Re: [PATCH v4 1/2] target/arm: kvm: Handle DABT with no valid ISS

[Beata Michalska]

Hi,

On Mon, 23 Mar 2020 at 12:44, Andrew Jones <address@hidden> wrote:
>
> On Mon, Mar 23, 2020 at 11:32:26AM +0000, Beata Michalska wrote:
> > On ARMv7 & ARMv8 some load/store instructions might trigger a data abort
> > exception with no valid ISS info to be decoded. The lack of decode info
> > makes it at least tricky to emulate those instruction which is one of the
> > (many) reasons why KVM will not even try to do so.
> >
> > Add support for handling those by requesting KVM to inject external
> > dabt into the quest.
> >
> > Signed-off-by: Beata Michalska <address@hidden>
> > ---
> >  target/arm/cpu.h     |  2 ++
> >  target/arm/kvm.c     | 54 
> > ++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  target/arm/kvm_arm.h | 11 +++++++++++
> >  3 files changed, 67 insertions(+)
> >
> > diff --git a/target/arm/cpu.h b/target/arm/cpu.h
> > index 4ffd991..4f834c1 100644
> > --- a/target/arm/cpu.h
> > +++ b/target/arm/cpu.h
> > @@ -560,6 +560,8 @@ typedef struct CPUARMState {
> >          uint64_t esr;
> >      } serror;
> >
> > +    uint8_t ext_dabt_pending; /* Request for injecting ext DABT */
> > +
> >      /* State of our input IRQ/FIQ/VIRQ/VFIQ lines */
> >      uint32_t irq_line_state;
> >
> > diff --git a/target/arm/kvm.c b/target/arm/kvm.c
> > index 85860e6..c088589 100644
> > --- a/target/arm/kvm.c
> > +++ b/target/arm/kvm.c
> > @@ -39,6 +39,7 @@ const KVMCapabilityInfo kvm_arch_required_capabilities[] 
> > = {
> >
> >  static bool cap_has_mp_state;
> >  static bool cap_has_inject_serror_esr;
> > +static bool cap_has_inject_ext_dabt;
> >
> >  static ARMHostCPUFeatures arm_host_cpu_features;
> >
> > @@ -244,6 +245,16 @@ int kvm_arch_init(MachineState *ms, KVMState *s)
> >          ret = -EINVAL;
> >      }
> >
> > +    if (kvm_check_extension(s, KVM_CAP_ARM_NISV_TO_USER)) {
> > +        if (kvm_vm_enable_cap(s, KVM_CAP_ARM_NISV_TO_USER, 0)) {
> > +            warn_report("Failed to enable DABT NISV cap");
>
> Shouldn't this be an error? If KVM says it has KVM_CAP_ARM_NISV_TO_USER,
> then I think it should always work to enable it, unless userspace passes
> the wrong flags. Currently flags must be zero, but if they were to change
> then we'll need to add the flags to vmstate and fail migration when they
> aren't compatible, and I guess that failure would occur here.
>
That's a fair point. From the kernel point of view this one is pretty
straightforward,
so it should not fail. I haven't used the error here as the lack of
this cap is not really
critical for guest but indeed it might be worth to have it here .

