On Fri, Jul 29, 2022 at 03:14:02PM +0800, Xiaoyao Li wrote: > On 5/30/2022 7:49 PM, Gerd Hoffmann wrote: > > On Thu, May 26, 2022 at 10:48:56AM +0800, Xiaoyao Li wrote: > > > On 5/24/2022 3:08 PM, Gerd Hoffmann wrote: > > > > On Thu, May 12, 2022 at 11:17:45AM +0800, Xiaoyao Li wrote: > > > > > TDX guest cannot go to real mode, so just skip the setup of isa-bios. > > > > > > > > Does isa-bios setup cause any actual problems? > > > > (same question for patch #19). > > There is no need for copying, end_of_1M is a alias memory region for > > end_of_4G, so the backing storage is the same. > > It is a reason that current alias approach cannot work for TDX. Because in > TDX a private page can be only mapped to one gpa. Ok, so memory aliasing not being supported by TDX is the underlying reason. > So for simplicity, I will > just skip isa-bios shadowing for TDX instead of implementing a non-alias + > memcpy approach. Makes sense given that tdx wouldn't use the mapping below 1M anyway. A comment explaining the tdx aliasing restriction would be good to make clear why the special case for tdx exists. take care, Gerd
