#include "qemu-common.h"
#include "sysemu.h"
#include "kvm.h"
+#include "kvm_i386.h"
#include "cpu.h"
#include "gdbstub.h"
#include "host-utils.h"
#include "hw/apic.h"
#include "ioport.h"
#include "hyperv.h"
+#include "hw/pci.h"
//#define DEBUG_KVM
static bool has_msr_hsave_pa;
static bool has_msr_tsc_deadline;
static bool has_msr_async_pf_en;
+static bool has_msr_pv_eoi_en;
static bool has_msr_misc_enable;
static int lm_capable_kernel;
+bool kvm_allows_irq0_override(void)
+{
+ return !kvm_irqchip_in_kernel() || kvm_has_gsi_routing();
+}
+
static struct kvm_cpuid2 *try_get_cpuid(KVMState *s, int max)
{
struct kvm_cpuid2 *cpuid;
* so add missing bits according to the AMD spec:
*/
cpuid_1_edx = kvm_arch_get_supported_cpuid(s, 1, 0, R_EDX);
- ret |= cpuid_1_edx & 0x183f7ff;
+ ret |= cpuid_1_edx & CPUID_EXT2_AMD_ALIASES;
break;
}
break;
return -ENOSYS;
}
-static void kvm_mce_inject(CPUState *env, target_phys_addr_t paddr, int code)
+static void kvm_mce_inject(CPUX86State *env, hwaddr paddr, int code)
{
uint64_t status = MCI_STATUS_VAL | MCI_STATUS_UC | MCI_STATUS_EN |
MCI_STATUS_MISCV | MCI_STATUS_ADDRV | MCI_STATUS_S;
exit(1);
}
-int kvm_arch_on_sigbus_vcpu(CPUState *env, int code, void *addr)
+int kvm_arch_on_sigbus_vcpu(CPUX86State *env, int code, void *addr)
{
ram_addr_t ram_addr;
- target_phys_addr_t paddr;
+ hwaddr paddr;
if ((env->mcg_cap & MCG_SER_P) && addr
&& (code == BUS_MCEERR_AR || code == BUS_MCEERR_AO)) {
{
if ((first_cpu->mcg_cap & MCG_SER_P) && addr && code == BUS_MCEERR_AO) {
ram_addr_t ram_addr;
- target_phys_addr_t paddr;
+ hwaddr paddr;
/* Hope we are lucky for AO MCE */
if (qemu_ram_addr_from_host(addr, &ram_addr) ||
return 0;
}
-static int kvm_inject_mce_oldstyle(CPUState *env)
+static int kvm_inject_mce_oldstyle(CPUX86State *env)
{
if (!kvm_has_vcpu_events() && env->exception_injected == EXCP12_MCHK) {
unsigned int bank, bank_num = env->mcg_cap & 0xff;
static void cpu_update_state(void *opaque, int running, RunState state)
{
- CPUState *env = opaque;
+ CPUX86State *env = opaque;
if (running) {
env->tsc_valid = false;
}
}
-int kvm_arch_init_vcpu(CPUState *env)
+int kvm_arch_init_vcpu(CPUX86State *env)
{
struct {
struct kvm_cpuid2 cpuid;
env->cpuid_features &= kvm_arch_get_supported_cpuid(s, 1, 0, R_EDX);
i = env->cpuid_ext_features & CPUID_EXT_HYPERVISOR;
+ j = env->cpuid_ext_features & CPUID_EXT_TSC_DEADLINE_TIMER;
env->cpuid_ext_features &= kvm_arch_get_supported_cpuid(s, 1, 0, R_ECX);
env->cpuid_ext_features |= i;
+ if (j && kvm_irqchip_in_kernel() &&
+ kvm_check_extension(s, KVM_CAP_TSC_DEADLINE_TIMER)) {
+ env->cpuid_ext_features |= CPUID_EXT_TSC_DEADLINE_TIMER;
+ }
env->cpuid_ext2_features &= kvm_arch_get_supported_cpuid(s, 0x80000001,
0, R_EDX);
has_msr_async_pf_en = c->eax & (1 << KVM_FEATURE_ASYNC_PF);
+ has_msr_pv_eoi_en = c->eax & (1 << KVM_FEATURE_PV_EOI);
+
