ACPI: regression: Failed to initialize GIC IRQ controller

[Ming Lei]

On Fri, Jul 10, 2015 at 4:23 PM, Marc Zyngier <marc.zyngier at arm.com> wrote:
> On 10/07/15 09:17, Suman Tripathi wrote:
>> On Fri, Jul 10, 2015 at 1:28 PM, Marc Zyngier <marc.zyngier at arm.com> wrote:
>>>
>>> On 10/07/15 08:45, Ming Lei wrote:
>>>> Hi,
>>>>
>>>> Commit 0cff8dc0099f6d4f(ACPICA: ACPI 6.0: Add changes for MADT table.)
>>>> causes the following failure on APM mustang board(arm64) when
>>>> booting via UEFI and ACPI:
>>>>
>>>> No valid GICC entries exist
>>>> ACPI: Failed to initialize GIC IRQ controller
>>>> Kernel panic - not syncing: No interrupt controller found.
>>>> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.2.0-rc1+ #45
>>>> Hardware name: APM X-Gene Mustang board (DT)
>>>> Call trace:
>>>> [<ffffffc000089b94>] dump_backtrace+0x0/0x12c
>>>> [<ffffffc000089cd0>] show_stack+0x10/0x1c
>>>> [<ffffffc0005fac18>] dump_stack+0x8c/0xdc
>>>> [<ffffffc0005f7218>] panic+0xe4/0x220
>>>> [<ffffffc00082631c>] init_IRQ+0x24/0x30
>>>> [<ffffffc00082486c>] start_kernel+0x274/0x3d8
>>>> ---[ end Kernel panic - not syncing: No interrupt controller found.
>>>
>>> Isn't that addressed by [1] which Catalin has queued for -rc2?
>>
>>
>> I am also seeing the same . But it fixes after I apply the parking
>> protocol  patch but that is not upstreamed.
>
> I'm more interested to find out if the patch I mentioned in my original
> email fixes it or not. An additional dependency on something that is not
> aimed for mainline yet doesn't really help.

Yes, Al's patches do fix the issue.

Thanks,
Ming