DVI output on i.MX51 EVP board not working?

[Philipp Zabel]

Hi Wladimir,

On Fri, 2017-03-31 at 07:36 +0200, Wladimir J. van der Laan wrote:
> Hello,
> 
> I can't get the DVI output to work on a i.MX51 EVP board (babbage). There
> seems to be no signal at all. The board is running the mainline kernel 4.11.0-rc4 as of a
> day ago (89970a0).
> 
> - The connector and cable works. u-boot (v2016.07-rc2) shows a penguin at boot.
>   However while the Linux kernel boots, the signal goes away: monitor (a fairly
>   old Dell 24" panel, max resolution 1600x1200, works fine with Linux/DRM with other
>   hardware) goes to sleep mode.
> 
> - Monitor is connected through HDMI<->DVI cable, could try another monitor if
>   that's potentially the issue (Could it be EDID related? I'd guess not as the timings
>   are hardcoded in the device tree).
> 
> - Haven't changed anything in the dtb with regard to port config or resolution
>   or timings. The default ones in imx51-babbage.dts should theoretically just
>   work. I checked and these are the same as in u-boot:
> 
>         timing0: dvi {
>                 clock-frequency = <65000000>;
>                 hactive = <1024>;
>                 vactive = <768>;
>                 hback-porch = <220>;
>                 hfront-porch = <40>;
>                 vback-porch = <21>;
>                 vfront-porch = <7>;
>                 hsync-len = <60>;
>                 vsync-len = <10>;
>         };
> 
> - Went as far back as kernel v4.0, even to v3.12 or so (commit 493a863, "ARM:
>   dts: imx51-babbage: Make DVI and WVGA panel functional"). No difference. So nothing to
>   bisect, unfortunately.

Looking at the hardware user's guide [1] chapter 9 (VGA and DVI out), I
notice that the TFP410PA bridge seems to be connected to the DISP2_DAT
data lines, but uses the DISP1_H/V sync signals (which I can't believe
is right), according to figure 9-1. But the device tree pin
configuration configures the DISP1_DAT data lines with pinctrl_ipu_disp1
instead. This pin configuration was moved there from imx51.dtsi, so I
think it was just never adapted for this boards' quirks.

[1] http://www.nxp.com/assets/documents/data/en/user-guides/evk_imx51_Hardware_UG.pdf

It might also be useful to find out which GPIOs are connected to the
TFP410 power down and reset lines.

> - In dmesg this is all the output I get:
> 
> [    0.801700] imx-ipuv3 40000000.ipu: no port at 0 node in /soc/ipu at 40000000, not using CSI0
> [    0.809826] imx-ipuv3 40000000.ipu: no port at 1 node in /soc/ipu at 40000000, not using CSI1
> [    0.818600] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
> [    0.825230] [drm] No driver support for vblank timestamp query.
> [    0.831481] imx-drm display-subsystem: bound imx-ipuv3-crtc.0 (ops ipu_crtc_ops)
> [    0.839132] imx-drm display-subsystem: bound imx-ipuv3-crtc.1 (ops ipu_crtc_ops)
> [    0.846617] imx-drm display-subsystem: bound display at di0 (ops imx_pd_ops)
> [    0.986410] Console: switching to colour frame buffer device 128x48
> [    1.032306] imx-drm display-subsystem: fb0:  frame buffer device
> [    1.039362] [drm] Initialized imxdrm 1.0.0 20120507 for display-subsystem on minor 0
> [    1.047276] imx-ipuv3 40000000.ipu: IPUv3EX probed
> 
> As far as I know this is nothing unexpected. CSI0 and CSI1 are not connected but DI0 should be the
> DVI port, which does get initialized. The size of the framebuffer device seems to indicate the resolution
> is set correctly (assuming 16x8 characters).
> 
> There was also a "timeout waiting for vsync" error, 
> 
> [    0.917254] WARNING: CPU: 0 PID: 1 at drivers/gpu/drm/drm_atomic_helper.c:1122 drm_atomic_helper_wait_for_vblanks.part.0+0x264/0x274
> [    0.917259] [CRTC:26] vblank wait timed out
> 
> which was fixed after applying Philipp Zabel's "drm/imx: don't wait for vblank and stop calling cleanup_planes in commit_tail" patch series.
> However that didn't solve the output issue.
> 
> I'm out of things to try to diagnose this, does anyone perhaps have an idea or seen this issue
> before?
> 
> Regards,
> Wladimir van der Laan
> 

regards
Philipp