Blame view
mm/percpu.c
63.4 KB
fbf59bc9d
|
1 2 3 4 5 6 7 8 9 10 |
/* * linux/mm/percpu.c - percpu memory allocator * * Copyright (C) 2009 SUSE Linux Products GmbH * Copyright (C) 2009 Tejun Heo <tj@kernel.org> * * This file is released under the GPLv2. * * This is percpu allocator which can handle both static and dynamic * areas. Percpu areas are allocated in chunks in vmalloc area. Each |
2f39e637e
|
11 12 13 14 15 16 17 |
* chunk is consisted of boot-time determined number of units and the * first chunk is used for static percpu variables in the kernel image * (special boot time alloc/init handling necessary as these areas * need to be brought up before allocation services are running). * Unit grows as necessary and all units grow or shrink in unison. * When a chunk is filled up, another chunk is allocated. ie. in * vmalloc area |
fbf59bc9d
|
18 19 20 21 22 23 24 25 |
* * c0 c1 c2 * ------------------- ------------------- ------------ * | u0 | u1 | u2 | u3 | | u0 | u1 | u2 | u3 | | u0 | u1 | u * ------------------- ...... ------------------- .... ------------ * * Allocation is done in offset-size areas of single unit space. Ie, * an area of 512 bytes at 6k in c1 occupies 512 bytes at 6k of c1:u0, |
2f39e637e
|
26 27 28 29 |
* c1:u1, c1:u2 and c1:u3. On UMA, units corresponds directly to * cpus. On NUMA, the mapping can be non-linear and even sparse. * Percpu access can be done by configuring percpu base registers * according to cpu to unit mapping and pcpu_unit_size. |
fbf59bc9d
|
30 |
* |
2f39e637e
|
31 32 |
* There are usually many small percpu allocations many of them being * as small as 4 bytes. The allocator organizes chunks into lists |
fbf59bc9d
|
33 34 35 36 37 38 39 40 41 42 43 |
* according to free size and tries to allocate from the fullest one. * Each chunk keeps the maximum contiguous area size hint which is * guaranteed to be eqaul to or larger than the maximum contiguous * area in the chunk. This helps the allocator not to iterate the * chunk maps unnecessarily. * * Allocation state in each chunk is kept using an array of integers * on chunk->map. A positive value in the map represents a free * region and negative allocated. Allocation inside a chunk is done * by scanning this map sequentially and serving the first matching * entry. This is mostly copied from the percpu_modalloc() allocator. |
e1b9aa3f4
|
44 45 |
* Chunks can be determined from the address using the index field * in the page struct. The index field contains a pointer to the chunk. |
fbf59bc9d
|
46 47 48 |
* * To use this allocator, arch code should do the followings. * |
fbf59bc9d
|
49 |
* - define __addr_to_pcpu_ptr() and __pcpu_ptr_to_addr() to translate |
e01009833
|
50 51 |
* regular address to percpu pointer and back if they need to be * different from the default |
fbf59bc9d
|
52 |
* |
8d408b4be
|
53 54 |
* - use pcpu_setup_first_chunk() during percpu area initialization to * setup the first chunk containing the kernel static percpu area |
fbf59bc9d
|
55 56 57 58 |
*/ #include <linux/bitmap.h> #include <linux/bootmem.h> |
fd1e8a1fe
|
59 |
#include <linux/err.h> |
fbf59bc9d
|
60 |
#include <linux/list.h> |
a530b7958
|
61 |
#include <linux/log2.h> |
fbf59bc9d
|
62 63 64 65 66 |
#include <linux/mm.h> #include <linux/module.h> #include <linux/mutex.h> #include <linux/percpu.h> #include <linux/pfn.h> |
fbf59bc9d
|
67 |
#include <linux/slab.h> |
ccea34b5d
|
68 |
#include <linux/spinlock.h> |
fbf59bc9d
|
69 |
#include <linux/vmalloc.h> |
a56dbddf0
|
70 |
#include <linux/workqueue.h> |
fbf59bc9d
|
71 72 |
#include <asm/cacheflush.h> |
e01009833
|
73 |
#include <asm/sections.h> |
fbf59bc9d
|
74 |
#include <asm/tlbflush.h> |
3b034b0d0
|
75 |
#include <asm/io.h> |
fbf59bc9d
|
76 |
|
fbf59bc9d
|
77 78 |
#define PCPU_SLOT_BASE_SHIFT 5 /* 1-31 shares the same slot */ #define PCPU_DFL_MAP_ALLOC 16 /* start a map with 16 ents */ |
e01009833
|
79 80 81 |
/* default addr <-> pcpu_ptr mapping, override in asm/percpu.h if necessary */ #ifndef __addr_to_pcpu_ptr #define __addr_to_pcpu_ptr(addr) \ |
43cf38eb5
|
82 83 84 |
(void __percpu *)((unsigned long)(addr) - \ (unsigned long)pcpu_base_addr + \ (unsigned long)__per_cpu_start) |
e01009833
|
85 86 87 |
#endif #ifndef __pcpu_ptr_to_addr #define __pcpu_ptr_to_addr(ptr) \ |
43cf38eb5
|
88 89 90 |
(void __force *)((unsigned long)(ptr) + \ (unsigned long)pcpu_base_addr - \ (unsigned long)__per_cpu_start) |
e01009833
|
91 |
#endif |
fbf59bc9d
|
92 93 |
struct pcpu_chunk { struct list_head list; /* linked to pcpu_slot lists */ |
fbf59bc9d
|
94 95 |
int free_size; /* free bytes in the chunk */ int contig_hint; /* max contiguous size hint */ |
bba174f5e
|
96 |
void *base_addr; /* base address of this chunk */ |
fbf59bc9d
|
97 98 99 |
int map_used; /* # of map entries used */ int map_alloc; /* # of map entries allocated */ int *map; /* allocation map */ |
6563297ce
|
100 |
struct vm_struct **vms; /* mapped vmalloc regions */ |
8d408b4be
|
101 |
bool immutable; /* no [de]population allowed */ |
ce3141a27
|
102 |
unsigned long populated[]; /* populated bitmap */ |
fbf59bc9d
|
103 |
}; |
40150d37b
|
104 105 |
static int pcpu_unit_pages __read_mostly; static int pcpu_unit_size __read_mostly; |
2f39e637e
|
106 |
static int pcpu_nr_units __read_mostly; |
6563297ce
|
107 |
static int pcpu_atom_size __read_mostly; |
40150d37b
|
108 109 |
static int pcpu_nr_slots __read_mostly; static size_t pcpu_chunk_struct_size __read_mostly; |
fbf59bc9d
|
110 |
|
2f39e637e
|
111 112 113 |
/* cpus with the lowest and highest unit numbers */ static unsigned int pcpu_first_unit_cpu __read_mostly; static unsigned int pcpu_last_unit_cpu __read_mostly; |
fbf59bc9d
|
114 |
/* the address of the first chunk which starts with the kernel static area */ |
40150d37b
|
115 |
void *pcpu_base_addr __read_mostly; |
fbf59bc9d
|
116 |
EXPORT_SYMBOL_GPL(pcpu_base_addr); |
fb435d523
|
117 118 |
static const int *pcpu_unit_map __read_mostly; /* cpu -> unit */ const unsigned long *pcpu_unit_offsets __read_mostly; /* cpu -> unit offset */ |
2f39e637e
|
119 |
|
6563297ce
|
120 121 122 123 |
/* group information, used for vm allocation */ static int pcpu_nr_groups __read_mostly; static const unsigned long *pcpu_group_offsets __read_mostly; static const size_t *pcpu_group_sizes __read_mostly; |
ae9e6bc9f
|
124 125 126 127 128 129 130 131 132 133 134 135 136 137 |
/* * The first chunk which always exists. Note that unlike other * chunks, this one can be allocated and mapped in several different * ways and thus often doesn't live in the vmalloc area. */ static struct pcpu_chunk *pcpu_first_chunk; /* * Optional reserved chunk. This chunk reserves part of the first * chunk and serves it for reserved allocations. The amount of * reserved offset is in pcpu_reserved_chunk_limit. When reserved * area doesn't exist, the following variables contain NULL and 0 * respectively. */ |
edcb46399
|
138 |
static struct pcpu_chunk *pcpu_reserved_chunk; |
edcb46399
|
139 |
static int pcpu_reserved_chunk_limit; |
fbf59bc9d
|
140 |
/* |
ccea34b5d
|
141 142 143 |
* Synchronization rules. * * There are two locks - pcpu_alloc_mutex and pcpu_lock. The former |
ce3141a27
|
144 145 146 |
* protects allocation/reclaim paths, chunks, populated bitmap and * vmalloc mapping. The latter is a spinlock and protects the index * data structures - chunk slots, chunks and area maps in chunks. |
ccea34b5d
|
147 148 149 |
* * During allocation, pcpu_alloc_mutex is kept locked all the time and * pcpu_lock is grabbed and released as necessary. All actual memory |
403a91b16
|
150 151 152 153 |
* allocations are done using GFP_KERNEL with pcpu_lock released. In * general, percpu memory can't be allocated with irq off but * irqsave/restore are still used in alloc path so that it can be used * from early init path - sched_init() specifically. |
ccea34b5d
|
154 155 156 157 158 159 160 161 162 |
* * Free path accesses and alters only the index data structures, so it * can be safely called from atomic context. When memory needs to be * returned to the system, free path schedules reclaim_work which * grabs both pcpu_alloc_mutex and pcpu_lock, unlinks chunks to be * reclaimed, release both locks and frees the chunks. Note that it's * necessary to grab both locks to remove a chunk from circulation as * allocation path might be referencing the chunk with only * pcpu_alloc_mutex locked. |
fbf59bc9d
|
163 |
*/ |
ccea34b5d
|
164 165 |
static DEFINE_MUTEX(pcpu_alloc_mutex); /* protects whole alloc and reclaim */ static DEFINE_SPINLOCK(pcpu_lock); /* protects index data structures */ |
fbf59bc9d
|
166 |
|
40150d37b
|
167 |
static struct list_head *pcpu_slot __read_mostly; /* chunk list slots */ |
fbf59bc9d
|
168 |
|
a56dbddf0
|
169 170 171 |
/* reclaim work to release fully free chunks, scheduled from free path */ static void pcpu_reclaim(struct work_struct *work); static DECLARE_WORK(pcpu_reclaim_work, pcpu_reclaim); |
d9b55eeb1
|
172 |
static int __pcpu_size_to_slot(int size) |
fbf59bc9d
|
173 |
{ |
cae3aeb83
|
174 |
int highbit = fls(size); /* size is in bytes */ |
fbf59bc9d
|
175 176 |
return max(highbit - PCPU_SLOT_BASE_SHIFT + 2, 1); } |
d9b55eeb1
|
177 178 179 180 181 182 |
static int pcpu_size_to_slot(int size) { if (size == pcpu_unit_size) return pcpu_nr_slots - 1; return __pcpu_size_to_slot(size); } |
fbf59bc9d
|
183 184 185 186 187 188 189 190 191 192 |
static int pcpu_chunk_slot(const struct pcpu_chunk *chunk) { if (chunk->free_size < sizeof(int) || chunk->contig_hint < sizeof(int)) return 0; return pcpu_size_to_slot(chunk->free_size); } static int pcpu_page_idx(unsigned int cpu, int page_idx) { |
2f39e637e
|
193 |
return pcpu_unit_map[cpu] * pcpu_unit_pages + page_idx; |
fbf59bc9d
|
194 195 196 197 198 |
} static unsigned long pcpu_chunk_addr(struct pcpu_chunk *chunk, unsigned int cpu, int page_idx) { |
bba174f5e
|
199 |
return (unsigned long)chunk->base_addr + pcpu_unit_offsets[cpu] + |
fb435d523
|
200 |
(page_idx << PAGE_SHIFT); |
fbf59bc9d
|
201 |
} |
ce3141a27
|
202 203 |
static struct page *pcpu_chunk_page(struct pcpu_chunk *chunk, unsigned int cpu, int page_idx) |
fbf59bc9d
|
204 |
{ |
ce3141a27
|
205 206 |
/* must not be used on pre-mapped chunk */ WARN_ON(chunk->immutable); |
c8a51be4c
|
207 |
|
ce3141a27
|
208 |
