Re: doc ipl image

[roger]

ok.

docipl /path/to/src/mainboard/asus/cua/ipl.S

and Python configure made the Makefiles, however dd fails with a
negative number with "docipl" in the config file.

So i'm doing this seperately and copying the created dociple to another
folder and then re-running the configure/make without the "docipl"
option to build the linuxbios.image and linuxkernel.image images.

I'm rebuilding the kernel again for the DoC (looks like it might be wise
to build mtd/mtdchar/doc2001 modules into the kernel) and am working on
a serial console for remote console.

So far, i've gotten nothing to boot.  Just the same scenario as a box
without a bios chip.  Although, I could have sworn i had some additional
hdd activity on the last attempt.

One additional question, if HAVE_FRAMEBUFFER=0, there still is console
activity on the monitor right?  (-i think i'm interpretting this
correctly-).

wew!  ain't google grand?!?  one can actually have a one-way conversion
with himself via email!!!

On Tue, 2003-06-10 at 22:24, roger wrote:
> Just a quick follow-up.  I noticed that I had to add the following line
> in the config file:
>
> docipl /pathto/ipl.S
>
> ok. now for ipl.S  (for a 440BX chipset?)

doc ipl image

roger  writes:
> ok.
>
> docipl /path/to/src/mainboard/asus/cua/ipl.S

But note that is for a sis630 chipset.

ipl.S must enable memory or there is no room to put LinuxBIOS,
and enabling memory is 512 bytes is a very challenging problem.

You can probably manage it if you are just hard coding the
necessary steps for one brand of memory.  Being more general
than that on a 440BX I suspect will be quite difficult..

Eric

doc ipl image

[ron minnich]

On 11 Jun 2003, Eric W. Biederman wrote:
> > docipl /path/to/src/mainboard/asus/cua/ipl.S
>
> But note that is for a sis630 chipset.

no, that is our asus cua support.

But, yes, ipl for a 440bx is going to REALLY be hard.

kron

doc ipl image

[Richard Smith]

ron minnich wrote:
> no, that is our asus cua support.
>
> But, yes, ipl for a 440bx is going to REALLY be hard.

If he needs to read the spd to get the DIMM settings I'd guess its not 
going to happen.

There's 512 bytes of code just in all the necesary PCI_CONFIG_WRITE 
macros alone much less reading the spd.

doc ipl image

[ron minnich]

On 10 Jun 2003, roger wrote:
> One additional question, if HAVE_FRAMEBUFFER=0, there still is console
> activity on the monitor right?  (-i think i'm interpretting this
> correctly-).

depends on the type of mainboard. Mostly, no.

ron