x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

When using eBPF to probe CS CALL instructions within a function,
a crash can be triggered.

The eBPF tool probes offset 257 of the __hrtimer_run_queues()
function:

<__hrtimer_run_queues+249>:  nopl   0x0(%rax,%rax,1)
<__hrtimer_run_queues+254>:  mov    %r14,%rdi
<__hrtimer_run_queues+257>:  cs call <__x86_indirect_thunk_r12>
<__hrtimer_run_queues+263>:  mov    %eax,%r12d
<__hrtimer_run_queues+266>:  xchg   %ax,%ax
<__hrtimer_run_queues+268>:  mov    %r13,%rdi

Which triggers this crash:

  BUG: unable to handle page fault for address: 00000000000f41c9
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 0 P4D 0
  Oops: 0002 [#1] SMP NOPTI
  CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P
  RIP: 0010:__hrtimer_run_queues+0x106/0x230

Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is
at the 6th byte of the above CS CALL instruction. Since the CS CALL
instruction occupies 6 bytes, the exception occurred in the middle of that
call instruction.

The root cause is that when using eBPF tools to probe in the middle of a
function, a kprobe with INT3 is used as the underlying implementation.

During single-step emulation of the original CALL instruction,
int3_emulate_call() assumes that the probed CALL instruction is 5 bytes
long. However, the actual CS-prefixed CALL instruction occupies 6 bytes,
so it constructs an incorrect exception return address. When the CPU
returns from the kprobe handler, the next instruction to be executed is at
the address of the last byte of that CS CALL instruction. Coincidentally,
starting from that address, the CPU fetches and decodes a completely
different instruction, which ultimately triggers a kernel crash.

Fix the issue by using the actual instruction length obtained from
the instruction decoder when constructing the exception return
address, rather than relying on the hardcoded CALL_INSN_SIZE macro.

[ mingo: Refined the changelog ]

Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step")
Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 1046 insertions, 0 deletions

diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c
new file mode 100644
index 000000000..e062ca19f
--- /dev/null
+++ b/arch/riscv/kvm/vcpu.c
@@ -0,0 +1,1046 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2019 Western Digital Corporation or its affiliates.
+ *
+ * Authors:
+ *     Anup Patel <anup.patel@wdc.com>
+ */
+
+#include <linux/bitops.h>
+#include <linux/errno.h>
+#include <linux/err.h>
+#include <linux/kdebug.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <linux/vmalloc.h>
+#include <linux/sched/signal.h>
+#include <linux/fs.h>
+#include <linux/kvm_host.h>
+#include <asm/cacheflush.h>
+#include <asm/kvm_mmu.h>
+#include <asm/kvm_nacl.h>
+#include <asm/kvm_vcpu_vector.h>
+
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+
+static DEFINE_PER_CPU(struct kvm_vcpu *, kvm_former_vcpu);
+
+void kvm_riscv_clear_former_vcpu(void)
+{
+	/*
+	 * Clear the per-CPU former VCPU pointer because hypervisor CSR state
+	 * will be lost. This ensures that the next VCPU entry will properly
+	 * restore all CSRs instead of incorrectly skipping CSR restoration
+	 * via the fast-path optimization.
+	 */
+	__this_cpu_write(kvm_former_vcpu, NULL);
+}
+
+const struct kvm_stats_desc kvm_vcpu_stats_desc[] = {
+	KVM_GENERIC_VCPU_STATS(),
+	STATS_DESC_COUNTER(VCPU, ecall_exit_stat),
+	STATS_DESC_COUNTER(VCPU, wfi_exit_stat),
+	STATS_DESC_COUNTER(VCPU, wrs_exit_stat),
+	STATS_DESC_COUNTER(VCPU, mmio_exit_user),
+	STATS_DESC_COUNTER(VCPU, mmio_exit_kernel),
+	STATS_DESC_COUNTER(VCPU, csr_exit_user),
+	STATS_DESC_COUNTER(VCPU, csr_exit_kernel),
+	STATS_DESC_COUNTER(VCPU, signal_exits),
+	STATS_DESC_COUNTER(VCPU, exits),
+	STATS_DESC_COUNTER(VCPU, instr_illegal_exits),
+	STATS_DESC_COUNTER(VCPU, load_misaligned_exits),
+	STATS_DESC_COUNTER(VCPU, store_misaligned_exits),
+	STATS_DESC_COUNTER(VCPU, load_access_exits),
+	STATS_DESC_COUNTER(VCPU, store_access_exits),
+};
+
+const struct kvm_stats_header kvm_vcpu_stats_header = {
+	.name_size = KVM_STATS_NAME_SIZE,
+	.num_desc = ARRAY_SIZE(kvm_vcpu_stats_desc),
+	.id_offset = sizeof(struct kvm_stats_header),
+	.desc_offset = sizeof(struct kvm_stats_header) + KVM_STATS_NAME_SIZE,
+	.data_offset = sizeof(struct kvm_stats_header) + KVM_STATS_NAME_SIZE +
+		       sizeof(kvm_vcpu_stats_desc),
+};
+
+static void kvm_riscv_vcpu_context_reset(struct kvm_vcpu *vcpu,
+					 bool kvm_sbi_reset)
+{
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+	struct kvm_cpu_context *cntx = &vcpu->arch.guest_context;
+	void *vector_datap = cntx->vector.datap;
+
+	memset(cntx, 0, sizeof(*cntx));
+	memset(csr, 0, sizeof(*csr));
+	memset(&vcpu->arch.smstateen_csr, 0, sizeof(vcpu->arch.smstateen_csr));
+	memset(&vcpu->arch.zicfiss_csr, 0, sizeof(vcpu->arch.zicfiss_csr));
+
+	/* Restore datap as it's not a part of the guest context. */
+	cntx->vector.datap = vector_datap;
+
+	if (kvm_sbi_reset)
+		kvm_riscv_vcpu_sbi_load_reset_state(vcpu);
+
+	/* Setup reset state of shadow SSTATUS and HSTATUS CSRs */
+	cntx->sstatus = SR_SPP | SR_SPIE;
+
+	cntx->hstatus |= HSTATUS_VTW;
+	cntx->hstatus |= HSTATUS_SPVP;
+	cntx->hstatus |= HSTATUS_SPV;
+}
+
+static void kvm_riscv_reset_vcpu(struct kvm_vcpu *vcpu, bool kvm_sbi_reset)
+{
+	unsigned long flags;
+	bool loaded;
+
+	/**
+	 * The preemption should be disabled here because it races with
+	 * kvm_sched_out/kvm_sched_in(called from preempt notifiers) which
+	 * also calls vcpu_load/put.
