1
linux/arch/powerpc/kvm/booke.c
Hollis Blanchard 7924bd4109 KVM: ppc: directly insert shadow mappings into the hardware TLB
Formerly, we used to maintain a per-vcpu shadow TLB and on every entry to the
guest would load this array into the hardware TLB. This consumed 1280 bytes of
memory (64 entries of 16 bytes plus a struct page pointer each), and also
required some assembly to loop over the array on every entry.

Instead of saving a copy in memory, we can just store shadow mappings directly
into the hardware TLB, accepting that the host kernel will clobber these as
part of the normal 440 TLB round robin. When we do that we need less than half
the memory, and we have decreased the exit handling time for all guest exits,
at the cost of increased number of TLB misses because the host overwrites some
guest entries.

These savings will be increased on processors with larger TLBs or which
implement intelligent flush instructions like tlbivax (which will avoid the
need to walk arrays in software).

In addition to that and to the code simplification, we have a greater chance of
leaving other host userspace mappings in the TLB, instead of forcing all
subsequent tasks to re-fault all their mappings.

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2008-12-31 16:55:09 +02:00

558 lines
15 KiB
C

/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License, version 2, as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* Copyright IBM Corp. 2007
*
* Authors: Hollis Blanchard <hollisb@us.ibm.com>
* Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
*/
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/kvm_host.h>
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/fs.h>
#include <asm/cputable.h>
#include <asm/uaccess.h>
#include <asm/kvm_ppc.h>
#include <asm/cacheflush.h>
#include <asm/kvm_44x.h>
#include "booke.h"
#include "44x_tlb.h"
unsigned long kvmppc_booke_handlers;
#define VM_STAT(x) offsetof(struct kvm, stat.x), KVM_STAT_VM
#define VCPU_STAT(x) offsetof(struct kvm_vcpu, stat.x), KVM_STAT_VCPU
struct kvm_stats_debugfs_item debugfs_entries[] = {
{ "mmio", VCPU_STAT(mmio_exits) },
{ "dcr", VCPU_STAT(dcr_exits) },
{ "sig", VCPU_STAT(signal_exits) },
{ "itlb_r", VCPU_STAT(itlb_real_miss_exits) },
{ "itlb_v", VCPU_STAT(itlb_virt_miss_exits) },
{ "dtlb_r", VCPU_STAT(dtlb_real_miss_exits) },
{ "dtlb_v", VCPU_STAT(dtlb_virt_miss_exits) },
{ "sysc", VCPU_STAT(syscall_exits) },
{ "isi", VCPU_STAT(isi_exits) },
{ "dsi", VCPU_STAT(dsi_exits) },
{ "inst_emu", VCPU_STAT(emulated_inst_exits) },
{ "dec", VCPU_STAT(dec_exits) },
{ "ext_intr", VCPU_STAT(ext_intr_exits) },
{ "halt_wakeup", VCPU_STAT(halt_wakeup) },
{ NULL }
};
/* TODO: use vcpu_printf() */
void kvmppc_dump_vcpu(struct kvm_vcpu *vcpu)
{
int i;
printk("pc: %08lx msr: %08lx\n", vcpu->arch.pc, vcpu->arch.msr);
printk("lr: %08lx ctr: %08lx\n", vcpu->arch.lr, vcpu->arch.ctr);
printk("srr0: %08lx srr1: %08lx\n", vcpu->arch.srr0, vcpu->arch.srr1);
printk("exceptions: %08lx\n", vcpu->arch.pending_exceptions);
for (i = 0; i < 32; i += 4) {
printk("gpr%02d: %08lx %08lx %08lx %08lx\n", i,
vcpu->arch.gpr[i],
vcpu->arch.gpr[i+1],
vcpu->arch.gpr[i+2],
