summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Bernhard Urban [Mon, 14 Dec 2009 03:25:14 +0000 (04:25 +0100)]
okay, here is a programmer (=kernel.py) which communicate to the pkernel
running on the board. there are some operation done to the flash,
however it isn't correct (nothing happens after a reset)
note: atm you have to program the pkernel with a windows programmer.
note: use simpleprg.mhx to test it. (on larger files it takes *very*
long, due to a built-in debug delay ;))
todo: after pkernel works correctly, we have to reduce the .mhx size.
when the mhx file is small enough we can use the bootrom hack to load it
directly into ram. after that it should be done :)
Bernhard Urban [Mon, 14 Dec 2009 02:04:17 +0000 (03:04 +0100)]
removed USTACK
Bernhard Urban [Mon, 14 Dec 2009 01:59:24 +0000 (02:59 +0100)]
changed from K to X -- even more DRAM \o/
Bernhard Urban [Mon, 14 Dec 2009 01:57:18 +0000 (02:57 +0100)]
removed "CONST" section
however, it is still there in the map... wtf?
Bernhard Urban [Mon, 14 Dec 2009 01:44:38 +0000 (02:44 +0100)]
clockstuff too
Bernhard Urban [Mon, 14 Dec 2009 01:40:44 +0000 (02:40 +0100)]
removed basically some clockrelated ifdef stuff
Bernhard Urban [Mon, 14 Dec 2009 01:34:00 +0000 (02:34 +0100)]
leds "show" 0xaa when finished flashing process
Bernhard Urban [Mon, 14 Dec 2009 01:33:18 +0000 (02:33 +0100)]
some cleanup
Bernhard Urban [Mon, 14 Dec 2009 01:20:45 +0000 (02:20 +0100)]
removed some uart strings and functions
-rw-r--r-- 1 lewurm lewurm 5206 2009-12-14 02:21 pkernel/pkernel.mhx
Bernhard Urban [Mon, 14 Dec 2009 01:12:32 +0000 (02:12 +0100)]
refactoring/reformatting fun...
dude, i think the don't even have a tab key in japan.
Bernhard Urban [Mon, 14 Dec 2009 00:33:17 +0000 (01:33 +0100)]
w00t, wtf IT WORKS NOW!!!1
BUT actually I have NO idea why!
Bernhard Urban [Sun, 13 Dec 2009 22:30:58 +0000 (23:30 +0100)]
we can write 0x00 at 0xf4000, but nothing else... :(
also added the a pragma section found in flash.c to all c-files... dunno
if it actually changes something.
Bernhard Urban [Sun, 13 Dec 2009 22:03:35 +0000 (23:03 +0100)]
nicer output; still don't work as desired :/
Bernhard Urban [Sun, 13 Dec 2009 21:37:19 +0000 (22:37 +0100)]
this commit won't work :(
basically he tries to flash a programm (here it is simpleprg) to
0xf4000 and erase it before. unfortunately it will fail -- maybe because
the pkernel is executed itself from there?!
however, the readme.txt of the flashdemo says no... dunno :(
Bernhard Urban [Sun, 13 Dec 2009 21:37:13 +0000 (22:37 +0100)]
prep for python kernelprog
Bernhard Urban [Sun, 13 Dec 2009 20:49:51 +0000 (21:49 +0100)]
root Makefile
Bernhard Urban [Sun, 13 Dec 2009 20:49:19 +0000 (21:49 +0100)]
moved pkernel in a own directory
Bernhard Urban [Sun, 13 Dec 2009 20:32:44 +0000 (21:32 +0100)]
merged flashdemo repo
Merge branch 'master' of wien.tomnetworks.com:flashdemo
Conflicts:
.gitignore
Bernhard Urban [Sun, 13 Dec 2009 07:28:39 +0000 (08:28 +0100)]
actually we have to write it in IRAM, thus we have to skip addresses for
section >= 0x148000 and also have to calculate a other writeaddress as
given in the mhx file. however, we need those addresses later for the
real flashing part (through pkernel)
theStack [Sun, 13 Dec 2009 07:08:53 +0000 (08:08 +0100)]
flashing will be done now in a simple loop
Bernhard Urban [Sun, 13 Dec 2009 07:06:43 +0000 (08:06 +0100)]
refactor fun
Bernhard Urban [Sun, 13 Dec 2009 07:03:18 +0000 (08:03 +0100)]
pkernel programmer init
theStack [Sun, 13 Dec 2009 07:01:11 +0000 (08:01 +0100)]
first version of mhx file support impl.
o it can only handle S2 records (which means max. 6 bytes for the addresses) by now,
but other data records shouldn't be used by the fujitsu converter anyway i guess?
o data is only read and printed to stdout, but not really used, that will be done
in the next commit
Bernhard Urban [Sun, 13 Dec 2009 04:58:33 +0000 (05:58 +0100)]
okay, big fat note: clear the fucking watchdog timer...... DAMN ;)
Bernhard Urban [Sun, 13 Dec 2009 02:19:02 +0000 (03:19 +0100)]
execute own code from iram, however READ attention text before doing
something stupid
also added some (useful?) examples...
theStack [Sat, 12 Dec 2009 04:15:53 +0000 (05:15 +0100)]
a call to the end of the IRAM (0x00033ffc)...
*really* starts the program that was flashed before - how crazy is that? does
that mean that instructions are called backwards or something like that? well,
at the moment i don't care new super mario bros is waiting xD
theStack [Sat, 12 Dec 2009 04:00:08 +0000 (05:00 +0100)]
impl. CALL and CHECKSUM bootloader cmds, untested
theStack [Sat, 12 Dec 2009 03:47:32 +0000 (04:47 +0100)]
add constants for device and baudrate
theStack [Sat, 12 Dec 2009 03:35:23 +0000 (04:35 +0100)]
add a test with some more data and local chksum
+ decreased sleep time (1ms instead of 10ms just works fine)
+ doubled baud rate (the highest possible by now as it seems?)
theStack [Sat, 12 Dec 2009 03:10:12 +0000 (04:10 +0100)]
initial commit
Bernhard Urban [Fri, 11 Dec 2009 23:33:26 +0000 (00:33 +0100)]
decreased speed for led counter and stop it at the end of main
Bernhard Urban [Fri, 11 Dec 2009 23:10:01 +0000 (00:10 +0100)]
lowercase plz
Bernhard Urban [Fri, 11 Dec 2009 23:08:47 +0000 (00:08 +0100)]
Revert "headerfile/makefile hax hax"
This reverts commit
828d4e7ae7bd728f23424ea58cc1b445a58328ac.
Bernhard Urban [Fri, 11 Dec 2009 23:06:18 +0000 (00:06 +0100)]
headerfile/makefile hax hax
Bernhard Urban [Fri, 11 Dec 2009 23:03:58 +0000 (00:03 +0100)]
right led port for bits pot blue board
theStack [Fri, 11 Dec 2009 22:57:46 +0000 (23:57 +0100)]
dos2unix on all files + tab madness
Bernhard Urban [Fri, 11 Dec 2009 22:13:51 +0000 (23:13 +0100)]
gitignore update
Bernhard Urban [Fri, 11 Dec 2009 21:59:40 +0000 (22:59 +0100)]
init