pmap.c revision 1.17
1 /* $NetBSD: pmap.c,v 1.17 2003/11/21 18:09:27 matt Exp $ */ 2 /*- 3 * Copyright (c) 2001 The NetBSD Foundation, Inc. 4 * All rights reserved. 5 * 6 * This code is derived from software contributed to The NetBSD Foundation 7 * by Matt Thomas <matt (at) 3am-software.com> of Allegro Networks, Inc. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. All advertising materials mentioning features or use of this software 18 * must display the following acknowledgement: 19 * This product includes software developed by the NetBSD 20 * Foundation, Inc. and its contributors. 21 * 4. Neither the name of The NetBSD Foundation nor the names of its 22 * contributors may be used to endorse or promote products derived 23 * from this software without specific prior written permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 26 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 27 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 28 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 29 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 30 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 31 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 32 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 33 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 34 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 35 * POSSIBILITY OF SUCH DAMAGE. 36 */ 37 38 /* 39 * Copyright (C) 1995, 1996 Wolfgang Solfrank. 40 * Copyright (C) 1995, 1996 TooLs GmbH. 41 * All rights reserved. 42 * 43 * Redistribution and use in source and binary forms, with or without 44 * modification, are permitted provided that the following conditions 45 * are met: 46 * 1. Redistributions of source code must retain the above copyright 47 * notice, this list of conditions and the following disclaimer. 48 * 2. Redistributions in binary form must reproduce the above copyright 49 * notice, this list of conditions and the following disclaimer in the 50 * documentation and/or other materials provided with the distribution. 51 * 3. All advertising materials mentioning features or use of this software 52 * must display the following acknowledgement: 53 * This product includes software developed by TooLs GmbH. 54 * 4. The name of TooLs GmbH may not be used to endorse or promote products 55 * derived from this software without specific prior written permission. 56 * 57 * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR 58 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 59 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 60 * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 61 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 62 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; 63 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, 64 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR 65 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF 66 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.17 2003/11/21 18:09:27 matt Exp $"); 71 72 #include "opt_altivec.h" 73 #include "opt_pmap.h" 74 #include <sys/param.h> 75 #include <sys/malloc.h> 76 #include <sys/proc.h> 77 #include <sys/user.h> 78 #include <sys/pool.h> 79 #include <sys/queue.h> 80 #include <sys/device.h> /* for evcnt */ 81 #include <sys/systm.h> 82 83 #if __NetBSD_Version__ < 105010000 84 #include <vm/vm.h> 85 #include <vm/vm_kern.h> 86 #define splvm() splimp() 87 #endif 88 89 #include <uvm/uvm.h> 90 91 #include <machine/pcb.h> 92 #include <machine/powerpc.h> 93 #include <powerpc/spr.h> 94 #include <powerpc/oea/sr_601.h> 95 #if __NetBSD_Version__ > 105010000 96 #include <powerpc/oea/bat.h> 97 #else 98 #include <powerpc/bat.h> 99 #endif 100 101 #if defined(DEBUG) || defined(PMAPCHECK) 102 #define STATIC 103 #else 104 #define STATIC static 105 #endif 106 107 #ifdef ALTIVEC 108 int pmap_use_altivec; 109 #endif 110 111 volatile struct pteg *pmap_pteg_table; 112 unsigned int pmap_pteg_cnt; 113 unsigned int pmap_pteg_mask; 114 paddr_t pmap_memlimit = -PAGE_SIZE; /* there is no limit */ 115 116 struct pmap kernel_pmap_; 117 unsigned int pmap_pages_stolen; 118 u_long pmap_pte_valid; 119 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 120 u_long pmap_pvo_enter_depth; 121 u_long pmap_pvo_remove_depth; 122 #endif 123 124 int physmem; 125 #ifndef MSGBUFADDR 126 extern paddr_t msgbuf_paddr; 127 #endif 128 129 static struct mem_region *mem, *avail; 130 static u_int mem_cnt, avail_cnt; 131 132 #ifdef __HAVE_PMAP_PHYSSEG 133 /* 134 * This is a cache of referenced/modified bits. 135 * Bits herein are shifted by ATTRSHFT. 136 */ 137 #define ATTR_SHFT 4 138 struct pmap_physseg pmap_physseg; 139 #endif 140 141 /* 142 * The following structure is exactly 32 bytes long (one cacheline). 143 */ 144 struct pvo_entry { 145 LIST_ENTRY(pvo_entry) pvo_vlink; /* Link to common virt page */ 146 TAILQ_ENTRY(pvo_entry) pvo_olink; /* Link to overflow entry */ 147 struct pte pvo_pte; /* Prebuilt PTE */ 148 pmap_t pvo_pmap; /* ptr to owning pmap */ 149 vaddr_t pvo_vaddr; /* VA of entry */ 150 #define PVO_PTEGIDX_MASK 0x0007 /* which PTEG slot */ 151 #define PVO_PTEGIDX_VALID 0x0008 /* slot is valid */ 152 #define PVO_WIRED 0x0010 /* PVO entry is wired */ 153 #define PVO_MANAGED 0x0020 /* PVO e. for managed page */ 154 #define PVO_EXECUTABLE 0x0040 /* PVO e. for executable page */ 155 #define PVO_ENTER_INSERT 0 /* PVO has been removed */ 156 #define PVO_SPILL_UNSET 1 /* PVO has been evicted */ 157 #define PVO_SPILL_SET 2 /* PVO has been spilled */ 158 #define PVO_SPILL_INSERT 3 /* PVO has been inserted */ 159 #define PVO_PMAP_PAGE_PROTECT 4 /* PVO has changed */ 160 #define PVO_PMAP_PROTECT 5 /* PVO has changed */ 161 #define PVO_REMOVE 6 /* PVO has been removed */ 162 #define PVO_WHERE_MASK 15 163 #define PVO_WHERE_SHFT 8 164 }; 165 #define PVO_VADDR(pvo) ((pvo)->pvo_vaddr & ~ADDR_POFF) 166 #define PVO_ISEXECUTABLE(pvo) ((pvo)->pvo_vaddr & PVO_EXECUTABLE) 167 #define PVO_PTEGIDX_GET(pvo) ((pvo)->pvo_vaddr & PVO_PTEGIDX_MASK) 168 #define PVO_PTEGIDX_ISSET(pvo) ((pvo)->pvo_vaddr & PVO_PTEGIDX_VALID) 169 #define PVO_PTEGIDX_CLR(pvo) \ 170 ((void)((pvo)->pvo_vaddr &= ~(PVO_PTEGIDX_VALID|PVO_PTEGIDX_MASK))) 171 #define PVO_PTEGIDX_SET(pvo,i) \ 172 ((void)((pvo)->pvo_vaddr |= (i)|PVO_PTEGIDX_VALID)) 173 #define PVO_WHERE(pvo,w) \ 174 ((pvo)->pvo_vaddr &= ~(PVO_WHERE_MASK << PVO_WHERE_SHFT), \ 175 (pvo)->pvo_vaddr |= ((PVO_ ## w) << PVO_WHERE_SHFT)) 176 177 TAILQ_HEAD(pvo_tqhead, pvo_entry); 178 struct pvo_tqhead *pmap_pvo_table; /* pvo entries by ptegroup index */ 179 struct pvo_head pmap_pvo_kunmanaged = LIST_HEAD_INITIALIZER(pmap_pvo_kunmanaged); /* list of unmanaged pages */ 180 struct pvo_head pmap_pvo_unmanaged = LIST_HEAD_INITIALIZER(pmap_pvo_unmanaged); /* list of unmanaged pages */ 181 182 struct pool pmap_pool; /* pool for pmap structures */ 183 struct pool pmap_upvo_pool; /* pool for pvo entries for unmanaged pages */ 184 struct pool pmap_mpvo_pool; /* pool for pvo entries for managed pages */ 185 186 /* 187 * We keep a cache of unmanaged pages to be used for pvo entries for 188 * unmanaged pages. 189 */ 190 struct pvo_page { 191 SIMPLEQ_ENTRY(pvo_page) pvop_link; 192 }; 193 SIMPLEQ_HEAD(pvop_head, pvo_page); 194 struct pvop_head pmap_upvop_head = SIMPLEQ_HEAD_INITIALIZER(pmap_upvop_head); 195 struct pvop_head pmap_mpvop_head = SIMPLEQ_HEAD_INITIALIZER(pmap_mpvop_head); 196 u_long pmap_upvop_free; 197 u_long pmap_upvop_maxfree; 198 u_long pmap_mpvop_free; 199 u_long pmap_mpvop_maxfree; 200 201 STATIC void *pmap_pool_ualloc(struct pool *, int); 202 STATIC void *pmap_pool_malloc(struct pool *, int); 203 204 STATIC void pmap_pool_ufree(struct pool *, void *); 205 STATIC void pmap_pool_mfree(struct pool *, void *); 206 207 static struct pool_allocator pmap_pool_mallocator = { 208 pmap_pool_malloc, pmap_pool_mfree, 0, 209 }; 210 211 static struct pool_allocator pmap_pool_uallocator = { 212 pmap_pool_ualloc, pmap_pool_ufree, 0, 213 }; 214 215 #if defined(DEBUG) || defined(PMAPCHECK) || defined(DDB) 216 void pmap_pte_print(volatile struct pte *); 217 #endif 218 219 #ifdef DDB 220 void pmap_pteg_check(void); 221 void pmap_pteg_dist(void); 222 void pmap_print_pte(pmap_t, vaddr_t); 223 void pmap_print_mmuregs(void); 224 #endif 225 226 #if defined(DEBUG) || defined(PMAPCHECK) 227 #ifdef PMAPCHECK 228 int pmapcheck = 1; 229 #else 230 int pmapcheck = 0; 231 #endif 232 void pmap_pvo_verify(void); 233 STATIC void pmap_pvo_check(const struct pvo_entry *); 234 #define PMAP_PVO_CHECK(pvo) \ 235 do { \ 236 if (pmapcheck) \ 237 pmap_pvo_check(pvo); \ 238 } while (0) 239 #else 240 #define PMAP_PVO_CHECK(pvo) do { } while (/*CONSTCOND*/0) 241 #endif 242 STATIC int pmap_pte_insert(int, struct pte *); 243 STATIC int pmap_pvo_enter(pmap_t, struct pool *, struct pvo_head *, 244 vaddr_t, paddr_t, register_t, int); 245 STATIC void pmap_pvo_remove(struct pvo_entry *, int); 246 STATIC struct pvo_entry *pmap_pvo_find_va(pmap_t, vaddr_t, int *); 247 STATIC volatile struct pte *pmap_pvo_to_pte(const struct pvo_entry *, int); 248 #define pmap_pvo_reclaim(pm) NULL 249 STATIC void pvo_set_exec(struct pvo_entry *); 250 STATIC void pvo_clear_exec(struct pvo_entry *); 251 252 STATIC void tlbia(void); 253 254 STATIC void pmap_release(pmap_t); 255 STATIC void *pmap_boot_find_memory(psize_t, psize_t, int); 256 257 #define VSID_NBPW (sizeof(uint32_t) * 8) 258 static uint32_t pmap_vsid_bitmap[NPMAPS / VSID_NBPW]; 259 260 static int pmap_initialized; 261 262 #if defined(DEBUG) || defined(PMAPDEBUG) 263 #define PMAPDEBUG_BOOT 0x0001 264 #define PMAPDEBUG_PTE 0x0002 265 #define PMAPDEBUG_EXEC 0x0008 266 #define PMAPDEBUG_PVOENTER 0x0010 267 #define PMAPDEBUG_PVOREMOVE 0x0020 268 #define PMAPDEBUG_ACTIVATE 0x0100 269 #define PMAPDEBUG_CREATE 0x0200 270 #define PMAPDEBUG_ENTER 0x1000 271 #define PMAPDEBUG_KENTER 0x2000 272 #define PMAPDEBUG_KREMOVE 0x4000 273 #define PMAPDEBUG_REMOVE 0x8000 274 unsigned int pmapdebug = 0; 275 # define DPRINTF(x) printf x 276 # define DPRINTFN(n, x) if (pmapdebug & PMAPDEBUG_ ## n) printf x 277 #else 278 # define DPRINTF(x) 279 # define DPRINTFN(n, x) 280 #endif 281 282 283 #ifdef PMAPCOUNTERS 284 #define PMAPCOUNT(ev) ((pmap_evcnt_ ## ev).ev_count++) 285 #define PMAPCOUNT2(ev) ((ev).ev_count++) 286 287 struct evcnt pmap_evcnt_mappings = 288 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 289 "pmap", "pages mapped"); 290 struct evcnt pmap_evcnt_unmappings = 291 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_mappings, 292 "pmap", "pages unmapped"); 293 294 struct evcnt pmap_evcnt_kernel_mappings = 295 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 296 "pmap", "kernel pages mapped"); 297 struct evcnt pmap_evcnt_kernel_unmappings = 298 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_kernel_mappings, 299 "pmap", "kernel pages unmapped"); 300 301 struct evcnt pmap_evcnt_mappings_replaced = 302 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 303 "pmap", "page mappings replaced"); 304 305 struct evcnt pmap_evcnt_exec_mappings = 306 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_mappings, 307 "pmap", "exec pages mapped"); 308 struct evcnt pmap_evcnt_exec_cached = 309 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_mappings, 310 "pmap", "exec pages cached"); 311 