return vmalloc_to_page((void *)pcpu_chunk_addr(chunk, cpu, page_idx)); |
fbf59bc9d
|
209 |
} |
e1b9aa3f4
|
210 211 212 213 214 215 216 217 218 219 220 |
/* set the pointer to a chunk in a page struct */ static void pcpu_set_page_chunk(struct page *page, struct pcpu_chunk *pcpu) { page->index = (unsigned long)pcpu; } /* obtain pointer to a chunk from a page struct */ static struct pcpu_chunk *pcpu_get_page_chunk(struct page *page) { return (struct pcpu_chunk *)page->index; } |
ce3141a27
|
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 |
static void pcpu_next_unpop(struct pcpu_chunk *chunk, int *rs, int *re, int end) { *rs = find_next_zero_bit(chunk->populated, end, *rs); *re = find_next_bit(chunk->populated, end, *rs + 1); } static void pcpu_next_pop(struct pcpu_chunk *chunk, int *rs, int *re, int end) { *rs = find_next_bit(chunk->populated, end, *rs); *re = find_next_zero_bit(chunk->populated, end, *rs + 1); } /* * (Un)populated page region iterators. Iterate over (un)populated * page regions betwen @start and @end in @chunk. @rs and @re should * be integer variables and will be set to start and end page index of * the current region. */ #define pcpu_for_each_unpop_region(chunk, rs, re, start, end) \ for ((rs) = (start), pcpu_next_unpop((chunk), &(rs), &(re), (end)); \ (rs) < (re); \ (rs) = (re) + 1, pcpu_next_unpop((chunk), &(rs), &(re), (end))) #define pcpu_for_each_pop_region(chunk, rs, re, start, end) \ for ((rs) = (start), pcpu_next_pop((chunk), &(rs), &(re), (end)); \ (rs) < (re); \ (rs) = (re) + 1, pcpu_next_pop((chunk), &(rs), &(re), (end))) |
fbf59bc9d
|
248 |
/** |
1880d93b8
|
249 250 |
* pcpu_mem_alloc - allocate memory * @size: bytes to allocate |
fbf59bc9d
|
251 |
* |
1880d93b8
|
252 253 254 |
* Allocate @size bytes. If @size is smaller than PAGE_SIZE, * kzalloc() is used; otherwise, vmalloc() is used. The returned * memory is always zeroed. |
fbf59bc9d
|
255 |
* |
ccea34b5d
|
256 257 258 |
* CONTEXT: * Does GFP_KERNEL allocation. * |
fbf59bc9d
|
259 |
* RETURNS: |
1880d93b8
|
260 |
* Pointer to the allocated area on success, NULL on failure. |
fbf59bc9d
|
261 |
*/ |
1880d93b8
|
262 |
static void *pcpu_mem_alloc(size_t size) |
fbf59bc9d
|
263 |
{ |
1880d93b8
|
264 265 266 267 268 269 270 271 272 |
if (size <= PAGE_SIZE) return kzalloc(size, GFP_KERNEL); else { void *ptr = vmalloc(size); if (ptr) memset(ptr, 0, size); return ptr; } } |
fbf59bc9d
|
273 |
|
1880d93b8
|
274 275 276 277 278 279 280 281 282 |
/** * pcpu_mem_free - free memory * @ptr: memory to free * @size: size of the area * * Free @ptr. @ptr should have been allocated using pcpu_mem_alloc(). */ static void pcpu_mem_free(void *ptr, size_t size) { |
fbf59bc9d
|
283 |
if (size <= PAGE_SIZE) |
1880d93b8
|
284 |
kfree(ptr); |
fbf59bc9d
|
285 |
else |
1880d93b8
|
286 |
vfree(ptr); |
fbf59bc9d
|
287 288 289 290 291 292 293 294 295 |
} /** * pcpu_chunk_relocate - put chunk in the appropriate chunk slot * @chunk: chunk of interest * @oslot: the previous slot it was on * * This function is called after an allocation or free changed @chunk. * New slot according to the changed state is determined and @chunk is |
edcb46399
|
296 297 |
* moved to the slot. Note that the reserved chunk is never put on * chunk slots. |
ccea34b5d
|
298 299 300 |
* * CONTEXT: * pcpu_lock. |
fbf59bc9d
|
301 302 303 304 |
*/ static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot) { int nslot = pcpu_chunk_slot(chunk); |
edcb46399
|
305 |
if (chunk != pcpu_reserved_chunk && oslot != nslot) { |
fbf59bc9d
|
306 307 308 309 310 311 |
if (oslot < nslot) list_move(&chunk->list, &pcpu_slot[nslot]); else list_move_tail(&chunk->list, &pcpu_slot[nslot]); } } |
fbf59bc9d
|
312 |
/** |
e1b9aa3f4
|
313 314 |
* pcpu_chunk_addr_search - determine chunk containing specified address * @addr: address for which the chunk needs to be determined. |
ccea34b5d
|
315 |
* |
fbf59bc9d
|
316 317 318 319 320 |
* RETURNS: * The address of the found chunk. */ static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr) { |
bba174f5e
|
321 |
void *first_start = pcpu_first_chunk->base_addr; |
fbf59bc9d
|
322 |
|
ae9e6bc9f
|
323 |
/* is it in the first chunk? */ |
79ba6ac82
|
324 |
if (addr >= first_start && addr < first_start + pcpu_unit_size) { |
ae9e6bc9f
|
325 326 |
/* is it in the reserved area? */ if (addr < first_start + pcpu_reserved_chunk_limit) |
edcb46399
|
327 |
return pcpu_reserved_chunk; |
ae9e6bc9f
|
328 |
return pcpu_first_chunk; |
edcb46399
|
329 |
} |
04a13c7c6
|
330 331 332 333 334 335 336 |
/* * The address is relative to unit0 which might be unused and * thus unmapped. Offset the address to the unit space of the * current processor before looking it up in the vmalloc * space. Note that any possible cpu id can be used here, so * there's no need to worry about preemption or cpu hotplug. */ |
5579fd7e6
|
337 |
addr += pcpu_unit_offsets[raw_smp_processor_id()]; |
e1b9aa3f4
|
338 |
return pcpu_get_page_chunk(vmalloc_to_page(addr)); |
fbf59bc9d
|
339 340 341 |
} /** |
833af8427
|
342 343 |
* pcpu_need_to_extend - determine whether chunk area map needs to be extended * @chunk: chunk of interest |
9f7dcf224
|
344 |
* |
833af8427
|
345 346 |
* Determine whether area map of @chunk needs to be extended to * accomodate a new allocation. |
9f7dcf224
|
347 |
* |
ccea34b5d
|
348 |
* CONTEXT: |
833af8427
|
349 |
* pcpu_lock. |
ccea34b5d
|
350 |
* |
9f7dcf224
|
351 |
* RETURNS: |
833af8427
|
352 353 |
* New target map allocation length if extension is necessary, 0 * otherwise. |
9f7dcf224
|
354 |
*/ |
833af8427
|
355 |
static int pcpu_need_to_extend(struct pcpu_chunk *chunk) |
9f7dcf224
|
356 357 |
{ int new_alloc; |
9f7dcf224
|
358 |
|
9f7dcf224
|
359 360 361 362 363 364 |
if (chunk->map_alloc >= chunk->map_used + 2) return 0; new_alloc = PCPU_DFL_MAP_ALLOC; while (new_alloc < chunk->map_used + 2) new_alloc *= 2; |
833af8427
|
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 |
return new_alloc; } /** * pcpu_extend_area_map - extend area map of a chunk * @chunk: chunk of interest * @new_alloc: new target allocation length of the area map * * Extend area map of @chunk to have @new_alloc entries. * * CONTEXT: * Does GFP_KERNEL allocation. Grabs and releases pcpu_lock. * * RETURNS: * 0 on success, -errno on failure. */ static int pcpu_extend_area_map(struct pcpu_chunk *chunk, int new_alloc) { int *old = NULL, *new = NULL; size_t old_size = 0, new_size = new_alloc * sizeof(new[0]); unsigned long flags; new = pcpu_mem_alloc(new_size); if (!new) |
9f7dcf224
|
389 |
return -ENOMEM; |
ccea34b5d
|
390 |
|
833af8427
|
391 392 393 394 395 |
/* acquire pcpu_lock and switch to new area map */ spin_lock_irqsave(&pcpu_lock, flags); if (new_alloc <= chunk->map_alloc) goto out_unlock; |
9f7dcf224
|
396 |
|
833af8427
|
397 398 |
old_size = chunk->map_alloc * sizeof(chunk->map[0]); memcpy(new, chunk->map, old_size); |
9f7dcf224
|
399 400 401 402 403 404 |
/* * map_alloc < PCPU_DFL_MAP_ALLOC indicates that the chunk is * one of the first chunks and still using static map. */ if (chunk->map_alloc >= PCPU_DFL_MAP_ALLOC) |
833af8427
|
405 |
old = chunk->map; |
9f7dcf224
|
406 407 408 |
chunk->map_alloc = new_alloc; chunk->map = new; |
833af8427
|
409 410 411 412 413 414 415 416 417 418 419 |
new = NULL; out_unlock: spin_unlock_irqrestore(&pcpu_lock, flags); /* * pcpu_mem_free() might end up calling vfree() which uses * IRQ-unsafe lock and thus can't be called under pcpu_lock. */ pcpu_mem_free(old, old_size); pcpu_mem_free(new, new_size); |
9f7dcf224
|
420 421 422 423 |
return 0; } /** |
fbf59bc9d
|
424 425 426 |
* pcpu_split_block - split a map block * @chunk: chunk of interest * @i: index of map block to split |
cae3aeb83
|
427 428 |
* @head: head size in bytes (can be 0) * @tail: tail size in bytes (can be 0) |
fbf59bc9d
|
429 430 431 432 433 434 435 436 437 |
* * Split the @i'th map block into two or three blocks. If @head is * non-zero, @head bytes block is inserted before block @i moving it * to @i+1 and reducing its size by @head bytes. * * If @tail is non-zero, the target block, which can be @i or @i+1 * depending on @head, is reduced by @tail bytes and @tail byte block * is inserted after the target block. * |
9f7dcf224
|
438 |
* @chunk->map must have enough free slots to accomodate the split. |
ccea34b5d
|
439 440 441 |
* * CONTEXT: * pcpu_lock. |
fbf59bc9d
|
442 |
*/ |
9f7dcf224
|
443 444 |
static void pcpu_split_block(struct pcpu_chunk *chunk, int i, int head, int tail) |
fbf59bc9d
|
445 446 |
{ int nr_extra = !!head + !!tail; |
1880d93b8
|
447 |
|
9f7dcf224
|
448 |
BUG_ON(chunk->map_alloc < chunk->map_used + nr_extra); |
fbf59bc9d
|
449 |
|
9f7dcf224
|
450 |
/* insert new subblocks */ |
fbf59bc9d
|
451 452 453 454 455 456 457 458 459 460 461 462 |
memmove(&chunk->map[i + nr_extra], &chunk->map[i], sizeof(chunk->map[0]) * (chunk->map_used - i)); chunk->map_used += nr_extra; if (head) { chunk->map[i + 1] = chunk->map[i] - head; chunk->map[i++] = head; } if (tail) { chunk->map[i++] -= tail; chunk->map[i] = tail; } |
fbf59bc9d
|
463 464 465 466 467 |
} /** * pcpu_alloc_area - allocate area from a pcpu_chunk * @chunk: chunk of interest |
cae3aeb83
|
468 |
* @size: wanted size in bytes |
fbf59bc9d
|
469 470 471 472 473 474 |
* @align: wanted align * * Try to allocate @size bytes area aligned at @align from @chunk. * Note that this function only allocates the offset. It doesn't * populate or map the area. * |
9f7dcf224
|
475 476 |
* @chunk->map must have at least two free slots. * |
ccea34b5d
|
477 478 479 |
* CONTEXT: * pcpu_lock. * |
fbf59bc9d
|
480 |
* RETURNS: |
9f7dcf224
|
481 482 |
* Allocated offset in @chunk on success, -1 if no matching area is * found. |
fbf59bc9d
|
483 484 485 486 487 488 |
*/ static int pcpu_alloc_area(struct pcpu_chunk *chunk, int size, int align) { int oslot = pcpu_chunk_slot(chunk); int max_contig = 0; int i, off; |
fbf59bc9d
|
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 |
