On Thu, 2012-11-22 at 15:04 +0100, Daniel Vetter wrote: > On Thu, Nov 22, 2012 at 2:50 PM, Imre Deak <imre.deak at intel.com> wrote: > >> > + enable_monotonic_timestamp(!(flags & TEST_TIMESTAMP_REAL)); > >> > >> I don't follow why we should enable monotonic timestamps for some tests > >> and not for some others? Shouldn't all tests with TEST_CHECK_TS just use > >> the same clock the kernel uses? > > > > The idea was to also test the compatibility mode, where we get real > > timestamps. > > Oh, I've missed that you write back the desired value. I don't think > we need to test the compat mode if it's not enabled, just test > whatever mode is enabled. That way we also check whether the getcap > interface works. Ok, will do so. We have to wait though until DRM_CAP_TIMESTAMP_MONOTONIC is added to libdrm. I've sent already a patch for this. --Imre