312 struct evcnt pmap_evcnt_exec_synced = 313 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 314 "pmap", "exec pages synced"); 315 struct evcnt pmap_evcnt_exec_synced_clear_modify = 316 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 317 "pmap", "exec pages synced (CM)"); 318 319 struct evcnt pmap_evcnt_exec_uncached_page_protect = 320 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 321 "pmap", "exec pages uncached (PP)"); 322 struct evcnt pmap_evcnt_exec_uncached_clear_modify = 323 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 324 "pmap", "exec pages uncached (CM)"); 325 struct evcnt pmap_evcnt_exec_uncached_zero_page = 326 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 327 "pmap", "exec pages uncached (ZP)"); 328 struct evcnt pmap_evcnt_exec_uncached_copy_page = 329 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, &pmap_evcnt_exec_mappings, 330 "pmap", "exec pages uncached (CP)"); 331 332 struct evcnt pmap_evcnt_updates = 333 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 334 "pmap", "updates"); 335 struct evcnt pmap_evcnt_collects = 336 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 337 "pmap", "collects"); 338 struct evcnt pmap_evcnt_copies = 339 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 340 "pmap", "copies"); 341 342 struct evcnt pmap_evcnt_ptes_spilled = 343 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 344 "pmap", "ptes spilled from overflow"); 345 struct evcnt pmap_evcnt_ptes_unspilled = 346 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 347 "pmap", "ptes not spilled"); 348 struct evcnt pmap_evcnt_ptes_evicted = 349 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 350 "pmap", "ptes evicted"); 351 352 struct evcnt pmap_evcnt_ptes_primary[8] = { 353 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 354 "pmap", "ptes added at primary[0]"), 355 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 356 "pmap", "ptes added at primary[1]"), 357 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 358 "pmap", "ptes added at primary[2]"), 359 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 360 "pmap", "ptes added at primary[3]"), 361 362 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 363 "pmap", "ptes added at primary[4]"), 364 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 365 "pmap", "ptes added at primary[5]"), 366 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 367 "pmap", "ptes added at primary[6]"), 368 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 369 "pmap", "ptes added at primary[7]"), 370 }; 371 struct evcnt pmap_evcnt_ptes_secondary[8] = { 372 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 373 "pmap", "ptes added at secondary[0]"), 374 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 375 "pmap", "ptes added at secondary[1]"), 376 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 377 "pmap", "ptes added at secondary[2]"), 378 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 379 "pmap", "ptes added at secondary[3]"), 380 381 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 382 "pmap", "ptes added at secondary[4]"), 383 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 384 "pmap", "ptes added at secondary[5]"), 385 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 386 "pmap", "ptes added at secondary[6]"), 387 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 388 "pmap", "ptes added at secondary[7]"), 389 }; 390 struct evcnt pmap_evcnt_ptes_removed = 391 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 392 "pmap", "ptes removed"); 393 struct evcnt pmap_evcnt_ptes_changed = 394 EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, 395 "pmap", "ptes changed"); 396 397 /* 398 * From pmap_subr.c 399 */ 400 extern struct evcnt pmap_evcnt_zeroed_pages; 401 extern struct evcnt pmap_evcnt_copied_pages; 402 extern struct evcnt pmap_evcnt_idlezeroed_pages; 403 #else 404 #define PMAPCOUNT(ev) ((void) 0) 405 #define PMAPCOUNT2(ev) ((void) 0) 406 #endif 407 408 #define TLBIE(va) __asm __volatile("tlbie %0" :: "r"(va)) 409 #define TLBSYNC() __asm __volatile("tlbsync") 410 #define SYNC() __asm __volatile("sync") 411 #define EIEIO() __asm __volatile("eieio") 412 #define MFMSR() mfmsr() 413 #define MTMSR(psl) mtmsr(psl) 414 #define MFPVR() mfpvr() 415 #define MFSRIN(va) mfsrin(va) 416 #define MFTB() mfrtcltbl() 417 418 static __inline register_t 419 mfsrin(vaddr_t va) 420 { 421 register_t sr; 422 __asm __volatile ("mfsrin %0,%1" : "=r"(sr) : "r"(va)); 423 return sr; 424 } 425 426 static __inline register_t 427 pmap_interrupts_off(void) 428 { 429 register_t msr = MFMSR(); 430 if (msr & PSL_EE) 431 MTMSR(msr & ~PSL_EE); 432 return msr; 433 } 434 435 static void 436 pmap_interrupts_restore(register_t msr) 437 { 438 if (msr & PSL_EE) 439 MTMSR(msr); 440 } 441 442 static __inline u_int32_t 443 mfrtcltbl(void) 444 { 445 446 if ((MFPVR() >> 16) == MPC601) 447 return (mfrtcl() >> 7); 448 else 449 return (mftbl()); 450 } 451 452 /* 453 * These small routines may have to be replaced, 454 * if/when we support processors other that the 604. 455 */ 456 457 void 458 tlbia(void) 459 { 460 caddr_t i; 461 462 SYNC(); 463 /* 464 * Why not use "tlbia"? Because not all processors implement it. 465 * 466 * This needs to be a per-cpu callback to do the appropriate thing 467 * for the CPU. XXX 468 */ 469 for (i = 0; i < (caddr_t)0x00040000; i += 0x00001000) { 470 TLBIE(i); 471 EIEIO(); 472 SYNC(); 473 } 474 TLBSYNC(); 475 SYNC(); 476 } 477 478 static __inline register_t 479 va_to_vsid(const struct pmap *pm, vaddr_t addr) 480 { 481 return (pm->pm_sr[addr >> ADDR_SR_SHFT] & SR_VSID); 482 } 483 484 static __inline register_t 485 va_to_pteg(const struct pmap *pm, vaddr_t addr) 486 { 487 register_t hash; 488 489 hash = va_to_vsid(pm, addr) ^ ((addr & ADDR_PIDX) >> ADDR_PIDX_SHFT); 490 return hash & pmap_pteg_mask; 491 } 492 493 #if defined(DEBUG) || defined(PMAPCHECK) || defined(DDB) 494 /* 495 * Given a PTE in the page table, calculate the VADDR that hashes to it. 496 * The only bit of magic is that the top 4 bits of the address doesn't 497 * technically exist in the PTE. But we know we reserved 4 bits of the 498 * VSID for it so that's how we get it. 499 */ 500 static vaddr_t 501 pmap_pte_to_va(volatile const struct pte *pt) 502 { 503 vaddr_t va; 504 uintptr_t ptaddr = (uintptr_t) pt; 505 506 if (pt->pte_hi & PTE_HID) 507 ptaddr ^= (pmap_pteg_mask * sizeof(struct pteg)); 508 509 /* PPC Bits 10-19 */ 510 va = ((pt->pte_hi >> PTE_VSID_SHFT) ^ (ptaddr / sizeof(struct pteg))) & 0x3ff; 511 va <<= ADDR_PIDX_SHFT; 512 513 /* PPC Bits 4-9 */ 514 va |= (pt->pte_hi & PTE_API) << ADDR_API_SHFT; 515 516 /* PPC Bits 0-3 */ 517 va |= VSID_TO_SR(pt->pte_hi >> PTE_VSID_SHFT) << ADDR_SR_SHFT; 518 519 return va; 520 } 521 #endif 522 523 static __inline struct pvo_head * 524 pa_to_pvoh(paddr_t pa, struct vm_page **pg_p) 525 { 526 #ifdef __HAVE_VM_PAGE_MD 527 struct vm_page *pg; 528 529 pg = PHYS_TO_VM_PAGE(pa); 530 if (pg_p != NULL) 531 *pg_p = pg; 532 if (pg == NULL) 533 return &pmap_pvo_unmanaged; 534 return &pg->mdpage.mdpg_pvoh; 535 #endif 536 #ifdef __HAVE_PMAP_PHYSSEG 537 int bank, pg; 538 539 bank = vm_physseg_find(atop(pa), &pg); 540 if (pg_p != NULL) 541 *pg_p = pg; 542 if (bank == -1) 543 return &pmap_pvo_unmanaged; 544 return &vm_physmem[bank].pmseg.pvoh[pg]; 545 #endif 546 } 547 548 static __inline struct pvo_head * 549 vm_page_to_pvoh(struct vm_page *pg) 550 { 551 #ifdef __HAVE_VM_PAGE_MD 552 return &pg->mdpage.mdpg_pvoh; 553 #endif 554 #ifdef __HAVE_PMAP_PHYSSEG 555 return pa_to_pvoh(VM_PAGE_TO_PHYS(pg), NULL); 556 #endif 557 } 558 559 560 #ifdef __HAVE_PMAP_PHYSSEG 561 static __inline char * 562 pa_to_attr(paddr_t pa) 563 { 564 int bank, pg; 565 566 bank = vm_physseg_find(atop(pa), &pg); 567 if (bank == -1) 568 return NULL; 569 return &vm_physmem[bank].pmseg.attrs[pg]; 570 } 571 #endif 572 573 static __inline void 574 pmap_attr_clear(struct vm_page *pg, int ptebit) 575 { 576 #ifdef __HAVE_PMAP_PHYSSEG 577 *pa_to_attr(VM_PAGE_TO_PHYS(pg)) &= ~(ptebit >> ATTR_SHFT); 578 #endif 579 #ifdef __HAVE_VM_PAGE_MD 580 pg->mdpage.mdpg_attrs &= ~ptebit; 581 #endif 582 } 583 584 static __inline int 585 pmap_attr_fetch(struct vm_page *pg) 586 { 587 #ifdef __HAVE_PMAP_PHYSSEG 588 return *pa_to_attr(VM_PAGE_TO_PHYS(pg)) << ATTR_SHFT; 589 #endif 590 #ifdef __HAVE_VM_PAGE_MD 591 return pg->mdpage.mdpg_attrs; 592 #endif 593 } 594 595 static __inline void 596 pmap_attr_save(struct vm_page *pg, int ptebit) 597 { 598 #ifdef __HAVE_PMAP_PHYSSEG 599 *pa_to_attr(VM_PAGE_TO_PHYS(pg)) |= (ptebit >> ATTR_SHFT); 600 #endif 601 #ifdef __HAVE_VM_PAGE_MD 602 pg->mdpage.mdpg_attrs |= ptebit; 603 #endif 604 } 605 606 static __inline int 607 pmap_pte_compare(const volatile struct pte *pt, const struct pte *pvo_pt) 608 { 609 if (pt->pte_hi == pvo_pt->pte_hi 610 #if 0 611 && ((pt->pte_lo ^ pvo_pt->pte_lo) & 612 ~(PTE_REF|PTE_CHG)) == 0 613 #endif 614 ) 615 return 1; 616 return 0; 617 } 618 619 static __inline void 620 pmap_pte_create(struct pte *pt, const struct pmap *pm, vaddr_t va, register_t pte_lo) 621 { 622 /* 623 * Construct the PTE. Default to IMB initially. Valid bit 624 * only gets set when the real pte is set in memory. 625 * 626 * Note: Don't set the valid bit for correct operation of tlb update. 627 */ 628 pt->pte_hi = (va_to_vsid(pm, va) << PTE_VSID_SHFT) 629 | (((va & ADDR_PIDX) >> (ADDR_API_SHFT - PTE_API_SHFT)) & PTE_API); 630 pt->pte_lo = pte_lo; 631 } 632 633 static __inline void 634 pmap_pte_synch(volatile struct pte *pt, struct pte *pvo_pt) 635 { 636 pvo_pt->pte_lo |= pt->pte_lo & (PTE_REF|PTE_CHG); 637 } 638 639 static __inline void 640 pmap_pte_clear(volatile struct pte *pt, vaddr_t va, int ptebit) 641 { 642 /* 643 * As shown in Section 7.6.3.2.3 644 */ 645 pt->pte_lo &= ~ptebit; 646 TLBIE(va); 647 SYNC(); 648 EIEIO(); 649 TLBSYNC(); 650 SYNC(); 651 } 652 653 static __inline void 654 pmap_pte_set(volatile struct pte *pt, struct pte *pvo_pt) 655 { 656 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 657 if (pvo_pt->pte_hi & PTE_VALID) 658 panic("pte_set: setting an already valid pte %p", pvo_pt); 659 #endif 660 pvo_pt->pte_hi |= PTE_VALID; 661 /* 662 * Update the PTE as defined in section 7.6.3.1 663 * Note that the REF/CHG bits are from pvo_pt and thus should 664 * have been saved so this routine can restore them (if desired). 665 */ 666 pt->pte_lo = pvo_pt->pte_lo; 667 EIEIO(); 668 pt->pte_hi = pvo_pt->pte_hi; 669 SYNC(); 670 pmap_pte_valid++; 671 } 672 673 static __inline void 674 pmap_pte_unset(volatile struct pte *pt, struct pte *pvo_pt, vaddr_t va) 675 { 676 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 677 if ((pvo_pt->pte_hi & PTE_VALID) == 0) 678 panic("pte_unset: attempt to unset an inactive pte#1 %p/%p", pvo_pt, pt); 679 if ((pt->pte_hi & PTE_VALID) == 0) 680 panic("pte_unset: attempt to unset an inactive pte#2 %p/%p", pvo_pt, pt); 681 #endif 682 683 pvo_pt->pte_hi &= ~PTE_VALID; 684 /* 685 * Force the ref & chg bits back into the PTEs. 686 */ 687 SYNC(); 688 /* 689 * Invalidate the pte ... (Section 7.6.3.3) 690 */ 691 pt->pte_hi &= ~PTE_VALID; 692 SYNC(); 693 TLBIE(va); 694 SYNC(); 695 EIEIO(); 696 TLBSYNC(); 697 SYNC(); 698 /* 699 * Save the ref & chg bits ... 700 */ 701 pmap_pte_synch(pt, pvo_pt); 702 pmap_pte_valid--; 703 } 704 705 static __inline void 706 pmap_pte_change(volatile struct pte *pt, struct pte *pvo_pt, vaddr_t va) 707 { 708 /* 709 * Invalidate the PTE 710 */ 711 pmap_pte_unset(pt, pvo_pt, va); 712 pmap_pte_set(pt, pvo_pt); 713 } 714 715 /* 716 * Try to insert the PTE @ *pvo_pt into the pmap_pteg_table at ptegidx 717 * (either primary or secondary location). 718 * 719 * Note: both the destination and source PTEs must not have PTE_VALID set. 720 */ 721 722 STATIC int 723 pmap_pte_insert(int ptegidx, struct pte *pvo_pt) 724 { 725 volatile struct pte *pt; 726 int i; 727 728 #if defined(DEBUG) 729 DPRINTFN(PTE, ("pmap_pte_insert: idx 0x%x, pte 0x%lx 0x%lx\n", 730 ptegidx, pvo_pt->pte_hi, pvo_pt->pte_lo)); 731 #endif 732 /* 733 * First try primary hash. 734 */ 735 for (pt = pmap_pteg_table[ptegidx].pt, i = 0; i < 8; i++, pt++) { 736 if ((pt->pte_hi & PTE_VALID) == 0) { 737 pvo_pt->pte_hi &= ~PTE_HID; 738 pmap_pte_set(pt, pvo_pt); 739 return i; 740 } 741 } 742 743 /* 744 * Now try secondary hash. 745 */ 746 ptegidx ^= pmap_pteg_mask; 747 for (pt = pmap_pteg_table[ptegidx].pt, i = 0; i < 8; i++, pt++) { 748 if ((pt->pte_hi & PTE_VALID) == 0) { 749 pvo_pt->pte_hi |= PTE_HID; 750 pmap_pte_set(pt, pvo_pt); 751 return i; 752 } 753 } 754 return -1; 755 } 756 757 /* 758 * Spill handler. 759 * 760 * Tries to spill a page table entry from the overflow area. 761 * This runs in either real mode (if dealing with a exception spill) 762 * or virtual mode when dealing with manually spilling one of the 763 * kernel's pte entries. In either case, interrupts are already 764 * disabled. 765 */ 766 767 int 768 pmap_pte_spill(struct pmap *pm, vaddr_t addr, boolean_t exec) 769 { 770 struct pvo_entry *source_pvo, *victim_pvo, *next_pvo; 771 struct pvo_entry *pvo; 772 /* XXX: gcc -- vpvoh is always set at either *1* or *2* */ 773 struct pvo_tqhead *pvoh, *vpvoh = NULL; 774 int ptegidx, i, j; 775 volatile struct pteg *pteg; 776 volatile struct pte *pt; 777 778 ptegidx = va_to_pteg(pm, addr); 779 780 /* 781 * Have to substitute some entry. Use the primary hash for this. 782 * Use low bits of timebase as random generator. Make sure we are 783 * not picking a kernel pte for replacement. 784 */ 785 pteg = &pmap_pteg_table[ptegidx]; 786 i = MFTB() & 7; 787 for (j = 0; j < 8; j++) { 788 pt = &pteg->pt[i]; 789 if ((pt->pte_hi & PTE_VALID) == 0 || 790 VSID_TO_HASH((pt->pte_hi & PTE_VSID) >> PTE_VSID_SHFT) 791 != KERNEL_VSIDBITS) 792 break; 793 i = (i + 1) & 7; 794 } 795 KASSERT(j < 8); 796 797 source_pvo = NULL; 798 victim_pvo = NULL; 799 pvoh = &pmap_pvo_table[ptegidx]; 800 TAILQ_FOREACH(pvo, pvoh, pvo_olink) { 801 802 /* 803 * We need to find pvo entry for this address... 804 */ 805 PMAP_PVO_CHECK(pvo); /* sanity check */ 806 807 /* 808 * If we haven't found the source and we come to a PVO with 809 * a valid PTE, then we know we can't find it because all 810 * evicted PVOs always are first in the list. 811 */ 812 if (source_pvo == NULL && (pvo->pvo_pte.pte_hi & PTE_VALID)) 813 break; 814 if (source_pvo == NULL && pm == pvo->pvo_pmap && 815 addr == PVO_VADDR(pvo)) { 816 817 /* 818 * Now we have found the entry to be spilled into the 819 * pteg. Attempt to insert it into the page table. 820 */ 821 j = pmap_pte_insert(ptegidx, &pvo->pvo_pte); 822 if (j >= 0) { 823 PVO_PTEGIDX_SET(pvo, j); 824 PMAP_PVO_CHECK(pvo); /* sanity check */ 825 PVO_WHERE(pvo, SPILL_INSERT); 826 pvo->pvo_pmap->pm_evictions--; 827 PMAPCOUNT(ptes_spilled); 828 PMAPCOUNT2(((pvo->pvo_pte.pte_hi & PTE_HID) 829 ? pmap_evcnt_ptes_secondary 830 : pmap_evcnt_ptes_primary)[j]); 831 832 /* 833 * Since we keep the evicted entries at the 834 * from of the PVO list, we need move this 835 * (now resident) PVO after the evicted 836 * entries. 837 */ 838 next_pvo = TAILQ_NEXT(pvo, pvo_olink); 839 840 /* 841 * If we don't have to move (either we were the 842 * last entry or the next entry was valid), 843 * don't change our position. Otherwise 844 * move ourselves to the tail of the queue. 845 */ 846 if (next_pvo != NULL && 847 !(next_pvo->pvo_pte.pte_hi & PTE_VALID)) { 848 TAILQ_REMOVE(pvoh, pvo, pvo_olink); 849 TAILQ_INSERT_TAIL(pvoh, pvo, pvo_olink); 850 } 851 return 1; 852 } 853 source_pvo = pvo; 854 if (exec && !PVO_ISEXECUTABLE(source_pvo)) { 855 return 0; 856 } 857 if (victim_pvo != NULL) 858 break; 859 } 860 861 /* 862 * We also need the pvo entry of the victim we are replacing 863 * so save the R & C bits of the PTE. 864 */ 865 if ((pt->pte_hi & PTE_HID) == 0 && victim_pvo == NULL && 866 pmap_pte_compare(pt, &pvo->pvo_pte)) { 867 vpvoh = pvoh; /* *1* */ 868 victim_pvo = pvo; 869 if (source_pvo != NULL) 870 break; 871 } 872 } 873 874 if (source_pvo == NULL) { 875 PMAPCOUNT(ptes_unspilled); 876 return 0; 877 } 878 879 if (victim_pvo == NULL) { 880 if ((pt->pte_hi & PTE_HID) == 0) 881 panic("pmap_pte_spill: victim p-pte (%p) has " 882 "no pvo entry!", pt); 883 884 /* 885 * If this is a secondary PTE, we need to search 886 * its primary pvo bucket for the matching PVO. 887 */ 888 vpvoh = &pmap_pvo_table[ptegidx ^ pmap_pteg_mask]; /* *2* */ 889 TAILQ_FOREACH(pvo, vpvoh, pvo_olink) { 890 PMAP_PVO_CHECK(pvo); /* sanity check */ 891 892 /* 893 * We also need the pvo entry of the victim we are 894 * replacing so save the R & C bits of the PTE. 895 */ 896 if (pmap_pte_compare(pt, &pvo->pvo_pte)) { 897 victim_pvo = pvo; 898 break; 899 } 900 } 901 if (victim_pvo == NULL) 902 panic("pmap_pte_spill: victim s-pte (%p) has " 903 "no pvo entry!", pt); 904 } 905 906 /* 907 * The victim should be not be a kernel PVO/PTE entry. 908 */ 909 KASSERT(victim_pvo->pvo_pmap != pmap_kernel()); 910 KASSERT(PVO_PTEGIDX_ISSET(victim_pvo)); 911 KASSERT(PVO_PTEGIDX_GET(victim_pvo) == i); 912 913 /* 914 * We are invalidating the TLB entry for the EA for the 915 * we are replacing even though its valid; If we don't 916 * we lose any ref/chg bit changes contained in the TLB 917 * entry. 918 */ 919 source_pvo->pvo_pte.pte_hi &= ~PTE_HID; 920 921 /* 922 * To enforce the PVO list ordering constraint that all 923 * evicted entries should come before all valid entries, 924 * move the source PVO to the tail of its list and the 925 * victim PVO to the head of its list (which might not be 926 * the same list, if the victim was using the secondary hash). 927 */ 928 TAILQ_REMOVE(pvoh, source_pvo, pvo_olink); 929 TAILQ_INSERT_TAIL(pvoh, source_pvo, pvo_olink); 930 TAILQ_REMOVE(vpvoh, victim_pvo, pvo_olink); 931 TAILQ_INSERT_HEAD(vpvoh, victim_pvo, pvo_olink); 932 pmap_pte_unset(pt, &victim_pvo->pvo_pte, victim_pvo->pvo_vaddr); 933 pmap_pte_set(pt, &source_pvo->pvo_pte); 934 victim_pvo->pvo_pmap->pm_evictions++; 935 source_pvo->pvo_pmap->pm_evictions--; 936 PVO_WHERE(victim_pvo, SPILL_UNSET); 937 PVO_WHERE(source_pvo, SPILL_SET); 938 939 PVO_PTEGIDX_CLR(victim_pvo); 940 PVO_PTEGIDX_SET(source_pvo, i); 941 PMAPCOUNT2(pmap_evcnt_ptes_primary[i]); 942 PMAPCOUNT(ptes_spilled); 943 PMAPCOUNT(ptes_evicted); 944 PMAPCOUNT(ptes_removed); 945 946 PMAP_PVO_CHECK(victim_pvo); 947 PMAP_PVO_CHECK(source_pvo); 948 return 1; 949 } 950 951 /* 952 * Restrict given range to physical memory 953 */ 954 void 955 pmap_real_memory(paddr_t *start, psize_t *size) 956 { 957 struct mem_region *mp; 958 959 for (mp = mem; mp->size; mp++) { 960 if (*start + *size > mp->start 961 && *start < mp->start + mp->size) { 962 if (*start < mp->start) { 963 *size -= mp->start - *start; 964 *start = mp->start; 965 } 966 if (*start + *size > mp->start + mp->size) 967 *size = mp->start + mp->size - *start; 968 return; 969 } 970 } 971 *size = 0; 972 } 973 974 /* 975 * Initialize anything else for pmap handling. 976 * Called during vm_init(). 977 */ 978 void 979 pmap_init(void) 980 { 981 #ifdef __HAVE_PMAP_PHYSSEG 982 struct pvo_tqhead *pvoh; 983 int bank; 984 long sz; 985 char *attr; 986 987 pvoh = pmap_physseg.pvoh; 988 attr = pmap_physseg.attrs; 989 for (bank = 0; bank < vm_nphysseg; bank++) { 990 sz = vm_physmem[bank].end - vm_physmem[bank].start; 991 vm_physmem[bank].pmseg.pvoh = pvoh; 992 vm_physmem[bank].pmseg.attrs = attr; 993 for (; sz > 0; sz--, pvoh++, attr++) { 994 TAILQ_INIT(pvoh); 995 *attr = 0; 996 } 997 } 998 #endif 999 1000 pool_init(&pmap_mpvo_pool, sizeof(struct pvo_entry), 1001 sizeof(struct pvo_entry), 0, 0, "pmap_mpvopl", 1002 &pmap_pool_mallocator); 1003 1004 pool_setlowat(&pmap_mpvo_pool, 1008); 1005 1006 pmap_initialized = 1; 1007 1008 #ifdef PMAPCOUNTERS 1009 evcnt_attach_static(&pmap_evcnt_mappings); 1010 evcnt_attach_static(&pmap_evcnt_mappings_replaced); 1011 evcnt_attach_static(&pmap_evcnt_unmappings); 1012 1013 evcnt_attach_static(&pmap_evcnt_kernel_mappings); 1014 evcnt_attach_static(&pmap_evcnt_kernel_unmappings); 1015 1016 evcnt_attach_static(&pmap_evcnt_exec_mappings); 1017 evcnt_attach_static(&pmap_evcnt_exec_cached); 1018 evcnt_attach_static(&pmap_evcnt_exec_synced); 1019 evcnt_attach_static(&pmap_evcnt_exec_synced_clear_modify); 1020 1021 evcnt_attach_static(&pmap_evcnt_exec_uncached_page_protect); 1022 evcnt_attach_static(&pmap_evcnt_exec_uncached_clear_modify); 1023 evcnt_attach_static(&pmap_evcnt_exec_uncached_zero_page); 1024 evcnt_attach_static(&pmap_evcnt_exec_uncached_copy_page); 1025 1026 evcnt_attach_static(&pmap_evcnt_zeroed_pages); 1027 evcnt_attach_static(&pmap_evcnt_copied_pages); 1028 evcnt_attach_static(&pmap_evcnt_idlezeroed_pages); 1029 1030 evcnt_attach_static(&pmap_evcnt_updates); 1031 evcnt_attach_static(&pmap_evcnt_collects); 1032 evcnt_attach_static(&pmap_evcnt_copies); 1033 1034 evcnt_attach_static(&pmap_evcnt_ptes_spilled); 1035 evcnt_attach_static(&pmap_evcnt_ptes_unspilled); 1036 evcnt_attach_static(&pmap_evcnt_ptes_evicted); 1037 evcnt_attach_static(&pmap_evcnt_ptes_removed); 1038 evcnt_attach_static(&pmap_evcnt_ptes_changed); 1039 evcnt_attach_static(&pmap_evcnt_ptes_primary[0]); 1040 evcnt_attach_static(&pmap_evcnt_ptes_primary[1]); 1041 evcnt_attach_static(&pmap_evcnt_ptes_primary[2]); 1042 evcnt_attach_static(&pmap_evcnt_ptes_primary[3]); 1043 evcnt_attach_static(&pmap_evcnt_ptes_primary[4]); 1044 evcnt_attach_static(&pmap_evcnt_ptes_primary[5]); 1045 evcnt_attach_static(&pmap_evcnt_ptes_primary[6]); 1046 evcnt_attach_static(&pmap_evcnt_ptes_primary[7]); 1047 evcnt_attach_static(&pmap_evcnt_ptes_secondary[0]); 1048 evcnt_attach_static(&pmap_evcnt_ptes_secondary[1]); 1049 evcnt_attach_static(&pmap_evcnt_ptes_secondary[2]); 1050 evcnt_attach_static(&pmap_evcnt_ptes_secondary[3]); 1051 evcnt_attach_static(&pmap_evcnt_ptes_secondary[4]); 1052 evcnt_attach_static(&pmap_evcnt_ptes_secondary[5]); 1053 evcnt_attach_static(&pmap_evcnt_ptes_secondary[6]); 1054 evcnt_attach_static(&pmap_evcnt_ptes_secondary[7]); 1055 #endif 1056 } 1057 1058 /* 1059 * How much virtual space does the kernel get? 1060 */ 1061 void 1062 pmap_virtual_space(vaddr_t *start, vaddr_t *end) 1063 { 1064 /* 1065 * For now, reserve one segment (minus some overhead) for kernel 1066 * virtual memory 1067 */ 1068 *start = VM_MIN_KERNEL_ADDRESS; 1069 *end = VM_MAX_KERNEL_ADDRESS; 1070 } 1071 1072 /* 1073 * Allocate, initialize, and return a new physical map. 1074 */ 1075 pmap_t 1076 pmap_create(void) 1077 { 1078 pmap_t pm; 1079 1080 pm = pool_get(&pmap_pool, PR_WAITOK); 1081 memset((caddr_t)pm, 0, sizeof *pm); 1082 pmap_pinit(pm); 1083 1084 DPRINTFN(CREATE,("pmap_create: pm %p:\n" 1085 "\t%06lx %06lx %06lx %06lx %06lx %06lx %06lx %06lx\n" 1086 "\t%06lx %06lx %06lx %06lx %06lx %06lx %06lx %06lx\n", pm, 1087 pm->pm_sr[0], pm->pm_sr[1], pm->pm_sr[2], pm->pm_sr[3], 1088 pm->pm_sr[4], pm->pm_sr[5], pm->pm_sr[6], pm->pm_sr[7], 1089 pm->pm_sr[8], pm->pm_sr[9], pm->pm_sr[10], pm->pm_sr[11], 1090 pm->pm_sr[12], pm->pm_sr[13], pm->pm_sr[14], pm->pm_sr[15])); 1091 return pm; 1092 } 1093 1094 /* 1095 * Initialize a preallocated and zeroed pmap structure. 