for (i = 0, off = 0; i < chunk->map_used; off += abs(chunk->map[i++])) { bool is_last = i + 1 == chunk->map_used; int head, tail; /* extra for alignment requirement */ head = ALIGN(off, align) - off; BUG_ON(i == 0 && head != 0); if (chunk->map[i] < 0) continue; if (chunk->map[i] < head + size) { max_contig = max(chunk->map[i], max_contig); continue; } /* * If head is small or the previous block is free, * merge'em. Note that 'small' is defined as smaller * than sizeof(int), which is very small but isn't too * uncommon for percpu allocations. */ if (head && (head < sizeof(int) || chunk->map[i - 1] > 0)) { if (chunk->map[i - 1] > 0) chunk->map[i - 1] += head; else { chunk->map[i - 1] -= head; chunk->free_size -= head; } chunk->map[i] -= head; off += head; head = 0; } /* if tail is small, just keep it around */ tail = chunk->map[i] - head - size; if (tail < sizeof(int)) tail = 0; /* split if warranted */ if (head || tail) { |
9f7dcf224
|
529 |
pcpu_split_block(chunk, i, head, tail); |
fbf59bc9d
|
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 |
if (head) { i++; off += head; max_contig = max(chunk->map[i - 1], max_contig); } if (tail) max_contig = max(chunk->map[i + 1], max_contig); } /* update hint and mark allocated */ if (is_last) chunk->contig_hint = max_contig; /* fully scanned */ else chunk->contig_hint = max(chunk->contig_hint, max_contig); chunk->free_size -= chunk->map[i]; chunk->map[i] = -chunk->map[i]; pcpu_chunk_relocate(chunk, oslot); return off; } chunk->contig_hint = max_contig; /* fully scanned */ pcpu_chunk_relocate(chunk, oslot); |
9f7dcf224
|
555 556 |
/* tell the upper layer that this chunk has no matching area */ return -1; |
fbf59bc9d
|
557 558 559 560 561 562 563 564 565 566 |
} /** * pcpu_free_area - free area to a pcpu_chunk * @chunk: chunk of interest * @freeme: offset of area to free * * Free area starting from @freeme to @chunk. Note that this function * only modifies the allocation map. It doesn't depopulate or unmap * the area. |
ccea34b5d
|
567 568 569 |
* * CONTEXT: * pcpu_lock. |
fbf59bc9d
|
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 |
*/ static void pcpu_free_area(struct pcpu_chunk *chunk, int freeme) { int oslot = pcpu_chunk_slot(chunk); int i, off; for (i = 0, off = 0; i < chunk->map_used; off += abs(chunk->map[i++])) if (off == freeme) break; BUG_ON(off != freeme); BUG_ON(chunk->map[i] > 0); chunk->map[i] = -chunk->map[i]; chunk->free_size += chunk->map[i]; /* merge with previous? */ if (i > 0 && chunk->map[i - 1] >= 0) { chunk->map[i - 1] += chunk->map[i]; chunk->map_used--; memmove(&chunk->map[i], &chunk->map[i + 1], (chunk->map_used - i) * sizeof(chunk->map[0])); i--; } /* merge with next? */ if (i + 1 < chunk->map_used && chunk->map[i + 1] >= 0) { chunk->map[i] += chunk->map[i + 1]; chunk->map_used--; memmove(&chunk->map[i + 1], &chunk->map[i + 2], (chunk->map_used - (i + 1)) * sizeof(chunk->map[0])); } chunk->contig_hint = max(chunk->map[i], chunk->contig_hint); pcpu_chunk_relocate(chunk, oslot); } /** |
ce3141a27
|
606 |
* pcpu_get_pages_and_bitmap - get temp pages array and bitmap |
fbf59bc9d
|
607 |
* @chunk: chunk of interest |
ce3141a27
|
608 609 |
* @bitmapp: output parameter for bitmap * @may_alloc: may allocate the array |
fbf59bc9d
|
610 |
* |
ce3141a27
|
611 612 613 614 615 616 617 618 619 620 621 622 |
* Returns pointer to array of pointers to struct page and bitmap, * both of which can be indexed with pcpu_page_idx(). The returned * array is cleared to zero and *@bitmapp is copied from * @chunk->populated. Note that there is only one array and bitmap * and access exclusion is the caller's responsibility. * * CONTEXT: * pcpu_alloc_mutex and does GFP_KERNEL allocation if @may_alloc. * Otherwise, don't care. * * RETURNS: * Pointer to temp pages array on success, NULL on failure. |
fbf59bc9d
|
623 |
*/ |
ce3141a27
|
624 625 626 |
static struct page **pcpu_get_pages_and_bitmap(struct pcpu_chunk *chunk, unsigned long **bitmapp, bool may_alloc) |
fbf59bc9d
|
627 |
{ |
ce3141a27
|
628 629 |
static struct page **pages; static unsigned long *bitmap; |
2f39e637e
|
630 |
size_t pages_size = pcpu_nr_units * pcpu_unit_pages * sizeof(pages[0]); |
ce3141a27
|
631 632 633 634 635 636 637 638 639 640 641 |
size_t bitmap_size = BITS_TO_LONGS(pcpu_unit_pages) * sizeof(unsigned long); if (!pages || !bitmap) { if (may_alloc && !pages) pages = pcpu_mem_alloc(pages_size); if (may_alloc && !bitmap) bitmap = pcpu_mem_alloc(bitmap_size); if (!pages || !bitmap) return NULL; } |
fbf59bc9d
|
642 |
|
ce3141a27
|
643 644 |
memset(pages, 0, pages_size); bitmap_copy(bitmap, chunk->populated, pcpu_unit_pages); |
8d408b4be
|
645 |
|
ce3141a27
|
646 647 648 |
*bitmapp = bitmap; return pages; } |
fbf59bc9d
|
649 |
|
ce3141a27
|
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 |
/** * pcpu_free_pages - free pages which were allocated for @chunk * @chunk: chunk pages were allocated for * @pages: array of pages to be freed, indexed by pcpu_page_idx() * @populated: populated bitmap * @page_start: page index of the first page to be freed * @page_end: page index of the last page to be freed + 1 * * Free pages [@page_start and @page_end) in @pages for all units. * The pages were allocated for @chunk. */ static void pcpu_free_pages(struct pcpu_chunk *chunk, struct page **pages, unsigned long *populated, int page_start, int page_end) { unsigned int cpu; int i; for_each_possible_cpu(cpu) { for (i = page_start; i < page_end; i++) { struct page *page = pages[pcpu_page_idx(cpu, i)]; if (page) __free_page(page); } } |
fbf59bc9d
|
676 677 678 |
} /** |
ce3141a27
|
679 680 681 682 683 684 685 686 687 688 |
* pcpu_alloc_pages - allocates pages for @chunk * @chunk: target chunk * @pages: array to put the allocated pages into, indexed by pcpu_page_idx() * @populated: populated bitmap * @page_start: page index of the first page to be allocated * @page_end: page index of the last page to be allocated + 1 * * Allocate pages [@page_start,@page_end) into @pages for all units. * The allocation is for @chunk. Percpu core doesn't care about the * content of @pages and will pass it verbatim to pcpu_map_pages(). |
fbf59bc9d
|
689 |
*/ |
ce3141a27
|
690 691 692 |
static int pcpu_alloc_pages(struct pcpu_chunk *chunk, struct page **pages, unsigned long *populated, int page_start, int page_end) |
fbf59bc9d
|
693 |
{ |
ce3141a27
|
694 |
const gfp_t gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_COLD; |
fbf59bc9d
|
695 696 |
unsigned int cpu; int i; |
ce3141a27
|
697 698 699 700 701 702 703 704 705 706 707 708 709 710 |
for_each_possible_cpu(cpu) { for (i = page_start; i < page_end; i++) { struct page **pagep = &pages[pcpu_page_idx(cpu, i)]; *pagep = alloc_pages_node(cpu_to_node(cpu), gfp, 0); if (!*pagep) { pcpu_free_pages(chunk, pages, populated, page_start, page_end); return -ENOMEM; } } } return 0; } |
fbf59bc9d
|
711 |
|
ce3141a27
|
712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 |
/** * pcpu_pre_unmap_flush - flush cache prior to unmapping * @chunk: chunk the regions to be flushed belongs to * @page_start: page index of the first page to be flushed * @page_end: page index of the last page to be flushed + 1 * * Pages in [@page_start,@page_end) of @chunk are about to be * unmapped. Flush cache. As each flushing trial can be very * expensive, issue flush on the whole region at once rather than * doing it for each cpu. This could be an overkill but is more * scalable. */ static void pcpu_pre_unmap_flush(struct pcpu_chunk *chunk, int page_start, int page_end) { |
2f39e637e
|
727 728 729 |
flush_cache_vunmap( pcpu_chunk_addr(chunk, pcpu_first_unit_cpu, page_start), pcpu_chunk_addr(chunk, pcpu_last_unit_cpu, page_end)); |
ce3141a27
|
730 731 732 733 734 735 |
} static void __pcpu_unmap_pages(unsigned long addr, int nr_pages) { unmap_kernel_range_noflush(addr, nr_pages << PAGE_SHIFT); } |
fbf59bc9d
|
736 |
|
ce3141a27
|
737 738 |
/** * pcpu_unmap_pages - unmap pages out of a pcpu_chunk |
fbf59bc9d
|
739 |
* @chunk: chunk of interest |
ce3141a27
|
740 741 |
* @pages: pages array which can be used to pass information to free * @populated: populated bitmap |
fbf59bc9d
|
742 743 |
* @page_start: page index of the first page to unmap * @page_end: page index of the last page to unmap + 1 |
fbf59bc9d
|
744 745 |
* * For each cpu, unmap pages [@page_start,@page_end) out of @chunk. |
ce3141a27
|
746 747 748 749 |
* Corresponding elements in @pages were cleared by the caller and can * be used to carry information to pcpu_free_pages() which will be * called after all unmaps are finished. The caller should call * proper pre/post flush functions. |
fbf59bc9d
|
750 |
*/ |
ce3141a27
|
751 752 753 |
static void pcpu_unmap_pages(struct pcpu_chunk *chunk, struct page **pages, unsigned long *populated, int page_start, int page_end) |
fbf59bc9d
|
754 |
{ |
fbf59bc9d
|
755 |
unsigned int cpu; |
ce3141a27
|
756 |
int i; |
fbf59bc9d
|
757 |
|
ce3141a27
|
758 759 760 |
for_each_possible_cpu(cpu) { for (i = page_start; i < page_end; i++) { struct page *page; |
fbf59bc9d
|
761 |
|
ce3141a27
|
762 763 764 |
page = pcpu_chunk_page(chunk, cpu, i); WARN_ON(!page); pages[pcpu_page_idx(cpu, i)] = page; |
fbf59bc9d
|
765 |
} |
ce3141a27
|
766 767 |
__pcpu_unmap_pages(pcpu_chunk_addr(chunk, cpu, page_start), page_end - page_start); |
fbf59bc9d
|
768 |
} |
ce3141a27
|
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 |
for (i = page_start; i < page_end; i++) __clear_bit(i, populated); } /** * pcpu_post_unmap_tlb_flush - flush TLB after unmapping * @chunk: pcpu_chunk the regions to be flushed belong to * @page_start: page index of the first page to be flushed * @page_end: page index of the last page to be flushed + 1 * * Pages [@page_start,@page_end) of @chunk have been unmapped. Flush * TLB for the regions. This can be skipped if the area is to be * returned to vmalloc as vmalloc will handle TLB flushing lazily. * * As with pcpu_pre_unmap_flush(), TLB flushing also is done at once * for the whole region. */ static void pcpu_post_unmap_tlb_flush(struct pcpu_chunk *chunk, int page_start, int page_end) { |
2f39e637e
|
789 790 791 |
flush_tlb_kernel_range( pcpu_chunk_addr(chunk, pcpu_first_unit_cpu, page_start), pcpu_chunk_addr(chunk, pcpu_last_unit_cpu, page_end)); |
fbf59bc9d
|
792 |
} |
c8a51be4c
|
793 794 795 796 797 |
static int __pcpu_map_pages(unsigned long addr, struct page **pages, int nr_pages) { return map_kernel_range_noflush(addr, nr_pages << PAGE_SHIFT, PAGE_KERNEL, pages); |
fbf59bc9d
|
798 799 800 |
} /** |
ce3141a27
|
801 |
* pcpu_map_pages - map pages into a pcpu_chunk |
fbf59bc9d
