Stacey Son | 6765e98 | 2023-09-25 21:26:57 +0300 | [diff] [blame] | 1 | /* |
| 2 | * memory management system call shims and definitions |
| 3 | * |
| 4 | * Copyright (c) 2013-15 Stacey D. Son |
| 5 | * |
| 6 | * This program is free software; you can redistribute it and/or modify |
| 7 | * it under the terms of the GNU General Public License as published by |
| 8 | * the Free Software Foundation; either version 2 of the License, or |
| 9 | * (at your option) any later version. |
| 10 | * |
| 11 | * This program is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 | * GNU General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU General Public License |
| 17 | * along with this program; if not, see <http://www.gnu.org/licenses/>. |
| 18 | */ |
| 19 | |
| 20 | /* |
| 21 | * Copyright (c) 1982, 1986, 1993 |
| 22 | * The Regents of the University of California. All rights reserved. |
| 23 | * |
| 24 | * Redistribution and use in source and binary forms, with or without |
| 25 | * modification, are permitted provided that the following conditions |
| 26 | * are met: |
| 27 | * 1. Redistributions of source code must retain the above copyright |
| 28 | * notice, this list of conditions and the following disclaimer. |
| 29 | * 2. Redistributions in binary form must reproduce the above copyright |
| 30 | * notice, this list of conditions and the following disclaimer in the |
| 31 | * documentation and/or other materials provided with the distribution. |
| 32 | * 4. Neither the name of the University nor the names of its contributors |
| 33 | * may be used to endorse or promote products derived from this software |
| 34 | * without specific prior written permission. |
| 35 | * |
| 36 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
| 37 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 38 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 39 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
| 40 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 41 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 42 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 43 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 44 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 45 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 46 | * SUCH DAMAGE. |
| 47 | */ |
| 48 | |
| 49 | #ifndef BSD_USER_BSD_MEM_H |
| 50 | #define BSD_USER_BSD_MEM_H |
| 51 | |
| 52 | #include <sys/types.h> |
| 53 | #include <sys/ipc.h> |
| 54 | #include <sys/mman.h> |
| 55 | #include <sys/shm.h> |
| 56 | #include <fcntl.h> |
| 57 | |
| 58 | #include "qemu-bsd.h" |
Philippe Mathieu-Daudé | 74781c0 | 2023-12-06 20:27:32 +0100 | [diff] [blame] | 59 | #include "exec/page-protection.h" |
Stacey Son | 6765e98 | 2023-09-25 21:26:57 +0300 | [diff] [blame] | 60 | |
| 61 | extern struct bsd_shm_regions bsd_shm_regions[]; |
| 62 | extern abi_ulong target_brk; |
| 63 | extern abi_ulong initial_target_brk; |
| 64 | |
Stacey Son | 87dcb4a | 2023-09-25 21:26:58 +0300 | [diff] [blame] | 65 | /* mmap(2) */ |
| 66 | static inline abi_long do_bsd_mmap(void *cpu_env, abi_long arg1, abi_long arg2, |
| 67 | abi_long arg3, abi_long arg4, abi_long arg5, abi_long arg6, abi_long arg7, |
