Re: [PATCH v6 3/3] ARM: dts: vf610: add Miscellaneous System Control Module (MSCM)

[Stefan Agner]

On 2015-03-11 01:48, Shawn Guo wrote:
> On Sun, Mar 01, 2015 at 11:41:29PM +0100, Stefan Agner wrote:
>> Add the Miscellaneous System Control Module (MSCM) to the base
>> device tree for Vybrid SoC's. This module contains registers
>> to get information of the individual and current (accessing)
>> CPU. In a second block, there is an interrupt router, which
>> handles the routing of the interrupts between the two CPU cores
>> on VF6xx variants of the SoC. However, also on single core
>> variants the interrupt router needs to be configured in order
>> to receive interrupts on the CPU's interrupt controller. Almost
>> all peripheral interrupts are routed through the router, hence
>> the MSCM module is the default interrupt parent for this SoC.
>>
>> In a earlier commit the interrupt nodes were moved out of the
>> peripheral nodes and specified in the CPU specific vf500.dtsi
>> device tree. This allowed to use the base device tree vfxxx.dtsi
>> also for a Cortex-M4 specific device tree, which uses different
>> interrupt nodes due to the NVIC interrupt controller. However,
>> since the interrupt parent for peripherals is the MSCM module
>> independently which CPU the device tree is used for, we can move
>> the interrupt nodes into the base device tree vfxxx.dtsi again.
>> Depending on which CPU this base device tree will be used with,
>> the correct parent interrupt controller has to be assigned to
>> the MSCM-IR node (GIC or NVIC). The driver takes care of the
>> parent interrupt controller specific needs (interrupt-cells).
>>
>> Acked-by: Marc Zyngier <marc.zyngier@xxxxxxx>
>> Signed-off-by: Stefan Agner <stefan@xxxxxxxx>
> 
> Stefan,
> 
> I guess this patch has a run-time dependency on the first two in the
> series, right?  Or put it another way, if I apply this single patch on
> my branch, the dtb and kernel built from the same branch do not work
> together, right?  If so, we will need to either wait for the first two
> hit mainline or pull Jason's irqchip/vybrid branch into my tree as
> prerequisite (irqchip/vybrid needs to be stable).
> 
> Shawn

Yes, that is true. The driver need to be in place in order to
successfully boot with the new device tree.

--
Stefan

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/