Solo tests with 1.1.5 fail

[Eric W. Biederman]

Stefan Reinauer <stepan at suse.de> writes:

> Hi,
> 
> I get reboots again, this time it's not dependent on the gcc version I
> use.. Eric, there are some weird bits in the attached log, can you help
> to interpret them? From what I see some the SuperIO setup has severe
> trouble, and some of the resources seem to be given out twice...

Going through the pieces.

It looks like you are using the NSC pc87360 code from the hdama,
given the solo has a different superio chip this may be part of
the problem.  I am moving to a complete superio implementation and
the difference is chips will likely become more significant.

This bit is very worrisome.  
> Allocating resources...
> PCI: 00:00.0 missing read_resources
> PCI: 00:01.0 missing read_resources
> PCI: 00:00.0 missing read_resources
> PCI: 00:01.0 missing read_resources

You should not have entries in the tree that don't have those
methods.  Nor should you have any devices on bus 0, except
the cpus right now.

It looks like this is probably due to listing the amd8151  
after the amd8111, instead of before in Config.lb

This bit is also worryisome:
> PCI: pci_scan_bus for bus 2
> PNP: 002e.0 [1002/5157] enabled
> PNP: 002e.8 [ffff/ffff] enabled
> PNP: 002e.9 [ffff/ffff] enabled
> PNP: 002e.a [ffff/ffff] enabled

PNP devices should not attempt to print out a pci vendor/device id.

I have appended my boot log for comparison purposes.  The NSC superio
code is currently grumpy because I am not allocating interrupts.
Which I think is what you were picking up on. But no interrupts are
harmless in this context.

I am very puzzled about what is actually triggering the unexpected
resets.  Though poorly setup hardware can't help.  I also don't
see the duplicate resource scenario you were reporting.

Eric