enlighten.c 41.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * Core of Xen paravirt_ops implementation.
 *
 * This file contains the xen_paravirt_ops structure itself, and the
 * implementations for:
 * - privileged instructions
 * - interrupt flags
 * - segment operations
 * - booting and setup
 *
 * Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
 */

#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/smp.h>
#include <linux/preempt.h>
18
#include <linux/hardirq.h>
19 20 21 22 23 24
#include <linux/percpu.h>
#include <linux/delay.h>
#include <linux/start_kernel.h>
#include <linux/sched.h>
#include <linux/bootmem.h>
#include <linux/module.h>
25 26 27
#include <linux/mm.h>
#include <linux/page-flags.h>
#include <linux/highmem.h>
28
#include <linux/console.h>
29 30 31 32

#include <xen/interface/xen.h>
#include <xen/interface/physdev.h>
#include <xen/interface/vcpu.h>
J
Jeremy Fitzhardinge 已提交
33
#include <xen/interface/sched.h>
34 35
#include <xen/features.h>
#include <xen/page.h>
J
Jeremy Fitzhardinge 已提交
36
#include <xen/hvc-console.h>
37 38 39 40 41 42 43

#include <asm/paravirt.h>
#include <asm/page.h>
#include <asm/xen/hypercall.h>
#include <asm/xen/hypervisor.h>
#include <asm/fixmap.h>
#include <asm/processor.h>
44
#include <asm/msr-index.h>
45 46 47
#include <asm/setup.h>
#include <asm/desc.h>
#include <asm/pgtable.h>
J
Jeremy Fitzhardinge 已提交
48
#include <asm/tlbflush.h>
J
Jeremy Fitzhardinge 已提交
49
#include <asm/reboot.h>
50 51

#include "xen-ops.h"
J
Jeremy Fitzhardinge 已提交
52
#include "mmu.h"
53 54 55 56 57 58
#include "multicalls.h"

EXPORT_SYMBOL_GPL(hypercall_page);

DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu);
DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info);
59

60 61 62 63 64 65 66 67 68 69 70 71
/*
 * Identity map, in addition to plain kernel map.  This needs to be
 * large enough to allocate page table pages to allocate the rest.
 * Each page can map 2MB.
 */
static pte_t level1_ident_pgt[PTRS_PER_PTE * 4] __page_aligned_bss;

#ifdef CONFIG_X86_64
/* l3 pud for userspace vsyscall mapping */
static pud_t level3_user_vsyscall[PTRS_PER_PUD] __page_aligned_bss;
#endif /* CONFIG_X86_64 */

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
/*
 * Note about cr3 (pagetable base) values:
 *
 * xen_cr3 contains the current logical cr3 value; it contains the
 * last set cr3.  This may not be the current effective cr3, because
 * its update may be being lazily deferred.  However, a vcpu looking
 * at its own cr3 can use this value knowing that it everything will
 * be self-consistent.
 *
 * xen_current_cr3 contains the actual vcpu cr3; it is set once the
 * hypercall to set the vcpu cr3 is complete (so it may be a little
 * out of date, but it will never be set early).  If one vcpu is
 * looking at another vcpu's cr3 value, it should use this variable.
 */
DEFINE_PER_CPU(unsigned long, xen_cr3);	 /* cr3 stored as physaddr */
DEFINE_PER_CPU(unsigned long, xen_current_cr3);	 /* actual vcpu cr3 */
88 89 90 91

struct start_info *xen_start_info;
EXPORT_SYMBOL_GPL(xen_start_info);

92
struct shared_info xen_dummy_shared_info;
93 94 95 96 97

/*
 * Point at some empty memory to start with. We map the real shared_info
 * page as soon as fixmap is up and running.
 */
98
struct shared_info *HYPERVISOR_shared_info = (void *)&xen_dummy_shared_info;
99 100 101 102 103 104 105 106 107 108 109 110 111 112

/*
 * Flag to determine whether vcpu info placement is available on all
 * VCPUs.  We assume it is to start with, and then set it to zero on
 * the first failure.  This is because it can succeed on some VCPUs
 * and not others, since it can involve hypervisor memory allocation,
 * or because the guest failed to guarantee all the appropriate
 * constraints on all VCPUs (ie buffer can't cross a page boundary).
 *
 * Note that any particular CPU may be using a placed vcpu structure,
 * but we can only optimise if the all are.
 *
 * 0: not available, 1: available
 */
113
static int have_vcpu_info_placement = 1;
114

115
static void xen_vcpu_setup(int cpu)
116
{
117 118 119 120
	struct vcpu_register_vcpu_info info;
	int err;
	struct vcpu_info *vcpup;

121
	BUG_ON(HYPERVISOR_shared_info == &xen_dummy_shared_info);
122
	per_cpu(xen_vcpu, cpu) = &HYPERVISOR_shared_info->vcpu_info[cpu];
123 124 125 126 127 128 129 130 131

	if (!have_vcpu_info_placement)
		return;		/* already tested, not available */

	vcpup = &per_cpu(xen_vcpu_info, cpu);

	info.mfn = virt_to_mfn(vcpup);
	info.offset = offset_in_page(vcpup);

132
	printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %llx, offset %d\n",
133 134 135 136 137 138 139 140 141 142 143 144 145 146
	       cpu, vcpup, info.mfn, info.offset);

	/* Check to see if the hypervisor will put the vcpu_info
	   structure where we want it, which allows direct access via
	   a percpu-variable. */
	err = HYPERVISOR_vcpu_op(VCPUOP_register_vcpu_info, cpu, &info);

	if (err) {
		printk(KERN_DEBUG "register_vcpu_info failed: err=%d\n", err);
		have_vcpu_info_placement = 0;
	} else {
		/* This cpu is using the registered vcpu info, even if
		   later ones fail to. */
		per_cpu(xen_vcpu, cpu) = vcpup;
147

148 149 150
		printk(KERN_DEBUG "cpu %d using vcpu_info at %p\n",
		       cpu, vcpup);
	}
151 152
}

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
/*
 * On restore, set the vcpu placement up again.
 * If it fails, then we're in a bad state, since
 * we can't back out from using it...
 */
void xen_vcpu_restore(void)
{
	if (have_vcpu_info_placement) {
		int cpu;

		for_each_online_cpu(cpu) {
			bool other_cpu = (cpu != smp_processor_id());

			if (other_cpu &&
			    HYPERVISOR_vcpu_op(VCPUOP_down, cpu, NULL))
				BUG();

			xen_vcpu_setup(cpu);

			if (other_cpu &&
			    HYPERVISOR_vcpu_op(VCPUOP_up, cpu, NULL))
				BUG();
		}

		BUG_ON(!have_vcpu_info_placement);
	}
}

181 182 183
static void __init xen_banner(void)
{
	printk(KERN_INFO "Booting paravirtualized kernel on %s\n",
184
	       pv_info.name);
185 186 187
	printk(KERN_INFO "Hypervisor signature: %s%s\n",
	       xen_start_info->magic,
	       xen_feature(XENFEAT_mmu_pt_update_preserve_ad) ? " (preserve-AD)" : "");
188 189
}

190 191
static void xen_cpuid(unsigned int *ax, unsigned int *bx,
		      unsigned int *cx, unsigned int *dx)
192 193 194 195 196 197 198
{
	unsigned maskedx = ~0;

	/*
	 * Mask out inconvenient features, to try and disable as many
	 * unsupported kernel subsystems as possible.
	 */
199
	if (*ax == 1)
200 201
		maskedx = ~((1 << X86_FEATURE_APIC) |  /* disable APIC */
			    (1 << X86_FEATURE_ACPI) |  /* disable ACPI */
202 203
			    (1 << X86_FEATURE_MCE)  |  /* disable MCE */
			    (1 << X86_FEATURE_MCA)  |  /* disable MCA */
204 205 206
			    (1 << X86_FEATURE_ACC));   /* thermal monitoring */

	asm(XEN_EMULATE_PREFIX "cpuid"
207 208 209 210 211 212
		: "=a" (*ax),
		  "=b" (*bx),
		  "=c" (*cx),
		  "=d" (*dx)
		: "0" (*ax), "2" (*cx));
	*dx &= maskedx;
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
}

static void xen_set_debugreg(int reg, unsigned long val)
{
	HYPERVISOR_set_debugreg(reg, val);
}

static unsigned long xen_get_debugreg(int reg)
{
	return HYPERVISOR_get_debugreg(reg);
}

static unsigned long xen_save_fl(void)
{
	struct vcpu_info *vcpu;
	unsigned long flags;

	vcpu = x86_read_percpu(xen_vcpu);
231

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
	/* flag has opposite sense of mask */
	flags = !vcpu->evtchn_upcall_mask;

