1 /* $NetBSD: pmap_segtab.c,v 1.14 2020/02/24 12:20:30 rin Exp $ */ 2 3 /*- 4 * Copyright (c) 1998, 2001 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center and by Chris G. Demetriou. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1992, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * This code is derived from software contributed to Berkeley by 38 * the Systems Programming Group of the University of Utah Computer 39 * Science Department and Ralph Campbell. 40 * 41 * Redistribution and use in source and binary forms, with or without 42 * modification, are permitted provided that the following conditions 43 * are met: 44 * 1. Redistributions of source code must retain the above copyright 45 * notice, this list of conditions and the following disclaimer. 46 * 2. Redistributions in binary form must reproduce the above copyright 47 * notice, this list of conditions and the following disclaimer in the 48 * documentation and/or other materials provided with the distribution. 49 * 3. Neither the name of the University nor the names of its contributors 50 * may be used to endorse or promote products derived from this software 51 * without specific prior written permission. 52 * 53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 63 * SUCH DAMAGE. 64 * 65 * @(#)pmap.c 8.4 (Berkeley) 1/26/94 66 */ 67 68 #include <sys/cdefs.h> 69 70 __KERNEL_RCSID(0, "$NetBSD: pmap_segtab.c,v 1.14 2020/02/24 12:20:30 rin Exp $"); 71 72 /* 73 * Manages physical address maps. 74 * 75 * In addition to hardware address maps, this 76 * module is called upon to provide software-use-only 77 * maps which may or may not be stored in the same 78 * form as hardware maps. These pseudo-maps are 79 * used to store intermediate results from copy 80 * operations to and from address spaces. 81 * 82 * Since the information managed by this module is 83 * also stored by the logical address mapping module, 84 * this module may throw away valid virtual-to-physical 85 * mappings at almost any time. However, invalidations 86 * of virtual-to-physical mappings must be done as 87 * requested. 88 * 89 * In order to cope with hardware architectures which 90 * make virtual-to-physical map invalidates expensive, 91 * this module may delay invalidate or reduced protection 92 * operations until such time as they are actually 93 * necessary. This module is given full information as 94 * to which processors are currently using which maps, 95 * and to when physical maps must be made correct. 96 */ 97 98 #define __PMAP_PRIVATE 99 100 #include "opt_multiprocessor.h" 101 102 #include <sys/param.h> 103 104 #include <sys/atomic.h> 105 #include <sys/mutex.h> 106 #include <sys/proc.h> 107 #include <sys/systm.h> 108 109 #include <uvm/uvm.h> 110 111 CTASSERT(NBPG >= sizeof(pmap_segtab_t)); 112 113 struct pmap_segtab_info { 114 pmap_segtab_t *free_segtab; /* free list kept locally */ 115 #ifdef DEBUG 116 uint32_t nget_segtab; 117 uint32_t nput_segtab; 118 uint32_t npage_segtab; 119 #define SEGTAB_ADD(n, v) (pmap_segtab_info.n ## _segtab += (v)) 120 #else 121 #define SEGTAB_ADD(n, v) ((void) 0) 122 #endif 123 #ifdef PMAP_PTP_CACHE 124 struct pgflist ptp_pgflist; /* Keep a list of idle page tables. */ 125 #endif 126 } pmap_segtab_info = { 127 #ifdef PMAP_PTP_CACHE 128 .ptp_pgflist = LIST_HEAD_INITIALIZER(pmap_segtab_info.ptp_pgflist), 129 #endif 130 }; 131 132 kmutex_t pmap_segtab_lock __cacheline_aligned; 133 134 static void 135 pmap_check_stp(pmap_segtab_t *stp, const char *caller, const char *why) 136 { 137 #ifdef DEBUG 138 for (size_t i = 0; i < PMAP_SEGTABSIZE; i++) { 139 if (stp->seg_tab[i] != 0) { 140 #ifdef DEBUG_NOISY 141 for (size_t j = i; j < PMAP_SEGTABSIZE; j++) 142 printf("%s: pm_segtab.seg_tab[%zu] = %p\n", 143 caller, j, stp->seg_tab[j]); 144 #endif 145 panic("%s: pm_segtab.seg_tab[%zu] != 0 (%p): %s", 146 caller, i, stp->seg_tab[i], why); 147 } 148 } 149 #endif 150 } 151 152 static inline struct vm_page * 153 pmap_pte_pagealloc(void) 154 { 155 struct vm_page *pg; 156 157 pg = PMAP_ALLOC_POOLPAGE(UVM_PGA_ZERO|UVM_PGA_USERESERVE); 158 if (pg) { 159 #ifdef UVM_PAGE_TRKOWN 160 pg->owner_tag = NULL; 161 #endif 162 UVM_PAGE_OWN(pg, "pmap-ptp"); 163 } 164 165 return pg; 166 } 167 168 static inline pt_entry_t * 169 pmap_segmap(struct pmap *pmap, vaddr_t va) 170 { 171 pmap_segtab_t *stp = pmap->pm_segtab; 172 KASSERTMSG(pmap != pmap_kernel() || !pmap_md_direct_mapped_vaddr_p(va), 173 "pmap %p va %#" PRIxVADDR, pmap, va); 174 #ifdef _LP64 175 stp = stp->seg_seg[(va >> XSEGSHIFT) & (NSEGPG - 1)]; 176 if (stp == NULL) 177 return NULL; 178 #endif 179 180 return stp->seg_tab[(va >> SEGSHIFT) & (PMAP_SEGTABSIZE - 1)]; 181 } 182 183 pt_entry_t * 184 pmap_pte_lookup(pmap_t pmap, vaddr_t va) 185 { 186 pt_entry_t *pte = pmap_segmap(pmap, va); 187 if (pte == NULL) 188 return NULL; 189 190 return pte + ((va >> PGSHIFT) & (NPTEPG - 1)); 191 } 192 193 static void 194 pmap_segtab_free(pmap_segtab_t *stp) 195 { 196 /* 197 * Insert the segtab into the segtab freelist. 198 */ 199 mutex_spin_enter(&pmap_segtab_lock); 200 stp->seg_seg[0] = pmap_segtab_info.free_segtab; 201 pmap_segtab_info.free_segtab = stp; 202 SEGTAB_ADD(nput, 1); 203 mutex_spin_exit(&pmap_segtab_lock); 204 } 205 206 static void 207 pmap_segtab_release(pmap_t pmap, pmap_segtab_t **stp_p, bool free_stp, 208 pte_callback_t callback, uintptr_t flags, 209 vaddr_t va, vsize_t vinc) 210 { 211 pmap_segtab_t *stp = *stp_p; 212 213 for (size_t i = (va / vinc) & (PMAP_SEGTABSIZE - 1); 214 i < PMAP_SEGTABSIZE; 215 i++, va += vinc) { 216 #ifdef _LP64 217 if (vinc > NBSEG) { 218 if (stp->seg_seg[i] != NULL) { 219 pmap_segtab_release(pmap, &stp->seg_seg[i], 220 true, callback, flags, va, vinc / NSEGPG); 221 KASSERT(stp->seg_seg[i] == NULL); 222 } 223 continue; 224 } 225 #endif 226 KASSERT(vinc == NBSEG); 227 228 /* get pointer to segment map */ 229 pt_entry_t *pte = stp->seg_tab[i]; 230 if (pte == NULL) 231 continue; 232 233 /* 234 * If our caller want a callback, do so. 235 */ 236 if (callback != NULL) { 237 (*callback)(pmap, va, va + vinc, pte, flags); 238 } 239 #ifdef DEBUG 240 for (size_t j = 0; j < NPTEPG; j++) { 241 if (!pte_zero_p(pte[j])) 242 panic("%s: pte entry %p not 0 (%#"PRIxPTE")", 243 __func__, &pte[j], pte_value(pte[j])); 244 } 245 #endif 246 // PMAP_UNMAP_POOLPAGE should handle any VCA issues itself 247 paddr_t pa = PMAP_UNMAP_POOLPAGE((vaddr_t)pte); 248 struct vm_page *pg = PHYS_TO_VM_PAGE(pa); 249 #ifdef PMAP_PTP_CACHE 250 mutex_spin_enter(&pmap_segtab_lock); 251 LIST_INSERT_HEAD(&pmap_segtab_info.ptp_pgflist, pg, pageq.list); 252 mutex_spin_exit(&pmap_segtab_lock); 253 #else 254 uvm_pagefree(pg); 255 #endif 256 257 stp->seg_tab[i] = NULL; 258 } 259 260 if (free_stp) { 261 pmap_check_stp(stp, __func__, 262 vinc == NBSEG ? "release seg" : "release xseg"); 263 pmap_segtab_free(stp); 264 *stp_p = NULL; 265 } 266 } 267 268 /* 269 * Create and return a physical map. 270 * 271 * If the size specified for the map 272 * is zero, the map is an actual physical 273 * map, and may be referenced by the 274 * hardware. 