panic: arm_unmask_irq [was: Re: TI platforms code update: switching to vendor FDT data]

[Ian Lepore]

On Sun, 2015-05-24 at 17:33 -0700, Marcel Moolenaar wrote:
> > On May 24, 2015, at 5:16 PM, Ian Lepore <ian at freebsd.org> wrote:
> > 
> > Shouldn't the driver attach order be controlled with BUS_PASS numbers
> > now?  That's been required with other socs that moved to the standard
> > dts data where we don't control the order of the nodes in the file.
> 
> The problem is different in my case, but yes it was the first
> thing I suspected without digging deeper:
> 
> fbsdvm64% svn diff sys/arm
> Index: sys/arm/ti/aintc.c
> ===================================================================
> --- sys/arm/ti/aintc.c	(revision 283321)
> +++ sys/arm/ti/aintc.c	(working copy)
> @@ -156,7 +156,8 @@
> 
>  static devclass_t ti_aintc_devclass;
> 
> -DRIVER_MODULE(aintc, simplebus, ti_aintc_driver, ti_aintc_devclass, 0, 0);
> +EARLY_DRIVER_MODULE(aintc, simplebus, ti_aintc_driver, ti_aintc_devclass,
> +    0, 0, BUS_PASS_ORDER_EARLY);
> 
>  int
>  arm_get_next_irq(int last_irq)
> 
> 
> --
> Marcel Moolenaar
> marcel at xcllnt.net
> 

For an interrupt controller that should probably be
BUS_PASS_INTERRUPT + BUS_PASS_ORDER_MIDDLE

We have to do some fine-grained tuning in some cases and there are only
a few predefined categories and they're spaced really close together.

-- Ian