Re: [PATCH] LoongArch: KVM: Add hypercall crash support
From: Huacai Chen
Date: Wed Oct 14 2026 - 00:30:37 EST
On Mon, Oct 12, 2026 at 8:55 AM Bibo Mao <maobibo@xxxxxxxxxxx> wrote:
>
>
>
> On 2026/10/11 下午10:06, Huacai Chen wrote:
> > On Thu, Oct 8, 2026 at 9:50 AM Bibo Mao <maobibo@xxxxxxxxxxx> wrote:
> >>
> >>
> >>
> >> On 2026/10/3 下午12:02, Huacai Chen wrote:
> >>> Hi, Bibo,
> >>>
> >>> On Wed, Sep 9, 2026 at 5:01 PM Bibo Mao <maobibo@xxxxxxxxxxx> wrote:
> >>>>
> >>>> Now generic pvpanic is supported in Linux kernel which will notify VMM
> >>>> with VM crash event. There is some limitations with pvpanic, such as it
> >>>> requires good kernel stack or CPU hardware context and panic() C function
> >>>> is in health state.
> >>>>
> >>>> Sometimes CPU hardware context or kernel/irq stack is corrupted. Hypercall
> >>>> code is more direct and low-level crash method, which is useful in
> >>>> exception entry with assmble language, here add one hypercall code to
> >>>> support VM crash function, so that VMM can capture CPU crash context
> >>>> in time.
> >>>>
> >>>> Signed-off-by: Bibo Mao <maobibo@xxxxxxxxxxx>
> >>>> ---
> >>>> arch/loongarch/include/asm/kvm_host.h | 1 +
> >>>> arch/loongarch/include/asm/kvm_para.h | 15 +++++++++++++++
> >>>> arch/loongarch/include/uapi/asm/kvm.h | 1 +
> >>>> arch/loongarch/include/uapi/asm/kvm_para.h | 1 +
> >>>> arch/loongarch/kvm/exit.c | 11 +++++++++++
> >>>> arch/loongarch/kvm/vm.c | 3 ++-
> >>>> 6 files changed, 31 insertions(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/arch/loongarch/include/asm/kvm_host.h b/arch/loongarch/include/asm/kvm_host.h
> >>>> index 65d91c3ce313..96f9f1317cad 100644
> >>>> --- a/arch/loongarch/include/asm/kvm_host.h
> >>>> +++ b/arch/loongarch/include/asm/kvm_host.h
> >>>> @@ -168,6 +168,7 @@ enum emulation_result {
> >>>> #define LOONGARCH_PV_FEAT_MASK (BIT(KVM_FEATURE_IPI) | \
> >>>> BIT(KVM_FEATURE_PREEMPT) | \
> >>>> BIT(KVM_FEATURE_STEAL_TIME) | \
> >>>> + BIT(KVM_FEATURE_CRASH) | \
> >>>> BIT(KVM_FEATURE_USER_HCALL) | \
> >>>> BIT(KVM_FEATURE_VIRT_EXTIOI))
> >>>>
> >>>> diff --git a/arch/loongarch/include/asm/kvm_para.h b/arch/loongarch/include/asm/kvm_para.h
> >>>> index fb17ba0fa101..3b183bb1aaa8 100644
> >>>> --- a/arch/loongarch/include/asm/kvm_para.h
> >>>> +++ b/arch/loongarch/include/asm/kvm_para.h
> >>>> @@ -14,6 +14,7 @@
> >>>> #define KVM_HCALL_CODE_SERVICE 0
> >>>> #define KVM_HCALL_CODE_SWDBG 1
> >>>> #define KVM_HCALL_CODE_USER_SERVICE 2
> >>>> +#define KVM_HCALL_CODE_CRASH 3
> >>>>
> >>>> #define KVM_HCALL_SERVICE HYPERCALL_ENCODE(HYPERVISOR_KVM, KVM_HCALL_CODE_SERVICE)
> >>>> #define KVM_HCALL_FUNC_IPI 1
> >>>> @@ -22,6 +23,7 @@
> >>>> #define KVM_HCALL_SWDBG HYPERCALL_ENCODE(HYPERVISOR_KVM, KVM_HCALL_CODE_SWDBG)
> >>>>
> >>>> #define KVM_HCALL_USER_SERVICE HYPERCALL_ENCODE(HYPERVISOR_KVM, KVM_HCALL_CODE_USER_SERVICE)
> >>>> +#define KVM_HCALL_CRASH HYPERCALL_ENCODE(HYPERVISOR_KVM, KVM_HCALL_CODE_CRASH)
> >>>>
> >>>> /*
> >>>> * LoongArch hypercall return code
> >>>> @@ -161,6 +163,19 @@ static __always_inline long kvm_hypercall5(u64 fid,
> >>>> return ret;
> >>>> }
> >>>>
> >>>> +static __always_inline long kvm_hypercall_crash(void)
> >>>> +{
> >>>> + register long ret asm("a0");
> >>>> +
> >>>> + __asm__ __volatile__(
> >>>> + "hvcl "__stringify(KVM_HCALL_CRASH)
> >>>> + : "=r" (ret)
> >>>> + : : "memory"
> >>>> + );
> >>>> +
> >>>> + return ret;
> >>>> +}
> >>> Why this function has no caller?
> >> Currently we directly modify guest kernel and hard-code the hypercall
> >> instruction to debug. This inline function is only example for howto
> >> usage, the normal usage such as checking KVM_FEATURE_USER_HCALL feature
> >> + kvm_hypercall_crash() is too complicated in exception entry assemble
> >> language.
> > There are no examples for KVM_HCALL_SWDBG and KVM_HCALL_USER_SERVICE,
> > I think KVM_HCALL_CRASH is no special.
