1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
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
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
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
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
|
/*
* Copyright (c) 1997 John S. Dyson
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice immediately at the beginning of the file, without modification,
* this list of conditions, and the following disclaimer.
* 2. Absolutely no warranty of function or purpose is made by the author
* John S. Dyson.
*
* $Id: vm_zone.c,v 1.17 1998/02/06 12:14:29 eivind Exp $
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/sysctl.h>
#include <vm/vm.h>
#include <vm/vm_object.h>
#include <vm/vm_prot.h>
#include <vm/vm_page.h>
#include <vm/vm_map.h>
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <vm/vm_zone.h>
static MALLOC_DEFINE(M_ZONE, "ZONE", "Zone header");
/*
* This file comprises a very simple zone allocator. This is used
* in lieu of the malloc allocator, where needed or more optimal.
*
* Note that the initial implementation of this had coloring, and
* absolutely no improvement (actually perf degradation) occurred.
*
* Note also that the zones are type stable. The only restriction is
* that the first two longwords of a data structure can be changed
* between allocations. Any data that must be stable between allocations
* must reside in areas after the first two longwords.
*
* zinitna, zinit, zbootinit are the initialization routines.
* zalloc, zfree, are the interrupt/lock unsafe allocation/free routines.
* zalloci, zfreei, are the interrupt/lock safe allocation/free routines.
*/
static struct vm_zone *zlist;
static int sysctl_vm_zone SYSCTL_HANDLER_ARGS;
/*
* Create a zone, but don't allocate the zone structure. If the
* zone had been previously created by the zone boot code, initialize
* various parts of the zone code.
*
* If waits are not allowed during allocation (e.g. during interrupt
* code), a-priori allocate the kernel virtual space, and allocate
* only pages when needed.
*
* Arguments:
* z pointer to zone structure.
* obj pointer to VM object (opt).
* name name of zone.
* size size of zone entries.
* nentries number of zone entries allocated (only ZONE_INTERRUPT.)
* flags ZONE_INTERRUPT -- items can be allocated at interrupt time.
* zalloc number of pages allocated when memory is needed.
*
* Note that when using ZONE_INTERRUPT, the size of the zone is limited
* by the nentries argument. The size of the memory allocatable is
* unlimited if ZONE_INTERRUPT is not set.
*
*/
int
zinitna(vm_zone_t z, vm_object_t obj, char *name, int size,
int nentries, int flags, int zalloc)
{
int totsize;
if ((z->zflags & ZONE_BOOT) == 0) {
z->zsize = (size + ZONE_ROUNDING - 1) & ~(ZONE_ROUNDING - 1);
simple_lock_init(&z->zlock);
z->zfreecnt = 0;
z->ztotal = 0;
z->zmax = 0;
z->zname = name;
z->znalloc = 0;
z->zitems = NULL;
if (zlist == 0) {
zlist = z;
} else {
z->znext = zlist;
zlist = z;
}
}
z->zflags |= flags;
/*
* If we cannot wait, allocate KVA space up front, and we will fill
* in pages as needed.
*/
if (z->zflags & ZONE_INTERRUPT) {
totsize = round_page(z->zsize * nentries);
z->zkva = kmem_alloc_pageable(kernel_map, totsize);
if (z->zkva == 0)
return 0;
z->zpagemax = totsize / PAGE_SIZE;
if (obj == NULL) {
z->zobj = vm_object_allocate(OBJT_DEFAULT, z->zpagemax);
} else {
z->zobj = obj;
_vm_object_allocate(OBJT_DEFAULT, z->zpagemax, obj);
}
z->zallocflag = VM_ALLOC_INTERRUPT;
z->zmax += nentries;
} else {
z->zallocflag = VM_ALLOC_SYSTEM;
z->zmax = 0;
}
if (z->zsize > PAGE_SIZE)
z->zfreemin = 1;
else
z->zfreemin = PAGE_SIZE / z->zsize;
z->zpagecount = 0;
if (zalloc)
z->zalloc = zalloc;
else
z->zalloc = 1;
return 1;
}
/*
* Subroutine same as zinitna, except zone data structure is allocated
* automatically by malloc. This routine should normally be used, except
* in certain tricky startup conditions in the VM system -- then
* zbootinit and zinitna can be used. Zinit is the standard zone
* initialization call.