	/* convert to IF type flag
	   -0 -> 0x00000000
	   -1 -> 0xffffffff
	*/
	return (-flags) & X86_EFLAGS_IF;
}

static void xen_restore_fl(unsigned long flags)
{
	struct vcpu_info *vcpu;

	/* convert from IF type flag */
	flags = !(flags & X86_EFLAGS_IF);
248 249 250 251 252

	/* There's a one instruction preempt window here.  We need to
	   make sure we're don't switch CPUs between getting the vcpu
	   pointer and updating the mask. */
	preempt_disable();
253 254
	vcpu = x86_read_percpu(xen_vcpu);
	vcpu->evtchn_upcall_mask = flags;
255
	preempt_enable_no_resched();
256

257 258
	/* Doesn't matter if we get preempted here, because any
	   pending event will get dealt with anyway. */
259

260 261 262
	if (flags == 0) {
		preempt_check_resched();
		barrier(); /* unmask then check (avoid races) */
263 264
		if (unlikely(vcpu->evtchn_upcall_pending))
			force_evtchn_callback();
265
	}
266 267 268 269
}

static void xen_irq_disable(void)
{
270 271 272
	/* There's a one instruction preempt window here.  We need to
	   make sure we're don't switch CPUs between getting the vcpu
	   pointer and updating the mask. */
273
	preempt_disable();
274
	x86_read_percpu(xen_vcpu)->evtchn_upcall_mask = 1;
275 276 277 278 279 280 281
	preempt_enable_no_resched();
}

static void xen_irq_enable(void)
{
	struct vcpu_info *vcpu;

282 283 284 285 286
	/* We don't need to worry about being preempted here, since
	   either a) interrupts are disabled, so no preemption, or b)
	   the caller is confused and is trying to re-enable interrupts
	   on an indeterminate processor. */

287 288 289
	vcpu = x86_read_percpu(xen_vcpu);
	vcpu->evtchn_upcall_mask = 0;

290 291
	/* Doesn't matter if we get preempted here, because any
	   pending event will get dealt with anyway. */
292

293
	barrier(); /* unmask then check (avoid races) */
294 295 296 297 298 299 300
	if (unlikely(vcpu->evtchn_upcall_pending))
		force_evtchn_callback();
}

static void xen_safe_halt(void)
{
	/* Blocking includes an implicit local_irq_enable(). */
J
Jeremy Fitzhardinge 已提交
301
	if (HYPERVISOR_sched_op(SCHEDOP_block, NULL) != 0)
302 303 304 305 306 307 308 309 310 311 312
		BUG();
}

static void xen_halt(void)
{
	if (irqs_disabled())
		HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL);
	else
		xen_safe_halt();
}

313
static void xen_leave_lazy(void)
314
{
315
	paravirt_leave_lazy(paravirt_get_lazy_mode());
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
	xen_mc_flush();
}

static unsigned long xen_store_tr(void)
{
	return 0;
}

static void xen_set_ldt(const void *addr, unsigned entries)
{
	struct mmuext_op *op;
	struct multicall_space mcs = xen_mc_entry(sizeof(*op));

	op = mcs.args;
	op->cmd = MMUEXT_SET_LDT;
331
	op->arg1.linear_addr = (unsigned long)addr;
332 333 334 335 336 337 338
	op->arg2.nr_ents = entries;

	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);

	xen_mc_issue(PARAVIRT_LAZY_CPU);
}

339
static void xen_load_gdt(const struct desc_ptr *dtr)
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
{
	unsigned long *frames;
	unsigned long va = dtr->address;
	unsigned int size = dtr->size + 1;
	unsigned pages = (size + PAGE_SIZE - 1) / PAGE_SIZE;
	int f;
	struct multicall_space mcs;

	/* A GDT can be up to 64k in size, which corresponds to 8192
	   8-byte entries, or 16 4k pages.. */

	BUG_ON(size > 65536);
	BUG_ON(va & ~PAGE_MASK);

	mcs = xen_mc_entry(sizeof(*frames) * pages);
	frames = mcs.args;

	for (f = 0; va < dtr->address + size; va += PAGE_SIZE, f++) {
		frames[f] = virt_to_mfn(va);
		make_lowmem_page_readonly((void *)va);
	}

	MULTI_set_gdt(mcs.mc, frames, size / sizeof(struct desc_struct));

	xen_mc_issue(PARAVIRT_LAZY_CPU);
}

static void load_TLS_descriptor(struct thread_struct *t,
				unsigned int cpu, unsigned int i)
{
	struct desc_struct *gdt = get_cpu_gdt_table(cpu);
	xmaddr_t maddr = virt_to_machine(&gdt[GDT_ENTRY_TLS_MIN+i]);
	struct multicall_space mc = __xen_mc_entry(0);

	MULTI_update_descriptor(mc.mc, maddr.maddr, t->tls_array[i]);
}

static void xen_load_tls(struct thread_struct *t, unsigned int cpu)
{
379 380 381 382 383 384 385 386
	/*
	 * XXX sleazy hack: If we're being called in a lazy-cpu zone,
	 * it means we're in a context switch, and %gs has just been
	 * saved.  This means we can zero it out to prevent faults on
	 * exit from the hypervisor if the next process has no %gs.
	 * Either way, it has been saved, and the new value will get
	 * loaded properly.  This will go away as soon as Xen has been
	 * modified to not save/restore %gs for normal hypercalls.
387 388 389 390 391 392 393 394
	 *
	 * On x86_64, this hack is not used for %gs, because gs points
	 * to KERNEL_GS_BASE (and uses it for PDA references), so we
	 * must not zero %gs on x86_64
	 *
	 * For x86_64, we need to zero %fs, otherwise we may get an
	 * exception between the new %fs descriptor being loaded and
	 * %fs being effectively cleared at __switch_to().
395
	 */
396 397
	if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU) {
#ifdef CONFIG_X86_32
398
		loadsegment(gs, 0);
399 400 401 402 403 404 405 406 407 408 409 410
#else
		loadsegment(fs, 0);
#endif
	}

	xen_mc_batch();

	load_TLS_descriptor(t, cpu, 0);
	load_TLS_descriptor(t, cpu, 1);
	load_TLS_descriptor(t, cpu, 2);

	xen_mc_issue(PARAVIRT_LAZY_CPU);
411 412
}

413 414 415 416 417 418 419 420
#ifdef CONFIG_X86_64
static void xen_load_gs_index(unsigned int idx)
{
	if (HYPERVISOR_set_segment_base(SEGBASE_GS_USER_SEL, idx))
		BUG();
}
#endif

421
static void xen_write_ldt_entry(struct desc_struct *dt, int entrynum,
422
				const void *ptr)
423 424 425
{
	unsigned long lp = (unsigned long)&dt[entrynum];
	xmaddr_t mach_lp = virt_to_machine(lp);
426
	u64 entry = *(u64 *)ptr;
427

428 429
	preempt_disable();

430 431 432
	xen_mc_flush();
	if (HYPERVISOR_update_descriptor(mach_lp.maddr, entry))
		BUG();
433 434

	preempt_enable();
435 436
}

437
static int cvt_gate_to_trap(int vector, const gate_desc *val,
438 439
			    struct trap_info *info)
{
440
	if (val->type != 0xf && val->type != 0xe)
441 442 443
		return 0;

	info->vector = vector;
444 445 446
	info->address = gate_offset(*val);
	info->cs = gate_segment(*val);
	info->flags = val->dpl;
447
	/* interrupt gates clear IF */
448
	if (val->type == 0xe)
449 450 451 452 453 454
		info->flags |= 4;

	return 1;
}

/* Locations of each CPU's IDT */
455
static DEFINE_PER_CPU(struct desc_ptr, idt_desc);
456 457 458

/* Set an IDT entry.  If the entry is part of the current IDT, then
   also update Xen. */
459
static void xen_write_idt_entry(gate_desc *dt, int entrynum, const gate_desc *g)
460 461
{
	unsigned long p = (unsigned long)&dt[entrynum];
462 463 464 465 466 467
	unsigned long start, end;

	preempt_disable();

	start = __get_cpu_var(idt_desc).address;
	end = start + __get_cpu_var(idt_desc).size + 1;
468 469 470

	xen_mc_flush();