+	 */
+	get_cpu();
+	loaded = (vcpu->cpu != -1);
+	if (loaded)
+		kvm_arch_vcpu_put(vcpu);
+
+	vcpu->arch.last_exit_cpu = -1;
+
+	kvm_riscv_vcpu_context_reset(vcpu, kvm_sbi_reset);
+
+	kvm_riscv_vcpu_fp_reset(vcpu);
+
+	kvm_riscv_vcpu_vector_reset(vcpu);
+
+	kvm_riscv_vcpu_timer_reset(vcpu);
+
+	kvm_riscv_vcpu_aia_reset(vcpu);
+
+	raw_spin_lock_irqsave(&vcpu->arch.irqs_pending_lock, flags);
+	bitmap_zero(vcpu->arch.irqs_pending, KVM_RISCV_VCPU_NR_IRQS);
+	bitmap_zero(vcpu->arch.irqs_pending_mask, KVM_RISCV_VCPU_NR_IRQS);
+	raw_spin_unlock_irqrestore(&vcpu->arch.irqs_pending_lock, flags);
+
+	kvm_riscv_vcpu_pmu_reset(vcpu);
+
+	vcpu->arch.hfence_head = 0;
+	vcpu->arch.hfence_tail = 0;
+	memset(vcpu->arch.hfence_queue, 0, sizeof(vcpu->arch.hfence_queue));
+
+	kvm_riscv_vcpu_sbi_reset(vcpu);
+
+	/* Reset the guest CSRs for hotplug usecase */
+	if (loaded)
+		kvm_arch_vcpu_load(vcpu, smp_processor_id());
+	put_cpu();
+}
+
+int kvm_arch_vcpu_precreate(struct kvm *kvm, unsigned int id)
+{
+	return 0;
+}
+
+int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)
+{
+	int rc;
+
+	spin_lock_init(&vcpu->arch.mp_state_lock);
+
+	/* Mark this VCPU never ran */
+	vcpu->arch.ran_atleast_once = false;
+
+	vcpu->arch.mmu_page_cache.gfp_zero = __GFP_ZERO;
+	bitmap_zero(vcpu->arch.isa, RISCV_ISA_EXT_MAX);
+
+	/* Setup VCPU config */
+	kvm_riscv_vcpu_config_init(vcpu);
+
+	/* Setup ISA features available to VCPU */
+	kvm_riscv_vcpu_setup_isa(vcpu);
+
+	/* Setup vendor, arch, and implementation details */
+	vcpu->arch.mvendorid = sbi_get_mvendorid();
+	vcpu->arch.marchid = sbi_get_marchid();
+	vcpu->arch.mimpid = sbi_get_mimpid();
+
+	/* Setup VCPU hfence queue */
+	spin_lock_init(&vcpu->arch.hfence_lock);
+	raw_spin_lock_init(&vcpu->arch.irqs_pending_lock);
+
+	spin_lock_init(&vcpu->arch.reset_state.lock);
+
+	rc = kvm_riscv_vcpu_alloc_vector_context(vcpu);
+	if (rc)
+		return rc;
+
+	/* Setup VCPU timer */
+	kvm_riscv_vcpu_timer_init(vcpu);
+
+	/* setup performance monitoring */
+	kvm_riscv_vcpu_pmu_init(vcpu);
+
+	/* Setup VCPU AIA */
+	kvm_riscv_vcpu_aia_init(vcpu);
+
+	/*
+	 * Setup SBI extensions
+	 * NOTE: This must be the last thing to be initialized.
+	 */
+	kvm_riscv_vcpu_sbi_init(vcpu);
+
+	/* Reset VCPU */
+	kvm_riscv_reset_vcpu(vcpu, false);
+
+	return 0;
+}
+
+void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
+{
+	/**
+	 * vcpu with id 0 is the designated boot cpu.
+	 * Keep all vcpus with non-zero id in power-off state so that
+	 * they can be brought up using SBI HSM extension.
