Re: [Qemu-devel] [PATCH v3 5/5] ppc: spapr: Enable FWNMI capability

[Aravinda Prasad <aravinda@linux.vnet.ibm.com>]

On Thursday 17 August 2017 07:29 AM, David Gibson wrote:
> On Wed, Aug 16, 2017 at 02:42:48PM +0530, Aravinda Prasad wrote:
>> Enable the KVM capability KVM_CAP_PPC_FWNMI so that
>> the KVM causes guest exit with NMI as exit reason
>> when it encounters a machine check exception on the
>> address belonging to a guest. Without this capability
>> enabled, KVM redirects machine check exceptions to
>> guest's 0x200 vector.
>>
>> Signed-off-by: Aravinda Prasad <aravinda@linux.vnet.ibm.com>
>> ---
>>  hw/ppc/spapr_rtas.c |   17 +++++++++++++++++
>>  target/ppc/kvm.c    |    2 ++
>>  2 files changed, 19 insertions(+)
>>
>> diff --git a/hw/ppc/spapr_rtas.c b/hw/ppc/spapr_rtas.c
>> index 2f3c47b..c2a361a 100644
>> --- a/hw/ppc/spapr_rtas.c
>> +++ b/hw/ppc/spapr_rtas.c
>> @@ -47,6 +47,8 @@
>>  #include "trace.h"
>>  #include "hw/ppc/fdt.h"
>>  
>> +extern int cap_fwnmi;
> 
> Ew.  as the style checker will tell you, bare externs in .c files are
> frowned upon.  And the cap_* variables xertainly shouldn't be exported
> from kvm.c.

Noted.

> 
>>  static void rtas_display_character(PowerPCCPU *cpu, sPAPRMachineState *spapr,
>>                                     uint32_t token, uint32_t nargs,
>>                                     target_ulong args,
>> @@ -354,7 +356,22 @@ static void rtas_ibm_nmi_register(PowerPCCPU *cpu,
>>                                    target_ulong args,
>>                                    uint32_t nret, target_ulong rets)
>>  {
>> +    int ret;
>> +    CPUState *cs = CPU(cpu);
>> +
>> +    if (!cap_fwnmi) {
>> +        rtas_st(rets, 0, RTAS_OUT_NOT_SUPPORTED);
>> +        return;
>> +    }
>> +
>>      spapr->guest_machine_check_addr = rtas_ld(args, 1);
>> +
>> +    ret = kvm_vcpu_enable_cap(cs, KVM_CAP_PPC_FWNMI, 0);
>> +    if (ret < 0) {
>> +        rtas_st(rets, 0, RTAS_OUT_HW_ERROR);
>> +        return;
>> +    }
>> +
> 
> Congratulations, you broke build on an x86 host.  Things that directly
> call KVM need to go in kvm.c with suitable stubs for non-kvm builds.

Sure.

Regards,
Aravinda

> 
>>      rtas_st(rets, 0, RTAS_OUT_SUCCESS);
>>  }
>>  
>> diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c
>> index 73f64ed..ca1ffa6 100644
>> --- a/target/ppc/kvm.c
>> +++ b/target/ppc/kvm.c
>> @@ -90,6 +90,7 @@ static int cap_htm;             /* Hardware transactional memory support */
>>  static int cap_mmu_radix;
>>  static int cap_mmu_hash_v3;
>>  static int cap_resize_hpt;
>> +int cap_fwnmi;
>>  
>>  static uint32_t debug_inst_opcode;
>>  
>> @@ -147,6 +148,7 @@ int kvm_arch_init(MachineState *ms, KVMState *s)
>>      cap_mmu_radix = kvm_vm_check_extension(s, KVM_CAP_PPC_MMU_RADIX);
>>      cap_mmu_hash_v3 = kvm_vm_check_extension(s, KVM_CAP_PPC_MMU_HASH_V3);
>>      cap_resize_hpt = kvm_vm_check_extension(s, KVM_CAP_SPAPR_RESIZE_HPT);
>> +    cap_fwnmi = kvm_check_extension(s, KVM_CAP_PPC_FWNMI);
>>  
>>      if (!cap_interrupt_level) {
>>          fprintf(stderr, "KVM: Couldn't find level irq capability. Expect the "
>>
> 

-- 
Regards,
Aravinda