1 /* $NetBSD: cpu.c,v 1.98 2014/02/12 23:24:09 dsl Exp $ */ 2 /* NetBSD: cpu.c,v 1.18 2004/02/20 17:35:01 yamt Exp */ 3 4 /*- 5 * Copyright (c) 2000 The NetBSD Foundation, Inc. 6 * Copyright (c) 2002, 2006, 2007 YAMAMOTO Takashi, 7 * All rights reserved. 8 * 9 * This code is derived from software contributed to The NetBSD Foundation 10 * by RedBack Networks Inc. 11 * 12 * Author: Bill Sommerfeld 13 * 14 * Redistribution and use in source and binary forms, with or without 15 * modification, are permitted provided that the following conditions 16 * are met: 17 * 1. Redistributions of source code must retain the above copyright 18 * notice, this list of conditions and the following disclaimer. 19 * 2. Redistributions in binary form must reproduce the above copyright 20 * notice, this list of conditions and the following disclaimer in the 21 * documentation and/or other materials provided with the distribution. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 24 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 25 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 26 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 27 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 28 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 29 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 30 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 31 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 32 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 33 * POSSIBILITY OF SUCH DAMAGE. 34 */ 35 36 /* 37 * Copyright (c) 1999 Stefan Grefen 38 * 39 * Redistribution and use in source and binary forms, with or without 40 * modification, are permitted provided that the following conditions 41 * are met: 42 * 1. Redistributions of source code must retain the above copyright 43 * notice, this list of conditions and the following disclaimer. 44 * 2. Redistributions in binary form must reproduce the above copyright 45 * notice, this list of conditions and the following disclaimer in the 46 * documentation and/or other materials provided with the distribution. 47 * 3. All advertising materials mentioning features or use of this software 48 * must display the following acknowledgement: 49 * This product includes software developed by the NetBSD 50 * Foundation, Inc. and its contributors. 51 * 4. Neither the name of The NetBSD Foundation nor the names of its 52 * contributors may be used to endorse or promote products derived 53 * from this software without specific prior written permission. 54 * 55 * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND ANY 56 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 57 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 58 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR AND CONTRIBUTORS BE LIABLE 59 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 60 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 61 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 62 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 63 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 64 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 65 * SUCH DAMAGE. 66 */ 67 68 #include <sys/cdefs.h> 69 __KERNEL_RCSID(0, "$NetBSD: cpu.c,v 1.98 2014/02/12 23:24:09 dsl Exp $"); 70 71 #include "opt_ddb.h" 72 #include "opt_multiprocessor.h" 73 #include "opt_mpbios.h" /* for MPDEBUG */ 74 #include "opt_mtrr.h" 75 #include "opt_xen.h" 76 77 #include "lapic.h" 78 #include "ioapic.h" 79 80 #include <sys/param.h> 81 #include <sys/proc.h> 82 #include <sys/systm.h> 83 #include <sys/device.h> 84 #include <sys/kmem.h> 85 #include <sys/cpu.h> 86 #include <sys/cpufreq.h> 87 #include <sys/atomic.h> 88 #include <sys/reboot.h> 89 #include <sys/idle.h> 90 91 #include <uvm/uvm.h> 92 93 #include <machine/cpufunc.h> 94 #include <machine/cpuvar.h> 95 #include <machine/pmap.h> 96 #include <machine/vmparam.h> 97 #include <machine/mpbiosvar.h> 98 #include <machine/pcb.h> 99 #include <machine/specialreg.h> 100 #include <machine/segments.h> 101 #include <machine/gdt.h> 102 #include <machine/mtrr.h> 103 #include <machine/pio.h> 104 105 #include <x86/fpu.h> 106 107 #include <xen/xen.h> 108 #include <xen/xen-public/vcpu.h> 109 #include <xen/vcpuvar.h> 110 111 #if NLAPIC > 0 112 #include <machine/apicvar.h> 113 #include <machine/i82489reg.h> 114 #include <machine/i82489var.h> 115 #endif 116 117 #include <dev/ic/mc146818reg.h> 