Re: [PATCH 1/5] i386: move hyperv_vendor_id initialization to x86_cpu_realizefn()

[Vitaly Kuznetsov]

Claudio Fontana <cfontana@suse.de> writes:

> On 3/10/21 12:43 PM, Vitaly Kuznetsov wrote:
>> Claudio Fontana <cfontana@suse.de> writes:
>> 
>>> On 11/19/20 11:32 AM, Vitaly Kuznetsov wrote:
>>>> As a preparation to expanding Hyper-V CPU features early, move
>>>> hyperv_vendor_id initialization to x86_cpu_realizefn(). Introduce
>>>> x86_cpu_hyperv_realize() to not not pollute x86_cpu_realizefn()
>>>> itself.
>>>>
>>>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>>>
>>>
>>> Hi Vitaly,
>>>
>>> If building for TCG-only, does the x86_cpu_hyperv_realize function and 
>>> other hyperv related functions
>>> make sense to keep in the build?
>>>
>>> Now that we have per-accelerator subdirs in target/i386, should the hyperv 
>>> code be moved over?
>>>
>> 
>> Hi Claudio,
>> 
>> I'm not exactly sure. On one hand, we only implement Hyper-V emulation
>> with KVM now.
>
>
> Hi Vitaly,
>
>
>> On the other hand Hyper-V specific CPU options are
>> available even without it (and as Igor feels strongly against
>> introducing custom option parsers, I don't see how we can forbid to use
>> them without KVM). x86_cpu_hyperv_realize() is the bare minimum which is
>> only needed to set our internal Hyper-V related data in a sane way,
>> e.g. set the default to 'cpu->hyperv_vendor_id'. The actual enablement
>> code is in target/i386/kvm.c already. Do you see anything besides
>> x86_cpu_hyperv_realize() which we could move there?
>
> Currently I don't,
> I see a lot of PROPs (hyperv_features), but I assume those are the ones you 
> mention cannot be moved right?
>

Right, not without introducing a custom parser checking if KVM is in use.

>
>> Could you please
>> take a look at v5
>> (https://mail.gnu.org/archive/html/qemu-devel/2021-03/msg00158.html)?
>
>
> I took a look at the series you pointed me to, and it seems it is all pretty 
> much kvm/ work to me.
>
> Which hypervisors do you think would theoretically make sense when looking at 
> hyperv emulation in the future?
>

In theory, I don't see why we can't emulate hyper-v in TCG, I'm just not
sure how practical that would be). There was a Windows Hypervisor
Platform accelerator (whpx) which uses genuine Hyper-V and can use
Hyper-V enlightenment option to set guest visible CPUIDs but honestly I
don't know much about whpx.

> Looking at the current code and your patches, in the series you link to,
> I'd be tempted to suggest moving the hyperv realize function inside
> kvm_cpu_realizefn in kvm/kvm-cpu.c .

No problem, I wouldn't be against that. I'm waiting for some feedback on
v5 and if I am to do v6 I'll include that. In case I got luck and v5 is
to get merged, I can send a follow-up patch.