+	 */
+	if (vcpu->vcpu_idx != 0)
+		kvm_riscv_vcpu_power_off(vcpu);
+}
+
+void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
+{
+	kvm_riscv_vcpu_sbi_deinit(vcpu);
+
+	/* Cleanup VCPU AIA context */
+	kvm_riscv_vcpu_aia_deinit(vcpu);
+
+	/* Cleanup VCPU timer */
+	kvm_riscv_vcpu_timer_deinit(vcpu);
+
+	kvm_riscv_vcpu_pmu_deinit(vcpu);
+
+	/* Free unused pages pre-allocated for G-stage page table mappings */
+	kvm_mmu_free_memory_cache(&vcpu->arch.mmu_page_cache);
+
+	/* Free vector context space for host and guest kernel */
+	kvm_riscv_vcpu_free_vector_context(vcpu);
+}
+
+int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu)
+{
+	return kvm_riscv_vcpu_timer_pending(vcpu);
+}
+
+int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
+{
+	return (kvm_riscv_vcpu_has_interrupts(vcpu, -1ULL) &&
+		!kvm_riscv_vcpu_stopped(vcpu) && !vcpu->arch.pause);
+}
+
+int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
+{
+	return kvm_vcpu_exiting_guest_mode(vcpu) == IN_GUEST_MODE;
+}
+
+bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu)
+{
+	return (vcpu->arch.guest_context.sstatus & SR_SPP) ? true : false;
+}
+
+#ifdef CONFIG_GUEST_PERF_EVENTS
+unsigned long kvm_arch_vcpu_get_ip(struct kvm_vcpu *vcpu)
+{
+	return vcpu->arch.guest_context.sepc;
+}
+#endif
+
+vm_fault_t kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf)
+{
+	return VM_FAULT_SIGBUS;
+}
+
+long kvm_arch_vcpu_unlocked_ioctl(struct file *filp, unsigned int ioctl,
+				  unsigned long arg)
+{
+	struct kvm_vcpu *vcpu = filp->private_data;
+	void __user *argp = (void __user *)arg;
+
+	if (ioctl == KVM_INTERRUPT) {
+		struct kvm_interrupt irq;
+
+		if (copy_from_user(&irq, argp, sizeof(irq)))
+			return -EFAULT;
+
+		if (irq.irq == KVM_INTERRUPT_SET)
+			return kvm_riscv_vcpu_set_interrupt(vcpu, IRQ_VS_EXT);
+		else
+			return kvm_riscv_vcpu_unset_interrupt(vcpu, IRQ_VS_EXT);
+	}
+
+	return -ENOIOCTLCMD;
+}
+
+long kvm_arch_vcpu_ioctl(struct file *filp,
+			 unsigned int ioctl, unsigned long arg)
+{
+	struct kvm_vcpu *vcpu = filp->private_data;
+	void __user *argp = (void __user *)arg;
+	long r = -EINVAL;
+
+	switch (ioctl) {
+	case KVM_SET_ONE_REG:
+	case KVM_GET_ONE_REG: {
+		struct kvm_one_reg reg;
+
+		r = -EFAULT;
+		if (copy_from_user(&reg, argp, sizeof(reg)))
+			break;
+
+		if (ioctl == KVM_SET_ONE_REG)
+			r = kvm_riscv_vcpu_set_reg(vcpu, &reg);
+		else
+			r = kvm_riscv_vcpu_get_reg(vcpu, &reg);
+		break;
+	}
+	case KVM_GET_REG_LIST: {
+		struct kvm_reg_list __user *user_list = argp;
+		struct kvm_reg_list reg_list;
+		unsigned int n;
+
+		r = -EFAULT;
+		if (copy_from_user(&reg_list, user_list, sizeof(reg_list)))
+			break;
+		n = reg_list.n;
+		reg_list.n = kvm_riscv_vcpu_num_regs(vcpu);
+		if (copy_to_user(user_list, &reg_list, sizeof(reg_list)))
+			break;
+		r = -E2BIG;
+		if (n < reg_list.n)
+			break;
+		r = kvm_riscv_vcpu_copy_reg_indices(vcpu, user_list->reg);
+		break;
+	}
+	default:
+		break;
+	}
+
+	return r;
+}
+
+int kvm_arch_vcpu_ioctl_get_sregs(struct kvm_vcpu *vcpu,
+				  struct kvm_sregs *sregs)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
+				  struct kvm_sregs *sregs)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_translate(struct kvm_vcpu *vcpu,
+				  struct kvm_translation *tr)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
+{
+	return -EINVAL;
+}
+
+int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
+{
+	return -EINVAL;
+}
+
+void kvm_riscv_vcpu_flush_interrupts(struct kvm_vcpu *vcpu)
+{
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+	unsigned long mask, val;
+	unsigned long flags;
+
+	raw_spin_lock_irqsave(&vcpu->arch.irqs_pending_lock, flags);
+
+	mask = vcpu->arch.irqs_pending_mask[0];
+	if (mask) {
+		vcpu->arch.irqs_pending_mask[0] = 0;
+		val = vcpu->arch.irqs_pending[0] & mask;
+
+		csr->hvip &= ~mask;
+		csr->hvip |= val;
+	}
+
+	/* Flush AIA high interrupts */
+	kvm_riscv_vcpu_aia_flush_interrupts(vcpu);
+
+	raw_spin_unlock_irqrestore(&vcpu->arch.irqs_pending_lock, flags);
+}
+
+void kvm_riscv_vcpu_sync_interrupts(struct kvm_vcpu *vcpu)
+{
+	unsigned long hvip;
+	unsigned long flags;
+	struct kvm_vcpu_arch *v = &vcpu->arch;
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	/* Read current HVIP and VSIE CSRs */
+	csr->vsie = ncsr_read(CSR_VSIE);
+
+	/* Sync-up HVIP.VSSIP bit changes does by Guest */
+	hvip = ncsr_read(CSR_HVIP);
+
+	raw_spin_lock_irqsave(&v->irqs_pending_lock, flags);
+
+	if ((csr->hvip ^ hvip) & (1UL << IRQ_VS_SOFT)) {
+		if (hvip & (1UL << IRQ_VS_SOFT)) {
+			if (!__test_and_set_bit(IRQ_VS_SOFT,
+						v->irqs_pending_mask))
+				__set_bit(IRQ_VS_SOFT, v->irqs_pending);
+		} else {
+			if (!__test_and_set_bit(IRQ_VS_SOFT,
+						v->irqs_pending_mask))
+				__clear_bit(IRQ_VS_SOFT, v->irqs_pending);
+		}
+	}
+
+	/* Sync up the HVIP.LCOFIP bit changes (only clear) by the guest */
+	if ((csr->hvip ^ hvip) & (1UL << IRQ_PMU_OVF)) {
+		if (!(hvip & (1UL << IRQ_PMU_OVF)) &&
+		    !__test_and_set_bit(IRQ_PMU_OVF, v->irqs_pending_mask))
+			__clear_bit(IRQ_PMU_OVF, v->irqs_pending);
+	}
+
+	/* Sync-up AIA high interrupts */
+	kvm_riscv_vcpu_aia_sync_interrupts(vcpu);
+
+	raw_spin_unlock_irqrestore(&v->irqs_pending_lock, flags);
+
+	/* Sync-up timer CSRs */
+	kvm_riscv_vcpu_timer_sync(vcpu);
+}
+
+int kvm_riscv_vcpu_set_interrupt(struct kvm_vcpu *vcpu, unsigned int irq)
+{
+	unsigned long flags;
+
+	/*
+	 * We only allow VS-mode software, timer, and external
+	 * interrupts when irq is one of the local interrupts
+	 * defined by RISC-V privilege specification.
+	 */
+	if (irq < IRQ_LOCAL_MAX &&
+	    irq != IRQ_VS_SOFT &&
+	    irq != IRQ_VS_TIMER &&
+	    irq != IRQ_VS_EXT &&
+	    irq != IRQ_PMU_OVF)
+		return -EINVAL;
+
+	raw_spin_lock_irqsave(&vcpu->arch.irqs_pending_lock, flags);
+	__set_bit(irq, vcpu->arch.irqs_pending);
+	__set_bit(irq, vcpu->arch.irqs_pending_mask);
+	raw_spin_unlock_irqrestore(&vcpu->arch.irqs_pending_lock, flags);
+
+	trace_kvm_vcpu_irq(vcpu->vcpu_id, irq, 1);
+
+	kvm_vcpu_kick(vcpu);
+
+	return 0;
+}
+
+int kvm_riscv_vcpu_unset_interrupt(struct kvm_vcpu *vcpu, unsigned int irq)
+{
+	unsigned long flags;
+
+	/*
+	 * We only allow VS-mode software, timer, counter overflow and external
+	 * interrupts when irq is one of the local interrupts
+	 * defined by RISC-V privilege specification.
+	 */
+	if (irq < IRQ_LOCAL_MAX &&
+	    irq != IRQ_VS_SOFT &&
+	    irq != IRQ_VS_TIMER &&
+	    irq != IRQ_VS_EXT &&
+	    irq != IRQ_PMU_OVF)
+		return -EINVAL;
+
+	raw_spin_lock_irqsave(&vcpu->arch.irqs_pending_lock, flags);
+	__clear_bit(irq, vcpu->arch.irqs_pending);
+	__set_bit(irq, vcpu->arch.irqs_pending_mask);
+	raw_spin_unlock_irqrestore(&vcpu->arch.irqs_pending_lock, flags);
+
+	trace_kvm_vcpu_irq(vcpu->vcpu_id, irq, 0);
+
+	return 0;
+}
+
+bool kvm_riscv_vcpu_has_interrupts(struct kvm_vcpu *vcpu, u64 mask)
+{
+	unsigned long flags;
+	unsigned long ie;
+	bool ret;
+
+	raw_spin_lock_irqsave(&vcpu->arch.irqs_pending_lock, flags);
+	ie = ((vcpu->arch.guest_csr.vsie & VSIP_VALID_MASK)
+		<< VSIP_TO_HVIP_SHIFT) & (unsigned long)mask;
+	ie |= vcpu->arch.guest_csr.vsie & ~IRQ_LOCAL_MASK &
+		(unsigned long)mask;
+	ret = vcpu->arch.irqs_pending[0] & ie;
+	raw_spin_unlock_irqrestore(&vcpu->arch.irqs_pending_lock, flags);
+
+	/* Check AIA high interrupts */
+	if (!ret)
+		ret = kvm_riscv_vcpu_aia_has_interrupts(vcpu, mask);
+
+	return ret;
+}
+
+void __kvm_riscv_vcpu_power_off(struct kvm_vcpu *vcpu)
+{
+	WRITE_ONCE(vcpu->arch.mp_state.mp_state, KVM_MP_STATE_STOPPED);
+	kvm_make_request(KVM_REQ_SLEEP, vcpu);
+	kvm_vcpu_kick(vcpu);
+}
+
+void kvm_riscv_vcpu_power_off(struct kvm_vcpu *vcpu)
+{
+	spin_lock(&vcpu->arch.mp_state_lock);
+	__kvm_riscv_vcpu_power_off(vcpu);
+	spin_unlock(&vcpu->arch.mp_state_lock);
+}
+
+void __kvm_riscv_vcpu_power_on(struct kvm_vcpu *vcpu)
+{
+	WRITE_ONCE(vcpu->arch.mp_state.mp_state, KVM_MP_STATE_RUNNABLE);
+	kvm_vcpu_wake_up(vcpu);
+}
+
+void kvm_riscv_vcpu_power_on(struct kvm_vcpu *vcpu)
+{
+	spin_lock(&vcpu->arch.mp_state_lock);
+	__kvm_riscv_vcpu_power_on(vcpu);
+	spin_unlock(&vcpu->arch.mp_state_lock);
+}
+
+bool kvm_riscv_vcpu_stopped(struct kvm_vcpu *vcpu)
+{
+	return READ_ONCE(vcpu->arch.mp_state.mp_state) == KVM_MP_STATE_STOPPED;
+}
+
+int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
+				    struct kvm_mp_state *mp_state)
+{
+	*mp_state = READ_ONCE(vcpu->arch.mp_state);
+
+	return 0;
+}
+
+int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
+				    struct kvm_mp_state *mp_state)
+{
+	int ret = 0;
+
+	spin_lock(&vcpu->arch.mp_state_lock);
+
+	switch (mp_state->mp_state) {
+	case KVM_MP_STATE_RUNNABLE:
+		WRITE_ONCE(vcpu->arch.mp_state, *mp_state);
+		break;
+	case KVM_MP_STATE_STOPPED:
+		__kvm_riscv_vcpu_power_off(vcpu);
+		break;
+	case KVM_MP_STATE_INIT_RECEIVED:
+		if (vcpu->kvm->arch.mp_state_reset)
+			kvm_riscv_reset_vcpu(vcpu, false);
+		else
+			ret = -EINVAL;
+		break;
+	default:
+		ret = -EINVAL;
+	}
+
+	spin_unlock(&vcpu->arch.mp_state_lock);
+
+	return ret;
+}
+
+int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
+					struct kvm_guest_debug *dbg)
+{