118 #include <dev/isa/isareg.h> 119 120 static int cpu_match(device_t, cfdata_t, void *); 121 static void cpu_attach(device_t, device_t, void *); 122 static void cpu_defer(device_t); 123 static int cpu_rescan(device_t, const char *, const int *); 124 static void cpu_childdetached(device_t, device_t); 125 static int vcpu_match(device_t, cfdata_t, void *); 126 static void vcpu_attach(device_t, device_t, void *); 127 static void cpu_attach_common(device_t, device_t, void *); 128 void cpu_offline_md(void); 129 130 struct cpu_softc { 131 device_t sc_dev; /* device tree glue */ 132 struct cpu_info *sc_info; /* pointer to CPU info */ 133 bool sc_wasonline; 134 }; 135 136 int mp_cpu_start(struct cpu_info *, vaddr_t); 137 void mp_cpu_start_cleanup(struct cpu_info *); 138 const struct cpu_functions mp_cpu_funcs = { mp_cpu_start, NULL, 139 mp_cpu_start_cleanup }; 140 141 CFATTACH_DECL2_NEW(cpu, sizeof(struct cpu_softc), 142 cpu_match, cpu_attach, NULL, NULL, cpu_rescan, cpu_childdetached); 143 144 CFATTACH_DECL_NEW(vcpu, sizeof(struct cpu_softc), 145 vcpu_match, vcpu_attach, NULL, NULL); 146 147 /* 148 * Statically-allocated CPU info for the primary CPU (or the only 149 * CPU, on uniprocessors). The CPU info list is initialized to 150 * point at it. 151 */ 152 #ifdef TRAPLOG 153 #include <machine/tlog.h> 154 struct tlog tlog_primary; 155 #endif 156 struct cpu_info cpu_info_primary __aligned(CACHE_LINE_SIZE) = { 157 .ci_dev = 0, 158 .ci_self = &cpu_info_primary, 159 .ci_idepth = -1, 160 .ci_curlwp = &lwp0, 161 .ci_curldt = -1, 162 #ifdef TRAPLOG 163 .ci_tlog = &tlog_primary, 164 #endif 165 166 }; 167 struct cpu_info phycpu_info_primary __aligned(CACHE_LINE_SIZE) = { 168 .ci_dev = 0, 169 .ci_self = &phycpu_info_primary, 170 }; 171 172 struct cpu_info *cpu_info_list = &cpu_info_primary; 173 struct cpu_info *phycpu_info_list = &phycpu_info_primary; 174 175 uint32_t cpu_feature[5]; /* X86 CPUID feature bits 176 * [0] basic features %edx 177 * [1] basic features %ecx 178 * [2] extended features %edx 179 * [3] extended features %ecx 180 * [4] VIA padlock features 181 */ 182 183 bool x86_mp_online; 184 paddr_t mp_trampoline_paddr = MP_TRAMPOLINE; 185 186 #if defined(MULTIPROCESSOR) 187 void cpu_hatch(void *); 188 static void cpu_boot_secondary(struct cpu_info *ci); 189 static void cpu_start_secondary(struct cpu_info *ci); 190 #endif /* MULTIPROCESSOR */ 191 192 static int 193 cpu_match(device_t parent, cfdata_t match, void *aux) 194 { 195 196 return 1; 197 } 198 199 static void 200 cpu_attach(device_t parent, device_t self, void *aux) 201 { 202 struct cpu_softc *sc = device_private(self); 203 struct cpu_attach_args *caa = aux; 204 struct cpu_info *ci; 205 uintptr_t ptr; 206 static int nphycpu = 0; 207 208 sc->sc_dev = self; 209 210 /* 211 * If we're an Application Processor, allocate a cpu_info 212 * If we're the first attached CPU use the primary cpu_info, 213 * otherwise allocate a new one 214 */ 215 aprint_naive("\n"); 216 aprint_normal("\n"); 217 if (nphycpu > 0) { 218 struct cpu_info *tmp; 219 ptr = (uintptr_t)kmem_zalloc(sizeof(*ci) + CACHE_LINE_SIZE - 1, 220 KM_SLEEP); 221 ci = (struct cpu_info *)roundup2(ptr, CACHE_LINE_SIZE); 222 ci->ci_curldt = -1; 223 224 tmp = phycpu_info_list; 225 while (tmp->ci_next) 226 tmp = tmp->ci_next; 227 228 tmp->ci_next = ci; 229 } else { 230 ci = &phycpu_info_primary; 231 } 232 233 ci->ci_self = ci; 234 sc->sc_info = ci; 235 236 ci->ci_dev = self; 237 ci->ci_acpiid = caa->cpu_id; 238 ci->ci_cpuid = caa->cpu_number; 239 ci->ci_vcpu = NULL; 240 ci->ci_index = nphycpu++; 241 242 if (!pmf_device_register(self, NULL, NULL)) 243 aprint_error_dev(self, "couldn't establish power handler\n"); 244 245 (void)config_defer(self, cpu_defer); 246 } 247 248 static void 249 cpu_defer(device_t self) 250 { 251 cpu_rescan(self, NULL, NULL); 252 } 253 254 static int 255 cpu_rescan(device_t self, const char *ifattr, const int *locators) 256 { 257 struct cpu_softc *sc = device_private(self); 258 struct cpufeature_attach_args cfaa; 259 struct cpu_info *ci = sc->sc_info; 260 261 memset(&cfaa, 0, sizeof(cfaa)); 262 cfaa.ci = ci; 263 264 if (ifattr_match(ifattr, "cpufeaturebus")) { 265 266 if (ci->ci_frequency == NULL) { 267 cfaa.name = "frequency"; 268 ci->ci_frequency = config_found_ia(self, 269 "cpufeaturebus", &cfaa, NULL); 270 } 271 } 272 273 return 0; 274 } 275 276 static void 277 cpu_childdetached(device_t self, device_t