| Message ID | 20190621093843.220980-12-marc.zyngier@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM: arm64: ARMv8.3 Nested Virtualization support | expand |
On 6/21/19 10:37 AM, Marc Zyngier wrote: > From: Jintack Lim <jintack.lim@linaro.org> > > Now that the psci call is done by the smc instruction when nested This suggests that we have support for PSCI calls using SMC as the conduit, but that is not the case, as the handle_smc function is not changed by this commit, and support for PSCI via SMC is added later in patch 22/59 "KVM: arm64: nv: Handle PSCI call via smc from the guest". Perhaps the commit message should be reworded to reflect that? > virtualization is enabled, it is clear that all hvc instruction from the > VM (including from the virtual EL2) are supposed to handled in the > virtual EL2. > > Signed-off-by: Jintack Lim <jintack.lim@linaro.org> > Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> > --- > arch/arm64/kvm/handle_exit.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c > index 516aead3c2a9..6c0ac52b34cc 100644 > --- a/arch/arm64/kvm/handle_exit.c > +++ b/arch/arm64/kvm/handle_exit.c > @@ -30,6 +30,7 @@ > #include <asm/kvm_coproc.h> > #include <asm/kvm_emulate.h> > #include <asm/kvm_mmu.h> > +#include <asm/kvm_nested.h> > #include <asm/debug-monitors.h> > #include <asm/traps.h> > > @@ -52,6 +53,12 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run) > kvm_vcpu_hvc_get_imm(vcpu)); > vcpu->stat.hvc_exit_stat++; > > + /* Forward hvc instructions to the virtual EL2 if the guest has EL2. */ > + if (nested_virt_in_use(vcpu)) { > + kvm_inject_nested_sync(vcpu, kvm_vcpu_get_hsr(vcpu)); > + return 1; > + } > + > ret = kvm_hvc_call_handler(vcpu); > if (ret < 0) { > vcpu_set_reg(vcpu, 0, ~0UL);
On 25/06/2019 14:13, Alexandru Elisei wrote: > On 6/21/19 10:37 AM, Marc Zyngier wrote: >> From: Jintack Lim <jintack.lim@linaro.org> >> >> Now that the psci call is done by the smc instruction when nested > This suggests that we have support for PSCI calls using SMC as the conduit, but > that is not the case, as the handle_smc function is not changed by this commit, > and support for PSCI via SMC is added later in patch 22/59 "KVM: arm64: nv: > Handle PSCI call via smc from the guest". Perhaps the commit message should be > reworded to reflect that? Sure. Thanks, M.
On 6/25/19 2:13 PM, Alexandru Elisei wrote: > On 6/21/19 10:37 AM, Marc Zyngier wrote: >> From: Jintack Lim <jintack.lim@linaro.org> >> >> Now that the psci call is done by the smc instruction when nested > This suggests that we have support for PSCI calls using SMC as the conduit, but > that is not the case, as the handle_smc function is not changed by this commit, > and support for PSCI via SMC is added later in patch 22/59 "KVM: arm64: nv: > Handle PSCI call via smc from the guest". Perhaps the commit message should be > reworded to reflect that? >> virtualization is enabled, it is clear that all hvc instruction from the >> VM (including from the virtual EL2) are supposed to handled in the >> virtual EL2. >> >> Signed-off-by: Jintack Lim <jintack.lim@linaro.org> >> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> >> --- >> arch/arm64/kvm/handle_exit.c | 7 +++++++ >> 1 file changed, 7 insertions(+) >> >> diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c >> index 516aead3c2a9..6c0ac52b34cc 100644 >> --- a/arch/arm64/kvm/handle_exit.c >> +++ b/arch/arm64/kvm/handle_exit.c >> @@ -30,6 +30,7 @@ >> #include <asm/kvm_coproc.h> >> #include <asm/kvm_emulate.h> >> #include <asm/kvm_mmu.h> >> +#include <asm/kvm_nested.h> >> #include <asm/debug-monitors.h> >> #include <asm/traps.h> >> >> @@ -52,6 +53,12 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run) >> kvm_vcpu_hvc_get_imm(vcpu)); >> vcpu->stat.hvc_exit_stat++; >> >> + /* Forward hvc instructions to the virtual EL2 if the guest has EL2. */ >> + if (nested_virt_in_use(vcpu)) { >> + kvm_inject_nested_sync(vcpu, kvm_vcpu_get_hsr(vcpu)); >> + return 1; >> + } According to ARM DDI 0487E.a, when HCR_EL2.HCD = 1, HVC instructions are undefined at EL2 and EL1. Thanks, Alex >> + >> ret = kvm_hvc_call_handler(vcpu); >> if (ret < 0) { >> vcpu_set_reg(vcpu, 0, ~0UL);
diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index 516aead3c2a9..6c0ac52b34cc 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -30,6 +30,7 @@ #include <asm/kvm_coproc.h> #include <asm/kvm_emulate.h> #include <asm/kvm_mmu.h> +#include <asm/kvm_nested.h> #include <asm/debug-monitors.h> #include <asm/traps.h> @@ -52,6 +53,12 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run) kvm_vcpu_hvc_get_imm(vcpu)); vcpu->stat.hvc_exit_stat++; + /* Forward hvc instructions to the virtual EL2 if the guest has EL2. */ + if (nested_virt_in_use(vcpu)) { + kvm_inject_nested_sync(vcpu, kvm_vcpu_get_hsr(vcpu)); + return 1; + } + ret = kvm_hvc_call_handler(vcpu); if (ret < 0) { vcpu_set_reg(vcpu, 0, ~0UL);