471
	native_write_idt_entry(dt, entrynum, g);
472 473 474 475 476 477

	if (p >= start && (p + 8) <= end) {
		struct trap_info info[2];

		info[1].address = 0;

478
		if (cvt_gate_to_trap(entrynum, g, &info[0]))
479 480 481
			if (HYPERVISOR_set_trap_table(info))
				BUG();
	}
482 483

	preempt_enable();
484 485
}

486
static void xen_convert_trap_info(const struct desc_ptr *desc,
J
Jeremy Fitzhardinge 已提交
487
				  struct trap_info *traps)
488 489 490
{
	unsigned in, out, count;

491
	count = (desc->size+1) / sizeof(gate_desc);
492 493 494
	BUG_ON(count > 256);

	for (in = out = 0; in < count; in++) {
495
		gate_desc *entry = (gate_desc*)(desc->address) + in;
496

497
		if (cvt_gate_to_trap(in, entry, &traps[out]))
498 499 500
			out++;
	}
	traps[out].address = 0;
J
Jeremy Fitzhardinge 已提交
501 502 503 504
}

void xen_copy_trap_info(struct trap_info *traps)
{
505
	const struct desc_ptr *desc = &__get_cpu_var(idt_desc);
J
Jeremy Fitzhardinge 已提交
506 507 508 509 510 511 512

	xen_convert_trap_info(desc, traps);
}

/* Load a new IDT into Xen.  In principle this can be per-CPU, so we
   hold a spinlock to protect the static traps[] array (static because
   it avoids allocation, and saves stack space). */
513
static void xen_load_idt(const struct desc_ptr *desc)
J
Jeremy Fitzhardinge 已提交
514 515 516 517 518 519
{
	static DEFINE_SPINLOCK(lock);
	static struct trap_info traps[257];

	spin_lock(&lock);

520 521
	__get_cpu_var(idt_desc) = *desc;

J
Jeremy Fitzhardinge 已提交
522
	xen_convert_trap_info(desc, traps);
523 524 525 526 527 528 529 530 531 532 533

	xen_mc_flush();
	if (HYPERVISOR_set_trap_table(traps))
		BUG();

	spin_unlock(&lock);
}

/* Write a GDT descriptor entry.  Ignore LDT descriptors, since
   they're handled differently. */
static void xen_write_gdt_entry(struct desc_struct *dt, int entry,
534
				const void *desc, int type)
535
{
536 537
	preempt_disable();

538 539 540
	switch (type) {
	case DESC_LDT:
	case DESC_TSS:
541 542 543 544 545 546 547
		/* ignore */
		break;

	default: {
		xmaddr_t maddr = virt_to_machine(&dt[entry]);

		xen_mc_flush();
548
		if (HYPERVISOR_update_descriptor(maddr.maddr, *(u64 *)desc))
549 550 551 552
			BUG();
	}

	}
553 554

	preempt_enable();
555 556
}

557
static void xen_load_sp0(struct tss_struct *tss,
558
			  struct thread_struct *thread)
559 560
{
	struct multicall_space mcs = xen_mc_entry(0);
561
	MULTI_stack_switch(mcs.mc, __KERNEL_DS, thread->sp0);
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578
	xen_mc_issue(PARAVIRT_LAZY_CPU);
}

static void xen_set_iopl_mask(unsigned mask)
{
	struct physdev_set_iopl set_iopl;

	/* Force the change at ring 0. */
	set_iopl.iopl = (mask == 0) ? 1 : (mask >> 12) & 3;
	HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl);
}

static void xen_io_delay(void)
{
}

#ifdef CONFIG_X86_LOCAL_APIC
579
static u32 xen_apic_read(unsigned long reg)
580 581 582
{
	return 0;
}
J
Jeremy Fitzhardinge 已提交
583

584
static void xen_apic_write(unsigned long reg, u32 val)
J
Jeremy Fitzhardinge 已提交
585 586 587 588
{
	/* Warn to see if there's any stray references */
	WARN_ON(1);
}
589 590 591 592
#endif

static void xen_flush_tlb(void)
{
J
Jeremy Fitzhardinge 已提交
593
	struct mmuext_op *op;
594 595 596 597 598
	struct multicall_space mcs;

	preempt_disable();

	mcs = xen_mc_entry(sizeof(*op));
599

J
Jeremy Fitzhardinge 已提交
600 601 602 603 604
	op = mcs.args;
	op->cmd = MMUEXT_TLB_FLUSH_LOCAL;
	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);

	xen_mc_issue(PARAVIRT_LAZY_MMU);
605 606

	preempt_enable();
607 608 609 610
}

static void xen_flush_tlb_single(unsigned long addr)
{
J
Jeremy Fitzhardinge 已提交
611
	struct mmuext_op *op;
612 613 614
	struct multicall_space mcs;

	preempt_disable();
615

616
	mcs = xen_mc_entry(sizeof(*op));
J
Jeremy Fitzhardinge 已提交
617 618 619 620 621 622
	op = mcs.args;
	op->cmd = MMUEXT_INVLPG_LOCAL;
	op->arg1.linear_addr = addr & PAGE_MASK;
	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);

	xen_mc_issue(PARAVIRT_LAZY_MMU);
623 624

	preempt_enable();
625 626
}

J
Jeremy Fitzhardinge 已提交
627 628 629
static void xen_flush_tlb_others(const cpumask_t *cpus, struct mm_struct *mm,
				 unsigned long va)
{
J
Jeremy Fitzhardinge 已提交
630 631 632 633
	struct {
		struct mmuext_op op;
		cpumask_t mask;
	} *args;
J
Jeremy Fitzhardinge 已提交
634
	cpumask_t cpumask = *cpus;
J
Jeremy Fitzhardinge 已提交
635
	struct multicall_space mcs;
J
Jeremy Fitzhardinge 已提交
636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651

	/*
	 * A couple of (to be removed) sanity checks:
	 *
	 * - current CPU must not be in mask
	 * - mask must exist :)
	 */
	BUG_ON(cpus_empty(cpumask));
	BUG_ON(cpu_isset(smp_processor_id(), cpumask));
	BUG_ON(!mm);

	/* If a CPU which we ran on has gone down, OK. */
	cpus_and(cpumask, cpumask, cpu_online_map);
	if (cpus_empty(cpumask))
		return;

J
Jeremy Fitzhardinge 已提交
652 653 654 655 656
	mcs = xen_mc_entry(sizeof(*args));
	args = mcs.args;
	args->mask = cpumask;
	args->op.arg2.vcpumask = &args->mask;

J
Jeremy Fitzhardinge 已提交
657
	if (va == TLB_FLUSH_ALL) {
J
Jeremy Fitzhardinge 已提交
658
		args->op.cmd = MMUEXT_TLB_FLUSH_MULTI;
J
Jeremy Fitzhardinge 已提交
659
	} else {
J
Jeremy Fitzhardinge 已提交
660 661
		args->op.cmd = MMUEXT_INVLPG_MULTI;
		args->op.arg1.linear_addr = va;
J
Jeremy Fitzhardinge 已提交
662 663
	}

J
Jeremy Fitzhardinge 已提交
664 665 666
	MULTI_mmuext_op(mcs.mc, &args->op, 1, NULL, DOMID_SELF);

	xen_mc_issue(PARAVIRT_LAZY_MMU);
J
Jeremy Fitzhardinge 已提交
667 668
}

669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692
static void xen_clts(void)
{
	struct multicall_space mcs;

	mcs = xen_mc_entry(0);

	MULTI_fpu_taskswitch(mcs.mc, 0);

	xen_mc_issue(PARAVIRT_LAZY_CPU);
}

static void xen_write_cr0(unsigned long cr0)
{
	struct multicall_space mcs;

	/* Only pay attention to cr0.TS; everything else is
	   ignored. */
	mcs = xen_mc_entry(0);

	MULTI_fpu_taskswitch(mcs.mc, (cr0 & X86_CR0_TS) != 0);

	xen_mc_issue(PARAVIRT_LAZY_CPU);
}

693 694 695 696 697
static void xen_write_cr2(unsigned long cr2)
{
	x86_read_percpu(xen_vcpu)->arch.cr2 = cr2;
}

698 699 700 701 702
static unsigned long xen_read_cr2(void)
{
	return x86_read_percpu(xen_vcpu)->arch.cr2;
}

703 704 705 706 707
static unsigned long xen_read_cr2_direct(void)
{
	return x86_read_percpu(xen_vcpu_info.arch.cr2);
}

708 709
static void xen_write_cr4(unsigned long cr4)
{
710 711 712 713
	cr4 &= ~X86_CR4_PGE;
	cr4 &= ~X86_CR4_PSE;

	native_write_cr4(cr4);
714 715 716 717 718 719 720
}

static unsigned long xen_read_cr3(void)
{
	return x86_read_percpu(xen_cr3);
}

721 722 723 724 725
static void set_current_cr3(void *v)
{
	x86_write_percpu(xen_current_cr3, (unsigned long)v);
}

726
static void __xen_write_cr3(bool kernel, unsigned long cr3)
727
{
728 729
	struct mmuext_op *op;
	struct multicall_space mcs;
730
	unsigned long mfn;
731

732 733 734 735
	if (cr3)
		mfn = pfn_to_mfn(PFN_DOWN(cr3));
	else
		mfn = 0;
736

737
	WARN_ON(mfn == 0 && kernel);
738

739
	mcs = __xen_mc_entry(sizeof(*op));
740

741
	op = mcs.args;
742
	op->cmd = kernel ? MMUEXT_NEW_BASEPTR : MMUEXT_NEW_USER_BASEPTR;
743
	op->arg1.mfn = mfn;
744

745
	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
746

747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776
	if (kernel) {
		x86_write_percpu(xen_cr3, cr3);

		/* Update xen_current_cr3 once the batch has actually
		   been submitted. */
		xen_mc_callback(set_current_cr3, (void *)cr3);
	}
}

static void xen_write_cr3(unsigned long cr3)
{
	BUG_ON(preemptible());

	xen_mc_batch();  /* disables interrupts */

	/* Update while interrupts are disabled, so its atomic with
	   respect to ipis */
	x86_write_percpu(xen_cr3, cr3);

	__xen_write_cr3(true, cr3);

#ifdef CONFIG_X86_64
	{
		pgd_t *user_pgd = xen_get_user_pgd(__va(cr3));
		if (user_pgd)
			__xen_write_cr3(false, __pa(user_pgd));
		else
			__xen_write_cr3(false, 0);
	}
#endif
777

778
	xen_mc_issue(PARAVIRT_LAZY_CPU);  /* interrupts restored */
779 780
}

781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808
static int xen_write_msr_safe(unsigned int msr, unsigned low, unsigned high)
{
	int ret;

	ret = 0;

	switch(msr) {
#ifdef CONFIG_X86_64
		unsigned which;
		u64 base;

	case MSR_FS_BASE:		which = SEGBASE_FS; goto set;
	case MSR_KERNEL_GS_BASE:	which = SEGBASE_GS_USER; goto set;
	case MSR_GS_BASE:		which = SEGBASE_GS_KERNEL; goto set;

	set:
		base = ((u64)high << 32) | low;
		if (HYPERVISOR_set_segment_base(which, base) != 0)
			ret = -EFAULT;
		break;
#endif
	default:
		ret = native_write_msr_safe(msr, low, high);
	}

	return ret;
}

809 810
/* Early in boot, while setting up the initial pagetable, assume
   everything is pinned. */
811
static __init void xen_alloc_pte_init(struct mm_struct *mm, u32 pfn)
812
{
813
#ifdef CONFIG_FLATMEM
814
	BUG_ON(mem_map);	/* should only be used early */
815
#endif
816 817 818
	make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
}

819
/* Early release_pte assumes that all pts are pinned, since there's
820
   only init_mm and anything attached to that is pinned. */
821
static void xen_release_pte_init(u32 pfn)
822 823 824 825
{
	make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
}

826
static void pin_pagetable_pfn(unsigned cmd, unsigned long pfn)
827 828
{
	struct mmuext_op op;
829
	op.cmd = cmd;
830 831 832 833 834
	op.arg1.mfn = pfn_to_mfn(pfn);
	if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF))
		BUG();
}

835 836
/* This needs to make sure the new pte page is pinned iff its being
   attached to a pinned pagetable. */
837
static void xen_alloc_ptpage(struct mm_struct *mm, u32 pfn, unsigned level)
838
{
839
	struct page *page = pfn_to_page(pfn);
840

841 842 843
	if (PagePinned(virt_to_page(mm->pgd))) {
		SetPagePinned(page);

844
		if (!PageHighMem(page)) {
845
			make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
846 847
			if (level == PT_PTE)
				pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
848
		} else
849 850 851 852
			/* make sure there are no stray mappings of
			   this page */
			kmap_flush_unused();
	}
853 854
}

855
static void xen_alloc_pte(struct mm_struct *mm, u32 pfn)
856
{
857
	xen_alloc_ptpage(mm, pfn, PT_PTE);
858 859
}

860
static void xen_alloc_pmd(struct mm_struct *mm, u32 pfn)
861
{
862
	xen_alloc_ptpage(mm, pfn, PT_PMD);
863 864
}

865 866 867 868 869 870 871 872 873 874
static int xen_pgd_alloc(struct mm_struct *mm)
{
	pgd_t *pgd = mm->pgd;
	int ret = 0;

	BUG_ON(PagePinned(virt_to_page(pgd)));

#ifdef CONFIG_X86_64
	{
		struct page *page = virt_to_page(pgd);
875
		pgd_t *user_pgd;
876 877 878

		BUG_ON(page->private != 0);

879 880 881 882 883 884 885 886 887 888
		ret = -ENOMEM;

		user_pgd = (pgd_t *)__get_free_page(GFP_KERNEL | __GFP_ZERO);
		page->private = (unsigned long)user_pgd;

		if (user_pgd != NULL) {
			user_pgd[pgd_index(VSYSCALL_START)] =
				__pgd(__pa(level3_user_vsyscall) | _PAGE_TABLE);
			ret = 0;
		}
889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906

		BUG_ON(PagePinned(virt_to_page(xen_get_user_pgd(pgd))));
	}
#endif

	return ret;
}

static void xen_pgd_free(struct mm_struct *mm, pgd_t *pgd)
{
#ifdef CONFIG_X86_64
	pgd_t *user_pgd = xen_get_user_pgd(pgd);

	if (user_pgd)
		free_page((unsigned long)user_pgd);
#endif
}

907
/* This should never happen until we're OK to use struct page */
908
static void xen_release_ptpage(u32 pfn, unsigned level)
909
{
910 911 912
	struct page *page = pfn_to_page(pfn);

	if (PagePinned(page)) {
913
		if (!PageHighMem(page)) {
914 915
			if (level == PT_PTE)
				pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
916
			make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
917
		}
918
		ClearPagePinned(page);
919
	}
920 921
}

922
static void xen_release_pte(u32 pfn)
923 924 925 926
{
	xen_release_ptpage(pfn, PT_PTE);
}

927
static void xen_release_pmd(u32 pfn)
928 929 930 931
{
	xen_release_ptpage(pfn, PT_PMD);
}

932 933 934 935 936 937 938 939 940 941 942 943
#if PAGETABLE_LEVELS == 4
static void xen_alloc_pud(struct mm_struct *mm, u32 pfn)
{
	xen_alloc_ptpage(mm, pfn, PT_PUD);
}

static void xen_release_pud(u32 pfn)
{
	xen_release_ptpage(pfn, PT_PUD);
}
#endif

