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

Re: [Xen-devel] [PATCH V2 3/3] xen/vm_event: Deny register writes if refused by vm_event reply



On 06/26/2015 12:39 PM, Jan Beulich wrote:
>>>> @@ -1249,15 +1250,16 @@ static void load_vvmcs_host_state(struct vcpu *v)
>>>> >>>          __vmwrite(vmcs_h2g_field[i].guest_field, r);
>>>> >>>      }
>>>> >>>  
>>>> >>> -    hvm_set_cr0(__get_vvmcs(vvmcs, HOST_CR0));
>>>> >>> -    hvm_set_cr4(__get_vvmcs(vvmcs, HOST_CR4));
>>>> >>> -    hvm_set_cr3(__get_vvmcs(vvmcs, HOST_CR3));
>>>> >>> +    hvm_set_cr0(__get_vvmcs(vvmcs, HOST_CR0), 1);
>>>> >>> +    hvm_set_cr4(__get_vvmcs(vvmcs, HOST_CR4), 1);
>>>> >>> +    hvm_set_cr3(__get_vvmcs(vvmcs, HOST_CR3), 1);
>>>> >>>  
>>>> >>>      control = __get_vvmcs(vvmcs, VM_EXIT_CONTROLS);
>>>> >>>      if ( control & VM_EXIT_LOAD_HOST_PAT )
>>>> >>>          hvm_set_guest_pat(v, __get_vvmcs(vvmcs, HOST_PAT));
>>>> >>>      if ( control & VM_EXIT_LOAD_PERF_GLOBAL_CTRL )
>>>> >>> -        hvm_msr_write_intercept(MSR_CORE_PERF_GLOBAL_CTRL, 
>>>> >>> __get_vvmcs(vvmcs, HOST_PERF_GLOBAL_CTRL));
>>>> >>> +        hvm_msr_write_intercept(MSR_CORE_PERF_GLOBAL_CTRL,
>>>> >>> +                                __get_vvmcs(vvmcs, 
>>>> >>> HOST_PERF_GLOBAL_CTRL), 1);
>>> >> 
>>> >> Considering these together with the above - do you really want/
>>> >> need to intercept and send events for both host and guest shadow
>>> >> state changes?
>> > 
>> > Guest MSRs should suffice indeed. I was just trying to be consistent,
>> > but no, the host changes should not be necessary.
> Aren't you inverting things here? We're talking about the guest's host
> vs (L2) guest values here afaict (note the __get_vvmcs() uses).

I am indeed, clarification appreciated.


Thanks,
Razvan

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.