From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eugeni Dodonov Subject: Re: [PATCH 01/25] drm/i915: add new Haswell DIP controls registers Date: Wed, 09 May 2012 22:02:05 -0300 Message-ID: <4FAB138D.1020204@linux.intel.com> References: <1336588652-702-1-git-send-email-eugeni.dodonov@intel.com> <1336588652-702-2-git-send-email-eugeni.dodonov@intel.com> Reply-To: eugeni.dodonov@intel.com Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTP id 148CF9E7BF for ; Wed, 9 May 2012 18:01:39 -0700 (PDT) In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org Errors-To: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org Cc: intel-gfx@lists.freedesktop.org, Eugeni Dodonov List-Id: intel-gfx@lists.freedesktop.org On 05/09/2012 09:34 PM, Paulo Zanoni wrote: > 2012/5/9 Eugeni Dodonov: >> Haswell has different DIP control registers and offsets which we need to >> use for infoframes, which this patch adds. >> >> Note that this does not adds full DIP frames support, but only the basic >> functionality necessary for HDMI to work in early enablement. >> >> Signed-off-by: Eugeni Dodonov >> --- > > Hi > > The i915_reg.h chunk seems correct, but I believe hsw_write_infoframe > is wrong. The DIP control register completely changed, and the way we > write the infoframes has also changed. Previously, we had just one > "data" register for the full 32-byte buffer and we controlled which > part of the buffer we wanted to write by the last 4 bits of the > "control" register. These last 4 bits on the control register don't > exist anymore (the whole register changed) and we have a ton of "data" > registers now. I didn't test anything I wrote here, I'm just assuming > based on the docs I have :) > > For basic HDMI enablement this function is not strictly necessary, so > maybe just adding a FIXME (like you did) and returning without doing > anything would be fine. I'm going to provide a patch for this. (Just following up on what we discussed with Paulo over irc). Without this, I am getting a nice black screen though . But the latest theory is that perhaps we just need to send any infoframe for the hardware to catch up and work in HDMI mode? Eugeni