944 945
#ifdef CONFIG_HIGHPTE
static void *xen_kmap_atomic_pte(struct page *page, enum km_type type)
946
{
947 948 949 950 951 952 953 954 955 956 957
	pgprot_t prot = PAGE_KERNEL;

	if (PagePinned(page))
		prot = PAGE_KERNEL_RO;

	if (0 && PageHighMem(page))
		printk("mapping highpte %lx type %d prot %s\n",
		       page_to_pfn(page), type,
		       (unsigned long)pgprot_val(prot) & _PAGE_RW ? "WRITE" : "READ");

	return kmap_atomic_prot(page, type, prot);
958
}
959
#endif
960

961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
static __init pte_t mask_rw_pte(pte_t *ptep, pte_t pte)
{
	/* If there's an existing pte, then don't allow _PAGE_RW to be set */
	if (pte_val_ma(*ptep) & _PAGE_PRESENT)
		pte = __pte_ma(((pte_val_ma(*ptep) & _PAGE_RW) | ~_PAGE_RW) &
			       pte_val_ma(pte));

	return pte;
}

/* Init-time set_pte while constructing initial pagetables, which
   doesn't allow RO pagetable pages to be remapped RW */
static __init void xen_set_pte_init(pte_t *ptep, pte_t pte)
{
	pte = mask_rw_pte(ptep, pte);

	xen_set_pte(ptep, pte);
}

980 981 982 983
static __init void xen_pagetable_setup_start(pgd_t *base)
{
}

984
void xen_setup_shared_info(void)
985 986
{
	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
987 988 989 990 991
		set_fixmap(FIX_PARAVIRT_BOOTMAP,
			   xen_start_info->shared_info);

		HYPERVISOR_shared_info =
			(struct shared_info *)fix_to_virt(FIX_PARAVIRT_BOOTMAP);
992 993 994 995
	} else
		HYPERVISOR_shared_info =
			(struct shared_info *)__va(xen_start_info->shared_info);

996 997 998 999
#ifndef CONFIG_SMP
	/* In UP this is as good a place as any to set up shared info */
	xen_setup_vcpu_info_placement();
#endif
J
Jeremy Fitzhardinge 已提交
1000 1001

	xen_setup_mfn_list_list();
1002 1003 1004 1005
}

static __init void xen_pagetable_setup_done(pgd_t *base)
{
1006
	xen_setup_shared_info();
1007
}
1008

1009 1010
static __init void xen_post_allocator_init(void)
{
1011
	pv_mmu_ops.set_pte = xen_set_pte;
1012 1013
	pv_mmu_ops.set_pmd = xen_set_pmd;
	pv_mmu_ops.set_pud = xen_set_pud;
1014 1015 1016
#if PAGETABLE_LEVELS == 4
	pv_mmu_ops.set_pgd = xen_set_pgd;
#endif
1017

1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
	/* This will work as long as patching hasn't happened yet
	   (which it hasn't) */
	pv_mmu_ops.alloc_pte = xen_alloc_pte;
	pv_mmu_ops.alloc_pmd = xen_alloc_pmd;
	pv_mmu_ops.release_pte = xen_release_pte;
	pv_mmu_ops.release_pmd = xen_release_pmd;
#if PAGETABLE_LEVELS == 4
	pv_mmu_ops.alloc_pud = xen_alloc_pud;
	pv_mmu_ops.release_pud = xen_release_pud;
#endif

1029 1030 1031
#ifdef CONFIG_X86_64
	SetPagePinned(virt_to_page(level3_user_vsyscall));
#endif
1032 1033 1034
	xen_mark_init_mm_pinned();
}

1035
/* This is called once we have the cpu_possible_map */
1036
void xen_setup_vcpu_info_placement(void)
1037 1038 1039 1040 1041 1042 1043 1044
{
	int cpu;

	for_each_possible_cpu(cpu)
		xen_vcpu_setup(cpu);

	/* xen_vcpu_setup managed to place the vcpu_info within the
	   percpu area for all cpus, so make use of it */
1045
#ifdef CONFIG_X86_32
1046 1047 1048
	if (have_vcpu_info_placement) {
		printk(KERN_INFO "Xen: using vcpu_info placement\n");

1049 1050 1051 1052 1053
		pv_irq_ops.save_fl = xen_save_fl_direct;
		pv_irq_ops.restore_fl = xen_restore_fl_direct;
		pv_irq_ops.irq_disable = xen_irq_disable_direct;
		pv_irq_ops.irq_enable = xen_irq_enable_direct;
		pv_mmu_ops.read_cr2 = xen_read_cr2_direct;
1054
	}
1055
#endif
1056 1057
}

1058 1059
static unsigned xen_patch(u8 type, u16 clobbers, void *insnbuf,
			  unsigned long addr, unsigned len)
1060 1061 1062 1063 1064 1065
{
	char *start, *end, *reloc;
	unsigned ret;

	start = end = reloc = NULL;

1066 1067
#define SITE(op, x)							\
	case PARAVIRT_PATCH(op.x):					\
1068 1069 1070 1071 1072 1073 1074 1075
	if (have_vcpu_info_placement) {					\
		start = (char *)xen_##x##_direct;			\
		end = xen_##x##_direct_end;				\
		reloc = xen_##x##_direct_reloc;				\
	}								\
	goto patch_site

	switch (type) {
1076
#ifdef CONFIG_X86_32
1077 1078 1079 1080
		SITE(pv_irq_ops, irq_enable);
		SITE(pv_irq_ops, irq_disable);
		SITE(pv_irq_ops, save_fl);
		SITE(pv_irq_ops, restore_fl);
1081
#endif /* CONFIG_X86_32 */
1082 1083 1084 1085 1086 1087
#undef SITE

	patch_site:
		if (start == NULL || (end-start) > len)
			goto default_patch;

1088
		ret = paravirt_patch_insns(insnbuf, len, start, end);
1089 1090 1091 1092 1093 1094 1095

		/* Note: because reloc is assigned from something that
		   appears to be an array, gcc assumes it's non-null,
		   but doesn't know its relationship with start and
		   end. */
		if (reloc > start && reloc < end) {
			int reloc_off = reloc - start;
1096 1097
			long *relocp = (long *)(insnbuf + reloc_off);
			long delta = start - (char *)addr;
1098 1099 1100 1101 1102 1103 1104

			*relocp += delta;
		}
		break;

	default_patch:
	default:
1105 1106
		ret = paravirt_patch_default(type, clobbers, insnbuf,
					     addr, len);
1107 1108 1109 1110 1111 1112
		break;
	}

	return ret;
}

1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
static void xen_set_fixmap(unsigned idx, unsigned long phys, pgprot_t prot)
{
	pte_t pte;

	phys >>= PAGE_SHIFT;

	switch (idx) {
	case FIX_BTMAP_END ... FIX_BTMAP_BEGIN:
#ifdef CONFIG_X86_F00F_BUG
	case FIX_F00F_IDT:
#endif
1124
#ifdef CONFIG_X86_32
1125 1126
	case FIX_WP_TEST:
	case FIX_VDSO:
1127 1128 1129 1130
	case FIX_KMAP_BEGIN ... FIX_KMAP_END:
#else
	case VSYSCALL_LAST_PAGE ... VSYSCALL_FIRST_PAGE:
#endif
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142
#ifdef CONFIG_X86_LOCAL_APIC
	case FIX_APIC_BASE:	/* maps dummy local APIC */
#endif
		pte = pfn_pte(phys, prot);
		break;

	default:
		pte = mfn_pte(phys, prot);
		break;
	}

	__native_set_fixmap(idx, pte);
1143 1144 1145 1146 1147 1148 1149 1150 1151

#ifdef CONFIG_X86_64
	/* Replicate changes to map the vsyscall page into the user
	   pagetable vsyscall mapping. */
	if (idx >= VSYSCALL_LAST_PAGE && idx <= VSYSCALL_FIRST_PAGE) {
		unsigned long vaddr = __fix_to_virt(idx);
		set_pte_vaddr_pud(level3_user_vsyscall, vaddr, pte);
	}
#endif
1152 1153
}

1154
static const struct pv_info xen_info __initdata = {
1155 1156 1157 1158
	.paravirt_enabled = 1,
	.shared_kernel_pmd = 0,

	.name = "Xen",
1159
};
1160

1161
static const struct pv_init_ops xen_init_ops __initdata = {
1162
	.patch = xen_patch,
1163

1164
	.banner = xen_banner,
1165 1166
	.memory_setup = xen_memory_setup,
	.arch_setup = xen_arch_setup,
1167
	.post_allocator_init = xen_post_allocator_init,
1168
};
1169

1170
static const struct pv_time_ops xen_time_ops __initdata = {
J
Jeremy Fitzhardinge 已提交
1171
	.time_init = xen_time_init,
1172

J
Jeremy Fitzhardinge 已提交
1173 1174
	.set_wallclock = xen_set_wallclock,
	.get_wallclock = xen_get_wallclock,
1175
	.get_tsc_khz = xen_tsc_khz,
J
Jeremy Fitzhardinge 已提交
1176
	.sched_clock = xen_sched_clock,
1177
};
J
Jeremy Fitzhardinge 已提交
1178

1179
static const struct pv_cpu_ops xen_cpu_ops __initdata = {
1180 1181 1182 1183 1184
	.cpuid = xen_cpuid,

	.set_debugreg = xen_set_debugreg,
	.get_debugreg = xen_get_debugreg,

1185
	.clts = xen_clts,
1186 1187

	.read_cr0 = native_read_cr0,
1188
	.write_cr0 = xen_write_cr0,
1189 1190 1191 1192 1193 1194 1195 1196

	.read_cr4 = native_read_cr4,
	.read_cr4_safe = native_read_cr4_safe,
	.write_cr4 = xen_write_cr4,

	.wbinvd = native_wbinvd,

	.read_msr = native_read_msr_safe,
1197
	.write_msr = xen_write_msr_safe,
1198 1199 1200
	.read_tsc = native_read_tsc,
	.read_pmc = native_read_pmc,

1201
	.iret = xen_iret,
1202
	.irq_enable_sysexit = xen_sysexit,
1203 1204 1205 1206
#ifdef CONFIG_X86_64
	.usergs_sysret32 = xen_sysret32,
	.usergs_sysret64 = xen_sysret64,
#endif
1207 1208 1209 1210 1211 1212

	.load_tr_desc = paravirt_nop,
	.set_ldt = xen_set_ldt,
	.load_gdt = xen_load_gdt,
	.load_idt = xen_load_idt,
	.load_tls = xen_load_tls,
1213 1214 1215
#ifdef CONFIG_X86_64
	.load_gs_index = xen_load_gs_index,
#endif
1216 1217 1218 1219 1220 1221 1222 1223

	.store_gdt = native_store_gdt,
	.store_idt = native_store_idt,
	.store_tr = xen_store_tr,

	.write_ldt_entry = xen_write_ldt_entry,
	.write_gdt_entry = xen_write_gdt_entry,
	.write_idt_entry = xen_write_idt_entry,
1224
	.load_sp0 = xen_load_sp0,
1225 1226 1227 1228

	.set_iopl_mask = xen_set_iopl_mask,
	.io_delay = xen_io_delay,

J
Jeremy Fitzhardinge 已提交
1229 1230 1231
	/* Xen takes care of %gs when switching to usermode for us */
	.swapgs = paravirt_nop,

1232 1233 1234 1235
	.lazy_mode = {
		.enter = paravirt_enter_lazy_cpu,
		.leave = xen_leave_lazy,
	},
1236 1237
};

1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
static void __init __xen_init_IRQ(void)
{
#ifdef CONFIG_X86_64
	int i;

	/* Create identity vector->irq map */
	for(i = 0; i < NR_VECTORS; i++) {
		int cpu;

		for_each_possible_cpu(cpu)
			per_cpu(vector_irq, cpu)[i] = i;
	}
#endif	/* CONFIG_X86_64 */

	xen_init_IRQ();
}

1255
static const struct pv_irq_ops xen_irq_ops __initdata = {
1256
	.init_IRQ = __xen_init_IRQ,
1257 1258 1259 1260 1261 1262
	.save_fl = xen_save_fl,
	.restore_fl = xen_restore_fl,
	.irq_disable = xen_irq_disable,
	.irq_enable = xen_irq_enable,
	.safe_halt = xen_safe_halt,
	.halt = xen_halt,
1263
#ifdef CONFIG_X86_64
1264
	.adjust_exception_frame = xen_adjust_exception_frame,
1265
#endif
1266
};
1267

1268
static const struct pv_apic_ops xen_apic_ops __initdata = {
1269
#ifdef CONFIG_X86_LOCAL_APIC
J
Jeremy Fitzhardinge 已提交
1270 1271
	.apic_write = xen_apic_write,
	.apic_write_atomic = xen_apic_write,
1272 1273 1274 1275 1276
	.apic_read = xen_apic_read,
	.setup_boot_clock = paravirt_nop,
	.setup_secondary_clock = paravirt_nop,
	.startup_ipi_hook = paravirt_nop,
#endif
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
};

static const struct pv_mmu_ops xen_mmu_ops __initdata = {
	.pagetable_setup_start = xen_pagetable_setup_start,
	.pagetable_setup_done = xen_pagetable_setup_done,

	.read_cr2 = xen_read_cr2,
	.write_cr2 = xen_write_cr2,

	.read_cr3 = xen_read_cr3,
	.write_cr3 = xen_write_cr3,
1288 1289 1290 1291

	.flush_tlb_user = xen_flush_tlb,
	.flush_tlb_kernel = xen_flush_tlb,
	.flush_tlb_single = xen_flush_tlb_single,
J
Jeremy Fitzhardinge 已提交
1292
	.flush_tlb_others = xen_flush_tlb_others,
1293 1294 1295 1296

	.pte_update = paravirt_nop,
	.pte_update_defer = paravirt_nop,

1297 1298
	.pgd_alloc = xen_pgd_alloc,
	.pgd_free = xen_pgd_free,
1299

1300 1301 1302 1303 1304
	.alloc_pte = xen_alloc_pte_init,
	.release_pte = xen_release_pte_init,
	.alloc_pmd = xen_alloc_pte_init,
	.alloc_pmd_clone = paravirt_nop,
	.release_pmd = xen_release_pte_init,
1305 1306 1307 1308

#ifdef CONFIG_HIGHPTE
	.kmap_atomic_pte = xen_kmap_atomic_pte,
#endif
1309

1310 1311 1312
#ifdef CONFIG_X86_64
	.set_pte = xen_set_pte,
#else
1313
	.set_pte = xen_set_pte_init,
1314
#endif
J
Jeremy Fitzhardinge 已提交
1315
	.set_pte_at = xen_set_pte_at,
1316
	.set_pmd = xen_set_pmd_hyper,
J
Jeremy Fitzhardinge 已提交
1317

1318 1319 1320
	.ptep_modify_prot_start = __ptep_modify_prot_start,
	.ptep_modify_prot_commit = __ptep_modify_prot_commit,

J
Jeremy Fitzhardinge 已提交
1321
	.pte_val = xen_pte_val,
1322
	.pte_flags = native_pte_val,
J
Jeremy Fitzhardinge 已提交
1323 1324 1325 1326 1327
	.pgd_val = xen_pgd_val,

	.make_pte = xen_make_pte,
	.make_pgd = xen_make_pgd,

1328
#ifdef CONFIG_X86_PAE
J
Jeremy Fitzhardinge 已提交
1329 1330 1331 1332
	.set_pte_atomic = xen_set_pte_atomic,
	.set_pte_present = xen_set_pte_at,
	.pte_clear = xen_pte_clear,
	.pmd_clear = xen_pmd_clear,
1333 1334
#endif	/* CONFIG_X86_PAE */
	.set_pud = xen_set_pud_hyper,
J
Jeremy Fitzhardinge 已提交
1335 1336 1337 1338

	.make_pmd = xen_make_pmd,
	.pmd_val = xen_pmd_val,

1339 1340 1341 1342 1343 1344 1345 1346 1347
#if PAGETABLE_LEVELS == 4
	.pud_val = xen_pud_val,
	.make_pud = xen_make_pud,
	.set_pgd = xen_set_pgd_hyper,

	.alloc_pud = xen_alloc_pte_init,
	.release_pud = xen_release_pte_init,
#endif	/* PAGETABLE_LEVELS == 4 */

J
Jeremy Fitzhardinge 已提交
1348 1349 1350 1351
	.activate_mm = xen_activate_mm,
	.dup_mmap = xen_dup_mmap,
	.exit_mmap = xen_exit_mmap,

1352 1353 1354 1355
	.lazy_mode = {
		.enter = paravirt_enter_lazy_mmu,
		.leave = xen_leave_lazy,
	},
1356 1357

	.set_fixmap = xen_set_fixmap,
1358 1359
};

J
Jeremy Fitzhardinge 已提交
1360 1361
static void xen_reboot(int reason)
{
J
Jeremy Fitzhardinge 已提交
1362 1363
	struct sched_shutdown r = { .reason = reason };

J
Jeremy Fitzhardinge 已提交
1364 1365 1366 1367
#ifdef CONFIG_SMP
	smp_send_stop();
#endif