vcpu->arch.gpr[i+3]);
}
}
static void kvmppc_booke_queue_irqprio(struct kvm_vcpu *vcpu,
unsigned int priority)
{
set_bit(priority, &vcpu->arch.pending_exceptions);
}
void kvmppc_core_queue_program(struct kvm_vcpu *vcpu)
{
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_PROGRAM);
}
void kvmppc_core_queue_dec(struct kvm_vcpu *vcpu)
{
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_DECREMENTER);
}
int kvmppc_core_pending_dec(struct kvm_vcpu *vcpu)
{
return test_bit(BOOKE_IRQPRIO_DECREMENTER, &vcpu->arch.pending_exceptions);
}
void kvmppc_core_queue_external(struct kvm_vcpu *vcpu,
struct kvm_interrupt *irq)
{
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_EXTERNAL);
}
/* Deliver the interrupt of the corresponding priority, if possible. */
static int kvmppc_booke_irqprio_deliver(struct kvm_vcpu *vcpu,
unsigned int priority)
{
int allowed = 0;
ulong msr_mask;
switch (priority) {
case BOOKE_IRQPRIO_PROGRAM:
case BOOKE_IRQPRIO_DTLB_MISS:
case BOOKE_IRQPRIO_ITLB_MISS:
case BOOKE_IRQPRIO_SYSCALL:
case BOOKE_IRQPRIO_DATA_STORAGE:
case BOOKE_IRQPRIO_INST_STORAGE:
case BOOKE_IRQPRIO_FP_UNAVAIL:
case BOOKE_IRQPRIO_AP_UNAVAIL:
case BOOKE_IRQPRIO_ALIGNMENT:
allowed = 1;
msr_mask = MSR_CE|MSR_ME|MSR_DE;
break;
case BOOKE_IRQPRIO_CRITICAL:
case BOOKE_IRQPRIO_WATCHDOG:
allowed = vcpu->arch.msr & MSR_CE;
msr_mask = MSR_ME;
break;
case BOOKE_IRQPRIO_MACHINE_CHECK:
allowed = vcpu->arch.msr & MSR_ME;
msr_mask = 0;
break;
case BOOKE_IRQPRIO_EXTERNAL:
case BOOKE_IRQPRIO_DECREMENTER:
case BOOKE_IRQPRIO_FIT:
allowed = vcpu->arch.msr & MSR_EE;
msr_mask = MSR_CE|MSR_ME|MSR_DE;
break;
case BOOKE_IRQPRIO_DEBUG:
allowed = vcpu->arch.msr & MSR_DE;
msr_mask = MSR_ME;
break;
}
if (allowed) {
vcpu->arch.srr0 = vcpu->arch.pc;
vcpu->arch.srr1 = vcpu->arch.msr;
vcpu->arch.pc = vcpu->arch.ivpr | vcpu->arch.ivor[priority];
kvmppc_set_msr(vcpu, vcpu->arch.msr & msr_mask);
clear_bit(priority, &vcpu->arch.pending_exceptions);
}
return allowed;
}
/* Check pending exceptions and deliver one, if possible. */
void kvmppc_core_deliver_interrupts(struct kvm_vcpu *vcpu)
{
unsigned long *pending = &vcpu->arch.pending_exceptions;
unsigned int priority;
priority = __ffs(*pending);
while (priority <= BOOKE_MAX_INTERRUPT) {
if (kvmppc_booke_irqprio_deliver(vcpu, priority))
break;
priority = find_next_bit(pending,
BITS_PER_BYTE * sizeof(*pending),
priority + 1);
}
}
/**
* kvmppc_handle_exit
*
* Return value is in the form (errcode<<2 | RESUME_FLAG_HOST | RESUME_FLAG_NV)
*/
int kvmppc_handle_exit(struct kvm_run *run, struct kvm_vcpu *vcpu,
unsigned int exit_nr)
{
enum emulation_result er;
int r = RESUME_HOST;
local_irq_enable();
run->exit_reason = KVM_EXIT_UNKNOWN;
run->ready_for_interrupt_injection = 1;
switch (exit_nr) {
case BOOKE_INTERRUPT_MACHINE_CHECK:
printk("MACHINE CHECK: %lx\n", mfspr(SPRN_MCSR));
kvmppc_dump_vcpu(vcpu);
r = RESUME_HOST;
break;
case BOOKE_INTERRUPT_EXTERNAL:
vcpu->stat.ext_intr_exits++;
if (need_resched())
cond_resched();
r = RESUME_GUEST;
break;
case BOOKE_INTERRUPT_DECREMENTER:
/* Since we switched IVPR back to the host's value, the host
* handled this interrupt the moment we enabled interrupts.
