[PATCH v4 22/46] KVM: PPC: Book3S HV P9: Stop handling hcalls in real-mode in the P9 path

[Nicholas Piggin]

Excerpts from Fabiano Rosas's message of March 24, 2021 8:57 am:
> Nicholas Piggin <npiggin at gmail.com> writes:
> 
>> In the interest of minimising the amount of code that is run in
>> "real-mode", don't handle hcalls in real mode in the P9 path.
>>
>> POWER8 and earlier are much more expensive to exit from HV real mode
>> and switch to host mode, because on those processors HV interrupts get
>> to the hypervisor with the MMU off, and the other threads in the core
>> need to be pulled out of the guest, and SLBs all need to be saved,
>> ERATs invalidated, and host SLB reloaded before the MMU is re-enabled
>> in host mode. Hash guests also require a lot of hcalls to run. The
>> XICS interrupt controller requires hcalls to run.
>>
>> By contrast, POWER9 has independent thread switching, and in radix mode
>> the hypervisor is already in a host virtual memory mode when the HV
>> interrupt is taken. Radix + xive guests don't need hcalls to handle
>> interrupts or manage translations.
>>
>> So it's much less important to handle hcalls in real mode in P9.
>>
>> Signed-off-by: Nicholas Piggin <npiggin at gmail.com>
>> ---
> 
> <snip>
> 
>> diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
>> index fa7614c37e08..17739aaee3d8 100644
>> --- a/arch/powerpc/kvm/book3s_hv.c
>> +++ b/arch/powerpc/kvm/book3s_hv.c
>> @@ -1142,12 +1142,13 @@ int kvmppc_pseries_do_hcall(struct kvm_vcpu *vcpu)
>>  }
>>
>>  /*
>> - * Handle H_CEDE in the nested virtualization case where we haven't
>> - * called the real-mode hcall handlers in book3s_hv_rmhandlers.S.
>> + * Handle H_CEDE in the P9 path where we don't call the real-mode hcall
>> + * handlers in book3s_hv_rmhandlers.S.
>> + *
>>   * This has to be done early, not in kvmppc_pseries_do_hcall(), so
>>   * that the cede logic in kvmppc_run_single_vcpu() works properly.
>>   */
>> -static void kvmppc_nested_cede(struct kvm_vcpu *vcpu)
>> +static void kvmppc_cede(struct kvm_vcpu *vcpu)
>>  {
>>  	vcpu->arch.shregs.msr |= MSR_EE;
>>  	vcpu->arch.ceded = 1;
>> @@ -1403,9 +1404,15 @@ static int kvmppc_handle_exit_hv(struct kvm_vcpu *vcpu,
>>  		/* hcall - punt to userspace */
>>  		int i;
>>
>> -		/* hypercall with MSR_PR has already been handled in rmode,
>> -		 * and never reaches here.
>> -		 */
>> +		if (unlikely(vcpu->arch.shregs.msr & MSR_PR)) {
>> +			/*
>> +			 * Guest userspace executed sc 1, reflect it back as a
>> +			 * privileged program check interrupt.
>> +			 */
>> +			kvmppc_core_queue_program(vcpu, SRR1_PROGPRIV);
>> +			r = RESUME_GUEST;
>> +			break;
>> +		}
> 
> This patch bypasses sc_1_fast_return so it breaks KVM-PR. L1 loops with
> the following output:
> 
> [    9.503929][ T3443] Couldn't emulate instruction 0x4e800020 (op 19 xop 16)
> [    9.503990][ T3443] kvmppc_exit_pr_progint: emulation at 48f4 failed (4e800020)
> [    9.504080][ T3443] Couldn't emulate instruction 0x4e800020 (op 19 xop 16)
> [    9.504170][ T3443] kvmppc_exit_pr_progint: emulation at 48f4 failed (4e800020)
> 
> 0x4e800020 is a blr after a sc 1 in SLOF.
> 
> For KVM-PR we need to inject a 0xc00 at some point, either here or
> before branching to no_try_real in book3s_hv_rmhandlers.S.

Ah, I didn't know about that PR KVM (I suppose I should test it but I 
haven't been able to get it running in the past).

Should be able to deal with that. This patch probably shouldn't change 
the syscall behaviour like this anyway.

Thanks,
Nick