Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1655732 > unrolled thread
| Started by | Bandan Das <bsd@redhat.com> |
|---|---|
| First post | 2017-06-01 22:30 +0200 |
| Last post | 2017-06-02 10:50 +0200 |
| Articles | 2 — 2 participants |
Back to article view | Back to linux.kernel
This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by
below is the oldest one visible, not the original post.
Re: [RFC 08/55] KVM: arm64: Set virtual EL2 context depending on the guest exception level Bandan Das <bsd@redhat.com> - 2017-06-01 22:30 +0200
Re: [RFC 08/55] KVM: arm64: Set virtual EL2 context depending on the guest exception level Marc Zyngier <marc.zyngier@arm.com> - 2017-06-02 10:50 +0200
| From | Bandan Das <bsd@redhat.com> |
|---|---|
| Date | 2017-06-01 22:30 +0200 |
| Subject | Re: [RFC 08/55] KVM: arm64: Set virtual EL2 context depending on the guest exception level |
| Message-ID | <tNETg-7v6-3@gated-at.bofh.it> |
Jintack Lim <jintack@cs.columbia.edu> writes:
> From: Christoffer Dall <christoffer.dall@linaro.org>
>
> Set up virutal EL2 context to hardware if the guest exception level is
> EL2.
>
> Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
> Signed-off-by: Jintack Lim <jintack@cs.columbia.edu>
> ---
> arch/arm64/kvm/context.c | 32 ++++++++++++++++++++++++++------
> 1 file changed, 26 insertions(+), 6 deletions(-)
>
> diff --git a/arch/arm64/kvm/context.c b/arch/arm64/kvm/context.c
> index 320afc6..acb4b1e 100644
> --- a/arch/arm64/kvm/context.c
> +++ b/arch/arm64/kvm/context.c
> @@ -25,10 +25,25 @@
> void kvm_arm_setup_shadow_state(struct kvm_vcpu *vcpu)
> {
> struct kvm_cpu_context *ctxt = &vcpu->arch.ctxt;
> + if (unlikely(vcpu_mode_el2(vcpu))) {
> + ctxt->hw_pstate = *vcpu_cpsr(vcpu) & ~PSR_MODE_MASK;
>
> - ctxt->hw_pstate = *vcpu_cpsr(vcpu);
> - ctxt->hw_sys_regs = ctxt->sys_regs;
> - ctxt->hw_sp_el1 = ctxt->gp_regs.sp_el1;
> + /*
> + * We emulate virtual EL2 mode in hardware EL1 mode using the
> + * same stack pointer mode as the guest expects.
> + */
> + if ((*vcpu_cpsr(vcpu) & PSR_MODE_MASK) == PSR_MODE_EL2h)
> + ctxt->hw_pstate |= PSR_MODE_EL1h;
> + else
> + ctxt->hw_pstate |= PSR_MODE_EL1t;
> +
I see vcpu_mode(el2) does
return mode == PSR_MODE_EL2h || mode == PSR_MODE_EL2t;
I can't seem to find this, what's the difference between
the modes: PSR_MODE_EL2h/EL2t ?
Bandan
> + ctxt->hw_sys_regs = ctxt->shadow_sys_regs;
> + ctxt->hw_sp_el1 = ctxt->el2_regs[SP_EL2];
> + } else {
> + ctxt->hw_pstate = *vcpu_cpsr(vcpu);
> + ctxt->hw_sys_regs = ctxt->sys_regs;
> + ctxt->hw_sp_el1 = ctxt->gp_regs.sp_el1;
> + }
> }
>
> /**
> @@ -38,9 +53,14 @@ void kvm_arm_setup_shadow_state(struct kvm_vcpu *vcpu)
> void kvm_arm_restore_shadow_state(struct kvm_vcpu *vcpu)
> {
> struct kvm_cpu_context *ctxt = &vcpu->arch.ctxt;
> -
> - *vcpu_cpsr(vcpu) = ctxt->hw_pstate;
> - ctxt->gp_regs.sp_el1 = ctxt->hw_sp_el1;
> + if (unlikely(vcpu_mode_el2(vcpu))) {
> + *vcpu_cpsr(vcpu) &= PSR_MODE_MASK;
> + *vcpu_cpsr(vcpu) |= ctxt->hw_pstate & ~PSR_MODE_MASK;
> + ctxt->el2_regs[SP_EL2] = ctxt->hw_sp_el1;
> + } else {
> + *vcpu_cpsr(vcpu) = ctxt->hw_pstate;
> + ctxt->gp_regs.sp_el1 = ctxt->hw_sp_el1;
> + }
> }
>
> void kvm_arm_init_cpu_context(kvm_cpu_context_t *cpu_ctxt)
[toc] | [next] | [standalone]
| From | Marc Zyngier <marc.zyngier@arm.com> |
|---|---|
| Date | 2017-06-02 10:50 +0200 |
| Subject | Re: [RFC 08/55] KVM: arm64: Set virtual EL2 context depending on the guest exception level |
| Message-ID | <tNQrn-74M-13@gated-at.bofh.it> |
| In reply to | #1655732 |
On 01/06/17 21:22, Bandan Das wrote:
> Jintack Lim <jintack@cs.columbia.edu> writes:
>
>> From: Christoffer Dall <christoffer.dall@linaro.org>
>>
>> Set up virutal EL2 context to hardware if the guest exception level is
>> EL2.
>>
>> Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
>> Signed-off-by: Jintack Lim <jintack@cs.columbia.edu>
>> ---
>> arch/arm64/kvm/context.c | 32 ++++++++++++++++++++++++++------
>> 1 file changed, 26 insertions(+), 6 deletions(-)
>>
>> diff --git a/arch/arm64/kvm/context.c b/arch/arm64/kvm/context.c
>> index 320afc6..acb4b1e 100644
>> --- a/arch/arm64/kvm/context.c
>> +++ b/arch/arm64/kvm/context.c
>> @@ -25,10 +25,25 @@
>> void kvm_arm_setup_shadow_state(struct kvm_vcpu *vcpu)
>> {
>> struct kvm_cpu_context *ctxt = &vcpu->arch.ctxt;
>> + if (unlikely(vcpu_mode_el2(vcpu))) {
>> + ctxt->hw_pstate = *vcpu_cpsr(vcpu) & ~PSR_MODE_MASK;
>>
>> - ctxt->hw_pstate = *vcpu_cpsr(vcpu);
>> - ctxt->hw_sys_regs = ctxt->sys_regs;
>> - ctxt->hw_sp_el1 = ctxt->gp_regs.sp_el1;
>> + /*
>> + * We emulate virtual EL2 mode in hardware EL1 mode using the
>> + * same stack pointer mode as the guest expects.
>> + */
>> + if ((*vcpu_cpsr(vcpu) & PSR_MODE_MASK) == PSR_MODE_EL2h)
>> + ctxt->hw_pstate |= PSR_MODE_EL1h;
>> + else
>> + ctxt->hw_pstate |= PSR_MODE_EL1t;
>> +
>
> I see vcpu_mode(el2) does
> return mode == PSR_MODE_EL2h || mode == PSR_MODE_EL2t;
>
> I can't seem to find this, what's the difference between
> the modes: PSR_MODE_EL2h/EL2t ?
The difference is the stack pointer that is getting used. When the CPU
is at ELxh, it uses SPx at ELx. When at ELxt, it uses SP0 (the userspace
stack pointer). See the definition of SPSR_EL2 in the ARMv8 ARM.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web