Re: [Xen-devel] [PATCH RFC v12 15/21] pvh: Set up more PV stuff in set_info_guest

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 20.09.13 at 17:12, George Dunlap <george.dunlap@xxxxxxxxxxxxx> wrote:
> On 20/09/13 15:58, Jan Beulich wrote:
>>>>> On 20.09.13 at 16:50, George Dunlap <george.dunlap@xxxxxxxxxxxxx> wrote:
>>> On 18/09/13 16:17, Jan Beulich wrote:
>>>>>>> On 13.09.13 at 18:25, George Dunlap <george.dunlap@xxxxxxxxxxxxx> wrote:
>>>>> +int vmx_pvh_vcpu_boot_set_info(struct vcpu *v,
>>>>> +                               struct vcpu_guest_context *ctxtp)
>>>>> +{
>>>>> +    if ( ctxtp->ldt_base || ctxtp->ldt_ents ||
>>>>> +         ctxtp->user_regs.cs || ctxtp->user_regs.ss || 
>>>>> ctxtp->user_regs.es ||
>>>>> +         ctxtp->user_regs.ds || ctxtp->user_regs.fs || 
>>>>> ctxtp->user_regs.gs ||
>>>>> +         *ctxtp->gdt_frames || ctxtp->gdt_ents ||
>>>> Don't know why I didn't spot this earlier, but the gdt_frames check
>>>> is pointless when gdt_ents is verified to be zero.
>>> You know, looking at this again -- is there a reason we can't just put
>>> this in hvm_set_info_guest()?  It's already only called from that one
>>> place, only a few lines before.  There doesn't really seem to be a need
>>> to have yet another function for just a few lines.
>> Indeed - only the VMX specific piece belongs here; everything
>> else should be in the generic wrapper (which perhaps then
>> won't be a simple wrapper anymore).
> 
> Or, we could do like is done for pv, and put the validity checks at the 
> top of arch_set_info_guest().  I think I like that option the best.

True.

Jan


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