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
|
#include "mmap.h"
#include "atomic.h"
#include "pk.h"
#include "boot.h"
#include "bits.h"
#include "mtrap.h"
#include <stdint.h>
#include <errno.h>
typedef struct {
uintptr_t addr;
size_t length;
file_t* file;
size_t offset;
unsigned refcnt;
int prot;
} vmr_t;
#define MAX_VMR (RISCV_PGSIZE / sizeof(vmr_t))
static spinlock_t vm_lock = SPINLOCK_INIT;
static vmr_t* vmrs;
static uintptr_t first_free_page;
static size_t next_free_page;
static size_t free_pages;
int demand_paging = 1; // unless -p flag is given
static uintptr_t __page_alloc()
{
kassert(next_free_page != free_pages);
uintptr_t addr = first_free_page + RISCV_PGSIZE * next_free_page++;
memset((void*)addr, 0, RISCV_PGSIZE);
return addr;
}
static vmr_t* __vmr_alloc(uintptr_t addr, size_t length, file_t* file,
size_t offset, unsigned refcnt, int prot)
{
if (!vmrs) {
spinlock_lock(&vm_lock);
if (!vmrs)
vmrs = (vmr_t*)__page_alloc();
spinlock_unlock(&vm_lock);
}
for (vmr_t* v = vmrs; v < vmrs + MAX_VMR; v++) {
if (v->refcnt == 0) {
if (file)
file_incref(file);
v->addr = addr;
v->length = length;
v->file = file;
v->offset = offset;
v->refcnt = refcnt;
v->prot = prot;
return v;
}
}
return NULL;
}
static void __vmr_decref(vmr_t* v, unsigned dec)
{
if ((v->refcnt -= dec) == 0)
{
if (v->file)
file_decref(v->file);
}
}
static size_t pte_ppn(pte_t pte)
{
return pte >> PTE_PPN_SHIFT;
}
static uintptr_t ppn(uintptr_t addr)
{
return addr >> RISCV_PGSHIFT;
}
static size_t pt_idx(uintptr_t addr, int level)
{
size_t idx = addr >> (RISCV_PGLEVEL_BITS*level + RISCV_PGSHIFT);
return idx & ((1 << RISCV_PGLEVEL_BITS) - 1);
}
static pte_t* __walk_create(uintptr_t addr);
static pte_t* __attribute__((noinline)) __continue_walk_create(uintptr_t addr, pte_t* pte)
{
*pte = ptd_create(ppn(__page_alloc()));
return __walk_create(addr);
}
static pte_t* __walk_internal(uintptr_t addr, int create)
{
pte_t* t = root_page_table;
for (int i = (VA_BITS - RISCV_PGSHIFT) / RISCV_PGLEVEL_BITS - 1; i > 0; i--) {
size_t idx = pt_idx(addr, i);
if (unlikely(!(t[idx] & PTE_V)))
return create ? __continue_walk_create(addr, &t[idx]) : 0;
t = (pte_t*)(pte_ppn(t[idx]) << RISCV_PGSHIFT);
}
return &t[pt_idx(addr, 0)];
}
static pte_t* __walk(uintptr_t addr)
{
return __walk_internal(addr, 0);
}
static pte_t* __walk_create(uintptr_t addr)
{
return __walk_internal(addr, 1);
}
static int __va_avail(uintptr_t vaddr)
{
pte_t* pte = __walk(vaddr);
return pte == 0 || *pte == 0;
}
static uintptr_t __vm_alloc(size_t npage)
{
uintptr_t start = current.brk, end = current.mmap_max - npage*RISCV_PGSIZE;
for (uintptr_t a = start; a <= end; a += RISCV_PGSIZE)
{
if (!__va_avail(a))
continue;
uintptr_t first = a, last = a + (npage-1) * RISCV_PGSIZE;
for (a = last; a > first && __va_avail(a); a -= RISCV_PGSIZE)
;
if (a > first)
continue;
return a;
}
return 0;
}
static inline pte_t prot_to_type(int prot, int user)
{
pte_t pte = 0;
if (prot & PROT_READ) pte |= PTE_R;
if (prot & PROT_WRITE) pte |= PTE_W;
if (prot & PROT_EXEC) pte |= PTE_X;
