From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gustavo Padovan Subject: Re: drm/exynos: Add atomic modesetting support Date: Tue, 12 May 2015 15:29:08 -0300 Message-ID: <20150512182908.GB3327@joana> References: <1430430651-22416-1-git-send-email-gustavo@padovan.org> <554D2FC7.1070208@gmx.net> <554DF9DE.7000200@gmx.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-qc0-f179.google.com ([209.85.216.179]:32830 "EHLO mail-qc0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933013AbbELS3M (ORCPT ); Tue, 12 May 2015 14:29:12 -0400 Received: by qcvo8 with SMTP id o8so9479207qcv.0 for ; Tue, 12 May 2015 11:29:11 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: linux-samsung-soc-owner@vger.kernel.org List-Id: linux-samsung-soc@vger.kernel.org To: Inki Dae Cc: Tobias Jakobi , "linux-samsung-soc@vger.kernel.org" , DRI mailing list 2015-05-10 Inki Dae : > 2015-05-09 21:13 GMT+09:00 Tobias Jakobi : > > Hello Inki, > > > > > > Inki Dae wrote: > >> Hi, > >> > >> 2015-05-09 6:51 GMT+09:00 Tobias Jakobi : > >>> Hello, > >>> > >>> I've tested this on my Hardkernel Odroid-X2 (connected via HDMI to a > >>> 1080p panel). > >>> > >>> Run the usual modetest tests (just primary plane, primary plane with > >>> vsync, primary plane with overlay, primary plane with overlay and video > >>> overlay, overlay partially outside of crtc area, etc.) and haven't > >>> noticed any issues so far. > >> > >> As I mentioned several times, it works well in case that only one crtc > >> driver is enabled. Could you check it again after you enable two or > >> more crtc drivers such as FIMD and HDMI or FIMD, HDMI and VIDI > >> together? For this, dts file for X2 should contain their device nodes > >> and also should be configurated though menuconfig. > > I've enabled VIDI and FIMD and confirmed that they should up properly in > > modetest before applying the series. > > > > Booting with the atomic series works fine, but I get a segfault when > > calling modetest. I've attached the kernel log below. > > Below panic issue is same as one I faced with at previous patch seris, > v3, which was invalid memory access - state->crtc is NULL whille > modetest is being performed. It seems that the last patch of v4 didn't > resolve this issue yet. The weird thing is that it works quite fine on my exynos 5 hardware. I'll take a look on what could be be happening to cause this kind of crash. Gustavo