|
802 |
* @chunk: chunk of interest |
ce3141a27
|
803 804 |
* @pages: pages array containing pages to be mapped * @populated: populated bitmap |
fbf59bc9d
|
805 806 807 |
* @page_start: page index of the first page to map * @page_end: page index of the last page to map + 1 * |
ce3141a27
|
808 809 810 811 812 813 814 |
* For each cpu, map pages [@page_start,@page_end) into @chunk. The * caller is responsible for calling pcpu_post_map_flush() after all * mappings are complete. * * This function is responsible for setting corresponding bits in * @chunk->populated bitmap and whatever is necessary for reverse * lookup (addr -> chunk). |
fbf59bc9d
|
815 |
*/ |
ce3141a27
|
816 817 818 |
static int pcpu_map_pages(struct pcpu_chunk *chunk, struct page **pages, unsigned long *populated, int page_start, int page_end) |
fbf59bc9d
|
819 |
{ |
ce3141a27
|
820 821 |
unsigned int cpu, tcpu; int i, err; |
8d408b4be
|
822 |
|
fbf59bc9d
|
823 |
for_each_possible_cpu(cpu) { |
c8a51be4c
|
824 |
err = __pcpu_map_pages(pcpu_chunk_addr(chunk, cpu, page_start), |
ce3141a27
|
825 |
&pages[pcpu_page_idx(cpu, page_start)], |
c8a51be4c
|
826 |
page_end - page_start); |
fbf59bc9d
|
827 |
if (err < 0) |
ce3141a27
|
828 |
goto err; |
c8a51be4c
|
829 |
} |
ce3141a27
|
830 831 832 833 834 835 |
/* mapping successful, link chunk and mark populated */ for (i = page_start; i < page_end; i++) { for_each_possible_cpu(cpu) pcpu_set_page_chunk(pages[pcpu_page_idx(cpu, i)], chunk); __set_bit(i, populated); |
fbf59bc9d
|
836 |
} |
fbf59bc9d
|
837 |
return 0; |
ce3141a27
|
838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 |
err: for_each_possible_cpu(tcpu) { if (tcpu == cpu) break; __pcpu_unmap_pages(pcpu_chunk_addr(chunk, tcpu, page_start), page_end - page_start); } return err; } /** * pcpu_post_map_flush - flush cache after mapping * @chunk: pcpu_chunk the regions to be flushed belong to * @page_start: page index of the first page to be flushed * @page_end: page index of the last page to be flushed + 1 * * Pages [@page_start,@page_end) of @chunk have been mapped. Flush * cache. * * As with pcpu_pre_unmap_flush(), TLB flushing also is done at once * for the whole region. */ static void pcpu_post_map_flush(struct pcpu_chunk *chunk, int page_start, int page_end) { |
2f39e637e
|
864 865 866 |
flush_cache_vmap( pcpu_chunk_addr(chunk, pcpu_first_unit_cpu, page_start), pcpu_chunk_addr(chunk, pcpu_last_unit_cpu, page_end)); |
c8a51be4c
|
867 |
} |
fbf59bc9d
|
868 869 870 871 |
/** * pcpu_depopulate_chunk - depopulate and unmap an area of a pcpu_chunk * @chunk: chunk to depopulate * @off: offset to the area to depopulate |
cae3aeb83
|
872 |
* @size: size of the area to depopulate in bytes |
fbf59bc9d
|
873 874 875 876 877 |
* @flush: whether to flush cache and tlb or not * * For each cpu, depopulate and unmap pages [@page_start,@page_end) * from @chunk. If @flush is true, vcache is flushed before unmapping * and tlb after. |
ccea34b5d
|
878 879 880 |
* * CONTEXT: * pcpu_alloc_mutex. |
fbf59bc9d
|
881 |
*/ |
ce3141a27
|
882 |
static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk, int off, int size) |
fbf59bc9d
|
883 884 885 |
{ int page_start = PFN_DOWN(off); int page_end = PFN_UP(off + size); |
ce3141a27
|
886 887 888 889 890 |
struct page **pages; unsigned long *populated; int rs, re; /* quick path, check whether it's empty already */ |
22b737f4c
|
891 892 893 894 |
rs = page_start; pcpu_next_unpop(chunk, &rs, &re, page_end); if (rs == page_start && re == page_end) return; |
fbf59bc9d
|
895 |
|
ce3141a27
|
896 897 |
/* immutable chunks can't be depopulated */ WARN_ON(chunk->immutable); |
fbf59bc9d
|
898 |
|
ce3141a27
|
899 900 901 902 903 904 905 |
/* * If control reaches here, there must have been at least one * successful population attempt so the temp pages array must * be available now. */ pages = pcpu_get_pages_and_bitmap(chunk, &populated, false); BUG_ON(!pages); |
fbf59bc9d
|
906 |
|
ce3141a27
|
907 908 |
/* unmap and free */ pcpu_pre_unmap_flush(chunk, page_start, page_end); |
fbf59bc9d
|
909 |
|
ce3141a27
|
910 911 |
pcpu_for_each_pop_region(chunk, rs, re, page_start, page_end) pcpu_unmap_pages(chunk, pages, populated, rs, re); |
fbf59bc9d
|
912 |
|
ce3141a27
|
913 914 915 916 |
/* no need to flush tlb, vmalloc will handle it lazily */ pcpu_for_each_pop_region(chunk, rs, re, page_start, page_end) pcpu_free_pages(chunk, pages, populated, rs, re); |
fbf59bc9d
|
917 |
|
ce3141a27
|
918 919 |
/* commit new bitmap */ bitmap_copy(chunk->populated, populated, pcpu_unit_pages); |
fbf59bc9d
|
920 921 922 923 924 925 |
} /** * pcpu_populate_chunk - populate and map an area of a pcpu_chunk * @chunk: chunk of interest * @off: offset to the area to populate |
cae3aeb83
|
926 |
* @size: size of the area to populate in bytes |
fbf59bc9d
|
927 928 929 |
* * For each cpu, populate and map pages [@page_start,@page_end) into * @chunk. The area is cleared on return. |
ccea34b5d
|
930 931 932 |
* * CONTEXT: * pcpu_alloc_mutex, does GFP_KERNEL allocation. |
fbf59bc9d
|
933 934 935 |
*/ static int pcpu_populate_chunk(struct pcpu_chunk *chunk, int off, int size) { |
fbf59bc9d
|
936 937 |
int page_start = PFN_DOWN(off); int page_end = PFN_UP(off + size); |
ce3141a27
|
938 939 940 |
int free_end = page_start, unmap_end = page_start; struct page **pages; unsigned long *populated; |
fbf59bc9d
|
941 |
unsigned int cpu; |
ce3141a27
|
942 |
int rs, re, rc; |
fbf59bc9d
|
943 |
|
ce3141a27
|
944 |
/* quick path, check whether all pages are already there */ |
22b737f4c
|
945 946 947 948 |
rs = page_start; pcpu_next_pop(chunk, &rs, &re, page_end); if (rs == page_start && re == page_end) goto clear; |
fbf59bc9d
|
949 |
|
ce3141a27
|
950 951 |
/* need to allocate and map pages, this chunk can't be immutable */ WARN_ON(chunk->immutable); |
fbf59bc9d
|
952 |
|
ce3141a27
|
953 954 955 |
pages = pcpu_get_pages_and_bitmap(chunk, &populated, true); if (!pages) return -ENOMEM; |
fbf59bc9d
|
956 |
|
ce3141a27
|
957 958 959 960 961 962 |
/* alloc and map */ pcpu_for_each_unpop_region(chunk, rs, re, page_start, page_end) { rc = pcpu_alloc_pages(chunk, pages, populated, rs, re); if (rc) goto err_free; free_end = re; |
fbf59bc9d
|
963 |
} |
ce3141a27
|
964 965 966 967 968 969 970 |
pcpu_for_each_unpop_region(chunk, rs, re, page_start, page_end) { rc = pcpu_map_pages(chunk, pages, populated, rs, re); if (rc) goto err_unmap; unmap_end = re; } pcpu_post_map_flush(chunk, page_start, page_end); |
fbf59bc9d
|
971 |
|
ce3141a27
|
972 973 974 |
/* commit new bitmap */ bitmap_copy(chunk->populated, populated, pcpu_unit_pages); clear: |
fbf59bc9d
|
975 |
for_each_possible_cpu(cpu) |
2f39e637e
|
976 |
memset((void *)pcpu_chunk_addr(chunk, cpu, 0) + off, 0, size); |
fbf59bc9d
|
977 |
return 0; |
ce3141a27
|
978 979 980 981 982 983 984 985 986 987 |
err_unmap: pcpu_pre_unmap_flush(chunk, page_start, unmap_end); pcpu_for_each_unpop_region(chunk, rs, re, page_start, unmap_end) pcpu_unmap_pages(chunk, pages, populated, rs, re); pcpu_post_unmap_tlb_flush(chunk, page_start, unmap_end); err_free: pcpu_for_each_unpop_region(chunk, rs, re, page_start, free_end) pcpu_free_pages(chunk, pages, populated, rs, re); return rc; |
fbf59bc9d
|
988 989 990 991 992 993 |
} static void free_pcpu_chunk(struct pcpu_chunk *chunk) { if (!chunk) return; |
6563297ce
|
994 995 |
if (chunk->vms) pcpu_free_vm_areas(chunk->vms, pcpu_nr_groups); |
1880d93b8
|
996 |
pcpu_mem_free(chunk->map, chunk->map_alloc * sizeof(chunk->map[0])); |
fbf59bc9d
|
997 998 999 1000 1001 1002 1003 1004 1005 1006 |
kfree(chunk); } static struct pcpu_chunk *alloc_pcpu_chunk(void) { struct pcpu_chunk *chunk; chunk = kzalloc(pcpu_chunk_struct_size, GFP_KERNEL); if (!chunk) return NULL; |
1880d93b8
|
1007 |
chunk->map = pcpu_mem_alloc(PCPU_DFL_MAP_ALLOC * sizeof(chunk->map[0])); |
fbf59bc9d
|
1008 1009 |
chunk->map_alloc = PCPU_DFL_MAP_ALLOC; chunk->map[chunk->map_used++] = pcpu_unit_size; |
6563297ce
|
1010 1011 1012 1013 |
chunk->vms = pcpu_get_vm_areas(pcpu_group_offsets, pcpu_group_sizes, pcpu_nr_groups, pcpu_atom_size, GFP_KERNEL); if (!chunk->vms) { |
fbf59bc9d
|
1014 1015 1016 1017 1018 1019 1020 |
free_pcpu_chunk(chunk); return NULL; } INIT_LIST_HEAD(&chunk->list); chunk->free_size = pcpu_unit_size; chunk->contig_hint = pcpu_unit_size; |
6563297ce
|
1021 |
chunk->base_addr = chunk->vms[0]->addr - pcpu_group_offsets[0]; |
fbf59bc9d
|
1022 1023 1024 1025 1026 |
return chunk; } /** |
edcb46399
|
1027 |
* pcpu_alloc - the percpu allocator |
cae3aeb83
|
1028 |
* @size: size of area to allocate in bytes |
fbf59bc9d
|
1029 |
* @align: alignment of area (max PAGE_SIZE) |
edcb46399
|
1030 |
* @reserved: allocate from the reserved chunk if available |
fbf59bc9d
|
1031 |
* |
ccea34b5d
|
1032 1033 1034 1035 |
* Allocate percpu area of @size bytes aligned at @align. * * CONTEXT: * Does GFP_KERNEL allocation. |
fbf59bc9d
|
1036 1037 1038 1039 |
* * RETURNS: * Percpu pointer to the allocated area on success, NULL on failure. */ |
43cf38eb5
|
1040 |
static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved) |
fbf59bc9d
|
1041 |
{ |
f2badb0c9
|
1042 |
static int warn_limit = 10; |
fbf59bc9d
|
1043 |
struct pcpu_chunk *chunk; |
f2badb0c9
|
1044 |
const char *err; |
833af8427
|
1045 |
int slot, off, new_alloc; |
403a91b16
|
1046 |
unsigned long flags; |
fbf59bc9d
|
1047 |
|
8d408b4be
|
1048 |
if (unlikely(!size || size > PCPU_MIN_UNIT_SIZE || align > PAGE_SIZE)) { |
fbf59bc9d
|
1049 1050 1051 1052 1053 |
WARN(true, "illegal size (%zu) or align (%zu) for " "percpu allocation ", size, align); return NULL; } |
ccea34b5d
|
1054 |
mutex_lock(&pcpu_alloc_mutex); |
403a91b16
|
1055 |
spin_lock_irqsave(&pcpu_lock, flags); |
fbf59bc9d
|
1056 |
|
edcb46399
|
1057 1058 1059 |
/* serve reserved allocations from the reserved chunk if available */ if (reserved && pcpu_reserved_chunk) { chunk = pcpu_reserved_chunk; |
833af8427
|
1060 1061 1062 |
if (size > chunk->contig_hint) { err = "alloc from reserved chunk failed"; |
ccea34b5d
|
1063 |
goto fail_unlock; |
f2badb0c9
|
1064 |
} |
833af8427
|
1065 1066 1067 1068 1069 1070 1071 1072 1073 |
while ((new_alloc = pcpu_need_to_extend(chunk))) { spin_unlock_irqrestore(&pcpu_lock, flags); if (pcpu_extend_area_map(chunk, new_alloc) < 0) { err = "failed to extend area map of reserved chunk"; goto fail_unlock_mutex; } spin_lock_irqsave(&pcpu_lock, flags); } |
