(const void *addr) { BUG_ON(in_nmi()); kmemleak_free(addr); if (!addr) return; __vfree_deferred(addr); } |
c67dc6247 mm/vmalloc: do no... |
2223 2224 2225 2226 2227 2228 2229 |
static void __vfree(const void *addr) { if (unlikely(in_interrupt())) __vfree_deferred(addr); else __vunmap(addr, 1); } |
1da177e4c Linux-2.6.12-rc2 |
2230 |
/** |
fa307474c mm: update the do... |
2231 2232 |
* vfree - Release memory allocated by vmalloc() * @addr: Memory base address |
1da177e4c Linux-2.6.12-rc2 |
2233 |
* |
fa307474c mm: update the do... |
2234 2235 2236 2237 |
* Free the virtually continuous memory area starting at @addr, as obtained * from one of the vmalloc() family of APIs. This will usually also free the * physical memory underlying the virtual allocation, but that memory is * reference counted, so it will not be freed until the last user goes away. |
1da177e4c Linux-2.6.12-rc2 |
2238 |
* |
fa307474c mm: update the do... |
2239 |
* If @addr is NULL, no operation is performed. |
c9fcee513 mm/vmalloc.c: add... |
2240 |
* |
fa307474c mm: update the do... |
2241 |
* Context: |
92eac1681 docs/mm: vmalloc:... |
2242 |
* May sleep if called *not* from interrupt context. |
fa307474c mm: update the do... |
2243 2244 2245 |
* Must not be called in NMI context (strictly speaking, it could be * if we have CONFIG_ARCH_HAVE_NMI_SAFE_CMPXCHG, but making the calling * conventions for vfree() arch-depenedent would be a really bad idea). |
1da177e4c Linux-2.6.12-rc2 |
2246 |
*/ |
b3bdda02a vmalloc: add cons... |
2247 |
void vfree(const void *addr) |
1da177e4c Linux-2.6.12-rc2 |
2248 |
{ |
32fcfd407 make vfree() safe... |
2249 |
BUG_ON(in_nmi()); |
89219d37a kmemleak: Add the... |
2250 2251 |
kmemleak_free(addr); |
a8dda165e vfree: add debug ... |
2252 |
might_sleep_if(!in_interrupt()); |
32fcfd407 make vfree() safe... |
2253 2254 |
if (!addr) return; |
c67dc6247 mm/vmalloc: do no... |
2255 2256 |
__vfree(addr); |
1da177e4c Linux-2.6.12-rc2 |
2257 |
} |
1da177e4c Linux-2.6.12-rc2 |
2258 2259 2260 |
EXPORT_SYMBOL(vfree); /** |
92eac1681 docs/mm: vmalloc:... |
2261 2262 |
* vunmap - release virtual mapping obtained by vmap() * @addr: memory base address |
1da177e4c Linux-2.6.12-rc2 |
2263 |
* |
92eac1681 docs/mm: vmalloc:... |
2264 2265 |
* Free the virtually contiguous memory area starting at @addr, * which was created from the page array passed to vmap(). |
1da177e4c Linux-2.6.12-rc2 |
2266 |
* |
92eac1681 docs/mm: vmalloc:... |
2267 |
* Must not be called in interrupt context. |
1da177e4c Linux-2.6.12-rc2 |
2268 |
*/ |
b3bdda02a vmalloc: add cons... |
2269 |
void vunmap(const void *addr) |
1da177e4c Linux-2.6.12-rc2 |
2270 2271 |
{ BUG_ON(in_interrupt()); |
34754b69a x86: make vmap ye... |
2272 |
might_sleep(); |
32fcfd407 make vfree() safe... |
2273 2274 |
if (addr) __vunmap(addr, 0); |
1da177e4c Linux-2.6.12-rc2 |
2275 |
} |
1da177e4c Linux-2.6.12-rc2 |
2276 2277 2278 |
EXPORT_SYMBOL(vunmap); /** |
92eac1681 docs/mm: vmalloc:... |
2279 2280 2281 2282 2283 2284 |
* vmap - map an array of pages into virtually contiguous space * @pages: array of page pointers * @count: number of pages to map * @flags: vm_area->flags * @prot: page protection for the mapping * |
b944afc9d mm: add a VM_MAP_... |
2285 2286 2287 2288 2289 |
* Maps @count pages from @pages into contiguous kernel virtual space. * If @flags contains %VM_MAP_PUT_PAGES the ownership of the pages array itself * (which must be kmalloc or vmalloc memory) and one reference per pages in it * are transferred from the caller to vmap(), and will be freed / dropped when * vfree() is called on the return value. |
a862f68a8 docs/core-api/mm:... |
2290 2291 |
* * Return: the address of the area or %NULL on failure |
1da177e4c Linux-2.6.12-rc2 |
2292 2293 |
*/ void *vmap(struct page **pages, unsigned int count, |
92eac1681 docs/mm: vmalloc:... |
2294 |
unsigned long flags, pgprot_t prot) |
1da177e4c Linux-2.6.12-rc2 |
2295 2296 |
{ struct vm_struct *area; |
65ee03c4b mm: fix overflow ... |
2297 |
unsigned long size; /* In bytes */ |
1da177e4c Linux-2.6.12-rc2 |
2298 |
|
34754b69a x86: make vmap ye... |
2299 |
might_sleep(); |
ca79b0c21 mm: convert total... |
2300 |
if (count > totalram_pages()) |
1da177e4c Linux-2.6.12-rc2 |
2301 |
return NULL; |
65ee03c4b mm: fix overflow ... |
2302 2303 |
size = (unsigned long)count << PAGE_SHIFT; area = get_vm_area_caller(size, flags, __builtin_return_address(0)); |
1da177e4c Linux-2.6.12-rc2 |
2304 2305 |
if (!area) return NULL; |
230169693 vmallocinfo: add ... |
2306 |
|
cca98e9f8 mm: enforce that ... |
2307 |
if (map_kernel_range((unsigned long)area->addr, size, pgprot_nx(prot), |
ed1f324c5 mm: remove map_vm... |
2308 |
pages) < 0) { |
1da177e4c Linux-2.6.12-rc2 |
2309 2310 2311 |
vunmap(area->addr); return NULL; } |
b4ecc2596 mm/vmalloc.c: fix... |
2312 |
if (flags & VM_MAP_PUT_PAGES) { |
b944afc9d mm: add a VM_MAP_... |
2313 |
area->pages = pages; |
b4ecc2596 mm/vmalloc.c: fix... |
2314 2315 |
area->nr_pages = count; } |
1da177e4c Linux-2.6.12-rc2 |
2316 2317 |
return area->addr; } |
1da177e4c Linux-2.6.12-rc2 |
2318 |
EXPORT_SYMBOL(vmap); |
3e9a9e256 mm: add a vmap_pf... |
2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 |
#ifdef CONFIG_VMAP_PFN struct vmap_pfn_data { unsigned long *pfns; pgprot_t prot; unsigned int idx; }; static int vmap_pfn_apply(pte_t *pte, unsigned long addr, void *private) { struct vmap_pfn_data *data = private; if (WARN_ON_ONCE(pfn_valid(data->pfns[data->idx]))) return -EINVAL; *pte = pte_mkspecial(pfn_pte(data->pfns[data->idx++], data->prot)); return 0; } /** * vmap_pfn - map an array of PFNs into virtually contiguous space * @pfns: array of PFNs * @count: number of pages to map * @prot: page protection for the mapping * * Maps @count PFNs from @pfns into contiguous kernel virtual space and returns * the start address of the mapping. */ void *vmap_pfn(unsigned long *pfns, unsigned int count, pgprot_t prot) { struct vmap_pfn_data data = { .pfns = pfns, .prot = pgprot_nx(prot) }; struct vm_struct *area; area = get_vm_area_caller(count * PAGE_SIZE, VM_IOREMAP, __builtin_return_address(0)); if (!area) return NULL; if (apply_to_page_range(&init_mm, (unsigned long)area->addr, count * PAGE_SIZE, vmap_pfn_apply, &data)) { free_vm_area(area); return NULL; } return area->addr; } EXPORT_SYMBOL_GPL(vmap_pfn); #endif /* CONFIG_VMAP_PFN */ |
e31d9eb5c make __vmalloc_ar... |
2363 |
