On 2017/02/24 17:21, GREen MOnster Games wrote: > > There is a lot more to supporting a computer architecture than just the > > actual CPU. > > Yes It's the start to begin, but It's not all, of course. > > > Yes, the cpu must of course have support, but all the chips > > around it for disk,console,net and how to find where in RAM they are mapped > > and so on is equally important, so if your device isn't on this list, > > I have all about the hw information, sdram 64 mb with 48.9 mb user > accessible, eeprom nor, vfat16 for the memory, consolle too. Probabily > I can try to connect the usb to make a cross compiling into them, I > think it's the better solution, maybe with Uboot or Redboot to start. > I have the problem to see how mbyte can I flashing into the eeprom, > 128 mbyte non volatile with 51 mb free user storage, but those are not > all to flashing. The big problem are the firmware, armish has the > instructions for ARMv4T but the wifi and videocard firwmare I think > it's a proble. > > eeprom: I think are 77 mb dedicated to flashing (NOR) > ram: sdram 64 mbyte with first vector 0x00000000 > > Graphical Controller: ATI Imageon 2282 ( Maybe I can modify the the > file config from xenocara) > Audio Controller: Asahi Kasei AK4641 ( I don't know if it is supported) > > Wifi controller: unknown IEEE 802.11b, IEEE 802.11g (????) > > > > > https://www.openbsd.org/armish.html, then chances are basically zero that it > > will work. >
armish is XScale 80219, ARMv5TE. Newer and faster than this, but even so, a full ports build on the thecus n2100 armish box took around 3 months when I last did one. Take a look at the commit log in the armv7 directories to get some kind of idea of the work involved in bringing up an architecture based on an existing similar-ish one. Though what you don't see there is the amount of work actually involved in reading specs, other code, reverse engineering, talking, debugging that goes into writing each diff. The developers who can do that are mostly looking at armv7 and particularly arm64 at the moment, where there is still a lot to be done. Honestly I don't think any are going to put in time to support a machine which is less useful than the zaurus unless there's something they find particularly interesting about it..