+	if (dbg->control & KVM_GUESTDBG_ENABLE)
+		vcpu->guest_debug = dbg->control;
+	else
+		vcpu->guest_debug = 0;
+
+	kvm_riscv_vcpu_config_guest_debug(vcpu);
+	return 0;
+}
+
+void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
+{
+	void *nsh;
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	/*
+	 * If VCPU is being reloaded on the same physical CPU and no
+	 * other KVM VCPU has run on this CPU since it was last put,
+	 * we can skip the expensive CSR and HGATP writes.
+	 *
+	 * Note: If a new CSR is added to this fast-path skip block,
+	 * make sure that 'csr_dirty' is set to true in any
+	 * ioctl (e.g., KVM_SET_ONE_REG) that modifies it.
+	 */
+	if (vcpu != __this_cpu_read(kvm_former_vcpu))
+		__this_cpu_write(kvm_former_vcpu, vcpu);
+	else if (vcpu->arch.last_exit_cpu == cpu && !vcpu->arch.csr_dirty)
+		goto csr_restore_done;
+
+	vcpu->arch.csr_dirty = false;
+
+	/*
+	 * Load VCPU config CSRs before other CSRs because
+	 * the read/write behaviour of certain CSRs change
+	 * based on VCPU config CSRs.
+	 */
+	kvm_riscv_vcpu_config_load(vcpu);
+
+	if (kvm_riscv_nacl_sync_csr_available()) {
+		nsh = nacl_shmem();
+		nacl_csr_write(nsh, CSR_VSSTATUS, csr->vsstatus);
+		nacl_csr_write(nsh, CSR_VSIE, csr->vsie);
+		nacl_csr_write(nsh, CSR_VSTVEC, csr->vstvec);
+		nacl_csr_write(nsh, CSR_VSSCRATCH, csr->vsscratch);
+		nacl_csr_write(nsh, CSR_VSEPC, csr->vsepc);
+		nacl_csr_write(nsh, CSR_VSCAUSE, csr->vscause);
+		nacl_csr_write(nsh, CSR_VSTVAL, csr->vstval);
+		nacl_csr_write(nsh, CSR_HVIP, csr->hvip);
+		nacl_csr_write(nsh, CSR_VSATP, csr->vsatp);
+	} else {
+		csr_write(CSR_VSSTATUS, csr->vsstatus);
+		csr_write(CSR_VSIE, csr->vsie);
+		csr_write(CSR_VSTVEC, csr->vstvec);
+		csr_write(CSR_VSSCRATCH, csr->vsscratch);
+		csr_write(CSR_VSEPC, csr->vsepc);
+		csr_write(CSR_VSCAUSE, csr->vscause);
+		csr_write(CSR_VSTVAL, csr->vstval);
+		csr_write(CSR_HVIP, csr->hvip);
+		csr_write(CSR_VSATP, csr->vsatp);
+	}
+
+	kvm_riscv_mmu_update_hgatp(vcpu);
+
+	kvm_riscv_vcpu_aia_load(vcpu, cpu);
+
+csr_restore_done:
+	kvm_riscv_vcpu_timer_restore(vcpu);
+
+	kvm_riscv_vcpu_host_fp_save(&vcpu->arch.host_context);
+	kvm_riscv_vcpu_guest_fp_restore(&vcpu->arch.guest_context,
+					vcpu->arch.isa);
+	get_cpu_vector_context();
+	kvm_riscv_vcpu_host_vector_save(&vcpu->arch.host_context);
+	kvm_riscv_vcpu_guest_vector_restore(&vcpu->arch.guest_context,
+					    vcpu->arch.isa);
+	put_cpu_vector_context();
+
+	kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
+
+	vcpu->cpu = cpu;
+}
+
+void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
+{
+	void *nsh;
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	vcpu->cpu = -1;
+
+	kvm_riscv_vcpu_aia_put(vcpu);
+
+	kvm_riscv_vcpu_guest_fp_save(&vcpu->arch.guest_context,
+				     vcpu->arch.isa);
+	kvm_riscv_vcpu_host_fp_restore(&vcpu->arch.host_context);
+
+	kvm_riscv_vcpu_timer_save(vcpu);
+	get_cpu_vector_context();
+	kvm_riscv_vcpu_guest_vector_save(&vcpu->arch.guest_context,
+					 vcpu->arch.isa);
+	kvm_riscv_vcpu_host_vector_restore(&vcpu->arch.host_context);
+	put_cpu_vector_context();
+
+	if (kvm_riscv_nacl_available()) {
+		nsh = nacl_shmem();
+		csr->vsstatus = nacl_csr_read(nsh, CSR_VSSTATUS);
+		csr->vsie = nacl_csr_read(nsh, CSR_VSIE);
+		csr->vstvec = nacl_csr_read(nsh, CSR_VSTVEC);
+		csr->vsscratch = nacl_csr_read(nsh, CSR_VSSCRATCH);
+		csr->vsepc = nacl_csr_read(nsh, CSR_VSEPC);
+		csr->vscause = nacl_csr_read(nsh, CSR_VSCAUSE);
+		csr->vstval = nacl_csr_read(nsh, CSR_VSTVAL);
+		csr->hvip = nacl_csr_read(nsh, CSR_HVIP);
+		csr->vsatp = nacl_csr_read(nsh, CSR_VSATP);
+	} else {
+		csr->vsstatus = csr_read(CSR_VSSTATUS);
+		csr->vsie = csr_read(CSR_VSIE);
+		csr->vstvec = csr_read(CSR_VSTVEC);
+		csr->vsscratch = csr_read(CSR_VSSCRATCH);
+		csr->vsepc = csr_read(CSR_VSEPC);
+		csr->vscause = csr_read(CSR_VSCAUSE);
+		csr->vstval = csr_read(CSR_VSTVAL);
+		csr->hvip = csr_read(CSR_HVIP);
+		csr->vsatp = csr_read(CSR_VSATP);
+	}
+}
+
+/**
+ * kvm_riscv_check_vcpu_requests - check and handle pending vCPU requests
+ * @vcpu:	the VCPU pointer
+ *
+ * Return: 1 if we should enter the guest
+ *	    0 if we should exit to userspace
+ */
+static int kvm_riscv_check_vcpu_requests(struct kvm_vcpu *vcpu)
+{
+	struct rcuwait *wait = kvm_arch_vcpu_get_wait(vcpu);
+
+	if (kvm_request_pending(vcpu)) {
+		if (kvm_check_request(KVM_REQ_SLEEP, vcpu)) {
+			kvm_vcpu_srcu_read_unlock(vcpu);
+			rcuwait_wait_event(wait,
+				(!kvm_riscv_vcpu_stopped(vcpu)) && (!vcpu->arch.pause),
+				TASK_INTERRUPTIBLE);
+			kvm_vcpu_srcu_read_lock(vcpu);
+
+			if (kvm_riscv_vcpu_stopped(vcpu) || vcpu->arch.pause) {
+				/*
+				 * Awaken to handle a signal, request to
+				 * sleep again later.
+				 */
+				kvm_make_request(KVM_REQ_SLEEP, vcpu);
+			}
+		}
+
+		if (kvm_check_request(KVM_REQ_VCPU_RESET, vcpu))
+			kvm_riscv_reset_vcpu(vcpu, true);
+
+		if (kvm_check_request(KVM_REQ_UPDATE_HGATP, vcpu))
+			kvm_riscv_mmu_update_hgatp(vcpu);
+
+		if (kvm_check_request(KVM_REQ_FENCE_I, vcpu))
+			kvm_riscv_fence_i_process(vcpu);
+
+		if (kvm_check_request(KVM_REQ_TLB_FLUSH, vcpu))
+			kvm_riscv_tlb_flush_process(vcpu);
+
+		if (kvm_check_request(KVM_REQ_HFENCE_VVMA_ALL, vcpu))
+			kvm_riscv_hfence_vvma_all_process(vcpu);
+
+		if (kvm_check_request(KVM_REQ_HFENCE, vcpu))
+			kvm_riscv_hfence_process(vcpu);
+
+		if (kvm_check_request(KVM_REQ_STEAL_UPDATE, vcpu))
+			kvm_riscv_vcpu_record_steal_time(vcpu);
+
+		if (kvm_dirty_ring_check_request(vcpu))
+			return 0;
+	}
+
+	return 1;
+}
+
+static void kvm_riscv_update_hvip(struct kvm_vcpu *vcpu)
+{
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	ncsr_write(CSR_HVIP, csr->hvip);
+	kvm_riscv_vcpu_aia_update_hvip(vcpu);
+}
+
+static __always_inline void kvm_riscv_vcpu_swap_in_guest_state(struct kvm_vcpu *vcpu)
+{
+	struct kvm_vcpu_zicfiss_csr *zicficsr = &vcpu->arch.zicfiss_csr;
+	struct kvm_vcpu_smstateen_csr *smcsr = &vcpu->arch.smstateen_csr;
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	vcpu->arch.host_scounteren = csr_swap(CSR_SCOUNTEREN, csr->scounteren);
+	vcpu->arch.host_senvcfg = csr_swap(CSR_SENVCFG, csr->senvcfg);
+	if (riscv_has_extension_unlikely(RISCV_ISA_EXT_SMSTATEEN))
+		vcpu->arch.host_sstateen0 = csr_swap(CSR_SSTATEEN0, smcsr->sstateen0);
+	if (riscv_has_extension_unlikely(RISCV_ISA_EXT_ZICFISS))
+		csr_write(CSR_SSP, zicficsr->ssp);
+}
+
+static __always_inline void kvm_riscv_vcpu_swap_in_host_state(struct kvm_vcpu *vcpu)
+{
+	struct kvm_vcpu_zicfiss_csr *zicficsr = &vcpu->arch.zicfiss_csr;
+	struct kvm_vcpu_smstateen_csr *smcsr = &vcpu->arch.smstateen_csr;
+	struct kvm_vcpu_csr *csr = &vcpu->arch.guest_csr;
+
+	csr->scounteren = csr_swap(CSR_SCOUNTEREN, vcpu->arch.host_scounteren);
+	csr->senvcfg = csr_swap(CSR_SENVCFG, vcpu->arch.host_senvcfg);
+	if (riscv_has_extension_unlikely(RISCV_ISA_EXT_SMSTATEEN))
+		smcsr->sstateen0 = csr_swap(CSR_SSTATEEN0, vcpu->arch.host_sstateen0);
+	if (riscv_has_extension_unlikely(RISCV_ISA_EXT_ZICFISS))
+		zicficsr->ssp = csr_swap(CSR_SSP, 0);
+}
+
+/*
+ * Actually run the vCPU, entering an RCU extended quiescent state (EQS) while
+ * the vCPU is running.