| 68 | abi_long arg8) |
| 69 | { |
| 70 | if (regpairs_aligned(cpu_env) != 0) { |
| 71 | arg6 = arg7; |
| 72 | arg7 = arg8; |
| 73 | } |
| 74 | return get_errno(target_mmap(arg1, arg2, arg3, |
| 75 | target_to_host_bitmask(arg4, mmap_flags_tbl), |
| 76 | arg5, target_arg64(arg6, arg7))); |
| 77 | } |
| 78 | |
| 79 | /* munmap(2) */ |
| 80 | static inline abi_long do_bsd_munmap(abi_long arg1, abi_long arg2) |
| 81 | { |
| 82 | return get_errno(target_munmap(arg1, arg2)); |
| 83 | } |
| 84 | |
Stacey Son | ecbe224 | 2023-09-25 21:26:59 +0300 | [diff] [blame] | 85 | /* mprotect(2) */ |
| 86 | static inline abi_long do_bsd_mprotect(abi_long arg1, abi_long arg2, |
| 87 | abi_long arg3) |
| 88 | { |
| 89 | return get_errno(target_mprotect(arg1, arg2, arg3)); |
| 90 | } |
| 91 | |
Stacey Son | f28a1e4 | 2023-09-25 21:27:00 +0300 | [diff] [blame] | 92 | /* msync(2) */ |
| 93 | static inline abi_long do_bsd_msync(abi_long addr, abi_long len, abi_long flags) |
| 94 | { |
| 95 | if (!guest_range_valid_untagged(addr, len)) { |
| 96 | /* It seems odd, but POSIX wants this to be ENOMEM */ |
| 97 | return -TARGET_ENOMEM; |
| 98 | } |
| 99 | |
| 100 | return get_errno(msync(g2h_untagged(addr), len, flags)); |
| 101 | } |
| 102 | |
Stacey Son | 0a49ef0 | 2023-09-25 21:27:01 +0300 | [diff] [blame] | 103 | /* mlock(2) */ |
| 104 | static inline abi_long do_bsd_mlock(abi_long arg1, abi_long arg2) |
| 105 | { |
| 106 | if (!guest_range_valid_untagged(arg1, arg2)) { |
| 107 | return -TARGET_EINVAL; |
| 108 | } |
| 109 | return get_errno(mlock(g2h_untagged(arg1), arg2)); |
| 110 | } |
| 111 | |
| 112 | /* munlock(2) */ |
| 113 | static inline abi_long do_bsd_munlock(abi_long arg1, abi_long arg2) |
| 114 | { |
| 115 | if (!guest_range_valid_untagged(arg1, arg2)) { |
| 116 | return -TARGET_EINVAL; |
| 117 | } |
| 118 | return get_errno(munlock(g2h_untagged(arg1), arg2)); |
| 119 | } |
| 120 | |
| 121 | /* mlockall(2) */ |
| 122 | static inline abi_long do_bsd_mlockall(abi_long arg1) |
| 123 | { |
| 124 | return get_errno(mlockall(arg1)); |
| 125 | } |
| 126 | |
| 127 | /* munlockall(2) */ |
| 128 | static inline abi_long do_bsd_munlockall(void) |
| 129 | { |
| 130 | return get_errno(munlockall()); |
| 131 | } |
| 132 | |
Karim Taha | 0c1ced4 | 2023-09-25 21:27:02 +0300 | [diff] [blame] | 133 | /* madvise(2) */ |
| 134 | static inline abi_long do_bsd_madvise(abi_long arg1, abi_long arg2, |
| 135 | abi_long arg3) |
| 136 | { |
| 137 | abi_ulong len; |
| 138 | int ret = 0; |
| 139 | abi_long start = arg1; |
| 140 | abi_long len_in = arg2; |
| 141 | abi_long advice = arg3; |
| 142 | |
| 143 | if (start & ~TARGET_PAGE_MASK) { |
| 144 | return -TARGET_EINVAL; |
| 145 | } |
| 146 | if (len_in == 0) { |
| 147 | return 0; |
| 148 | } |
| 149 | len = TARGET_PAGE_ALIGN(len_in); |
| 150 | if (len == 0 || !guest_range_valid_untagged(start, len)) { |
| 151 | return -TARGET_EINVAL; |
| 152 | } |
| 153 | |
| 154 | /* |
| 155 | * Most advice values are hints, so ignoring and returning success is ok. |
| 156 | * |
| 157 | * However, some advice values such as MADV_DONTNEED, are not hints and |
| 158 | * need to be emulated. |
| 159 | * |