275 * 276 * If the size specified is non-zero, 277 * the map will be used in software only, and 278 * is bounded by that size. 279 */ 280 static pmap_segtab_t * 281 pmap_segtab_alloc(void) 282 { 283 pmap_segtab_t *stp; 284 bool found_on_freelist = false; 285 286 again: 287 mutex_spin_enter(&pmap_segtab_lock); 288 if (__predict_true((stp = pmap_segtab_info.free_segtab) != NULL)) { 289 pmap_segtab_info.free_segtab = stp->seg_seg[0]; 290 stp->seg_seg[0] = NULL; 291 SEGTAB_ADD(nget, 1); 292 found_on_freelist = true; 293 } 294 mutex_spin_exit(&pmap_segtab_lock); 295 296 if (__predict_false(stp == NULL)) { 297 struct vm_page * const stp_pg = pmap_pte_pagealloc(); 298 299 if (__predict_false(stp_pg == NULL)) { 300 /* 301 * XXX What else can we do? Could we deadlock here? 302 */ 303 uvm_wait("segtab"); 304 goto again; 305 } 306 SEGTAB_ADD(npage, 1); 307 const paddr_t stp_pa = VM_PAGE_TO_PHYS(stp_pg); 308 309 stp = (pmap_segtab_t *)PMAP_MAP_POOLPAGE(stp_pa); 310 const size_t n = NBPG / sizeof(*stp); 311 if (n > 1) { 312 /* 313 * link all the segtabs in this page together 314 */ 315 for (size_t i = 1; i < n - 1; i++) { 316 stp[i].seg_seg[0] = &stp[i+1]; 317 } 318 /* 319 * Now link the new segtabs into the free segtab list. 320 */ 321 mutex_spin_enter(&pmap_segtab_lock); 322 stp[n-1].seg_seg[0] = pmap_segtab_info.free_segtab; 323 pmap_segtab_info.free_segtab = stp + 1; 324 SEGTAB_ADD(nput, n - 1); 325 mutex_spin_exit(&pmap_segtab_lock); 326 } 327 } 328 329 pmap_check_stp(stp, __func__, 330 found_on_freelist ? "from free list" : "allocated"); 331 332 return stp; 333 } 334 335 /* 336 * Allocate the top segment table for the pmap. 337 */ 338 void 339 pmap_segtab_init(pmap_t pmap) 340 { 341 342 pmap->pm_segtab = pmap_segtab_alloc(); 343 } 344 345 /* 346 * Retire the given physical map from service. 347 * Should only be called if the map contains 348 * no valid mappings. 349 */ 350 void 351 pmap_segtab_destroy(pmap_t pmap, pte_callback_t func, uintptr_t flags) 352 { 353 if (pmap->pm_segtab == NULL) 354 return; 355 356 #ifdef _LP64 357 const vsize_t vinc = NBXSEG; 358 #else 359 const vsize_t vinc = NBSEG; 360 #endif 361 pmap_segtab_release(pmap, &pmap->pm_segtab, 362 func == NULL, func, flags, pmap->pm_minaddr, vinc); 363 } 364 365 /* 366 * Make a new pmap (vmspace) active for the given process. 367 */ 368 void 369 pmap_segtab_activate(struct pmap *pm, struct lwp *l) 370 { 371 if (l == curlwp) { 372 struct cpu_info * const ci = l->l_cpu; 373 KASSERT(pm == l->l_proc->p_vmspace->vm_map.pmap); 374 if (pm == pmap_kernel()) { 375 ci->ci_pmap_user_segtab = PMAP_INVALID_SEGTAB_ADDRESS; 376 #ifdef _LP64 377 ci->ci_pmap_user_seg0tab = PMAP_INVALID_SEGTAB_ADDRESS; 378 #endif 379 } else { 380 ci->ci_pmap_user_segtab = pm->pm_segtab; 381 #ifdef _LP64 382 ci->ci_pmap_user_seg0tab = pm->pm_segtab->seg_seg[0]; 383 #endif 384 } 385 } 386 } 387 388 /* 389 * Act on the given range of addresses from the specified map. 390 * 391 * It is assumed that the start and end are properly rounded to 392 * the page size. 393 */ 394 void 395 pmap_pte_process(pmap_t pmap, vaddr_t sva, vaddr_t eva, 396 pte_callback_t callback, uintptr_t flags) 397 { 398 #if 0 399 printf("%s: %p, %"PRIxVADDR", %"PRIxVADDR", %p, %"PRIxPTR"\n", 400 __func__, pmap, sva, eva, callback, flags); 401 #endif 402 while (sva < eva) { 403 vaddr_t lastseg_va = pmap_trunc_seg(sva) + NBSEG; 404 if (lastseg_va == 0 || lastseg_va > eva) 405 lastseg_va = eva; 406 407 /* 408 * If VA belongs to an unallocated segment, 409 * skip to the next segment boundary. 