child) 278 { 279 struct cpu_softc *sc = device_private(self); 280 struct cpu_info *ci = sc->sc_info; 281 282 if (ci->ci_frequency == child) 283 ci->ci_frequency = NULL; 284 } 285 286 static int 287 vcpu_match(device_t parent, cfdata_t match, void *aux) 288 { 289 struct vcpu_attach_args *vcaa = aux; 290 struct vcpu_runstate_info vcr; 291 int error; 292 293 if (strcmp(vcaa->vcaa_name, match->cf_name) == 0) { 294 error = HYPERVISOR_vcpu_op(VCPUOP_get_runstate_info, 295 vcaa->vcaa_caa.cpu_number, 296 &vcr); 297 switch (error) { 298 case 0: 299 return 1; 300 case -ENOENT: 301 return 0; 302 default: 303 panic("Unknown hypervisor error %d returned on vcpu runstate probe\n", error); 304 } 305 } 306 307 return 0; 308 } 309 310 static void 311 vcpu_attach(device_t parent, device_t self, void *aux) 312 { 313 struct vcpu_attach_args *vcaa = aux; 314 315 KASSERT(vcaa->vcaa_caa.cpu_func == NULL); 316 vcaa->vcaa_caa.cpu_func = &mp_cpu_funcs; 317 cpu_attach_common(parent, self, &vcaa->vcaa_caa); 318 319 if (!pmf_device_register(self, NULL, NULL)) 320 aprint_error_dev(self, "couldn't establish power handler\n"); 321 } 322 323 static int 324 vcpu_is_up(struct cpu_info *ci) 325 { 326 KASSERT(ci != NULL); 327 return HYPERVISOR_vcpu_op(VCPUOP_is_up, ci->ci_cpuid, NULL); 328 } 329 330 static void 331 cpu_vm_init(struct cpu_info *ci) 332 { 333 int ncolors = 2, i; 334 335 for (i = CAI_ICACHE; i <= CAI_L2CACHE; i++) { 336 struct x86_cache_info *cai; 337 int tcolors; 338 339 cai = &ci->ci_cinfo[i]; 340 341 tcolors = atop(cai->cai_totalsize); 342 switch(cai->cai_associativity) { 343 case 0xff: 344 tcolors = 1; /* fully associative */ 345 break; 346 case 0: 347 case 1: 348 break; 349 default: 350 tcolors /= cai->cai_associativity; 351 } 352 ncolors = max(ncolors, tcolors); 353 } 354 355 /* 356 * Knowing the size of the largest cache on this CPU, potentially 357 * re-color our pages. 358 */ 359 aprint_debug_dev(ci->ci_dev, "%d page colors\n", ncolors); 360 uvm_page_recolor(ncolors); 361 pmap_tlb_cpu_init(ci); 362 } 363 364 static void 365 cpu_attach_common(device_t parent, device_t self, void *aux) 366 { 367 struct cpu_softc *sc = device_private(self); 368 struct cpu_attach_args *caa = aux; 369 struct cpu_info *ci; 370 uintptr_t ptr; 371 int cpunum = caa->cpu_number; 372 static bool again = false; 373 374 sc->sc_dev = self; 375 376 /* 377 * If we're an Application Processor, allocate a cpu_info 378 * structure, otherwise use the primary's. 379 */ 380 if (caa->cpu_role == CPU_ROLE_AP) { 381 aprint_naive(": Application Processor\n"); 382 ptr = (uintptr_t)kmem_alloc(sizeof(*ci) + CACHE_LINE_SIZE - 1, 383 KM_SLEEP); 384 ci = (struct cpu_info *)roundup2(ptr, CACHE_LINE_SIZE); 385 memset(ci, 0, sizeof(*ci)); 386 #ifdef TRAPLOG 387 ci->ci_tlog_base = kmem_zalloc(sizeof(struct tlog), KM_SLEEP); 388 #endif 389 } else { 390 aprint_naive(": %s Processor\n", 391 caa->cpu_role == CPU_ROLE_SP ? "Single" : "Boot"); 392 ci = &cpu_info_primary; 393 } 394 395 ci->ci_self = ci; 396 sc->sc_info = ci; 397 ci->ci_dev = self; 398 ci->ci_cpuid = cpunum; 399 400 KASSERT(HYPERVISOR_shared_info != NULL); 401 KASSERT(cpunum < XEN_LEGACY_MAX_VCPUS); 402 ci->ci_vcpu = &HYPERVISOR_shared_info->vcpu_info[cpunum]; 403 404 KASSERT(ci->ci_func == 0); 405 ci->ci_func = caa->cpu_func; 406 407 /* Must be called before mi_cpu_attach(). */ 408 cpu_vm_init(ci); 409 410 if (caa->cpu_role == CPU_ROLE_AP) { 411 int error; 412 413 error = mi_cpu_attach(ci); 414 415 KASSERT(ci->ci_data.cpu_idlelwp != NULL); 416 if (error != 0) { 417 aprint_normal("\n"); 418 aprint_error_dev(self, 419 "mi_cpu_attach failed with %d\n", error); 420 return; 421 } 422 423 } else { 424 KASSERT(ci->ci_data.cpu_idlelwp != NULL); 425 } 426 427 KASSERT(ci->ci_cpuid == ci->ci_index); 428 pmap_reference(pmap_kernel()); 429 ci->ci_pmap = pmap_kernel(); 430 ci->ci_tlbstate = TLBSTATE_STALE; 431 432 /* 433 * Boot processor may not be attached first, but the below 434 * must be done to allow booting other processors. 