[PATCH 2/8] dt-bindings: spi: fsl-qspi: support SpacemiT K1

[Conor Dooley]

On Mon, Oct 20, 2025 at 07:26:17PM +0100, Mark Brown wrote:
> On Mon, Oct 20, 2025 at 01:06:46PM -0500, Alex Elder wrote:
> > On 10/20/25 12:39 PM, Conor Dooley wrote:
> 
> > > > +          - spacemit,k1-qspi
> 
> > > Are the newly added resets mandatory for the spacemit platform?
> 
> > This is interesting.  I never even tried it without specifying them.
> 
> > I just tried it, and at least on my system QSPI functioned without
> > defining these resets.  I will ask SpacemiT about this.  If they are
> > not needed I will omit the first patch (which added optional resets),
> > and won't use them.
> 
> It might be safer to describe them, otherwise things are vulnerable to
> issues like the bootloader not leaving things in a predictable state.

Yeah, if a linux driver requires that a bootloader set up a clock or
de-assert a reset etc, then the binding should mark them required since,
as you say, a bootloader change might do away with that de-assertion.
Additionally, the stage doing that de-assertion etc could be U-Boot
or barebox, which import devicetrees from Linux, so making sure that
the resets are present has that benefit too.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-riscv/attachments/20251020/6ac0fd39/attachment.sig>