* Now we just offer it a chance to reschedule the guest. */
vcpu->stat.dec_exits++;
if (need_resched())
cond_resched();
r = RESUME_GUEST;
break;
case BOOKE_INTERRUPT_PROGRAM:
if (vcpu->arch.msr & MSR_PR) {
/* Program traps generated by user-level software must be handled
* by the guest kernel. */
vcpu->arch.esr = vcpu->arch.fault_esr;
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_PROGRAM);
r = RESUME_GUEST;
break;
}
er = kvmppc_emulate_instruction(run, vcpu);
switch (er) {
case EMULATE_DONE:
/* Future optimization: only reload non-volatiles if
* they were actually modified by emulation. */
vcpu->stat.emulated_inst_exits++;
r = RESUME_GUEST_NV;
break;
case EMULATE_DO_DCR:
run->exit_reason = KVM_EXIT_DCR;
vcpu->stat.dcr_exits++;
r = RESUME_HOST;
break;
case EMULATE_FAIL:
/* XXX Deliver Program interrupt to guest. */
printk(KERN_CRIT "%s: emulation at %lx failed (%08x)\n",
__func__, vcpu->arch.pc, vcpu->arch.last_inst);
/* For debugging, encode the failing instruction and
* report it to userspace. */
run->hw.hardware_exit_reason = ~0ULL << 32;
run->hw.hardware_exit_reason |= vcpu->arch.last_inst;
r = RESUME_HOST;
break;
default:
BUG();
}
break;
case BOOKE_INTERRUPT_FP_UNAVAIL:
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_FP_UNAVAIL);
r = RESUME_GUEST;
break;
case BOOKE_INTERRUPT_DATA_STORAGE:
vcpu->arch.dear = vcpu->arch.fault_dear;
vcpu->arch.esr = vcpu->arch.fault_esr;
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_DATA_STORAGE);
vcpu->stat.dsi_exits++;
r = RESUME_GUEST;
break;
case BOOKE_INTERRUPT_INST_STORAGE:
vcpu->arch.esr = vcpu->arch.fault_esr;
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_INST_STORAGE);
vcpu->stat.isi_exits++;
r = RESUME_GUEST;
break;
case BOOKE_INTERRUPT_SYSCALL:
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_SYSCALL);
vcpu->stat.syscall_exits++;
r = RESUME_GUEST;
break;
/* XXX move to a 440-specific file. */
case BOOKE_INTERRUPT_DTLB_MISS: {
struct kvmppc_vcpu_44x *vcpu_44x = to_44x(vcpu);
struct kvmppc_44x_tlbe *gtlbe;
unsigned long eaddr = vcpu->arch.fault_dear;
int gtlb_index;
gfn_t gfn;
/* Check the guest TLB. */
gtlb_index = kvmppc_44x_dtlb_index(vcpu, eaddr);
if (gtlb_index < 0) {
/* The guest didn't have a mapping for it. */
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_DTLB_MISS);
vcpu->arch.dear = vcpu->arch.fault_dear;
vcpu->arch.esr = vcpu->arch.fault_esr;
vcpu->stat.dtlb_real_miss_exits++;
r = RESUME_GUEST;
break;
}
gtlbe = &vcpu_44x->guest_tlb[gtlb_index];
vcpu->arch.paddr_accessed = tlb_xlate(gtlbe, eaddr);
gfn = vcpu->arch.paddr_accessed >> PAGE_SHIFT;
if (kvm_is_visible_gfn(vcpu->kvm, gfn)) {
/* The guest TLB had a mapping, but the shadow TLB
* didn't, and it is RAM. This could be because:
* a) the entry is mapping the host kernel, or
* b) the guest used a large mapping which we're faking
* Either way, we need to satisfy the fault without
* invoking the guest. */
