linuxbios on epia-m -- not starting

[Niki Waibel]

linuxbios does not start ...
i can see:
===
LinuxBIOS-1.0.0 Mon Sep 22 02:17:36 CEST 2003 starting...
Copying LinuxBIOS to ram.
Jumping to LinuxBIOS.
===
that's it.
first i tought it is a mem problem (512MB Infineon CL2).
i changed 0x350 to 0x150 in src/northbridge/via/vt8623/raminit.inc
and i've activated ../freebios/src/ram/ramtest.inc
in ../freebios/src/mainboard/via/epia-m/Config.
it seems that everything is ok:
===
LinuxBIOS-1.0.0 Mon Sep 22 12:38:53 CEST 2003 starting...                       
Testing SDRAM : 00000000-0009ffff                                               
SDRAM fill:                                                                     
0009ffff                                                                        
SDRAM verify:                                                                   
0009ffff                                                                        
Done.                                                                           
Copying LinuxBIOS to ram.                                                       
Jumping to LinuxBIOS.                                                     
===

so i tried to figure out where the stop occurs.
i verified if the code is copied correctly:
===
LinuxBIOS-1.0.0 Mon Sep 22 02:17:36 CEST 2003 starting...                       
Copying LinuxBIOS to ram.                                                       
NIKI: Jumping to LinuxBIOS.                                                     
src:  000f0b70
dst:  00004000
size: 0000aa8c
src_code: 010f2efa00405215
dst_code: 010f2efa00405215
===
-> fine.

unfort i dont know much about x86 asm...

but it seems that the problem is somewhere in
c_start.s ...

if i put
        movl $0xfffe0000, %edi
        jmp *%edi
to the top of the _start: function then linuxbios loops.
(i just wanted to check if the copied code is executed...)

is that a better way to find out what's wrong?

by the way:
===
gcc -x assembler-with-cpp -DASSEMBLY -E ... crt0.S > crt0.s 
gcc ... -o crt0.o crt0.s
crt0.S: Assembler messages:
crt0.S:163: Warning: indirect jmp without `*'
===
instead of:
        leal    _iseg, %edi
        jmp     %edi
shouldn't it be:
        leal    _iseg, %edi
        jmp     *%edi
?

niki w. waibel

[ron minnich]

I think you should get memtest86, set it up to run on serial console, and 
see if it works or not.

ron

[Niki Waibel]

> I think you should get memtest86, set it up to run on serial console, and 
> see if it works or not.

hmmm -- you mean with lilo?
i've never used memtest86.
will give it a try.

niki

[ron minnich]

On Mon, 22 Sep 2003, Niki Waibel wrote:
> i've never used memtest86.

make memtest86 the payload. be sure to build with serial console only.

ron

[Niki Waibel]

> > i've never used memtest86.
>
> make memtest86 the payload. be sure to build with serial console only.

no, no, that does not work.
the payload is never loaded/reached.
the trouble is caused earier!
right after crt0.s a jmp to the copied mem
is done. the copied mem is c_start.s at the beginning.
there it seems to stop.
(otherwise i'd see the messages from hardwaremain.c,
am i wrong?)

anyway -- running memtest86 with lilo right now and
it seems that everything is fine (test4 is running.
it will take some time to finnish). i dont think it
is a mem problem (but i might be wrong).

niki

[Steve Gehlbach]

Niki Waibel wrote:
> anyway -- running memtest86 with lilo right now and
> it seems that everything is fine (test4 is running.
> it will take some time to finnish). i dont think it
> is a mem problem (but i might be wrong).

But the issue is, is linuxbios setting up ram correctly.  I posted some 
time ago some simple assy language instructions to check each byte as it 
is copied to ram.  Or make your own, but as the C-code is copied test 
each byte and see if it matches.

Did you try both ram slots?  I think only one works with a single SDRAM. 
  Also set CONFIG_COMPRESS=0 is simpler code. What size SDRAM?

-Steve

[Niki Waibel]

> > anyway -- running memtest86 with lilo right now and
> > it seems that everything is fine (test4 is running.
> > it will take some time to finnish). i dont think it
> > is a mem problem (but i might be wrong).
>
> But the issue is, is linuxbios setting up ram correctly.  I posted some 
> time ago some simple assy language instructions to check each byte as it 
> is copied to ram.  Or make your own, but as the C-code is copied test 
> each byte and see if it matches.
>
> Did you try both ram slots?

there is only one in the epia-m (600mhz).
> I think only one works with a single SDRAM. 
>   Also set CONFIG_COMPRESS=0 is simpler code. What size SDRAM?

i use CONFIG_COMPRESS=0.
size is 512MByte CL2.

have to leave company now.
pls Cc ``niki dot waibel at gmx dot net''.

thanks, niki

[Steve Gehlbach]

Niki Waibel wrote:
> size is 512MByte CL2.

The only thing I can think of is to try a different/ smaller SDRAM; 64M 
or 128M.  The RAM setup code is not always that smart about 
configuration.  But has anyone else got the epia-m working?  I have only 
used EPIA-800.

-Steve