| 160 | * A straight passthrough for those may not be safe because qemu sometimes |
| 161 | * turns private file-backed mappings into anonymous mappings. |
| 162 | * If all guest pages have PAGE_PASSTHROUGH set, mappings have the |
| 163 | * same semantics for the host as for the guest. |
| 164 | * |
| 165 | * MADV_DONTNEED is passed through, if possible. |
| 166 | * If passthrough isn't possible, we nevertheless (wrongly!) return |
| 167 | * success, which is broken but some userspace programs fail to work |
| 168 | * otherwise. Completely implementing such emulation is quite complicated |
| 169 | * though. |
| 170 | */ |
| 171 | mmap_lock(); |
| 172 | switch (advice) { |
| 173 | case MADV_DONTNEED: |
| 174 | if (page_check_range(start, len, PAGE_PASSTHROUGH)) { |
| 175 | ret = get_errno(madvise(g2h_untagged(start), len, advice)); |
| 176 | if (ret == 0) { |
| 177 | page_reset_target_data(start, start + len - 1); |
| 178 | } |
| 179 | } |
| 180 | } |
| 181 | mmap_unlock(); |
| 182 | |
| 183 | return ret; |
| 184 | } |
| 185 | |
Stacey Son | 0a49ef0 | 2023-09-25 21:27:01 +0300 | [diff] [blame] | 186 | /* minherit(2) */ |
| 187 | static inline abi_long do_bsd_minherit(abi_long addr, abi_long len, |
| 188 | abi_long inherit) |
| 189 | { |
| 190 | return get_errno(minherit(g2h_untagged(addr), len, inherit)); |
| 191 | } |
| 192 | |
Stacey Son | 83b045a | 2023-09-25 21:27:03 +0300 | [diff] [blame] | 193 | /* mincore(2) */ |
| 194 | static inline abi_long do_bsd_mincore(abi_ulong target_addr, abi_ulong len, |
| 195 | abi_ulong target_vec) |
| 196 | { |
| 197 | abi_long ret; |
| 198 | void *p; |
| 199 | abi_ulong vec_len = DIV_ROUND_UP(len, TARGET_PAGE_SIZE); |
| 200 | |
| 201 | if (!guest_range_valid_untagged(target_addr, len) |
| 202 | || !page_check_range(target_addr, len, PAGE_VALID)) { |
| 203 | return -TARGET_EFAULT; |
| 204 | } |
| 205 | |
| 206 | p = lock_user(VERIFY_WRITE, target_vec, vec_len, 0); |
| 207 | if (p == NULL) { |
| 208 | return -TARGET_EFAULT; |
| 209 | } |
| 210 | ret = get_errno(mincore(g2h_untagged(target_addr), len, p)); |
| 211 | unlock_user(p, target_vec, vec_len); |
| 212 | |
| 213 | return ret; |
| 214 | } |
| 215 | |
Stacey Son | a99d740 | 2023-09-25 21:27:04 +0300 | [diff] [blame] | 216 | /* do_brk() must return target values and target errnos. */ |
| 217 | static inline abi_long do_obreak(abi_ulong brk_val) |
| 218 | { |
| 219 | abi_long mapped_addr; |
| 220 | abi_ulong new_brk; |
| 221 | abi_ulong old_brk; |
| 222 | |
| 223 | /* brk pointers are always untagged */ |
| 224 | |
| 225 | /* do not allow to shrink below initial brk value */ |
| 226 | if (brk_val < initial_target_brk) { |
| 227 | return target_brk; |
| 228 | } |
| 229 | |
| 230 | new_brk = TARGET_PAGE_ALIGN(brk_val); |
| 231 | old_brk = TARGET_PAGE_ALIGN(target_brk); |
| 232 | |
| 233 | /* new and old target_brk might be on the same page */ |
| 234 | if (new_brk == old_brk) { |
| 235 | target_brk = brk_val; |
| 236 | return target_brk; |
| 237 | } |
| 238 | |
Michael Tokarev | 9bfba08 | 2023-11-14 19:06:14 +0300 | [diff] [blame] | 239 | /* Release heap if necessary */ |
Stacey Son | a99d740 | 2023-09-25 21:27:04 +0300 | [diff] [blame] | 240 | if (new_brk < old_brk) { |
