[RFC PATCH 0/2] Propose critical clocks

[Marco Felsch]

Hi Peng,

On 22-09-13, Peng Fan wrote:
> On 9/13/2022 6:21 PM, Marco Felsch wrote:
> > Hi,
> > 
> > this proposal is to mark clocks as critical. It is somehow inspired by
> > the regulator-always-on property. Since sometimes we can end in circular
> > dependcies if we wanna solve the dependcies for a specific clock
> > provider.
> > 
> > The property is generic so it can be used by every hw clock provider. So
> > it can be seen as generic implementation to [1].
> 
> Thanks for working on a generic solution, I think your proposal could also
> help [1] and try to resolve same issue as [2].

Didn't noticed that there was already a on going discussion on this
topic. Maybe a combination of my proposal and [2] is the solution, but I
have no idea if I my solution can work on a ID based array. Let's see
what the maintainers say.

> [1] https://lore.kernel.org/all/CAJ+vNU1Za2CPGVX3q4HKufsxbL5zRrk1B5CWFpKritetrTs4dA@mail.gmail.com/
> [2] https://lore.kernel.org/all/20220517235919.200375-1-marex@denx.de/T/#m52d6d0831bf43d5f293e35cb27f3021f278d0564
> 
> Thanks,
> Peng.
> 
> 
> > 
> > [1] https://lore.kernel.org/linux-clk/20220913092136.1706263-1-peng.fan@oss.nxp.com/
> > 
> > Marco Felsch (2):
> >    clk: add support for critical always-on clocks
> >    arm64: dts: imx8mm-evk: mark 32k pmic clock as always-on
> > 
> >   arch/arm64/boot/dts/freescale/imx8mm-evk.dtsi |  1 +
> >   drivers/clk/clk.c                             | 18 ++++++++++++++++++
> >   2 files changed, 19 insertions(+)
> > 
>