On 08/31/18 16:01, Bjorn Andersson wrote:
> On Fri 31 Aug 15:46 PDT 2018, Frank Rowand wrote:
>>> + spmi_bus: spmi@fc4cf000 {
>>> + compatible = "qcom,spmi-pmic-arb";
>>> + reg-names = "core", "intr", "cnfg";
>>> + reg = <0xfc4cf000 0x1000>,
>>> + <0xfc4cb000 0x1000>,
>>> + <0xfc4ca000 0x1000>;
>>> + interrupt-names = "periph_irq";
>>
>>> + interrupts = <0 190 0>;
>>
>> The final value in this interrupts property means IRQ_TYPE_NONE.
>>
>> A WARN_ON() was added early this year to complain about use of
>> IRQ_TYPE_NONE: 83a86fbb5b56 "irqchip/gic: Loudly complain about
>> the use of IRQ_TYPE_NONE", resulting in many warnings spewing
>> forth when I boot an APQ8074 Dragonboard. I am trying to
>> determine whether the warning is overly aggressive, or whether
>> the IRQ TYPE is incorrectly specified for the spmi node.
>>
>> The interrupt-parent for the spmi node is intc: interrupt-controller@f9000000,
>> which has compatible = "qcom,msm-qgic2". I do not know the architecture
>> or implementation of this interrupt controller. Is an IRQ_TYPE_NONE
>> valid in this case, or should a specific type be provided?
>>
>
> No, IRQ_TYPE_NONE isn't valid and the WARN_ON() is reasonable.
>
> Please change it to IRQ_TYPE_LEVEL_HIGH. And while you're at it, replace
> the first 0 with GIC_SPI.
>
>
> If you have more of these warnings you can most likely look at e.g.
> msm8916 (arm64) to find the right flags.
Thanks, I'll create a patch.
