On 07/10/2015 04:35 AM, Lengyel, Tamas wrote:
>
> @@ -546,6 +652,23 @@ int main(int argc, char *argv[])
> }
>
> break;
> + case VM_EVENT_REASON_SINGLESTEP:
> + printf("Singlestep: rip=%016"PRIx64", vcpu %d\n",
> + req.regs.x86.rip,
> + req.vcpu_id);
> +
> + if ( altp2m )
> + {
> + printf("\tSwitching altp2m to view %u!\n",
> altp2m_view_id);
> +
> + rsp.reason = VM_EVENT_REASON_MEM_ACCESS;
>
>
> So this was a workaround for v3 of the series that is no longer
> necessary - it's probably cleaner to have the same reason set for the
> response as the request was. It's not against any rule, so the code is
> still correct and works, it's just not best practice. So in case there
> is another round on the series, it could be fixed then.
With or without that change (but preferably with it):
Reviewed-by: Razvan Cojocaru <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xen.org/xen-devel