if (pte == 0) pte = PTE_R;
if (user) pte |= PTE_U;
return pte;
}
int __valid_user_range(uintptr_t vaddr, size_t len)
{
if (vaddr + len < vaddr)
return 0;
return vaddr + len <= current.mmap_max;
}
static int __handle_page_fault(uintptr_t vaddr, int prot)
{
uintptr_t vpn = vaddr >> RISCV_PGSHIFT;
vaddr = vpn << RISCV_PGSHIFT;
pte_t* pte = __walk(vaddr);
if (pte == 0 || *pte == 0 || !__valid_user_range(vaddr, 1))
return -1;
else if (!(*pte & PTE_V))
{
uintptr_t ppn = vpn + (first_free_paddr / RISCV_PGSIZE);
vmr_t* v = (vmr_t*)*pte;
*pte = pte_create(ppn, prot_to_type(PROT_READ|PROT_WRITE, 0));
flush_tlb();
if (v->file)
{
size_t flen = MIN(RISCV_PGSIZE, v->length - (vaddr - v->addr));
ssize_t ret = file_pread(v->file, (void*)vaddr, flen, vaddr - v->addr + v->offset);
kassert(ret > 0);
if (ret < RISCV_PGSIZE)
memset((void*)vaddr + ret, 0, RISCV_PGSIZE - ret);
}
else
memset((void*)vaddr, 0, RISCV_PGSIZE);
__vmr_decref(v, 1);
*pte = pte_create(ppn, prot_to_type(v->prot, 1));
}
pte_t perms = pte_create(0, prot_to_type(prot, 1));
if ((*pte & perms) != perms)
return -1;
flush_tlb();
return 0;
}
int handle_page_fault(uintptr_t vaddr, int prot)
{
spinlock_lock(&vm_lock);
int ret = __handle_page_fault(vaddr, prot);
spinlock_unlock(&vm_lock);
return ret;
}
static void __do_munmap(uintptr_t addr, size_t len)
{
for (uintptr_t a = addr; a < addr + len; a += RISCV_PGSIZE)
{
pte_t* pte = __walk(a);
if (pte == 0 || *pte == 0)
continue;
if (!(*pte & PTE_V))
__vmr_decref((vmr_t*)*pte, 1);
*pte = 0;
}
flush_tlb(); // TODO: shootdown
}
uintptr_t __do_mmap(uintptr_t addr, size_t length, int prot, int flags, file_t* f, off_t offset)
{
size_t npage = (length-1)/RISCV_PGSIZE+1;
if (flags & MAP_FIXED)
{
if ((addr & (RISCV_PGSIZE-1)) || !__valid_user_range(addr, length))
return (uintptr_t)-1;
}
else if ((addr = __vm_alloc(npage)) == 0)
return (uintptr_t)-1;
vmr_t* v = __vmr_alloc(addr, length, f, offset, npage, prot);
if (!v)
return (uintptr_t)-1;
for (uintptr_t a = addr; a < addr + length; a += RISCV_PGSIZE)
{
pte_t* pte = __walk_create(a);
kassert(pte);
if (*pte)
__do_munmap(a, RISCV_PGSIZE);
*pte = (pte_t)v;
}
if (!demand_paging || (flags & MAP_POPULATE))
for (uintptr_t a = addr; a < addr + length; a += RISCV_PGSIZE)
kassert(__handle_page_fault(a, prot) == 0);
return addr;
}
int do_munmap(uintptr_t addr, size_t length)
{
if ((addr & (RISCV_PGSIZE-1)) || !__valid_user_range(addr, length))
return -EINVAL;
spinlock_lock(&vm_lock);
__do_munmap(addr, length);
spinlock_unlock(&vm_lock);
return 0;
}
uintptr_t do_mmap(uintptr_t addr, size_t length, int prot, int flags, int fd, off_t offset)
{
if (!(flags & MAP_PRIVATE) || length == 0 || (offset & (RISCV_PGSIZE-1)))
return -EINVAL;
file_t* f = NULL;
if (!(flags & MAP_ANONYMOUS) && (f = file_get(fd)) == NULL)
return -EBADF;
spinlock_lock(&vm_lock);
addr = __do_mmap(addr, length, prot, flags, f, offset);
if (addr < current.brk_max)
current.brk_max = addr;
spinlock_unlock(&vm_lock);
if (f) file_decref(f);
return addr;
}
uintptr_t __do_brk(size_t addr)
