panel-simple-dp-aux: "DP AUX done_probing() can't defer" on MT8186 w/ Collabora kernel

Chen-Yu Tsai wenst at chromium.org
Fri Aug 2 07:16:06 UTC 2024


On Tue, Jul 30, 2024 at 7:24 AM Dmitry Baryshkov
<dmitry.baryshkov at linaro.org> wrote:
>
> On Sat, Jul 27, 2024 at 08:33:10PM GMT, Alper Nebi Yasak wrote:
> > Hi,
> >
> > I have a MT8186 "Magneton" Chromebook that I'm trying to boot a kernel
> > based on Collabora's for-kernelci branch [1], using a config from
> > postmarketOS [2] (intended for that), on a Debian sid installation.
> > This sometimes fails to enable the display with:
> >
> > > panel-simple-dp-aux aux-0-0058: DP AUX done_probing() can't defer
>
> The error is caused by mtk_dp_edp_link_panel() returning -EPROBE_DEFER
> (most likely because something is missing). However this callback must
> not defer, it's called at at a different point.
>
> The actual issue seems to be in these lines. The callback tries to bind
> all components, however it should not as the DPI isn't available yet.

My guess is that I upstreamed an incomplete external display pipeline.
The pipeline is still missing a representation for a mux that routes
the signal to either the left or right side type-C port.

I will send a patch to disable it.

ChenYu

> [drm:drm_bridge_attach] *ERROR* failed to attach bridge /soc/dpi at 1400a000 to encoder TMDS-32: -517
> [drm:mtk_dsi_host_attach] *ERROR* failed to add dsi_host component: -517
> anx7625 0-0058: [drm:anx7625_link_bridge] *ERROR* fail to attach dsi to host.
>
> >
> > I know this is a weird out-of-tree combination of things but I've been
> > asked to report this on the mailing list by wens on the #linux-mediatek
> > IRC channel.
> >
> > [1] https://gitlab.collabora.com/google/chromeos-kernel/-/commits/for-kernelci/
> >
> > [2] https://gitlab.com/postmarketOS/pmaports/-/blob/master/device/testing/linux-postmarketos-mediatek-mt81xx/config-postmarketos-mediatek-mt81xx.aarch64
> >
> > Serial output from the system starting with dmesg and some more info copied below.
>
>
> [trimmed the log]
>
> --
> With best wishes
> Dmitry


More information about the dri-devel mailing list