> > Our ZY1000 product uses OpenOCD like a library and builds on it. > > We could certainly build on a library that catered for non-MCU > devices too :-) > > Could you be a bit more specific as to what you imagine this would > entail on either end? >
If you are talking about openocd vs. urjtag, I can't really comment on the code, as I haven't done much development. If you are talking about non-mcu devices, I think what most people would use is something that will take a .hex or .elf file and squirt it over the jtag but with I2C or SPI EEPROM or NAND/NOR flash programming algoritms instead of jtag, I sort of thought there was some support for that already, I don't have any such devices. Just thought it was an obvious convergence. If you are talking about the lib functions, I'd just like to see a simple (and maybe this is impossible) command-line that will flash an lpc2000 over jtag without starting up the daemon (currently I'm stuck always calling my programs main.c, because main.elf is in my script file) - something canonical over the command line would be more applicable to calling something along the lines of: openocd -interface ft2232 -ft2232_layout "olimex-jtag" -reset_config none -jtag_device 4 0x1 0xf 0xe -target create target0 arm7tdmi -endian little -chain-position 0 -variant arm7tdmi-s_r4 -flash bank lpc2000 0x0 0x7d000 0 0 0 lpc2000_v2 14745 -arm7_9 dcc_downloads enable -flash probe 0 -flash auto_erase on -flash write_image $(TARGET).elf No idea if command lines are still limited to 255 chars, no doubt that would put a kink in my idea. I think the main problems with this working smoothly for me is that I'm a hardware jock origonally, so I don't have TRST or SRST hooked up in any of my jtag headers, and there's been alot of keyword/architecture changes since the examples folder was updated. Rick was helping me update things, but the serial bootloader on the lpc's works well enough I haven't had any pressure to figure it out yet I think the current setup makes good sense for debugging, which is the main use of openocd. I try to avoid debugging as a matter of routine, however. All the 8-bit micros were pretty much useless with debugging realtime anyway. Maybe the trace stuff on the arm7 makes that untrue... Best, Steve
_______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development