+ *
+ * This must be noinstr as instrumentation may make use of RCU, and this is not
+ * safe during the EQS.
+ */
+static void noinstr kvm_riscv_vcpu_enter_exit(struct kvm_vcpu *vcpu,
+					      struct kvm_cpu_trap *trap)
+{
+	void *nsh;
+	struct kvm_cpu_context *gcntx = &vcpu->arch.guest_context;
+	struct kvm_cpu_context *hcntx = &vcpu->arch.host_context;
+
+	/*
+	 * We save trap CSRs (such as SEPC, SCAUSE, STVAL, HTVAL, and
+	 * HTINST) here because we do local_irq_enable() after this
+	 * function in kvm_arch_vcpu_ioctl_run() which can result in
+	 * an interrupt immediately after local_irq_enable() and can
+	 * potentially change trap CSRs.
+	 */
+
+	kvm_riscv_vcpu_swap_in_guest_state(vcpu);
+	guest_state_enter_irqoff();
+
+	/* sstatus.VS != SR_VS_OFF is guaranteed when NEED_RESTORE is set */
+	if (current->thread.riscv_v_flags & RISCV_V_VCPU_NEED_RESTORE) {
+		current->thread.riscv_v_flags &= ~RISCV_V_VCPU_NEED_RESTORE;
+		current->thread.riscv_v_flags |= RISCV_V_VCPU_CTX;
+		__kvm_riscv_vector_restore(gcntx);
+		gcntx->sstatus = (gcntx->sstatus & ~SR_VS) | SR_VS_CLEAN;
+	}
+
+	if (kvm_riscv_nacl_sync_sret_available()) {
+		nsh = nacl_shmem();
+
+		if (kvm_riscv_nacl_autoswap_csr_available()) {
+			hcntx->hstatus =
+				nacl_csr_read(nsh, CSR_HSTATUS);
+			nacl_scratch_write_long(nsh,
+						SBI_NACL_SHMEM_AUTOSWAP_OFFSET +
+						SBI_NACL_SHMEM_AUTOSWAP_HSTATUS,
+						gcntx->hstatus);
+			nacl_scratch_write_long(nsh,
+						SBI_NACL_SHMEM_AUTOSWAP_OFFSET,
+						SBI_NACL_SHMEM_AUTOSWAP_FLAG_HSTATUS);
+		} else if (kvm_riscv_nacl_sync_csr_available()) {
+			hcntx->hstatus = nacl_csr_swap(nsh,
+						       CSR_HSTATUS, gcntx->hstatus);
+		} else {
+			hcntx->hstatus = csr_swap(CSR_HSTATUS, gcntx->hstatus);
+		}
+
+		nacl_scratch_write_longs(nsh,
+					 SBI_NACL_SHMEM_SRET_OFFSET +
+					 SBI_NACL_SHMEM_SRET_X(1),
+					 &gcntx->ra,
+					 SBI_NACL_SHMEM_SRET_X_LAST);
+
+		__kvm_riscv_nacl_switch_to(&vcpu->arch, SBI_EXT_NACL,
+					   SBI_EXT_NACL_SYNC_SRET);
+
+		if (kvm_riscv_nacl_autoswap_csr_available()) {
+			nacl_scratch_write_long(nsh,
+						SBI_NACL_SHMEM_AUTOSWAP_OFFSET,
+						0);
+			gcntx->hstatus = nacl_scratch_read_long(nsh,
+								SBI_NACL_SHMEM_AUTOSWAP_OFFSET +
+								SBI_NACL_SHMEM_AUTOSWAP_HSTATUS);
+		} else {
+			gcntx->hstatus = csr_swap(CSR_HSTATUS, hcntx->hstatus);
+		}
+
+		trap->htval = nacl_csr_read(nsh, CSR_HTVAL);
+		trap->htinst = nacl_csr_read(nsh, CSR_HTINST);
+	} else {
+		hcntx->hstatus = csr_swap(CSR_HSTATUS, gcntx->hstatus);
+
+		__kvm_riscv_switch_to(&vcpu->arch);
+
+		gcntx->hstatus = csr_swap(CSR_HSTATUS, hcntx->hstatus);
+
+		trap->htval = csr_read(CSR_HTVAL);
+		trap->htinst = csr_read(CSR_HTINST);
+	}
+
+	trap->sepc = gcntx->sepc;
+	trap->scause = csr_read(CSR_SCAUSE);
+	trap->stval = csr_read(CSR_STVAL);
+
+	vcpu->arch.last_exit_cpu = vcpu->cpu;
+	guest_state_exit_irqoff();
+	kvm_riscv_vcpu_swap_in_host_state(vcpu);
+}
+
+int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)
+{
+	int ret;
+	struct kvm_cpu_trap trap;
+	struct kvm_run *run = vcpu->run;
+
+	if (!vcpu->arch.ran_atleast_once)
+		kvm_riscv_vcpu_config_ran_once(vcpu);
+
+	/* Mark this VCPU ran at least once */
+	vcpu->arch.ran_atleast_once = true;
+
+	kvm_vcpu_srcu_read_lock(vcpu);
+
+	switch (run->exit_reason) {
+	case KVM_EXIT_MMIO:
