Martin,
    I've found the problem I created regarding the entrypoint thread. The AM33XX has a set of control module registers and pin-muxing registers which require privileged access. To allow these registers to be accessed from user-land, I created two kernel calls to access them and added them to the end of of the call list in kernel.h in core/includes. That was a while ago. Now there appears to be a second list in core_interface.h that extends that list. So, I clobbered the first to calls in core_interface, one of which created a thread and returned an id.

The fix to my exension of the kernel calls gets me through to init being called and passed the available memory resources. I've an issue now that I think I'll be able to find as config appears to trying to access a valid memory location that has no translation table entry.

Thanks for your help.

        Bob


On 08/19/2014 12:36 PM, Bob Stewart wrote:
Thanks for the reply Martin.
Two headlines are indeed printed from git log | grep "get a thread cap in Thread_base constructor" --

    hw: get a thread cap in Thread_base constructor
    hw: get a thread cap in Thread_base constructor

This morning I pulled from the current master, merged, buit, and got the same error output on the entrypoint thread.

Thanks for the info on getting an active thread list. I'll instrument call_access_thread_regs when I get back to that development system.

Thanks for the help,

    Bob


On 08/19/2014 08:42 AM, Martin Stein wrote:
git log | grep "get a thread cap in Thread_base constructor"