>
> Ciao,
>
> Claudio
>
>
>
>> 
>> Thanks!
>> 
>> 
>>> Thanks,
>>>
>>> Claudio
>>>
>>>
>>>> ---
>>>>  target/i386/cpu.c | 23 ++++++++++++++++++++++-
>>>>  target/i386/cpu.h |  3 ++-
>>>>  target/i386/kvm.c | 25 ++++++++++---------------
>>>>  3 files changed, 34 insertions(+), 17 deletions(-)
>>>>
>>>> diff --git a/target/i386/cpu.c b/target/i386/cpu.c
>>>> index 5a8c96072e41..2a6885753378 100644
>>>> --- a/target/i386/cpu.c
>>>> +++ b/target/i386/cpu.c
>>>> @@ -6509,6 +6509,25 @@ static void x86_cpu_filter_features(X86CPU *cpu, 
>>>> bool verbose)
>>>>      }
>>>>  }
>>>>  
>>>> +static void x86_cpu_hyperv_realize(X86CPU *cpu)
>>>> +{
>>>> +    size_t len;
>>>> +
>>>> +    /* Hyper-V vendor id */
>>>> +    if (!cpu->hyperv_vendor) {
>>>> +        memcpy(cpu->hyperv_vendor_id, "Microsoft Hv", 12);
>>>> +    } else {
>>>> +        len = strlen(cpu->hyperv_vendor);
>>>> +
>>>> +        if (len > 12) {
>>>> +            warn_report("hv-vendor-id truncated to 12 characters");
>>>> +            len = 12;
>>>> +        }
>>>> +        memset(cpu->hyperv_vendor_id, 0, 12);
>>>> +        memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
>>>> +    }
>>>> +}
>>>> +
>>>>  static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
>>>>  {
>>>>      CPUState *cs = CPU(dev);
>>>> @@ -6680,6 +6699,8 @@ static void x86_cpu_realizefn(DeviceState *dev, 
>>>> Error **errp)
>>>>          env->cache_info_amd.l3_cache = &legacy_l3_cache;
>>>>      }
>>>>  
>>>> +    /* Process Hyper-V enlightenments */
>>>> +    x86_cpu_hyperv_realize(cpu);
>>>>  
>>>>      cpu_exec_realizefn(cs, &local_err);
>>>>      if (local_err != NULL) {
>>>> @@ -7198,7 +7219,7 @@ static Property x86_cpu_properties[] = {
>>>>      DEFINE_PROP_UINT32("min-xlevel2", X86CPU, env.cpuid_min_xlevel2, 0),
>>>>      DEFINE_PROP_UINT64("ucode-rev", X86CPU, ucode_rev, 0),
>>>>      DEFINE_PROP_BOOL("full-cpuid-auto-level", X86CPU, 
>>>> full_cpuid_auto_level, true),
>>>> -    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor_id),
>>>> +    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor),
>>>>      DEFINE_PROP_BOOL("cpuid-0xb", X86CPU, enable_cpuid_0xb, true),
>>>>      DEFINE_PROP_BOOL("lmce", X86CPU, enable_lmce, false),
>>>>      DEFINE_PROP_BOOL("l3-cache", X86CPU, enable_l3_cache, true),
>>>> diff --git a/target/i386/cpu.h b/target/i386/cpu.h
>>>> index 88e8586f8fb4..be640bf45c29 100644
>>>> --- a/target/i386/cpu.h
>>>> +++ b/target/i386/cpu.h
>>>> @@ -1655,11 +1655,12 @@ struct X86CPU {
>>>>      uint64_t ucode_rev;
>>>>  
>>>>      uint32_t hyperv_spinlock_attempts;
>>>> -    char *hyperv_vendor_id;
>>>> +    char *hyperv_vendor;
>>>>      bool hyperv_synic_kvm_only;
>>>>      uint64_t hyperv_features;
>>>>      bool hyperv_passthrough;
>>>>      OnOffAuto hyperv_no_nonarch_cs;
>>>> +    uint32_t hyperv_vendor_id[3];
>>>>  
>>>>      bool check_cpuid;
>>>>      bool enforce_cpuid;
>>>> diff --git a/target/i386/kvm.c b/target/i386/kvm.c
>>>> index a2934dda027c..788a2cf2ec51 100644
>>>> --- a/target/i386/kvm.c
>>>> +++ b/target/i386/kvm.c
>>>> @@ -1205,6 +1205,13 @@ static int hyperv_handle_properties(CPUState *cs,
>>>>          memcpy(cpuid_ent, &cpuid->entries[0],
>>>>                 cpuid->nent * sizeof(cpuid->entries[0]));
>>>>  
>>>> +        c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
>>>> +        if (c) {
>>>> +            cpu->hyperv_vendor_id[0] = c->ebx;
>>>> +            cpu->hyperv_vendor_id[1] = c->ecx;
>>>> +            cpu->hyperv_vendor_id[2] = c->edx;
>>>> +        }
>>>> +
>>>>          c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
>>>>          if (c) {
>>>>              env->features[FEAT_HYPERV_EAX] = c->eax;
>>>> @@ -1279,23 +1286,11 @@ static int hyperv_handle_properties(CPUState *cs,
>>>>  
>>>>      c = &cpuid_ent[cpuid_i++];
>>>>      c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
>>>> -    if (!cpu->hyperv_vendor_id) {
>>>> -        memcpy(signature, "Microsoft Hv", 12);
>>>> -    } else {
>>>> -        size_t len = strlen(cpu->hyperv_vendor_id);
>>>> -
>>>> -        if (len > 12) {
>>>> -            error_report("hv-vendor-id truncated to 12 characters");
>>>> -            len = 12;
>>>> -        }
>>>> -        memset(signature, 0, 12);
>>>> -        memcpy(signature, cpu->hyperv_vendor_id, len);
>>>> -    }
>>>>      c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
>>>>          HV_CPUID_NESTED_FEATURES : HV_CPUID_IMPLEMENT_LIMITS;
>>>> -    c->ebx = signature[0];
>>>> -    c->ecx = signature[1];
>>>> -    c->edx = signature[2];
>>>> +    c->ebx = cpu->hyperv_vendor_id[0];
>>>> +    c->ecx = cpu->hyperv_vendor_id[1];
>>>> +    c->edx = cpu->hyperv_vendor_id[2];
>>>>  
>>>>      c = &cpuid_ent[cpuid_i++];
>>>>      c->function = HV_CPUID_INTERFACE;
>>>>
>>>
>> 
>

-- 
Vitaly