Re: [PATCH v1 0/3] Add Luckfox Omni3576 Carrier Board support for RK3576

From: "Heiko Stübner" <heiko@sntech.de>
To: John Clark <inindev@gmail.com>,
	Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Cc: robh@kernel.org, krzk+dt@kernel.org, conor+dt@kernel.org,
	detlev.casanova@collabora.com,
	linux-arm-kernel@lists.infradead.org,
	linux-rockchip@lists.infradead.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org, John Clark <inindev@gmail.com>,
	sebastian.reichel@collabora.com
Subject: Re: [PATCH v1 0/3] Add Luckfox Omni3576 Carrier Board support for RK3576
Date: Sun, 04 May 2025 10:08:09 +0200	[thread overview]
Message-ID: <3352030.oiGErgHkdL@diego> (raw)
In-Reply-To: <7462454.lOV4Wx5bFT@workhorse>

Hi,

Am Sonntag, 4. Mai 2025, 01:39:56 Mitteleuropäische Sommerzeit schrieb Nicolas Frattaroli:
> On Friday, 2 May 2025 22:55:30 Central European Summer Time John Clark wrote:
> > Disabled features:
> > - eMMC: Not populated on the tested board; believed to be functional but disabled.
> > - HDMI: Disabled due to mainline driver maturity issues (basic 1080p output works; audio and 4K untested).
> 
> that's not a good reason to disable the node, in my eyes. HDMI 2.0 4K output 
> should work AFAIK, HDMI audio will land with the already posted SAI audio 
> controller patch set. Is there any specific issue like an SoC lock-up you ran 
> into that is keeping you from enabling it? If so, please do let the mailing list 
> know.

Also a more general point, Devicetree is supposed to describe the hardware
and both emmc and hdmi do work on a hw-level and have reviewed bindings,
so the node can be enabled and also any functionality improvement should
just affect the kernel driver-side.

And in my testing on the rk3576 firefly, I also got a working 3440x1440
on my display. So not a full 4k, but nearly there :-) .

> - RK3576 USB Type-C OTG stuff: the USB 2 PHY driver needed some changes to make
>   this work. I've yet to send out v2 that reworks the way we detect a type-C
>   port, but that part of the code is already done. I was basically just letting
>   it stew on my local machine for a bit longer to see if I can figure out why
>   superspeed only works in one cable orientation,

That is probably the same issue, I have on rk3588, the type-c phy not
changing its settings and waiting for the dwc3 to re-start, which it never
does.

See [0], [1] ... but importantly Ondřej Jirman's response on why my approach
is probably wrong [2].

At this time, I don't know when I get around to look at that again, so
if someone wants to give it a try, feel free to :-)

Heiko

[0] https://lore.kernel.org/r/20250225184519.3586926-1-heiko@sntech.de
[1] https://lore.kernel.org/r/20250226103810.3746018-1-heiko@sntech.de
[2] https://lore.kernel.org/r/h57ok2hw6os7bcafqkrqknfvm7hnu25m2oe54qmrsuzdwqlos3@m4och2fcdm7s