[E3-hacking] [E2 monitor] v0.25

Matt Evans matt at axio.ms
Thu Jul 7 01:52:52 BST 2005


OK I've /finally/ scraped this off my harddisc and uploaded it...


The main changes are improvements/bugfixes to the XMODEM code 
(particularly in cancelling the transfer if an error is detected), 
avoiding using the wrong signed type which makes stuff break every 256 
blocks... and some much nicer exception debug output.  (The register 
dump provided when an exception is taken should now be accurate for all 
running modes.)

A couple of people have reported 115200 being flaky for them; the baud 
rate divisor could be 1 off for them.  It's spot on for my cable - but 
then my cable could just be clocking itself a little slower than normal 
;)  In uart.S:setup_uart try using 27 instead of 26 for the divisor if 
you are having problems at 115200.

No 'fire' optimisation yet ;) -- Does anyone have any explanations for 
why the pixels are in the strange order in memory?  (They are not in 
increasing-nybble address, or big-endian nybble address in little 
endian words.. etc.  See cons.c)

Also does anyone know whether SPI is used for anything on the E2?  (The 
pins are wired - somewhere - but they could just be in GPIO mode.)  I'd 
like to add an MMC card to rapidly download/prototype code...



More information about the e3-hacking mailing list