kvmppc_mmu_map(vcpu, eaddr, vcpu->arch.paddr_accessed, gtlbe->tid,
gtlbe->word2, get_tlb_bytes(gtlbe), gtlb_index);
vcpu->stat.dtlb_virt_miss_exits++;
r = RESUME_GUEST;
} else {
/* Guest has mapped and accessed a page which is not
* actually RAM. */
r = kvmppc_emulate_mmio(run, vcpu);
vcpu->stat.mmio_exits++;
}
break;
}
/* XXX move to a 440-specific file. */
case BOOKE_INTERRUPT_ITLB_MISS: {
struct kvmppc_vcpu_44x *vcpu_44x = to_44x(vcpu);
struct kvmppc_44x_tlbe *gtlbe;
unsigned long eaddr = vcpu->arch.pc;
gpa_t gpaddr;
gfn_t gfn;
int gtlb_index;
r = RESUME_GUEST;
/* Check the guest TLB. */
gtlb_index = kvmppc_44x_itlb_index(vcpu, eaddr);
if (gtlb_index < 0) {
/* The guest didn't have a mapping for it. */
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_ITLB_MISS);
vcpu->stat.itlb_real_miss_exits++;
break;
}
vcpu->stat.itlb_virt_miss_exits++;
gtlbe = &vcpu_44x->guest_tlb[gtlb_index];
gpaddr = tlb_xlate(gtlbe, eaddr);
gfn = gpaddr >> PAGE_SHIFT;
if (kvm_is_visible_gfn(vcpu->kvm, gfn)) {
/* The guest TLB had a mapping, but the shadow TLB
* didn't. This could be because:
* a) the entry is mapping the host kernel, or
* b) the guest used a large mapping which we're faking
* Either way, we need to satisfy the fault without
* invoking the guest. */
kvmppc_mmu_map(vcpu, eaddr, gpaddr, gtlbe->tid,
gtlbe->word2, get_tlb_bytes(gtlbe), gtlb_index);
} else {
/* Guest mapped and leaped at non-RAM! */
kvmppc_booke_queue_irqprio(vcpu, BOOKE_IRQPRIO_MACHINE_CHECK);
}
break;
}
case BOOKE_INTERRUPT_DEBUG: {
u32 dbsr;
vcpu->arch.pc = mfspr(SPRN_CSRR0);
/* clear IAC events in DBSR register */
dbsr = mfspr(SPRN_DBSR);
dbsr &= DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4;
mtspr(SPRN_DBSR, dbsr);
run->exit_reason = KVM_EXIT_DEBUG;
r = RESUME_HOST;
break;
}
default:
printk(KERN_EMERG "exit_nr %d\n", exit_nr);
BUG();
}
local_irq_disable();
kvmppc_core_deliver_interrupts(vcpu);
if (!(r & RESUME_HOST)) {
/* To avoid clobbering exit_reason, only check for signals if
* we aren't already exiting to userspace for some other
* reason. */
if (signal_pending(current)) {
run->exit_reason = KVM_EXIT_INTR;
r = (-EINTR << 2) | RESUME_HOST | (r & RESUME_FLAG_NV);
vcpu->stat.signal_exits++;
}
}
return r;
}
/* Initial guest state: 16MB mapping 0 -> 0, PC = 0, MSR = 0, R1 = 16MB */
int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
{
vcpu->arch.pc = 0;
vcpu->arch.msr = 0;
vcpu->arch.gpr[1] = (16<<20) - 8; /* -8 for the callee-save LR slot */
vcpu->arch.shadow_pid = 1;
/* Eye-catching number so we know if the guest takes an interrupt
* before it's programmed its own IVPR. */
vcpu->arch.ivpr = 0x55550000;
return kvmppc_core_vcpu_setup(vcpu);
}
int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
{
int i;
regs->pc = vcpu->arch.pc;
regs->cr = vcpu->arch.cr;
regs->ctr = vcpu->arch.ctr;
regs->lr = vcpu->arch.lr;
regs->xer = vcpu->arch.xer;
regs->msr = vcpu->arch.msr;
