On Friday 25 September 2009, Dirk Behme wrote: > > Patch in attachment results in errors:
> -jtag configure $_CHIPNAME.jrc -event post-reset "runtest 100" > +jtag configure $_CHIPNAME.jrc -event post-reset \ > + "runtest 100; omap3_dbginit" Yeah, annoyingly enough post-reset fires four times during normal startup ... at least the first few are not safe to do more than the "runtest", because the scan chain isn't yet validated. The runtest needs (AFAICT) to be issued about that early, although maybe only onece ... to make sure the ICEpick gets powered up. Then sometimes after the scan chain is validated, the guts of the dbginit need to be issued. It needs a new TAP-level event, since it must talk to the ICEpick not to a (disabled) ARM target. .... or so I understand things. At the moment I'm poking at that startup processing since it's not behaving right for me on OMAP5912 or DaVinci DM355. That seems like a regression. Once it behaves again, adding such a new event will be easy enough. - Dave _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