1096 */ 1097 void 1098 pmap_pinit(pmap_t pm) 1099 { 1100 register_t entropy = MFTB(); 1101 register_t mask; 1102 int i; 1103 1104 /* 1105 * Allocate some segment registers for this pmap. 1106 */ 1107 pm->pm_refs = 1; 1108 for (i = 0; i < NPMAPS; i += VSID_NBPW) { 1109 static register_t pmap_vsidcontext; 1110 register_t hash; 1111 unsigned int n; 1112 1113 /* Create a new value by multiplying by a prime adding in 1114 * entropy from the timebase register. This is to make the 1115 * VSID more random so that the PT Hash function collides 1116 * less often. (note that the prime causes gcc to do shifts 1117 * instead of a multiply) 1118 */ 1119 pmap_vsidcontext = (pmap_vsidcontext * 0x1105) + entropy; 1120 hash = pmap_vsidcontext & (NPMAPS - 1); 1121 if (hash == 0) /* 0 is special, avoid it */ 1122 continue; 1123 n = hash >> 5; 1124 mask = 1L << (hash & (VSID_NBPW-1)); 1125 hash = pmap_vsidcontext; 1126 if (pmap_vsid_bitmap[n] & mask) { /* collision? */ 1127 /* anything free in this bucket? */ 1128 if (~pmap_vsid_bitmap[n] == 0) { 1129 entropy = hash >> PTE_VSID_SHFT; 1130 continue; 1131 } 1132 i = ffs(~pmap_vsid_bitmap[n]) - 1; 1133 mask = 1L << i; 1134 hash &= ~(VSID_NBPW-1); 1135 hash |= i; 1136 } 1137 /* 1138 * Make sure clear out SR_KEY_LEN bits because we put our 1139 * our data in those bits (to identify the segment). 1140 */ 1141 hash &= PTE_VSID >> (PTE_VSID_SHFT + SR_KEY_LEN); 1142 pmap_vsid_bitmap[n] |= mask; 1143 for (i = 0; i < 16; i++) 1144 pm->pm_sr[i] = VSID_MAKE(i, hash) | SR_PRKEY | 1145 SR_NOEXEC; 1146 return; 1147 } 1148 panic("pmap_pinit: out of segments"); 1149 } 1150 1151 /* 1152 * Add a reference to the given pmap. 1153 */ 1154 void 1155 pmap_reference(pmap_t pm) 1156 { 1157 pm->pm_refs++; 1158 } 1159 1160 /* 1161 * Retire the given pmap from service. 1162 * Should only be called if the map contains no valid mappings. 1163 */ 1164 void 1165 pmap_destroy(pmap_t pm) 1166 { 1167 if (--pm->pm_refs == 0) { 1168 pmap_release(pm); 1169 pool_put(&pmap_pool, pm); 1170 } 1171 } 1172 1173 /* 1174 * Release any resources held by the given physical map. 1175 * Called when a pmap initialized by pmap_pinit is being released. 1176 */ 1177 void 1178 pmap_release(pmap_t pm) 1179 { 1180 int idx, mask; 1181 1182 if (pm->pm_sr[0] == 0) 1183 panic("pmap_release"); 1184 idx = VSID_TO_HASH(pm->pm_sr[0]) & (NPMAPS-1); 1185 mask = 1 << (idx % VSID_NBPW); 1186 idx /= VSID_NBPW; 1187 pmap_vsid_bitmap[idx] &= ~mask; 1188 } 1189 1190 /* 1191 * Copy the range specified by src_addr/len 1192 * from the source map to the range dst_addr/len 1193 * in the destination map. 1194 * 1195 * This routine is only advisory and need not do anything. 1196 */ 1197 void 1198 pmap_copy(pmap_t dst_pmap, pmap_t src_pmap, vaddr_t dst_addr, 1199 vsize_t len, vaddr_t src_addr) 1200 { 1201 PMAPCOUNT(copies); 1202 } 1203 1204 /* 1205 * Require that all active physical maps contain no 1206 * incorrect entries NOW. 1207 */ 1208 void 1209 pmap_update(struct pmap *pmap) 1210 { 1211 PMAPCOUNT(updates); 1212 TLBSYNC(); 1213 } 1214 1215 /* 1216 * Garbage collects the physical map system for 1217 * pages which are no longer used. 1218 * Success need not be guaranteed -- that is, there 1219 * may well be pages which are not referenced, but 1220 * others may be collected. 1221 * Called by the pageout daemon when pages are scarce. 1222 */ 1223 void 1224 pmap_collect(pmap_t pm) 1225 { 1226 PMAPCOUNT(collects); 1227 } 1228 1229 static __inline int 1230 pmap_pvo_pte_index(const struct pvo_entry *pvo, int ptegidx) 1231 { 1232 int pteidx; 1233 /* 1234 * We can find the actual pte entry without searching by 1235 * grabbing the PTEG index from 3 unused bits in pte_lo[11:9] 1236 * and by noticing the HID bit. 1237 */ 1238 pteidx = ptegidx * 8 + PVO_PTEGIDX_GET(pvo); 1239 if (pvo->pvo_pte.pte_hi & PTE_HID) 1240 pteidx ^= pmap_pteg_mask * 8; 1241 return pteidx; 1242 } 1243 1244 volatile struct pte * 1245 pmap_pvo_to_pte(const struct pvo_entry *pvo, int pteidx) 1246 { 1247 volatile struct pte *pt; 1248 1249 #if !defined(DIAGNOSTIC) && !defined(DEBUG) && !defined(PMAPCHECK) 1250 if ((pvo->pvo_pte.pte_hi & PTE_VALID) == 0) 1251 return NULL; 1252 #endif 1253 1254 /* 1255 * If we haven't been supplied the ptegidx, calculate it. 1256 */ 1257 if (pteidx == -1) { 1258 int ptegidx; 1259 ptegidx = va_to_pteg(pvo->pvo_pmap, pvo->pvo_vaddr); 1260 pteidx = pmap_pvo_pte_index(pvo, ptegidx); 1261 } 1262 1263 pt = &pmap_pteg_table[pteidx >> 3].pt[pteidx & 7]; 1264 1265 #if !defined(DIAGNOSTIC) && !defined(DEBUG) && !defined(PMAPCHECK) 1266 return pt; 1267 #else 1268 if ((pvo->pvo_pte.pte_hi & PTE_VALID) && !PVO_PTEGIDX_ISSET(pvo)) { 1269 panic("pmap_pvo_to_pte: pvo %p: has valid pte in " 1270 "pvo but no valid pte index", pvo); 1271 } 1272 if ((pvo->pvo_pte.pte_hi & PTE_VALID) == 0 && PVO_PTEGIDX_ISSET(pvo)) { 1273 panic("pmap_pvo_to_pte: pvo %p: has valid pte index in " 1274 "pvo but no valid pte", pvo); 1275 } 1276 1277 if ((pt->pte_hi ^ (pvo->pvo_pte.pte_hi & ~PTE_VALID)) == PTE_VALID) { 1278 if ((pvo->pvo_pte.pte_hi & PTE_VALID) == 0) { 1279 #if defined(DEBUG) || defined(PMAPCHECK) 1280 pmap_pte_print(pt); 1281 #endif 1282 panic("pmap_pvo_to_pte: pvo %p: has valid pte in " 1283 "pmap_pteg_table %p but invalid in pvo", 1284 pvo, pt); 1285 } 1286 if (((pt->pte_lo ^ pvo->pvo_pte.pte_lo) & ~(PTE_CHG|PTE_REF)) != 0) { 1287 #if defined(DEBUG) || defined(PMAPCHECK) 1288 pmap_pte_print(pt); 1289 #endif 1290 panic("pmap_pvo_to_pte: pvo %p: pvo pte does " 1291 "not match pte %p in pmap_pteg_table", 1292 pvo, pt); 1293 } 1294 return pt; 1295 } 1296 1297 if (pvo->pvo_pte.pte_hi & PTE_VALID) { 1298 #if defined(DEBUG) || defined(PMAPCHECK) 1299 pmap_pte_print(pt); 1300 #endif 1301 panic("pmap_pvo_to_pte: pvo %p: has nomatching pte %p in " 1302 "pmap_pteg_table but valid in pvo", pvo, pt); 1303 } 1304 return NULL; 1305 #endif /* !(!DIAGNOSTIC && !DEBUG && !PMAPCHECK) */ 1306 } 1307 1308 struct pvo_entry * 1309 pmap_pvo_find_va(pmap_t pm, vaddr_t va, int *pteidx_p) 1310 { 1311 struct pvo_entry *pvo; 1312 int ptegidx; 1313 1314 va &= ~ADDR_POFF; 1315 ptegidx = va_to_pteg(pm, va); 1316 1317 TAILQ_FOREACH(pvo, &pmap_pvo_table[ptegidx], pvo_olink) { 1318 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1319 if ((uintptr_t) pvo >= SEGMENT_LENGTH) 1320 panic("pmap_pvo_find_va: invalid pvo %p on " 1321 "list %#x (%p)", pvo, ptegidx, 1322 &pmap_pvo_table[ptegidx]); 1323 #endif 1324 if (pvo->pvo_pmap == pm && PVO_VADDR(pvo) == va) { 1325 if (pteidx_p) 1326 *pteidx_p = pmap_pvo_pte_index(pvo, ptegidx); 1327 return pvo; 1328 } 1329 } 1330 return NULL; 1331 } 1332 1333 #if defined(DEBUG) || defined(PMAPCHECK) 1334 void 1335 pmap_pvo_check(const struct pvo_entry *pvo) 1336 { 1337 struct pvo_head *pvo_head; 1338 struct pvo_entry *pvo0; 1339 volatile struct pte *pt; 1340 int failed = 0; 1341 1342 if ((uintptr_t)(pvo+1) >= SEGMENT_LENGTH) 1343 panic("pmap_pvo_check: pvo %p: invalid address", pvo); 1344 1345 if ((uintptr_t)(pvo->pvo_pmap+1) >= SEGMENT_LENGTH) { 1346 printf("pmap_pvo_check: pvo %p: invalid pmap address %p\n", 1347 pvo, pvo->pvo_pmap); 1348 failed = 1; 1349 } 1350 1351 if ((uintptr_t)TAILQ_NEXT(pvo, pvo_olink) >= SEGMENT_LENGTH || 1352 (((uintptr_t)TAILQ_NEXT(pvo, pvo_olink)) & 0x1f) != 0) { 1353 printf("pmap_pvo_check: pvo %p: invalid ovlink address %p\n", 1354 pvo, TAILQ_NEXT(pvo, pvo_olink)); 1355 failed = 1; 1356 } 1357 1358 if ((uintptr_t)LIST_NEXT(pvo, pvo_vlink) >= SEGMENT_LENGTH || 1359 (((uintptr_t)LIST_NEXT(pvo, pvo_vlink)) & 0x1f) != 0) { 1360 printf("pmap_pvo_check: pvo %p: invalid ovlink address %p\n", 1361 pvo, LIST_NEXT(pvo, pvo_vlink)); 1362 failed = 1; 1363 } 1364 1365 if (pvo->pvo_vaddr & PVO_MANAGED) { 1366 pvo_head = pa_to_pvoh(pvo->pvo_pte.pte_lo & PTE_RPGN, NULL); 1367 } else { 1368 if (pvo->pvo_vaddr < VM_MIN_KERNEL_ADDRESS) { 1369 printf("pmap_pvo_check: pvo %p: non kernel address " 1370 "on kernel unmanaged list\n", pvo); 1371 failed = 1; 1372 } 1373 pvo_head = &pmap_pvo_kunmanaged; 1374 } 1375 LIST_FOREACH(pvo0, pvo_head, pvo_vlink) { 1376 if (pvo0 == pvo) 1377 break; 1378 } 1379 if (pvo0 == NULL) { 1380 printf("pmap_pvo_check: pvo %p: not present " 1381 "on its vlist head %p\n", pvo, pvo_head); 1382 failed = 1; 1383 } 1384 if (pvo != pmap_pvo_find_va(pvo->pvo_pmap, pvo->pvo_vaddr, NULL)) { 1385 printf("pmap_pvo_check: pvo %p: not present " 1386 "on its olist head\n", pvo); 1387 failed = 1; 1388 } 1389 pt = pmap_pvo_to_pte(pvo, -1); 1390 if (pt == NULL) { 1391 if (pvo->pvo_pte.pte_hi & PTE_VALID) { 1392 printf("pmap_pvo_check: pvo %p: pte_hi VALID but " 1393 "no PTE\n", pvo); 1394 failed = 1; 1395 } 1396 } else { 1397 if ((uintptr_t) pt < (uintptr_t) &pmap_pteg_table[0] || 1398 (uintptr_t) pt >= 1399 (uintptr_t) &pmap_pteg_table[pmap_pteg_cnt]) { 1400 printf("pmap_pvo_check: pvo %p: pte %p not in " 1401 "pteg table\n", pvo, pt); 1402 failed = 1; 1403 } 1404 if (((((uintptr_t) pt) >> 3) & 7) != PVO_PTEGIDX_GET(pvo)) { 1405 printf("pmap_pvo_check: pvo %p: pte_hi VALID but " 1406 "no PTE\n", pvo); 1407 failed = 1; 1408 } 1409 if (pvo->pvo_pte.pte_hi != pt->pte_hi) { 1410 printf("pmap_pvo_check: pvo %p: pte_hi differ: " 1411 "%#lx/%#lx\n", pvo, pvo->pvo_pte.pte_hi, pt->pte_hi); 1412 failed = 1; 1413 } 1414 if (((pvo->pvo_pte.pte_lo ^ pt->pte_lo) & 1415 (PTE_PP|PTE_WIMG|PTE_RPGN)) != 0) { 1416 printf("pmap_pvo_check: pvo %p: pte_lo differ: " 1417 "%#lx/%#lx\n", pvo, 1418 pvo->pvo_pte.pte_lo & (PTE_PP|PTE_WIMG|PTE_RPGN), 1419 pt->pte_lo & (PTE_PP|PTE_WIMG|PTE_RPGN)); 1420 failed = 1; 1421 } 1422 if ((pmap_pte_to_va(pt) ^ PVO_VADDR(pvo)) & 0x0fffffff) { 1423 printf("pmap_pvo_check: pvo %p: PTE %p derived VA %#lx" 1424 " doesn't not match PVO's VA %#lx\n", 1425 pvo, pt, pmap_pte_to_va(pt), PVO_VADDR(pvo)); 1426 failed = 1; 1427 } 1428 if (failed) 1429 pmap_pte_print(pt); 1430 } 1431 if (failed) 1432 panic("pmap_pvo_check: pvo %p, pm %p: bugcheck!", pvo, 1433 pvo->pvo_pmap); 1434 } 1435 #endif /* DEBUG || PMAPCHECK */ 1436 1437 /* 1438 * This returns whether this is the first mapping of a page. 1439 */ 1440 int 1441 pmap_pvo_enter(pmap_t pm, struct pool *pl, struct pvo_head *pvo_head, 1442 vaddr_t va, paddr_t pa, register_t pte_lo, int flags) 1443 { 1444 struct pvo_entry *pvo; 1445 struct pvo_tqhead *pvoh; 1446 register_t msr; 1447 int ptegidx; 1448 int i; 1449 int poolflags = PR_NOWAIT; 1450 1451 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1452 if (pmap_pvo_remove_depth > 0) 1453 panic("pmap_pvo_enter: called while pmap_pvo_remove active!"); 1454 if (++pmap_pvo_enter_depth > 1) 1455 panic("pmap_pvo_enter: called recursively!"); 1456 #endif 1457 1458 /* 1459 * Compute the PTE Group index. 1460 */ 1461 va &= ~ADDR_POFF; 1462 ptegidx = va_to_pteg(pm, va); 1463 1464 msr = pmap_interrupts_off(); 1465 /* 1466 * Remove any existing mapping for this page. Reuse the 1467 * pvo entry if there a mapping. 