cpu_x86_cpuid(env, 0, 0, &limit, &unused, &unused, &unused);
for (i = 0; i <= limit; i++) {
qemu_add_vm_change_state_handler(cpu_update_state, env);
+ cpuid_data.cpuid.padding = 0;
r = kvm_vcpu_ioctl(env, KVM_SET_CPUID2, &cpuid_data);
if (r) {
return r;
return 0;
}
-void kvm_arch_reset_vcpu(CPUState *env)
+void kvm_arch_reset_vcpu(CPUX86State *env)
{
+ X86CPU *cpu = x86_env_get_cpu(env);
+
env->exception_injected = -1;
env->interrupt_injected = -1;
env->xcr0 = 1;
if (kvm_irqchip_in_kernel()) {
- env->mp_state = cpu_is_bsp(env) ? KVM_MP_STATE_RUNNABLE :
+ env->mp_state = cpu_is_bsp(cpu) ? KVM_MP_STATE_RUNNABLE :
KVM_MP_STATE_UNINITIALIZED;
} else {
env->mp_state = KVM_MP_STATE_RUNNABLE;
lhs->g = (flags & DESC_G_MASK) != 0;
lhs->avl = (flags & DESC_AVL_MASK) != 0;
lhs->unusable = 0;
+ lhs->padding = 0;
}
static void get_seg(SegmentCache *lhs, const struct kvm_segment *rhs)
}
}
-static int kvm_getput_regs(CPUState *env, int set)
+static int kvm_getput_regs(CPUX86State *env, int set)
{
struct kvm_regs regs;
int ret = 0;
return ret;
}
-static int kvm_put_fpu(CPUState *env)
+static int kvm_put_fpu(CPUX86State *env)
{
struct kvm_fpu fpu;
int i;
#define XSAVE_XSTATE_BV 128
#define XSAVE_YMMH_SPACE 144
-static int kvm_put_xsave(CPUState *env)
+static int kvm_put_xsave(CPUX86State *env)
{
struct kvm_xsave* xsave = env->kvm_xsave_buf;
uint16_t cwd, swd, twd;
return r;
}
-static int kvm_put_xcrs(CPUState *env)
+static int kvm_put_xcrs(CPUX86State *env)
{
struct kvm_xcrs xcrs;
return kvm_vcpu_ioctl(env, KVM_SET_XCRS, &xcrs);
}
-static int kvm_put_sregs(CPUState *env)
+static int kvm_put_sregs(CPUX86State *env)
{
struct kvm_sregs sregs;
sregs.idt.limit = env->idt.limit;
sregs.idt.base = env->idt.base;
+ memset(sregs.idt.padding, 0, sizeof sregs.idt.padding);
sregs.gdt.limit = env->gdt.limit;
sregs.gdt.base = env->gdt.base;
+ memset(sregs.gdt.padding, 0, sizeof sregs.gdt.padding);
sregs.cr0 = env->cr[0];
sregs.cr2 = env->cr[2];
entry->data = value;
}
-static int kvm_put_msrs(CPUState *env, int level)
+static int kvm_put_msrs(CPUX86State *env, int level)
{
struct {
struct kvm_msrs info;
kvm_msr_entry_set(&msrs[n++], MSR_KVM_ASYNC_PF_EN,
env->async_pf_en_msr);
}
+ if (has_msr_pv_eoi_en) {
+ kvm_msr_entry_set(&msrs[n++], MSR_KVM_PV_EOI_EN,
+ env->pv_eoi_en_msr);
+ }
if (hyperv_hypercall_available()) {
kvm_msr_entry_set(&msrs[n++], HV_X64_MSR_GUEST_OS_ID, 0);
kvm_msr_entry_set(&msrs[n++], HV_X64_MSR_HYPERCALL, 0);
}
-static int kvm_get_fpu(CPUState *env)
+static int kvm_get_fpu(CPUX86State *env)
{
struct kvm_fpu fpu;
int i, ret;
return 0;
}
-static int kvm_get_xsave(CPUState *env)
+static int kvm_get_xsave(CPUX86State *env)
{
struct kvm_xsave* xsave = env->kvm_xsave_buf;
int ret, i;
return 0;
}
-static int kvm_get_xcrs(CPUState *env)
+static int kvm_get_xcrs(CPUX86State *env)
{
int i, ret;
struct kvm_xcrs xcrs;
return 0;
}
-static int kvm_get_sregs(CPUState *env)
+static int kvm_get_sregs(CPUX86State *env)
{
struct kvm_sregs sregs;
uint32_t hflags;
return 0;
}
-static int kvm_get_msrs(CPUState *env)
+static int kvm_get_msrs(CPUX86State *env)
{
struct {
struct kvm_msrs info;
if (has_msr_async_pf_en) {
msrs[n++].index = MSR_KVM_ASYNC_PF_EN;
}
+ if (has_msr_pv_eoi_en) {
+ msrs[n++].index = MSR_KVM_PV_EOI_EN;
+ }
if (env->mcg_cap) {
msrs[n++].index = MSR_MCG_STATUS;
case MSR_KVM_ASYNC_PF_EN:
env->async_pf_en_msr = msrs[i].data;
break;
+ case MSR_KVM_PV_EOI_EN:
+ env->pv_eoi_en_msr = msrs[i].data;
+ break;
}
}
return 0;
}
-static int kvm_put_mp_state(CPUState *env)
+static int kvm_put_mp_state(CPUX86State *env)
{
struct kvm_mp_state mp_state = { .mp_state = env->mp_state };
return kvm_vcpu_ioctl(env, KVM_SET_MP_STATE, &mp_state);
}
-static int kvm_get_mp_state(CPUState *env)
+static int kvm_get_mp_state(CPUX86State *env)
{
struct kvm_mp_state mp_state;
int ret;
return 0;
}
-static int kvm_get_apic(CPUState *env)
+static int kvm_get_apic(CPUX86State *env)
{
DeviceState *apic = env->apic_state;
struct kvm_lapic_state kapic;
return 0;
}
-static int kvm_put_apic(CPUState *env)
+static int kvm_put_apic(CPUX86State *env)
{
DeviceState *apic = env->apic_state;
struct kvm_lapic_state kapic;
return 0;
}
-static int kvm_put_vcpu_events(CPUState *env, int level)
+static int kvm_put_vcpu_events(CPUX86State *env, int level)
{
struct kvm_vcpu_events events;
events.exception.nr = env->exception_injected;
events.exception.has_error_code = env->has_error_code;
events.exception.error_code = env->error_code;
+ events.exception.pad = 0;
events.interrupt.injected = (env->interrupt_injected >= 0);
events.interrupt.nr = env->interrupt_injected;
events.nmi.injected = env->nmi_injected;
events.nmi.pending = env->nmi_pending;
events.nmi.masked = !!(env->hflags2 & HF2_NMI_MASK);
+ events.nmi.pad = 0;
events.sipi_vector = env->sipi_vector;
return kvm_vcpu_ioctl(env, KVM_SET_VCPU_EVENTS, &events);
}
-static int kvm_get_vcpu_events(CPUState *env)
+static int kvm_get_vcpu_events(CPUX86State *env)
{
struct kvm_vcpu_events events;
int ret;
return 0;
}
-static int kvm_guest_debug_workarounds(CPUState *env)
+static int kvm_guest_debug_workarounds(CPUX86State *env)
{
int ret = 0;
unsigned long reinject_trap = 0;
return ret;
}
-static int kvm_put_debugregs(CPUState *env)
+static int kvm_put_debugregs(CPUX86State *env)
{
struct kvm_debugregs dbgregs;
int i;
return kvm_vcpu_ioctl(env, KVM_SET_DEBUGREGS, &dbgregs);
}
-static int kvm_get_debugregs(CPUState *env)
+static int kvm_get_debugregs(CPUX86State *env)
{
struct kvm_debugregs dbgregs;
int i, ret;
return 0;
}
-int kvm_arch_put_registers(CPUState *env, int level)
+int kvm_arch_put_registers(CPUX86State *env, int level)
{
int ret;
return 0;
}
-int kvm_arch_get_registers(CPUState *env)
+int kvm_arch_get_registers(CPUX86State *env)
{
int ret;
return 0;
}
-void kvm_arch_pre_run(CPUState *env, struct kvm_run *run)
+void kvm_arch_pre_run(CPUX86State *env, struct kvm_run *run)
{
int ret;
}
}
-void kvm_arch_post_run(CPUState *env, struct kvm_run *run)
+void kvm_arch_post_run(CPUX86State *env, struct kvm_run *run)
{
if (run->if_flag) {
env->eflags |= IF_MASK;
cpu_set_apic_base(env->apic_state, run->apic_base);
}
-int kvm_arch_process_async_events(CPUState *env)
+int kvm_arch_process_async_events(CPUX86State *env)
{
+ X86CPU *cpu = x86_env_get_cpu(env);
+
if (env->interrupt_request & CPU_INTERRUPT_MCE) {
/* We must not raise CPU_INTERRUPT_MCE if it's not supported. */
assert(env->mcg_cap);
return 0;
}
+ if (env->interrupt_request & CPU_INTERRUPT_POLL) {
+ env->interrupt_request &= ~CPU_INTERRUPT_POLL;
+ apic_poll_irq(env->apic_state);
+ }
if (((env->interrupt_request & CPU_INTERRUPT_HARD) &&
(env->eflags & IF_MASK)) ||
(env->interrupt_request & CPU_INTERRUPT_NMI)) {
}
if (env->interrupt_request & CPU_INTERRUPT_INIT) {
kvm_cpu_synchronize_state(env);
- do_cpu_init(env);
+ do_cpu_init(cpu);
}
if (env->interrupt_request & CPU_INTERRUPT_SIPI) {
kvm_cpu_synchronize_state(env);
- do_cpu_sipi(env);
+ do_cpu_sipi(cpu);
}
if (env->interrupt_request & CPU_INTERRUPT_TPR) {
env->interrupt_request &= ~CPU_INTERRUPT_TPR;
return env->halted;
}
-static int kvm_handle_halt(CPUState *env)
+static int kvm_handle_halt(CPUX86State *env)
{
if (!((env->interrupt_request & CPU_INTERRUPT_HARD) &&
(env->eflags & IF_MASK)) &&
return 0;
}
-static int kvm_handle_tpr_access(CPUState *env)
+static int kvm_handle_tpr_access(CPUX86State *env)
{
struct kvm_run *run = env->kvm_run;
return 1;
}
-int kvm_arch_insert_sw_breakpoint(CPUState *env, struct kvm_sw_breakpoint *bp)
+int kvm_arch_insert_sw_breakpoint(CPUX86State *env, struct kvm_sw_breakpoint *bp)
{
static const uint8_t int3 = 0xcc;
return 0;
}
-int kvm_arch_remove_sw_breakpoint(CPUState *env, struct kvm_sw_breakpoint *bp)
+int kvm_arch_remove_sw_breakpoint(CPUX86State *env, struct kvm_sw_breakpoint *bp)
{
uint8_t int3;
return ret;
}
-void kvm_arch_update_guest_debug(CPUState *env, struct kvm_guest_debug *dbg)
+void kvm_arch_update_guest_debug(CPUX86State *env, struct kvm_guest_debug *dbg)
{
const uint8_t type_code[] = {
[GDB_BREAKPOINT_HW] = 0x0,
#define VMX_INVALID_GUEST_STATE 0x80000021
-int kvm_arch_handle_exit(CPUState *env, struct kvm_run *run)
+int kvm_arch_handle_exit(CPUX86State *env, struct kvm_run *run)
{
uint64_t code;
int ret;
return ret;
}
-bool kvm_arch_stop_on_emulation_error(CPUState *env)
+bool kvm_arch_stop_on_emulation_error(CPUX86State *env)
{
kvm_cpu_synchronize_state(env);
return !(env->cr[0] & CR0_PE_MASK) ||
*/
no_hpet = 1;
}
+ /* We know at this point that we're using the in-kernel
+ * irqchip, so we can use irqfds, and on x86 we know
+ * we can use msi via irqfd and GSI routing.
