On 13/03/20 00:27, Sean Christopherson wrote: > Start chipping away at the crustiness in the nVMX tests by refactoring > "struct vmentry_failure" into "struct vmentry_result", with the full > VM-Exit stored in vmentry_result. Capturing the exit reason allows for a > variety of cleanups and consolidations. > > This series really only dives into the v1 tests. I'd like to also clean > up the v2 tests, e.g. take the expected exit reason in enter_guest() so > that the expected behavior is more obvious, but that's a more invasive > cleanup for another day. > > Sean Christopherson (8): > nVMX: Eliminate superfluous entry_failure_handler() wrapper > nVMX: Refactor VM-Entry "failure" struct into "result" > nVMX: Consolidate non-canonical code in test_canonical() > nVMX: Drop redundant check for guest termination > nVMX: Expose __enter_guest() and consolidate guest state test code > nVMX: Pass exit reason union to v1 exit handlers > nVMX: Pass exit reason union to is_hypercall() > nVMX: Pass exit reason enum to print_vmexit_info() > > x86/vmx.c | 191 +++++++++++-------------- > x86/vmx.h | 50 +++++-- > x86/vmx_tests.c | 366 +++++++++++++++++++----------------------------- > 3 files changed, 263 insertions(+), 344 deletions(-) > Queued, thanks. Paolo
