Re: [PATCH v8 4/7] xen/arm: Add virtual GICv3 support

[Julien Grall <julien.grall@linaro.org>]

On 07/28/2014 04:26 PM, Ian Campbell wrote:
> On Mon, 2014-07-28 at 16:08 +0100, Julien Grall wrote:
>> On 07/28/2014 03:55 PM, Ian Campbell wrote:
>>> On Mon, 2014-07-28 at 15:35 +0100, Ian Campbell wrote:
>>>> FYI I'm seeing a trap in gicv3_init when running on a model. I haven't
>>>> tracked it down to an actual location yet though. I'll let you know what
>>>> I find.
>>>
>>> The access to ICC_BPR1_EL1 in gicv3_cpu_init is faulting for some
>>> reason. do_hyp_trap is then access current-> but it is too soon to do
>>> that, hence a data abort occurs which is what I was seeing.
>>
>> In which "case" does Xen end up? The sysreg one?
> 
> Yes.
> 
>> If so, it looks wrong to me that an error who come from the hypervisor
>> is using the current variable.
> 
> Correct, although it's an unexpected occurrence in the normal case.
> 
> If we intended to "probe" for system registers then we'd need to handle
> this case explicitly. As it stands it just lead to a rather odd failure
> mode. We could perhaps add some debugging checks on regs->cpsr mode.

I think it's a good idea to add debugging checks. It would avoid
spending time to understand what's going.

But it's not part of this series.

Regards,

-- 
Julien Grall