RE: [PATCH v15 6/9] arm64/kvm: Add hypercall service for kvm ptp.
From: Jianyong Wu
Date: Tue Nov 24 2020 - 00:21:28 EST
Hi Marc,
> -----Original Message-----
> From: Marc Zyngier <maz@xxxxxxxxxx>
> Sent: Monday, November 23, 2020 6:44 PM
> To: Jianyong Wu <Jianyong.Wu@xxxxxxx>
> Cc: netdev@xxxxxxxxxxxxxxx; yangbo.lu@xxxxxxx; john.stultz@xxxxxxxxxx;
> tglx@xxxxxxxxxxxxx; pbonzini@xxxxxxxxxx; sean.j.christopherson@xxxxxxxxx;
> richardcochran@xxxxxxxxx; Mark Rutland <Mark.Rutland@xxxxxxx>;
> will@xxxxxxxxxx; Suzuki Poulose <Suzuki.Poulose@xxxxxxx>; Andre
> Przywara <Andre.Przywara@xxxxxxx>; Steven Price
> <Steven.Price@xxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx; linux-arm-
> kernel@xxxxxxxxxxxxxxxxxxx; kvmarm@xxxxxxxxxxxxxxxxxxxxx;
> kvm@xxxxxxxxxxxxxxx; Steve Capper <Steve.Capper@xxxxxxx>; Justin He
> <Justin.He@xxxxxxx>; nd <nd@xxxxxxx>
> Subject: Re: [PATCH v15 6/9] arm64/kvm: Add hypercall service for kvm ptp.
>
> On 2020-11-11 06:22, Jianyong Wu wrote:
> > ptp_kvm will get this service through SMCC call.
> > The service offers wall time and cycle count of host to guest.
> > The caller must specify whether they want the host cycle count or the
> > difference between host cycle count and cntvoff.
> >
> > Signed-off-by: Jianyong Wu <jianyong.wu@xxxxxxx>
> > ---
> > arch/arm64/kvm/hypercalls.c | 61
> +++++++++++++++++++++++++++++++++++++
> > include/linux/arm-smccc.h | 17 +++++++++++
> > 2 files changed, 78 insertions(+)
> >
> > diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c
> > index b9d8607083eb..f7d189563f3d 100644
> > --- a/arch/arm64/kvm/hypercalls.c
> > +++ b/arch/arm64/kvm/hypercalls.c
> > @@ -9,6 +9,51 @@
> > #include <kvm/arm_hypercalls.h>
> > #include <kvm/arm_psci.h>
> >
> > +static void kvm_ptp_get_time(struct kvm_vcpu *vcpu, u64 *val) {
> > + struct system_time_snapshot systime_snapshot;
> > + u64 cycles = ~0UL;
> > + u32 feature;
> > +
> > + /*
> > + * system time and counter value must captured in the same
> > + * time to keep consistency and precision.
> > + */
> > + ktime_get_snapshot(&systime_snapshot);
> > +
> > + // binding ptp_kvm clocksource to arm_arch_counter
> > + if (systime_snapshot.cs_id != CSID_ARM_ARCH_COUNTER)
> > + return;
> > +
> > + val[0] = upper_32_bits(systime_snapshot.real);
> > + val[1] = lower_32_bits(systime_snapshot.real);
>
> What is the endianness of these values? I can't see it defined anywhere, and
> this is likely not to work if guest and hypervisor don't align.
>
> > +
> > + /*
> > + * which of virtual counter or physical counter being
> > + * asked for is decided by the r1 value of SMCCC
> > + * call. If no invalid r1 value offered, default cycle
> > + * value(-1) will be returned.
> > + * Note: keep in mind that feature is u32 and smccc_get_arg1
> > + * will return u64, so need auto cast here.
> > + */
> > + feature = smccc_get_arg1(vcpu);
> > + switch (feature) {
> > + case ARM_PTP_VIRT_COUNTER:
> > + cycles = systime_snapshot.cycles - vcpu_read_sys_reg(vcpu,
> > CNTVOFF_EL2);
> > + break;
> > + case ARM_PTP_PHY_COUNTER:
> > + cycles = systime_snapshot.cycles;
> > + break;
> > + case ARM_PTP_NONE_COUNTER:
>
> What is this "NONE" counter?