static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, |
3722e13cf mm/vmalloc: don't... |
2364 |
pgprot_t prot, int node) |
1da177e4c Linux-2.6.12-rc2 |
2365 |
{ |
930f036b4 mm, vmalloc: cons... |
2366 |
const gfp_t nested_gfp = (gfp_mask & GFP_RECLAIM_MASK) | __GFP_ZERO; |
f255935b9 mm: cleanup the g... |
2367 2368 2369 |
unsigned int nr_pages = get_vm_area_size(area) >> PAGE_SHIFT; unsigned int array_size = nr_pages * sizeof(struct page *), i; struct page **pages; |
1da177e4c Linux-2.6.12-rc2 |
2370 |
|
f255935b9 mm: cleanup the g... |
2371 2372 2373 |
gfp_mask |= __GFP_NOWARN; if (!(gfp_mask & (GFP_DMA | GFP_DMA32))) gfp_mask |= __GFP_HIGHMEM; |
1da177e4c Linux-2.6.12-rc2 |
2374 |
|
1da177e4c Linux-2.6.12-rc2 |
2375 |
/* Please note that the recursion is strictly bounded. */ |
8757d5fa6 [PATCH] mm: fix o... |
2376 |
if (array_size > PAGE_SIZE) { |
f255935b9 mm: cleanup the g... |
2377 2378 |
pages = __vmalloc_node(array_size, 1, nested_gfp, node, area->caller); |
286e1ea3a [PATCH] vmalloc()... |
2379 |
} else { |
976d6dfbb vmalloc(): adjust... |
2380 |
pages = kmalloc_node(array_size, nested_gfp, node); |
286e1ea3a [PATCH] vmalloc()... |
2381 |
} |
7ea362427 mm/vmalloc.c: mov... |
2382 2383 |
if (!pages) { |
1da177e4c Linux-2.6.12-rc2 |
2384 2385 2386 2387 |
remove_vm_area(area->addr); kfree(area); return NULL; } |
1da177e4c Linux-2.6.12-rc2 |
2388 |
|
7ea362427 mm/vmalloc.c: mov... |
2389 2390 |
area->pages = pages; area->nr_pages = nr_pages; |
1da177e4c Linux-2.6.12-rc2 |
2391 |
for (i = 0; i < area->nr_pages; i++) { |
bf53d6f8f vmalloc: clean up... |
2392 |
struct page *page; |
4b90951c0 mm/vmalloc: use N... |
2393 |
if (node == NUMA_NO_NODE) |
f255935b9 mm: cleanup the g... |
2394 |
page = alloc_page(gfp_mask); |
930fc45a4 [PATCH] vmalloc_node |
2395 |
else |
f255935b9 mm: cleanup the g... |
2396 |
page = alloc_pages_node(node, gfp_mask, 0); |
bf53d6f8f vmalloc: clean up... |
2397 2398 |
if (unlikely(!page)) { |
82afbc32f mm/vmalloc.c: upd... |
2399 |
/* Successfully allocated i pages, free them in __vfree() */ |
1da177e4c Linux-2.6.12-rc2 |
2400 |
area->nr_pages = i; |
97105f0ab mm: vmalloc: show... |
2401 |
atomic_long_add(area->nr_pages, &nr_vmalloc_pages); |
1da177e4c Linux-2.6.12-rc2 |
2402 2403 |
goto fail; } |
bf53d6f8f vmalloc: clean up... |
2404 |
area->pages[i] = page; |
dcf61ff06 mm/vmalloc.c: rem... |
2405 |
if (gfpflags_allow_blocking(gfp_mask)) |
660654f90 mm/vmalloc.c: add... |
2406 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
2407 |
} |
97105f0ab mm: vmalloc: show... |
2408 |
atomic_long_add(area->nr_pages, &nr_vmalloc_pages); |
1da177e4c Linux-2.6.12-rc2 |
2409 |
|
ed1f324c5 mm: remove map_vm... |
2410 2411 |
if (map_kernel_range((unsigned long)area->addr, get_vm_area_size(area), prot, pages) < 0) |
1da177e4c Linux-2.6.12-rc2 |
2412 |
goto fail; |
ed1f324c5 mm: remove map_vm... |
2413 |
|
1da177e4c Linux-2.6.12-rc2 |
2414 2415 2416 |
return area->addr; fail: |
a8e99259e mm, page_alloc: w... |
2417 |
warn_alloc(gfp_mask, NULL, |
7877cdcc3 mm: consolidate w... |
2418 |
"vmalloc: allocation failure, allocated %ld of %ld bytes", |
22943ab11 mm: print vmalloc... |
2419 |
(area->nr_pages*PAGE_SIZE), area->size); |
c67dc6247 mm/vmalloc: do no... |
2420 |
__vfree(area->addr); |
1da177e4c Linux-2.6.12-rc2 |
2421 2422 2423 2424 |
return NULL; } /** |
92eac1681 docs/mm: vmalloc:... |
2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 |
* __vmalloc_node_range - allocate virtually contiguous memory * @size: allocation size * @align: desired alignment * @start: vm area range start * @end: vm area range end * @gfp_mask: flags for the page level allocator * @prot: protection mask for the allocated pages * @vm_flags: additional vm area flags (e.g. %VM_NO_GUARD) * @node: node to use for allocation or NUMA_NO_NODE * @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. |
a862f68a8 docs/core-api/mm:... |
2439 2440 |
* * Return: the address of the area or %NULL on failure |
1da177e4c Linux-2.6.12-rc2 |
2441 |
*/ |
d0a21265d mm: unify module_... |
2442 2443 |
void *__vmalloc_node_range(unsigned long size, unsigned long align, unsigned long start, unsigned long end, gfp_t gfp_mask, |
cb9e3c292 mm: vmalloc: pass... |
2444 2445 |
pgprot_t prot, unsigned long vm_flags, int node, const void *caller) |
1da177e4c Linux-2.6.12-rc2 |
2446 2447 |
{ struct vm_struct *area; |
89219d37a kmemleak: Add the... |
2448 2449 |
void *addr; unsigned long real_size = size; |
1da177e4c Linux-2.6.12-rc2 |
2450 2451 |
size = PAGE_ALIGN(size); |
ca79b0c21 mm: convert total... |
2452 |
if (!size || (size >> PAGE_SHIFT) > totalram_pages()) |
de7d2b567 mm/vmalloc.c: rep... |
2453 |
goto fail; |
1da177e4c Linux-2.6.12-rc2 |
2454 |
|
d98c9e83b kasan: fix crashe... |
2455 |
area = __get_vm_area_node(real_size, align, VM_ALLOC | VM_UNINITIALIZED | |
cb9e3c292 mm: vmalloc: pass... |
2456 |
vm_flags, start, end, node, gfp_mask, caller); |
1da177e4c Linux-2.6.12-rc2 |
2457 |
if (!area) |
de7d2b567 mm/vmalloc.c: rep... |
2458 |
goto fail; |
1da177e4c Linux-2.6.12-rc2 |
2459 |
|
3722e13cf mm/vmalloc: don't... |
2460 |
addr = __vmalloc_area_node(area, gfp_mask, prot, node); |
1368edf06 mm: vmalloc: chec... |
2461 |
if (!addr) |
b82225f3f revert mm/vmalloc... |
2462 |
return NULL; |
89219d37a kmemleak: Add the... |
2463 2464 |
/* |
20fc02b47 mm/vmalloc.c: ren... |
2465 2466 |
* In this function, newly allocated vm_struct has VM_UNINITIALIZED * flag. It means that vm_struct is not fully initialized. |
4341fa454 mm, vmalloc: remo... |
2467 |
* Now, it is fully initialized, so remove this flag here. |
f5252e009 mm: avoid null po... |
2468 |
*/ |
20fc02b47 mm/vmalloc.c: ren... |
2469 |
clear_vm_uninitialized_flag(area); |
f5252e009 mm: avoid null po... |
2470 |
|
94f4a1618 mm: kmemleak: tre... |
2471 |
kmemleak_vmalloc(area, size, gfp_mask); |
89219d37a kmemleak: Add the... |
2472 2473 |
return addr; |
de7d2b567 mm/vmalloc.c: rep... |
2474 2475 |
fail: |
a8e99259e mm, page_alloc: w... |
2476 |
warn_alloc(gfp_mask, NULL, |
7877cdcc3 mm: consolidate w... |
2477 |
"vmalloc: allocation failure: %lu bytes", real_size); |
de7d2b567 mm/vmalloc.c: rep... |
2478 |
return NULL; |
1da177e4c Linux-2.6.12-rc2 |
2479 |
} |
d0a21265d mm: unify module_... |
2480 |
/** |
92eac1681 docs/mm: vmalloc:... |
2481 2482 2483 2484 |
* __vmalloc_node - allocate virtually contiguous memory * @size: allocation size * @align: desired alignment * @gfp_mask: flags for the page level allocator |
92eac1681 docs/mm: vmalloc:... |
2485 2486 |
* @node: node to use for allocation or NUMA_NO_NODE * @caller: caller's return address |
a7c3e901a mm: introduce kv[... |
2487 |
* |
f38fcb9c1 mm: remove the pr... |
2488 2489 |
* Allocate enough pages to cover @size from the page level allocator with * @gfp_mask flags. Map them into contiguous kernel virtual space. |
a7c3e901a mm: introduce kv[... |
2490 |
* |
92eac1681 docs/mm: vmalloc:... |
2491 2492 |
* Reclaim modifiers in @gfp_mask - __GFP_NORETRY, __GFP_RETRY_MAYFAIL * and __GFP_NOFAIL are not supported |
a7c3e901a mm: introduce kv[... |
2493 |
* |
92eac1681 docs/mm: vmalloc:... |
2494 2495 |
* Any use of gfp flags outside of GFP_KERNEL should be consulted * with mm people. |
a862f68a8 docs/core-api/mm:... |
2496 2497 |
* * Return: pointer to the allocated memory or %NULL on error |
d0a21265d mm: unify module_... |
2498 |
*/ |
2b9059489 mm: remove __vmal... |
2499 |
void *__vmalloc_node(unsigned long size, unsigned long align, |
f38fcb9c1 mm: remove the pr... |
2500 |
gfp_t gfp_mask, int node, const void *caller) |
d0a21265d mm: unify module_... |
2501 2502 |
{ return __vmalloc_node_range(size, align, VMALLOC_START, VMALLOC_END, |
f38fcb9c1 mm: remove the pr... |
2503 |
gfp_mask, PAGE_KERNEL, 0, node, caller); |
d0a21265d mm: unify module_... |
2504 |
} |
c3f896dcf mm: switch the te... |
2505 2506 2507 2508 2509 2510 2511 2512 |
/* * This is only for performance analysis of vmalloc and stress purpose. * It is required by vmalloc test module, therefore do not use it other * than that. */ #ifdef CONFIG_TEST_VMALLOC_MODULE EXPORT_SYMBOL_GPL(__vmalloc_node); #endif |
d0a21265d mm: unify module_... |
2513 |
|
88dca4ca5 mm: remove the pg... |
2514 |
void *__vmalloc(unsigned long size, gfp_t gfp_mask) |
930fc45a4 [PATCH] vmalloc_node |
2515 |
{ |
f38fcb9c1 mm: remove the pr... |
2516 |
return __vmalloc_node(size, 1, gfp_mask, NUMA_NO_NODE, |
230169693 vmallocinfo: add ... |
2517 |
__builtin_return_address(0)); |
930fc45a4 [PATCH] vmalloc_node |
2518 |
} |
1da177e4c Linux-2.6.12-rc2 |
2519 2520 2521 |
EXPORT_SYMBOL(__vmalloc); /** |
92eac1681 docs/mm: vmalloc:... |
2522 2523 2524 2525 2526 |
* vmalloc - allocate virtually contiguous memory * @size: allocation size * * Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. |
1da177e4c Linux-2.6.12-rc2 |
2527 |
* |
92eac1681 docs/mm: vmalloc:... |
2528 2529 |
* For tight control over page level allocator and protection flags * use __vmalloc() instead. |
a862f68a8 docs/core-api/mm:... |
2530 2531 |
* * Return: pointer to the allocated memory or %NULL on error |
1da177e4c Linux-2.6.12-rc2 |
2532 2533 2534 |
*/ void *vmalloc(unsigned long size) { |
4d39d7285 mm: remove both i... |
2535 2536 |
return __vmalloc_node(size, 1, GFP_KERNEL, NUMA_NO_NODE, __builtin_return_address(0)); |
1da177e4c Linux-2.6.12-rc2 |
2537 |
} |
1da177e4c Linux-2.6.12-rc2 |
2538 |
EXPORT_SYMBOL(vmalloc); |
930fc45a4 [PATCH] vmalloc_node |
2539 |
/** |
92eac1681 docs/mm: vmalloc:... |
2540 2541 2542 2543 2544 2545 2546 2547 2548 |
* 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. |
a862f68a8 docs/core-api/mm:... |
2549 2550 |
* * Return: pointer to the allocated memory or %NULL on error |
e1ca7788d mm: add vzalloc()... |
2551 2552 2553 |
*/ void *vzalloc(unsigned long size) { |
4d39d7285 mm: remove both i... |
2554 2555 |
return __vmalloc_node(size, 1, GFP_KERNEL | __GFP_ZERO, NUMA_NO_NODE, __builtin_return_address(0)); |
e1ca7788d mm: add vzalloc()... |
2556 2557 2558 2559 |
} EXPORT_SYMBOL(vzalloc); /** |
ead04089b [PATCH] Fix kerne... |
2560 2561 |
* vmalloc_user - allocate zeroed virtually contiguous memory for userspace * @size: allocation size |
833423143 [PATCH] mm: intro... |
2562 |
* |
ead04089b [PATCH] Fix kerne... |
2563 2564 |
* The resulting memory area is zeroed so it can be mapped to userspace * without leaking data. |
a862f68a8 docs/core-api/mm:... |
2565 2566 |
* * Return: pointer to the allocated memory or %NULL on error |
833423143 [PATCH] mm: intro... |
2567 2568 2569 |
*/ void *vmalloc_user(unsigned long size) { |
bc84c5352 mm/vmalloc: pass ... |
2570 2571 2572 2573 |
return __vmalloc_node_range(size, SHMLBA, VMALLOC_START, VMALLOC_END, GFP_KERNEL | __GFP_ZERO, PAGE_KERNEL, VM_USERMAP, NUMA_NO_NODE, __builtin_return_address(0)); |
833423143 [PATCH] mm: intro... |
2574 2575 2576 2577 |
} EXPORT_SYMBOL(vmalloc_user); /** |
92eac1681 docs/mm: vmalloc:... |
2578 2579 2580 |
* vmalloc_node - allocate memory on a specific node * @size: allocation size * @node: numa node |
930fc45a4 [PATCH] vmalloc_node |
2581 |
* |
92eac1681 docs/mm: vmalloc:... |
2582 2583 |
* Allocate enough pages to cover @size from the page level * allocator and map them into contiguous kernel virtual space. |
930fc45a4 [PATCH] vmalloc_node |
2584 |
* |
92eac1681 docs/mm: vmalloc:... |
2585 2586 |
* For tight control over page level allocator and protection flags * use __vmalloc() instead. |
a862f68a8 docs/core-api/mm:... |
2587 2588 |
* * Return: pointer to the allocated memory or %NULL on error |
930fc45a4 [PATCH] vmalloc_node |
2589 2590 2591 |
*/ void *vmalloc_node(unsigned long size, int node) { |
f38fcb9c1 mm: remove the pr... |
2592 2593 |
return __vmalloc_node(size, 1, GFP_KERNEL, node, __builtin_return_address(0)); |
930fc45a4 [PATCH] vmalloc_node |
2594 2595 |
} EXPORT_SYMBOL(vmalloc_node); |
e1ca7788d mm: add vzalloc()... |
2596 2597 2598 2599 2600 2601 2602 2603 2604 |
/** * 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. * |
a862f68a8 docs/core-api/mm:... |
2605 |
* Return: pointer to the allocated memory or %NULL on error |
e1ca7788d mm: add vzalloc()... |
2606 2607 2608 |
*/ void *vzalloc_node(unsigned long size, int node) { |
4d39d7285 mm: remove both i... |
2609 2610 |
return __vmalloc_node(size, 1, GFP_KERNEL | __GFP_ZERO, node, __builtin_return_address(0)); |
e1ca7788d mm: add vzalloc()... |
2611 2612 |
} EXPORT_SYMBOL(vzalloc_node); |
0d08e0d3a [PATCH] x86-64: F... |
2613 |
#if defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA32) |
698d0831b vmalloc: fix __GF... |
2614 |
#define GFP_VMALLOC32 (GFP_DMA32 | GFP_KERNEL) |
0d08e0d3a [PATCH] x86-64: F... |
2615 |
#elif defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA) |
698d0831b vmalloc: fix __GF... |
2616 |
#define GFP_VMALLOC32 (GFP_DMA | GFP_KERNEL) |
0d08e0d3a [PATCH] x86-64: F... |
2617 |
#else |
698d0831b vmalloc: fix __GF... |
2618 2619 2620 2621 2622 |
/* * 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 [PATCH] x86-64: F... |
2623 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
2624 |
/** |
92eac1681 docs/mm: vmalloc:... |
2625 2626 |
* vmalloc_32 - allocate virtually contiguous memory (32bit addressable) * @size: allocation size |
1da177e4c Linux-2.6.12-rc2 |
2627 |
* |
92eac1681 docs/mm: vmalloc:... |
2628 2629 |
* Allocate enough 32bit PA addressable pages to cover @size from the * page level allocator and map them into contiguous kernel virtual space. |
a862f68a8 docs/core-api/mm:... |
2630 2631 |
* * Return: pointer to the allocated memory or %NULL on error |
1da177e4c Linux-2.6.12-rc2 |
2632 2633 2634 |
*/ void *vmalloc_32(unsigned long size) { |
f38fcb9c1 mm: remove the pr... |
2635 2636 |
return __vmalloc_node(size, 1, GFP_VMALLOC32, NUMA_NO_NODE, __builtin_return_address(0)); |
1da177e4c Linux-2.6.12-rc2 |
2637 |
} |
1da177e4c Linux-2.6.12-rc2 |
2638 |
EXPORT_SYMBOL(vmalloc_32); |
833423143 [PATCH] mm: intro... |
2639 |
/** |
ead04089b [PATCH] Fix kerne... |
2640 |
* vmalloc_32_user - allocate zeroed virtually contiguous 32bit memory |
92eac1681 docs/mm: vmalloc:... |
2641 |
* @size: allocation size |
ead04089b [PATCH] Fix kerne... |
2642 2643 2644 |
* * The resulting memory area is 32bit addressable and zeroed so it can be * mapped to userspace without leaking data. |
a862f68a8 docs/core-api/mm:... |
2645 2646 |
* * Return: pointer to the allocated memory or %NULL on error |
833423143 [PATCH] mm: intro... |
2647 2648 2649 |
*/ void *vmalloc_32_user(unsigned long size) { |
bc84c5352 mm/vmalloc: pass ... |
2650 2651 2652 2653 |
return __vmalloc_node_range(size, SHMLBA, VMALLOC_START, VMALLOC_END, GFP_VMALLOC32 | __GFP_ZERO, PAGE_KERNEL, VM_USERMAP, NUMA_NO_NODE, __builtin_return_address(0)); |
833423143 [PATCH] mm: intro... |
2654 2655 |
} EXPORT_SYMBOL(vmalloc_32_user); |
d0107eb07 kcore: fix vread/... |
2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 |
/* * 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 mm/vmalloc: use o... |
2668 |
offset = offset_in_page(addr); |
d0107eb07 kcore: fix vread/... |
2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 |
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 mm: remove the se... |
2685 |
void *map = kmap_atomic(p); |
d0107eb07 kcore: fix vread/... |
2686 |
memcpy(buf, map + offset, length); |
9b04c5fec mm: remove the se... |
2687 |
kunmap_atomic(map); |
d0107eb07 kcore: fix vread/... |
2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 |
} 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 mm/vmalloc: use o... |
2706 |
offset = offset_in_page(addr); |
d0107eb07 kcore: fix vread/... |
2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 |
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 mm: remove the se... |
2723 |
void *map = kmap_atomic(p); |
d0107eb07 kcore: fix vread/... |
2724 |
memcpy(map + offset, buf, length); |
9b04c5fec mm: remove the se... |
2725 |
kunmap_atomic(map); |
d0107eb07 kcore: fix vread/... |
2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 |
} addr += length; buf += length; copied += length; count -= length; } return copied; } /** |
92eac1681 docs/mm: vmalloc:... |
2736 2737 2738 2739 2740 |
* vread() - read vmalloc area in a safe way. * @buf: buffer for reading data * @addr: vm address. * @count: number of bytes to be read. * |
92eac1681 docs/mm: vmalloc:... |
2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 |
* 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 * vm_struct area, returns 0. @buf should be kernel's buffer. * * 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 |
d9009d67f mm/vmalloc.c: spe... |
2753 |
* any information, as /dev/kmem. |
a862f68a8 docs/core-api/mm:... |
2754 2755 2756 2757 |
* * Return: number of bytes for which addr and buf should be increased * (same number as @count) or %0 if [addr...addr+count) doesn't * include any intersection with valid vmalloc area |
d0107eb07 kcore: fix vread/... |
2758 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
2759 2760 |
long vread(char *buf, char *addr, unsigned long count) { |
e81ce85f9 mm, vmalloc: iter... |
2761 2762 |
struct vmap_area *va; struct vm_struct *vm; |
1da177e4c Linux-2.6.12-rc2 |
2763 |
char *vaddr, *buf_start = buf; |
d0107eb07 kcore: fix vread/... |
2764 |
unsigned long buflen = count; |
1da177e4c Linux-2.6.12-rc2 |
2765 2766 2767 2768 2769 |
unsigned long n; /* Don't allow overflow */ if ((unsigned long) addr + count < count) count = -(unsigned long) addr; |
e81ce85f9 mm, vmalloc: iter... |
2770 2771 2772 2773 |
spin_lock(&vmap_area_lock); list_for_each_entry(va, &vmap_area_list, list) { if (!count) break; |
688fcbfc0 mm/vmalloc: modif... |
2774 |
if (!va->vm) |
e81ce85f9 mm, vmalloc: iter... |
2775 2776 2777 2778 |
continue; vm = va->vm; vaddr = (char *) vm->addr; |
762216ab4 mm/vmalloc: use w... |
2779 |
if (addr >= vaddr + get_vm_area_size(vm)) |
1da177e4c Linux-2.6.12-rc2 |
2780 2781 2782 2783 2784 2785 2786 2787 2788 |
continue; while (addr < vaddr) { if (count == 0) goto finished; *buf = '\0'; buf++; addr++; count--; } |
762216ab4 mm/vmalloc: use w... |
2789 |
n = vaddr + get_vm_area_size(vm) - addr; |
d0107eb07 kcore: fix vread/... |
2790 2791 |
if (n > count) n = count; |
e81ce85f9 mm, vmalloc: iter... |
2792 |
if (!(vm->flags & VM_IOREMAP)) |
d0107eb07 kcore: fix vread/... |
2793 2794 2795 2796 2797 2798 |
aligned_vread(buf, addr, n); else /* IOREMAP area is treated as memory hole */ memset(buf, 0, n); buf += n; addr += n; count -= n; |
1da177e4c Linux-2.6.12-rc2 |
2799 2800 |
} finished: |
e81ce85f9 mm, vmalloc: iter... |
2801 |
spin_unlock(&vmap_area_lock); |
d0107eb07 kcore: fix vread/... |
2802 2803 2804 2805 2806 2807 2808 2809 |
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 Linux-2.6.12-rc2 |
2810 |
} |
d0107eb07 kcore: fix vread/... |
2811 |
/** |
92eac1681 docs/mm: vmalloc:... |
2812 2813 2814 2815 2816 |
* vwrite() - write vmalloc area in a safe way. * @buf: buffer for source data * @addr: vm address. * @count: number of bytes to be read. * |
92eac1681 docs/mm: vmalloc:... |
2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 |
* 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 * vm_struct area, returns 0. @buf should be kernel's buffer. * * 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 |
