From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Nelson Date: Fri, 13 Sep 2013 07:51:22 -0700 Subject: [U-Boot] [PATCH v2 1/2] mx6sabresd: Avoid hang when HDMI cable is not connected In-Reply-To: References: <1378934070-11721-1-git-send-email-festevam@gmail.com> <52318D9E.1@denx.de> <5231D78A.5000002@boundarydevices.com> <5232C9EC.604@denx.de> Message-ID: <5233266A.7050302@boundarydevices.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Hi Fabio, On 09/13/2013 06:58 AM, Fabio Estevam wrote: > Hi Stefano, > > On Fri, Sep 13, 2013 at 5:16 AM, Stefano Babic wrote: >> I admit that using DT also for u-boot seems a better solution. You're >> right about i.MX6 in kernel, but on the other hand I like how it is >> described for i.MX28 boards. The display-timings node contain all >> information we need. It would be nice to have the same for i.MX6. > > We currently have the same in DT for mx6 as well. Check this commit, > for example: > https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/arch/arm/boot/dts/imx6q-sabrelite.dts?id=722cfacd2e719b6447fb4bd5cd3372725876336d > Nice! So, how should we get this done? We have support for parsing individual lines of DT, and the primary thing(s) needed by U-Boot's display are in the timing block of the device tree, but I wonder whether it makes sense to implement a full parser for that, or a simpler parser for fb_videomode. Since both cfb_console in U-Boot and the of_get_fb_videomode() routine in the kernel use fb_videomode, it should be straightforward to hand the information off. I don't think the outer information (the lvds-channel at 0 and ldb blocks in imx6q-sabrelite.dts) should be parsed by U-Boot as device-tree code since that path leads to having a full device-tree compiler that seems inappropriate. Let me know your thoughts. Eric