435 */ 436 if (!again) { 437 atomic_or_32(&ci->ci_flags, CPUF_PRESENT | CPUF_PRIMARY); 438 /* Basic init. */ 439 cpu_intr_init(ci); 440 cpu_get_tsc_freq(ci); 441 cpu_init(ci); 442 pmap_cpu_init_late(ci); 443 444 /* Every processor needs to init it's own ipi h/w (similar to lapic) */ 445 xen_ipi_init(); 446 447 /* Make sure DELAY() is initialized. */ 448 DELAY(1); 449 again = true; 450 } 451 452 /* further PCB init done later. */ 453 454 switch (caa->cpu_role) { 455 case CPU_ROLE_SP: 456 atomic_or_32(&ci->ci_flags, CPUF_SP); 457 cpu_identify(ci); 458 x86_cpu_idle_init(); 459 460 break; 461 462 case CPU_ROLE_BP: 463 atomic_or_32(&ci->ci_flags, CPUF_BSP); 464 cpu_identify(ci); 465 x86_cpu_idle_init(); 466 467 break; 468 469 case CPU_ROLE_AP: 470 atomic_or_32(&ci->ci_flags, CPUF_AP); 471 472 /* 473 * report on an AP 474 */ 475 476 #if defined(MULTIPROCESSOR) 477 /* interrupt handler stack */ 478 cpu_intr_init(ci); 479 480 /* Setup per-cpu memory for gdt */ 481 gdt_alloc_cpu(ci); 482 483 pmap_cpu_init_late(ci); 484 cpu_start_secondary(ci); 485 486 if (ci->ci_flags & CPUF_PRESENT) { 487 struct cpu_info *tmp; 488 489 cpu_identify(ci); 490 tmp = cpu_info_list; 491 while (tmp->ci_next) 492 tmp = tmp->ci_next; 493 494 tmp->ci_next = ci; 495 } 496 #else 497 aprint_error(": not started\n"); 498 #endif 499 break; 500 501 default: 502 aprint_normal("\n"); 503 panic("unknown processor type??\n"); 504 } 505 506 #ifdef MPVERBOSE 507 if (mp_verbose) { 508 struct lwp *l = ci->ci_data.cpu_idlelwp; 509 struct pcb *pcb = lwp_getpcb(l); 510 511 aprint_verbose_dev(self, 512 "idle lwp at %p, idle sp at 0x%p\n", 513 l, 514 #ifdef i386 515 (void *)pcb->pcb_esp 516 #else /* i386 */ 517 (void *)pcb->pcb_rsp 518 #endif /* i386 */ 519 ); 520 521 } 522 #endif /* MPVERBOSE */ 523 } 524 525 /* 526 * Initialize the processor appropriately. 527 */ 528 529 void 530 cpu_init(struct cpu_info *ci) 531 { 532 533 /* 534 * If we have FXSAVE/FXRESTOR, use them. 535 */ 536 if (cpu_feature[0] & CPUID_FXSR) { 537 lcr4(rcr4() | CR4_OSFXSR); 538 539 /* 540 * If we have SSE/SSE2, enable XMM exceptions. 541 */ 542 if (cpu_feature[0] & (CPUID_SSE|CPUID_SSE2)) 543 lcr4(rcr4() | CR4_OSXMMEXCPT); 544 } 545 546 #ifdef __x86_64__ 547 /* No user PGD mapped for this CPU yet */ 548 ci->ci_xen_current_user_pgd = 0; 549 #endif 550 #if defined(__x86_64__) || defined(PAE) 551 mutex_init(&ci->ci_kpm_mtx, MUTEX_DEFAULT, IPL_VM); 552 #endif 553 554 atomic_or_32(&ci->ci_flags, CPUF_RUNNING); 555 } 556 557 558 #ifdef MULTIPROCESSOR 559 560 void 561 cpu_boot_secondary_processors(void) 562 { 563 struct cpu_info *ci; 564 u_long i; 565 for (i = 0; i < maxcpus; i++) { 566 ci = cpu_lookup(i); 567 if (ci == NULL) 568 continue; 569 if (ci->ci_data.cpu_idlelwp == NULL) 570 continue; 571 if ((ci->ci_flags & CPUF_PRESENT) == 0) 572 continue; 573 if (ci->ci_flags & (CPUF_BSP|CPUF_SP|CPUF_PRIMARY)) 574 continue; 575 cpu_boot_secondary(ci); 576 } 577 578 x86_mp_online = true; 579 } 580 581 static void 582 cpu_init_idle_lwp(struct cpu_info *ci) 583 { 584 struct lwp *l = ci->ci_data.cpu_idlelwp; 585 struct pcb *pcb = lwp_getpcb(l); 586 587 pcb->pcb_cr0 = rcr0(); 588 } 589 590 void 591 cpu_init_idle_lwps(void) 592 { 593 struct cpu_info *ci; 594 u_long i; 595 596 for (i = 0; i < maxcpus; i++) { 597 ci = cpu_lookup(i); 598 if (ci == NULL) 599 continue; 600 if (ci->ci_data.cpu_idlelwp == NULL) 601 continue; 602 if ((ci->ci_flags & CPUF_PRESENT) == 0) 603 continue; 604 cpu_init_idle_lwp(ci); 605 } 606 } 607 608 static void 609 cpu_start_secondary(struct cpu_info *ci) 610 { 611 int i; 612 613 aprint_debug_dev(ci->ci_dev, "starting\n"); 614 615 ci->ci_curlwp = ci->ci_data.cpu_idlelwp; 616 617 if (CPU_STARTUP(ci, (vaddr_t) cpu_hatch) != 0) { 618 return; 619 } 620 621 /* 622 * wait for it to become ready 623 */ 624 for (i = 100000; (!(ci->ci_flags & CPUF_PRESENT)) && i > 0; i--) { 625 delay(10); 626 } 627 if ((ci->ci_flags & CPUF_PRESENT) == 0) { 628 aprint_error_dev(ci->ci_dev, "failed to become ready\n"); 629 #if defined(MPDEBUG) && defined(DDB) 630 printf("dropping into debugger; continue from here to resume boot\n"); 631 Debugger(); 632 #endif 633 } 634 635 CPU_START_CLEANUP(ci); 636 } 637 638 void 639 cpu_boot_secondary(struct cpu_info *ci) 640 { 641 int i; 642 atomic_or_32(&ci->ci_flags, CPUF_GO); 643 for (i = 100000; (!(ci->ci_flags & CPUF_RUNNING)) && i > 0; i--) { 644 delay(10); 645 } 646 if ((ci->ci_flags & CPUF_RUNNING) == 0) { 647 aprint_error_dev(ci->ci_dev, "CPU failed to start\n"); 648 #if defined(MPDEBUG) && defined(DDB) 649 printf("dropping into debugger; continue from here to resume boot\n"); 650 Debugger(); 651 #endif 652 } 653 } 654 655 /* 656 * APs end up here immediately after initialisation and VCPUOP_up in 657 * mp_cpu_start(). 658 * At this point, we are running in the idle pcb/idle stack of the new 659 * CPU. This function jumps to the idle loop and starts looking for 660 * work. 661 */ 662 extern void x86_64_tls_switch(struct lwp *); 663 void 664 cpu_hatch(void *v) 665 { 666 struct cpu_info *ci = (struct cpu_info *)v; 667 struct pcb *pcb; 668 int s, i; 669 670 /* Setup TLS and kernel GS/FS */ 671 cpu_init_msrs(ci, true); 672 cpu_init_idt(); 673 gdt_init_cpu(ci); 674 675 cpu_probe(ci); 676 677 atomic_or_32(&ci->ci_flags, CPUF_PRESENT); 678 679 while ((ci->ci_flags & CPUF_GO) == 0) { 680 /* Don't use delay, boot CPU may be patching the text. */ 681 for (i = 10000; i != 0; i--) 682 x86_pause(); 683 } 684 685 /* Because the text may have been patched in x86_patch(). */ 686 x86_flush(); 687 tlbflushg(); 688 689 KASSERT((ci->ci_flags & CPUF_RUNNING) == 0); 690 691 pcb = lwp_getpcb(curlwp); 692 pcb->pcb_cr3 = pmap_pdirpa(pmap_kernel(), 0); 693 pcb = lwp_getpcb(ci->ci_data.cpu_idlelwp); 694 695 xen_ipi_init(); 696 697 xen_initclocks(); 698 699 #ifdef __x86_64__ 700 fpuinit(ci); 701 #endif 702 703 lldt(GSEL(GLDT_SEL, SEL_KPL)); 704 705 cpu_init(ci); 706 cpu_get_tsc_freq(ci); 707 708 s = splhigh(); 709 x86_enable_intr(); 710 splx(s); 711 712 aprint_debug_dev(ci->ci_dev, "running\n"); 713 714 cpu_switchto(NULL, ci->ci_data.cpu_idlelwp, true); 715 716 idle_loop(NULL); 717 KASSERT(false); 718 } 719 720 #if defined(DDB) 721 722 #include <ddb/db_output.h> 723 #include <machine/db_machdep.h> 724 725 /* 726 * Dump CPU information from ddb. 727 */ 728 void 729 cpu_debug_dump(void) 730 { 731 struct cpu_info *ci; 732 CPU_INFO_ITERATOR cii; 733 734 db_printf("addr dev id flags ipis curlwp fpcurlwp\n"); 735 for (CPU_INFO_FOREACH(cii, ci)) { 736 db_printf("%p %s %ld %x %x %10p %10p\n", 737 ci, 738 ci->ci_dev == NULL ? "BOOT" : device_xname(ci->ci_dev), 739 (long)ci->ci_cpuid, 740 ci->ci_flags, ci->ci_ipis, 741 ci->ci_curlwp, 742 ci->ci_fpcurlwp); 743 } 744 } 745 #endif /* DDB */ 746 747 #endif /* MULTIPROCESSOR */ 748 749 extern void hypervisor_callback(void); 750 extern void failsafe_callback(void); 751 #ifdef __x86_64__ 752 typedef void (vector)(void); 753 extern vector Xsyscall, Xsyscall32; 754 #endif 755 756 /* 757 * Setup the "trampoline". On Xen, we setup nearly all cpu context 758 * outside a trampoline, so we prototype and call targetip like so: 759 * void targetip(struct cpu_info *); 760 */ 761 762 static void 763 gdt_prepframes(paddr_t *frames, vaddr_t base, uint32_t entries) 764 { 765 int i; 766 for (i = 0; i < roundup(entries, PAGE_SIZE) >> PAGE_SHIFT; i++) { 767 768 frames[i] = ((paddr_t) xpmap_ptetomach( 769 (pt_entry_t *) (base + (i << PAGE_SHIFT)))) 770 >> PAGE_SHIFT; 771 772 /* Mark Read-only */ 773 pmap_pte_clearbits(kvtopte(base + (i << PAGE_SHIFT)), 774 PG_RW); 775 } 776 } 777 778 #ifdef __x86_64__ 779 extern char *ldtstore; 780 781 static void 782 xen_init_amd64_vcpuctxt(struct cpu_info *ci, 783 struct vcpu_guest_context *initctx, 784 void targetrip(struct cpu_info *)) 785 { 786 /* page frames to point at GDT */ 787 extern int gdt_size; 788 paddr_t frames[16]; 789 psize_t gdt_ents; 790 791 struct lwp *l; 792 struct pcb *pcb; 793 794 volatile struct vcpu_info *vci; 795 796 KASSERT(ci != NULL); 797 KASSERT(ci != &cpu_info_primary); 798 KASSERT(initctx != NULL); 799 KASSERT(targetrip != NULL); 800 801 memset(initctx, 0, sizeof *initctx); 802 803 gdt_ents = roundup(gdt_size, PAGE_SIZE) >> PAGE_SHIFT; 804 KASSERT(gdt_ents <= 16); 805 806 gdt_prepframes(frames, (vaddr_t) ci->ci_gdt, gdt_ents); 807 808 /* Initialise the vcpu context: We use idle_loop()'s pcb context. */ 809 810 l = ci->ci_data.cpu_idlelwp; 811 812 KASSERT(l != NULL); 813 pcb = lwp_getpcb(l); 814 KASSERT(pcb != NULL); 815 816 /* resume with interrupts off */ 817 vci = ci->ci_vcpu; 818 vci->evtchn_upcall_mask = 1; 819 xen_mb(); 820 821 /* resume in kernel-mode */ 822 initctx->flags = VGCF_in_kernel | VGCF_online; 823 824 /* Stack and entry points: 825 * We arrange for the stack frame for cpu_hatch() to 826 * appear as a callee frame of lwp_trampoline(). Being a 827 * leaf frame prevents trampling on any of the MD stack setup 828 * that x86/vm_machdep.c:cpu_lwp_fork() does for idle_loop() 829 */ 830 831 initctx->user_regs.rdi = (uint64_t) ci; /* targetrip(ci); */ 832 initctx->user_regs.rip = (vaddr_t) targetrip; 833 834 initctx->user_regs.cs = GSEL(GCODE_SEL, SEL_KPL); 835 836 initctx->user_regs.rflags = pcb->pcb_flags; 837 initctx->user_regs.rsp = pcb->pcb_rsp; 838 839 /* Data