On Thursday 01 October 2009, Magnus Lundin wrote: > > For testing and comments
As for the latter ... this looks sufficiently generic that I don't see why it should even be needed. IMO the core code *should* do all of this without needing a method! The fact that cache validation *isn't* automatic makes some problems, for example. Getting a generic hook to do that stuff -- and for all the registers on a target, not just that handful -- ought to be uncomplicated. On the flip side, the SRST stuff ... clearly is a NOP for most Cortex-A8 parts. I do seem to have *one* board with a JTAG connector that wires it up; but I don't have a level-shifter to plug into that connector, sigh... this part gets complicated. It's where I had suggested we have an escape to a new TCL event handler, which would be supplied for use instead of SRST on boards which need it (e.g. they use a TI-14 JTAG connector). - Dave _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