*/
vm_zone_t
zinit(char *name, int size, int nentries, int flags, int zalloc)
{
vm_zone_t z;
z = (vm_zone_t) malloc(sizeof (struct vm_zone), M_ZONE, M_NOWAIT);
if (z == NULL)
return NULL;
z->zflags = 0;
if (zinitna(z, NULL, name, size, nentries, flags, zalloc) == 0) {
free(z, M_ZONE);
return NULL;
}
return z;
}
/*
* Initialize a zone before the system is fully up. This routine should
* only be called before full VM startup.
*/
void
zbootinit(vm_zone_t z, char *name, int size, void *item, int nitems)
{
int i;
z->zname = name;
z->zsize = size;
z->zpagemax = 0;
z->zobj = NULL;
z->zflags = ZONE_BOOT;
z->zfreemin = 0;
z->zallocflag = 0;
z->zpagecount = 0;
z->zalloc = 0;
z->znalloc = 0;
simple_lock_init(&z->zlock);
bzero(item, nitems * z->zsize);
z->zitems = NULL;
for (i = 0; i < nitems; i++) {
((void **) item)[0] = z->zitems;
#if defined(DIAGNOSTIC)
((void **) item)[1] = (void *) ZENTRY_FREE;
#endif
z->zitems = item;
(char *) item += z->zsize;
}
z->zfreecnt = nitems;
z->zmax = nitems;
z->ztotal = nitems;
if (zlist == 0) {
zlist = z;
} else {
z->znext = zlist;
zlist = z;
}
}
/*
* Zone critical region locks.
*/
static inline int
zlock(vm_zone_t z)
{
int s;
s = splhigh();
simple_lock(&z->zlock);
return s;
}
static inline void
zunlock(vm_zone_t z, int s)
{
simple_unlock(&z->zlock);
splx(s);
}
/*
* void *zalloc(vm_zone_t zone) --
* Returns an item from a specified zone.
*
* void zfree(vm_zone_t zone, void *item) --
* Frees an item back to a specified zone.
*
* void *zalloci(vm_zone_t zone) --
* Returns an item from a specified zone, interrupt safe.
*
* void zfreei(vm_zone_t zone, void *item) --
* Frees an item back to a specified zone, interrupt safe.
*
*/
/*
* Zone allocator/deallocator. These are interrupt / (or potentially SMP)
* safe. The raw zalloc/zfree routines are in the vm_zone header file,
* and are not interrupt safe, but are fast.
*/
void *
zalloci(vm_zone_t z)
{
int s;
void *item;
s = zlock(z);
item = _zalloc(z);
zunlock(z, s);
return item;
}
void
zfreei(vm_zone_t z, void *item)
{
int s;
s = zlock(z);
_zfree(z, item);
zunlock(z, s);
return;
}
/*
* Internal zone routine. Not to be called from external (non vm_zone) code.
*/
void *
_zget(vm_zone_t z)
{
int i;
vm_page_t m;
int nitems, nbytes;
void *item;
if (z == NULL)
panic("zget: null zone");
if (z->zflags & ZONE_INTERRUPT) {
item = (char *) z->zkva + z->zpagecount * PAGE_SIZE;
for (i = 0; ((i < z->zalloc) && (z->zpagecount < z->zpagemax));
i++) {
vm_offset_t zkva;
m = vm_page_alloc(z->zobj, z->zpagecount,
z->zallocflag);
if (m == NULL)
break;
zkva = z->zkva + z->zpagecount * PAGE_SIZE;
pmap_kenter(zkva, VM_PAGE_TO_PHYS(m));
bzero((caddr_t) zkva, PAGE_SIZE);
z->zpagecount++;
}
nitems = (i * PAGE_SIZE) / z->zsize;
} else {
nbytes = z->zalloc * PAGE_SIZE;
/*
* Check to see if the kernel map is already locked. We could allow
* for recursive locks, but that eliminates a valuable debugging
* mechanism, and opens up the kernel map for potential corruption
* by inconsistent data structure manipulation. We could also use
* the interrupt allocation mechanism, but that has size limitations.
