wad at laptop.org
Thu Apr 19 09:17:04 EDT 2007
Today's debugging went even better than yesterday!
Most of the peripherals have been tested and are working.
Yesterday's big problems, talking to the CAFE ASIC and getting
video through the DCON, were both firmware problems and have
The complete suite of firmware diagnostics passes on the board that
we have been testing the longest.
Linux boots -
a) Build 386 works verbatim, except that
1a) There is a slight DCON timing issue when output is going
2a) Sugar comes up, but the DCON screen is messed up -
rectangular regions are in the wrong places
b) With build 386 but the kernel replaced by the latest master RPM,
1b) The lxfb problem cited in 1a) is fixed
2b) X does not start. The log says: Chipset not GEODELX!!!
What is still giving us problems:
- On about half of the boards we are having problems with the
DCON video out
- The camera sometimes generates slightly corrupted data
- One board had two single bit memory errors in twenty-four hours
of memory tests. Two others showed no errors. We will be running
the memory test on a larger number of boards with chips from all
three vendors over the next few days.
What hasn't been tested yet:
- Long term NAND read/write tests (quick tests pass OK, and booting
the OS from NAND works)
wad, Richard, and Mitch
More information about the Devel