linux-kernel - Re: [PATCH 07/23] KVM: MMU: remove kvm_mmu_calc_root_page

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 10 Feb 2022 18:43:00 +0100
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Sean Christopherson <seanjc@...gle.com>
Cc:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
        dmatlack@...gle.com, vkuznets@...hat.com
Subject: Re: [PATCH 07/23] KVM: MMU: remove kvm_mmu_calc_root_page_role

On 2/10/22 18:29, Sean Christopherson wrote:
> On Thu, Feb 10, 2022, Paolo Bonzini wrote:
>> On 2/10/22 01:47, Sean Christopherson wrote:
>>> The nested mess is likely easily solved, I don't see any obvious issue with swapping
>>> the order.  But I still don't love the subtlety.  I do like shaving cycles, just
>>> not the subtlety...
>>
>> Not so easily, but it's doable and it's essentially what I did in the other
>> series (the one that reworks the root cache).
> 
> Sounds like I should reveiw that series first?

Yeah, this one is still a nice step in the direction of guest pt walk 
from shadow pt build(*), but with no immediate use for TDP MMU root reuse.

The original idea was to use the MMU role to decide whether to do 
kvm_mmu_unload(), but that would have still been a bandaid---inefficient 
and a bad idea overall.  Patches 6+7 of this series (once fixed, because 
they were buggy as hell) turned out to be enough to tame the PGD cache 
and remove kvm_mmu_unload() altogether from kvm_mmu_reset_context().

Paolo

(*) Your idea of detecting stale roots is quite easily implemented on 
top of these, for example, because a root is stale if and only if the 
root_role changes.