[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ...delo getting me frustrated :-)



On Mon, 2002-12-09 at 23:46, Thiemo Seufer wrote:
> > yeah, that's what's written in the firmware specs.. :-/
> You have specs? :-)

well, mainly the 
"TurboChannel firmware specification"

quite an interesting read...

...you could find that at DEC or mirrored here
http://decstation.unix-ag.org/docs/dec_docs/


> > but for some strange reason it never returned something != -1 for that
> > particular decstation...
> > 
> > > > ...why is it stopping here now? :-/
> > > You might be interested in my improved version of delo, source is available
> > > at http://www.csv.ica.uni-stuttgart.de/homes/ths/linux-mips/delo
> > >
> > > It fixes several quirks, maybe also this one. I can't promise it will
> > > work for you, though. :-) If bootinit always returns -1, it won't.
> > 
> > thx, that helped me alot!
> > 
> > alas your version had some ext2fs trouble (and I had to make the code
> > ignore negative bootinit() retvals); 
> 
> What trouble specifically?

the ext2fs returned an error when trying to load the delo.conf..

it was
ec      EXT2_ET_SHORT_READ,
        "Attempt to read block from filesystem resulted in short read"
iirc...

I thought it had to do w/ delo_read's logic... but didn't analyze it
further, instead I already suspected what the problem affecting me with
delo-0.7 was... the elf handling... and concentrated on trying that one
out...

> > in the meantime I figured out, that the problem lays in 0.7's loadelf.c
> > which causes sections of the ELF kernel to be written to memory starting
> > from address 0 ... which caused the firmware to halt...
> > 
> > your code instead uses the program header tables which fixes the problem
> > (for me at least)... well... until your version works for me, I've put
> > up my hacked up code at
> > 
> > http://www.ifs.tuwien.ac.at/~hvr/linux-mipsel/
> > 
> > maybe it'll help others...
> 
> Can you can provide a patch between 0.7 and your version without the
> seemingly ident-related clutter? I would like to include the relevant
> parts in my version.

your code already does the right thing... (except for the disk I/O
maybe, there seems to be some bug... btw, while at it, maybe grub's fs
code might be interesting...)

I did just replaced the code in copyelf.c with something corresponding
to your your ELF handling in loadelf.c, and pulled bootinit() into
main(); and some other minor stuff I added while debugging delo...

sorry for indenting; (btw, you'd just have had to run 'indent' on
delo-0.7 and diff against it):
http://www.ifs.tuwien.ac.at/~hvr/linux-mipsel/delo-0.7indent-0.7hvr.diff

...lemme know when I can test/try out a new version of yours :-)

regards,
-- 
Herbert Valerio Riedel       /    Phone: (EUROPE) +43-1-58801-18840
Email: hvr@hvrlab.org       /    Finger hvr@gnu.org for GnuPG Public Key
GnuPG Key Fingerprint: 7BB9 2D6C D485 CE64 4748  5F65 4981 E064 883F 4142



Reply to: