PSIM and MP tests -- anybody got working?

[Joel Sherrill]

runtest seems to be able to drive powerpc-rtems-run but I'm not having
any luck with getting a device tree file that gdb likes.  Any luck on your
side?


Karel Gardas wrote:

> On Mon, 23 Jan 2006, Joel Sherrill wrote:
>
>> Karel Gardas wrote:
>>
>>> On Mon, 23 Jan 2006, Joel Sherrill wrote:
>>>
>>>> I haven't tested this in a while so will try to get to it today on 
>>>> at least 4.6.
>>>
>>>
>>>
>>> HEAD runs, at least I got something from it. Anyway, something is 
>>> broken either in simulator or in RTEMS since all (just 3 tests run 
>>> so far) MP-related outputs include this:
>>>
>>> IRQ vector 32 already connected
>>> Unable to initialize system clock
>>> rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 1
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .rtems-4.6.99.2(PowerPC/PowerPC 603e/psim) PANIC ERROR 12
>>> .
>>>
>>> before actual output is printed.
>>>
>>> Any idea what's going wrong here?
>>>
>> I was afraid of this.  shmdr has not been used with new execption model
>> and assumes it can "wrap" the clock tick driver ISR.  It was written 
>> before
>> the Timer Manager exists.  Instead of its clock tick ISR trick, it 
>> should install
>> an RTEMS timer to run once a tick when polled.
>
>
> I assume you are refering to c/src/libchip/shmdr directory and poll.c 
> and setckvec.c files there, right?
>
> BTW: Is there any platform other than PSIM used to test this
> functionality? What about x86/SMP, which we can emulate in Qemu? i.e. 
> while fixing this I'd like to see if I break something or not...
>
> Thanks,
> Karel
> -- 
> Karel Gardas                  kgardas at objectsecurity.com
> ObjectSecurity Ltd.           http://www.objectsecurity.com