Re: [PATCH 1/4] clocksource: Add brcm,bcm6345-timer device tree binding

[Jonas Gorski]

On Sat, Nov 21, 2015 at 8:02 PM, Simon Arlott <simon@xxxxxxxxxxx> wrote:
> Add device tree binding for the BCM6345 timer. This is required for the
> BCM6345 watchdog which needs to respond to one of the timer interrupts.
>
> Signed-off-by: Simon Arlott <simon@xxxxxxxxxxx>
> ---
>  .../bindings/timer/brcm,bcm6345-timer.txt          | 57 ++++++++++++++++++++++
>  1 file changed, 57 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/timer/brcm,bcm6345-timer.txt
>
> diff --git a/Documentation/devicetree/bindings/timer/brcm,bcm6345-timer.txt b/Documentation/devicetree/bindings/timer/brcm,bcm6345-timer.txt
> new file mode 100644
> index 0000000..2593907
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/timer/brcm,bcm6345-timer.txt
> @@ -0,0 +1,57 @@
> +Broadcom BCM6345 Timer
> +
> +This block is a timer that is connected to one interrupt on the main interrupt
> +controller and functions as a programmable interrupt controller for timer events.
> +
> +- 3 to 4 independent timers with their own maskable level interrupt bit (but not
> +  per CPU because there is only one parent interrupt and the timers share it)
> +

This is true for the 6345 one but not the 6318/63148/63381 one, which
has one interrupt per timer (+ one extra watchdog interrupt on
6318/63148), and the parent interrupt controller supports affinity. So
you could e.g. route the timer 0 irq to cpu 0 and timer1 irq on cpu 1.

> +- 1 watchdog timer with an unmaskable level interrupt
> +
> +- Contains one enable/status word pair
> +
> +- No atomic set/clear operations
> +
> +The lack of per CPU ability of timers makes them unusable as a set of
> +clockevent devices, otherwise they could be attached to the remaining
> +interrupts.
> +
> +The BCM6318 also has a separate interrupt for every timer except the watchdog.
> +
> +Required properties:
> +
> +- compatible: should be "brcm,bcm<soc>-timer", "brcm,bcm6345-timer"

Since bcm6318 uses a slightly different register layout than the
earlier SoCs, I'd argue that using bcm6345-timer as a compatible for
bcm6318 is wrong.


Jonas
--
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/