mbox series

[v2,0/5] Optimize clear dirty log

Message ID 20230203192822.106773-1-vipinsh@google.com
Headers
Series Optimize clear dirty log |

Message

Vipin Sharma Feb. 3, 2023, 7:28 p.m. UTC 
  Hi,

This patch series has optimized control flow of clearing dirty log and
improved its performance by ~38%. Patch 2 has more details about
optimization.

It also got rid of many variants of the handle_changed_spte family of
functions and converged logic to one handle_changed_spte() function. It
also remove tdp_mmu_set_spte_no_[acc_track|dirty_log] and various
booleans for controlling them.

Thanks
Vipin

v2:
- Clear dirty log and age gfn range does not go through
  handle_changed_spte, they handle their SPTE changes locally to improve
  their speed.
- Clear only specific bits atomically when updating SPTEs in clearing
  dirty log and aging gfn range functions.
- Removed tdp_mmu_set_spte_no_[acc_track|dirty_log] APIs.
- Converged all handle_changed_spte related functions to one place.

v1: https://lore.kernel.org/lkml/20230125213857.824959-1-vipinsh@google.com/

Vipin Sharma (5):
  KVM: x86/mmu: Make separate function to check for SPTEs atomic write
    conditions
  KVM: x86/mmu: Optimize SPTE change flow for clear-dirty-log
  KVM: x86/mmu: Optimize SPTE change for aging gfn range
  KVM: x86/mmu: Remove handle_changed_spte_dirty_log()
  KVM: x86/mmu: Merge all handle_changed_pte* functions.

 arch/x86/kvm/mmu/tdp_iter.h |  29 +++++--
 arch/x86/kvm/mmu/tdp_mmu.c  | 163 +++++++++++-------------------------
 2 files changed, 72 insertions(+), 120 deletions(-)