J
Jeremy Fitzhardinge 已提交
1368
	if (HYPERVISOR_sched_op(SCHEDOP_shutdown, &r))
J
Jeremy Fitzhardinge 已提交
1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
		BUG();
}

static void xen_restart(char *msg)
{
	xen_reboot(SHUTDOWN_reboot);
}

static void xen_emergency_restart(void)
{
	xen_reboot(SHUTDOWN_reboot);
}

static void xen_machine_halt(void)
{
	xen_reboot(SHUTDOWN_poweroff);
}

static void xen_crash_shutdown(struct pt_regs *regs)
{
	xen_reboot(SHUTDOWN_crash);
}

static const struct machine_ops __initdata xen_machine_ops = {
	.restart = xen_restart,
	.halt = xen_machine_halt,
	.power_off = xen_machine_halt,
	.shutdown = xen_machine_halt,
	.crash_shutdown = xen_crash_shutdown,
	.emergency_restart = xen_emergency_restart,
};

1401

1402 1403
static void __init xen_reserve_top(void)
{
1404
#ifdef CONFIG_X86_32
1405 1406 1407 1408 1409 1410 1411
	unsigned long top = HYPERVISOR_VIRT_START;
	struct xen_platform_parameters pp;

	if (HYPERVISOR_xen_version(XENVER_platform_parameters, &pp) == 0)
		top = pp.virt_start;

	reserve_top_address(-top + 2 * PAGE_SIZE);
1412
#endif	/* CONFIG_X86_32 */
1413 1414
}

J
Jeremy Fitzhardinge 已提交
1415 1416 1417 1418 1419 1420
/*
 * Like __va(), but returns address in the kernel mapping (which is
 * all we have until the physical memory mapping has been set up.
 */
static void *__ka(phys_addr_t paddr)
{
1421
#ifdef CONFIG_X86_64
J
Jeremy Fitzhardinge 已提交
1422
	return (void *)(paddr + __START_KERNEL_map);
1423 1424 1425
#else
	return __va(paddr);
#endif
J
Jeremy Fitzhardinge 已提交
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
}

/* Convert a machine address to physical address */
static unsigned long m2p(phys_addr_t maddr)
{
	phys_addr_t paddr;

	maddr &= PTE_MASK;
	paddr = mfn_to_pfn(maddr >> PAGE_SHIFT) << PAGE_SHIFT;

	return paddr;
}

/* Convert a machine address to kernel virtual */
static void *m2v(phys_addr_t maddr)
{
	return __ka(m2p(maddr));
}

1445
#ifdef CONFIG_X86_64
J
Jeremy Fitzhardinge 已提交
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
static void walk(pgd_t *pgd, unsigned long addr)
{
	unsigned l4idx = pgd_index(addr);
	unsigned l3idx = pud_index(addr);
	unsigned l2idx = pmd_index(addr);
	unsigned l1idx = pte_index(addr);
	pgd_t l4;
	pud_t l3;
	pmd_t l2;
	pte_t l1;

	xen_raw_printk("walk %p, %lx -> %d %d %d %d\n",
		       pgd, addr, l4idx, l3idx, l2idx, l1idx);

	l4 = pgd[l4idx];
	xen_raw_printk("  l4: %016lx\n", l4.pgd);
	xen_raw_printk("      %016lx\n", pgd_val(l4));

	l3 = ((pud_t *)(m2v(l4.pgd)))[l3idx];
	xen_raw_printk("  l3: %016lx\n", l3.pud);
	xen_raw_printk("      %016lx\n", pud_val(l3));

	l2 = ((pmd_t *)(m2v(l3.pud)))[l2idx];
	xen_raw_printk("  l2: %016lx\n", l2.pmd);
	xen_raw_printk("      %016lx\n", pmd_val(l2));

	l1 = ((pte_t *)(m2v(l2.pmd)))[l1idx];
	xen_raw_printk("  l1: %016lx\n", l1.pte);
	xen_raw_printk("      %016lx\n", pte_val(l1));
}
1476
#endif
J
Jeremy Fitzhardinge 已提交
1477 1478 1479 1480 1481 1482

static void set_page_prot(void *addr, pgprot_t prot)
{
	unsigned long pfn = __pa(addr) >> PAGE_SHIFT;
	pte_t pte = pfn_pte(pfn, prot);

1483
	xen_raw_printk("addr=%p pfn=%lx mfn=%lx prot=%016llx pte=%016llx\n",
J
Jeremy Fitzhardinge 已提交
1484 1485 1486 1487 1488 1489 1490
		       addr, pfn, get_phys_to_machine(pfn),
		       pgprot_val(prot), pte.pte);

	if (HYPERVISOR_update_va_mapping((unsigned long)addr, pte, 0))
		BUG();
}

1491
static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn)
1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
{
	unsigned pmdidx, pteidx;
	unsigned ident_pte;
	unsigned long pfn;

	ident_pte = 0;
	pfn = 0;
	for(pmdidx = 0; pmdidx < PTRS_PER_PMD && pfn < max_pfn; pmdidx++) {
		pte_t *pte_page;

		/* Reuse or allocate a page of ptes */
1503 1504
		if (pmd_present(pmd[pmdidx]))
			pte_page = m2v(pmd[pmdidx].pmd);
1505 1506 1507 1508 1509 1510 1511 1512
		else {
			/* Check for free pte pages */
			if (ident_pte == ARRAY_SIZE(level1_ident_pgt))
				break;

			pte_page = &level1_ident_pgt[ident_pte];
			ident_pte += PTRS_PER_PTE;

1513
			pmd[pmdidx] = __pmd(__pa(pte_page) | _PAGE_TABLE);
1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
		}

		/* Install mappings */
		for(pteidx = 0; pteidx < PTRS_PER_PTE; pteidx++, pfn++) {
			pte_t pte;

			if (pfn > max_pfn_mapped)
				max_pfn_mapped = pfn;

			if (!pte_none(pte_page[pteidx]))
				continue;

			pte = pfn_pte(pfn, PAGE_KERNEL_EXEC);
			pte_page[pteidx] = pte;
		}
	}

	for(pteidx = 0; pteidx < ident_pte; pteidx += PTRS_PER_PTE)
		set_page_prot(&level1_ident_pgt[pteidx], PAGE_KERNEL_RO);
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546

	set_page_prot(pmd, PAGE_KERNEL_RO);
}

#ifdef CONFIG_X86_64
static void convert_pfn_mfn(void *v)
{
	pte_t *pte = v;
	int i;

	/* All levels are converted the same way, so just treat them
	   as ptes. */
	for(i = 0; i < PTRS_PER_PTE; i++)
		pte[i] = xen_make_pte(pte[i].pte);
1547 1548
}

J
Jeremy Fitzhardinge 已提交
1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
/*
 * Set up the inital kernel pagetable.
 *
 * We can construct this by grafting the Xen provided pagetable into
 * head_64.S's preconstructed pagetables.  We copy the Xen L2's into
 * level2_ident_pgt, level2_kernel_pgt and level2_fixmap_pgt.  This
 * means that only the kernel has a physical mapping to start with -
 * but that's enough to get __va working.  We need to fill in the rest
 * of the physical mapping once some sort of allocator has been set
 * up.
 */
1560
static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pfn)
J
Jeremy Fitzhardinge 已提交
1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582
{
	pud_t *l3;
	pmd_t *l2;

	/* Zap identity mapping */
	init_level4_pgt[0] = __pgd(0);

	/* Pre-constructed entries are in pfn, so convert to mfn */
	convert_pfn_mfn(init_level4_pgt);
	convert_pfn_mfn(level3_ident_pgt);
	convert_pfn_mfn(level3_kernel_pgt);

	l3 = m2v(pgd[pgd_index(__START_KERNEL_map)].pgd);
	l2 = m2v(l3[pud_index(__START_KERNEL_map)].pud);

	memcpy(level2_ident_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);
	memcpy(level2_kernel_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);

	l3 = m2v(pgd[pgd_index(__START_KERNEL_map + PMD_SIZE)].pgd);
	l2 = m2v(l3[pud_index(__START_KERNEL_map + PMD_SIZE)].pud);
	memcpy(level2_fixmap_pgt, l2, sizeof(pmd_t) * PTRS_PER_PMD);

1583
	/* Set up identity map */
1584
	xen_map_identity_early(level2_ident_pgt, max_pfn);
1585

