From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marcel Ziswiler Date: Wed, 14 Sep 2016 21:44:40 +0000 Subject: [U-Boot] [PATCH v1 3/5] colibri_t20: fix display configuration In-Reply-To: References: <1473437441-938-1-git-send-email-marcel.ziswiler@toradex.com> <1473437441-938-4-git-send-email-marcel.ziswiler@toradex.com> <1473866438.14918.14.camel@toradex.com> Message-ID: <1473889479.10781.12.camel@toradex.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Wed, 2016-09-14 at 17:19 +0000, Stephen Warren wrote: > On 09/14/2016 09:20 AM, Marcel Ziswiler wrote: > > > > On Mon, 2016-09-12 at 12:18 -0600, Stephen Warren wrote: > > > > > > On 09/09/2016 10:10 AM, Marcel Ziswiler wrote: > > > > > > > > > > > > Without this patch the following error will be shown: > > > > > > > > stdio_add_devices: Video device failed (ret=-22) > > > > > > > > As commit ec5507707a1d1e84056a6c864338f95f6118d3ca (video: > > > > tegra: Move > > > > to using simple-panel and pwm-backlight) states the Colibri T20 > > > > needs > > > > updating too which this patch finally attempts doing. > > > > > > > > Please note that the current U-Boot implementation requires a > > > > dummy > > > > GPIO e.g. for a fixed backlight regulator to be explicitly > > > > defined in > > > > order to work unlike in the Linux kernel where this is taken > > > > care of > > > > automatically. > > > > > > > > > > > > > + compatible = "regulator-fixed"; > > > > + reg = <0>; > > > > + regulator-name = "Dummy"; > > > > + /* Dummy N/C */ > > > > + gpio = <&gpio TEGRA_GPIO(V, 7) > > > > GPIO_ACTIVE_HIGH>; > ?>> > > > > > > > > This is wrong. If that GPIO isn't actually part of the backlight, > > > the DT > > > should not say that it is. The gpio property is optional > > > according to > > > the DT binding documentation, so this shouldn't be needed. > > Well, I guess then it's lying. > Does "it" mean the binding? Please note that the binding defines how > the? > DT should be structured and how code interpreting the DT should > operate.? > The binding isn't derived from the code, but rather the code is > derived? > from the binding. In theory I agree but in practical speak this is wishful thinking as we just first handedly saw now. > ?> If I leave it away I get the following: > > > > > > stdio_add_devices: Video device failed (ret=-38) > > > > And it won't quite work. > That sounds like a bug in the U-Boot regulator driver. I believe you? > should fix that, rather than working around the bug in DT. Yes, I am pretty sure it won't be the last bug I uncover. Unfortunately right now I do not feel like fixing all of U-Boot. This series just tries to fix a very few select things (e.g. display and USB) which used to work just fine before some bigger agendas came along and broke them.