[PATCH 2/4] thunderbolt: add support optional dedicated wakeup irq

mika.westerberg at linux.intel.com mika.westerberg at linux.intel.com
Tue Nov 3 03:59:42 PST 2026


Hi,

On Tue, Nov 03, 2026 at 09:05:45AM +0000, Chunfeng Yun (云春峰) wrote:
> On Fri, 2026-10-30 at 09:48 +0100, Mika Westerberg wrote:
> > On Fri, Oct 30, 2026 at 04:25:37PM +0800, Chunfeng Yun wrote:
> > > There is a dedicated wakeup irq on nvidia n1/n2 SoCs.
> > 
> > Okay but why it is needed?
> due to mac/phy is in power domain which its power/clock will be turn
> off when enter into sleep mode, so can not use the mac's irq to trigger
> wakeup signal, instead add a dedicated irq to send wake signal which
> stay in always-on power domain.
>
> >  Can you add this to the commit message?
> OK, will add more info
> 
> > 
> > And why you did not add it already in the first commit that added the
> > platform driver there anyway?
> OK.
> because I test base funtion first, then support suspend/resume

It's fine to test it how you like but split the patches logically when you
submit upstream, and this one belongs to the first patch IMHO.



More information about the linux-arm-kernel mailing list