Re: [Xen-devel] [PATCH for-next v3 9/9] x86: introduce CONFIG_HYPERV and detection code

[Jan Beulich <jbeulich@suse.com>]

On 21.11.2019 17:27, Wei Liu wrote:
> On Fri, Nov 15, 2019 at 03:07:29PM +0100, Jan Beulich wrote:
>> On 21.10.2019 17:57, Wei Liu wrote:
>>> --- a/xen/arch/x86/Kconfig
>>> +++ b/xen/arch/x86/Kconfig
>>> @@ -164,6 +164,15 @@ endchoice
>>>  config GUEST
>>>  	bool
>>>  
>>> +config HYPERV_GUEST
>>> +	def_bool n
>>> +	select GUEST
>>> +	prompt "Hyper-V Guest"
>>
>> Please can you avoid following the bad example XEN_GUEST gives (and
>> perhaps even take the opportunity here or in the earlier patch
>> adding GUEST to change that one as well)? What you want is
>>
>> config HYPERV_GUEST
>> 	bool "Hyper-V Guest"
>> 	select GUEST
> 
> Ack.
> 
>>
>>> --- /dev/null
>>> +++ b/xen/arch/x86/guest/hyperv/hyperv.c
>>> @@ -0,0 +1,54 @@
>>> +/******************************************************************************
>>> + * arch/x86/guest/hyperv/hyperv.c
>>> + *
>>> + * Support for detecting and running under Hyper-V.
>>> + *
>>> + * This program is free software; you can redistribute it and/or modify
>>> + * it under the terms of the GNU General Public License as published by
>>> + * the Free Software Foundation; either version 2 of the License, or
>>> + * (at your option) any later version.
>>> + *
>>> + * This program is distributed in the hope that it will be useful,
>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>>> + * GNU General Public License for more details.
>>> + *
>>> + * You should have received a copy of the GNU General Public License
>>> + * along with this program; If not, see <http://www.gnu.org/licenses/>.
>>> + *
>>> + * Copyright (c) 2019 Microsoft.
>>> + */
>>> +#include <xen/init.h>
>>> +
>>> +#include <asm/guest.h>
>>> +
>>> +bool __init hyperv_probe(void)
>>> +{
>>> +    uint32_t eax, ebx, ecx, edx;
>>> +
>>> +    cpuid(0x40000000, &eax, &ebx, &ecx, &edx);
>>> +    if ( !((ebx == 0x7263694d) &&  /* "Micr" */
>>> +           (ecx == 0x666f736f) &&  /* "osof" */
>>> +           (edx == 0x76482074)) )  /* "t Hv" */
>>> +        return false;
>>> +
>>> +    cpuid(0x40000001, &eax, &ebx, &ecx, &edx);
>>> +    if ( eax != 0x31237648 )    /* Hv#1 */
>>> +        return false;
>>> +
>>> +    return true;
>>> +}
>>> +
>>> +struct hypervisor_ops hyperv_ops = {
>>
>> const again.
>>
>>> --- a/xen/arch/x86/guest/hypervisor.c
>>> +++ b/xen/arch/x86/guest/hypervisor.c
>>> @@ -43,6 +43,14 @@ bool hypervisor_probe(void)
>>>      }
>>>  #endif
>>>  
>>> +#ifdef CONFIG_HYPERV_GUEST
>>> +    if ( hyperv_probe() )
>>> +    {
>>> +        hops = &hyperv_ops;
>>> +        return true;
>>> +    }
>>> +#endif
>>
>> This recurring #ifdef CONFIG_*_GUEST is going to start looking ugly
>> the latest when one or two more get added. Perhaps better providing
>> *_probe() stubs returning false, and (like we do elsewhere) rely on
>> DCE to get rid of the *_ops reference? (And really you already have
>> such a stub - all you need to do is put the hyperv_ops declaration
>> outside the #ifdef (but read on).
>>
>> Also how about having *_probe() return the address of *_ops, such
>> that the latter could all become static?
> 
> Previously you made a suggestion to make probe return the name of the
> hypervisor. Here you ask for address of ops. I actually prefer the
> method suggested here, but this means I will need to keep
> hypervisor_name around.

Is there actually any user of the name field other than the caller
of probe? If not, surely that caller could access the name field
without a hypervisor_name() wrapper.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel