John McCarthy wrote: > Also, there are some older AMD flashes (e.g 5v) that apparently don't > support this programming mode which I expect will break with this patch. > I don't see any easy way to detect these flash types except by CFI id. > Any suggestions on how to handle this in my patch? > > The simplest fix would be to add another option to the cfi flash type to > enable or disable this feature, leaving it up to the end user to figure > out if their flash supports UnlockBypass mode. > > Messier would be to build a table of CFI id's with/without this mode. > I don't have the info to build this table except for the few flash > devices I've got data sheets for.
I definitely think a table that whitelists only those IDs where we know the device supports the bypass mode is in order - there are too many devices that will not work otherwise. Using a whitelist, a missing entry will lead to non-optimal performance, but nothing will break. (BTW, I think you are meaning Jedec IDs, not CFI IDs, right?). Also, I would like to limit parameters in the config file to those absolutely needed - ie. the user should not need to configure things that can be auto-detected. I regularly work with boards where one of a series of allowed flashs may be used as alternatives, and is is annoying if I need to edit the config file every time I switch between two boards of the same kind. cu Michael _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
