On 29 August 2011 02:00, Andrew Leech <[email protected]> wrote:
> Hi Simon,
>

Hello,

>
> Actually, your problem may be directly  caused by the delay settings in the
> file:
> jtag_nsrst_delay 1000
> jtag_ntrst_delay 0
>
> try setting both to 500, and bigger/smaller settings. The delay required
> will be dependent on internal chip delays as well as speed of any external
> reset hardware (I've just got caps with pullups on the my reset lines, not
> even any proper supervisor chips - extreme cost cutting).
>

I had similar thoughts before posting and played around with these
settings, but without any change in behaviour.

Following your comments above, I've gone back and experimented further
with a wider range of values from 0 to 2 seconds.

I dropped jtag_nsrst_delay down to 0 because I wanted to see if I got
a JTAG communications failure due to the JTAG clock been stopped on
the LPC3131 immediately after reset, and sure enough, that is exactly
what happened. It would seem therefore that the halt not been
successfully signalled to OpenOCD is not as a result of a JTAG
communications failure as such as I don't see any JTAG communications
failure error messages when jtag_nsrst_delay is set to a reasonable
value.

However, regardless of the delay settings, OpenOCD does not see the
halt, although it does appear to actually happen on the board as the
onboard test blinker does stop running.

> For your reference, my init scrip currently used (a3p060_lpc3131_jtagkey.cfg
> - not in openocd distro):
> Note: I have an fpga in my jtag chain as well as the lpc3131
>

[snip settings]

>
> All the extra setting of pc and arm core mode may be more than required in
> places, but this had given me a perfectly reliable connection sequence in
> eclipse. Earlier with less/different commands it would sometimes connect,
> other times have halt timed out etc.
>

Thanks. It's useful to see what you have needed to set in order to
have a reliable debugging session.

> Good Luck,
> Andrew Leech
>

Thank you for posting your suggestions.

Given I have a workaround for this problem and a setup which allows me
to get on with the rest of the project, I think I am going to use it
for now. If I look into this again in the future (I don't like
unsolved mysteries :-) ) and find anything (ie: a OpenOCD or
configuration issue) I will post it to the list.

Simon.

-- 
Simon Clubley
[email protected]
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to