| 241 | target_munmap(new_brk, old_brk - new_brk); |
| 242 | |
| 243 | target_brk = brk_val; |
| 244 | return target_brk; |
| 245 | } |
| 246 | |
| 247 | mapped_addr = target_mmap(old_brk, new_brk - old_brk, |
| 248 | PROT_READ | PROT_WRITE, |
| 249 | MAP_FIXED | MAP_EXCL | MAP_ANON | MAP_PRIVATE, |
| 250 | -1, 0); |
| 251 | |
| 252 | if (mapped_addr == old_brk) { |
| 253 | target_brk = brk_val; |
| 254 | return target_brk; |
| 255 | } |
| 256 | |
| 257 | /* For everything else, return the previous break. */ |
| 258 | return target_brk; |
| 259 | } |
| 260 | |
Stacey Son | 4f0be68 | 2023-09-25 21:27:05 +0300 | [diff] [blame] | 261 | /* shm_open(2) */ |
| 262 | static inline abi_long do_bsd_shm_open(abi_ulong arg1, abi_long arg2, |
| 263 | abi_long arg3) |
| 264 | { |
| 265 | int ret; |
| 266 | void *p; |
| 267 | |
| 268 | if (arg1 == (uintptr_t)SHM_ANON) { |
| 269 | p = SHM_ANON; |
| 270 | } else { |
| 271 | p = lock_user_string(arg1); |
| 272 | if (p == NULL) { |
| 273 | return -TARGET_EFAULT; |
| 274 | } |
| 275 | } |
| 276 | ret = get_errno(shm_open(p, target_to_host_bitmask(arg2, fcntl_flags_tbl), |
| 277 | arg3)); |
| 278 | |
| 279 | if (p != SHM_ANON) { |
| 280 | unlock_user(p, arg1, 0); |
| 281 | } |
| 282 | |
| 283 | return ret; |
| 284 | } |
| 285 | |
Stacey Son | 9d14db1 | 2023-09-25 21:27:06 +0300 | [diff] [blame] | 286 | /* shm_unlink(2) */ |
| 287 | static inline abi_long do_bsd_shm_unlink(abi_ulong arg1) |
| 288 | { |
| 289 | int ret; |
| 290 | void *p; |
| 291 | |
| 292 | p = lock_user_string(arg1); |
| 293 | if (p == NULL) { |
| 294 | return -TARGET_EFAULT; |
| 295 | } |
| 296 | ret = get_errno(shm_unlink(p)); /* XXX path(p)? */ |
| 297 | unlock_user(p, arg1, 0); |
| 298 | |
| 299 | return ret; |
| 300 | } |
| 301 | |
| 302 | /* shmget(2) */ |
| 303 | static inline abi_long do_bsd_shmget(abi_long arg1, abi_ulong arg2, |
| 304 | abi_long arg3) |
| 305 | { |
| 306 | return get_errno(shmget(arg1, arg2, arg3)); |
| 307 | } |
| 308 | |
Stacey Son | f9bbe3c | 2023-09-25 21:27:07 +0300 | [diff] [blame] | 309 | /* shmctl(2) */ |
| 310 | static inline abi_long do_bsd_shmctl(abi_long shmid, abi_long cmd, |
| 311 | abi_ulong buff) |
| 312 | { |
| 313 | struct shmid_ds dsarg; |
| 314 | abi_long ret = -TARGET_EINVAL; |
| 315 | |
| 316 | cmd &= 0xff; |
| 317 | |
| 318 | switch (cmd) { |
| 319 | case IPC_STAT: |
| 320 | if (target_to_host_shmid_ds(&dsarg, buff)) { |
| 321 | return -TARGET_EFAULT; |
| 322 | } |
| 323 | ret = get_errno(shmctl(shmid, cmd, &dsarg)); |
| 324 | if (host_to_target_shmid_ds(buff, &dsarg)) { |
| 325 | return -TARGET_EFAULT; |
| 326 | } |
| 327 | break; |
| 328 | |
| 329 | case IPC_SET: |
| 330 | if (target_to_host_shmid_ds(&dsarg, buff)) { |
| 331 | return -TARGET_EFAULT; |
| 332 | } |
| 333 | ret = get_errno(shmctl(shmid, cmd, &dsarg)); |
| 334 | break; |
| 335 | |
| 336 | case IPC_RMID: |
| 337 | ret = get_errno(shmctl(shmid, cmd, NULL)); |
| 338 | break; |
| 339 | |
| 340 | default: |
| 341 | ret = -TARGET_EINVAL; |
| 342 | break; |
| 343 | } |
| 344 | |
| 345 | return ret; |
| 346 | } |
| 347 | |
Stacey Son | 4e00b7d | 2023-09-25 21:27:08 +0300 | [diff] [blame] | 348 | /* shmat(2) */ |
