STM32 board support eDP bridge?

[LLe.1]

Hi,

We are using stm32mp157c MCU/board and it has DSI output built-in. Now we want to add display port support via DSI to eDP bridge SN65DSI86. It seems that stm32 does not support Displayport output or it has never been done. At least the driver dw_mipi_dsi-stm.c should be updated not to require a panel.

Any idea how this can be achieved please?

[LLe.1]

@Erwan SZYMANSKI​ As of now it looks like an issue with stm dsi driver that it does not retry or called too early, I think the bridge works as intended. But here it is mate, I did not apply any patch to the bridge, only dsi. Thanks again for help.

[LLe.1]

Here is my patched generic dsi driver

[LLe.1]

stm dsi driver with some debug prints, otherwise intact

[LLe.1]

@Erwan SZYMANSKI (ST Employee)​ As of now it looks like an issue with stm dsi driver that it does not retry or called too early, I think the bridge works as intended. But here it is mate, I did not apply any patch to the bridge, only dsi. Thanks again for help.

[LLe.1]

Here is my patched generic dsi driver

[LLe.1]

stm dsi driver with some debug prints, otherwise intact

[Erwan SZYMANSKI]

@LLe.1​ ,

Well we are facing the known issue between I2C and DSI probe (which is one of the things better managed in kernel 5.15 / OSTL 4.0). In fact, in a perfect world, your bridge is supposed to probe before your DSI and the, all should be good. But in your case, it is not.

If we want to do a quick hotfix to test, maybe you could try to do the following thing. Find the exact return that fails thanks to your logs and try to return a -EPROBE_DEFER instead. It will ask the DSI to retry its probe later.

Regards.

Erwan

[Erwan SZYMANSKI]

Hi @LLe.1​ ,

Do you have any update from your side ? I saw that some patches on the bridge driver also deal with your probe ordering issue, such as this patch.

Maybe it could be interesting to take a look at it.

Regards.

Erwan.

[LLe.1]

@Erwan SZYMANSKI​ sorry for late reply, I was in holiday. I checked the code and the function actually returns correct flag -EPROBE_DEFER so it is not that. However, I think the latest patch you mentioned may solve the issue, thanks for finding out. I will try it soon and let you know.

[LLe.1]

@Erwan SZYMANSKI​ the patch doesn't help because dw_mipi_dsi_bridge_attach() is called even before ti-bridge probe is called.

I start to think that I will need somehow to get new version of the bridge driver -> new kernel + drm. Or what is your suggestion?