On Thu, Sep 17, 2015 at 04:47:06PM +0100, Thomas Wood wrote: > On 17 September 2015 at 13:43, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: > > On Thu, Sep 17, 2015 at 01:16:18PM +0100, Thomas Wood wrote: > >> On 17 September 2015 at 13:09, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: > >> > On Thu, Sep 17, 2015 at 12:42:44PM +0100, Thomas Wood wrote: > >> >> Add a script to take a piglit results file and create a list of tests > >> >> that ran in under 60 seconds. This list can be used by the --test-list > >> >> option of piglit. > >> > > >> > And passed? > >> > >> Any status except "incomplete", which normally means a reboot was > >> required. It might also be worth noting that tests skipped on one > >> platform may not skip on another, so the output only really applies to > >> the platform the results were initially produced on. > > > > My point is that the timing can only be expected to be consistent for a > > pass. A failure may be quick, but success may take a few hours. > > I think including failed tests depends on whether the "quick" run > should be able to detect fixes from previous runs. Perhaps an option > to exclude anything other than successful tests might be useful. There > is also the timeout mechanism in piglit/igt that will prevent any test > from running for more than 10 minutes. Yeah, for the fast run we also need to reduce the piglit timeout to 60s (and kill anything that timed out ofc too). Otherwise there's indeed a gap where a testcase that takes too long could sneak in. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