* Luckily, we have kmem_map that is a submap of kernel map available
* for memory allocation, and manipulation of that map doesn't affect
* the kernel map structures themselves.
*
* We can wait, so just do normal map allocation in the appropriate
* map.
*/
if (lockstatus(&kernel_map->lock)) {
int s;
s = splvm();
item = (void *) kmem_malloc(kmem_map, nbytes, M_WAITOK);
splx(s);
} else {
item = (void *) kmem_alloc(kernel_map, nbytes);
}
bzero(item, nbytes);
nitems = nbytes / z->zsize;
}
z->ztotal += nitems;
/*
* Save one for immediate allocation
*/
if (nitems != 0) {
nitems -= 1;
for (i = 0; i < nitems; i++) {
((void **) item)[0] = z->zitems;
#if defined(DIAGNOSTIC)
((void **) item)[1] = (void *) ZENTRY_FREE;
#endif
z->zitems = item;
(char *) item += z->zsize;
}
z->zfreecnt += nitems;
} else if (z->zfreecnt > 0) {
item = z->zitems;
z->zitems = ((void **) item)[0];
#if defined(DIAGNOSTIC)
if (((void **) item)[1] != (void *) ZENTRY_FREE)
zerror(ZONE_ERROR_NOTFREE);
((void **) item)[1] = 0;
#endif
z->zfreecnt--;
} else {
item = NULL;
}
return item;
}
static int
sysctl_vm_zone SYSCTL_HANDLER_ARGS
{
int error=0;
vm_zone_t curzone, nextzone;
char tmpbuf[128];
char tmpname[14];
sprintf(tmpbuf, "\nITEM SIZE LIMIT USED FREE REQUESTS\n");
error = SYSCTL_OUT(req, tmpbuf, strlen(tmpbuf));
if (error)
return (error);
for (curzone = zlist; curzone; curzone = nextzone) {
int i;
int len;
int offset;
nextzone = curzone->znext;
len = strlen(curzone->zname);
if (len >= (sizeof(tmpname) - 1))
len = (sizeof(tmpname) - 1);
for(i = 0; i < sizeof(tmpname) - 1; i++)
tmpname[i] = ' ';
tmpname[i] = 0;
memcpy(tmpname, curzone->zname, len);
tmpname[len] = ':';
offset = 0;
if (curzone == zlist) {
offset = 1;
tmpbuf[0] = '\n';
}
sprintf(tmpbuf + offset,
"%s %6.6u, %8.8u, %6.6u, %6.6u, %8.8u\n",
tmpname, curzone->zsize, curzone->zmax,
(curzone->ztotal - curzone->zfreecnt),
curzone->zfreecnt, curzone->znalloc);
len = strlen((char *)tmpbuf);
if (nextzone == NULL)
tmpbuf[len - 1] = 0;
error = SYSCTL_OUT(req, tmpbuf, len);
if (error)
return (error);
}
return (0);
}
#if defined(DIAGNOSTIC)
void
zerror(int error)
{
char *msg;
switch (error) {
case ZONE_ERROR_INVALID:
msg = "zone: invalid zone";
break;
case ZONE_ERROR_NOTFREE:
msg = "zone: entry not free";
break;
case ZONE_ERROR_ALREADYFREE:
msg = "zone: freeing free entry";
break;
default:
msg = "zone: invalid error";
break;
}
panic(msg);
}
#endif
SYSCTL_OID(_kern, OID_AUTO, zone, CTLTYPE_STRING|CTLFLAG_RD, \
NULL, 0, sysctl_vm_zone, "A", "Zone Info");
|