segments */ 840 initctx->user_regs.ss = GSEL(GDATA_SEL, SEL_KPL); 841 initctx->user_regs.es = GSEL(GDATA_SEL, SEL_KPL); 842 initctx->user_regs.ds = GSEL(GDATA_SEL, SEL_KPL); 843 844 /* GDT */ 845 memcpy(initctx->gdt_frames, frames, sizeof frames); 846 initctx->gdt_ents = gdt_ents; 847 848 /* LDT */ 849 initctx->ldt_base = (unsigned long) ldtstore; 850 initctx->ldt_ents = LDT_SIZE >> 3; 851 852 /* Kernel context state */ 853 initctx->kernel_ss = GSEL(GDATA_SEL, SEL_KPL); 854 initctx->kernel_sp = pcb->pcb_rsp0; 855 initctx->ctrlreg[0] = pcb->pcb_cr0; 856 initctx->ctrlreg[1] = 0; /* "resuming" from kernel - no User cr3. */ 857 initctx->ctrlreg[2] = (vaddr_t) targetrip; 858 /* 859 * Use pmap_kernel() L4 PD directly, until we setup the 860 * per-cpu L4 PD in pmap_cpu_init_late() 861 */ 862 initctx->ctrlreg[3] = xen_pfn_to_cr3(x86_btop(xpmap_ptom(ci->ci_kpm_pdirpa))); 863 initctx->ctrlreg[4] = CR4_PAE | CR4_OSFXSR | CR4_OSXMMEXCPT; 864 865 866 /* Xen callbacks */ 867 initctx->event_callback_eip = (unsigned long) hypervisor_callback; 868 initctx->failsafe_callback_eip = (unsigned long) failsafe_callback; 869 initctx->syscall_callback_eip = (unsigned long) Xsyscall; 870 871 return; 872 } 873 #else /* i386 */ 874 extern union descriptor *ldt; 875 extern void Xsyscall(void); 876 877 static void 878 xen_init_i386_vcpuctxt(struct cpu_info *ci, 879 struct vcpu_guest_context *initctx, 880 void targeteip(struct cpu_info *)) 881 { 882 /* page frames to point at GDT */ 883 extern int gdt_size; 884 paddr_t frames[16]; 885 psize_t gdt_ents; 886 887 struct lwp *l; 888 struct pcb *pcb; 889 890 volatile struct vcpu_info *vci; 891 892 KASSERT(ci != NULL); 893 KASSERT(ci != &cpu_info_primary); 894 KASSERT(initctx != NULL); 895 KASSERT(targeteip != NULL); 896 897 memset(initctx, 0, sizeof *initctx); 898 899 gdt_ents = roundup(gdt_size, PAGE_SIZE) >> PAGE_SHIFT; 900 KASSERT(gdt_ents <= 16); 901 902 gdt_prepframes(frames, (vaddr_t) ci->ci_gdt, gdt_ents); 903 904 /* 905 * Initialise the vcpu context: 906 * We use this cpu's idle_loop() pcb context. 907 */ 908 909 l = ci->ci_data.cpu_idlelwp; 910 911 KASSERT(l != NULL); 912 pcb = lwp_getpcb(l); 913 KASSERT(pcb != NULL); 914 915 /* resume with interrupts off */ 916 vci = ci->ci_vcpu; 917 vci->evtchn_upcall_mask = 1; 918 xen_mb(); 919 920 /* resume in kernel-mode */ 921 initctx->flags = VGCF_in_kernel | VGCF_online; 922 923 /* Stack frame setup for cpu_hatch(): 924 * We arrange for the stack frame for cpu_hatch() to 925 * appear as a callee frame of lwp_trampoline(). Being a 926 * leaf frame prevents trampling on any of the MD stack setup 927 * that x86/vm_machdep.c:cpu_lwp_fork() does for idle_loop() 928 */ 929 930 initctx->user_regs.esp = pcb->pcb_esp - 4; /* Leave word for 931 arg1 */ 932 { /* targeteip(ci); */ 933 uint32_t *arg = (uint32_t *) initctx->user_regs.esp; 934 arg[1] = (uint32_t) ci; /* arg1 */ 935 936 } 937 938 initctx->user_regs.eip = (vaddr_t) targeteip; 939 initctx->user_regs.cs = GSEL(GCODE_SEL, SEL_KPL); 940 initctx->user_regs.eflags |= pcb->pcb_iopl; 941 942 /* Data segments */ 943 initctx->user_regs.ss = GSEL(GDATA_SEL, SEL_KPL); 944 initctx->user_regs.es = GSEL(GDATA_SEL, SEL_KPL); 945 initctx->user_regs.ds = GSEL(GDATA_SEL, SEL_KPL); 946 initctx->user_regs.fs = GSEL(GDATA_SEL, SEL_KPL); 947 948 /* GDT */ 949 memcpy(initctx->gdt_frames, frames, sizeof frames); 950 initctx->gdt_ents = gdt_ents; 951 952 /* LDT */ 953 initctx->ldt_base = (unsigned long) ldt; 954 initctx->ldt_ents = NLDT; 955 956 /* Kernel context state */ 957 initctx->kernel_ss = GSEL(GDATA_SEL, SEL_KPL); 958 initctx->kernel_sp = pcb->pcb_esp0; 959 initctx->ctrlreg[0] = pcb->pcb_cr0; 960 initctx->ctrlreg[1] = 0; /* "resuming" from kernel - no User cr3. */ 961 initctx->ctrlreg[2] = (vaddr_t) targeteip; 962 #ifdef PAE 963 initctx->ctrlreg[3] = xen_pfn_to_cr3(x86_btop(xpmap_ptom(ci->ci_pae_l3_pdirpa))); 964 #else /* PAE */ 965 initctx->ctrlreg[3] = xen_pfn_to_cr3(x86_btop(xpmap_ptom(pcb->pcb_cr3))); 966 #endif /* PAE */ 967 initctx->ctrlreg[4] = /* CR4_PAE | */CR4_OSFXSR | CR4_OSXMMEXCPT; 968 969 970 /* Xen callbacks */ 971 initctx->event_callback_eip = (unsigned long) hypervisor_callback; 972 initctx->event_callback_cs = GSEL(GCODE_SEL, SEL_KPL); 973 initctx->failsafe_callback_eip = (unsigned long) failsafe_callback; 974 initctx->failsafe_callback_cs = GSEL(GCODE_SEL, SEL_KPL); 975 976 return; 977 } 978 #endif /* __x86_64__ */ 979 980 int 981 mp_cpu_start(struct cpu_info *ci, vaddr_t target) 982 { 983 984 int hyperror; 985 struct vcpu_guest_context vcpuctx; 986 987 KASSERT(ci != NULL); 988 KASSERT(ci != &cpu_info_primary); 989 KASSERT(ci->ci_flags & CPUF_AP); 990 991 #ifdef __x86_64__ 992 xen_init_amd64_vcpuctxt(ci, &vcpuctx, (void (*)(struct cpu_info *))target); 993 #else /* i386 */ 994 xen_init_i386_vcpuctxt(ci, &vcpuctx, (void (*)(struct cpu_info *))target); 995 #endif /* __x86_64__ */ 996 997 /* Initialise the given vcpu to execute cpu_hatch(ci); */ 998 if ((hyperror = HYPERVISOR_vcpu_op(VCPUOP_initialise, ci->ci_cpuid, &vcpuctx))) { 999 aprint_error(": context initialisation failed. errno = %d\n", hyperror); 1000 return hyperror; 1001 } 1002 1003 /* Start it up */ 1004 1005 /* First bring it down */ 1006 if ((hyperror = HYPERVISOR_vcpu_op(VCPUOP_down, ci->ci_cpuid, NULL))) { 1007 aprint_error(": VCPUOP_down hypervisor command failed. errno = %d\n", hyperror); 1008 return hyperror; 1009 } 1010 1011 if ((hyperror = HYPERVISOR_vcpu_op(VCPUOP_up, ci->ci_cpuid, NULL))) { 1012 aprint_error(": VCPUOP_up hypervisor command failed. errno = %d\n", hyperror); 1013 return hyperror; 1014 } 1015 1016 if (!vcpu_is_up(ci)) { 1017 aprint_error(": did not come up\n"); 1018 return -1; 1019 } 1020 1021 return 0; 1022 } 1023 1024 void 1025 mp_cpu_start_cleanup(struct cpu_info *ci) 1026 { 1027 if (vcpu_is_up(ci)) { 1028 aprint_debug_dev(ci->ci_dev, "is started.\n"); 1029 } 1030 else { 1031 aprint_error_dev(ci->ci_dev, "did not start up.\n"); 1032 } 1033 1034 } 1035 1036 void 1037 cpu_init_msrs(struct cpu_info *ci, bool full) 1038 { 1039 #ifdef __x86_64__ 1040 if (full) { 1041 HYPERVISOR_set_segment_base (SEGBASE_FS, 0); 1042 HYPERVISOR_set_segment_base (SEGBASE_GS_KERNEL, (uint64_t) ci); 1043 HYPERVISOR_set_segment_base (SEGBASE_GS_USER, 0); 1044 } 1045 #endif /* __x86_64__ */ 1046 1047 if (cpu_feature[2] & CPUID_NOX) 1048 wrmsr(MSR_EFER, rdmsr(MSR_EFER) | EFER_NXE); 1049 1050 } 1051 1052 void 1053 cpu_offline_md(void) 1054 { 1055 int s; 1056 1057 s = splhigh(); 1058 fpusave_cpu(true); 1059 splx(s); 1060 } 1061 1062 void 1063 cpu_get_tsc_freq(struct cpu_info *ci) 1064 { 1065 uint32_t vcpu_tversion; 1066 const volatile vcpu_time_info_t *tinfo = &ci->ci_vcpu->time; 1067 1068 vcpu_tversion = tinfo->version; 1069 while (tinfo->version == vcpu_tversion); /* Wait for a time update. XXX: timeout ? */ 1070 1071 uint64_t freq = 1000000000ULL << 32; 1072 freq = freq / (uint64_t)tinfo->tsc_to_system_mul; 1073 if ( tinfo->tsc_shift < 0 ) 1074 freq = freq << -tinfo->tsc_shift; 1075 else 1076 freq = freq >> tinfo->tsc_shift; 1077 ci->ci_data.cpu_cc_freq = freq; 1078 } 1079 1080 void 1081 x86_cpu_idle_xen(void) 1082 { 1083 struct cpu_info *ci = curcpu(); 1084 1085 KASSERT(ci->ci_ilevel == IPL_NONE); 1086 1087 x86_disable_intr(); 1088 if (!__predict_false(ci->ci_want_resched)) { 1089 idle_block(); 1090 } else { 1091 x86_enable_intr(); 1092 } 1093 } 1094 1095 /* 1096 * Loads pmap for the current CPU. 1097 */ 1098 void 1099 cpu_load_pmap(struct pmap *pmap, struct pmap *oldpmap) 1100 { 1101 KASSERT(pmap != pmap_kernel()); 1102 1103 #if defined(__x86_64__) || defined(PAE) 1104 struct cpu_info *ci = curcpu(); 1105 cpuid_t cid = cpu_index(ci); 1106 1107 mutex_enter(&ci->ci_kpm_mtx); 1108 /* make new pmap visible to xen_kpm_sync() */ 1109 kcpuset_atomic_set(pmap->pm_xen_ptp_cpus, cid); 1110 #endif 1111 #ifdef i386 1112 #ifdef PAE 1113 { 1114 int i; 1115 paddr_t l3_pd = xpmap_ptom_masked(ci->ci_pae_l3_pdirpa); 1116 /* don't update the kernel L3 slot */ 1117 for (i = 0 ; i < PDP_SIZE - 1; i++) { 1118 xpq_queue_pte_update(l3_pd + i * sizeof(pd_entry_t), 1119 xpmap_ptom(pmap->pm_pdirpa[i]) | PG_V); 1120 } 1121 tlbflush(); 1122 } 1123 #else /* PAE */ 1124 lcr3(pmap_pdirpa(pmap, 0)); 1125 #endif /* PAE */ 1126 #endif /* i386 */ 1127 1128 #ifdef __x86_64__ 1129 { 1130 int i; 1131 pd_entry_t *new_pgd; 1132 paddr_t l4_pd_ma; 1133 1134 l4_pd_ma = xpmap_ptom_masked(ci->ci_kpm_pdirpa); 1135 1136 /* 1137 * Map user space address in kernel space and load 1138 * user cr3 1139 */ 1140 new_pgd = pmap->pm_pdir; 1141 KASSERT(pmap == ci->ci_pmap); 1142 1143 /* Copy user pmap L4 PDEs (in user addr. range) to per-cpu L4 */ 1144 for (i = 0; i < PDIR_SLOT_PTE; i++) { 1145 KASSERT(pmap != pmap_kernel() || new_pgd[i] == 0); 1146 if (ci->ci_kpm_pdir[i] != new_pgd[i]) { 1147 xpq_queue_pte_update( 1148 l4_pd_ma + i * sizeof(pd_entry_t), 1149 new_pgd[i]); 1150 } 1151 } 1152 1153 xen_set_user_pgd(pmap_pdirpa(pmap, 0)); 1154 ci->ci_xen_current_user_pgd = pmap_pdirpa(pmap, 0); 1155 1156 tlbflush(); 1157 } 1158 1159 #endif /* __x86_64__ */ 1160 #if defined(__x86_64__) || defined(PAE) 1161 /* old pmap no longer visible to xen_kpm_sync() */ 1162 if (oldpmap != pmap_kernel()) { 1163 kcpuset_atomic_clear(oldpmap->pm_xen_ptp_cpus, cid); 1164 } 1165 mutex_exit(&ci->ci_kpm_mtx); 1166 #endif 1167 } 1168 1169 /* 1170 * pmap_cpu_init_late: perform late per-CPU initialization. 