For those who do not know - I am a new commit person for OpenOCD.
I have just committed a large (in size) but small (in scope) patch. The
patch is 300K, I did not attach it.
There may be some "bumps" in the road, they should be sorted out
*QUICKLY*.
If you are reading this and
* you are a "packager" type for others - please don't jump on this
change for a week or two.
* FYI - Those that package openocd - in case you have not heard we are
working towards a 1.0 release.
meanwhile - if you can test something - please do - I've tested all I
could test (see the list below).
-Overview of changes follow -
========================
(A) I have updated the DOCUMENTATION - to reflect all of the changes below.
Yes - There is now updated documentation! Be Happy!
Hopefully soon the WEB version of the docs will be updated.
(I don't know how to do that yet)
I tried to explain many of the new features in depth and in detail.
(Could somebody proof read/spell check/grammar check - I am notorious
please...)
This change alone - made the PATCH file huge.
========================
(B) The primary goal is/was to support "jtag route controllers" aka: JRC.
Example: TI OMAP3530 - (aka: BeagleBoard) has one, as does the STR912
which was "a hack" to get working.
What makes the patch seem large is this:
(1) A common "struct fields" no longer has a member: "int device"
(2) Many functions had an "int chain_pos" parameter (various names)
(3) The integer index for(;;) loops over the chain are gone.
They have been replaced with a "jtag_tap_t *tapptr"
and a new function: "jtag_NextEnabledTap()"
========================
(C) To make the JRC stuff work right the "jtag_device" command is gone.
And was replaced with "jtag newtap"
The old command syntax was:
jtag_device IRLEN IRCAPTURE IRMASK
The new command is:
jtag newtap CHIP TAP -irmask NUMBER -ircapture NUMBER -irmask
NUMBER .... other parameters
This effects almost all ".cfg" files for targets.
========================
(D) Testing - yes, I have done quite a bit.
I also have created quite a few new CONFIG fles
And split the "target" directory into "board" and "target".
I tested these changes with:
CortexM3
ST STM3210E-Eval (stm32 cortexM3)
ST STM32F10x-128k-eval
Arm7TDMI
Sharp LH379532 eval board
Olimex SAM7x256
ARM Evaluator 7T
Arm926
LogicPD IMX27
Arm920t
Atmel AT91RM9200-DK (I do not have the new EK board).
By test: I mean - for all the above boards.
The TAPIDs verify. I can type reset, halt, and single step
through code using OpenOCD.
A few of them, I tested further with GDB.
========================
(E) What appears to be broken is my IAR Kick Start ST-912 eval board.
Actually I think my configuration is broken.
I could not make this work with older versions of OpenOCD.
===
*END*
===
-Duane.
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development
