Comment # 3 on bug 1212833 from Anthony Iliopoulos

looks like irqbalance reading /proc/interrupts is causing a null ptr deref at
000000000000005a

(the rcu stall warnings is a side-effect, due to irqbalance crashing while
being in a rcu reader critical section, and can be ignored).

show_interrupts+0x24c corresponds kernel/irq/proc:529:

> seq_printf(p, "  %s", action->name);

and that implies that action->next was pointing to 0xa.

Normally action->next is set when the irq line is shared, not sure what may be
corrupting it here.

You could perhaps add the following boot params:

> kprobe_event="p request_threaded_irq irq=$arg1 handler=$arg2 thread_fn=$arg3 devname=+0($arg5):string dev_id=$arg6;p free_irq irq=$arg1 dev_id=$arg2;p __setup_irq irq=$arg1 desc=$arg2 new=$arg3 desc_action=+160($arg2)" ftrace_dump_on_oops trace_options=stacktrace

(append the above to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub and run
update-grub)

after rebooting (and assuming the issue is reproduced), please attach the
ftrace output (cat /sys/kernel/tracing/trace or from dmesg logs)