> > +        } else {
> > +            /* Set status for supporting the external dabt injection */
> > +            cap_has_inject_ext_dabt = kvm_check_extension(s,
> > +                                    KVM_CAP_ARM_INJECT_EXT_DABT);
> > +        }
> > +    }
> > +
> >      return ret;
> >  }
> >
> > @@ -703,9 +714,16 @@ int kvm_put_vcpu_events(ARMCPU *cpu)
> >          events.exception.serror_esr = env->serror.esr;
> >      }
> >
> > +    if (cap_has_inject_ext_dabt) {
> > +        events.exception.ext_dabt_pending = env->ext_dabt_pending;
> > +    }
> > +
> >      ret = kvm_vcpu_ioctl(CPU(cpu), KVM_SET_VCPU_EVENTS, &events);
> >      if (ret) {
> >          error_report("failed to put vcpu events");
> > +    } else {
> > +        /* Clear instantly if the call was successful */
> > +        env->ext_dabt_pending = 0;
> >      }
> >
> >      return ret;
> > @@ -819,6 +837,11 @@ int kvm_arch_handle_exit(CPUState *cs, struct kvm_run 
> > *run)
> >              ret = EXCP_DEBUG;
> >          } /* otherwise return to guest */
> >          break;
> > +    case KVM_EXIT_ARM_NISV:
> > +        /* External DABT with no valid iss to decode */
> > +        ret = kvm_arm_handle_dabt_nisv(cs, run->arm_nisv.esr_iss,
> > +                                       run->arm_nisv.fault_ipa);
> > +        break;
> >      default:
> >          qemu_log_mask(LOG_UNIMP, "%s: un-handled exit reason %d\n",
> >                        __func__, run->exit_reason);
> > @@ -953,3 +976,34 @@ int kvm_arch_msi_data_to_gsi(uint32_t data)
> >  {
> >      return (data - 32) & 0xffff;
> >  }
> > +
> > +int kvm_arm_handle_dabt_nisv(CPUState *cs, uint64_t esr_iss,
> > +                             uint64_t fault_ipa)
> > +{
> > +    ARMCPU *cpu = ARM_CPU(cs);
> > +    CPUARMState *env = &cpu->env;
> > +
> > +   /*
> > +    * ISS [23:14] is invalid so there is a limited info
> > +    * on what has just happened so the only *useful* thing that can
> > +    * be retrieved from ISS is WnR & DFSC (though in some cases WnR
> > +    * might be less of a value as well)
> > +    */
> > +
> > +    /*
> > +     * Set pending ext dabt and trigger SET_EVENTS so that
> > +     * KVM can inject the abort
> > +     */
> > +    if (cap_has_inject_ext_dabt) {
> > +        kvm_cpu_synchronize_state(cs);
>
> I guess this is just an expensive 'vcpu_dirty=true', which the comment
> above justifies, but not super clearly. Can you try to clarify the
> comment some more?  I also wonder if we shouldn't add a KVM function
> that just marks a vcpu dirty, rather than fetching all registers.
>
I can definitely adjust the comments here to explain the case more
clearly. And I would definitely vote for moving the flag
setting/clearing outside the
sync so that it can be triggered separately without involving all the series
of potentially unnecessary ioctls. WIll draft that in the next iteration.

Thanks
Beata


> > +        env->ext_dabt_pending = 1;
> > +    } else {
> > +        error_report("Data abort exception triggered by guest memory 
> > access "
> > +                     "at physical address: 0x"  TARGET_FMT_lx,
> > +                     (target_ulong)fault_ipa);
> > +        error_printf("KVM unable to emulate faulting instruction.\n");
> > +        return -1;
> > +    }
> > +
> > +    return 0;
> > +}
> > diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h
> > index ae9e075..39472d5 100644
> > --- a/target/arm/kvm_arm.h
> > +++ b/target/arm/kvm_arm.h
> > @@ -450,6 +450,17 @@ struct kvm_guest_debug_arch;
> >  void kvm_arm_copy_hw_debug_data(struct kvm_guest_debug_arch *ptr);
> >
> >  /**
> > + * kvm_arm_handle_dabt_nisv:
> > + * @cs: CPUState
> > + * @esr_iss: ISS encoding (limited) for the exception from Data Abort
> > + *           ISV bit set to '0b0' -> no valid instruction syndrome
> > + * @fault_ipa: faulting address for the synch data abort
> > + *
> > + * Returns: 0 if the exception has been handled, < 0 otherwise
> > + */
> > +int kvm_arm_handle_dabt_nisv(CPUState *cs, uint64_t esr_iss,
> > +                             uint64_t fault_ipa);
> > +/**
> >   * its_class_name:
> >   *
> >   * Return the ITS class name to use depending on whether KVM acceleration
> > --
> > 2.7.4
> >
>
> Thanks,
> drew
>