Started by pentium, September 15, 2008, 11:26:02 pm
Quote from: "barcher174"Good news, I've managed to get a part number off one of the remaining chips.
Quote from: "barcher174"Hi allI have some more debug I want to do on my side, but if I'm unsuccessful I'll pass the ball along. --Brian
Quote from: "gtnicol"There are multiple versions of the DayDream... I assume they're all pretty much the same. I have another one or two here...
Quote from: "gtnicol"My one is rev 1.34 and both those chips have been sanded to remove the identifier.
Quote from: "barcher174"As long as there are multiple people working on this, let me share what I know. The daydream is using the UART on the DSP port to communicate. I think the only reason the DSP port is used is the addition of 12V to power the device. Otherwise they probably would have just used a standard serial port. The pinout is as follows: Pin 1 SCK Pin 9 GND Pin 2 SRD Pin 10 GND Pin 3 STD Pin 11 GND Pin 4 SCLK Pin 12 SC2 --> Pin 5 RxD Pin 13 SC1 --> Pin 6 TxD Pin 14 SC0 --> Pin 7 +12V, .5A Pin 15 GND Pin 8 -12V, .1A The daydream appears to be a "dumb" device. As far as I can tell it waits for predetermined sequence from the daydream kernel over serial and then begins dumping the rom row by row to be read into memory. I have been exploring this handshake using a USB serial adapter as I documented in another thread. You can set a boot loop by using 'sd daydream' as your default boot device. When you do this you get a repeating pattern in a serial console. Example at 115200:xx~x~xxx~xNxpx@x~x>xNxxpx0x@xx~x~~x>x>~xNxN~xx~xpxp~x0x0~x@x@~xxx~x~~x~~x~xx>x>~x>~x>xxNxN~xN~xNxxx~x~xxxpxp~xp~xpxx0x0~x0~x0xx@x@~x@~x@xxx~xxx~x~~x~~x~xx~x~cx~xx~`xx~x~xxxxcx|x`xfxf~xf~xfxxfxfgxf|xf`xx~x~xxxxcx|x`xxxx~xx~xxxxxxxcxxxxx`xx~x~xxxxcx|x`x`x`~x`~x`xx`x`cx`|x``xx~x~xxxxqxx8q8It of course is not in ASCII, so I've been looking for a serial debugger I can use on the Mac (probably going to need to install windows). After I have the hex sequence to send to the device, I will hook up the device to a serial console and record the output. I'm pretty sure we then just need to feed this back when prompted by the daydream kernel. There is then no need to reverse engineer the actual hardware. Anyone can replicate what I've done so far without the daydream hardware. If you're able to get the output from the next into a serial debugger, please contact me. Would this NeXT dev tool help its in the archives :http://www.nextcomputers.org/NeXTfiles/Software/NEXTSTEP/Developer/Tools/Gtools/HexEditor.READMEBest regards Rob Blessin--Brian