edcb46399
|
1074 1075 1076 |
off = pcpu_alloc_area(chunk, size, align); if (off >= 0) goto area_found; |
833af8427
|
1077 |
|
f2badb0c9
|
1078 |
err = "alloc from reserved chunk failed"; |
ccea34b5d
|
1079 |
goto fail_unlock; |
edcb46399
|
1080 |
} |
ccea34b5d
|
1081 |
restart: |
edcb46399
|
1082 |
/* search through normal chunks */ |
fbf59bc9d
|
1083 1084 1085 1086 |
for (slot = pcpu_size_to_slot(size); slot < pcpu_nr_slots; slot++) { list_for_each_entry(chunk, &pcpu_slot[slot], list) { if (size > chunk->contig_hint) continue; |
ccea34b5d
|
1087 |
|
833af8427
|
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 |
new_alloc = pcpu_need_to_extend(chunk); if (new_alloc) { spin_unlock_irqrestore(&pcpu_lock, flags); if (pcpu_extend_area_map(chunk, new_alloc) < 0) { err = "failed to extend area map"; goto fail_unlock_mutex; } spin_lock_irqsave(&pcpu_lock, flags); /* * pcpu_lock has been dropped, need to * restart cpu_slot list walking. */ goto restart; |
ccea34b5d
|
1102 |
} |
fbf59bc9d
|
1103 1104 1105 |
off = pcpu_alloc_area(chunk, size, align); if (off >= 0) goto area_found; |
fbf59bc9d
|
1106 1107 1108 1109 |
} } /* hmmm... no space left, create a new chunk */ |
403a91b16
|
1110 |
spin_unlock_irqrestore(&pcpu_lock, flags); |
ccea34b5d
|
1111 |
|
fbf59bc9d
|
1112 |
chunk = alloc_pcpu_chunk(); |
f2badb0c9
|
1113 1114 |
if (!chunk) { err = "failed to allocate new chunk"; |
ccea34b5d
|
1115 |
goto fail_unlock_mutex; |
f2badb0c9
|
1116 |
} |
ccea34b5d
|
1117 |
|
403a91b16
|
1118 |
spin_lock_irqsave(&pcpu_lock, flags); |
fbf59bc9d
|
1119 |
pcpu_chunk_relocate(chunk, -1); |
ccea34b5d
|
1120 |
goto restart; |
fbf59bc9d
|
1121 1122 |
area_found: |
403a91b16
|
1123 |
spin_unlock_irqrestore(&pcpu_lock, flags); |
ccea34b5d
|
1124 |
|
fbf59bc9d
|
1125 1126 |
/* populate, map and clear the area */ if (pcpu_populate_chunk(chunk, off, size)) { |
403a91b16
|
1127 |
spin_lock_irqsave(&pcpu_lock, flags); |
fbf59bc9d
|
1128 |
pcpu_free_area(chunk, off); |
f2badb0c9
|
1129 |
err = "failed to populate"; |
ccea34b5d
|
1130 |
goto fail_unlock; |
fbf59bc9d
|
1131 |
} |
ccea34b5d
|
1132 |
mutex_unlock(&pcpu_alloc_mutex); |
bba174f5e
|
1133 1134 |
/* return address relative to base address */ return __addr_to_pcpu_ptr(chunk->base_addr + off); |
ccea34b5d
|
1135 1136 |
fail_unlock: |
403a91b16
|
1137 |
spin_unlock_irqrestore(&pcpu_lock, flags); |
ccea34b5d
|
1138 1139 |
fail_unlock_mutex: mutex_unlock(&pcpu_alloc_mutex); |
f2badb0c9
|
1140 1141 1142 1143 1144 1145 1146 1147 1148 |
if (warn_limit) { pr_warning("PERCPU: allocation failed, size=%zu align=%zu, " "%s ", size, align, err); dump_stack(); if (!--warn_limit) pr_info("PERCPU: limit reached, disable warning "); } |
ccea34b5d
|
1149 |
return NULL; |
fbf59bc9d
|
1150 |
} |
edcb46399
|
1151 1152 1153 1154 1155 1156 1157 1158 1159 |
/** * __alloc_percpu - allocate dynamic percpu area * @size: size of area to allocate in bytes * @align: alignment of area (max PAGE_SIZE) * * Allocate percpu area of @size bytes aligned at @align. Might * sleep. Might trigger writeouts. * |
ccea34b5d
|
1160 1161 1162 |
* CONTEXT: * Does GFP_KERNEL allocation. * |
edcb46399
|
1163 1164 1165 |
* RETURNS: * Percpu pointer to the allocated area on success, NULL on failure. */ |
43cf38eb5
|
1166 |
void __percpu *__alloc_percpu(size_t size, size_t align) |
edcb46399
|
1167 1168 1169 |
{ return pcpu_alloc(size, align, false); } |
fbf59bc9d
|
1170 |
EXPORT_SYMBOL_GPL(__alloc_percpu); |
edcb46399
|
1171 1172 1173 1174 1175 1176 1177 1178 1179 |
/** * __alloc_reserved_percpu - allocate reserved percpu area * @size: size of area to allocate in bytes * @align: alignment of area (max PAGE_SIZE) * * Allocate percpu area of @size bytes aligned at @align from reserved * percpu area if arch has set it up; otherwise, allocation is served * from the same dynamic area. Might sleep. Might trigger writeouts. * |
ccea34b5d
|
1180 1181 1182 |
* CONTEXT: * Does GFP_KERNEL allocation. * |
edcb46399
|
1183 1184 1185 |
* RETURNS: * Percpu pointer to the allocated area on success, NULL on failure. */ |
43cf38eb5
|
1186 |
void __percpu *__alloc_reserved_percpu(size_t size, size_t align) |
edcb46399
|
1187 1188 1189 |
{ return pcpu_alloc(size, align, true); } |
a56dbddf0
|
1190 1191 1192 1193 1194 |
/** * pcpu_reclaim - reclaim fully free chunks, workqueue function * @work: unused * * Reclaim all fully free chunks except for the first one. |
ccea34b5d
|
1195 1196 1197 |
* * CONTEXT: * workqueue context. |
a56dbddf0
|
1198 1199 |
*/ static void pcpu_reclaim(struct work_struct *work) |
fbf59bc9d
|
1200 |
{ |
a56dbddf0
|
1201 1202 1203 |
LIST_HEAD(todo); struct list_head *head = &pcpu_slot[pcpu_nr_slots - 1]; struct pcpu_chunk *chunk, *next; |
ccea34b5d
|
1204 1205 |
mutex_lock(&pcpu_alloc_mutex); spin_lock_irq(&pcpu_lock); |
a56dbddf0
|
1206 1207 1208 1209 1210 1211 1212 |
list_for_each_entry_safe(chunk, next, head, list) { WARN_ON(chunk->immutable); /* spare the first one */ if (chunk == list_first_entry(head, struct pcpu_chunk, list)) continue; |
a56dbddf0
|
1213 1214 |
list_move(&chunk->list, &todo); } |
ccea34b5d
|
1215 |
spin_unlock_irq(&pcpu_lock); |
a56dbddf0
|
1216 1217 |
list_for_each_entry_safe(chunk, next, &todo, list) { |
ce3141a27
|
1218 |
pcpu_depopulate_chunk(chunk, 0, pcpu_unit_size); |
a56dbddf0
|
1219 1220 |
free_pcpu_chunk(chunk); } |
971f3918a
|
1221 1222 |
mutex_unlock(&pcpu_alloc_mutex); |
fbf59bc9d
|
1223 1224 1225 1226 1227 1228 |
} /** * free_percpu - free percpu area * @ptr: pointer to area to free * |
ccea34b5d
|
1229 1230 1231 1232 |
* Free percpu area @ptr. * * CONTEXT: * Can be called from atomic context. |
fbf59bc9d
|
1233 |
*/ |
43cf38eb5
|
1234 |
void free_percpu(void __percpu *ptr) |
fbf59bc9d
|
1235 |
{ |
129182e56
|
1236 |
void *addr; |
fbf59bc9d
|
1237 |
struct pcpu_chunk *chunk; |
ccea34b5d
|
1238 |
unsigned long flags; |
fbf59bc9d
|
1239 1240 1241 1242 |
int off; if (!ptr) return; |
129182e56
|
1243 |
addr = __pcpu_ptr_to_addr(ptr); |
ccea34b5d
|
1244 |
spin_lock_irqsave(&pcpu_lock, flags); |
fbf59bc9d
|
1245 1246 |
chunk = pcpu_chunk_addr_search(addr); |
bba174f5e
|
1247 |
off = addr - chunk->base_addr; |
fbf59bc9d
|
1248 1249 |
pcpu_free_area(chunk, off); |
a56dbddf0
|
1250 |
/* if there are more than one fully free chunks, wake up grim reaper */ |
fbf59bc9d
|
1251 1252 |
if (chunk->free_size == pcpu_unit_size) { struct pcpu_chunk *pos; |
a56dbddf0
|
1253 |
list_for_each_entry(pos, &pcpu_slot[pcpu_nr_slots - 1], list) |
fbf59bc9d
|
1254 |
if (pos != chunk) { |
a56dbddf0
|
1255 |
schedule_work(&pcpu_reclaim_work); |
fbf59bc9d
|
1256 1257 1258 |
break; } } |
ccea34b5d
|
1259 |
spin_unlock_irqrestore(&pcpu_lock, flags); |
fbf59bc9d
|
1260 1261 |
} EXPORT_SYMBOL_GPL(free_percpu); |
3b034b0d0
|
1262 |
/** |
10fad5e46
|
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 |
* is_kernel_percpu_address - test whether address is from static percpu area * @addr: address to test * * Test whether @addr belongs to in-kernel static percpu area. Module * static percpu areas are not considered. For those, use * is_module_percpu_address(). * * RETURNS: * %true if @addr is from in-kernel static percpu area, %false otherwise. */ bool is_kernel_percpu_address(unsigned long addr) { const size_t static_size = __per_cpu_end - __per_cpu_start; void __percpu *base = __addr_to_pcpu_ptr(pcpu_base_addr); unsigned int cpu; for_each_possible_cpu(cpu) { void *start = per_cpu_ptr(base, cpu); if ((void *)addr >= start && (void *)addr < start + static_size) return true; } return false; } /** |
3b034b0d0
|
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 |
* per_cpu_ptr_to_phys - convert translated percpu address to physical address * @addr: the address to be converted to physical address * * Given @addr which is dereferenceable address obtained via one of * percpu access macros, this function translates it into its physical * address. The caller is responsible for ensuring @addr stays valid * until this function finishes. * * RETURNS: * The physical address for @addr. */ phys_addr_t per_cpu_ptr_to_phys(void *addr) { if ((unsigned long)addr < VMALLOC_START || (unsigned long)addr >= VMALLOC_END) return __pa(addr); else return page_to_phys(vmalloc_to_page(addr)); } |
033e48fb8
|
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 |
static inline size_t pcpu_calc_fc_sizes(size_t static_size, size_t reserved_size, ssize_t *dyn_sizep) { size_t size_sum; size_sum = PFN_ALIGN(static_size + reserved_size + (*dyn_sizep >= 0 ? *dyn_sizep : 0)); if (*dyn_sizep != 0) *dyn_sizep = size_sum - static_size - reserved_size; return size_sum; } |
fbf59bc9d
|
1321 |
/** |
fd1e8a1fe
|
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 |
* pcpu_alloc_alloc_info - allocate percpu allocation info * @nr_groups: the number of groups * @nr_units: the number of units * * Allocate ai which is large enough for @nr_groups groups containing * @nr_units units. The returned ai's groups[0].cpu_map points to the * cpu_map array which is long enough for @nr_units and filled with * NR_CPUS. It's the caller's responsibility to initialize cpu_map * pointer of other groups. * * RETURNS: * Pointer to the allocated pcpu_alloc_info on success, NULL on * failure. */ struct pcpu_alloc_info * __init pcpu_alloc_alloc_info(int nr_groups, int nr_units) { struct pcpu_alloc_info *ai; size_t base_size, ai_size; void *ptr; int unit; base_size = ALIGN(sizeof(*ai) + nr_groups * sizeof(ai->groups[0]), __alignof__(ai->groups[0].cpu_map[0])); ai_size = base_size + nr_units * sizeof(ai->groups[0].cpu_map[0]); ptr = alloc_bootmem_nopanic(PFN_ALIGN(ai_size)); if (!ptr) return NULL; ai = ptr; ptr += base_size; ai->groups[0].cpu_map = ptr; for (unit = 0; unit < nr_units; unit++) ai->groups[0].cpu_map[unit] = NR_CPUS; ai->nr_groups = nr_groups; ai->__ai_size = PFN_ALIGN(ai_size); return ai; } /** * pcpu_free_alloc_info - free percpu allocation info * @ai: pcpu_alloc_info to free * * Free @ai which was allocated by pcpu_alloc_alloc_info(). */ void __init pcpu_free_alloc_info(struct pcpu_alloc_info *ai) { free_bootmem(__pa(ai), ai->__ai_size); } /** * pcpu_build_alloc_info - build alloc_info considering distances between CPUs |
