Re: [Xen-devel] [PATCH 1/3] vvmx: set vmxon_region_pa of vcpu out of VMX operation to an invalid address

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 14.12.16 at 02:37, <haozhong.zhang@xxxxxxxxx> wrote:
> On 12/13/16 08:21 -0700, Jan Beulich wrote:
>>>>> On 13.12.16 at 13:16, <haozhong.zhang@xxxxxxxxx> wrote:
>>> --- a/xen/arch/x86/hvm/vmx/vvmx.c
>>> +++ b/xen/arch/x86/hvm/vmx/vvmx.c
>>> @@ -32,6 +32,18 @@ static DEFINE_PER_CPU(u64 *, vvmcs_buf);
>>>
>>>  static void nvmx_purge_vvmcs(struct vcpu *v);
>>>
>>> +/*
>>> + * When a vcpu is out of VMXON region, set its vmxon_region_pa to
>>> + * INVALID_VMXON_REGION_PA. We cannot use 0, because 0 is also a valid
>>> + * VMXON region address.
>>> + */
>>> +#define INVALID_VMXON_REGION_PA (~0UL)
>>
>>And btw, having looked at patch 2 - any reason you can't simply
>>re-use VMCX_EADDR here?
>>
> 
> I just find INVALID_PADDR defined along with type paddr_t:
> 
> typedef unsigned long paddr_t;
> #define INVALID_PADDR (~0UL)
> 
> Which one, INVALID_PADDR or VMCX_EADDR, would be better?

The former then, I would say. And perhaps ditch VMCX_EADDR in
a separate patch then too, in favor of the more general one.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel