Re: [PATCH] kvm: x86: Improve virtual machine startup performance

[Paolo Bonzini <pbonzini@redhat.com>]

On 3/8/22 03:03, Sean Christopherson wrote:
> On Thu, Mar 03, 2022, Hao Peng wrote:
>> On Thu, Mar 3, 2022 at 9:29 AM Sean Christopherson <seanjc@google.com> wrote:
>>>
>>> On Wed, Mar 02, 2022, Hao Peng wrote:
>>>> Thanks for pointing this out. However, other than shadow_root_level,
>>>> other fields of context will not
>>>> change during the entire operation, such as
>>>> page_fault/sync_page/direct_map and so on under
>>>> the condition of tdp_mmu.
>>>> Is this patch still viable after careful confirmation of the fields
>>>> that won't be modified?
>>>
>>> No, passing around the "init" flag is a hack.
>>>
>>> But, we can achieve what you want simply by initializing the constant data once
>>> per vCPU.  There's a _lot_ of state that is constant for a given MMU now that KVM
>>> uses separate MMUs for L1 vs. L2 when TDP is enabled.  I should get patches posted
>>> tomorrow, just need to test (famous last words).
> 
> Famous last words indeed.  Long story short, the patches were mostly easy, but I
> wandered deep into a rabbit hole when trying to make ->inject_page_fault() constant
> per MMU.  I'll get something posted this week, though exactly what that something is
> remains to be seen :-)

This is exactly what I have posted a few weeks ago:

https://patchew.org/linux/20220221162243.683208-1-pbonzini@redhat.com/

See in particular

   KVM: nVMX/nSVM: do not monkey-patch inject_page_fault callback
   KVM: x86/mmu: initialize constant-value fields just once

Paolo