MMCCAM - Re: onboard wireless on rpi4

[Andriy Gapon]

On 07/09/2020 19:18, Bjoern A. Zeeb wrote:
> On 7 Sep 2020, at 15:21, bob prohaska wrote:
> 
>> On Mon, Sep 07, 2020 at 03:14:28PM +0200, Ronald Klop wrote:
>>>
>>> I've MMCCAM running since 12.1-BETA1 on rpi3B+. No problems so far.
>>> Apparently I'm to only one :-)
>>> https://dmesgd.nycbug.org/index.cgi?do=index&fts=mmccam
>>>
>>
>> I just gave it a spin on a Pi3B (not +) and was rewarded with a
>> panic. The machine boots from a USB mechanical disk. Sources are
>> at 365355. Here's the console output:
>>
>> http://www.zefox.net/~fbsd/rpi3/GENERIC-MMCCAM/console
>>
>> I'll give the corresponding vanilla GENERIC kernel a try shortly, but 364900
>> boots without trouble.
> 
> This seems to be a FDT/loader thing;  others will know more about what it means.

Bjoern, I think that you didn't scroll down enough.
The real panic is down there.

It looks like a page fault or something like that in sddastart.
Perhaps a stack overflow?

Kernel page fault with the following non-sleepable locks held:
...
#2 0xffff00000073d35c at data_abort+0xe4
#3 0xffff00000073cd38 at do_el1h_sync+0x14c
#4 0xffff000000724078 at handle_el1h_sync+0x78
#5 0xffff0000000f7d68 at sddastart+0xa8

Bob, I noticed this:
(sdda0:sdhci_slot0:0:0:0): Partition  0 -> 8265462

Perhaps r365445 might be of some help.

P.S.
I found this article about ARMv8 and handle_el1h_sync:
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/semihalf-arm-blog-2-dead-board-and-stack-growth
It has some theory but not sure if it help in practice.

-- 
Andriy Gapon