1468 */ 1469 TAILQ_FOREACH(pvo, &pmap_pvo_table[ptegidx], pvo_olink) { 1470 if (pvo->pvo_pmap == pm && PVO_VADDR(pvo) == va) { 1471 #ifdef DEBUG 1472 if ((pmapdebug & PMAPDEBUG_PVOENTER) && 1473 ((pvo->pvo_pte.pte_lo ^ (pa|pte_lo)) & 1474 ~(PTE_REF|PTE_CHG)) == 0 && 1475 va < VM_MIN_KERNEL_ADDRESS) { 1476 printf("pmap_pvo_enter: pvo %p: dup %#lx/%#lx\n", 1477 pvo, pvo->pvo_pte.pte_lo, pte_lo|pa); 1478 printf("pmap_pvo_enter: pte_hi=%#lx sr=%#lx\n", 1479 pvo->pvo_pte.pte_hi, 1480 pm->pm_sr[va >> ADDR_SR_SHFT]); 1481 pmap_pte_print(pmap_pvo_to_pte(pvo, -1)); 1482 #ifdef DDBX 1483 Debugger(); 1484 #endif 1485 } 1486 #endif 1487 PMAPCOUNT(mappings_replaced); 1488 pmap_pvo_remove(pvo, -1); 1489 break; 1490 } 1491 } 1492 1493 /* 1494 * If we aren't overwriting an mapping, try to allocate 1495 */ 1496 pmap_interrupts_restore(msr); 1497 pvo = pool_get(pl, poolflags); 1498 msr = pmap_interrupts_off(); 1499 if (pvo == NULL) { 1500 pvo = pmap_pvo_reclaim(pm); 1501 if (pvo == NULL) { 1502 if ((flags & PMAP_CANFAIL) == 0) 1503 panic("pmap_pvo_enter: failed"); 1504 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1505 pmap_pvo_enter_depth--; 1506 #endif 1507 pmap_interrupts_restore(msr); 1508 return ENOMEM; 1509 } 1510 } 1511 pvo->pvo_vaddr = va; 1512 pvo->pvo_pmap = pm; 1513 pvo->pvo_vaddr &= ~ADDR_POFF; 1514 if (flags & VM_PROT_EXECUTE) { 1515 PMAPCOUNT(exec_mappings); 1516 pvo_set_exec(pvo); 1517 } 1518 if (flags & PMAP_WIRED) 1519 pvo->pvo_vaddr |= PVO_WIRED; 1520 if (pvo_head != &pmap_pvo_kunmanaged) { 1521 pvo->pvo_vaddr |= PVO_MANAGED; 1522 PMAPCOUNT(mappings); 1523 } else { 1524 PMAPCOUNT(kernel_mappings); 1525 } 1526 pmap_pte_create(&pvo->pvo_pte, pm, va, pa | pte_lo); 1527 1528 LIST_INSERT_HEAD(pvo_head, pvo, pvo_vlink); 1529 if (pvo->pvo_pte.pte_lo & PVO_WIRED) 1530 pvo->pvo_pmap->pm_stats.wired_count++; 1531 pvo->pvo_pmap->pm_stats.resident_count++; 1532 #if defined(DEBUG) 1533 if (pm != pmap_kernel() && va < VM_MIN_KERNEL_ADDRESS) 1534 DPRINTFN(PVOENTER, 1535 ("pmap_pvo_enter: pvo %p: pm %p va %#lx pa %#lx\n", 1536 pvo, pm, va, pa)); 1537 #endif 1538 1539 /* 1540 * We hope this succeeds but it isn't required. 1541 */ 1542 pvoh = &pmap_pvo_table[ptegidx]; 1543 i = pmap_pte_insert(ptegidx, &pvo->pvo_pte); 1544 if (i >= 0) { 1545 PVO_PTEGIDX_SET(pvo, i); 1546 PVO_WHERE(pvo, ENTER_INSERT); 1547 PMAPCOUNT2(((pvo->pvo_pte.pte_hi & PTE_HID) 1548 ? pmap_evcnt_ptes_secondary : pmap_evcnt_ptes_primary)[i]); 1549 TAILQ_INSERT_TAIL(pvoh, pvo, pvo_olink); 1550 } else { 1551 /* 1552 * Since we didn't have room for this entry (which makes it 1553 * and evicted entry), place it at the head of the list. 1554 */ 1555 TAILQ_INSERT_HEAD(pvoh, pvo, pvo_olink); 1556 PMAPCOUNT(ptes_evicted); 1557 pm->pm_evictions++; 1558 /* 1559 * If this is a kernel page, make sure it's active. 1560 */ 1561 if (pm == pmap_kernel()) { 1562 i = pmap_pte_spill(pm, va, FALSE); 1563 KASSERT(i); 1564 } 1565 } 1566 PMAP_PVO_CHECK(pvo); /* sanity check */ 1567 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1568 pmap_pvo_enter_depth--; 1569 #endif 1570 pmap_interrupts_restore(msr); 1571 return 0; 1572 } 1573 1574 void 1575 pmap_pvo_remove(struct pvo_entry *pvo, int pteidx) 1576 { 1577 volatile struct pte *pt; 1578 int ptegidx; 1579 1580 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1581 if (++pmap_pvo_remove_depth > 1) 1582 panic("pmap_pvo_remove: called recursively!"); 1583 #endif 1584 1585 /* 1586 * If we haven't been supplied the ptegidx, calculate it. 1587 */ 1588 if (pteidx == -1) { 1589 ptegidx = va_to_pteg(pvo->pvo_pmap, pvo->pvo_vaddr); 1590 pteidx = pmap_pvo_pte_index(pvo, ptegidx); 1591 } else { 1592 ptegidx = pteidx >> 3; 1593 if (pvo->pvo_pte.pte_hi & PTE_HID) 1594 ptegidx ^= pmap_pteg_mask; 1595 } 1596 PMAP_PVO_CHECK(pvo); /* sanity check */ 1597 1598 /* 1599 * If there is an active pte entry, we need to deactivate it 1600 * (and save the ref & chg bits). 1601 */ 1602 pt = pmap_pvo_to_pte(pvo, pteidx); 1603 if (pt != NULL) { 1604 pmap_pte_unset(pt, &pvo->pvo_pte, pvo->pvo_vaddr); 1605 PVO_WHERE(pvo, REMOVE); 1606 PVO_PTEGIDX_CLR(pvo); 1607 PMAPCOUNT(ptes_removed); 1608 } else { 1609 KASSERT(pvo->pvo_pmap->pm_evictions > 0); 1610 pvo->pvo_pmap->pm_evictions--; 1611 } 1612 1613 /* 1614 * Account for executable mappings. 1615 */ 1616 if (PVO_ISEXECUTABLE(pvo)) 1617 pvo_clear_exec(pvo); 1618 1619 /* 1620 * Update our statistics. 1621 */ 1622 pvo->pvo_pmap->pm_stats.resident_count--; 1623 if (pvo->pvo_pte.pte_lo & PVO_WIRED) 1624 pvo->pvo_pmap->pm_stats.wired_count--; 1625 1626 /* 1627 * Save the REF/CHG bits into their cache if the page is managed. 1628 */ 1629 if (pvo->pvo_vaddr & PVO_MANAGED) { 1630 register_t ptelo = pvo->pvo_pte.pte_lo; 1631 struct vm_page *pg = PHYS_TO_VM_PAGE(ptelo & PTE_RPGN); 1632 1633 if (pg != NULL) { 1634 pmap_attr_save(pg, ptelo & (PTE_REF|PTE_CHG)); 1635 } 1636 PMAPCOUNT(unmappings); 1637 } else { 1638 PMAPCOUNT(kernel_unmappings); 1639 } 1640 1641 /* 1642 * Remove the PVO from its lists and return it to the pool. 1643 */ 1644 LIST_REMOVE(pvo, pvo_vlink); 1645 TAILQ_REMOVE(&pmap_pvo_table[ptegidx], pvo, pvo_olink); 1646 pool_put(pvo->pvo_vaddr & PVO_MANAGED 1647 ? &pmap_mpvo_pool : &pmap_upvo_pool, pvo); 1648 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 1649 pmap_pvo_remove_depth--; 1650 #endif 1651 } 1652 1653 /* 1654 * Mark a mapping as executable. 1655 * If this is the first executable mapping in the segment, 1656 * clear the noexec flag. 1657 */ 1658 STATIC void 1659 pvo_set_exec(struct pvo_entry *pvo) 1660 { 1661 struct pmap *pm = pvo->pvo_pmap; 1662 int sr; 1663 1664 if (pm == pmap_kernel() || PVO_ISEXECUTABLE(pvo)) { 1665 return; 1666 } 1667 pvo->pvo_vaddr |= PVO_EXECUTABLE; 1668 sr = PVO_VADDR(pvo) >> ADDR_SR_SHFT; 1669 if (pm->pm_exec[sr]++ == 0) { 1670 pm->pm_sr[sr] &= ~SR_NOEXEC; 1671 } 1672 } 1673 1674 /* 1675 * Mark a mapping as non-executable. 1676 * If this was the last executable mapping in the segment, 1677 * set the noexec flag. 1678 */ 1679 STATIC void 1680 pvo_clear_exec(struct pvo_entry *pvo) 1681 { 1682 struct pmap *pm = pvo->pvo_pmap; 1683 int sr; 1684 1685 if (pm == pmap_kernel() || !PVO_ISEXECUTABLE(pvo)) { 1686 return; 1687 } 1688 pvo->pvo_vaddr &= ~PVO_EXECUTABLE; 1689 sr = PVO_VADDR(pvo) >> ADDR_SR_SHFT; 1690 if (--pm->pm_exec[sr] == 0) { 1691 pm->pm_sr[sr] |= SR_NOEXEC; 1692 } 1693 } 1694 1695 /* 1696 * Insert physical page at pa into the given pmap at virtual address va. 1697 */ 1698 int 1699 pmap_enter(pmap_t pm, vaddr_t va, paddr_t pa, vm_prot_t prot, int flags) 1700 { 1701 struct mem_region *mp; 1702 struct pvo_head *pvo_head; 1703 struct vm_page *pg; 1704 struct pool *pl; 1705 register_t pte_lo; 1706 int error; 1707 u_int pvo_flags; 1708 u_int was_exec = 0; 1709 1710 if (__predict_false(!pmap_initialized)) { 1711 pvo_head = &pmap_pvo_kunmanaged; 1712 pl = &pmap_upvo_pool; 1713 pvo_flags = 0; 1714 pg = NULL; 1715 was_exec = PTE_EXEC; 1716 } else { 1717 pvo_head = pa_to_pvoh(pa, &pg); 1718 pl = &pmap_mpvo_pool; 1719 pvo_flags = PVO_MANAGED; 1720 } 1721 1722 DPRINTFN(ENTER, 1723 ("pmap_enter(%p, 0x%lx, 0x%lx, 0x%x, 0x%x):", 1724 pm, va, pa, prot, flags)); 1725 1726 /* 1727 * If this is a managed page, and it's the first reference to the 1728 * page clear the execness of the page. Otherwise fetch the execness. 1729 */ 1730 if (pg != NULL) 1731 was_exec = pmap_attr_fetch(pg) & PTE_EXEC; 1732 1733 DPRINTFN(ENTER, (" was_exec=%d", was_exec)); 1734 1735 /* 1736 * Assume the page is cache inhibited and access is guarded unless 1737 * it's in our available memory array. If it is in the memory array, 1738 * asssume it's in memory coherent memory. 1739 */ 1740 pte_lo = PTE_IG; 1741 if ((flags & PMAP_NC) == 0) { 1742 for (mp = mem; mp->size; mp++) { 1743 if (pa >= mp->start && pa < mp->start + mp->size) { 1744 pte_lo = PTE_M; 1745 break; 1746 } 1747 } 1748 } 1749 1750 if (prot & VM_PROT_WRITE) 1751 pte_lo |= PTE_BW; 1752 else 1753 pte_lo |= PTE_BR; 1754 1755 /* 1756 * If this was in response to a fault, "pre-fault" the PTE's 1757 * changed/referenced bit appropriately. 1758 */ 1759 if (flags & VM_PROT_WRITE) 1760 pte_lo |= PTE_CHG; 1761 if (flags & (VM_PROT_READ|VM_PROT_WRITE)) 1762 pte_lo |= PTE_REF; 1763 1764 /* 1765 * We need to know if this page can be executable 1766 */ 1767 flags |= (prot & VM_PROT_EXECUTE); 1768 1769 /* 1770 * Record mapping for later back-translation and pte spilling. 1771 * This will overwrite any existing mapping. 1772 */ 1773 error = pmap_pvo_enter(pm, pl, pvo_head, va, pa, pte_lo, flags); 1774 1775 /* 1776 * Flush the real page from the instruction cache if this page is 1777 * mapped executable and cacheable and has not been flushed since 1778 * the last time it was modified. 1779 */ 1780 if (error == 0 && 1781 (flags & VM_PROT_EXECUTE) && 1782 (pte_lo & PTE_I) == 0 && 1783 was_exec == 0) { 1784 DPRINTFN(ENTER, (" syncicache")); 1785 PMAPCOUNT(exec_synced); 1786 pmap_syncicache(pa, PAGE_SIZE); 1787 if (pg != NULL) { 1788 pmap_attr_save(pg, PTE_EXEC); 1789 PMAPCOUNT(exec_cached); 1790 #if defined(DEBUG) || defined(PMAPDEBUG) 1791 if (pmapdebug & PMAPDEBUG_ENTER) 1792 printf(" marked-as-exec"); 1793 else if (pmapdebug & PMAPDEBUG_EXEC) 1794 printf("[pmap_enter: %#lx: marked-as-exec]\n", 1795 pg->phys_addr); 1796 1797 #endif 1798 } 1799 } 1800 1801 DPRINTFN(ENTER, (": error=%d\n", error)); 1802 1803 return error; 1804 } 1805 1806 void 1807 pmap_kenter_pa(vaddr_t va, paddr_t pa, vm_prot_t prot) 1808 { 1809 struct mem_region *mp; 1810 register_t pte_lo; 1811 int error; 1812 1813 if (va < VM_MIN_KERNEL_ADDRESS) 1814 panic("pmap_kenter_pa: attempt to enter " 1815 "non-kernel address %#lx!", va); 1816 1817 DPRINTFN(KENTER, 1818 ("pmap_kenter_pa(%#lx,%#lx,%#x)\n", va, pa, prot)); 1819 1820 /* 1821 * Assume the page is cache inhibited and access is guarded unless 1822 * it's in our available memory array. If it is in the memory array, 1823 * asssume it's in memory coherent memory. 1824 */ 1825 pte_lo = PTE_IG; 1826 if ((prot & PMAP_NC) == 0) { 1827 for (mp = mem; mp->size; mp++) { 1828 if (pa >= mp->start && pa < mp->start + mp->size) { 1829 pte_lo = PTE_M; 1830 break; 1831 } 1832 } 1833 } 1834 1835 if (prot & VM_PROT_WRITE) 1836 pte_lo |= PTE_BW; 1837 else 1838 pte_lo |= PTE_BR; 1839 1840 /* 1841 * We don't care about REF/CHG on PVOs on the unmanaged list. 1842 */ 1843 error = pmap_pvo_enter(pmap_kernel(), &pmap_upvo_pool, 1844 &pmap_pvo_kunmanaged, va, pa, pte_lo, prot|PMAP_WIRED); 1845 1846 if (error != 0) 1847 panic("pmap_kenter_pa: failed to enter va %#lx pa %#lx: %d", 1848 va, pa, error); 1849 } 1850 1851 void 1852 pmap_kremove(vaddr_t va, vsize_t len) 1853 { 1854 if (va < VM_MIN_KERNEL_ADDRESS) 1855 panic("pmap_kremove: attempt to remove " 1856 "non-kernel address %#lx!", va); 1857 1858 DPRINTFN(KREMOVE,("pmap_kremove(%#lx,%#lx)\n", va, len)); 1859 pmap_remove(pmap_kernel(), va, va + len); 1860 } 1861 1862 /* 1863 * Remove the given range of mapping entries. 1864 */ 1865 void 1866 pmap_remove(pmap_t pm, vaddr_t va, vaddr_t endva) 1867 { 1868 struct pvo_entry *pvo; 1869 register_t msr; 1870 int pteidx; 1871 1872 msr = pmap_interrupts_off(); 1873 for (; va < endva; va += PAGE_SIZE) { 1874 pvo = pmap_pvo_find_va(pm, va, &pteidx); 1875 if (pvo != NULL) { 1876 pmap_pvo_remove(pvo, pteidx); 1877 } 1878 } 1879 pmap_interrupts_restore(msr); 1880 } 1881 1882 /* 1883 * Get the physical page address for the given pmap/virtual address. 