On Tue, Jun 2, 2009 at 8:45 AM, David Brownell <davi...@pacbell.net> wrote: > On Monday 01 June 2009, Ųyvind Harboe wrote: >> - RESET is an illegal end state. It's split out as a separate thing for lots >> of >> reasons, including that this triggers events that listeners are listening >> for, etc. > > That makes sense as a policy. It's not enforced today > in "irscan", "drscan", or "endstate" commands though. > > That seems a bit more on the "post-0.2.0" side of things. > Though I'd not object to taking RESET out of the list > of states that are "safe", so it triggers a warning...
RESET *could* be made a legal state from the users point of view, even if it is handled quite differently inside OpenOCD innards. >> - DRPAUSE makes no sense for IRSCAN as an endstate according to your >> new and improved rules... Ditto for DRSCAN > > Those rules just say "don't do anything known to be broken". But > those transitions are legal, stable, safe -- not broken: > > IRSCAN: ... Update-IR --> Select-DR-Scan > --> Capture-DR --> Exit1-DR --> Pause-DR > DRSCAN: ... Update-DR --> Select-DR-Scan --> Select-IR-Scan > --> Capture-IR --> Exit1-IR --> Pause-IR > > I don't think the API should make a policy decision to prevent > those particular transitions, if some wierdo needs them. > > Let them be safe, legal, and rare ... but possible, without any > extra roadblocks. Folk *do* have TAP engines that respond to > unusual state transitions. (Although maybe all we'd really > need to worry about is the IEEE 1149.7 ZBS ...) How about removing the -endstate option entirely for IRSCAN/DRSCAN and force IRPAUSE/DRPAUSE as endstates? >From there the user can *move* into whatever state he wants... -- Øyvind Harboe Embedded software and hardware consulting services http://consulting.zylin.com _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
