On 17.02.2012, at 22:55, Scott Wood wrote:
> On 02/17/2012 11:13 AM, Alexander Graf wrote:
>> When one vcpu wants to kick another, it can issue a special IPI instruction
>> called msgsnd. This patch emulates this instruction, its clearing counterpart
>> and the infrastructure required to actually trigger that interrupt inside
>> a guest vcpu.
>>
>> With this patch, SMP guests on e500mc work.
>>
>> Signed-off-by: Alexander Graf <agraf@xxxxxxx>
>> ---
>> arch/powerpc/kvm/booke.c | 6 +++
>> arch/powerpc/kvm/e500_emulate.c | 68 +++++++++++++++++++++++++++++++++++++++
>> 2 files changed, 74 insertions(+), 0 deletions(-)
>>
>> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c
>> index 3dd200d..ce1599d 100644
>> --- a/arch/powerpc/kvm/booke.c
>> +++ b/arch/powerpc/kvm/booke.c
>> @@ -326,6 +326,9 @@ static int kvmppc_booke_irqprio_deliver(struct kvm_vcpu *vcpu,
>> int_class = INT_CLASS_NONCRIT;
>> break;
>> case BOOKE_IRQPRIO_CRITICAL:
>> +#ifdef CONFIG_KVM_E500MC
>> + case BOOKE_IRQPRIO_DBELL_CRIT:
>> +#endif
>> allowed = vcpu->arch.shared->msr & MSR_CE;
>> allowed = allowed && !crit;
>> msr_mask = MSR_GS | MSR_ME;
>> @@ -342,6 +345,9 @@ static int kvmppc_booke_irqprio_deliver(struct kvm_vcpu *vcpu,
>> keep_irq = true;
>> /* fall through */
>> case BOOKE_IRQPRIO_EXTERNAL:
>> +#ifdef CONFIG_KVM_E500MC
>> + case BOOKE_IRQPRIO_DBELL:
>> +#endif
>
> This isn't e500mc specific -- it's in the ISA as "Embedded.Processor
> Control".
>
> Any harm in just removing the ifdef (similar to tlbilx)?
Well, to me this is more of an indication "this should become a runtime check one day" in case we want to combine the two targets. On e500v2, we don't know what a doorbell interrupt is, so we really shouldn't be delivering one either.
>
>> allowed = vcpu->arch.shared->msr & MSR_EE;
>> allowed = allowed && !crit;
>> msr_mask = MSR_GS | MSR_CE | MSR_ME | MSR_DE;
>> diff --git a/arch/powerpc/kvm/e500_emulate.c b/arch/powerpc/kvm/e500_emulate.c
>> index 98b6c1c..29d5604 100644
>> --- a/arch/powerpc/kvm/e500_emulate.c
>> +++ b/arch/powerpc/kvm/e500_emulate.c
>> @@ -14,16 +14,74 @@
>>
>> #include <asm/kvm_ppc.h>
>> #include <asm/disassemble.h>
>> +#include <asm/dbell.h>
>>
>> #include "booke.h"
>> #include "e500.h"
>>
>> +#define XOP_MSGSND 206
>> +#define XOP_MSGCLR 238
>> #define XOP_TLBIVAX 786
>> #define XOP_TLBSX 914
>> #define XOP_TLBRE 946
>> #define XOP_TLBWE 978
>> #define XOP_TLBILX 18
>>
>> +#ifdef CONFIG_KVM_E500MC
>> +static int dbell2prio(ulong param)
>> +{
>> + int msg = param & PPC_DBELL_TYPE(-1);
>
> Maybe introduce PPC_DBELL_TYPE_MASK or GET_PPC_DBELL_TYPE?
TYPE_MASK I'd say.
>
>> + int prio = -1;
>> +
>> + switch (msg) {
>> + case PPC_DBELL_TYPE(PPC_DBELL):
>> + prio = BOOKE_IRQPRIO_DBELL;
>> + break;
>> + case PPC_DBELL_TYPE(PPC_DBELL_CRIT):
>> + prio = BOOKE_IRQPRIO_DBELL_CRIT;
>> + break;
>> + default:
>> + break;
>> + }
>> +
>> + return prio;
>> +}
>> +
>> +static int kvmppc_e500_emul_msgclr(struct kvm_vcpu *vcpu, int rb)
>> +{
>> + ulong param = vcpu->arch.gpr[rb];
>> + int prio = dbell2prio(param);
>> +
>> + if (prio < 0)
>> + return EMULATE_FAIL;
>> +
>> + clear_bit(prio, &vcpu->arch.pending_exceptions);
>> + return EMULATE_DONE;
>> +}
>> +
>> +static int kvmppc_e500_emul_msgsnd(struct kvm_vcpu *vcpu, int rb)
>> +{
>> + ulong param = vcpu->arch.gpr[rb];
>> + int prio = dbell2prio(rb);
>> + int pir = param & 0x3fff;
>
> Introduce PPC_DBELL_PIR_MASK or GET_PPC_DBELL_PIR?
Good point :)
>
>> + int i;
>> + struct kvm_vcpu *cvcpu;
>> +
>> + if (prio < 0)
>> + return EMULATE_FAIL;
>> +
>> + kvm_for_each_vcpu(i, cvcpu, vcpu->kvm) {
>> + int cpir = cvcpu->arch.shared->pir;
>> + if ((param & PPC_DBELL_MSG_BRDCAST) || (cpir == pir)) {
>> + set_bit(prio, &cvcpu->arch.pending_exceptions);
>> + kvm_vcpu_kick(cvcpu);
>> + }
>> + }
>
> Should this be a kvm_make_request instead (with a separate
> pending_doorbell bool in vcpu that msgclr can act on), considering
> earlier discussion of phasing out atomics on pending_exceptions, in
> favor of requests?
Yeah, I was thinking about that too, but right now we already have some direct use of pending_exceptions in different places around the code. So today, that is our public interface.
I'd rather go in and clean up the whole thing to make pending_exceptions private in a separate cleanup round, rather than having it part of e500mc support.
Alex
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
