On Thursday 25 February 2010, David Brownell wrote: > Here's a works-with-JTAG version of that. I'll hold off > a few days before committing it, to collect comments, but > I don't currently have any reservations about these patches.
Hmm, no comments yet. It's almost as if the folk who know about any part of this haven't had a chance to digest it yet. :) I'll hold off till Monday ... or longer, if some comments come along which need to be addressed. I think Simon has looked longest, and most effectively, at SWD support for OpenOCD. But Spen is also I think quite familiar with the ADIv5 infrastructure, and there are also a few other folk with something more than vague familiarity with it. > 1 Add a new new interface abstracting DAP operations, > and implements the JTAG ops using existing code > > 2 Switch to one of the operations in the new interface, > and make it so the JTAG implementation is no longer > exposed > > 3 Similarly for the two DP operations (register read/write); > this is internal to the ADIv5 code > > 4 And for the AP operations (register r/w) ... this is mostly > but not completely internal to ADIv5 Also ... I'm not sure I said it before, but the abstract DAP operations accessed in patches #3 and #4 are close matches to SWD protocol operations. So for those not very fluent in ADIv5, the way to understand what these patches enable is to imagine those calls just going out over SWD, instead of through JTAG operations using the APACC and DPACC scan chains. - Dave _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
