Messages in this thread | | | Subject | Re: [RFC v2-fix-v5 1/1] x86: Skip WBINVD instruction for VM guest | From | Dave Hansen <> | Date | Wed, 9 Jun 2021 16:55:39 -0700 |
| |
On 6/9/21 2:42 PM, Kuppuswamy, Sathyanarayanan wrote: > On 6/9/21 2:38 PM, Dan Williams wrote: >>> In TDX guests, these WBINVD operations cause #VE exceptions. For debug, >>> it would be ideal for the #VE handler to be able to WARN() when an >>> unexpected WBINVD occurs. (<--- problem #2) >> ...but it doesn't WARN() it triggers unhandled #VE, unless I missed >> another patch that precedes this that turns it into a WARN()? If a >> code path expects WBINVD for correct operation and the guest can't >> execute that sounds fatal, not a WARN to me. > > Yes. It is not WARN. It is a fatal unhandled exception.
That makes the problem statement a wee bit different, but it should still be pretty easy to explain:
In TDX guests, these WBINVD operations cause #VE exceptions. While some #VE exceptions can be handled, there is no recourse for a TDX guest to handle a WBINVD and it will panic(). (<--- problem #2)
| |