d9009d67f mm/vmalloc.c: spe... |
2829 |
* any information, as /dev/kmem. |
a862f68a8 docs/core-api/mm:... |
2830 2831 2832 2833 |
* * Return: number of bytes for which addr and buf should be * increased (same number as @count) or %0 if [addr...addr+count) * doesn't include any intersection with valid vmalloc area |
d0107eb07 kcore: fix vread/... |
2834 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
2835 2836 |
long vwrite(char *buf, char *addr, unsigned long count) { |
e81ce85f9 mm, vmalloc: iter... |
2837 2838 |
struct vmap_area *va; struct vm_struct *vm; |
d0107eb07 kcore: fix vread/... |
2839 2840 2841 |
char *vaddr; unsigned long n, buflen; int copied = 0; |
1da177e4c Linux-2.6.12-rc2 |
2842 2843 2844 2845 |
/* Don't allow overflow */ if ((unsigned long) addr + count < count) count = -(unsigned long) addr; |
d0107eb07 kcore: fix vread/... |
2846 |
buflen = count; |
1da177e4c Linux-2.6.12-rc2 |
2847 |
|
e81ce85f9 mm, vmalloc: iter... |
2848 2849 2850 2851 |
spin_lock(&vmap_area_lock); list_for_each_entry(va, &vmap_area_list, list) { if (!count) break; |
688fcbfc0 mm/vmalloc: modif... |
2852 |
if (!va->vm) |
e81ce85f9 mm, vmalloc: iter... |
2853 2854 2855 2856 |
continue; vm = va->vm; vaddr = (char *) vm->addr; |
762216ab4 mm/vmalloc: use w... |
2857 |
if (addr >= vaddr + get_vm_area_size(vm)) |
1da177e4c Linux-2.6.12-rc2 |
2858 2859 2860 2861 2862 2863 2864 2865 |
continue; while (addr < vaddr) { if (count == 0) goto finished; buf++; addr++; count--; } |
762216ab4 mm/vmalloc: use w... |
2866 |
n = vaddr + get_vm_area_size(vm) - addr; |
d0107eb07 kcore: fix vread/... |
2867 2868 |
if (n > count) n = count; |
e81ce85f9 mm, vmalloc: iter... |
2869 |
if (!(vm->flags & VM_IOREMAP)) { |
d0107eb07 kcore: fix vread/... |
2870 2871 2872 2873 2874 2875 |
aligned_vwrite(buf, addr, n); copied++; } buf += n; addr += n; count -= n; |
1da177e4c Linux-2.6.12-rc2 |
2876 2877 |
} finished: |
e81ce85f9 mm, vmalloc: iter... |
2878 |
spin_unlock(&vmap_area_lock); |
d0107eb07 kcore: fix vread/... |
2879 2880 2881 |
if (!copied) return 0; return buflen; |
1da177e4c Linux-2.6.12-rc2 |
2882 |
} |
833423143 [PATCH] mm: intro... |
2883 2884 |
/** |
92eac1681 docs/mm: vmalloc:... |
2885 2886 2887 2888 |
* 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 |
bdebd6a28 vmalloc: fix rema... |
2889 |
* @pgoff: offset from @kaddr to start at |
92eac1681 docs/mm: vmalloc:... |
2890 |
* @size: size of map area |
7682486b3 mm: fix various k... |
2891 |
* |
92eac1681 docs/mm: vmalloc:... |
2892 |
* Returns: 0 for success, -Exxx on failure |
833423143 [PATCH] mm: intro... |
2893 |
* |
92eac1681 docs/mm: vmalloc:... |
2894 2895 2896 2897 |
* 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 [PATCH] mm: intro... |
2898 |
* |
92eac1681 docs/mm: vmalloc:... |
2899 |
* Similar to remap_pfn_range() (see mm/memory.c) |
833423143 [PATCH] mm: intro... |
2900 |
*/ |
e69e9d4ae vmalloc: introduc... |
2901 |
int remap_vmalloc_range_partial(struct vm_area_struct *vma, unsigned long uaddr, |
bdebd6a28 vmalloc: fix rema... |
2902 2903 |
void *kaddr, unsigned long pgoff, unsigned long size) |
833423143 [PATCH] mm: intro... |
2904 2905 |
{ struct vm_struct *area; |
bdebd6a28 vmalloc: fix rema... |
2906 2907 2908 2909 2910 |
unsigned long off; unsigned long end_index; if (check_shl_overflow(pgoff, PAGE_SHIFT, &off)) return -EINVAL; |
833423143 [PATCH] mm: intro... |
2911 |
|
e69e9d4ae vmalloc: introduc... |
2912 2913 2914 |
size = PAGE_ALIGN(size); if (!PAGE_ALIGNED(uaddr) || !PAGE_ALIGNED(kaddr)) |
833423143 [PATCH] mm: intro... |
2915 |
return -EINVAL; |
e69e9d4ae vmalloc: introduc... |
2916 |
area = find_vm_area(kaddr); |
833423143 [PATCH] mm: intro... |
2917 |
if (!area) |
db64fe022 mm: rewrite vmap ... |
2918 |
return -EINVAL; |
833423143 [PATCH] mm: intro... |
2919 |
|
fe9041c24 vmalloc: lift the... |
2920 |
if (!(area->flags & (VM_USERMAP | VM_DMA_COHERENT))) |
db64fe022 mm: rewrite vmap ... |
2921 |
return -EINVAL; |
833423143 [PATCH] mm: intro... |
2922 |
|
bdebd6a28 vmalloc: fix rema... |
2923 2924 |
if (check_add_overflow(size, off, &end_index) || end_index > get_vm_area_size(area)) |
db64fe022 mm: rewrite vmap ... |
2925 |
return -EINVAL; |
bdebd6a28 vmalloc: fix rema... |
2926 |
kaddr += off; |
833423143 [PATCH] mm: intro... |
2927 |
|
833423143 [PATCH] mm: intro... |
2928 |
do { |
e69e9d4ae vmalloc: introduc... |
2929 |
struct page *page = vmalloc_to_page(kaddr); |
db64fe022 mm: rewrite vmap ... |
2930 |
int ret; |
833423143 [PATCH] mm: intro... |
2931 2932 2933 2934 2935 |
ret = vm_insert_page(vma, uaddr, page); if (ret) return ret; uaddr += PAGE_SIZE; |
e69e9d4ae vmalloc: introduc... |
2936 2937 2938 |
kaddr += PAGE_SIZE; size -= PAGE_SIZE; } while (size > 0); |
833423143 [PATCH] mm: intro... |
2939 |
|
314e51b98 mm: kill vma flag... |
2940 |
vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP; |
833423143 [PATCH] mm: intro... |
2941 |
|
db64fe022 mm: rewrite vmap ... |
2942 |
return 0; |
833423143 [PATCH] mm: intro... |
2943 |
} |
e69e9d4ae vmalloc: introduc... |
2944 2945 2946 |
EXPORT_SYMBOL(remap_vmalloc_range_partial); /** |
92eac1681 docs/mm: vmalloc:... |
2947 2948 2949 2950 |
* 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 |
e69e9d4ae vmalloc: introduc... |
2951 |
* |
92eac1681 docs/mm: vmalloc:... |
2952 |
* Returns: 0 for success, -Exxx on failure |
e69e9d4ae vmalloc: introduc... |
2953 |
* |
92eac1681 docs/mm: vmalloc:... |
2954 2955 2956 |
* 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. |
e69e9d4ae vmalloc: introduc... |
2957 |
* |
92eac1681 docs/mm: vmalloc:... |
2958 |
* Similar to remap_pfn_range() (see mm/memory.c) |
e69e9d4ae vmalloc: introduc... |
2959 2960 2961 2962 2963 |
*/ int remap_vmalloc_range(struct vm_area_struct *vma, void *addr, unsigned long pgoff) { return remap_vmalloc_range_partial(vma, vma->vm_start, |
bdebd6a28 vmalloc: fix rema... |
2964 |
addr, pgoff, |
e69e9d4ae vmalloc: introduc... |
2965 2966 |
vma->vm_end - vma->vm_start); } |
833423143 [PATCH] mm: intro... |
2967 |
EXPORT_SYMBOL(remap_vmalloc_range); |
5f4352fbf Allocate and free... |
2968 2969 2970 2971 2972 2973 2974 2975 |
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 vmalloc: show vma... |
2976 |
|
4f8b02b4e vmalloc: pcpu_get... |
2977 |
#ifdef CONFIG_SMP |
ca23e405e vmalloc: implemen... |
2978 2979 |
