Re: [PATCH 1/9] dt-bindings: bus: Add SpacemiT K3 eSPI controller

From: Conor Dooley

Date: Thu Nov 19 2026 - 06:10:24 EST


On Thu, Nov 19, 2026 at 11:02:03AM +0000, Conor Dooley wrote:
> On Thu, Nov 19, 2026 at 05:05:53PM +0800, Wang Yang wrote:
> > Hi Conor, Aurelien,
> >
> > > How is this "eSPI" different to the eSPI described in fsl,espi.yaml and
> > > implemented by a driver in drivers/spi?
> >
> > Thank you, Aurelien, for clarifying this. K3 uses the Intel-defined
> > eSPI interface intended to replace LPC. It is a different protocol from the
> > Freescale/NXP SPI controller described by fsl,espi.yaml.
>
> Ah I think this might be confusing af, because that device also
> supports what it calls "eLBC"...

As in, I got confused while looking at the documentation for the fsl
device between what they were calling eSPI and eLBC, and also didn't
realised that it was lBc not lPc.. Too many similar-ish terms...

>
> > On K3, the controller exposes a CPU-accessible window for the eSPI
> > Peripheral Channel. The EC client accesses its command, response
> > and memory-map regions through this window, rather than exchanging
> > ordinary SPI messages.
> >
> > The proposed controller driver handles clocks, resets, link
> > configuration and window setup. The EC transport handles the EC
> > register mapping and host commands. The distinction is therefore
> > the protocol and access model, rather than the use of MMIO itself.
> >
> > I will make the Intel eSPI terminology and this access model
> > explicit in the binding description.
>
> Please take a look at a similar series from Aspeed and see if you can
> cooperate on creating something common between you:
> https://lore.kernel.org/all/20260313-upstream_espi-v1-0-9504428e1f43@xxxxxxxxxxxxxx/

Attachment: signature.asc
Description: PGP signature