Re: Modified version of ppcboot.c

[Nicholai B. <nic...@ch...>]

Hello Giorgio

Thanks for your contribution. Nice to see someone else working on boothack
:-)

Den 24-Jun-00, skrev Giorgio Terzi:

GT> Hello Nicholai ! 
GT> i wish submit to you, Jesper and the user-group a modified version
GT> of ppcboot that now works pretty with my a4K-604e/200.

Hehe...we changed to another list and we are now using a cvs for the kernel
sources as well as for boothack. Please take a look at
linux-apus.sourceforge.net for more information. And join the devlist! :-)

GT> 
GT> 1) I have found that the asm instruction " mtspr 280,3 " in the ppc
GT> registers
GT>   disable routine (after progress('s') ) was responsible of the boot
GT> malfunctions i had.
GT>   I have only out commented it and now all works fine.   
GT> 

I have no idea why this fails on your machine. But as long as it works
better I guess we should remove the instruction. Comments anyone?

GT> 2) I have added a function and some code in the "Inhibition routine"
GT> that
GT>   delays to the last the SYS: device shutdown so no other device will
GT> complain
GT>   popping up requesters.
GT> 

Hmmm...nice but the fix isn't very clean as you remarked. Wouldn't it be
better to handle this directly in the loop instead of this messy function? I
think that could be done quite easily. Call FindDosEntry before entering the
loop and then jump over the sys entry in the tree and inhibit it after all
other filesystems. I guess that's basically what you do but I've got a
feeling it could be done with less effort...but I might be wrong.

GT> Nicholai i wish thank you a lot for the good suggestions to setup a gcc
GT> compiler 
GT> environment. (It is not so friendly to set :-) );
GT> 

I know. It is more difficult than you first think. Nice to have someone
else who can compile it.

GT> Problem: as you will see in the DOS function i have written there is a
GT> problem in offsets calculation (in my case the Node structure) and i
GT> have written some dirty code using NUMERIC offsets because symbolic
GT> offests fail. I think there is an aligning problem but i do not know gcc
GT> options so if someone
GT> can suggest me any solution will be welcome!
GT> 

Haven't looked into this yet.

GT> I attach the source and executables of ppcboot.c because is the only
GT> source i have
GT> changed.
GT> 
GT> I have delayed my IOBlix serial driver but soon as possible i shall
GT> submit the diff
GT> file now i have learnt how to do!
GT> 

We are looking forward to your contribution :-)). Please post to the devlist
when you feel you have something to share.

/Nicholai