static struct vmap_area *node_to_va(struct rb_node *n) { |
4583e7731 mm/vmalloc.c: use... |
2980 |
return rb_entry_safe(n, struct vmap_area, rb_node); |
ca23e405e vmalloc: implemen... |
2981 2982 2983 |
} /** |
68ad4a330 mm/vmalloc.c: kee... |
2984 2985 |
* pvm_find_va_enclose_addr - find the vmap_area @addr belongs to * @addr: target address |
ca23e405e vmalloc: implemen... |
2986 |
* |
68ad4a330 mm/vmalloc.c: kee... |
2987 2988 2989 2990 |
* Returns: vmap_area if it is found. If there is no such area * the first highest(reverse order) vmap_area is returned * i.e. va->va_start < addr && va->va_end < addr or NULL * if there are no any areas before @addr. |
ca23e405e vmalloc: implemen... |
2991 |
*/ |
68ad4a330 mm/vmalloc.c: kee... |
2992 2993 |
static struct vmap_area * pvm_find_va_enclose_addr(unsigned long addr) |
ca23e405e vmalloc: implemen... |
2994 |
{ |
68ad4a330 mm/vmalloc.c: kee... |
2995 2996 2997 2998 2999 |
struct vmap_area *va, *tmp; struct rb_node *n; n = free_vmap_area_root.rb_node; va = NULL; |
ca23e405e vmalloc: implemen... |
3000 3001 |
while (n) { |
68ad4a330 mm/vmalloc.c: kee... |
3002 3003 3004 3005 3006 |
tmp = rb_entry(n, struct vmap_area, rb_node); if (tmp->va_start <= addr) { va = tmp; if (tmp->va_end >= addr) break; |
ca23e405e vmalloc: implemen... |
3007 |
n = n->rb_right; |
68ad4a330 mm/vmalloc.c: kee... |
3008 3009 3010 |
} else { n = n->rb_left; } |
ca23e405e vmalloc: implemen... |
3011 |
} |
68ad4a330 mm/vmalloc.c: kee... |
3012 |
return va; |
ca23e405e vmalloc: implemen... |
3013 3014 3015 |
} /** |
68ad4a330 mm/vmalloc.c: kee... |
3016 3017 3018 3019 3020 |
* pvm_determine_end_from_reverse - find the highest aligned address * of free block below VMALLOC_END * @va: * in - the VA we start the search(reverse order); * out - the VA with the highest aligned end address. |
ca23e405e vmalloc: implemen... |
3021 |
* |
68ad4a330 mm/vmalloc.c: kee... |
3022 |
* Returns: determined end address within vmap_area |
ca23e405e vmalloc: implemen... |
3023 |
*/ |
68ad4a330 mm/vmalloc.c: kee... |
3024 3025 |
static unsigned long pvm_determine_end_from_reverse(struct vmap_area **va, unsigned long align) |
ca23e405e vmalloc: implemen... |
3026 |
{ |
68ad4a330 mm/vmalloc.c: kee... |
3027 |
unsigned long vmalloc_end = VMALLOC_END & ~(align - 1); |
ca23e405e vmalloc: implemen... |
3028 |
unsigned long addr; |
68ad4a330 mm/vmalloc.c: kee... |
3029 3030 3031 3032 3033 3034 3035 |
if (likely(*va)) { list_for_each_entry_from_reverse((*va), &free_vmap_area_list, list) { addr = min((*va)->va_end & ~(align - 1), vmalloc_end); if ((*va)->va_start < addr) return addr; } |
ca23e405e vmalloc: implemen... |
3036 |
} |
68ad4a330 mm/vmalloc.c: kee... |
3037 |
return 0; |
ca23e405e vmalloc: implemen... |
3038 3039 3040 3041 3042 3043 3044 3045 |
} /** * 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 vmalloc: implemen... |
3046 3047 3048 3049 3050 3051 |
* * 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 mm: remove gfp ma... |
3052 3053 3054 3055 |
* 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 vmalloc: implemen... |
3056 |
* |
68ad4a330 mm/vmalloc.c: kee... |
3057 3058 3059 3060 3061 3062 |
* Despite its complicated look, this allocator is rather simple. It * does everything top-down and scans free blocks from the end looking * for matching base. While scanning, if any of the areas do not fit the * base address is pulled down to fit the area. Scanning is repeated till * all the areas fit and then all necessary data structures are inserted * and the result is returned. |
ca23e405e vmalloc: implemen... |
3063 3064 3065 |
*/ struct vm_struct **pcpu_get_vm_areas(const unsigned long *offsets, const size_t *sizes, int nr_vms, |
ec3f64fc9 mm: remove gfp ma... |
3066 |
size_t align) |
ca23e405e vmalloc: implemen... |
3067 3068 3069 |
{ const unsigned long vmalloc_start = ALIGN(VMALLOC_START, align); const unsigned long vmalloc_end = VMALLOC_END & ~(align - 1); |
68ad4a330 mm/vmalloc.c: kee... |
3070 |
struct vmap_area **vas, *va; |
ca23e405e vmalloc: implemen... |
3071 3072 |
struct vm_struct **vms; int area, area2, last_area, term_area; |
253a496d8 kasan: don't assu... |
3073 |
unsigned long base, start, size, end, last_end, orig_start, orig_end; |
ca23e405e vmalloc: implemen... |
3074 |
bool purged = false; |
68ad4a330 mm/vmalloc.c: kee... |
3075 |
enum fit_type type; |
ca23e405e vmalloc: implemen... |
3076 |
|
ca23e405e vmalloc: implemen... |
3077 |
/* verify parameters and allocate data structures */ |
891c49abf mm/vmalloc: use o... |
3078 |
BUG_ON(offset_in_page(align) || !is_power_of_2(align)); |
ca23e405e vmalloc: implemen... |
3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 |
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 mm/vmalloc.c: hal... |
3090 |
for (area2 = area + 1; area2 < nr_vms; area2++) { |
ca23e405e vmalloc: implemen... |
3091 3092 |
unsigned long start2 = offsets[area2]; unsigned long end2 = start2 + sizes[area2]; |
c568da282 mm/vmalloc.c: hal... |
3093 |
BUG_ON(start2 < end && start < end2); |
ca23e405e vmalloc: implemen... |
3094 3095 3096 3097 3098 3099 3100 3101 |
} } last_end = offsets[last_area] + sizes[last_area]; if (vmalloc_end - vmalloc_start < last_end) { WARN_ON(true); return NULL; } |
4d67d8605 mm: use kcalloc()... |
3102 3103 |
vms = kcalloc(nr_vms, sizeof(vms[0]), GFP_KERNEL); vas = kcalloc(nr_vms, sizeof(vas[0]), GFP_KERNEL); |
ca23e405e vmalloc: implemen... |
3104 |
if (!vas || !vms) |
f1db7afd9 mm/vmalloc.c: eli... |
3105 |
goto err_free2; |
ca23e405e vmalloc: implemen... |
3106 3107 |
for (area = 0; area < nr_vms; area++) { |
68ad4a330 mm/vmalloc.c: kee... |
3108 |
vas[area] = kmem_cache_zalloc(vmap_area_cachep, GFP_KERNEL); |
ec3f64fc9 mm: remove gfp ma... |
3109 |
vms[area] = kzalloc(sizeof(struct vm_struct), GFP_KERNEL); |
ca23e405e vmalloc: implemen... |
3110 3111 3112 3113 |
if (!vas[area] || !vms[area]) goto err_free; } retry: |
e36176be1 mm/vmalloc: rewor... |
3114 |
spin_lock(&free_vmap_area_lock); |
ca23e405e vmalloc: implemen... |
3115 3116 3117 3118 3119 |
/* start scanning - we scan from the top, begin with the last area */ area = term_area = last_area; start = offsets[area]; end = start + sizes[area]; |
68ad4a330 mm/vmalloc.c: kee... |
3120 3121 |
va = pvm_find_va_enclose_addr(vmalloc_end); base = pvm_determine_end_from_reverse(&va, align) - end; |
ca23e405e vmalloc: implemen... |
3122 3123 |