+		/* Process MMIO value returned from user-space */
+		ret = kvm_riscv_vcpu_mmio_return(vcpu, vcpu->run);
+		break;
+	case KVM_EXIT_RISCV_SBI:
+		/* Process SBI value returned from user-space */
+		ret = kvm_riscv_vcpu_sbi_return(vcpu, vcpu->run);
+		break;
+	case KVM_EXIT_RISCV_CSR:
+		/* Process CSR value returned from user-space */
+		ret = kvm_riscv_vcpu_csr_return(vcpu, vcpu->run);
+		break;
+	default:
+		ret = 0;
+		break;
+	}
+	if (ret) {
+		kvm_vcpu_srcu_read_unlock(vcpu);
+		return ret;
+	}
+
+	if (!vcpu->wants_to_run) {
+		kvm_vcpu_srcu_read_unlock(vcpu);
+		return -EINTR;
+	}
+
+	vcpu_load(vcpu);
+
+	kvm_sigset_activate(vcpu);
+
+	ret = 1;
+	run->exit_reason = KVM_EXIT_UNKNOWN;
+	while (ret > 0) {
+		/* Check conditions before entering the guest */
+		ret = kvm_xfer_to_guest_mode_handle_work(vcpu);
+		if (ret)
+			continue;
+		ret = 1;
+
+		kvm_riscv_gstage_vmid_update(vcpu);
+
+		ret = kvm_riscv_check_vcpu_requests(vcpu);
+		if (ret <= 0)
+			continue;
+
+		preempt_disable();
+
+		/* Update AIA HW state before entering guest */
+		ret = kvm_riscv_vcpu_aia_update(vcpu);
+		if (ret <= 0) {
+			preempt_enable();
+			continue;
+		}
+
+		local_irq_disable();
+
+		/*
+		 * Ensure we set mode to IN_GUEST_MODE after we disable
+		 * interrupts and before the final VCPU requests check.
+		 * See the comment in kvm_vcpu_exiting_guest_mode() and
+		 * Documentation/virt/kvm/vcpu-requests.rst
+		 */
+		vcpu->mode = IN_GUEST_MODE;
+
+		kvm_vcpu_srcu_read_unlock(vcpu);
+		smp_mb__after_srcu_read_unlock();
+
+		/*
+		 * We might have got VCPU interrupts updated asynchronously
+		 * so update it in HW.
+		 */
+		kvm_riscv_vcpu_flush_interrupts(vcpu);
+
+		/* Update HVIP CSR for current CPU */
+		kvm_riscv_update_hvip(vcpu);
+
+		if (kvm_riscv_gstage_vmid_ver_changed(&vcpu->kvm->arch.vmid) ||
+		    kvm_request_pending(vcpu) ||
+		    xfer_to_guest_mode_work_pending()) {
+			vcpu->mode = OUTSIDE_GUEST_MODE;
+			local_irq_enable();
+			preempt_enable();
+			kvm_vcpu_srcu_read_lock(vcpu);
+			continue;
+		}
+
+		/*
+		 * Sanitize VMID mappings cached (TLB) on current CPU
+		 *
+		 * Note: This should be done after G-stage VMID has been
+		 * updated using kvm_riscv_gstage_vmid_ver_changed()
+		 */
+		kvm_riscv_local_tlb_sanitize(vcpu);
+
+		trace_kvm_entry(vcpu);
+
+		guest_timing_enter_irqoff();
+
+		kvm_riscv_vcpu_enter_exit(vcpu, &trap);
+
+		vcpu->mode = OUTSIDE_GUEST_MODE;
+		vcpu->stat.exits++;
+
+		/* Syncup interrupts state with HW */
+		kvm_riscv_vcpu_sync_interrupts(vcpu);
+
+		/*
+		 * We must ensure that any pending interrupts are taken before
+		 * we exit guest timing so that timer ticks are accounted as
+		 * guest time. Transiently unmask interrupts so that any
+		 * pending interrupts are taken.
+		 *
+		 * There's no barrier which ensures that pending interrupts are
+		 * recognised, so we just hope that the CPU takes any pending
+		 * interrupts between the enable and disable.
+		 */
+		local_irq_enable();
+		local_irq_disable();
+
+		guest_timing_exit_irqoff();
+
+		local_irq_enable();
+
+		trace_kvm_exit(&trap);
+
+		preempt_enable();
+
+		kvm_vcpu_srcu_read_lock(vcpu);
+
+		ret = kvm_riscv_vcpu_exit(vcpu, run, &trap);
+	}
+
+	kvm_sigset_deactivate(vcpu);
+
+	vcpu_put(vcpu);
+
+	kvm_vcpu_srcu_read_unlock(vcpu);
+
+	return ret;
+}