Re: [PATCH v2 3/3] riscv: dts: spacemit: enable USB3 on OrangePi RV2

From: Yixun Lan <dlan@kernel.org>
To: Chukun Pan <amadeus@jmu.edu.cn>
Cc: alex@ghiti.fr, aou@eecs.berkeley.edu, conor+dt@kernel.org,
	devicetree@vger.kernel.org, krzk+dt@kernel.org,
	linux-kernel@vger.kernel.org, linux-riscv@lists.infradead.org,
	palmer@dabbelt.com, pjw@kernel.org, robh@kernel.org,
	spacemit@lists.linux.dev
Subject: Re: [PATCH v2 3/3] riscv: dts: spacemit: enable USB3 on OrangePi RV2
Date: Fri, 3 Apr 2026 08:16:44 +0800	[thread overview]
Message-ID: <20260403001644-GKC1016296@kernel.org> (raw)
In-Reply-To: <20260402130608.133154-1-amadeus@jmu.edu.cn>

Hi Chukun,

On 21:06 Thu 02 Apr     , Chukun Pan wrote:
> Hi,
> 
> > Can you work with Han for adding USB support[1]? this will simply
> > distribute our effort, and make the review process even harder
> 
> Sorry, I didn't consider this. 
> Could you drop this patch ([PATCH v2 3/3]...)? 
> The first two patches should not cause conflicts.
> 
> > > +	vbus-supply = <&vcc5v0_usb30>;
> > IMO, the vbus doesn't directly tie to dwc3 host, but to HUB's port
> > so I think this is still wrong, although it may work on the board..
> 
> We can switch to the onboard_usb_dev driver after it's merged.
Sure, I think that's a more elegant solution

> Keeping vbus always-on may cause unnecessary waste.
> I won't insist if you think this is wrong.
Currently either way isn't ideal.. but binding vbus to the host is kind
of acceptable to me, if no ojection, I can take this, then we can adjust
later once the USB HUB vbus support[1] landing in tree
https://lore.kernel.org/all/20260223-v6-16-topic-usb-onboard-dev-v5-0-28d3018a8026@pengutronix.de/ [1]

btw, Could you take a look at Han's patch? if you have final say?
https://lore.kernel.org/spacemit/20260403000539-GKB1016296@kernel.org/ [2]

I'd like to push this series to v7.1, it's almost running out of time..

> 
> Thanks,
> Chukun

-- 
Yixun Lan (dlan)