while (true) { |
ca23e405e vmalloc: implemen... |
3124 3125 3126 3127 |
/* * base might have underflowed, add last_end before * comparing. */ |
68ad4a330 mm/vmalloc.c: kee... |
3128 3129 |
if (base + last_end < vmalloc_start + last_end) goto overflow; |
ca23e405e vmalloc: implemen... |
3130 3131 |
/* |
68ad4a330 mm/vmalloc.c: kee... |
3132 |
* Fitting base has not been found. |
ca23e405e vmalloc: implemen... |
3133 |
*/ |
68ad4a330 mm/vmalloc.c: kee... |
3134 3135 |
if (va == NULL) goto overflow; |
ca23e405e vmalloc: implemen... |
3136 3137 |
/* |
d8cc323d9 mm/vmalloc: fix a... |
3138 |
* If required width exceeds current VA block, move |
5336e52c9 mm/vmalloc.c: fix... |
3139 3140 3141 3142 3143 3144 3145 3146 3147 |
* base downwards and then recheck. */ if (base + end > va->va_end) { base = pvm_determine_end_from_reverse(&va, align) - end; term_area = area; continue; } /* |
68ad4a330 mm/vmalloc.c: kee... |
3148 |
* If this VA does not fit, move base downwards and recheck. |
ca23e405e vmalloc: implemen... |
3149 |
*/ |
5336e52c9 mm/vmalloc.c: fix... |
3150 |
if (base + start < va->va_start) { |
68ad4a330 mm/vmalloc.c: kee... |
3151 3152 |
va = node_to_va(rb_prev(&va->rb_node)); base = pvm_determine_end_from_reverse(&va, align) - end; |
ca23e405e vmalloc: implemen... |
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 |
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; |
68ad4a330 mm/vmalloc.c: kee... |
3164 |
|
ca23e405e vmalloc: implemen... |
3165 3166 |
start = offsets[area]; end = start + sizes[area]; |
68ad4a330 mm/vmalloc.c: kee... |
3167 |
va = pvm_find_va_enclose_addr(base + end); |
ca23e405e vmalloc: implemen... |
3168 |
} |
68ad4a330 mm/vmalloc.c: kee... |
3169 |
|
ca23e405e vmalloc: implemen... |
3170 3171 |
/* we've found a fitting base, insert all va's */ for (area = 0; area < nr_vms; area++) { |
68ad4a330 mm/vmalloc.c: kee... |
3172 |
int ret; |
ca23e405e vmalloc: implemen... |
3173 |
|
68ad4a330 mm/vmalloc.c: kee... |
3174 3175 |
start = base + offsets[area]; size = sizes[area]; |
ca23e405e vmalloc: implemen... |
3176 |
|
68ad4a330 mm/vmalloc.c: kee... |
3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 |
va = pvm_find_va_enclose_addr(start); if (WARN_ON_ONCE(va == NULL)) /* It is a BUG(), but trigger recovery instead. */ goto recovery; type = classify_va_fit_type(va, start, size); if (WARN_ON_ONCE(type == NOTHING_FIT)) /* It is a BUG(), but trigger recovery instead. */ goto recovery; ret = adjust_va_to_fit_type(va, start, size, type); if (unlikely(ret)) goto recovery; /* Allocated area. */ va = vas[area]; va->va_start = start; va->va_end = start + size; |
68ad4a330 mm/vmalloc.c: kee... |
3195 |
} |
ca23e405e vmalloc: implemen... |
3196 |
|
e36176be1 mm/vmalloc: rewor... |
3197 |
spin_unlock(&free_vmap_area_lock); |
ca23e405e vmalloc: implemen... |
3198 |
|
253a496d8 kasan: don't assu... |
3199 3200 3201 3202 3203 3204 3205 3206 |
/* populate the kasan shadow space */ for (area = 0; area < nr_vms; area++) { if (kasan_populate_vmalloc(vas[area]->va_start, sizes[area])) goto err_free_shadow; kasan_unpoison_vmalloc((void *)vas[area]->va_start, sizes[area]); } |
ca23e405e vmalloc: implemen... |
3207 |
/* insert all vm's */ |
e36176be1 mm/vmalloc: rewor... |
3208 3209 3210 3211 3212 |
spin_lock(&vmap_area_lock); for (area = 0; area < nr_vms; area++) { insert_vmap_area(vas[area], &vmap_area_root, &vmap_area_list); setup_vmalloc_vm_locked(vms[area], vas[area], VM_ALLOC, |
3645cb4a4 mm, vmalloc: call... |
3213 |
pcpu_get_vm_areas); |
e36176be1 mm/vmalloc: rewor... |
3214 3215 |
} spin_unlock(&vmap_area_lock); |
ca23e405e vmalloc: implemen... |
3216 3217 3218 |
kfree(vas); return vms; |
68ad4a330 mm/vmalloc.c: kee... |
3219 |
recovery: |
e36176be1 mm/vmalloc: rewor... |
3220 3221 3222 3223 3224 3225 |
/* * Remove previously allocated areas. There is no * need in removing these areas from the busy tree, * because they are inserted only on the final step * and when pcpu_get_vm_areas() is success. */ |
68ad4a330 mm/vmalloc.c: kee... |
3226 |
while (area--) { |
253a496d8 kasan: don't assu... |
3227 3228 3229 3230 |
orig_start = vas[area]->va_start; orig_end = vas[area]->va_end; va = merge_or_add_vmap_area(vas[area], &free_vmap_area_root, &free_vmap_area_list); |
9c801f61d mm/vmalloc.c: rem... |
3231 3232 3233 |
if (va) kasan_release_vmalloc(orig_start, orig_end, va->va_start, va->va_end); |
68ad4a330 mm/vmalloc.c: kee... |
3234 3235 3236 3237 |
vas[area] = NULL; } overflow: |
e36176be1 mm/vmalloc: rewor... |
3238 |
spin_unlock(&free_vmap_area_lock); |
68ad4a330 mm/vmalloc.c: kee... |
3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 |
if (!purged) { purge_vmap_area_lazy(); purged = true; /* Before "retry", check if we recover. */ for (area = 0; area < nr_vms; area++) { if (vas[area]) continue; vas[area] = kmem_cache_zalloc( vmap_area_cachep, GFP_KERNEL); if (!vas[area]) goto err_free; } goto retry; } |
ca23e405e vmalloc: implemen... |
3256 3257 |
err_free: for (area = 0; area < nr_vms; area++) { |
68ad4a330 mm/vmalloc.c: kee... |
3258 3259 |
if (vas[area]) kmem_cache_free(vmap_area_cachep, vas[area]); |
f1db7afd9 mm/vmalloc.c: eli... |
3260 |
kfree(vms[area]); |
ca23e405e vmalloc: implemen... |
3261 |
} |
f1db7afd9 mm/vmalloc.c: eli... |
3262 |
err_free2: |
ca23e405e vmalloc: implemen... |
3263 3264 3265 |
kfree(vas); kfree(vms); return NULL; |
253a496d8 kasan: don't assu... |
3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 |
err_free_shadow: spin_lock(&free_vmap_area_lock); /* * We release all the vmalloc shadows, even the ones for regions that * hadn't been successfully added. This relies on kasan_release_vmalloc * being able to tolerate this case. */ for (area = 0; area < nr_vms; area++) { orig_start = vas[area]->va_start; orig_end = vas[area]->va_end; va = merge_or_add_vmap_area(vas[area], &free_vmap_area_root, &free_vmap_area_list); |
9c801f61d mm/vmalloc.c: rem... |
3279 3280 3281 |
if (va) kasan_release_vmalloc(orig_start, orig_end, va->va_start, va->va_end); |
253a496d8 kasan: don't assu... |
3282 3283 3284 3285 3286 3287 3288 |
vas[area] = NULL; kfree(vms[area]); } spin_unlock(&free_vmap_area_lock); kfree(vas); kfree(vms); return NULL; |
ca23e405e vmalloc: implemen... |
3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 |
} /** * 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 vmalloc: pcpu_get... |
3306 |
#endif /* CONFIG_SMP */ |
