From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga06.intel.com (mga06.intel.com [134.134.136.31]) by gabe.freedesktop.org (Postfix) with ESMTPS id 5E7EB6E0E6 for ; Tue, 17 Mar 2020 12:36:54 +0000 (UTC) Date: Tue, 17 Mar 2020 14:36:50 +0200 From: Ville =?iso-8859-1?Q?Syrj=E4l=E4?= Message-ID: <20200317123650.GT13686@intel.com> References: <20200307211751.22125-1-mohammed.khajapasha@intel.com> <20200307211751.22125-4-mohammed.khajapasha@intel.com> <87d09lnca7.fsf@intel.com> <1EBC98102E7E77409022937977AC68920C3E4633@BGSMSX110.gar.corp.intel.com> <878sk8oc32.fsf@intel.com> <20200310121445.GD13686@intel.com> <1EBC98102E7E77409022937977AC68920C3EDCBE@BGSMSX110.gar.corp.intel.com> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <1EBC98102E7E77409022937977AC68920C3EDCBE@BGSMSX110.gar.corp.intel.com> Subject: Re: [igt-dev] [PATCH i-g-t 3/3] kms/test: Pass correct pipe value to print pipe name List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" To: "Khajapasha, Mohammed" Cc: "igt-dev@lists.freedesktop.org" List-ID: On Tue, Mar 17, 2020 at 05:11:27AM +0000, Khajapasha, Mohammed wrote: > = > = > > -----Original Message----- > > From: Ville Syrj=E4l=E4 > > Sent: Tuesday, March 10, 2020 5:45 PM > > To: Jani Nikula > > Cc: Khajapasha, Mohammed ; igt- > > dev@lists.freedesktop.org > > Subject: Re: [igt-dev] [PATCH i-g-t 3/3] kms/test: Pass correct pipe va= lue to > > print pipe name > > = > > On Tue, Mar 10, 2020 at 10:20:33AM +0200, Jani Nikula wrote: > > > On Tue, 10 Mar 2020, "Khajapasha, Mohammed" > > wrote: > > > > Are you saying, we should get rid of pipe enum and make it array > > > > index based? Is my understanding correct? At present, every IGT > > > > subtest iterates over pipe list using "enum pipe" and this is > > > > getting passed down. > > > > > > I am saying do not conflate enum pipe and crtc index. > > > > > > If you call something a pipe, it should be the pipe, not crtc index. > > > And vice versa. > > > > > > It gets really confusing really quick if you call something a pipe, > > > and then go on to convert it into a pipe, as if it wasn't a pipe alre= ady. > > > > > > In i915 we've removed all assumptions that pipe =3D=3D crtc index. So > > > should igt. > > = > > I tend to agree, especially as other hw doens't use the "pipe" > > terminology, nor do they necessarily refer to them with A/B/C/... > > = > > The ugly part would be that we'd have maybe rename all tests with a pipe > > in the name. Should we just use the crtc index as a bare number in those > > cases? > = > As per my understandings, yes, we need to use bare number of crtc index > to get the pipe in display->pipes[] array for for_each_pipe() macros. > = > > = > > This would also have some complications for being able to quickly see > > which pipe had a specific failure as you have to dig through the logs t= o get > > that (though we'd anyway need to do that if the pipes are fused off in a > > way that makes the crtc index not match the pipe). > > = > > Also in some cases we may actually want to know which pipe we're dealing > > with, but currently there is no convenient way for userspace to find out > > which pipe maps to which crtc. > > > In this patch series, we are able to map the kernel pipe to IGT pipe using > GET_PIPE_FROM_CRTC_ID ioctl, with this we are able to print the logs for = failure > pipes in IGT. Right. I forgot that we had that ioctl. -- = Ville Syrj=E4l=E4 Intel _______________________________________________ igt-dev mailing list igt-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/igt-dev