Re: [PATCH 1/2] serial: msm-geni: allow invalid clock

[Sumit Garg <sumit.garg@kernel.org>]

On Wed, Apr 08, 2026 at 11:16:24AM +0200, Casey Connolly wrote:
> Hi Sumit,
> 
> On 07/04/2026 13:38, Sumit Garg wrote:
> > On Fri, Mar 20, 2026 at 04:45:40PM +0100, Casey Connolly via B4 Relay wrote:
> >> From: Casey Connolly <casey.connolly@linaro.org>
> >>
> >> Pre-relocation we probably won't have a clock but it's usually
> >> been enabled for us already, or worst case we will enable it after
> >> relocation.
> >>
> > 
> > Not sure if this is the right approach, since the UART driver might just
> > hang without clock being enabled. What's the motivation for this change?
> 
> Hmm, I didn't consider that. The motivation here is to handle CCF, since
> we don't want to support clock pre-relocation. In pretty much all cases
> today the serial clocks are already configured but you're right that we
> ought not to depend on that.

I think this again comes down to that how CCF framework can enable
U-Boot SPL use-cases or where in U-Boot proper you get to work with an
uninitialized UART port for an OEM board variant.

> 
> I can't really think of another good solution than hardcoding this clock
> configuration somewhere :/
> 

That's true since we rely totally on DT based driver probing in U-Boot.

> Any ideas?
> 

Have you tried working with DT properties like bootph-pre-ram etc. that
allows only say UART driver to probe before relocation with CCF?

-Sumit