Dear developers,
I'm using str9comstick and, having to debug some problems, I would like to
know the following things:
- nSRST only resets the core or also resets the embeddedICE (e.g. clear all
breakpoints)?
- the reset halt sequence can be done in this way?
- reset all (nSRST and nTRST asserted)
- run only the embeddedICE (nTRST de-asserted)
- set a breakpoint at location 0x00000000 (or similar request to
embeddedICE in order to stop the CPU before any instruction execution)
- release the CPU from reset (nSRST de-asserted)
- the CPU halts before executing the first instruction
My program (I will eventually post it) runs only if I do this:
- jtag_reset 1 1
- jtag_reset 0 0
or
- jtag_reset 1 0
If I do this:
- reset (halt)
- resume
it stops at a breakpoint that I haven't setted and from that moment the JTAG
seems not work (unless a new reset).
This sound like OpenOCD problem (that's the reason why I posted it here) but
I'm not sure of that.
For those interested I modified OpenOCD source code a little to get it
compile with Visual C++ (VC6).
I also added breakpoint save/restore/list capability.
I modified the telnet server in order to manage F1 - F12 keys, each running
a F1.script - F12.script files. This is usefull to quickly run configurable
commands/scripts using keyboard (e.g. halt/reset/resume/load
image/erase_all/...).
I have also started an AGDI driver to interface Keil with OpenOCD/gdbserver:
I can actually debug at source level bu it is not yet stable (I will release
it to opensource)
Thank you,
Dario
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development