a10aa5798 vmalloc: show vma... |
3307 3308 3309 |
#ifdef CONFIG_PROC_FS static void *s_start(struct seq_file *m, loff_t *pos) |
e36176be1 mm/vmalloc: rewor... |
3310 |
__acquires(&vmap_purge_lock) |
d4033afdf mm, vmalloc: iter... |
3311 |
__acquires(&vmap_area_lock) |
a10aa5798 vmalloc: show vma... |
3312 |
{ |
e36176be1 mm/vmalloc: rewor... |
3313 |
mutex_lock(&vmap_purge_lock); |
d4033afdf mm, vmalloc: iter... |
3314 |
spin_lock(&vmap_area_lock); |
e36176be1 mm/vmalloc: rewor... |
3315 |
|
3f5000693 mm/vmalloc.c: sim... |
3316 |
return seq_list_start(&vmap_area_list, *pos); |
a10aa5798 vmalloc: show vma... |
3317 3318 3319 3320 |
} static void *s_next(struct seq_file *m, void *p, loff_t *pos) { |
3f5000693 mm/vmalloc.c: sim... |
3321 |
return seq_list_next(p, &vmap_area_list, pos); |
a10aa5798 vmalloc: show vma... |
3322 3323 3324 |
} static void s_stop(struct seq_file *m, void *p) |
d4033afdf mm, vmalloc: iter... |
3325 |
__releases(&vmap_area_lock) |
4a9d8b078 mm/vmalloc: Fix u... |
3326 |
__releases(&vmap_purge_lock) |
a10aa5798 vmalloc: show vma... |
3327 |
{ |
d4033afdf mm, vmalloc: iter... |
3328 |
spin_unlock(&vmap_area_lock); |
4a9d8b078 mm/vmalloc: Fix u... |
3329 |
mutex_unlock(&vmap_purge_lock); |
a10aa5798 vmalloc: show vma... |
3330 |
} |
a47a126ad vmallocinfo: add ... |
3331 3332 |
static void show_numa_info(struct seq_file *m, struct vm_struct *v) { |
e5adfffc8 mm: use IS_ENABLE... |
3333 |
if (IS_ENABLED(CONFIG_NUMA)) { |
a47a126ad vmallocinfo: add ... |
3334 3335 3336 3337 |
unsigned int nr, *counters = m->private; if (!counters) return; |
af12346cd mm/vmalloc: rever... |
3338 3339 |
if (v->flags & VM_UNINITIALIZED) return; |
7e5b528b4 mm/vmalloc.c: fix... |
3340 3341 |
/* Pair with smp_wmb() in clear_vm_uninitialized_flag() */ smp_rmb(); |
af12346cd mm/vmalloc: rever... |
3342 |
|
a47a126ad vmallocinfo: add ... |
3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 |
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]); } } |
dd3b8353b mm/vmalloc: do no... |
3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 |
static void show_purge_info(struct seq_file *m) { struct llist_node *head; struct vmap_area *va; head = READ_ONCE(vmap_purge_list.first); if (head == NULL) return; llist_for_each_entry(va, head, purge_list) { seq_printf(m, "0x%pK-0x%pK %7ld unpurged vm_area ", (void *)va->va_start, (void *)va->va_end, va->va_end - va->va_start); } } |
a10aa5798 vmalloc: show vma... |
3369 3370 |
static int s_show(struct seq_file *m, void *p) { |
3f5000693 mm/vmalloc.c: sim... |
3371 |
struct vmap_area *va; |
d4033afdf mm, vmalloc: iter... |
3372 |
struct vm_struct *v; |
3f5000693 mm/vmalloc.c: sim... |
3373 |
va = list_entry(p, struct vmap_area, list); |
c2ce8c142 mm/vmalloc: fix s... |
3374 |
/* |
688fcbfc0 mm/vmalloc: modif... |
3375 3376 |
* s_show can encounter race with remove_vm_area, !vm on behalf * of vmap area is being tear down or vm_map_ram allocation. |
c2ce8c142 mm/vmalloc: fix s... |
3377 |
*/ |
688fcbfc0 mm/vmalloc: modif... |
3378 |
if (!va->vm) { |
dd3b8353b mm/vmalloc: do no... |
3379 3380 |
seq_printf(m, "0x%pK-0x%pK %7ld vm_map_ram ", |
78c72746f vmalloc: show laz... |
3381 |
(void *)va->va_start, (void *)va->va_end, |
dd3b8353b mm/vmalloc: do no... |
3382 |
va->va_end - va->va_start); |
78c72746f vmalloc: show laz... |
3383 |
|
d4033afdf mm, vmalloc: iter... |
3384 |
return 0; |
78c72746f vmalloc: show laz... |
3385 |
} |
d4033afdf mm, vmalloc: iter... |
3386 3387 |
v = va->vm; |
a10aa5798 vmalloc: show vma... |
3388 |
|
45ec16908 mm: use %pK for /... |
3389 |
seq_printf(m, "0x%pK-0x%pK %7ld", |
a10aa5798 vmalloc: show vma... |
3390 |
v->addr, v->addr + v->size, v->size); |
62c70bce8 mm: convert sprin... |
3391 3392 |
if (v->caller) seq_printf(m, " %pS", v->caller); |
230169693 vmallocinfo: add ... |
3393 |
|
a10aa5798 vmalloc: show vma... |
3394 3395 3396 3397 |
if (v->nr_pages) seq_printf(m, " pages=%d", v->nr_pages); if (v->phys_addr) |
199eaa05a mm: cleanups for ... |
3398 |
seq_printf(m, " phys=%pa", &v->phys_addr); |
a10aa5798 vmalloc: show vma... |
3399 3400 |
if (v->flags & VM_IOREMAP) |
f4527c908 mm/vmalloc.c: rep... |
3401 |
seq_puts(m, " ioremap"); |
a10aa5798 vmalloc: show vma... |
3402 3403 |
if (v->flags & VM_ALLOC) |
f4527c908 mm/vmalloc.c: rep... |
3404 |
seq_puts(m, " vmalloc"); |
a10aa5798 vmalloc: show vma... |
3405 3406 |
if (v->flags & VM_MAP) |
f4527c908 mm/vmalloc.c: rep... |
3407 |
seq_puts(m, " vmap"); |
a10aa5798 vmalloc: show vma... |
3408 3409 |
if (v->flags & VM_USERMAP) |
f4527c908 mm/vmalloc.c: rep... |
3410 |
seq_puts(m, " user"); |
a10aa5798 vmalloc: show vma... |
3411 |
|
fe9041c24 vmalloc: lift the... |
3412 3413 |
if (v->flags & VM_DMA_COHERENT) seq_puts(m, " dma-coherent"); |
244d63ee3 mm, vmalloc: remo... |
3414 |
if (is_vmalloc_addr(v->pages)) |
f4527c908 mm/vmalloc.c: rep... |
3415 |
seq_puts(m, " vpages"); |
a10aa5798 vmalloc: show vma... |
3416 |
|
a47a126ad vmallocinfo: add ... |
3417 |
show_numa_info(m, v); |
a10aa5798 vmalloc: show vma... |
3418 3419 |
seq_putc(m, ' '); |
dd3b8353b mm/vmalloc: do no... |
3420 3421 3422 3423 3424 3425 3426 3427 3428 |
/* * As a final step, dump "unpurged" areas. Note, * that entire "/proc/vmallocinfo" output will not * be address sorted, because the purge list is not * sorted. */ if (list_is_last(&va->list, &vmap_area_list)) show_purge_info(m); |
a10aa5798 vmalloc: show vma... |
3429 3430 |
return 0; } |
5f6a6a9c4 proc: move /proc/... |
3431 |
static const struct seq_operations vmalloc_op = { |
a10aa5798 vmalloc: show vma... |
3432 3433 3434 3435 3436 |
.start = s_start, .next = s_next, .stop = s_stop, .show = s_show, }; |
5f6a6a9c4 proc: move /proc/... |
3437 |
|
5f6a6a9c4 proc: move /proc/... |
3438 3439 |
static int __init proc_vmalloc_init(void) { |
fddda2b7b proc: introduce p... |
3440 |
if (IS_ENABLED(CONFIG_NUMA)) |
0825a6f98 mm: use octal not... |
3441 |
proc_create_seq_private("vmallocinfo", 0400, NULL, |
44414d82c proc: introduce p... |
3442 3443 |
&vmalloc_op, nr_node_ids * sizeof(unsigned int), NULL); |
fddda2b7b proc: introduce p... |
3444 |
else |
0825a6f98 mm: use octal not... |
3445 |
proc_create_seq("vmallocinfo", 0400, NULL, &vmalloc_op); |
5f6a6a9c4 proc: move /proc/... |
3446 3447 3448 |
return 0; } module_init(proc_vmalloc_init); |
db3808c1b mm, vmalloc: move... |
3449 |
|
a10aa5798 vmalloc: show vma... |
3450 |
#endif |