J
Jeremy Fitzhardinge 已提交
1586 1587 1588 1589
	/* Make pagetable pieces RO */
	set_page_prot(init_level4_pgt, PAGE_KERNEL_RO);
	set_page_prot(level3_ident_pgt, PAGE_KERNEL_RO);
	set_page_prot(level3_kernel_pgt, PAGE_KERNEL_RO);
1590
	set_page_prot(level3_user_vsyscall, PAGE_KERNEL_RO);
J
Jeremy Fitzhardinge 已提交
1591 1592 1593 1594
	set_page_prot(level2_kernel_pgt, PAGE_KERNEL_RO);
	set_page_prot(level2_fixmap_pgt, PAGE_KERNEL_RO);

	/* Pin down new L4 */
1595 1596
	pin_pagetable_pfn(MMUEXT_PIN_L4_TABLE,
			  PFN_DOWN(__pa_symbol(init_level4_pgt)));
J
Jeremy Fitzhardinge 已提交
1597 1598 1599 1600 1601 1602

	/* Unpin Xen-provided one */
	pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd)));

	/* Switch over */
	pgd = init_level4_pgt;
1603 1604 1605 1606 1607 1608 1609 1610 1611

	/*
	 * At this stage there can be no user pgd, and no page
	 * structure to attach it to, so make sure we just set kernel
	 * pgd.
	 */
	xen_mc_batch();
	__xen_write_cr3(true, __pa(pgd));
	xen_mc_issue(PARAVIRT_LAZY_CPU);
J
Jeremy Fitzhardinge 已提交
1612

1613 1614 1615 1616
	reserve_early(__pa(xen_start_info->pt_base),
		      __pa(xen_start_info->pt_base +
			   xen_start_info->nr_pt_frames * PAGE_SIZE),
		      "XEN PAGETABLES");
J
Jeremy Fitzhardinge 已提交
1617 1618 1619

	return pgd;
}
1620 1621 1622
#else	/* !CONFIG_X86_64 */
static pmd_t level2_kernel_pgt[PTRS_PER_PMD] __page_aligned_bss;

1623
static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pfn)
J
Jeremy Fitzhardinge 已提交
1624
{
1625 1626
	pmd_t *kernel_pmd;

J
Jeremy Fitzhardinge 已提交
1627 1628 1629 1630
	init_pg_tables_start = __pa(pgd);
	init_pg_tables_end = __pa(pgd) + xen_start_info->nr_pt_frames*PAGE_SIZE;
	max_pfn_mapped = PFN_DOWN(init_pg_tables_end + 512*1024);

1631 1632
	kernel_pmd = m2v(pgd[KERNEL_PGD_BOUNDARY].pgd);
	memcpy(level2_kernel_pgt, kernel_pmd, sizeof(pmd_t) * PTRS_PER_PMD);
1633

1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650
	xen_map_identity_early(level2_kernel_pgt, max_pfn);

	memcpy(swapper_pg_dir, pgd, sizeof(pgd_t) * PTRS_PER_PGD);
	set_pgd(&swapper_pg_dir[KERNEL_PGD_BOUNDARY],
			__pgd(__pa(level2_kernel_pgt) | _PAGE_PRESENT));

	set_page_prot(level2_kernel_pgt, PAGE_KERNEL_RO);
	set_page_prot(swapper_pg_dir, PAGE_KERNEL_RO);
	set_page_prot(empty_zero_page, PAGE_KERNEL_RO);

	pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd)));

	xen_write_cr3(__pa(swapper_pg_dir));

	pin_pagetable_pfn(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(swapper_pg_dir)));

	return swapper_pg_dir;
J
Jeremy Fitzhardinge 已提交
1651 1652 1653
}
#endif	/* CONFIG_X86_64 */

1654 1655 1656 1657 1658 1659 1660 1661
/* First C function to be called on Xen boot */
asmlinkage void __init xen_start_kernel(void)
{
	pgd_t *pgd;

	if (!xen_start_info)
		return;

J
Jeremy Fitzhardinge 已提交
1662
	BUG_ON(memcmp(xen_start_info->magic, "xen-3", 5) != 0);
1663

1664 1665
	xen_setup_features();

1666
	/* Install Xen paravirt ops */
1667 1668 1669 1670 1671 1672 1673 1674
	pv_info = xen_info;
	pv_init_ops = xen_init_ops;
	pv_time_ops = xen_time_ops;
	pv_cpu_ops = xen_cpu_ops;
	pv_irq_ops = xen_irq_ops;
	pv_apic_ops = xen_apic_ops;
	pv_mmu_ops = xen_mmu_ops;

1675 1676 1677 1678 1679
	if (xen_feature(XENFEAT_mmu_pt_update_preserve_ad)) {
		pv_mmu_ops.ptep_modify_prot_start = xen_ptep_modify_prot_start;
		pv_mmu_ops.ptep_modify_prot_commit = xen_ptep_modify_prot_commit;
	}

J
Jeremy Fitzhardinge 已提交
1680 1681
	machine_ops = xen_machine_ops;

1682 1683 1684
#ifdef CONFIG_X86_64
	/* Disable until direct per-cpu data access. */
	have_vcpu_info_placement = 0;
1685
	x86_64_init_pda();
1686 1687
#endif

1688
	xen_smp_init();
1689 1690 1691

	/* Get mfn list */
	if (!xen_feature(XENFEAT_auto_translated_physmap))
1692
		xen_build_dynamic_phys_to_machine();
1693 1694 1695

	pgd = (pgd_t *)xen_start_info->pt_base;

J
Jeremy Fitzhardinge 已提交
1696 1697 1698 1699 1700 1701 1702 1703 1704 1705
	/* Prevent unwanted bits from being set in PTEs. */
	__supported_pte_mask &= ~_PAGE_GLOBAL;
	if (!is_initial_xendomain())
		__supported_pte_mask &= ~(_PAGE_PWT | _PAGE_PCD);

	/* Don't do the full vcpu_info placement stuff until we have a
	   possible map and a non-dummy shared_info. */
	per_cpu(xen_vcpu, 0) = &HYPERVISOR_shared_info->vcpu_info[0];

	xen_raw_console_write("mapping kernel into physical memory\n");
1706
	pgd = xen_setup_kernel_pagetable(pgd, xen_start_info->nr_pages);
1707

J
Jeremy Fitzhardinge 已提交
1708
	init_mm.pgd = pgd;
1709 1710 1711

	/* keep using Xen gdt for now; no urgent need to change it */

1712
	pv_info.kernel_rpl = 1;
1713
	if (xen_feature(XENFEAT_supervisor_mode_kernel))
1714
		pv_info.kernel_rpl = 0;
1715 1716

	/* set the limit of our address space */
1717
	xen_reserve_top();
1718

1719
#ifdef CONFIG_X86_32
1720 1721 1722 1723
	/* set up basic CPUID stuff */
	cpu_detect(&new_cpu_data);
	new_cpu_data.hard_math = 1;
	new_cpu_data.x86_capability[0] = cpuid_edx(1);
1724
#endif
1725 1726

	/* Poke various useful things into boot_params */
1727 1728 1729 1730
	boot_params.hdr.type_of_loader = (9 << 4) | 0;
	boot_params.hdr.ramdisk_image = xen_start_info->mod_start
		? __pa(xen_start_info->mod_start) : 0;
	boot_params.hdr.ramdisk_size = xen_start_info->mod_len;
1731
	boot_params.hdr.cmd_line_ptr = __pa(xen_start_info->cmd_line);
1732

1733
	if (!is_initial_xendomain()) {
1734
		add_preferred_console("xenboot", 0, NULL);
1735
		add_preferred_console("tty", 0, NULL);
1736
		add_preferred_console("hvc", 0, NULL);
1737
	}
1738

J
Jeremy Fitzhardinge 已提交
1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749
	xen_raw_console_write("about to get started...\n");

#if 0
	xen_raw_printk("&boot_params=%p __pa(&boot_params)=%lx __va(__pa(&boot_params))=%lx\n",
		       &boot_params, __pa_symbol(&boot_params),
		       __va(__pa_symbol(&boot_params)));

	walk(pgd, &boot_params);
	walk(pgd, __va(__pa(&boot_params)));
#endif

1750
	/* Start the world */
1751
#ifdef CONFIG_X86_32
1752
	i386_start_kernel();
1753
#else
J
Jeremy Fitzhardinge 已提交
1754
	x86_64_start_reservations((char *)__pa_symbol(&boot_params));
1755
#endif
1756
}