+ */
+ kvm_irqfds_allowed = true;
+ kvm_msi_via_irqfd_allowed = true;
+ kvm_gsi_routing_allowed = true;
+}
+
+/* Classic KVM device assignment interface. Will remain x86 only. */
+int kvm_device_pci_assign(KVMState *s, PCIHostDeviceAddress *dev_addr,
+ uint32_t flags, uint32_t *dev_id)
+{
+ struct kvm_assigned_pci_dev dev_data = {
+ .segnr = dev_addr->domain,
+ .busnr = dev_addr->bus,
+ .devfn = PCI_DEVFN(dev_addr->slot, dev_addr->function),
+ .flags = flags,
+ };
+ int ret;
+
+ dev_data.assigned_dev_id =
+ (dev_addr->domain << 16) | (dev_addr->bus << 8) | dev_data.devfn;
+
+ ret = kvm_vm_ioctl(s, KVM_ASSIGN_PCI_DEVICE, &dev_data);
+ if (ret < 0) {
+ return ret;
+ }
+
+ *dev_id = dev_data.assigned_dev_id;
+
+ return 0;
+}
+
+int kvm_device_pci_deassign(KVMState *s, uint32_t dev_id)
+{
+ struct kvm_assigned_pci_dev dev_data = {
+ .assigned_dev_id = dev_id,
+ };
+
+ return kvm_vm_ioctl(s, KVM_DEASSIGN_PCI_DEVICE, &dev_data);
+}
+
+static int kvm_assign_irq_internal(KVMState *s, uint32_t dev_id,
+ uint32_t irq_type, uint32_t guest_irq)
+{
+ struct kvm_assigned_irq assigned_irq = {
+ .assigned_dev_id = dev_id,
+ .guest_irq = guest_irq,
+ .flags = irq_type,
+ };
+
+ if (kvm_check_extension(s, KVM_CAP_ASSIGN_DEV_IRQ)) {
+ return kvm_vm_ioctl(s, KVM_ASSIGN_DEV_IRQ, &assigned_irq);
+ } else {
+ return kvm_vm_ioctl(s, KVM_ASSIGN_IRQ, &assigned_irq);
+ }
+}
+
+int kvm_device_intx_assign(KVMState *s, uint32_t dev_id, bool use_host_msi,
+ uint32_t guest_irq)
+{
+ uint32_t irq_type = KVM_DEV_IRQ_GUEST_INTX |
+ (use_host_msi ? KVM_DEV_IRQ_HOST_MSI : KVM_DEV_IRQ_HOST_INTX);
+
+ return kvm_assign_irq_internal(s, dev_id, irq_type, guest_irq);
+}
+
+int kvm_device_intx_set_mask(KVMState *s, uint32_t dev_id, bool masked)
+{
+ struct kvm_assigned_pci_dev dev_data = {
+ .assigned_dev_id = dev_id,
+ .flags = masked ? KVM_DEV_ASSIGN_MASK_INTX : 0,
+ };
+
+ return kvm_vm_ioctl(s, KVM_ASSIGN_SET_INTX_MASK, &dev_data);
+}
+
+static int kvm_deassign_irq_internal(KVMState *s, uint32_t dev_id,
+ uint32_t type)
+{
+ struct kvm_assigned_irq assigned_irq = {
+ .assigned_dev_id = dev_id,
+ .flags = type,
+ };
+
+ return kvm_vm_ioctl(s, KVM_DEASSIGN_DEV_IRQ, &assigned_irq);
+}
+
+int kvm_device_intx_deassign(KVMState *s, uint32_t dev_id, bool use_host_msi)
+{
+ return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_INTX |
+ (use_host_msi ? KVM_DEV_IRQ_HOST_MSI : KVM_DEV_IRQ_HOST_INTX));
+}
+
+int kvm_device_msi_assign(KVMState *s, uint32_t dev_id, int virq)
+{
+ return kvm_assign_irq_internal(s, dev_id, KVM_DEV_IRQ_HOST_MSI |
+ KVM_DEV_IRQ_GUEST_MSI, virq);
+}
+
+int kvm_device_msi_deassign(KVMState *s, uint32_t dev_id)
+{
+ return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_MSI |
+ KVM_DEV_IRQ_HOST_MSI);
+}
+
+bool kvm_device_msix_supported(KVMState *s)
+{
+ /* The kernel lacks a corresponding KVM_CAP, so we probe by calling
+ * KVM_ASSIGN_SET_MSIX_NR with an invalid parameter. */
+ return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_NR, NULL) == -EFAULT;
+}
+
+int kvm_device_msix_init_vectors(KVMState *s, uint32_t dev_id,
+ uint32_t nr_vectors)
+{
+ struct kvm_assigned_msix_nr msix_nr = {
+ .assigned_dev_id = dev_id,
+ .entry_nr = nr_vectors,
+ };
+
+ return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_NR, &msix_nr);
+}
+
+int kvm_device_msix_set_vector(KVMState *s, uint32_t dev_id, uint32_t vector,
+ int virq)
+{
+ struct kvm_assigned_msix_entry msix_entry = {
+ .assigned_dev_id = dev_id,
+ .gsi = virq,
+ .entry = vector,
+ };
+
+ return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_ENTRY, &msix_entry);
+}
+
+int kvm_device_msix_assign(KVMState *s, uint32_t dev_id)
+{
+ return kvm_assign_irq_internal(s, dev_id, KVM_DEV_IRQ_HOST_MSIX |
+ KVM_DEV_IRQ_GUEST_MSIX, 0);
+}
+
+int kvm_device_msix_deassign(KVMState *s, uint32_t dev_id)
+{
+ return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_MSIX |
+ KVM_DEV_IRQ_HOST_MSIX);
}
projects / qemu.git / blobdiff