1171 * Short note about percpu PDIR pages: 1172 * Both the PAE and __x86_64__ architectures have per-cpu PDIR 1173 * tables. This is to get around Xen's pagetable setup constraints for 1174 * PAE (multiple L3[3]s cannot point to the same L2 - Xen 1175 * will refuse to pin a table setup this way.) and for multiple cpus 1176 * to map in different user pmaps on __x86_64__ (see: cpu_load_pmap()) 1177 * 1178 * What this means for us is that the PDIR of the pmap_kernel() is 1179 * considered to be a canonical "SHADOW" PDIR with the following 1180 * properties: 1181 * - Its recursive mapping points to itself 1182 * - per-cpu recursive mappings point to themselves on __x86_64__ 1183 * - per-cpu L4 pages' kernel entries are expected to be in sync with 1184 * the shadow 1185 */ 1186 1187 void 1188 pmap_cpu_init_late(struct cpu_info *ci) 1189 { 1190 #if defined(PAE) || defined(__x86_64__) 1191 /* 1192 * The BP has already its own PD page allocated during early 1193 * MD startup. 1194 */ 1195 1196 #if defined(__x86_64__) 1197 /* Setup per-cpu normal_pdes */ 1198 int i; 1199 extern pd_entry_t * const normal_pdes[]; 1200 for (i = 0;i < PTP_LEVELS - 1;i++) { 1201 ci->ci_normal_pdes[i] = normal_pdes[i]; 1202 } 1203 #endif /* __x86_64__ */ 1204 1205 if (ci == &cpu_info_primary) 1206 return; 1207 1208 KASSERT(ci != NULL); 1209 1210 #if defined(PAE) 1211 cpu_alloc_l3_page(ci); 1212 KASSERT(ci->ci_pae_l3_pdirpa != 0); 1213 1214 /* Initialise L2 entries 0 - 2: Point them to pmap_kernel() */ 1215 int i; 1216 for (i = 0 ; i < PDP_SIZE - 1; i++) { 1217 ci->ci_pae_l3_pdir[i] = 1218 xpmap_ptom_masked(pmap_kernel()->pm_pdirpa[i]) | PG_V; 1219 } 1220 #endif /* PAE */ 1221 1222 ci->ci_kpm_pdir = (pd_entry_t *)uvm_km_alloc(kernel_map, PAGE_SIZE, 0, 1223 UVM_KMF_WIRED | UVM_KMF_ZERO | UVM_KMF_NOWAIT); 1224 1225 if (ci->ci_kpm_pdir == NULL) { 1226 panic("%s: failed to allocate L4 per-cpu PD for CPU %d\n", 1227 __func__, cpu_index(ci)); 1228 } 1229 ci->ci_kpm_pdirpa = vtophys((vaddr_t) ci->ci_kpm_pdir); 1230 KASSERT(ci->ci_kpm_pdirpa != 0); 1231 1232 #if defined(__x86_64__) 1233 /* 1234 * Copy over the pmap_kernel() shadow L4 entries 1235 */ 1236 1237 memcpy(ci->ci_kpm_pdir, pmap_kernel()->pm_pdir, PAGE_SIZE); 1238 1239 /* Recursive kernel mapping */ 1240 ci->ci_kpm_pdir[PDIR_SLOT_PTE] = xpmap_ptom_masked(ci->ci_kpm_pdirpa) | PG_k | PG_V; 1241 #elif defined(PAE) 1242 /* Copy over the pmap_kernel() shadow L2 entries that map the kernel */ 1243 memcpy(ci->ci_kpm_pdir, pmap_kernel()->pm_pdir + PDIR_SLOT_KERN, nkptp[PTP_LEVELS - 1] * sizeof(pd_entry_t)); 1244 #endif /* __x86_64__ else PAE */ 1245 1246 /* Xen wants R/O */ 1247 pmap_protect(pmap_kernel(), (vaddr_t)ci->ci_kpm_pdir, 1248 (vaddr_t)ci->ci_kpm_pdir + PAGE_SIZE, VM_PROT_READ); 1249 pmap_update(pmap_kernel()); 1250 #if defined(PAE) 1251 /* Initialise L3 entry 3. This mapping is shared across all 1252 * pmaps and is static, ie; loading a new pmap will not update 1253 * this entry. 1254 */ 1255 1256 ci->ci_pae_l3_pdir[3] = xpmap_ptom_masked(ci->ci_kpm_pdirpa) | PG_k | PG_V; 1257 1258 /* Mark L3 R/O (Xen wants this) */ 1259 pmap_protect(pmap_kernel(), (vaddr_t)ci->ci_pae_l3_pdir, 1260 (vaddr_t)ci->ci_pae_l3_pdir + PAGE_SIZE, VM_PROT_READ); 1261 pmap_update(pmap_kernel()); 1262 1263 xpq_queue_pin_l3_table(xpmap_ptom_masked(ci->ci_pae_l3_pdirpa)); 1264 1265 #elif defined(__x86_64__) 1266 xpq_queue_pin_l4_table(xpmap_ptom_masked(ci->ci_kpm_pdirpa)); 1267 #endif /* PAE , __x86_64__ */ 1268 #endif /* defined(PAE) || defined(__x86_64__) */ 1269 } 1270 1271 /* 1272 * Notify all other cpus to halt. 1273 */ 1274 1275 void 1276 cpu_broadcast_halt(void) 1277 { 1278 xen_broadcast_ipi(XEN_IPI_HALT); 1279 } 1280 1281 /* 1282 * Send a dummy ipi to a cpu. 1283 */ 1284 1285 void 1286 cpu_kick(struct cpu_info *ci) 1287 { 1288 (void)xen_send_ipi(ci, XEN_IPI_KICK); 1289 } 1290