Yeah, there is no counter named "NONE". this is not a counter, and only means no counter data needed for guest and just do nothing.
If no this arm here, it will go to the default one and return "NOT_SUPPORTED"
>
> > + break;
> > + default:
> > + val[0] = SMCCC_RET_NOT_SUPPORTED;
> > + break;
> > + }
> > + val[2] = upper_32_bits(cycles);
> > + val[3] = lower_32_bits(cycles);
>
> Same problem as above.
>
> > +}
> > +
> > int kvm_hvc_call_handler(struct kvm_vcpu *vcpu) {
> > u32 func_id = smccc_get_function(vcpu); @@ -79,6 +124,22 @@ int
> > kvm_hvc_call_handler(struct kvm_vcpu *vcpu)
> > break;
> > case ARM_SMCCC_VENDOR_HYP_KVM_FEATURES_FUNC_ID:
> > val[0] = BIT(ARM_SMCCC_KVM_FUNC_FEATURES);
> > + val[0] |= BIT(ARM_SMCCC_KVM_FUNC_KVM_PTP);
> > + break;
> > + /*
> > + * This serves virtual kvm_ptp.
> > + * Four values will be passed back.
> > + * reg0 stores high 32-bits of host ktime;
> > + * reg1 stores low 32-bits of host ktime;
> > + * For ARM_PTP_VIRT_COUNTER:
> > + * reg2 stores high 32-bits of difference of host cycles and cntvoff;
> > + * reg3 stores low 32-bits of difference of host cycles and cntvoff.
> > + * For ARM_PTP_PHY_COUNTER:
> > + * reg2 stores the high 32-bits of host cycles;
> > + * reg3 stores the low 32-bits of host cycles.
> > + */
> > + case ARM_SMCCC_VENDOR_HYP_KVM_PTP_FUNC_ID:
> > + kvm_ptp_get_time(vcpu, val);
> > break;
> > default:
> > return kvm_psci_call(vcpu);
> > diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
> > index d75408141137..a03c5dd409d3 100644
> > --- a/include/linux/arm-smccc.h
> > +++ b/include/linux/arm-smccc.h
> > @@ -103,6 +103,7 @@
> >
> > /* KVM "vendor specific" services */
> > #define ARM_SMCCC_KVM_FUNC_FEATURES 0
> > +#define ARM_SMCCC_KVM_FUNC_KVM_PTP 1
>
> I think having KVM once in the name is enough.
>
> > #define ARM_SMCCC_KVM_FUNC_FEATURES_2 127
> > #define ARM_SMCCC_KVM_NUM_FUNCS 128
> >
> > @@ -114,6 +115,22 @@
> >
> > #define SMCCC_ARCH_WORKAROUND_RET_UNAFFECTED 1
> >
> > +/*
> > + * ptp_kvm is a feature used for time sync between vm and host.
> > + * ptp_kvm module in guest kernel will get service from host using
> > + * this hypercall ID.
> > + */
> > +#define ARM_SMCCC_VENDOR_HYP_KVM_PTP_FUNC_ID
> \
> > + ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,
> \
> > + ARM_SMCCC_SMC_32,
> \
> > + ARM_SMCCC_OWNER_VENDOR_HYP,
> \
> > + ARM_SMCCC_KVM_FUNC_KVM_PTP)
> > +
> > +/* ptp_kvm counter type ID */
> > +#define ARM_PTP_VIRT_COUNTER 0
> > +#define ARM_PTP_PHY_COUNTER 1
> > +#define ARM_PTP_NONE_COUNTER 2
>
> The architecture definitely doesn't have this last counter.
Yeah, this is just represent no counter data needed from guest.
Some annotation should be added here.
Thanks
Jianyong
>
> > +
> > /* Paravirtualised time calls (defined by ARM DEN0057A) */
> > #define ARM_SMCCC_HV_PV_TIME_FEATURES
> \
> > ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,
> \
>
> Thanks,
>
> M.
> --
> Jazz is not dead. It just smells funny...