On Mon, Jul 06, 2015 at 10:17:44AM +0800, [email protected] wrote: > From: Shannon Zhao <[email protected]> > > Since the reset value of PMUSERENR_EL0 is UNKNOWN, use reset_unknown for > its reset handler. Add access handler which emulates writing and reading > PMUSERENR_EL0 register. > > Signed-off-by: Shannon Zhao <[email protected]> > --- > arch/arm64/kvm/sys_regs.c | 15 ++++++++++++++- > include/kvm/arm_pmu.h | 1 + > 2 files changed, 15 insertions(+), 1 deletion(-) > > diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c > index ec80937..d5984d0 100644 > --- a/arch/arm64/kvm/sys_regs.c > +++ b/arch/arm64/kvm/sys_regs.c > @@ -522,6 +522,19 @@ static bool access_pmovsclr(struct kvm_vcpu *vcpu, > return true; > } > > +/* PMUSERENR_EL0 accessor. */ > +static bool access_pmuserenr(struct kvm_vcpu *vcpu, > + const struct sys_reg_params *p, > + const struct sys_reg_desc *r) > +{ > + if (p->is_write) > + vcpu->arch.pmu.user_enable |= *vcpu_reg(vcpu, p->Rt); > + else > + *vcpu_reg(vcpu, p->Rt) = vcpu->arch.pmu.user_enable; > + > + return true; > +} > + > /* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */ > #define DBG_BCR_BVR_WCR_WVR_EL1(n) \ > /* DBGBVRn_EL1 */ \ > @@ -746,7 +759,7 @@ static const struct sys_reg_desc sys_reg_descs[] = { > access_pmxevcntr, reset_unknown, PMXEVCNTR_EL0 }, > /* PMUSERENR_EL0 */ > { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000), > - trap_raz_wi }, > + access_pmuserenr, reset_unknown, PMUSERENR_EL0 }, > /* PMOVSSET_EL0 */ > { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011), > access_pmovsset, reset_unknown, PMOVSSET_EL0 }, > diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h > index dee8356..4f3d8a6 100644 > --- a/include/kvm/arm_pmu.h > +++ b/include/kvm/arm_pmu.h > @@ -39,6 +39,7 @@ struct kvm_pmu { > /* IRQ pending flag */ > bool irq_pending; > struct irq_work irq_work; > + u32 user_enable;
why not store this in PMUSERENR_EL0 and get VM migration of this state included for free? Also, I assume the functionality to respect these flags are implemented in a later patch or simply not supported? It would have been good to note this in the cover letter. -Christoffer > struct kvm_pmc pmc[ARMV8_MAX_COUNTERS]; > #endif > }; > -- > 2.1.0 > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
