On Mon, Jan 23, 2017 at 11:20:42AM -0500, Sean Paul wrote: > On Mon, Jan 23, 2017 at 05:01:56PM +0100, Tobias Jakobi wrote: > > Sean Paul wrote: > > > <snip> > > > > > >>>>> As of now, I don't see any case. even without Maarten's patch set, it works well - actually, I had a test with atomic test app more than 10 hours.. > > >>>> Can you provide this test application? In particular I'm asking this > > >>>> because libdrm currently doesn't provide any tests using the atomic API. > > >>>> So this application might be of interest also for other people. > > >>> > > >>> Below is the app I tested. Know that this application is from chromiumOS tree and I just fixed some parts for internal test. > > >>> https://review.tizen.org/git/?p=atform/upstream/libdrm.git;a=commitdiff;h�3bd95f2c5a9b4b69062a3ff008947054b94f55 > > >> Thanks, any chance this is going to be submitted upstream? > > >> > > > > > > Probably not in its current form. I just wrote it to quickly test out > > > some stuff we didn't yet support in CrOS. I don't really think it's fit > > > for inclusion upstream. > > Thanks for the clarification! Just voicing my interest here to have > > something like this upstream. I mean with atomic now being mandatory for > > new DRM kernel drivers and all... :-) > > Agreed that tests are important, I'm just hesitant to sling my spaghetti code > around too widely :) > > The igt suite should have you covered for basic testing and more. CrOS also has a > bsdrm test suite that may or may not become atomic-aware in the future. Both are > better options that my atomictest, IMO. +1 on igt. I kinda would like to nuke all other test apps we have floating around, to avoid duplicating effort. Unfortunately we dont have a full replacement for modetest in igt yet for manual testing. The testcases are there, and igt supports manually/interactive testing so you can check the results by hand, it's just not wired up yet for the tests that would be equivalent to modetest ... hint, hint :-) -Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
