Jeff Williams wrote:
> Hi Duane,
>
> I've attached a patch against svn trunk revision 1405.
Thanks - I and others will have a look.
> I am new to ARM in the last few weeks, and I only have a Segger JLink
> since this is what comes with Freescale's DSK. I think this is going
> to be a crawl, walk, run deal where the crawl is just getting up on
> the silicon at all. Next big step is to work on flashing the
> SPI-based serial flash that comes with it. Tho' I think many of my
> changes will make the MCu more stable on any of the p-converters.
There are TWO methods:
(A) programing a spi flash - like Xilinx does with their dongle, for use
with an FPGA configuration.
(B) Download a small "helper program" to the target, and then
downloading the data to that helper program.
Which do you plan on doing?
To date, OpenOCD has never done (A) - I'm planning on doing exactly that
fairly soon.
Method (B) has not to date been done with SPI flash, it has been done
with standard flash of all types. See the "run algorithim" routines as
examples. The other primary reason why (B) has not been done, is that
every single chip has a different SPI flash controller, and every board
has yet another SPI flash configuration {Did you use GPIO for the CS
pin, and which one?}.
If you plan on (B) please think of the host side operation in "generic
terms", as a semi-spi master. I have some ideas, but will not be back
online till tonight.
> This patch might be a bit controversial around one of the changes I
> made, but I think it's right since it brings OpenOCD in-line with the
> actual source documents from ARM. I also sort of made it a bit easier
> to grok state transitions, and that might be a little controversial as
> well.
Sounds interesting. Thanks
> Jeff
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