edcb46399
|
1378 |
* @reserved_size: the size of reserved percpu area in bytes |
cafe8816b
|
1379 |
* @dyn_size: free size for dynamic allocation in bytes, -1 for auto |
fd1e8a1fe
|
1380 1381 |
* @atom_size: allocation atom size * @cpu_distance_fn: callback to determine distance between cpus, optional |
033e48fb8
|
1382 |
* |
fd1e8a1fe
|
1383 1384 1385 |
* This function determines grouping of units, their mappings to cpus * and other parameters considering needed percpu size, allocation * atom size and distances between CPUs. |
033e48fb8
|
1386 |
* |
fd1e8a1fe
|
1387 1388 1389 1390 1391 |
* Groups are always mutliples of atom size and CPUs which are of * LOCAL_DISTANCE both ways are grouped together and share space for * units in the same group. The returned configuration is guaranteed * to have CPUs on different nodes on different groups and >=75% usage * of allocated virtual address space. |
033e48fb8
|
1392 1393 |
* * RETURNS: |
fd1e8a1fe
|
1394 1395 |
* On success, pointer to the new allocation_info is returned. On * failure, ERR_PTR value is returned. |
033e48fb8
|
1396 |
*/ |
fd1e8a1fe
|
1397 1398 1399 1400 |
struct pcpu_alloc_info * __init pcpu_build_alloc_info( size_t reserved_size, ssize_t dyn_size, size_t atom_size, pcpu_fc_cpu_distance_fn_t cpu_distance_fn) |
033e48fb8
|
1401 1402 1403 1404 |
{ static int group_map[NR_CPUS] __initdata; static int group_cnt[NR_CPUS] __initdata; const size_t static_size = __per_cpu_end - __per_cpu_start; |
fd1e8a1fe
|
1405 |
int group_cnt_max = 0, nr_groups = 1, nr_units = 0; |
033e48fb8
|
1406 1407 |
size_t size_sum, min_unit_size, alloc_size; int upa, max_upa, uninitialized_var(best_upa); /* units_per_alloc */ |
fd1e8a1fe
|
1408 |
int last_allocs, group, unit; |
033e48fb8
|
1409 |
unsigned int cpu, tcpu; |
fd1e8a1fe
|
1410 1411 |
struct pcpu_alloc_info *ai; unsigned int *cpu_map; |
033e48fb8
|
1412 |
|
fb59e72e7
|
1413 1414 1415 |
/* this function may be called multiple times */ memset(group_map, 0, sizeof(group_map)); memset(group_cnt, 0, sizeof(group_map)); |
033e48fb8
|
1416 1417 |
/* * Determine min_unit_size, alloc_size and max_upa such that |
fd1e8a1fe
|
1418 |
* alloc_size is multiple of atom_size and is the smallest |
033e48fb8
|
1419 1420 1421 |
* which can accomodate 4k aligned segments which are equal to * or larger than min_unit_size. */ |
fd1e8a1fe
|
1422 |
size_sum = pcpu_calc_fc_sizes(static_size, reserved_size, &dyn_size); |
033e48fb8
|
1423 |
min_unit_size = max_t(size_t, size_sum, PCPU_MIN_UNIT_SIZE); |
fd1e8a1fe
|
1424 |
alloc_size = roundup(min_unit_size, atom_size); |
033e48fb8
|
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 |
upa = alloc_size / min_unit_size; while (alloc_size % upa || ((alloc_size / upa) & ~PAGE_MASK)) upa--; max_upa = upa; /* group cpus according to their proximity */ for_each_possible_cpu(cpu) { group = 0; next_group: for_each_possible_cpu(tcpu) { if (cpu == tcpu) break; |
fd1e8a1fe
|
1437 |
if (group_map[tcpu] == group && cpu_distance_fn && |
033e48fb8
|
1438 1439 1440 |
(cpu_distance_fn(cpu, tcpu) > LOCAL_DISTANCE || cpu_distance_fn(tcpu, cpu) > LOCAL_DISTANCE)) { group++; |
fd1e8a1fe
|
1441 |
nr_groups = max(nr_groups, group + 1); |
033e48fb8
|
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 |
goto next_group; } } group_map[cpu] = group; group_cnt[group]++; group_cnt_max = max(group_cnt_max, group_cnt[group]); } /* * Expand unit size until address space usage goes over 75% * and then as much as possible without using more address * space. */ last_allocs = INT_MAX; for (upa = max_upa; upa; upa--) { int allocs = 0, wasted = 0; if (alloc_size % upa || ((alloc_size / upa) & ~PAGE_MASK)) continue; |
fd1e8a1fe
|
1461 |
for (group = 0; group < nr_groups; group++) { |
033e48fb8
|
1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 |
int this_allocs = DIV_ROUND_UP(group_cnt[group], upa); allocs += this_allocs; wasted += this_allocs * upa - group_cnt[group]; } /* * Don't accept if wastage is over 25%. The * greater-than comparison ensures upa==1 always * passes the following check. */ if (wasted > num_possible_cpus() / 3) continue; /* and then don't consume more memory */ if (allocs > last_allocs) break; last_allocs = allocs; best_upa = upa; } |
fd1e8a1fe
|
1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 |
upa = best_upa; /* allocate and fill alloc_info */ for (group = 0; group < nr_groups; group++) nr_units += roundup(group_cnt[group], upa); ai = pcpu_alloc_alloc_info(nr_groups, nr_units); if (!ai) return ERR_PTR(-ENOMEM); cpu_map = ai->groups[0].cpu_map; for (group = 0; group < nr_groups; group++) { ai->groups[group].cpu_map = cpu_map; cpu_map += roundup(group_cnt[group], upa); } ai->static_size = static_size; ai->reserved_size = reserved_size; ai->dyn_size = dyn_size; ai->unit_size = alloc_size / upa; ai->atom_size = atom_size; ai->alloc_size = alloc_size; for (group = 0, unit = 0; group_cnt[group]; group++) { struct pcpu_group_info *gi = &ai->groups[group]; /* * Initialize base_offset as if all groups are located * back-to-back. The caller should update this to * reflect actual allocation. */ gi->base_offset = unit * ai->unit_size; |
033e48fb8
|
1513 |
|
033e48fb8
|
1514 1515 |
for_each_possible_cpu(cpu) if (group_map[cpu] == group) |
fd1e8a1fe
|
1516 1517 1518 |
gi->cpu_map[gi->nr_units++] = cpu; gi->nr_units = roundup(gi->nr_units, upa); unit += gi->nr_units; |
033e48fb8
|
1519 |
} |
fd1e8a1fe
|
1520 |
BUG_ON(unit != nr_units); |
033e48fb8
|
1521 |
|
fd1e8a1fe
|
1522 |
return ai; |
033e48fb8
|
1523 |
} |
fd1e8a1fe
|
1524 1525 1526 1527 1528 1529 1530 1531 1532 |
/** * pcpu_dump_alloc_info - print out information about pcpu_alloc_info * @lvl: loglevel * @ai: allocation info to dump * * Print out information about @ai using loglevel @lvl. */ static void pcpu_dump_alloc_info(const char *lvl, const struct pcpu_alloc_info *ai) |
033e48fb8
|
1533 |
{ |
fd1e8a1fe
|
1534 |
int group_width = 1, cpu_width = 1, width; |
033e48fb8
|
1535 |
char empty_str[] = "--------"; |
fd1e8a1fe
|
1536 1537 1538 1539 1540 1541 1542 |
int alloc = 0, alloc_end = 0; int group, v; int upa, apl; /* units per alloc, allocs per line */ v = ai->nr_groups; while (v /= 10) group_width++; |
033e48fb8
|
1543 |
|
fd1e8a1fe
|
1544 |
v = num_possible_cpus(); |
033e48fb8
|
1545 |
while (v /= 10) |
fd1e8a1fe
|
1546 1547 |
cpu_width++; empty_str[min_t(int, cpu_width, sizeof(empty_str) - 1)] = '\0'; |
033e48fb8
|
1548 |
|
fd1e8a1fe
|
1549 1550 1551 |
upa = ai->alloc_size / ai->unit_size; width = upa * (cpu_width + 1) + group_width + 3; apl = rounddown_pow_of_two(max(60 / width, 1)); |
033e48fb8
|
1552 |
|
fd1e8a1fe
|
1553 1554 1555 |
printk("%spcpu-alloc: s%zu r%zu d%zu u%zu alloc=%zu*%zu", lvl, ai->static_size, ai->reserved_size, ai->dyn_size, ai->unit_size, ai->alloc_size / ai->atom_size, ai->atom_size); |
033e48fb8
|
1556 |
|
fd1e8a1fe
|
1557 1558 1559 1560 1561 1562 1563 1564 |
for (group = 0; group < ai->nr_groups; group++) { const struct pcpu_group_info *gi = &ai->groups[group]; int unit = 0, unit_end = 0; BUG_ON(gi->nr_units % upa); for (alloc_end += gi->nr_units / upa; alloc < alloc_end; alloc++) { if (!(alloc % apl)) { |
033e48fb8
|
1565 1566 |
printk(" "); |
fd1e8a1fe
|
1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 |
printk("%spcpu-alloc: ", lvl); } printk("[%0*d] ", group_width, group); for (unit_end += upa; unit < unit_end; unit++) if (gi->cpu_map[unit] != NR_CPUS) printk("%0*d ", cpu_width, gi->cpu_map[unit]); else printk("%s ", empty_str); |
033e48fb8
|
1577 |
} |
033e48fb8
|
1578 1579 1580 1581 |
} printk(" "); } |
033e48fb8
|
1582 |
|
fbf59bc9d
|
1583 |
/** |
8d408b4be
|
1584 |
* pcpu_setup_first_chunk - initialize the first percpu chunk |
fd1e8a1fe
|
1585 |
* @ai: pcpu_alloc_info describing how to percpu area is shaped |
38a6be525
|
1586 |
* @base_addr: mapped address |
8d408b4be
|
1587 1588 1589 |
* * Initialize the first percpu chunk which contains the kernel static * perpcu area. This function is to be called from arch percpu area |
38a6be525
|
1590 |
* setup path. |
8d408b4be
|
1591 |
* |
fd1e8a1fe
|
1592 1593 1594 1595 1596 1597 |
* @ai contains all information necessary to initialize the first * chunk and prime the dynamic percpu allocator. * * @ai->static_size is the size of static percpu area. * * @ai->reserved_size, if non-zero, specifies the amount of bytes to |
edcb46399
|
1598 1599 1600 1601 1602 1603 1604 |
* reserve after the static area in the first chunk. This reserves * the first chunk such that it's available only through reserved * percpu allocation. This is primarily used to serve module percpu * static areas on architectures where the addressing model has * limited offset range for symbol relocations to guarantee module * percpu symbols fall inside the relocatable range. * |
fd1e8a1fe
|
1605 1606 1607 |
* @ai->dyn_size determines the number of bytes available for dynamic * allocation in the first chunk. The area between @ai->static_size + * @ai->reserved_size + @ai->dyn_size and @ai->unit_size is unused. |
6074d5b0a
|
1608 |
* |
fd1e8a1fe
|
1609 1610 1611 |
* @ai->unit_size specifies unit size and must be aligned to PAGE_SIZE * and equal to or larger than @ai->static_size + @ai->reserved_size + * @ai->dyn_size. |
8d408b4be
|
1612 |
* |
fd1e8a1fe
|
1613 1614 |
* @ai->atom_size is the allocation atom size and used as alignment * for vm areas. |
8d408b4be
|
1615 |
* |
fd1e8a1fe
|
1616 1617 1618 1619 1620 1621 1622 1623 1624 |
* @ai->alloc_size is the allocation size and always multiple of * @ai->atom_size. This is larger than @ai->atom_size if * @ai->unit_size is larger than @ai->atom_size. * * @ai->nr_groups and @ai->groups describe virtual memory layout of * percpu areas. Units which should be colocated are put into the * same group. Dynamic VM areas will be allocated according to these * groupings. If @ai->nr_groups is zero, a single group containing * all units is assumed. |
8d408b4be
|
1625 |
* |
38a6be525
|
1626 1627 |
* The caller should have mapped the first chunk at @base_addr and * copied static data to each unit. |
fbf59bc9d
|
1628 |
* |
edcb46399
|
1629 1630 1631 1632 1633 1634 1635 |
