Re: [PATCH 5/5] x86/PV32: avoid TLB flushing after mod_l3_entry()

[Jan Beulich <jbeulich@xxxxxxxx>]

On 11.01.2021 15:23, Roger Pau Monné wrote:
> On Tue, Nov 03, 2020 at 11:58:16AM +0100, Jan Beulich wrote:
>> 32-bit guests may not depend upon the side effect of using ordinary
>> 4-level paging when running on a 64-bit hypervisor. For L3 entry updates
>> to take effect, they have to use a CR3 reload. Therefore there's no need
>> to issue a paging structure invalidating TLB flush in this case.
> 
> I assume it's fine for the Xen linear page tables to be lkely out of
> sync during the windows between the entry update and the CR3 reload?

Yes, because ...

> I wonder, won't something similar also apply to 64bit and L4 entries?

... unlike 64-bit paging, PAE paging special cases the treatment
of the 4 top level table entries. On bare metal they get loaded
by the CPU upon CR3 load, not when walking page tables.

Jan