On Fri, Feb 10, 2023 at 05:46:21PM -0800, Vipin Sharma wrote: > Do atomic-AND to clear the dirty state of SPTEs. Optimize clear-dirty-log > flow by avoiding to go through __handle_changed_spte() and directly call > kvm_set_pfn_dirty() instead. > > Atomic-AND allows to fetch the latest value in SPTE, clear only its > dirty state and set the new SPTE value. This optimization avoids > executing unnecessary checks by not calling __handle_changed_spte(). > > With the removal of tdp_mmu_set_spte_no_dirty_log(), "record_dirty_log" > parameter in __tdp_mmu_set_spte() is now obsolete. It will always be set > to true by its caller. This dead code will be cleaned up in future > commits. > > Tested on a VM (160 vCPUs, 160 GB memory) and found that performance of clear > dirty log stage improved by ~40% in dirty_log_perf_test > > Before optimization: > -------------------- > Iteration 1 clear dirty log time: 3.638543593s > Iteration 2 clear dirty log time: 3.145032742s > Iteration 3 clear dirty log time: 3.142340358s > Clear dirty log over 3 iterations took 9.925916693s. (Avg 3.308638897s/iteration) > > After optimization: > ------------------- > Iteration 1 clear dirty log time: 2.318988110s > Iteration 2 clear dirty log time: 1.794470164s > Iteration 3 clear dirty log time: 1.791668628s > Clear dirty log over 3 iterations took 5.905126902s. (Avg 1.968375634s/iteration) > > Signed-off-by: Vipin Sharma <vipinsh@xxxxxxxxxx> Reviewed-by: David Matlack <dmatlack@xxxxxxxxxx>