| 349 | static inline abi_long do_bsd_shmat(int shmid, abi_ulong shmaddr, int shmflg) |
| 350 | { |
| 351 | abi_ulong raddr; |
| 352 | abi_long ret; |
| 353 | struct shmid_ds shm_info; |
| 354 | |
| 355 | /* Find out the length of the shared memory segment. */ |
| 356 | ret = get_errno(shmctl(shmid, IPC_STAT, &shm_info)); |
| 357 | if (is_error(ret)) { |
| 358 | /* Can't get the length */ |
| 359 | return ret; |
| 360 | } |
| 361 | |
| 362 | if (!guest_range_valid_untagged(shmaddr, shm_info.shm_segsz)) { |
| 363 | return -TARGET_EINVAL; |
| 364 | } |
| 365 | |
| 366 | WITH_MMAP_LOCK_GUARD() { |
| 367 | void *host_raddr; |
| 368 | |
| 369 | if (shmaddr) { |
| 370 | host_raddr = shmat(shmid, (void *)g2h_untagged(shmaddr), shmflg); |
| 371 | } else { |
| 372 | abi_ulong mmap_start; |
| 373 | |
| 374 | mmap_start = mmap_find_vma(0, shm_info.shm_segsz); |
| 375 | |
| 376 | if (mmap_start == -1) { |
| 377 | return -TARGET_ENOMEM; |
| 378 | } |
| 379 | host_raddr = shmat(shmid, g2h_untagged(mmap_start), |
| 380 | shmflg | SHM_REMAP); |
| 381 | } |
| 382 | |
| 383 | if (host_raddr == (void *)-1) { |
| 384 | return get_errno(-1); |
| 385 | } |
| 386 | raddr = h2g(host_raddr); |
| 387 | |
| 388 | page_set_flags(raddr, raddr + shm_info.shm_segsz - 1, |
| 389 | PAGE_VALID | PAGE_RESET | PAGE_READ | |
| 390 | (shmflg & SHM_RDONLY ? 0 : PAGE_WRITE)); |
| 391 | |
| 392 | for (int i = 0; i < N_BSD_SHM_REGIONS; i++) { |
| 393 | if (bsd_shm_regions[i].start == 0) { |
| 394 | bsd_shm_regions[i].start = raddr; |
| 395 | bsd_shm_regions[i].size = shm_info.shm_segsz; |
| 396 | break; |
| 397 | } |
| 398 | } |
| 399 | } |
| 400 | |
| 401 | return raddr; |
| 402 | } |
| 403 | |
| 404 | /* shmdt(2) */ |
| 405 | static inline abi_long do_bsd_shmdt(abi_ulong shmaddr) |
| 406 | { |
| 407 | abi_long ret; |
| 408 | |
| 409 | WITH_MMAP_LOCK_GUARD() { |
| 410 | int i; |
| 411 | |
| 412 | for (i = 0; i < N_BSD_SHM_REGIONS; ++i) { |
| 413 | if (bsd_shm_regions[i].start == shmaddr) { |
| 414 | break; |
| 415 | } |
| 416 | } |
| 417 | |
| 418 | if (i == N_BSD_SHM_REGIONS) { |
| 419 | return -TARGET_EINVAL; |
| 420 | } |
| 421 | |
| 422 | ret = get_errno(shmdt(g2h_untagged(shmaddr))); |
| 423 | if (ret == 0) { |
| 424 | abi_ulong size = bsd_shm_regions[i].size; |
| 425 | |
| 426 | bsd_shm_regions[i].start = 0; |
| 427 | page_set_flags(shmaddr, shmaddr + size - 1, 0); |
| 428 | mmap_reserve(shmaddr, size); |
| 429 | } |
| 430 | } |
| 431 | |
| 432 | return ret; |
| 433 | } |
| 434 | |
Warner Losh | dfa1d91 | 2023-09-25 21:27:09 +0300 | [diff] [blame] | 435 | static inline abi_long do_bsd_vadvise(void) |
| 436 | { |
| 437 | /* See sys_ovadvise() in vm_unix.c */ |
| 438 | return -TARGET_EINVAL; |
| 439 | } |
| 440 | |
| 441 | static inline abi_long do_bsd_sbrk(void) |
| 442 | { |
| 443 | /* see sys_sbrk() in vm_mmap.c */ |
| 444 | return -TARGET_EOPNOTSUPP; |
| 445 | } |
| 446 | |
| 447 | static inline abi_long do_bsd_sstk(void) |
| 448 | { |
| 449 | /* see sys_sstk() in vm_mmap.c */ |
| 450 | return -TARGET_EOPNOTSUPP; |
| 451 | } |
| 452 | |
Stacey Son | 6765e98 | 2023-09-25 21:26:57 +0300 | [diff] [blame] | 453 | #endif /* BSD_USER_BSD_MEM_H */ |