[PATCH v5 2/3] Irqchip/ast2700-intc: add debugfs support for routing/protection display

[Ryan Chen]

> Subject: Re: [PATCH v5 2/3] Irqchip/ast2700-intc: add debugfs support for
> routing/protection display
> 
> On Wed, Oct 22 2025 at 14:55, Ryan Chen wrote:
> 
> The prefix is: irqchip/ast....:

I'll use irqchip/aspeed-intc:

> 
> > AST2700 INTC0/INTC1 nodes ("aspeed,ast2700-intc0/1") not only include
> > the interrupt controller child node ("aspeed,ast2700-intc-ic"), but
> > also provide interrupt routing and register protection features.
> 
> Lacks a new line to open a new paragraph.

I'll split the opening sentence and the follow-up paragraph in
the commit message.
> 
> > Adds debugfs entries for interrupt routing and protection status for
> 
> Add

Will update

> 
> > AST2700 INTC0/INTC1.
> 
> But what you are failing to explain is why this is required and useful. Just
> adding code because we can is not a real good reason.

The INTC0/INTC1 routing and protection registers are set up by early firmware
and are not modified by Linux. 

When routing is off, interrupts get merged/misdirected silently.
A minimal kernel-side visibility helps confirm that the upstream interrupt
topology (e.g., which INTC1 groups are funneled to which INTC0 inputs / GIC SPIs)
matches what the DT and drivers assume.

The protection bits gate which CPU can write/read specific INTC 
windows (PSP/SSP/TSP). When those are mis-programmed, 
Linux can neither steer nor even read status in places it expects to. 
A raw readout lets us verify that Linux isn't boxed out by mistake.
> 
> Thanks,
> 
>         tglx