410 */ 411 pt_entry_t * const ptep = pmap_pte_lookup(pmap, sva); 412 if (ptep != NULL) { 413 /* 414 * Callback to deal with the ptes for this segment. 415 */ 416 (*callback)(pmap, sva, lastseg_va, ptep, flags); 417 } 418 /* 419 * In theory we could release pages with no entries, 420 * but that takes more effort than we want here. 421 */ 422 sva = lastseg_va; 423 } 424 } 425 426 /* 427 * Return a pointer for the pte that corresponds to the specified virtual 428 * address (va) in the target physical map, allocating if needed. 429 */ 430 pt_entry_t * 431 pmap_pte_reserve(pmap_t pmap, vaddr_t va, int flags) 432 { 433 pmap_segtab_t *stp = pmap->pm_segtab; 434 pt_entry_t *pte; 435 436 pte = pmap_pte_lookup(pmap, va); 437 if (__predict_false(pte == NULL)) { 438 #ifdef _LP64 439 pmap_segtab_t ** const stp_p = 440 &stp->seg_seg[(va >> XSEGSHIFT) & (NSEGPG - 1)]; 441 if (__predict_false((stp = *stp_p) == NULL)) { 442 pmap_segtab_t *nstp = pmap_segtab_alloc(); 443 #ifdef MULTIPROCESSOR 444 pmap_segtab_t *ostp = atomic_cas_ptr(stp_p, NULL, nstp); 445 if (__predict_false(ostp != NULL)) { 446 pmap_check_stp(nstp, __func__, "reserve"); 447 pmap_segtab_free(nstp); 448 nstp = ostp; 449 } 450 #else 451 *stp_p = nstp; 452 #endif /* MULTIPROCESSOR */ 453 stp = nstp; 454 } 455 KASSERT(stp == pmap->pm_segtab->seg_seg[(va >> XSEGSHIFT) & (NSEGPG - 1)]); 456 #endif /* _LP64 */ 457 struct vm_page *pg = NULL; 458 #ifdef PMAP_PTP_CACHE 459 mutex_spin_enter(&pmap_segtab_lock); 460 if ((pg = LIST_FIRST(&pmap_segtab_info.ptp_pgflist)) != NULL) { 461 LIST_REMOVE(pg, pageq.list); 462 KASSERT(LIST_FIRST(&pmap_segtab_info.ptp_pgflist) != pg); 463 } 464 mutex_spin_exit(&pmap_segtab_lock); 465 #endif 466 if (pg == NULL) 467 pg = pmap_pte_pagealloc(); 468 if (pg == NULL) { 469 if (flags & PMAP_CANFAIL) 470 return NULL; 471 panic("%s: cannot allocate page table page " 472 "for va %" PRIxVADDR, __func__, va); 473 } 474 475 const paddr_t pa = VM_PAGE_TO_PHYS(pg); 476 pte = (pt_entry_t *)PMAP_MAP_POOLPAGE(pa); 477 pt_entry_t ** const pte_p = 478 &stp->seg_tab[(va >> SEGSHIFT) & (PMAP_SEGTABSIZE - 1)]; 479 #ifdef MULTIPROCESSOR 480 pt_entry_t *opte = atomic_cas_ptr(pte_p, NULL, pte); 481 /* 482 * If another thread allocated the segtab needed for this va 483 * free the page we just allocated. 484 */ 485 if (__predict_false(opte != NULL)) { 486 #ifdef PMAP_PTP_CACHE 487 mutex_spin_enter(&pmap_segtab_lock); 488 LIST_INSERT_HEAD(&pmap_segtab_info.ptp_pgflist, 489 pg, pageq.list); 490 mutex_spin_exit(&pmap_segtab_lock); 491 #else 492 PMAP_UNMAP_POOLPAGE((vaddr_t)pte); 493 uvm_pagefree(pg); 494 #endif 495 pte = opte; 496 } 497 #else 498 *pte_p = pte; 499 #endif 500 KASSERT(pte == stp->seg_tab[(va >> SEGSHIFT) & (PMAP_SEGTABSIZE - 1)]); 501 502 #ifdef DEBUG 503 for (size_t i = 0; i < NPTEPG; i++) { 504 if (!pte_zero_p(pte[i])) 505 panic("%s: new segmap %p not empty @ %zu", 506 __func__, pte, i); 507 } 508 #endif 509 pte += (va >> PGSHIFT) & (NPTEPG - 1); 510 } 511 512 return pte; 513 } 514