[PATCH 3.17-rc4 v7 0/6] arm: Implement arch_trigger_all_cpu_backtrace

[Daniel Thompson]

On 15/10/14 00:31, Russell King - ARM Linux wrote:
> On Tue, Oct 14, 2014 at 04:37:51PM -0600, Daniel Drake wrote:
>> Hi,
>>
>> Thanks a lot for working on this!
>>
>> On Wed, Sep 17, 2014 at 10:10 AM, Daniel Thompson
>> <daniel.thompson at linaro.org> wrote:
>>> Changes *before* v1:
>>>
>>> * This patchset is a hugely cut-down successor to "[PATCH v11 00/19]
>>>   arm: KGDB NMI/FIQ support". Thanks to Thomas Gleixner for suggesting
>>>   the new structure. For historic details see:
>>>         https://lkml.org/lkml/2014/9/2/227
>>
>> What's the right way to extend your work in order to get a NMI-like
>> watchdog hard lockup detector similar to the one on x86?
>>
>> I'm testing your patches on Exynos4412 and I guess in their current
>> state they don't go quite this deep, as the only callers of
>> trigger_all_cpu_backtrace() are sysrq, hung_task and spinlock debug
>> code - none of which seem as fail-safe as a trigger like a
>> pre-programmed watchdog NMI interrupt would be.
>>
>> Do I need to find a way to get CONFIG_FIQ available on this platform
>> first? and/or CONFIG_HARDLOCKUP_DETECTOR?
> 
> The blocker on this work right now is the annoying Versatile Express
> platform, which pretty much means that we currently can't push the
> code into the GIC to support FIQs.  As long as adding FIQ support to
> the GIC results in the Versatile Express becoming non-bootable, the
> idea of using FIQs is a total non-starter.
> 
> Or we decide that we dump the platform completely (which will upset
> a number of developers.)
> 
> I have patches I'm using for trigger_all_cpu_backtrace() which I'm
> maintaining privately in my tree until we can get the FIQ situation
> sorted.

I do hope to gain (remote) access to a vexpress at some point just to
pick at this issue a little.

That said your previous description of the issue and of the GIC version
register leaves very little to explore!

In the end I'm expecting to have to use some kind of black list logic. I
assuming that vexpress by its nature as a bring up platform is more
likely to exhibit problems in this sort of area and therefore a black
list is more appropriate than a white list.


Daniel.