> >
> > So if you agree, I will remove this example when applying.
> Removinge the example usage with KVM_HCALL_CRASH is fine to me.
Now all KVM patches for 7.4 is merged to loongarch-kvm. You can test
it if you want.
Huacai
>
> Regards
> Bibo Mao
> >
> >
> > Huacai
> >
> >>
> >> Regards
> >> Bibo Mao
> >>>
> >>>
> >>> Huacai
> >>>
> >>>> +
> >>>> #ifdef CONFIG_PARAVIRT
> >>>> bool kvm_para_available(void);
> >>>> unsigned int kvm_arch_para_features(void);
> >>>> diff --git a/arch/loongarch/include/uapi/asm/kvm.h b/arch/loongarch/include/uapi/asm/kvm.h
> >>>> index cd0b5c11ca9c..b222bfc48042 100644
> >>>> --- a/arch/loongarch/include/uapi/asm/kvm.h
> >>>> +++ b/arch/loongarch/include/uapi/asm/kvm.h
> >>>> @@ -106,6 +106,7 @@ struct kvm_fpu {
> >>>> #define KVM_LOONGARCH_VM_FEAT_PTW 8
> >>>> #define KVM_LOONGARCH_VM_FEAT_MSGINT 9
> >>>> #define KVM_LOONGARCH_VM_FEAT_PV_PREEMPT 10
> >>>> +#define KVM_LOONGARCH_VM_FEAT_CRASH 11
> >>>>
> >>>> /* Device Control API on vcpu fd */
> >>>> #define KVM_LOONGARCH_VCPU_CPUCFG 0
> >>>> diff --git a/arch/loongarch/include/uapi/asm/kvm_para.h b/arch/loongarch/include/uapi/asm/kvm_para.h
> >>>> index d28cbcadd276..bf1a0c826ca6 100644
> >>>> --- a/arch/loongarch/include/uapi/asm/kvm_para.h
> >>>> +++ b/arch/loongarch/include/uapi/asm/kvm_para.h
> >>>> @@ -16,6 +16,7 @@
> >>>> #define KVM_FEATURE_IPI 1
> >>>> #define KVM_FEATURE_STEAL_TIME 2
> >>>> #define KVM_FEATURE_PREEMPT 3
> >>>> +#define KVM_FEATURE_CRASH 4
> >>>> /* BIT 24 - 31 are features configurable by user space vmm */
> >>>> #define KVM_FEATURE_VIRT_EXTIOI 24
> >>>> #define KVM_FEATURE_USER_HCALL 25
> >>>> diff --git a/arch/loongarch/kvm/exit.c b/arch/loongarch/kvm/exit.c
> >>>> index 4f58e6e2cf86..3b6daa946c76 100644
> >>>> --- a/arch/loongarch/kvm/exit.c
> >>>> +++ b/arch/loongarch/kvm/exit.c
> >>>> @@ -911,6 +911,17 @@ static int kvm_handle_hypercall(struct kvm_vcpu *vcpu, int ecode)
> >>>> vcpu->run->hypercall.ret = KVM_HCALL_INVALID_CODE;
> >>>> ret = RESUME_HOST;
> >>>> break;
> >>>> + case KVM_HCALL_CRASH:
> >>>> + if (!kvm_guest_has_pv_feature(vcpu, KVM_FEATURE_CRASH)) {
> >>>> + kvm_write_reg(vcpu, LOONGARCH_GPR_A0, KVM_HCALL_INVALID_CODE);
> >>>> + break;
> >>>> + }
> >>>> +
> >>>> + vcpu->run->exit_reason = KVM_EXIT_SYSTEM_EVENT;
> >>>> + vcpu->run->system_event.type = KVM_SYSTEM_EVENT_CRASH;
> >>>> + vcpu->run->system_event.ndata = 0;
> >>>> + ret = RESUME_HOST;
> >>>> + break;
> >>>> case KVM_HCALL_SWDBG:
> >>>> /* KVM_HCALL_SWDBG only in effective when SW_BP is enabled */
> >>>> if (vcpu->guest_debug & KVM_GUESTDBG_SW_BP_MASK) {
> >>>> diff --git a/arch/loongarch/kvm/vm.c b/arch/loongarch/kvm/vm.c
> >>>> index 6dabb227a732..e5d10ceb5e67 100644
> >>>> --- a/arch/loongarch/kvm/vm.c
> >>>> +++ b/arch/loongarch/kvm/vm.c
> >>>> @@ -49,8 +49,9 @@ static void kvm_vm_init_features(struct kvm *kvm)
> >>>> kvm->arch.kvm_features |= BIT(KVM_LOONGARCH_VM_FEAT_PMU);
> >>>>
> >>>> /* Enable all PV features by default */
> >>>> - kvm->arch.pv_features |= BIT(KVM_FEATURE_IPI);
> >>>> + kvm->arch.pv_features |= BIT(KVM_FEATURE_IPI) | BIT(KVM_FEATURE_CRASH);
> >>>> kvm->arch.kvm_features |= BIT(KVM_LOONGARCH_VM_FEAT_PV_IPI);
> >>>> + kvm->arch.kvm_features |= BIT(KVM_LOONGARCH_VM_FEAT_CRASH);
> >>>> if (kvm_pvtime_supported()) {
> >>>> kvm->arch.pv_features |= BIT(KVM_FEATURE_PREEMPT);
> >>>> kvm->arch.pv_features |= BIT(KVM_FEATURE_STEAL_TIME);
> >>>>
> >>>> base-commit: df2908090cda368b01ff43709f51890076c56157
> >>>> --
> >>>> 2.39.3
> >>>>
> >>>>
> >>
>