qemu-s390x
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2 08/13] s390x: protvirt: Add new VCPU reset functions


From: Janosch Frank
Subject: Re: [PATCH v2 08/13] s390x: protvirt: Add new VCPU reset functions
Date: Fri, 29 Nov 2019 12:21:20 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.1.1

On 11/29/19 11:47 AM, David Hildenbrand wrote:
> On 29.11.19 10:48, Janosch Frank wrote:
>> CPU resets for protected guests need to be done via Ultravisor
>> calls. Hence we need a way to issue these calls for each reset.
>>
>> As we formerly had only one reset function and it was called for
>> initial, as well as for the clear reset, we now need a new interface.
>>
>> Signed-off-by: Janosch Frank <address@hidden>
>> ---
>>  target/s390x/cpu.c       | 14 ++++++++++++--
>>  target/s390x/kvm-stub.c  | 10 +++++++++-
>>  target/s390x/kvm.c       | 38 ++++++++++++++++++++++++++++++++------
>>  target/s390x/kvm_s390x.h |  4 +++-
>>  4 files changed, 56 insertions(+), 10 deletions(-)
>>
>> diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
>> index bd39cb54b7..52fefa1586 100644
>> --- a/target/s390x/cpu.c
>> +++ b/target/s390x/cpu.c
>> @@ -131,8 +131,18 @@ static void s390_cpu_reset(CPUState *s, cpu_reset_type 
>> type)
>>      }
>>  
>>      /* Reset state inside the kernel that we cannot access yet from QEMU. */
>> -    if (kvm_enabled() && type != S390_CPU_RESET_NORMAL) {
>> -        kvm_s390_reset_vcpu(cpu);
>> +    if (kvm_enabled()) {
>> +        switch (type) {
>> +        case S390_CPU_RESET_CLEAR:
>> +            kvm_s390_reset_vcpu_clear(cpu);
>> +            break;
>> +        case S390_CPU_RESET_INITIAL:
>> +            kvm_s390_reset_vcpu_initial(cpu);
>> +            break;
>> +        case S390_CPU_RESET_NORMAL:
>> +            kvm_s390_reset_vcpu_normal(cpu);
>> +            break;
>> +        }
> 
> I would have assumed you only have to do that for pv? For ordinary
> guests we can avoid unnecessary ioctls IMHO.

Remember the reset normal IRQ fix in KVM?
Without it we're not architecture compliant.

> 
>>      }
>>  }
>>  
>> diff --git a/target/s390x/kvm-stub.c b/target/s390x/kvm-stub.c
>> index 5152e2bdf1..c4cd497f85 100644
>> --- a/target/s390x/kvm-stub.c
>> +++ b/target/s390x/kvm-stub.c
>> @@ -83,7 +83,15 @@ void kvm_s390_cmma_reset(void)
>>  {
>>  }
>>  
>> -void kvm_s390_reset_vcpu(S390CPU *cpu)
>> +void kvm_s390_reset_vcpu_initial(S390CPU *cpu)
>> +{
>> +}
>> +
>> +void kvm_s390_reset_vcpu_clear(S390CPU *cpu)
>> +{
>> +}
>> +
>> +void kvm_s390_reset_vcpu_normal(S390CPU *cpu)
>>  {
>>  }
>>  
>> diff --git a/target/s390x/kvm.c b/target/s390x/kvm.c
>> index b802d02ff5..5b1ed3acb4 100644
>> --- a/target/s390x/kvm.c
>> +++ b/target/s390x/kvm.c
>> @@ -154,6 +154,7 @@ static int cap_ri;
>>  static int cap_gs;
>>  static int cap_hpage_1m;
>>  static int cap_protvirt;
>> +static int cap_vcpu_resets;
>>  
>>  static int active_cmma;
>>  
>> @@ -346,6 +347,7 @@ int kvm_arch_init(MachineState *ms, KVMState *s)
>>      cap_mem_op = kvm_check_extension(s, KVM_CAP_S390_MEM_OP);
>>      cap_s390_irq = kvm_check_extension(s, KVM_CAP_S390_INJECT_IRQ);
>>      cap_protvirt = kvm_check_extension(s, KVM_CAP_S390_PROTECTED);
>> +    cap_vcpu_resets = kvm_check_extension(s, KVM_CAP_S390_VCPU_RESETS);
>>  
>>      if (!kvm_check_extension(s, KVM_CAP_S390_GMAP)
>>          || !kvm_check_extension(s, KVM_CAP_S390_COW)) {
>> @@ -407,20 +409,44 @@ int kvm_arch_destroy_vcpu(CPUState *cs)
>>      return 0;
>>  }
>>  
>> -void kvm_s390_reset_vcpu(S390CPU *cpu)
>> +static void kvm_s390_reset_vcpu(S390CPU *cpu, unsigned long type)
>>  {
>>      CPUState *cs = CPU(cpu);
>>  
>> -    /* The initial reset call is needed here to reset in-kernel
>> -     * vcpu data that we can't access directly from QEMU
>> -     * (i.e. with older kernels which don't support sync_regs/ONE_REG).
>> -     * Before this ioctl cpu_synchronize_state() is called in common kvm
>> -     * code (kvm-all) */
>> +    /*
>> +     * The reset call is needed here to reset in-kernel vcpu data that
>> +     * we can't access directly from QEMU (i.e. with older kernels
>> +     * which don't support sync_regs/ONE_REG).  Before this ioctl
>> +     * cpu_synchronize_state() is called in common kvm code
>> +     * (kvm-all).
>> +     */
>> +    if (cap_vcpu_resets) {
>> +        if (kvm_vcpu_ioctl(cs, KVM_S390_VCPU_RESET, type)) {
>> +            error_report("CPU reset type %ld failed on CPU %i",
>> +                         type, cs->cpu_index);
>> +        }
>> +        return;> +    }
> 
> This is broken for S390_CPU_RESET_NORMAL where we don't do a
> KVM_S390_INITIAL_RESET for !pv ...
> 
> Can't we limit that new handling to pv only and keep it out of this path?

Look above

> 
> 
> [...]
> 
> 


Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]