Does somebody know the destiny of this item?

[Till Straumann]

Joel Sherrill wrote:
> Leon Pollak wrote:
> 
>> Hello.
>>     One month ago Till Straumann wrote about the bug in 
>> 'new_exception_processing' code. Where there some actions about this 
>> since that time or the problem exists in MPC860 BSPs?
>>
> 
> The fix is in the libbsp/powerpc/shared code plus some tinkering in the 
> score so I think the
> MPC860 BSPs are fixed.

I fixed all libbsp/powerpc/xxx  BSPs which are using new exception
processing, not only the 'shared' one - although I could test only
the 'shared' BSP. However, there are no real differences between the
affected files. Ideally, that code would be inherited rather than
copied...

Note that I also inserted a few lines of paranoia code into 'score'
(I believe) - so that an unfixed new-exception BSP will panic.

-- Till

> 
> --joel
> 
>>     Thanks ahead.
>>
>> ------------------------------------------------------------------------------- 
>>
>> On Thursday 10 July 2003 21:00, Till Straumann wrote:
>> I recently experienced corruption of the 'tick' queue
>> on my PPC. I finally found out one of the reasons:
>>
>> _ISR_Nest_level
>>
>> on which the watchdog timer queues rely
>> is still not maintained on 'new_exception_processing' style
>> PPC BSPs other than mpc8260ads!!!
>>
>> PR288 never made it to other PPC BSPs.
>>
>> watch out
>>
>> -- Till
>>
>> As a quick workaround, you can delete the line
>>
>> _ISR_Flash()
>>
>> in the file cpukit/score/src/watchdoginsert.c
>> (your IRQ latency will go up)
>>
>> (old_exception_processing BSPs are not flawed)
>>
>> T.
>>
>>
>>  
>>
> 
> 
>