Re: [Xen-devel] [PATCH v5] arch/x86: Add registers to vm_event

[Tamas K Lengyel <tamas@xxxxxxxxxxxxx>]

On Tue, Oct 30, 2018 at 7:19 AM Jan Beulich <JBeulich@xxxxxxxx> wrote:
>
> >>> On 30.10.18 at 13:26, <rcojocaru@xxxxxxxxxxxxxxx> wrote:
> > On 10/30/18 1:33 PM, Jan Beulich wrote:
> >>>>> On 30.10.18 at 11:07, <aisaila@xxxxxxxxxxxxxxx> wrote:
> >>> --- a/xen/include/public/vm_event.h
> >>> +++ b/xen/include/public/vm_event.h
> >>> @@ -29,7 +29,7 @@
> >>>
> >>>  #include "xen.h"
> >>>
> >>> -#define VM_EVENT_INTERFACE_VERSION 0x00000003
> >>> +#define VM_EVENT_INTERFACE_VERSION 0x00000004
> >>>
> >>>  #if defined(__XEN__) || defined(__XEN_TOOLS__)
> >>>
> >>> @@ -157,6 +157,12 @@
> >>>  #define VM_EVENT_X86_CR4    2
> >>>  #define VM_EVENT_X86_XCR0   3
> >>>
> >>> +/* The limit field is right-shifted by 12 bits if .ar.g is set. */
> >>> +struct x86_selector_reg {
> >>
> >> I'm sorry for not having noticed this earlier, but this needs proper
> >> prefixing: Matching struct vm_event_regs_x86, it should at least
> >> be prefixed by vm_event_. Strictly speaking xen_ as the very
> >> first thing would also be required, but I'll leave that to the VM
> >> event maintainers to decide. With this
> >> Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
> >
> > I am fine with either approach (so leaving it as it is not a problem),
> > so we'll go with Tamas' preference. Tamas, what's your opinion?
>
> FAOD leaving as is is not an option - at least vm_event_ needs to
> be added, to not chance collision with a future addition in Xen itself.

I agree, we should keep things prefixed with vm_event_ in this header.

Tamas

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