1884 */ 1885 boolean_t 1886 pmap_extract(pmap_t pm, vaddr_t va, paddr_t *pap) 1887 { 1888 struct pvo_entry *pvo; 1889 register_t msr; 1890 1891 /* 1892 * If this is a kernel pmap lookup, also check the battable 1893 * and if we get a hit, translate the VA to a PA using the 1894 * BAT entries. Don't check for VM_MAX_KENREL_ADDRESS is 1895 * that will wrap back to 0. 1896 */ 1897 if (pm == pmap_kernel() && 1898 (va < VM_MIN_KERNEL_ADDRESS || 1899 (KERNEL2_SR < 15 && VM_MAX_KERNEL_ADDRESS <= va))) { 1900 register_t batu = battable[va >> ADDR_SR_SHFT].batu; 1901 KASSERT((va >> ADDR_SR_SHFT) != USER_SR); 1902 if (BAT_VALID_P(batu,0) && BAT_VA_MATCH_P(batu,va)) { 1903 register_t batl = battable[va >> ADDR_SR_SHFT].batl; 1904 register_t mask = (~(batu & BAT_BL) << 15) & ~0x1ffffL; 1905 *pap = (batl & mask) | (va & ~mask); 1906 return TRUE; 1907 } 1908 return FALSE; 1909 } 1910 1911 msr = pmap_interrupts_off(); 1912 pvo = pmap_pvo_find_va(pm, va & ~ADDR_POFF, NULL); 1913 if (pvo != NULL) { 1914 PMAP_PVO_CHECK(pvo); /* sanity check */ 1915 *pap = (pvo->pvo_pte.pte_lo & PTE_RPGN) | (va & ADDR_POFF); 1916 } 1917 pmap_interrupts_restore(msr); 1918 return pvo != NULL; 1919 } 1920 1921 /* 1922 * Lower the protection on the specified range of this pmap. 1923 */ 1924 void 1925 pmap_protect(pmap_t pm, vaddr_t va, vaddr_t endva, vm_prot_t prot) 1926 { 1927 struct pvo_entry *pvo; 1928 volatile struct pte *pt; 1929 register_t msr; 1930 int pteidx; 1931 1932 /* 1933 * Since this routine only downgrades protection, we should 1934 * always be called with at least one bit not set. 1935 */ 1936 KASSERT(prot != VM_PROT_ALL); 1937 1938 /* 1939 * If there is no protection, this is equivalent to 1940 * remove the pmap from the pmap. 1941 */ 1942 if ((prot & VM_PROT_READ) == 0) { 1943 pmap_remove(pm, va, endva); 1944 return; 1945 } 1946 1947 msr = pmap_interrupts_off(); 1948 for (; va < endva; va += PAGE_SIZE) { 1949 pvo = pmap_pvo_find_va(pm, va, &pteidx); 1950 if (pvo == NULL) 1951 continue; 1952 PMAP_PVO_CHECK(pvo); /* sanity check */ 1953 1954 /* 1955 * Revoke executable if asked to do so. 1956 */ 1957 if ((prot & VM_PROT_EXECUTE) == 0) 1958 pvo_clear_exec(pvo); 1959 1960 #if 0 1961 /* 1962 * If the page is already read-only, no change 1963 * needs to be made. 1964 */ 1965 if ((pvo->pvo_pte.pte_lo & PTE_PP) == PTE_BR) 1966 continue; 1967 #endif 1968 /* 1969 * Grab the PTE pointer before we diddle with 1970 * the cached PTE copy. 1971 */ 1972 pt = pmap_pvo_to_pte(pvo, pteidx); 1973 /* 1974 * Change the protection of the page. 1975 */ 1976 pvo->pvo_pte.pte_lo &= ~PTE_PP; 1977 pvo->pvo_pte.pte_lo |= PTE_BR; 1978 1979 /* 1980 * If the PVO is in the page table, update 1981 * that pte at well. 1982 */ 1983 if (pt != NULL) { 1984 pmap_pte_change(pt, &pvo->pvo_pte, pvo->pvo_vaddr); 1985 PVO_WHERE(pvo, PMAP_PROTECT); 1986 PMAPCOUNT(ptes_changed); 1987 } 1988 1989 PMAP_PVO_CHECK(pvo); /* sanity check */ 1990 } 1991 pmap_interrupts_restore(msr); 1992 } 1993 1994 void 1995 pmap_unwire(pmap_t pm, vaddr_t va) 1996 { 1997 struct pvo_entry *pvo; 1998 register_t msr; 1999 2000 msr = pmap_interrupts_off(); 2001 pvo = pmap_pvo_find_va(pm, va, NULL); 2002 if (pvo != NULL) { 2003 if (pvo->pvo_vaddr & PVO_WIRED) { 2004 pvo->pvo_vaddr &= ~PVO_WIRED; 2005 pm->pm_stats.wired_count--; 2006 } 2007 PMAP_PVO_CHECK(pvo); /* sanity check */ 2008 } 2009 pmap_interrupts_restore(msr); 2010 } 2011 2012 /* 2013 * Lower the protection on the specified physical page. 2014 */ 2015 void 2016 pmap_page_protect(struct vm_page *pg, vm_prot_t prot) 2017 { 2018 struct pvo_head *pvo_head; 2019 struct pvo_entry *pvo, *next_pvo; 2020 volatile struct pte *pt; 2021 register_t msr; 2022 2023 KASSERT(prot != VM_PROT_ALL); 2024 msr = pmap_interrupts_off(); 2025 2026 /* 2027 * When UVM reuses a page, it does a pmap_page_protect with 2028 * VM_PROT_NONE. At that point, we can clear the exec flag 2029 * since we know the page will have different contents. 2030 */ 2031 if ((prot & VM_PROT_READ) == 0) { 2032 DPRINTFN(EXEC, ("[pmap_page_protect: %#lx: clear-exec]\n", 2033 pg->phys_addr)); 2034 if (pmap_attr_fetch(pg) & PTE_EXEC) { 2035 PMAPCOUNT(exec_uncached_page_protect); 2036 pmap_attr_clear(pg, PTE_EXEC); 2037 } 2038 } 2039 2040 pvo_head = vm_page_to_pvoh(pg); 2041 for (pvo = LIST_FIRST(pvo_head); pvo != NULL; pvo = next_pvo) { 2042 next_pvo = LIST_NEXT(pvo, pvo_vlink); 2043 PMAP_PVO_CHECK(pvo); /* sanity check */ 2044 2045 /* 2046 * Downgrading to no mapping at all, we just remove the entry. 2047 */ 2048 if ((prot & VM_PROT_READ) == 0) { 2049 pmap_pvo_remove(pvo, -1); 2050 continue; 2051 } 2052 2053 /* 2054 * If EXEC permission is being revoked, just clear the 2055 * flag in the PVO. 2056 */ 2057 if ((prot & VM_PROT_EXECUTE) == 0) 2058 pvo_clear_exec(pvo); 2059 2060 /* 2061 * If this entry is already RO, don't diddle with the 2062 * page table. 2063 */ 2064 if ((pvo->pvo_pte.pte_lo & PTE_PP) == PTE_BR) { 2065 PMAP_PVO_CHECK(pvo); 2066 continue; 2067 } 2068 2069 /* 2070 * Grab the PTE before the we diddle the bits so 2071 * pvo_to_pte can verify the pte contents are as 2072 * expected. 2073 */ 2074 pt = pmap_pvo_to_pte(pvo, -1); 2075 pvo->pvo_pte.pte_lo &= ~PTE_PP; 2076 pvo->pvo_pte.pte_lo |= PTE_BR; 2077 if (pt != NULL) { 2078 pmap_pte_change(pt, &pvo->pvo_pte, pvo->pvo_vaddr); 2079 PVO_WHERE(pvo, PMAP_PAGE_PROTECT); 2080 PMAPCOUNT(ptes_changed); 2081 } 2082 PMAP_PVO_CHECK(pvo); /* sanity check */ 2083 } 2084 pmap_interrupts_restore(msr); 2085 } 2086 2087 /* 2088 * Activate the address space for the specified process. If the process 2089 * is the current process, load the new MMU context. 2090 */ 2091 void 2092 pmap_activate(struct lwp *l) 2093 { 2094 struct pcb *pcb = &l->l_addr->u_pcb; 2095 pmap_t pmap = l->l_proc->p_vmspace->vm_map.pmap; 2096 2097 DPRINTFN(ACTIVATE, 2098 ("pmap_activate: lwp %p (curlwp %p)\n", l, curlwp)); 2099 2100 /* 2101 * XXX Normally performed in cpu_fork(). 2102 */ 2103 pcb->pcb_pm = pmap; 2104 2105 /* 2106 * In theory, the SR registers need only be valid on return 2107 * to user space wait to do them there. 2108 */ 2109 if (l == curlwp) { 2110 /* Store pointer to new current pmap. */ 2111 curpm = pmap; 2112 } 2113 } 2114 2115 /* 2116 * Deactivate the specified process's address space. 2117 */ 2118 void 2119 pmap_deactivate(struct lwp *l) 2120 { 2121 } 2122 2123 boolean_t 2124 pmap_query_bit(struct vm_page *pg, int ptebit) 2125 { 2126 struct pvo_entry *pvo; 2127 volatile struct pte *pt; 2128 register_t msr; 2129 2130 if (pmap_attr_fetch(pg) & ptebit) 2131 return TRUE; 2132 2133 msr = pmap_interrupts_off(); 2134 LIST_FOREACH(pvo, vm_page_to_pvoh(pg), pvo_vlink) { 2135 PMAP_PVO_CHECK(pvo); /* sanity check */ 2136 /* 2137 * See if we saved the bit off. If so cache, it and return 2138 * success. 2139 */ 2140 if (pvo->pvo_pte.pte_lo & ptebit) { 2141 pmap_attr_save(pg, ptebit); 2142 PMAP_PVO_CHECK(pvo); /* sanity check */ 2143 pmap_interrupts_restore(msr); 2144 return TRUE; 2145 } 2146 } 2147 /* 2148 * No luck, now go thru the hard part of looking at the ptes 2149 * themselves. Sync so any pending REF/CHG bits are flushed 2150 * to the PTEs. 2151 */ 2152 SYNC(); 2153 LIST_FOREACH(pvo, vm_page_to_pvoh(pg), pvo_vlink) { 2154 PMAP_PVO_CHECK(pvo); /* sanity check */ 2155 /* 2156 * See if this pvo have a valid PTE. If so, fetch the 2157 * REF/CHG bits from the valid PTE. If the appropriate 2158 * ptebit is set, cache, it and return success. 2159 */ 2160 pt = pmap_pvo_to_pte(pvo, -1); 2161 if (pt != NULL) { 2162 pmap_pte_synch(pt, &pvo->pvo_pte); 2163 if (pvo->pvo_pte.pte_lo & ptebit) { 2164 pmap_attr_save(pg, ptebit); 2165 PMAP_PVO_CHECK(pvo); /* sanity check */ 2166 pmap_interrupts_restore(msr); 2167 return TRUE; 2168 } 2169 } 2170 } 2171 pmap_interrupts_restore(msr); 2172 return FALSE; 2173 } 2174 2175 boolean_t 2176 pmap_clear_bit(struct vm_page *pg, int ptebit) 2177 { 2178 struct pvo_head *pvoh = vm_page_to_pvoh(pg); 2179 struct pvo_entry *pvo; 2180 volatile struct pte *pt; 2181 register_t msr; 2182 int rv = 0; 2183 2184 msr = pmap_interrupts_off(); 2185 2186 /* 2187 * Fetch the cache value 2188 */ 2189 rv |= pmap_attr_fetch(pg); 2190 2191 /* 2192 * Clear the cached value. 2193 */ 2194 pmap_attr_clear(pg, ptebit); 2195 2196 /* 2197 * Sync so any pending REF/CHG bits are flushed to the PTEs (so we 2198 * can reset the right ones). Note that since the pvo entries and 2199 * list heads are accessed via BAT0 and are never placed in the 2200 * page table, we don't have to worry about further accesses setting 2201 * the REF/CHG bits. 2202 */ 2203 SYNC(); 2204 2205 /* 2206 * For each pvo entry, clear pvo's ptebit. If this pvo have a 2207 * valid PTE. If so, clear the ptebit from the valid PTE. 2208 */ 2209 LIST_FOREACH(pvo, pvoh, pvo_vlink) { 2210 PMAP_PVO_CHECK(pvo); /* sanity check */ 2211 pt = pmap_pvo_to_pte(pvo, -1); 2212 if (pt != NULL) { 2213 /* 2214 * Only sync the PTE if the bit we are looking 2215 * for is not already set. 2216 */ 2217 if ((pvo->pvo_pte.pte_lo & ptebit) == 0) 2218 pmap_pte_synch(pt, &pvo->pvo_pte); 2219 /* 2220 * If the bit we are looking for was already set, 2221 * clear that bit in the pte. 2222 */ 2223 if (pvo->pvo_pte.pte_lo & ptebit) 2224 pmap_pte_clear(pt, PVO_VADDR(pvo), ptebit); 2225 } 2226 rv |= pvo->pvo_pte.pte_lo & (PTE_CHG|PTE_REF); 2227 pvo->pvo_pte.pte_lo &= ~ptebit; 2228 PMAP_PVO_CHECK(pvo); /* sanity check */ 2229 } 2230 pmap_interrupts_restore(msr); 2231 2232 /* 2233 * If we are clearing the modify bit and this page was marked EXEC 2234 * and the user of the page thinks the page was modified, then we 2235 * need to clean it from the icache if it's mapped or clear the EXEC 2236 * bit if it's not mapped. The page itself might not have the CHG 2237 * bit set if the modification was done via DMA to the page. 2238 */ 2239 if ((ptebit & PTE_CHG) && (rv & PTE_EXEC)) { 2240 if (LIST_EMPTY(pvoh)) { 2241 DPRINTFN(EXEC, ("[pmap_clear_bit: %#lx: clear-exec]\n", 2242 pg->phys_addr)); 2243 pmap_attr_clear(pg, PTE_EXEC); 2244 PMAPCOUNT(exec_uncached_clear_modify); 2245 } else { 2246 DPRINTFN(EXEC, ("[pmap_clear_bit: %#lx: syncicache]\n", 2247 pg->phys_addr)); 2248 pmap_syncicache(pg->phys_addr, PAGE_SIZE); 2249 PMAPCOUNT(exec_synced_clear_modify); 2250 } 2251 } 2252 return (rv & ptebit) != 0; 2253 } 2254 2255 void 2256 pmap_procwr(struct proc *p, vaddr_t va, size_t len) 2257 { 2258 struct pvo_entry *pvo; 2259 size_t offset = va & ADDR_POFF; 2260 int s; 2261 2262 s = splvm(); 2263 while (len > 0) { 2264 size_t seglen = PAGE_SIZE - offset; 2265 if (seglen > len) 2266 seglen = len; 2267 pvo = pmap_pvo_find_va(p->p_vmspace->vm_map.pmap, va, NULL); 2268 if (pvo != NULL && PVO_ISEXECUTABLE(pvo)) { 2269 pmap_syncicache( 2270 (pvo->pvo_pte.pte_lo & PTE_RPGN) | offset, seglen); 2271 PMAP_PVO_CHECK(pvo); 2272 } 2273 va += seglen; 2274 len -= seglen; 2275 offset = 0; 2276 } 2277 splx(s); 2278 } 2279 2280 #if defined(DEBUG) || defined(PMAPCHECK) || defined(DDB) 2281 void 2282 pmap_pte_print(volatile struct pte *pt) 2283 { 2284 printf("PTE %p: ", pt); 2285 /* High word: */ 2286 printf("0x%08lx: [", pt->pte_hi); 2287 printf("%c ", (pt->pte_hi & PTE_VALID) ? 'v' : 'i'); 2288 printf("%c ", (pt->pte_hi & PTE_HID) ? 'h' : '-'); 2289 printf("0x%06lx 0x%02lx", 2290 (pt->pte_hi &~ PTE_VALID)>>PTE_VSID_SHFT, 2291 pt->pte_hi & PTE_API); 2292 printf(" (va 0x%08lx)] ", pmap_pte_to_va(pt)); 2293 /* Low word: */ 2294 printf(" 0x%08lx: [", pt->pte_lo); 2295 printf("0x%05lx... ", pt->pte_lo >> 12); 2296 printf("%c ", (pt->pte_lo & PTE_REF) ? 'r' : 'u'); 2297 printf("%c ", (pt->pte_lo & PTE_CHG) ? 'c' : 'n'); 2298 printf("%c", (pt->pte_lo & PTE_W) ? 'w' : '.'); 2299 printf("%c", (pt->pte_lo & PTE_I) ? 'i' : '.'); 2300 printf("%c", (pt->pte_lo & PTE_M) ? 'm' : '.'); 2301 printf("%c ", (pt->pte_lo & PTE_G) ? 'g' : '.'); 2302 switch (pt->pte_lo & PTE_PP) { 2303 case PTE_BR: printf("br]\n"); break; 2304 case PTE_BW: printf("bw]\n"); break; 2305 case PTE_SO: printf("so]\n"); break; 2306 case PTE_SW: printf("sw]\n"); break; 2307 } 2308 } 2309 #endif 2310 2311 #if defined(DDB) 2312 void 2313 pmap_pteg_check(void) 2314 { 2315 volatile struct pte *pt; 2316 int i; 2317 int ptegidx; 2318 u_int p_valid = 0; 2319 u_int s_valid = 0; 2320 u_int invalid = 0; 2321 2322 for (ptegidx = 0; ptegidx < pmap_pteg_cnt; ptegidx++) { 2323 for (pt = pmap_pteg_table[ptegidx].pt, i = 8; --i >= 0; pt++) { 2324 if (pt->pte_hi & PTE_VALID) { 2325 if (pt->pte_hi & PTE_HID) 2326 s_valid++; 2327 else 2328 p_valid++; 2329 } else 2330 invalid++; 2331 } 2332 } 2333 printf("pteg_check: v(p) %#x (%d), v(s) %#x (%d), i %#x (%d)\n", 2334 p_valid, p_valid, s_valid, s_valid, 2335 invalid, invalid); 2336 } 2337 2338 void 2339 pmap_print_mmuregs(void) 2340 { 2341 int i; 2342 u_int cpuvers; 2343 vaddr_t addr; 2344 register_t soft_sr[16]; 2345 struct bat soft_ibat[4]; 2346 struct bat soft_dbat[4]; 2347 register_t sdr1; 2348 2349 cpuvers = MFPVR() >> 16; 2350 2351 __asm __volatile ("mfsdr1 %0" : "=r"(sdr1)); 2352 2353 addr = 0; 2354 for (i=0; i<16; i++) { 2355 soft_sr[i] = MFSRIN(addr); 2356 addr += (1 << ADDR_SR_SHFT); 2357 } 2358 2359 /* read iBAT (601: uBAT) registers */ 2360 __asm __volatile ("mfibatu %0,0" : "=r"(soft_ibat[0].batu)); 2361 __asm __volatile ("mfibatl %0,0" : "=r"(soft_ibat[0].batl)); 2362 __asm __volatile ("mfibatu %0,1" : "=r"(soft_ibat[1].batu)); 2363 __asm __volatile ("mfibatl %0,1" : "=r"(soft_ibat[1].batl)); 2364 __asm __volatile ("mfibatu %0,2" : "=r"(soft_ibat[2].batu)); 2365 __asm __volatile ("mfibatl %0,2" : "=r"(soft_ibat[2].batl)); 2366 __asm __volatile ("mfibatu %0,3" : "=r"(soft_ibat[3].batu)); 2367 __asm __volatile ("mfibatl %0,3" : "=r"(soft_ibat[3].batl)); 2368 2369 2370 if (cpuvers != MPC601) { 2371 /* read dBAT registers */ 2372 __asm __volatile ("mfdbatu %0,0" : "=r"(soft_dbat[0].batu)); 2373 __asm __volatile ("mfdbatl %0,0" : "=r"(soft_dbat[0].batl)); 2374 __asm __volatile ("mfdbatu %0,1" : "=r"(soft_dbat[1].batu)); 2375 __asm __volatile ("mfdbatl %0,1" : "=r"(soft_dbat[1].batl)); 2376 __asm __volatile ("mfdbatu %0,2" : "=r"(soft_dbat[2].batu)); 2377 __asm __volatile ("mfdbatl %0,2" : "=r"(soft_dbat[2].batl)); 2378 __asm __volatile ("mfdbatu %0,3" : "=r"(soft_dbat[3].batu)); 2379 __asm __volatile ("mfdbatl %0,3" : "=r"(soft_dbat[3].batl)); 2380 } 2381 2382 printf("SDR1:\t%#lx\n", sdr1); 2383 printf("SR[]:\t"); 2384 for (i=0; i<4; i++) 2385 printf("0x%08lx, ", soft_sr[i]); 2386 printf("\n\t"); 2387 for ( ; i<8; i++) 2388 printf("0x%08lx, ", soft_sr[i]); 2389 printf("\n\t"); 2390 for ( ; i<12; i++) 2391 printf("0x%08lx, ", soft_sr[i]); 2392 printf("\n\t"); 2393 for ( ; i<16; i++) 2394 printf("0x%08lx, ", soft_sr[i]); 2395 printf("\n"); 2396 2397 printf("%cBAT[]:\t", cpuvers == MPC601 ? 'u' : 'i'); 2398 for (i=0; i<4; i++) { 2399 printf("0x%08lx 0x%08lx, ", 2400 soft_ibat[i].batu, soft_ibat[i].batl); 2401 if (i == 1) 2402 printf("\n\t"); 2403 } 2404 if (cpuvers != MPC601) { 2405 printf("\ndBAT[]:\t"); 2406 for (i=0; i<4; i++) { 2407 printf("0x%08lx 0x%08lx, ", 2408 soft_dbat[i].batu, soft_dbat[i].batl); 2409 if (i == 1) 2410 printf("\n\t"); 2411 } 2412 } 2413 printf("\n"); 2414 } 2415 2416 void 2417 pmap_print_pte(pmap_t pm, vaddr_t va) 2418 { 2419 struct pvo_entry *pvo; 2420 volatile struct pte *pt; 2421 int pteidx; 2422 2423 pvo = pmap_pvo_find_va(pm, va, &pteidx); 2424 if (pvo != NULL) { 2425 pt = pmap_pvo_to_pte(pvo, pteidx); 2426 if (pt != NULL) { 2427 printf("VA %#lx -> %p -> %s %#lx, %#lx\n", 2428 va, pt, 2429 pt->pte_hi & PTE_HID ? "(sec)" : "(pri)", 2430 pt->pte_hi, pt->pte_lo); 2431 } else { 2432 printf("No valid PTE found\n"); 2433 } 2434 } else { 2435 printf("Address not in pmap\n"); 2436 } 2437 } 2438 2439 void 2440 pmap_pteg_dist(void) 2441 { 2442 struct pvo_entry *pvo; 2443 int ptegidx; 2444 int depth; 2445 int max_depth = 0; 2446 unsigned int depths[64]; 2447 2448 memset(depths, 0, sizeof(depths)); 2449 for (ptegidx = 0; ptegidx < pmap_pteg_cnt; ptegidx++) { 2450 depth = 0; 2451 TAILQ_FOREACH(pvo, &pmap_pvo_table[ptegidx], pvo_olink) { 2452 depth++; 2453 } 2454 if (depth > max_depth) 2455 max_depth = depth; 2456 if (depth > 63) 2457 depth = 63; 2458 depths[depth]++; 2459 } 2460 2461 for (depth = 0; depth < 64; depth++) { 2462 printf(" [%2d]: %8u", depth, depths[depth]); 2463 if ((depth & 3) == 3) 2464 printf("\n"); 2465 if (depth == max_depth) 2466 break; 2467 } 2468 if ((depth & 3) != 3) 2469 printf("\n"); 2470 printf("Max depth found was %d\n", max_depth); 2471 } 2472 #endif /* DEBUG */ 2473 2474 #if defined(PMAPCHECK) || defined(DEBUG) 2475 void 2476 pmap_pvo_verify(void) 2477 { 2478 int ptegidx; 2479 int s; 2480 2481 s = splvm(); 2482 for (ptegidx = 0; ptegidx < pmap_pteg_cnt; ptegidx++) { 2483 struct pvo_entry *pvo; 2484 TAILQ_FOREACH(pvo, &pmap_pvo_table[ptegidx], pvo_olink) { 2485 if ((uintptr_t) pvo >= SEGMENT_LENGTH) 2486 panic("pmap_pvo_verify: invalid pvo %p " 2487 "on list %#x", pvo, ptegidx); 2488 pmap_pvo_check(pvo); 2489 } 2490 } 2491 splx(s); 2492 } 2493 #endif /* PMAPCHECK */ 2494 2495 2496 void * 2497 pmap_pool_ualloc(struct pool *pp, int flags) 2498 { 2499 struct pvo_page *pvop; 2500 2501 pvop = SIMPLEQ_FIRST(&pmap_upvop_head); 2502 if (pvop != NULL) { 2503 pmap_upvop_free--; 2504 SIMPLEQ_REMOVE_HEAD(&pmap_upvop_head, pvop_link); 2505 return pvop; 2506 } 2507 if (uvm.page_init_done != TRUE) { 2508 return (void *) uvm_pageboot_alloc(PAGE_SIZE); 2509 } 2510 return pmap_pool_malloc(pp, flags); 2511 } 2512 2513 void * 2514 pmap_pool_malloc(struct pool *pp, int flags) 2515 { 2516 struct pvo_page *pvop; 2517 struct vm_page *pg; 2518 2519 pvop = SIMPLEQ_FIRST(&pmap_mpvop_head); 2520 if (pvop != NULL) { 2521 pmap_mpvop_free--; 2522 SIMPLEQ_REMOVE_HEAD(&pmap_mpvop_head, pvop_link); 2523 return pvop; 2524 } 2525 again: 2526 pg = uvm_pagealloc_strat(NULL, 0, NULL, UVM_PGA_USERESERVE, 2527 UVM_PGA_STRAT_ONLY, VM_FREELIST_FIRST256); 2528 if (__predict_false(pg == NULL)) { 2529 if (flags & PR_WAITOK) { 2530 uvm_wait("plpg"); 2531 goto again; 2532 } else { 2533 return (0); 2534 } 2535 } 2536 return (void *) VM_PAGE_TO_PHYS(pg); 2537 } 2538 2539 void 2540 pmap_pool_ufree(struct pool *pp, void *va) 2541 { 2542 struct pvo_page *pvop; 2543 #if 0 2544 if (PHYS_TO_VM_PAGE((paddr_t) va) != NULL) { 2545 pmap_pool_mfree(va, size, tag); 2546 return; 2547 } 2548 #endif 2549 pvop = va; 2550 SIMPLEQ_INSERT_HEAD(&pmap_upvop_head, pvop, pvop_link); 2551 pmap_upvop_free++; 2552 if (pmap_upvop_free > pmap_upvop_maxfree) 2553 pmap_upvop_maxfree = pmap_upvop_free; 2554 } 2555 2556 void 2557 pmap_pool_mfree(struct pool *pp, void *va) 2558 { 2559 struct pvo_page *pvop; 2560 2561 pvop = va; 2562 SIMPLEQ_INSERT_HEAD(&pmap_mpvop_head, pvop, pvop_link); 2563 pmap_mpvop_free++; 2564 if (pmap_mpvop_free > pmap_mpvop_maxfree) 2565 pmap_mpvop_maxfree = pmap_mpvop_free; 2566 #if 0 2567 uvm_pagefree(PHYS_TO_VM_PAGE((paddr_t) va)); 2568 #endif 2569 } 2570 2571 /* 2572 * This routine in bootstraping to steal to-be-managed memory (which will 2573 * then be unmanaged). We use it to grab from the first 256MB for our 2574 * pmap needs and above 256MB for other stuff. 2575 */ 2576 vaddr_t 2577 pmap_steal_memory(vsize_t vsize, vaddr_t *vstartp, vaddr_t *vendp) 2578 { 2579 vsize_t size; 2580 vaddr_t va; 2581 paddr_t pa = 0; 2582 int npgs, bank; 2583 struct vm_physseg *ps; 2584 2585 if (uvm.page_init_done == TRUE) 2586 panic("pmap_steal_memory: called _after_ bootstrap"); 2587 2588 *vstartp = VM_MIN_KERNEL_ADDRESS; 2589 *vendp = VM_MAX_KERNEL_ADDRESS; 2590 2591 size = round_page(vsize); 2592 npgs = atop(size); 2593 2594 /* 2595 * PA 0 will never be among those given to UVM so we can use it 2596 * to indicate we couldn't steal any memory. 2597 */ 2598 for (ps = vm_physmem, bank = 0; bank < vm_nphysseg; bank++, ps++) { 2599 if (ps->free_list == VM_FREELIST_FIRST256 && 2600 ps->avail_end - ps->avail_start >= npgs) { 2601 pa = ptoa(ps->avail_start); 2602 break; 2603 } 2604 } 2605 2606 if (pa == 0) 2607 panic("pmap_steal_memory: no approriate memory to steal!"); 2608 2609 ps->avail_start += npgs; 2610 ps->start += npgs; 2611 2612 /* 2613 * If we've used up all the pages in the segment, remove it and 2614 * compact the list. 2615 */ 2616 if (ps->avail_start == ps->end) { 2617 /* 2618 * If this was the last one, then a very bad thing has occurred 2619 */ 2620 if (--vm_nphysseg == 0) 2621 panic("pmap_steal_memory: out of memory!"); 2622 2623 printf("pmap_steal_memory: consumed bank %d\n", bank); 2624 for (; bank < vm_nphysseg; bank++, ps++) { 2625 ps[0] = ps[1]; 2626 } 2627 } 2628 2629 va = (vaddr_t) pa; 2630 memset((caddr_t) va, 0, size); 2631 pmap_pages_stolen += npgs; 2632 #ifdef DEBUG 2633 if (pmapdebug && npgs > 1) { 2634 u_int cnt = 0; 2635 for (bank = 0, ps = vm_physmem; bank < vm_nphysseg; bank++, ps++) 2636 cnt += ps->avail_end - ps->avail_start; 2637 printf("pmap_steal_memory: stole %u (total %u) pages (%u left)\n", 2638 npgs, pmap_pages_stolen, cnt); 2639 } 2640 #endif 2641 2642 return va; 2643 } 2644 2645 /* 2646 * Find a chuck of memory with right size and alignment. 2647 */ 2648 void * 2649 pmap_boot_find_memory(psize_t size, psize_t alignment, int at_end) 2650 { 2651 struct mem_region *mp; 2652 paddr_t s, e; 2653 int i, j; 2654 2655 size = round_page(size); 2656 2657 DPRINTFN(BOOT, 2658 ("pmap_boot_find_memory: size=%lx, alignment=%lx, at_end=%d", 2659 size, alignment, at_end)); 2660 2661 if (alignment < PAGE_SIZE || (alignment & (alignment-1)) != 0) 2662 panic("pmap_boot_find_memory: invalid alignment %lx", 2663 alignment); 2664 2665 if (at_end) { 2666 if (alignment != PAGE_SIZE) 2667 panic("pmap_boot_find_memory: invalid ending " 2668 "alignment %lx", alignment); 2669 2670 for (mp = &avail[avail_cnt-1]; mp >= avail; mp--) { 2671 s = mp->start + mp->size - size; 2672 if (s >= mp->start && mp->size >= size) { 2673 DPRINTFN(BOOT,(": %lx\n", s)); 2674 DPRINTFN(BOOT, 2675 ("pmap_boot_find_memory: b-avail[%d] start " 2676 "0x%lx size 0x%lx\n", mp - avail, 2677 mp->start, mp->size)); 2678 mp->size -= size; 2679 DPRINTFN(BOOT, 2680 ("pmap_boot_find_memory: a-avail[%d] start " 2681 "0x%lx size 0x%lx\n", mp - avail, 2682 mp->start, mp->size)); 2683 return (void *) s; 2684 } 2685 } 2686 panic("pmap_boot_find_memory: no available memory"); 2687 } 2688 2689 for (mp = avail, i = 0; i < avail_cnt; i++, mp++) { 2690 s = (mp->start + alignment - 1) & ~(alignment-1); 2691 e = s + size; 2692 2693 /* 2694 * Is the calculated region entirely within the region? 2695 */ 2696 if (s < mp->start || e > mp->start + mp->size) 2697 continue; 2698 2699 DPRINTFN(BOOT,(": %lx\n", s)); 2700 if (s == mp->start) { 2701 /* 2702 * If the block starts at the beginning of region, 2703 * adjust the size & start. (the region may now be 2704 * zero in length) 2705 */ 2706 DPRINTFN(BOOT, 2707 ("pmap_boot_find_memory: b-avail[%d] start " 2708 "0x%lx size 0x%lx\n", i, mp->start, mp->size)); 2709 mp->start += size; 2710 mp->size -= size; 2711 DPRINTFN(BOOT, 2712 ("pmap_boot_find_memory: a-avail[%d] start " 2713 "0x%lx size 0x%lx\n", i, mp->start, mp->size)); 2714 } else if (e == mp->start + mp->size) { 2715 /* 2716 * If the block starts at the beginning of region, 2717 * adjust only the size. 2718 */ 2719 DPRINTFN(BOOT, 2720 ("pmap_boot_find_memory: b-avail[%d] start " 2721 "0x%lx size 0x%lx\n", i, mp->start, mp->size)); 2722 mp->size -= size; 2723 DPRINTFN(BOOT, 2724 ("pmap_boot_find_memory: a-avail[%d] start " 2725 "0x%lx size 0x%lx\n", i, mp->start, mp->size)); 2726 } else { 2727 /* 2728 * Block is in the middle of the region, so we 2729 * have to split it in two. 2730 */ 2731 for (j = avail_cnt; j > i + 1; j--) { 2732 avail[j] = avail[j-1]; 2733 } 2734 DPRINTFN(BOOT, 2735 ("pmap_boot_find_memory: b-avail[%d] start " 2736 "0x%lx size 0x%lx\n", i, mp->start, mp->size)); 2737 mp[1].start = e; 2738 mp[1].size = mp[0].start + mp[0].size - e; 2739 mp[0].size = s - mp[0].start; 2740 avail_cnt++; 2741 for (; i < avail_cnt; i++) { 2742 DPRINTFN(BOOT, 2743 ("pmap_boot_find_memory: a-avail[%d] " 2744 "start 0x%lx size 0x%lx\n", i, 2745 avail[i].start, avail[i].size)); 2746 } 2747 } 2748 return (void *) s; 2749 } 2750 panic("pmap_boot_find_memory: not enough memory for " 2751 "%lx/%lx allocation?", size, alignment); 2752 } 2753 2754 /* 2755 * This is not part of the defined PMAP interface and is specific to the 2756 * PowerPC architecture. This is called during initppc, before the system 2757 * is really initialized. 2758 */ 2759 void 2760 pmap_bootstrap(paddr_t kernelstart, paddr_t kernelend) 2761 { 2762 struct mem_region *mp, tmp; 2763 paddr_t s, e; 2764 psize_t size; 2765 int i, j; 2766 2767 /* 2768 * Get memory. 2769 */ 2770 mem_regions(&mem, &avail); 2771 #if defined(DEBUG) 2772 if (pmapdebug & PMAPDEBUG_BOOT) { 2773 printf("pmap_bootstrap: memory configuration:\n"); 2774 for (mp = mem; mp->size; mp++) { 2775 printf("pmap_bootstrap: mem start 0x%lx size 0x%lx\n", 2776 mp->start, mp->size); 2777 } 2778 for (mp = avail; mp->size; mp++) { 2779 printf("pmap_bootstrap: avail start 0x%lx size 0x%lx\n", 2780 mp->start, mp->size); 2781 } 2782 } 2783 #endif 2784 2785 /* 2786 * Find out how much physical memory we have and in how many chunks. 2787 */ 2788 for (mem_cnt = 0, mp = mem; mp->size; mp++) { 2789 if (mp->start >= pmap_memlimit) 2790 continue; 2791 if (mp->start + mp->size > pmap_memlimit) { 2792 size = pmap_memlimit - mp->start; 2793 physmem += btoc(size); 2794 } else { 2795 physmem += btoc(mp->size); 2796 } 2797 mem_cnt++; 2798 } 2799 2800 /* 2801 * Count the number of available entries. 2802 */ 2803 for (avail_cnt = 0, mp = avail; mp->size; mp++) 2804 avail_cnt++; 2805 2806 /* 2807 * Page align all regions. 2808 */ 2809 kernelstart = trunc_page(kernelstart); 2810 kernelend = round_page(kernelend); 2811 for (mp = avail, i = 0; i < avail_cnt; i++, mp++) { 2812 s = round_page(mp->start); 2813 mp->size -= (s - mp->start); 2814 mp->size = trunc_page(mp->size); 2815 mp->start = s; 2816 e = mp->start + mp->size; 2817 2818 DPRINTFN(BOOT, 2819 ("pmap_bootstrap: b-avail[%d] start 0x%lx size 0x%lx\n", 2820 i, mp->start, mp->size)); 2821 2822 /* 2823 * Don't allow the end to run beyond our artificial limit 2824 */ 2825 if (e > pmap_memlimit) 2826 e = pmap_memlimit; 2827 2828 /* 2829 * Is this region empty or strange? skip it. 2830 */ 2831 if (e <= s) { 2832 mp->start = 0; 2833 mp->size = 0; 2834 continue; 2835 } 2836 2837 /* 2838 * Does this overlap the beginning of kernel? 2839 * Does extend past the end of the kernel? 2840 */ 2841 else if (s < kernelstart && e > kernelstart) { 2842 if (e > kernelend) { 2843 avail[avail_cnt].start = kernelend; 2844 avail[avail_cnt].size = e - kernelend; 2845 avail_cnt++; 2846 } 2847 mp->size = kernelstart - s; 2848 } 2849 /* 2850 * Check whether this region overlaps the end of the kernel. 2851 */ 2852 else if (s < kernelend && e > kernelend) { 2853 mp->start = kernelend; 2854 mp->size = e - kernelend; 2855 } 2856 /* 2857 * Look whether this regions is completely inside the kernel. 2858 * Nuke it if it does. 2859 */ 2860 else if (s >= kernelstart && e <= kernelend) { 2861 mp->start = 0; 2862 mp->size = 0; 2863 } 2864 /* 2865 * If the user imposed a memory limit, enforce it. 2866 */ 2867 else if (s >= pmap_memlimit) { 2868 mp->start = -PAGE_SIZE; /* let's know why */ 2869 mp->size = 0; 2870 } 2871 else { 2872 mp->start = s; 2873 mp->size = e - s; 2874 } 2875 DPRINTFN(BOOT, 2876 ("pmap_bootstrap: a-avail[%d] start 0x%lx size 0x%lx\n", 2877 i, mp->start, mp->size)); 2878 } 2879 2880 /* 2881 * Move (and uncount) all the null return to the end. 2882 */ 2883 for (mp = avail, i = 0; i < avail_cnt; i++, mp++) { 2884 if (mp->size == 0) { 2885 tmp = avail[i]; 2886 avail[i] = avail[--avail_cnt]; 2887 avail[avail_cnt] = avail[i]; 2888 } 2889 } 2890 2891 /* 2892 * (Bubble)sort them into asecnding order. 2893 */ 2894 for (i = 0; i < avail_cnt; i++) { 2895 for (j = i + 1; j < avail_cnt; j++) { 2896 if (avail[i].start > avail[j].start) { 2897 tmp = avail[i]; 2898 avail[i] = avail[j]; 2899 avail[j] = tmp; 2900 } 2901 } 2902 } 2903 2904 /* 2905 * Make sure they don't overlap. 2906 */ 2907 for (mp = avail, i = 0; i < avail_cnt - 1; i++, mp++) { 2908 if (mp[0].start + mp[0].size > mp[1].start) { 2909 mp[0].size = mp[1].start - mp[0].start; 2910 } 2911 DPRINTFN(BOOT, 2912 ("pmap_bootstrap: avail[%d] start 0x%lx size 0x%lx\n", 2913 i, mp->start, mp->size)); 2914 } 2915 DPRINTFN(BOOT, 2916 ("pmap_bootstrap: avail[%d] start 0x%lx size 0x%lx\n", 2917 i, mp->start, mp->size)); 2918 2919 #ifdef PTEGCOUNT 2920 pmap_pteg_cnt = PTEGCOUNT; 2921 #else /* PTEGCOUNT */ 2922 pmap_pteg_cnt = 0x1000; 2923 2924 while (pmap_pteg_cnt < physmem) 2925 pmap_pteg_cnt <<= 1; 2926 2927 pmap_pteg_cnt >>= 1; 2928 #endif /* PTEGCOUNT */ 2929 2930 /* 2931 * Find suitably aligned memory for PTEG hash table. 2932 */ 2933 size = pmap_pteg_cnt * sizeof(struct pteg); 2934 pmap_pteg_table = pmap_boot_find_memory(size, size, 0); 2935 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 2936 if ( (uintptr_t) pmap_pteg_table + size > SEGMENT_LENGTH) 2937 panic("pmap_bootstrap: pmap_pteg_table end (%p + %lx) > 256MB", 2938 pmap_pteg_table, size); 2939 #endif 2940 2941 memset((void *)pmap_pteg_table, 0, pmap_pteg_cnt * sizeof(struct pteg)); 2942 pmap_pteg_mask = pmap_pteg_cnt - 1; 2943 2944 /* 2945 * We cannot do pmap_steal_memory here since UVM hasn't been loaded 2946 * with pages. So we just steal them before giving them to UVM. 2947 */ 2948 size = sizeof(pmap_pvo_table[0]) * pmap_pteg_cnt; 2949 pmap_pvo_table = pmap_boot_find_memory(size, PAGE_SIZE, 0); 2950 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 2951 if ( (uintptr_t) pmap_pvo_table + size > SEGMENT_LENGTH) 2952 panic("pmap_bootstrap: pmap_pvo_table end (%p + %lx) > 256MB", 2953 pmap_pvo_table, size); 2954 #endif 2955 2956 for (i = 0; i < pmap_pteg_cnt; i++) 2957 TAILQ_INIT(&pmap_pvo_table[i]); 2958 2959 #ifndef MSGBUFADDR 2960 /* 2961 * Allocate msgbuf in high memory. 2962 */ 2963 msgbuf_paddr = 2964 (paddr_t) pmap_boot_find_memory(MSGBUFSIZE, PAGE_SIZE, 1); 2965 #endif 2966 2967 #ifdef __HAVE_PMAP_PHYSSEG 2968 { 2969 u_int npgs = 0; 2970 for (i = 0, mp = avail; i < avail_cnt; i++, mp++) 2971 npgs += btoc(mp->size); 2972 size = (sizeof(struct pvo_head) + 1) * npgs; 2973 pmap_physseg.pvoh = pmap_boot_find_memory(size, PAGE_SIZE, 0); 2974 pmap_physseg.attrs = (char *) &pmap_physseg.pvoh[npgs]; 2975 #if defined(DIAGNOSTIC) || defined(DEBUG) || defined(PMAPCHECK) 2976 if ((uintptr_t)pmap_physseg.pvoh + size > SEGMENT_LENGTH) 2977 panic("pmap_bootstrap: PVO list end (%p + %lx) > 256MB", 2978 pmap_physseg.pvoh, size); 2979 #endif 2980 } 2981 #endif 2982 2983 for (mp = avail, i = 0; i < avail_cnt; mp++, i++) { 2984 paddr_t pfstart = atop(mp->start); 2985 paddr_t pfend = atop(mp->start + mp->size); 2986 if (mp->size == 0) 2987 continue; 2988 if (mp->start + mp->size <= SEGMENT_LENGTH) { 2989 uvm_page_physload(pfstart, pfend, pfstart, pfend, 2990 VM_FREELIST_FIRST256); 2991 } else if (mp->start >= SEGMENT_LENGTH) { 2992 uvm_page_physload(pfstart, pfend, pfstart, pfend, 2993 VM_FREELIST_DEFAULT); 2994 } else { 2995 pfend = atop(SEGMENT_LENGTH); 2996 uvm_page_physload(pfstart, pfend, pfstart, pfend, 2997 VM_FREELIST_FIRST256); 2998 pfstart = atop(SEGMENT_LENGTH); 2999 pfend = atop(mp->start + mp->size); 3000 uvm_page_physload(pfstart, pfend, pfstart, pfend, 3001 VM_FREELIST_DEFAULT); 3002 } 3003 } 3004 3005 /* 3006 * Make sure kernel vsid is allocated as well as VSID 0. 3007 */ 3008 pmap_vsid_bitmap[(KERNEL_VSIDBITS & (NPMAPS-1)) / VSID_NBPW] 3009 |= 1 << (KERNEL_VSIDBITS % VSID_NBPW); 3010 pmap_vsid_bitmap[0] |= 1; 3011 3012 /* 3013 * Initialize kernel pmap and hardware. 3014 */ 3015 for (i = 0; i < 16; i++) { 3016 pmap_kernel()->pm_sr[i] = EMPTY_SEGMENT; 3017 __asm __volatile ("mtsrin %0,%1" 3018 :: "r"(EMPTY_SEGMENT), "r"(i << ADDR_SR_SHFT)); 3019 } 3020 3021 pmap_kernel()->pm_sr[KERNEL_SR] = KERNEL_SEGMENT|SR_SUKEY|SR_PRKEY; 3022 __asm __volatile ("mtsr %0,%1" 3023 :: "n"(KERNEL_SR), "r"(KERNEL_SEGMENT)); 3024 #ifdef KERNEL2_SR 3025 pmap_kernel()->pm_sr[KERNEL2_SR] = KERNEL2_SEGMENT|SR_SUKEY|SR_PRKEY; 3026 __asm __volatile ("mtsr %0,%1" 3027 :: "n"(KERNEL2_SR), "r"(KERNEL2_SEGMENT)); 3028 #endif 3029 for (i = 0; i < 16; i++) { 3030 if (iosrtable[i] & SR601_T) { 3031 pmap_kernel()->pm_sr[i] = iosrtable[i]; 3032 __asm __volatile ("mtsrin %0,%1" 3033 :: "r"(iosrtable[i]), "r"(i << ADDR_SR_SHFT)); 3034 } 3035 } 3036 3037 __asm __volatile ("sync; mtsdr1 %0; isync" 3038 :: "r"((uintptr_t)pmap_pteg_table | (pmap_pteg_mask >> 10))); 3039 tlbia(); 3040 3041 #ifdef ALTIVEC 3042 pmap_use_altivec = cpu_altivec; 3043 #endif 3044 3045 #ifdef DEBUG 3046 if (pmapdebug & PMAPDEBUG_BOOT) { 3047 u_int cnt; 3048 int bank; 3049 char pbuf[9]; 3050 for (cnt = 0, bank = 0; bank < vm_nphysseg; bank++) { 3051 cnt += vm_physmem[bank].avail_end - vm_physmem[bank].avail_start; 3052 printf("pmap_bootstrap: vm_physmem[%d]=%#lx-%#lx/%#lx\n", 3053 bank, 3054 ptoa(vm_physmem[bank].avail_start), 3055 ptoa(vm_physmem[bank].avail_end), 3056 ptoa(vm_physmem[bank].avail_end - vm_physmem[bank].avail_start)); 3057 } 3058 format_bytes(pbuf, sizeof(pbuf), ptoa((u_int64_t) cnt)); 3059 printf("pmap_bootstrap: UVM memory = %s (%u pages)\n", 3060 pbuf, cnt); 3061 } 3062 #endif 3063 3064 pool_init(&pmap_upvo_pool, sizeof(struct pvo_entry), 3065 sizeof(struct pvo_entry), 0, 0, "pmap_upvopl", 3066 &pmap_pool_uallocator); 3067 3068 pool_setlowat(&pmap_upvo_pool, 252); 3069 3070 pool_init(&pmap_pool, sizeof(struct pmap), 3071 sizeof(void *), 0, 0, "pmap_pl", &pmap_pool_uallocator); 3072 } 3073
Indexes created Sat Sep 27 22:09:54 GMT 2025