On Fri, 2014-04-11 at 02:57 -0600, Daniel Vetter wrote: > On Fri, Apr 11, 2014 at 08:56:28AM +0800, Zhao Yakui wrote: > > On Thu, 2014-04-10 at 03:04 -0600, Daniel Vetter wrote: > > > On Thu, Apr 10, 2014 at 04:28:34PM +0800, Zhao Yakui wrote: > > > > BTW: Does it need to check all the flags defined in i915_drm.h or the > > > > exported flag returned by i915_get_parameter? > > > > > > I don't have i915_get_parameter anywhere in my sources, so no idea what > > > you mean ... > > > > Sorry that the function should be i915_getparam. It is called by the > > I915_GETPARAM ioctl to query the flag supported by the driver. > > Ah, now I understand. The idea is to test all fields of the structure > exhaustively (so also rsvd to make sure it's 0). Well except for the > buffer count field since we have tests for that already. > > For the reasons see my two blog posts on the topic: > > http://blog.ffwll.ch/2013/11/testing-requirements-for-drmi915.html > > http://blog.ffwll.ch/2013/11/botching-up-ioctls.html OK. It seems that the case needs to check more fields than the exported flag. I will take a look at your blog and understand how to write the test case. Thanks. Yakui > > Cheers, Daniel _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
