[kernel 5.4 | potential upstream bug] irq polling for non-existent IRQ77 on PHY
Just for awareness https://lists.openwrt.org/pipermail/openwrt-devel/2020-June/023954.html
device tree
interrupt-parent = <&gpio1>;
interrupts = <14 IRQ_TYPE_LEVEL_LOW>;
does not spawn a IRQ77 child however for phy@1
cat /proc/interrupts | grep 8-0071
76: 9 0 f1018140.gpio 14 Level 8-0071
and leads potentially to the reported output.
With SFP the issue does not exhibit since IRQ77 gets spawned
cat /proc/interrupts | grep 8-0071
76: 9 0 f1018140.gpio 14 Level 8-0071
77: 0 0 8-0071 4 Edge sfp
78: 0 0 8-0071 3 Edge sfp
79: 0 6 8-0071 0 Edge sfp
due to device tree
tx-fault-gpios = <&sfpgpio 0 GPIO_ACTIVE_HIGH>;
tx-disable-gpios = <&sfpgpio 1 GPIO_ACTIVE_HIGH>;
los-gpios = <&sfpgpio 3 GPIO_ACTIVE_HIGH>;
nobody cared (try booting with the "irqpoll" option)
gets printed by https://github.com/torvalds/linux/blob/v5.4/kernel/irq/spurious.c#L201
There are some code changes between 4.14 and 5.4 that might explain why the issue does not exhibit on 4.14 but on 5.4.