Re: [PATCH 4/6] riscv: dts: spacemit: enable QSPI and add SPI NOR on Milk-V Jupiter

From: Yixun Lan <dlan@kernel.org>
To: Yao Zi <me@ziyao.cc>
Cc: Aurelien Jarno <aurelien@aurel32.net>,
	linux-kernel@vger.kernel.org, Rob Herring <robh@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Paul Walmsley <pjw@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	"open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
	<devicetree@vger.kernel.org>,
	"open list:RISC-V ARCHITECTURE" <linux-riscv@lists.infradead.org>,
	"open list:RISC-V SPACEMIT SoC Support"
	<spacemit@lists.linux.dev>
Subject: Re: [PATCH 4/6] riscv: dts: spacemit: enable QSPI and add SPI NOR on Milk-V Jupiter
Date: Wed, 25 Mar 2026 05:59:37 +0800	[thread overview]
Message-ID: <20260324215937-GKA767626@kernel.org> (raw)
In-Reply-To: <acLE5xnEBoCtW4Vw@pie>

Hi Yao,

On 17:07 Tue 24 Mar     , Yao Zi wrote:
> On Tue, Mar 24, 2026 at 05:02:20PM +0800, Yixun Lan wrote:
> > Hi Aurelien,
> > 
> > On 21:28 Sun 22 Mar     , Aurelien Jarno wrote:
> > > Add the QSPI controller node for the Milk-V Jupiter board and describe
> > > the attached SPI NOR flash (GD25Q64E).
> > > 
> > > The flash supports a frequency up to 133MHz (80 MHz for reads), and the
> > > SoC supports a frequency up to 104 MHz. However tests have shown that
> > > the flash is not reliably detected above 26.5 MHz, consistent with
> > > frequency used in the vendor kernel. Therefore, use this frequency.
> > > 
> > ..
> > > The m25p,fast-read properties is taken from the vendor kernel.
> > > 
> > So long as this is verified and works fine on board?
> > 
> > > Add a corresponding flash partition layout, matching the layout and the
> > > names used in the vendor U-Boot.
> > > 
> > ..
> > > Also add the bootph-pre-ram property to make the device tree usable by
> > > early firmware/bootloaders without modification, as U-Boot is stored on
> > > this NOR flash.
> > Is the dtb file actually used by U-Boot? I'd highly doubt about this,
> > if not the case or has not been tested, I'd suggest then not to add
> > this property..
> 
> Currently no, but it would be a bogus if we could do it at introduction
> of the device, if it's really necessary in pre-DRAM stages of
> bootloaders, e.g. U-Boot SPL. This would reduce the duplicated work of
> downstream projects if they decide to switch to Linux upstream
> devicetree.
> 
All above arguments are based on assumption, I object to add things randomly
without real test proved.

Let's wait once U-Boot actually switch to use upstream devicetree.

> As SpacemiT K1 supports booting from flash, the description of flash
> nodes are likely useful for pre-DRAM bootloaders.
> 
This is still a 'likely' ..

> > > 
> > > Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
> 
> Best regards,
> Yao Zi

-- 
Yixun Lan (dlan)