* If the first chunk ends up with both reserved and dynamic areas, it * is served by two chunks - one to serve the core static and reserved * areas and the other for the dynamic area. They share the same vm * and page map but uses different area allocation map to stay away * from each other. The latter chunk is circulated in the chunk slots * and available for dynamic allocation like any other chunks. * |
fbf59bc9d
|
1636 |
* RETURNS: |
fb435d523
|
1637 |
* 0 on success, -errno on failure. |
fbf59bc9d
|
1638 |
*/ |
fb435d523
|
1639 1640 |
int __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai, void *base_addr) |
fbf59bc9d
|
1641 |
{ |
635b75fc1
|
1642 |
static char cpus_buf[4096] __initdata; |
edcb46399
|
1643 |
static int smap[2], dmap[2]; |
fd1e8a1fe
|
1644 1645 |
size_t dyn_size = ai->dyn_size; size_t size_sum = ai->static_size + ai->reserved_size + dyn_size; |
edcb46399
|
1646 |
struct pcpu_chunk *schunk, *dchunk = NULL; |
6563297ce
|
1647 1648 |
unsigned long *group_offsets; size_t *group_sizes; |
fb435d523
|
1649 |
unsigned long *unit_off; |
fbf59bc9d
|
1650 |
unsigned int cpu; |
fd1e8a1fe
|
1651 1652 |
int *unit_map; int group, unit, i; |
fbf59bc9d
|
1653 |
|
635b75fc1
|
1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 |
cpumask_scnprintf(cpus_buf, sizeof(cpus_buf), cpu_possible_mask); #define PCPU_SETUP_BUG_ON(cond) do { \ if (unlikely(cond)) { \ pr_emerg("PERCPU: failed to initialize, %s", #cond); \ pr_emerg("PERCPU: cpu_possible_mask=%s ", cpus_buf); \ pcpu_dump_alloc_info(KERN_EMERG, ai); \ BUG(); \ } \ } while (0) |
2f39e637e
|
1665 |
/* sanity checks */ |
edcb46399
|
1666 1667 |
BUILD_BUG_ON(ARRAY_SIZE(smap) >= PCPU_DFL_MAP_ALLOC || ARRAY_SIZE(dmap) >= PCPU_DFL_MAP_ALLOC); |
635b75fc1
|
1668 1669 1670 1671 1672 1673 |
PCPU_SETUP_BUG_ON(ai->nr_groups <= 0); PCPU_SETUP_BUG_ON(!ai->static_size); PCPU_SETUP_BUG_ON(!base_addr); PCPU_SETUP_BUG_ON(ai->unit_size < size_sum); PCPU_SETUP_BUG_ON(ai->unit_size & ~PAGE_MASK); PCPU_SETUP_BUG_ON(ai->unit_size < PCPU_MIN_UNIT_SIZE); |
8d408b4be
|
1674 |
|
6563297ce
|
1675 1676 1677 |
/* process group information and build config tables accordingly */ group_offsets = alloc_bootmem(ai->nr_groups * sizeof(group_offsets[0])); group_sizes = alloc_bootmem(ai->nr_groups * sizeof(group_sizes[0])); |
fd1e8a1fe
|
1678 |
unit_map = alloc_bootmem(nr_cpu_ids * sizeof(unit_map[0])); |
fb435d523
|
1679 |
unit_off = alloc_bootmem(nr_cpu_ids * sizeof(unit_off[0])); |
2f39e637e
|
1680 |
|
fd1e8a1fe
|
1681 |
for (cpu = 0; cpu < nr_cpu_ids; cpu++) |
ffe0d5a57
|
1682 |
unit_map[cpu] = UINT_MAX; |
fd1e8a1fe
|
1683 |
pcpu_first_unit_cpu = NR_CPUS; |
2f39e637e
|
1684 |
|
fd1e8a1fe
|
1685 1686 |
for (group = 0, unit = 0; group < ai->nr_groups; group++, unit += i) { const struct pcpu_group_info *gi = &ai->groups[group]; |
2f39e637e
|
1687 |
|
6563297ce
|
1688 1689 |
group_offsets[group] = gi->base_offset; group_sizes[group] = gi->nr_units * ai->unit_size; |
fd1e8a1fe
|
1690 1691 1692 1693 |
for (i = 0; i < gi->nr_units; i++) { cpu = gi->cpu_map[i]; if (cpu == NR_CPUS) continue; |
8d408b4be
|
1694 |
|
635b75fc1
|
1695 1696 1697 |
PCPU_SETUP_BUG_ON(cpu > nr_cpu_ids); PCPU_SETUP_BUG_ON(!cpu_possible(cpu)); PCPU_SETUP_BUG_ON(unit_map[cpu] != UINT_MAX); |
fbf59bc9d
|
1698 |
|
fd1e8a1fe
|
1699 |
unit_map[cpu] = unit + i; |
fb435d523
|
1700 |
unit_off[cpu] = gi->base_offset + i * ai->unit_size; |
fd1e8a1fe
|
1701 1702 1703 |
if (pcpu_first_unit_cpu == NR_CPUS) pcpu_first_unit_cpu = cpu; } |
2f39e637e
|
1704 |
} |
fd1e8a1fe
|
1705 1706 1707 1708 |
pcpu_last_unit_cpu = cpu; pcpu_nr_units = unit; for_each_possible_cpu(cpu) |
635b75fc1
|
1709 1710 1711 1712 1713 |
PCPU_SETUP_BUG_ON(unit_map[cpu] == UINT_MAX); /* we're done parsing the input, undefine BUG macro and dump config */ #undef PCPU_SETUP_BUG_ON pcpu_dump_alloc_info(KERN_INFO, ai); |
fd1e8a1fe
|
1714 |
|
6563297ce
|
1715 1716 1717 |
pcpu_nr_groups = ai->nr_groups; pcpu_group_offsets = group_offsets; pcpu_group_sizes = group_sizes; |
fd1e8a1fe
|
1718 |
pcpu_unit_map = unit_map; |
fb435d523
|
1719 |
pcpu_unit_offsets = unit_off; |
2f39e637e
|
1720 1721 |
/* determine basic parameters */ |
fd1e8a1fe
|
1722 |
pcpu_unit_pages = ai->unit_size >> PAGE_SHIFT; |
d9b55eeb1
|
1723 |
pcpu_unit_size = pcpu_unit_pages << PAGE_SHIFT; |
6563297ce
|
1724 |
pcpu_atom_size = ai->atom_size; |
ce3141a27
|
1725 1726 |
pcpu_chunk_struct_size = sizeof(struct pcpu_chunk) + BITS_TO_LONGS(pcpu_unit_pages) * sizeof(unsigned long); |
cafe8816b
|
1727 |
|
d9b55eeb1
|
1728 1729 1730 1731 1732 |
/* * Allocate chunk slots. The additional last slot is for * empty chunks. */ pcpu_nr_slots = __pcpu_size_to_slot(pcpu_unit_size) + 2; |
fbf59bc9d
|
1733 1734 1735 |
pcpu_slot = alloc_bootmem(pcpu_nr_slots * sizeof(pcpu_slot[0])); for (i = 0; i < pcpu_nr_slots; i++) INIT_LIST_HEAD(&pcpu_slot[i]); |
edcb46399
|
1736 1737 1738 1739 1740 1741 1742 |
/* * Initialize static chunk. If reserved_size is zero, the * static chunk covers static area + dynamic allocation area * in the first chunk. If reserved_size is not zero, it * covers static area + reserved area (mostly used for module * static percpu allocation). */ |
2441d15c9
|
1743 1744 |
schunk = alloc_bootmem(pcpu_chunk_struct_size); INIT_LIST_HEAD(&schunk->list); |
bba174f5e
|
1745 |
schunk->base_addr = base_addr; |
61ace7fa2
|
1746 1747 |
schunk->map = smap; schunk->map_alloc = ARRAY_SIZE(smap); |
38a6be525
|
1748 |
schunk->immutable = true; |
ce3141a27
|
1749 |
bitmap_fill(schunk->populated, pcpu_unit_pages); |
edcb46399
|
1750 |
|
fd1e8a1fe
|
1751 1752 |
if (ai->reserved_size) { schunk->free_size = ai->reserved_size; |
ae9e6bc9f
|
1753 |
pcpu_reserved_chunk = schunk; |
fd1e8a1fe
|
1754 |
pcpu_reserved_chunk_limit = ai->static_size + ai->reserved_size; |
edcb46399
|
1755 1756 1757 1758 |
} else { schunk->free_size = dyn_size; dyn_size = 0; /* dynamic area covered */ } |
2441d15c9
|
1759 |
schunk->contig_hint = schunk->free_size; |
fbf59bc9d
|
1760 |
|
fd1e8a1fe
|
1761 |
schunk->map[schunk->map_used++] = -ai->static_size; |
61ace7fa2
|
1762 1763 |
if (schunk->free_size) schunk->map[schunk->map_used++] = schunk->free_size; |
edcb46399
|
1764 1765 |
/* init dynamic chunk if necessary */ if (dyn_size) { |
ce3141a27
|
1766 |
dchunk = alloc_bootmem(pcpu_chunk_struct_size); |
edcb46399
|
1767 |
INIT_LIST_HEAD(&dchunk->list); |
bba174f5e
|
1768 |
dchunk->base_addr = base_addr; |
edcb46399
|
1769 1770 |
dchunk->map = dmap; dchunk->map_alloc = ARRAY_SIZE(dmap); |
38a6be525
|
1771 |
dchunk->immutable = true; |
ce3141a27
|
1772 |
bitmap_fill(dchunk->populated, pcpu_unit_pages); |
edcb46399
|
1773 1774 1775 1776 1777 |
dchunk->contig_hint = dchunk->free_size = dyn_size; dchunk->map[dchunk->map_used++] = -pcpu_reserved_chunk_limit; dchunk->map[dchunk->map_used++] = dchunk->free_size; } |
2441d15c9
|
1778 |
/* link the first chunk in */ |
ae9e6bc9f
|
1779 1780 |
pcpu_first_chunk = dchunk ?: schunk; pcpu_chunk_relocate(pcpu_first_chunk, -1); |
fbf59bc9d
|
1781 1782 |
/* we're done */ |
bba174f5e
|
1783 |
pcpu_base_addr = base_addr; |
fb435d523
|
1784 |
return 0; |
fbf59bc9d
|
1785 |
} |
66c3a7577
|
1786 |
|
f58dc01ba
|
1787 1788 1789 1790 |
const char *pcpu_fc_names[PCPU_FC_NR] __initdata = { [PCPU_FC_AUTO] = "auto", [PCPU_FC_EMBED] = "embed", [PCPU_FC_PAGE] = "page", |
f58dc01ba
|
1791 |
}; |
66c3a7577
|
1792 |
|
f58dc01ba
|
1793 |
enum pcpu_fc pcpu_chosen_fc __initdata = PCPU_FC_AUTO; |
66c3a7577
|
1794 |
|
f58dc01ba
|
1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 |
static int __init percpu_alloc_setup(char *str) { if (0) /* nada */; #ifdef CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK else if (!strcmp(str, "embed")) pcpu_chosen_fc = PCPU_FC_EMBED; #endif #ifdef CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK else if (!strcmp(str, "page")) pcpu_chosen_fc = PCPU_FC_PAGE; #endif |
f58dc01ba
|
1807 1808 1809 |
else pr_warning("PERCPU: unknown allocator %s specified ", str); |
66c3a7577
|
1810 |
|
f58dc01ba
|
1811 |
return 0; |
66c3a7577
|
1812 |
} |
f58dc01ba
|
1813 |
early_param("percpu_alloc", percpu_alloc_setup); |
66c3a7577
|
1814 |
|
08fc45806
|
1815 1816 |
#if defined(CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK) || \ !defined(CONFIG_HAVE_SETUP_PER_CPU_AREA) |
66c3a7577
|
1817 1818 |
/** * pcpu_embed_first_chunk - embed the first percpu chunk into bootmem |
66c3a7577
|
1819 1820 |
* @reserved_size: the size of reserved percpu area in bytes * @dyn_size: free size for dynamic allocation in bytes, -1 for auto |
c8826dd53
|
1821 1822 1823 1824 |
* @atom_size: allocation atom size * @cpu_distance_fn: callback to determine distance between cpus, optional * @alloc_fn: function to allocate percpu page * @free_fn: funtion to free percpu page |
66c3a7577
|
1825 1826 1827 1828 1829 |
* * This is a helper to ease setting up embedded first percpu chunk and * can be called where pcpu_setup_first_chunk() is expected. * * If this function is used to setup the first chunk, it is allocated |
c8826dd53
|
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 |
* by calling @alloc_fn and used as-is without being mapped into * vmalloc area. Allocations are always whole multiples of @atom_size * aligned to @atom_size. * * This enables the first chunk to piggy back on the linear physical * mapping which often uses larger page size. Please note that this * can result in very sparse cpu->unit mapping on NUMA machines thus * requiring large vmalloc address space. Don't use this allocator if * vmalloc space is not orders of magnitude larger than distances * between node memory addresses (ie. 32bit NUMA machines). |
66c3a7577
|
1840 1841 |
* * When @dyn_size is positive, dynamic area might be larger than |
788e5abc5
|
1842 1843 1844 |
* specified to fill page alignment. When @dyn_size is auto, * @dyn_size is just big enough to fill page alignment after static * and reserved areas. |
66c3a7577
|
1845 1846 |
* * If the needed size is smaller than the minimum or specified unit |
c8826dd53
|
1847 |
* size, the leftover is returned using @free_fn. |
66c3a7577
|
1848 1849 |
* * RETURNS: |
fb435d523
|
1850 |
* 0 on success, -errno on failure. |
66c3a7577
|
1851 |
*/ |
c8826dd53
|
1852 1853 1854 1855 1856 |
int __init pcpu_embed_first_chunk(size_t reserved_size, ssize_t dyn_size, size_t atom_size, pcpu_fc_cpu_distance_fn_t cpu_distance_fn, pcpu_fc_alloc_fn_t alloc_fn, pcpu_fc_free_fn_t free_fn) |
66c3a7577
|
1857 |