{
uintptr_t newbrk = addr;
if (addr < current.brk_min)
newbrk = current.brk_min;
else if (addr > current.brk_max)
newbrk = current.brk_max;
if (current.brk == 0)
current.brk = ROUNDUP(current.brk_min, RISCV_PGSIZE);
uintptr_t newbrk_page = ROUNDUP(newbrk, RISCV_PGSIZE);
if (current.brk > newbrk_page)
__do_munmap(newbrk_page, current.brk - newbrk_page);
else if (current.brk < newbrk_page)
kassert(__do_mmap(current.brk, newbrk_page - current.brk, -1, MAP_FIXED|MAP_PRIVATE|MAP_ANONYMOUS, 0, 0) == current.brk);
current.brk = newbrk_page;
return newbrk;
}
uintptr_t do_brk(size_t addr)
{
spinlock_lock(&vm_lock);
addr = __do_brk(addr);
spinlock_unlock(&vm_lock);
return addr;
}
uintptr_t do_mremap(uintptr_t addr, size_t old_size, size_t new_size, int flags)
{
return -ENOSYS;
}
uintptr_t do_mprotect(uintptr_t addr, size_t length, int prot)
{
uintptr_t res = 0;
if ((addr) & (RISCV_PGSIZE-1))
return -EINVAL;
spinlock_lock(&vm_lock);
for (uintptr_t a = addr; a < addr + length; a += RISCV_PGSIZE)
{
pte_t* pte = __walk(a);
if (pte == 0 || *pte == 0) {
res = -ENOMEM;
break;
}
if (!(*pte & PTE_V)) {
vmr_t* v = (vmr_t*)*pte;
if((v->prot ^ prot) & ~v->prot){
//TODO:look at file to find perms
res = -EACCES;
break;
}
v->prot = prot;
} else {
if (!(*pte & PTE_U) ||
((prot & PROT_READ) && !(*pte & PTE_R)) ||
((prot & PROT_WRITE) && !(*pte & PTE_W)) ||
((prot & PROT_EXEC) && !(*pte & PTE_X))) {
//TODO:look at file to find perms
res = -EACCES;
break;
}
*pte = pte_create(pte_ppn(*pte), prot_to_type(prot, 1));
}
}
spinlock_unlock(&vm_lock);
return res;
}
void __map_kernel_range(uintptr_t vaddr, uintptr_t paddr, size_t len, int prot)
{
uintptr_t n = ROUNDUP(len, RISCV_PGSIZE) / RISCV_PGSIZE;
uintptr_t offset = paddr - vaddr;
for (uintptr_t a = vaddr, i = 0; i < n; i++, a += RISCV_PGSIZE)
{
pte_t* pte = __walk_create(a);
kassert(pte);
*pte = pte_create((a + offset) >> RISCV_PGSHIFT, prot_to_type(prot, 0));
}
}
void populate_mapping(const void* start, size_t size, int prot)
{
uintptr_t a0 = ROUNDDOWN((uintptr_t)start, RISCV_PGSIZE);
for (uintptr_t a = a0; a < (uintptr_t)start+size; a += RISCV_PGSIZE)
{
if (prot & PROT_WRITE)
atomic_add((int*)a, 0);
else
atomic_read((int*)a);
}
}
uintptr_t pk_vm_init()
{
size_t mem_pages = mem_size >> RISCV_PGSHIFT;
free_pages = MAX(8, mem_pages >> (RISCV_PGLEVEL_BITS-1));
first_free_page = first_free_paddr;
first_free_paddr += free_pages * RISCV_PGSIZE;
root_page_table = (void*)__page_alloc();
__map_kernel_range(DRAM_BASE, DRAM_BASE, first_free_paddr - DRAM_BASE, PROT_READ|PROT_WRITE|PROT_EXEC);
current.mmap_max = current.brk_max =
MIN(DRAM_BASE, mem_size - (first_free_paddr - DRAM_BASE));
size_t stack_size = RISCV_PGSIZE * 64;
size_t stack_bottom = __do_mmap(current.mmap_max - stack_size, stack_size, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE|MAP_ANONYMOUS|MAP_FIXED, 0, 0);
kassert(stack_bottom != (uintptr_t)-1);
current.stack_top = stack_bottom + stack_size;
uintptr_t kernel_stack_top = __page_alloc() + RISCV_PGSIZE;
return kernel_stack_top;
}
|