Jeff,
I think we all need to take a deep breath for a second. The project is
managed in SVN, there can be a revert of changes to a different revision
and Zach or some one else can repeat the warning removal process.
Warning removal is much more trackable than the stuff you have
accomplished. I think that is why Øyvind requested a patch against
your version of OpenOCD. It may be salvageable if there is a rollback
of current work. All is not lost.
// Dean
Jeff Williams wrote:
If MC1322x standard then good luck getting it to work. Trust me, this
is not a trivial matter. It took me almost a month to figure it out,
and this is after smart, seasoned people had already broken their
picks on it.
But there's no point in debating - the fact is that increased
efficiency, support for a new IC, stability improvements to JLink,
better alignment with ARM technical documentation, and a significant
upgrade to state and system debugging are now all in pieces, and there
is no resource to recover it. Those changes, a month's work, are lost
for good. What's to debate? The group decided they didn't like/want
my contribution, and now I'm moving on. I don't need to be convinced
that it was a good choice, because it's immaterial. To my mind there
were at least 10 different ways to manage this, but fundamentally the
group is happy with the status quo and I regret having wasted my time.
Jeff
El Apr 21, 2009, a las 09:53 , Magnus Lundin escribió:
Hi
Yes there are significant changes, I can not comment about the
improvements in efficency. The BIG problem is that it broke existing
code, mainly the ftdi2232 interface. This is not perfect by any
means
and needs work, but it is a main workhorse in OpenOCD. Duane has put
work into getting the patch to work with this but has not been
succesful. So IMHO this patch was not ready for inclusion. I think
Zacks
work is very important, it breaks some pending patches, even for me in
other ares, but it also creates a better foundation to work from.
As to the MC1322x support, I have read a lot of the list
communications
going back to march and also checked the Freescale manual. It seems
MC1322x should be a fairly standard ARM7TDMI-S. So MC1322x should
work
with OpenOCD with perhaps some configuration and startup tweaks. The
problem is the JLink interface, it is known to never have worked
100% .
So my suggestion for going forward is:
- Zack keeps cleaning up the code.
- Test and debug MC1322x support using a FT2232 based interface.
- Test and debug JLink code on target we have lots of experience with
- Update the jag/tap code, but this can only be done if it works with
all already supported interfaces.
Best regards,
Magnus
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