{ |
c8826dd53
|
1858 1859 |
void *base = (void *)ULONG_MAX; void **areas = NULL; |
fd1e8a1fe
|
1860 |
struct pcpu_alloc_info *ai; |
6ea529a20
|
1861 |
size_t size_sum, areas_size, max_distance; |
c8826dd53
|
1862 |
int group, i, rc; |
66c3a7577
|
1863 |
|
c8826dd53
|
1864 1865 |
ai = pcpu_build_alloc_info(reserved_size, dyn_size, atom_size, cpu_distance_fn); |
fd1e8a1fe
|
1866 1867 |
if (IS_ERR(ai)) return PTR_ERR(ai); |
66c3a7577
|
1868 |
|
fd1e8a1fe
|
1869 |
size_sum = ai->static_size + ai->reserved_size + ai->dyn_size; |
c8826dd53
|
1870 |
areas_size = PFN_ALIGN(ai->nr_groups * sizeof(void *)); |
fa8a7094b
|
1871 |
|
c8826dd53
|
1872 1873 |
areas = alloc_bootmem_nopanic(areas_size); if (!areas) { |
fb435d523
|
1874 |
rc = -ENOMEM; |
c8826dd53
|
1875 |
goto out_free; |
fa8a7094b
|
1876 |
} |
66c3a7577
|
1877 |
|
c8826dd53
|
1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 |
/* allocate, copy and determine base address */ for (group = 0; group < ai->nr_groups; group++) { struct pcpu_group_info *gi = &ai->groups[group]; unsigned int cpu = NR_CPUS; void *ptr; for (i = 0; i < gi->nr_units && cpu == NR_CPUS; i++) cpu = gi->cpu_map[i]; BUG_ON(cpu == NR_CPUS); /* allocate space for the whole group */ ptr = alloc_fn(cpu, gi->nr_units * ai->unit_size, atom_size); if (!ptr) { rc = -ENOMEM; goto out_free_areas; } areas[group] = ptr; |
fd1e8a1fe
|
1895 |
|
c8826dd53
|
1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 |
base = min(ptr, base); for (i = 0; i < gi->nr_units; i++, ptr += ai->unit_size) { if (gi->cpu_map[i] == NR_CPUS) { /* unused unit, free whole */ free_fn(ptr, ai->unit_size); continue; } /* copy and return the unused part */ memcpy(ptr, __per_cpu_load, ai->static_size); free_fn(ptr + size_sum, ai->unit_size - size_sum); } |
fa8a7094b
|
1908 |
} |
66c3a7577
|
1909 |
|
c8826dd53
|
1910 |
/* base address is now known, determine group base offsets */ |
6ea529a20
|
1911 1912 |
max_distance = 0; for (group = 0; group < ai->nr_groups; group++) { |
c8826dd53
|
1913 |
ai->groups[group].base_offset = areas[group] - base; |
1a0c3298d
|
1914 1915 |
max_distance = max_t(size_t, max_distance, ai->groups[group].base_offset); |
6ea529a20
|
1916 1917 1918 1919 1920 |
} max_distance += ai->unit_size; /* warn if maximum distance is further than 75% of vmalloc space */ if (max_distance > (VMALLOC_END - VMALLOC_START) * 3 / 4) { |
1a0c3298d
|
1921 |
pr_warning("PERCPU: max_distance=0x%zx too large for vmalloc " |
6ea529a20
|
1922 1923 1924 1925 1926 1927 1928 1929 1930 |
"space 0x%lx ", max_distance, VMALLOC_END - VMALLOC_START); #ifdef CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK /* and fail if we have fallback */ rc = -EINVAL; goto out_free; #endif } |
c8826dd53
|
1931 |
|
004018e2c
|
1932 1933 |
pr_info("PERCPU: Embedded %zu pages/cpu @%p s%zu r%zu d%zu u%zu ", |
fd1e8a1fe
|
1934 1935 |
PFN_DOWN(size_sum), base, ai->static_size, ai->reserved_size, ai->dyn_size, ai->unit_size); |
d4b95f803
|
1936 |
|
fb435d523
|
1937 |
rc = pcpu_setup_first_chunk(ai, base); |
c8826dd53
|
1938 1939 1940 1941 1942 1943 1944 |
goto out_free; out_free_areas: for (group = 0; group < ai->nr_groups; group++) free_fn(areas[group], ai->groups[group].nr_units * ai->unit_size); out_free: |
fd1e8a1fe
|
1945 |
pcpu_free_alloc_info(ai); |
c8826dd53
|
1946 1947 |
if (areas) free_bootmem(__pa(areas), areas_size); |
fb435d523
|
1948 |
return rc; |
d4b95f803
|
1949 |
} |
08fc45806
|
1950 1951 |
#endif /* CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK || !CONFIG_HAVE_SETUP_PER_CPU_AREA */ |
d4b95f803
|
1952 |
|
08fc45806
|
1953 |
#ifdef CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK |
d4b95f803
|
1954 |
/** |
00ae4064b
|
1955 |
* pcpu_page_first_chunk - map the first chunk using PAGE_SIZE pages |
d4b95f803
|
1956 1957 1958 1959 1960 |
* @reserved_size: the size of reserved percpu area in bytes * @alloc_fn: function to allocate percpu page, always called with PAGE_SIZE * @free_fn: funtion to free percpu page, always called with PAGE_SIZE * @populate_pte_fn: function to populate pte * |
00ae4064b
|
1961 1962 |
* This is a helper to ease setting up page-remapped first percpu * chunk and can be called where pcpu_setup_first_chunk() is expected. |
d4b95f803
|
1963 1964 1965 1966 1967 |
* * This is the basic allocator. Static percpu area is allocated * page-by-page into vmalloc area. * * RETURNS: |
fb435d523
|
1968 |
* 0 on success, -errno on failure. |
d4b95f803
|
1969 |
*/ |
fb435d523
|
1970 1971 1972 1973 |
int __init pcpu_page_first_chunk(size_t reserved_size, pcpu_fc_alloc_fn_t alloc_fn, pcpu_fc_free_fn_t free_fn, pcpu_fc_populate_pte_fn_t populate_pte_fn) |
d4b95f803
|
1974 |
{ |
8f05a6a65
|
1975 |
static struct vm_struct vm; |
fd1e8a1fe
|
1976 |
struct pcpu_alloc_info *ai; |
00ae4064b
|
1977 |
char psize_str[16]; |
ce3141a27
|
1978 |
int unit_pages; |
d4b95f803
|
1979 |
size_t pages_size; |
ce3141a27
|
1980 |
struct page **pages; |
fb435d523
|
1981 |
int unit, i, j, rc; |
d4b95f803
|
1982 |
|
00ae4064b
|
1983 |
snprintf(psize_str, sizeof(psize_str), "%luK", PAGE_SIZE >> 10); |
fd1e8a1fe
|
1984 1985 1986 1987 1988 1989 1990 |
ai = pcpu_build_alloc_info(reserved_size, -1, PAGE_SIZE, NULL); if (IS_ERR(ai)) return PTR_ERR(ai); BUG_ON(ai->nr_groups != 1); BUG_ON(ai->groups[0].nr_units != num_possible_cpus()); unit_pages = ai->unit_size >> PAGE_SHIFT; |
d4b95f803
|
1991 1992 |
/* unaligned allocations can't be freed, round up to page size */ |
fd1e8a1fe
|
1993 1994 |
pages_size = PFN_ALIGN(unit_pages * num_possible_cpus() * sizeof(pages[0])); |
ce3141a27
|
1995 |
pages = alloc_bootmem(pages_size); |
d4b95f803
|
1996 |
|
8f05a6a65
|
1997 |
/* allocate pages */ |
d4b95f803
|
1998 |
j = 0; |
fd1e8a1fe
|
1999 |
for (unit = 0; unit < num_possible_cpus(); unit++) |
ce3141a27
|
2000 |
for (i = 0; i < unit_pages; i++) { |
fd1e8a1fe
|
2001 |
unsigned int cpu = ai->groups[0].cpu_map[unit]; |
d4b95f803
|
2002 |
void *ptr; |
3cbc85652
|
2003 |
ptr = alloc_fn(cpu, PAGE_SIZE, PAGE_SIZE); |
d4b95f803
|
2004 |
if (!ptr) { |
00ae4064b
|
2005 2006 2007 |
pr_warning("PERCPU: failed to allocate %s page " "for cpu%u ", psize_str, cpu); |
d4b95f803
|
2008 2009 |
goto enomem; } |
ce3141a27
|
2010 |
pages[j++] = virt_to_page(ptr); |
d4b95f803
|
2011 |
} |
8f05a6a65
|
2012 2013 |
/* allocate vm area, map the pages and copy static data */ vm.flags = VM_ALLOC; |
fd1e8a1fe
|
2014 |
vm.size = num_possible_cpus() * ai->unit_size; |
8f05a6a65
|
2015 |
vm_area_register_early(&vm, PAGE_SIZE); |
fd1e8a1fe
|
2016 |
for (unit = 0; unit < num_possible_cpus(); unit++) { |
1d9d32572
|
2017 |
unsigned long unit_addr = |
fd1e8a1fe
|
2018 |
(unsigned long)vm.addr + unit * ai->unit_size; |
8f05a6a65
|
2019 |
|
ce3141a27
|
2020 |
for (i = 0; i < unit_pages; i++) |
8f05a6a65
|
2021 2022 2023 |
populate_pte_fn(unit_addr + (i << PAGE_SHIFT)); /* pte already populated, the following shouldn't fail */ |
fb435d523
|
2024 2025 2026 2027 2028 |
rc = __pcpu_map_pages(unit_addr, &pages[unit * unit_pages], unit_pages); if (rc < 0) panic("failed to map percpu area, err=%d ", rc); |
66c3a7577
|
2029 |
|
8f05a6a65
|
2030 2031 2032 2033 2034 2035 2036 2037 2038 |
/* * FIXME: Archs with virtual cache should flush local * cache for the linear mapping here - something * equivalent to flush_cache_vmap() on the local cpu. * flush_cache_vmap() can't be used as most supporting * data structures are not set up yet. */ /* copy static data */ |
fd1e8a1fe
|
2039 |
memcpy((void *)unit_addr, __per_cpu_load, ai->static_size); |
66c3a7577
|
2040 2041 2042 |
} /* we're ready, commit */ |
1d9d32572
|
2043 2044 |
pr_info("PERCPU: %d %s pages/cpu @%p s%zu r%zu d%zu ", |
fd1e8a1fe
|
2045 2046 |
unit_pages, psize_str, vm.addr, ai->static_size, ai->reserved_size, ai->dyn_size); |
d4b95f803
|
2047 |
|
fb435d523
|
2048 |
rc = pcpu_setup_first_chunk(ai, vm.addr); |
d4b95f803
|
2049 2050 2051 2052 |
goto out_free_ar; enomem: while (--j >= 0) |
ce3141a27
|
2053 |
free_fn(page_address(pages[j]), PAGE_SIZE); |
fb435d523
|
2054 |
rc = -ENOMEM; |
d4b95f803
|
2055 |
out_free_ar: |
ce3141a27
|
2056 |
free_bootmem(__pa(pages), pages_size); |
fd1e8a1fe
|
2057 |
pcpu_free_alloc_info(ai); |
fb435d523
|
2058 |
return rc; |
d4b95f803
|
2059 |
} |
08fc45806
|
2060 |
#endif /* CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK */ |
d4b95f803
|
2061 |
|
8c4bfc6e8
|
2062 |
/* |
e74e39620
|
2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 |
* Generic percpu area setup. * * The embedding helper is used because its behavior closely resembles * the original non-dynamic generic percpu area setup. This is * important because many archs have addressing restrictions and might * fail if the percpu area is located far away from the previous * location. As an added bonus, in non-NUMA cases, embedding is * generally a good idea TLB-wise because percpu area can piggy back * on the physical linear memory mapping which uses large page * mappings on applicable archs. */ #ifndef CONFIG_HAVE_SETUP_PER_CPU_AREA unsigned long __per_cpu_offset[NR_CPUS] __read_mostly; EXPORT_SYMBOL(__per_cpu_offset); |
c8826dd53
|
2077 2078 2079 2080 2081 |
static void * __init pcpu_dfl_fc_alloc(unsigned int cpu, size_t size, size_t align) { return __alloc_bootmem_nopanic(size, align, __pa(MAX_DMA_ADDRESS)); } |
66c3a7577
|
2082 |
|
c8826dd53
|
2083 2084 2085 2086 |
static void __init pcpu_dfl_fc_free(void *ptr, size_t size) { free_bootmem(__pa(ptr), size); } |
e74e39620
|
2087 2088 |
void __init setup_per_cpu_areas(void) { |
e74e39620
|
2089 2090 |
unsigned long delta; unsigned int cpu; |
fb435d523
|
2091 |
int rc; |
e74e39620
|
2092 2093 2094 2095 2096 |
/* * Always reserve area for module percpu variables. That's * what the legacy allocator did. */ |
fb435d523
|
2097 |
rc = pcpu_embed_first_chunk(PERCPU_MODULE_RESERVE, |
c8826dd53
|
2098 2099 |
PERCPU_DYNAMIC_RESERVE, PAGE_SIZE, NULL, pcpu_dfl_fc_alloc, pcpu_dfl_fc_free); |
fb435d523
|
2100 |
if (rc < 0) |
e74e39620
|
2101 2102 2103 2104 |
panic("Failed to initialized percpu areas."); delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start; for_each_possible_cpu(cpu) |
fb435d523
|
2105 |
__per_cpu_offset[cpu] = delta + pcpu_unit_offsets[cpu]; |
66c3a7577
|
2106 |
} |
e74e39620
|
2107 |
#endif /* CONFIG_HAVE_SETUP_PER_CPU_AREA */ |