Re: [Mspgcc-users] GoodFET support to mspdebug

[Robert S. <rsp...@ze...>]

Hi Harri,

Harri Haataja wrote:
> Is it possible to add GoodFET support to mspdebug?

Daniel Beer wrote:
> It's certainly possible. I was planning on it, but unfortunately  
> haven't had time so far.

I think implementing support for a new programmer or debugger is as
simple as writing some code with the API described by the device_t
struct in device.h.  The contents of device_t is extremely
self-explanatory (see below).

(This is for Harri's benefit... I'm not trying to tell you about your
own code Daniel ;)

> Is there a document on the protocol available anywhere? 

It appears there is documentation here:
http://goodfet.sourceforge.net/apps/jtag430/

> If it's quite  similar to the TI FET, you could probably just modify
> fet.c.

bsl.c is also a relatively short reference on implementing a
programmer...

Cheers,

Rob

----

struct device {
	/* Breakpoint table. This should not be modified directly.
	 * Instead, you should use the device_setbrk() helper function. This
	 * will set the appropriate flags and ensure that the breakpoint is
	 * reloaded before the next run.
	 */
	int max_breakpoints;
	struct device_breakpoint breakpoints[DEVICE_MAX_BREAKPOINTS];

	/* Close the connection to the device and destroy the driver object */
	void (*destroy)(device_t dev);

	/* Read/write memory */
	int (*readmem)(device_t dev, address_t addr,
		       uint8_t *mem, address_t len);
	int (*writemem)(device_t dev, address_t addr,
			const uint8_t *mem, address_t len);

	/* Erase memory */
	int (*erase)(device_t dev, device_erase_type_t type,
		     address_t address);

	/* Read/write registers */
	int (*getregs)(device_t dev, address_t *regs);
	int (*setregs)(device_t dev, const address_t *regs);

	/* CPU control */
	int (*ctl)(device_t dev, device_ctl_t op);

	/* Wait a little while for the CPU to change state */
	device_status_t (*poll)(device_t dev);
};