On Wed, 10 Jan 2024 18:00:40 -0800, Sean Christopherson wrote:
> This series is the result of digging into why deleting a memslot, which on
> x86 forces all vCPUs to reload a new MMU root, causes noticeably more jitter
> in vCPUs and other tasks when running with the TDP MMU than the Shadow MMU
> (with TDP enabled).
>
> Patch 1 addresses the most obvious issue by simply zapping at a finer
> granularity so that if a different task, e.g. a vCPU, wants to run on the
> pCPU doing the zapping, it doesn't have to wait for KVM to zap an entire
> 1GiB region, which can take a hundreds of microseconds (or more). The
> shadow MMU checks for need_resched() (and mmu_lock contention, see below)
> every 10 zaps, which is why the shadow MMU doesn't induce the same level
> of jitter.
>
> [...]
Applied to kvm-x86 mmu, thanks!
[1/8] KVM: x86/mmu: Zap invalidated TDP MMU roots at 4KiB granularity
https://github.com/kvm-x86/linux/commit/8ca983631f3c
[2/8] KVM: x86/mmu: Don't do TLB flush when zappings SPTEs in invalid roots
https://github.com/kvm-x86/linux/commit/fcdffe97f80e
[3/8] KVM: x86/mmu: Allow passing '-1' for "all" as_id for TDP MMU iterators
https://github.com/kvm-x86/linux/commit/6577f1efdff4
[4/8] KVM: x86/mmu: Skip invalid roots when zapping leaf SPTEs for GFN range
https://github.com/kvm-x86/linux/commit/99b85fda91b1
[5/8] KVM: x86/mmu: Skip invalid TDP MMU roots when write-protecting SPTEs
https://github.com/kvm-x86/linux/commit/d746182337c2
[6/8] KVM: x86/mmu: Check for usable TDP MMU root while holding mmu_lock for read
https://github.com/kvm-x86/linux/commit/f5238c2a60f1
[7/8] KVM: x86/mmu: Alloc TDP MMU roots while holding mmu_lock for read
https://github.com/kvm-x86/linux/commit/dab285e4ec73
[8/8] KVM: x86/mmu: Free TDP MMU roots while holding mmy_lock for read
https://github.com/kvm-x86/linux/commit/576a15de8d29
--
https://github.com/kvm-x86/linux/tree/next
