#12 jornada820 support + piggy loader + small fixes
After a lot of stupidity on my behalf, here is
a working port of blob to the Jornada 820.
* I added the usual target-specific files for the new
architecture jornada.
* I completed some documentation (README, porting.txt)
with some information erikm gave me on IRC.
* I somehow managed to "port" ledasm.S to the Jornada,
by drawing lines on the WinCE-preconfigured
framebuffer instead of trying to blink a non-existing
LED.
* I added support for a new load_kernel method: loading
the kernel from some piggy-back BLOB on blob.
* I modified rest-ld-script, trampoline.S and chain.S,
so blob knows to move around just the right size
needed, not a blind 64KB of data (needed for piggy).
* I left some debug information #if 0'ed, that I'd like
to remove only after it has been committed to CVS, so I
can recover it if I ever need it.
Logged In: YES
user_id=94240
* Made the cvs diff output friendly to patch -p0.
* fixed bug in call.c
* fixed bugs and potential issues in previous patch.
* Found that memory cell at 0xc005c080 is constantly
changing, probably due to some DMA going on. No fix yet.
diff from CVS for jornada820 support
third take at a jornada820 port, diff against cvs
Logged In: YES
user_id=94240
* Cleaned up my patch somewhat.
* Added diag/lcd.c settings
* Added dirty workaround in memory.c for the uncontrolled
1ms timer at 0xc005c080.
I can now have Linux die in head-armv.S -- woohoo!
Logged In: YES
user_id=94240
* a better patch for memory.c
take 4
diff from recent CVS blob-jornada820-cvs-6.diff.gz
Logged In: YES
user_id=94240
Patch updated with recent CVS.
Obviously, you reimplemented some features yourself in the
mean time.
Your loss.
Logged In: YES
user_id=94240
Once again, merged with current CVS.
Why make me suffer so much?
jornada820 + piggy