regs->srr0 = vcpu->arch.srr0;
regs->srr1 = vcpu->arch.srr1;
regs->pid = vcpu->arch.pid;
regs->sprg0 = vcpu->arch.sprg0;
regs->sprg1 = vcpu->arch.sprg1;
regs->sprg2 = vcpu->arch.sprg2;
regs->sprg3 = vcpu->arch.sprg3;
regs->sprg5 = vcpu->arch.sprg4;
regs->sprg6 = vcpu->arch.sprg5;
regs->sprg7 = vcpu->arch.sprg6;
for (i = 0; i < ARRAY_SIZE(regs->gpr); i++)
regs->gpr[i] = vcpu->arch.gpr[i];
return 0;
}
int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
{
int i;
vcpu->arch.pc = regs->pc;
vcpu->arch.cr = regs->cr;
vcpu->arch.ctr = regs->ctr;
vcpu->arch.lr = regs->lr;
vcpu->arch.xer = regs->xer;
kvmppc_set_msr(vcpu, regs->msr);
vcpu->arch.srr0 = regs->srr0;
vcpu->arch.srr1 = regs->srr1;
vcpu->arch.sprg0 = regs->sprg0;
vcpu->arch.sprg1 = regs->sprg1;
vcpu->arch.sprg2 = regs->sprg2;
vcpu->arch.sprg3 = regs->sprg3;
vcpu->arch.sprg5 = regs->sprg4;
vcpu->arch.sprg6 = regs->sprg5;
vcpu->arch.sprg7 = regs->sprg6;
for (i = 0; i < ARRAY_SIZE(vcpu->arch.gpr); i++)
vcpu->arch.gpr[i] = regs->gpr[i];
return 0;
}
int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
struct kvm_sregs *sregs)
{
return -ENOTSUPP;
}
int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
struct kvm_sregs *sregs)
{
return -ENOTSUPP;
}
int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
return -ENOTSUPP;
}
int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
{
return -ENOTSUPP;
}
int kvm_arch_vcpu_ioctl_translate(struct kvm_vcpu *vcpu,
struct kvm_translation *tr)
{
return kvmppc_core_vcpu_translate(vcpu, tr);
}
int kvmppc_booke_init(void)
{
unsigned long ivor[16];
unsigned long max_ivor = 0;
int i;
/* We install our own exception handlers by hijacking IVPR. IVPR must
* be 16-bit aligned, so we need a 64KB allocation. */
kvmppc_booke_handlers = __get_free_pages(GFP_KERNEL | __GFP_ZERO,
VCPU_SIZE_ORDER);
if (!kvmppc_booke_handlers)
return -ENOMEM;
/* XXX make sure our handlers are smaller than Linux's */
/* Copy our interrupt handlers to match host IVORs. That way we don't
* have to swap the IVORs on every guest/host transition. */
ivor[0] = mfspr(SPRN_IVOR0);
ivor[1] = mfspr(SPRN_IVOR1);
ivor[2] = mfspr(SPRN_IVOR2);
ivor[3] = mfspr(SPRN_IVOR3);
ivor[4] = mfspr(SPRN_IVOR4);
ivor[5] = mfspr(SPRN_IVOR5);
ivor[6] = mfspr(SPRN_IVOR6);
ivor[7] = mfspr(SPRN_IVOR7);
ivor[8] = mfspr(SPRN_IVOR8);
ivor[9] = mfspr(SPRN_IVOR9);
ivor[10] = mfspr(SPRN_IVOR10);
ivor[11] = mfspr(SPRN_IVOR11);
ivor[12] = mfspr(SPRN_IVOR12);
ivor[13] = mfspr(SPRN_IVOR13);
ivor[14] = mfspr(SPRN_IVOR14);
ivor[15] = mfspr(SPRN_IVOR15);
for (i = 0; i < 16; i++) {
if (ivor[i] > max_ivor)
max_ivor = ivor[i];
memcpy((void *)kvmppc_booke_handlers + ivor[i],
kvmppc_handlers_start + i * kvmppc_handler_len,
kvmppc_handler_len);
}
flush_icache_range(kvmppc_booke_handlers,
kvmppc_booke_handlers + max_ivor + kvmppc_handler_len);
return 0;
}
void __exit kvmppc_booke_exit(void)
{
free_pages(kvmppc_booke_handlers, VCPU_SIZE_ORDER);
kvm_exit();
}