MMU issues with AM335X and 14.05

[Martin Stein]

Hi Bob,

On 09.08.2014 22:21, Bob Stewart wrote:
> I went back to the 14.05 issues today and found I could get the kernel 
> initialization to complete successfully if I reverted the S bit to 
> "unshared" in the memory attributes in a Section entry create. Prior to 
> 14.05 this bit was set to "unshared" and was presumably changed in 14.05 
> to allow for multiple processors accessing the same memory regions.

We had an issue (https://github.com/genodelabs/genode/issues/1181)
recently that the shared-bit should be set only when using SMP. The
related changes are in the current state of our master branch
(https://github.com/genodelabs/genode/tree/master) but not in version
14.05. Could you please give it a try?

> In addition, after completing kernel initialization, core's "main" 
> function is entered, the info message for creating local services shows 
> up, a translation for the top of RAM (0x80000000) is created, then the 
> message "failed to communicate thread event" occurs and init is never 
> called. Any thoughts on why that message is appearing would be 
> appreciated. It appears to be coming from initialization of the root 
> interfaces.

This seems to be a page fault in core. Normally core should never
trigger a page fault because there's no one to handle it. So the kernel
doesn't know who to inform about it and thus prints this message. To
prevent this situation, memory regions statically needed by core
(program image, MMIO regions) get mapped 1:1 in the 'Core_pd'
constructor in 'base-hw/src/core/kernel/kernel.cc' using, among others,
the platform specific method 'Platform::_core_only_mmio_regions'. I
assume that your port misses a region in this function. You can get
further information about the page fault by printing things like
'_fault_addr', '_fault_writes', 'char const * Thread::label()', and
'unsigned long Thread::ip' in the method 'Thread::_mmu_exception()' in
'base-hw/src/core/arm/cpu_support.cc' right before '_fault.submit()'.

Cheers
Martin