From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-pb0-f46.google.com ([209.85.160.46]:44760 "EHLO mail-pb0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753882Ab2JJAWx (ORCPT ); Tue, 9 Oct 2012 20:22:53 -0400 Received: by mail-pb0-f46.google.com with SMTP id rr4so47986pbb.19 for ; Tue, 09 Oct 2012 17:22:52 -0700 (PDT) From: Kevin Hilman To: Sakari Ailus Cc: paul@pwsan.com, laurent.pinchart@ideasonboard.com, linux-media@vger.kernel.org, linux-omap@vger.kernel.org Subject: Re: [PATCH v2 1/2] omap3: Provide means for changing CSI2 PHY configuration References: <20120926215001.GA14107@valkosipuli.retiisi.org.uk> <1348696236-3470-1-git-send-email-sakari.ailus@iki.fi> <87zk3vz7yb.fsf@deeprootsystems.com> <20121009223340.GM14107@valkosipuli.retiisi.org.uk> Date: Tue, 09 Oct 2012 17:22:52 -0700 In-Reply-To: <20121009223340.GM14107@valkosipuli.retiisi.org.uk> (Sakari Ailus's message of "Wed, 10 Oct 2012 01:33:40 +0300") Message-ID: <87pq4ryy3n.fsf@deeprootsystems.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-media-owner@vger.kernel.org List-ID: Sakari Ailus writes: > Hi Kevin, > > Thanks for the comments! > > On Tue, Oct 09, 2012 at 01:50:04PM -0700, Kevin Hilman wrote: >> Hi Sakari, >> >> Sakari Ailus writes: >> >> > The OMAP 3630 has configuration how the ISP CSI-2 PHY pins are connected to >> > the actual CSI-2 receivers outside the ISP itself. Allow changing this >> > configuration from the ISP driver. >> > >> > Signed-off-by: Sakari Ailus >> >> These control module registers (CSIRXFE, CAMERA_PHY_CTRL) are in the >> CORE powerdomain, so they will be lost during off-mode transitions. So, >> I suspect you'll also want to add them to the save/restore functions in >> control.c in order for this to work across off-mode transitions. > > I've got another patch that implements this in the ISP driver instead. > > Oops, sorry. I should've looked at v3. > The ISP also can't wake up the MPU from the off mode, so I don't think > losing the register contents is necessarily an issue. The registers will be > written to a new value whenever streaming is started. Perhaps adding a note > about that would be worthwhile. Yes, I suggest mentioning that these register are not saved/restored and why would be helpful. Kevin