Re: [ECOS] Real Time Clock

[Gary Thomas <gary@mlbassoc.com>]

On Thu, 2003-10-23 at 05:44, James Yates wrote:
> I have a platform which when boots, copies code that was placed to be run from RAM, from a flash device to the correct location in RAM. The VBR is then modified correctly and a jump is made to the first address contained in the first entry of the vector table.
> 
> The app I have built which I store in flash is a simple eCos app consisting of 2 threads and I am trying to get timeslicing working. I am a bit unsure of the build options I should set since although this is running from RAM it is not running through a debug stub or rom monitor.
> 
> Should I disable all ROM Monitor Support in this case and link it as if it was to be placed in ROM but modify the corresponding memory layouts to actualy place it in RAM? This is a bit of a special scenario for eCos but unfortunately I have to build it this way.
> I seem to get an exception when the scheduler is started, I think the RTC is getting an interrupt but the VSR doesn't have an entry in. So when the interrupt happens, it atempts to jump to a routine that doesn't exist. Provided the HAL_CLOCK_INITIALIZE, HAL_CLOCK_RESET, HAL_CLOCK_VAR_INITIALIZE and HAL_CLOCK_READ macros are valid, should the realtime clock code in the kernel handle everything else for me or do I need to make any special calls to setup the RTC isr?

If you don't have an [eCos compatible] ROM monitor running, then
set your application thusly.  Make sure that CYGSEM_HAL_USE_ROM_MONITOR
is not set and it should take over everything, including all of the
interrupt & exception processing.  This is invariant of the ROM/RAM
startup type setting (although RAM mode often implies USE_ROM_MONITOR)

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss