This patchset is to verify if dirty logging works properly by checking page stats from the per-VM interface. We discovered one performance bug in disallowed_hugepage_adjust() which prevents KVM from recovering large pages for the guest. The selftest logic added later could help validate the problem. The patchset borrowes two patches come from Ben's series: "[PATCH 00/13] KVM: x86: Add a cap to disable NX hugepages on a VM" [1], which completes the selftest library functions to use the stats interface. [1] https://lore.kernel.org/all/20220310164532.1821490-2-bgardon@xxxxxxxxxx/T/ v2 -> v1: - Update the commit message. [dmatlack] - Update the comments in patch 3/4 to clarify the motivation. [bgardon] - Add another iteration in dirty_log_perf_test to regain pages [bgardon] v1: https://lore.kernel.org/all/20220321002638.379672-1-mizhang@xxxxxxxxxx/ Ben Gardon (2): selftests: KVM: Dump VM stats in binary stats test selftests: KVM: Test reading a single stat Mingwei Zhang (2): KVM: x86/mmu: explicitly check nx_hugepage in disallowed_hugepage_adjust() selftests: KVM: use dirty logging to check if page stats work correctly arch/x86/kvm/mmu/mmu.c | 14 +- .../selftests/kvm/dirty_log_perf_test.c | 53 +++++ .../selftests/kvm/include/kvm_util_base.h | 2 + .../selftests/kvm/kvm_binary_stats_test.c | 6 + tools/testing/selftests/kvm/lib/kvm_util.c | 196 ++++++++++++++++++ 5 files changed, 269 insertions(+), 2 deletions(-) -- 2.35.1.1021.g381101b075-goog
