The solution for deinitialization is trivial - fix generic ft22332_quit() to match general needs of well designed interfaces (i.e. putting Hi-Z on FT chip is enough to cut it off electrically from target / produce Hi-Z on cable pinout). If any other dongle needs some specific deinitialization define its own quit() in layout. Generic ft2232_quit() (or its wrapper) can chceck if there is layout_quit() defined and if so runs it, if not run further standard ft2232_quit().
There seems to be general misunderstanding/disagreement among us about how the interface should behave on quit.. In my opinion (and Laurent I think) it should "dissapear" from target as it was not connected, Michael wants it to stay as it was on quit (i.e. keep target reset). This is the standard I am talking about - one common behavior for all interfaces that we should define. Why I want to happen this way? Because I want target to behave as usual when I connect my cable (i.e. no reset). So I want it to behave as usual when I end up working with it. If I want to stay in some specific state impacting target, I can simply leave OpenOCD session active... This is sensible and intuitive. -- CeDeROM, SQ7MHZ, http://www.tomek.cedro.info _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
