Lpc2000

At this point, I just use the Philips Flash Utility to get the code 
into the part, and load the debug symbols using axd. I have used axd 
to debug out of flash for years...no problem.

--- In lpc2000@yahoogroups.com, "Amit" <emailakb@y...> wrote:
> 
> How are you debugging from Flash using AxD? 
> Have you implemented the Flash programming algo for the LPC in AxD?
> 
> I was under the impression that the Flash programming interface 
has 
> to be implemented for the LPC by customers.
> 
> amit
> 
> --- In lpc2000@yahoogroups.com, "tah2k" <tah2k@y...> wrote:
> > 
> > Still not up and running with my Arm tools and the Keil MCB2130.
> > 
> > I setup the Arm AXD debugger to run the following script on 
startup:
> > 
> > loadsymbols C:\Projects\Sampler_2138\AppFlash\Output\AppFlash.axf
> > setpc 0x0
> > sreg cpsr 0xd3
> > spp vector_catch 0
> > spp semihosting_enabled 0
> > 
> > It basically loads the debug symbols for the image in the 2138 
> > flash, initializes the CPSR, sets the pc to 0, and disables 
> > breakpoints at the Arm exceptions.
> > 
> > The debugger loads the source file correctly and the PC is set 
at 
> > 0x0 which is a branch to my startup handler at 0x20. However, if 
I 
> > run, or singlestep, or do anything, I get a "Data abort".
> > 
> > Also, if I try and display memory starting at 0x0, the debugger 
> > gives an error "Memory access failed" and only gives the 
contents 
> of 
> > addresses 0-20, the rest 0's.
> > 
> > I thought perhaps the data abort had something to do with the 
> > remapping done by the Philips boot loader, but I checked the 
MEMMAP 
> > flag and it is set to 1, so I am running the flash at 0. 
> > 
> > I'm new to the part and evaluating it against the AT91 
pocessors, 
> > and I am fresh out of ideas. The memory bus is internal so its 
not 
> > like I can hook up my logic analyzer too get some ideas.
> > 
> > Is it possible the Philips memory accelerator is partly to 
blame? 
> > Tehre must be otehrs using Arm tools with this part so I am 
hoping 
> > it is something trivial.
> > 
> > Any ideas are appreciated.
> > Tim