New l3-noc error with CPUFREQ_DT built-in with v4.0-rc1

[Tony Lindgren]

* Felipe Balbi <balbi at ti.com> [150223 19:19]:
> On Mon, Feb 23, 2015 at 07:01:42PM -0800, Tony Lindgren wrote:
> > * Tony Lindgren <tony at atomide.com> [150223 18:43]:
> > 
> > Looks like the address is 0 for "Custom Error". Anyways, reverting
> 
> yeah, that's because the error comes from l4per2, not l3 :-)

Right so it seems :)
 
> > a fix for similar issue found on omap3 so far seems to help, that's
> > 3d009c8c61f9 ("gpio: omap: Fix bad device access with setup_irq()").
> 
> if we revert that, we regress omap3, right ?

Yes we saw that getting triggered on omap3/4 before 3d009c8c61f9.

Now looking at the stack trace again, it actually has something:

...
[<c05999a4>] (__irq_svc) from [<c0599164>] (_raw_spin_unlock_irqrestore+0x34/0x44)
[<c0599164>] (_raw_spin_unlock_irqrestore) from [<c0027120>] (omap_hwmod_idle+0x34/0x44)
[<c0027120>] (omap_hwmod_idle) from [<c00283f8>] (omap_device_idle+0x38/0x78)
[<c00283f8>] (omap_device_idle) from [<c0028454>] (_od_runtime_suspend+0x1c/0x24)
[<c0028454>] (_od_runtime_suspend) from [<c03b5214>] (__rpm_callback+0x2c/0x60)
[<c03b5214>] (__rpm_callback) from [<c03b5268>] (rpm_callback+0x20/0x80)
[<c03b5268>] (rpm_callback) from [<c03b56b8>] (rpm_suspend+0xe8/0x55c)
[<c03b56b8>] (rpm_suspend) from [<c03b6bdc>] (pm_runtime_work+0x74/0xa8)
[<c03b6bdc>] (pm_runtime_work) from [<c0054608>] (process_one_work+0x1b0/0x4a0)
...

If it's the gpio-omap, there's probably some confusion still in
the driver regarding the GPIO bank idle status. Anyways, will look
more into it tomorrow.

Regards,

Tony