Re: [Xen-devel] [PATCH v1 2/6] vmx: add raw_vmx_msr_policy

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 06.07.17 at 11:29, <sergey.dyasli@xxxxxxxxxx> wrote:
> On Tue, 2017-07-04 at 08:15 -0600, Jan Beulich wrote:
>> > > > On 26.06.17 at 12:44, <sergey.dyasli@xxxxxxxxxx> wrote:
>> > 
>> > @@ -611,6 +624,9 @@ int vmx_cpu_up(void)
>> >  
>> >      BUG_ON(!(read_cr4() & X86_CR4_VMXE));
>> >  
>> > +    if ( (rc = calculate_raw_policy(false)) != 0 )
>> > +        return rc;
>> > +
>> >      /* 
>> >       * Ensure the current processor operating mode meets 
>> >       * the requred CRO fixed bits in VMX operation. 
>> 
>> Btw., is it intentional that the function is being invoked for the BSP a
>> second time here (after start_vmx() did so already), with the flag
>> now being passed with the wrong value?
> 
> Unfortunately, I couldn't find a better way of detecting if the code is 
> running
> on the boot CPU. And I decided to use the existing practice of passing a flag.

This passing of a flag is fine; what I'm uncomfortable with is that the
second invocation on the BSP will say it's not on the BSP. While this
looks to be benign, it would feel better if there wasn't such a second
invocation at all.

Jan


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