linux bios kernel for m758lmr+

[James Alton]

Hi,

I'm trying to get *any* linux kernel to compile for the m758lmr+
motherboard and disk on chip millenium. I tried 2.4.7, 2.4.17, and
2.4.19, but none of them compiled. I even tried messing with the code
and fixing it up, but still am unable to compile. Does anyone know of
the best kernel to use for sis630 chipset, and do you have to patch it?
(I have been patching the kernel every time I have tried to compile, but
am not sure this is even necessary.)

-James Alton
[e-mail]

[ollie lho]

On Thu, 2002-11-07 at 10:41, James Alton wrote:
> Hi,
>
> I'm trying to get *any* linux kernel to compile for the m758lmr+
> motherboard and disk on chip millenium. I tried 2.4.7, 2.4.17, and
> 2.4.19, but none of them compiled. I even tried messing with the code
> and fixing it up, but still am unable to compile. Does anyone know of
> the best kernel to use for sis630 chipset, and do you have to patch it?
> (I have been patching the kernel every time I have tried to compile, but
> am not sure this is even necessary.)

What is the error message ?? All the kernel you metioned work fine
for me.

Ollie

[Adam Sulmicki]

> the best kernel to use for sis630 chipset, and do you have to patch it?

normally I would say "yes", except that I just today ran unmodifed linux
and unmodifed OpenBSD on top of LinuxBIOS.
ops I have said too much :P
can you all please look at this red dot here.

Intel 440MX project

[Richard A. Smith]

Ok... Finally after almost 1.5 years.  We (Bitworks) got the go
(complete with a PO) to build a system that uses linuxbios.

Its going to be a mostly based on our 440BX based system but we want
to use the the 440MX for to the low power CPU use, built in AC97 and
North-South bridge integration into one BGA.

The 440MX seems to have some nifty features as well.  Like the
ability to assign the top 4Megs of address space to the 8-bit X-bus. 
Which is good since I have a lot of stuff I need to get into flash.

An interesting feature that I'll have to add to LinuxBIOS fir this
project is digital signature checking (via SHA1 and DSS) of both the
bios image and the payload.   My plan is to raid GPGP for most of
this.  Our customers Public keys and the BIOS image check will be
stored in ROM space that can't be updated except by removing the chip
and reprogramming.  Pretty much everything else that loads will be
verified against those keys.

Since the hardware isn't designed yet and I want to start working on
the firmware.  I'm looking for motherboards that use the 440MX
chipset.  They will probally be a laptop MB. My initial searches on
google and pricewatch didn't yeild much.  Searching for 440MX
generates lots of noise since there is a 440MX Matrox Video card. 
Filtering for Mainbords out of that has been elusive so far.

Anybody know of a product that used the 440MX chipset or have any
experiece with that chipset?

[Ronald G. Minnich]

send me your output from the compile.

ron

[James Alton]

ron,

Ok. I am going to try 2.4.19 this time round. Updating
freebios/linuxbios as I type this with cvs update -P.

1. tar -xvf linux-2.4.19.tar (linux kernel is now installed at
/usr/src/linux-2.4.19)
2. mv linux-2.4.19 linux   (to satisfy all the scripts that require
this)
3. cd linux; cp
/home/jalton/freebios/src/kernel_patches/linux-2.4.19-sis.patch .; cp
/home/jalton/freebios/src/kernel_patches/config-2.4.19-sis .config;
4. (still in /usr/src/linux) patch -p1 < linux-2.4.19-sis.patch
5. (patches scroll by, no errors listed)
6. make menuconfig (I enter this, don't change anything, and say save
changes. To satisfy autoconf makefile stuffs)
7. make dep && make bzImage

a few minutes later...

make[1]: Leaving directory `/usr/src/linux/arch/i386/lib'
ld -m elf_i386 -T /usr/src/linux/arch/i386/vmlinux.lds -e stext
arch/i386/kernel/head.o arch/i386/kernel/init_task.o init/main.o
init/version.o init/do_mounts.o \
        --start-group \
        arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o
mm/mm.o fs/fs.o ipc/ipc.o \
         drivers/char/char.o drivers/block/block.o drivers/misc/misc.o
drivers/net/net.o drivers/media/media.o drivers/ide/idedriver.o
drivers/cdrom/driver.o drivers/sound/sounddrivers.o drivers/pci/driver.o
drivers/video/video.o \
        net/network.o \
        /usr/src/linux/arch/i386/lib/lib.a /usr/src/linux/lib/lib.a
/usr/src/linux/arch/i386/lib/lib.a \
        --end-group \
        -o vmlinux
drivers/video/video.o(.data+0x1bd4): undefined reference to `local
symbols in discarded section .text.exit'
make: *** [vmlinux] Error 1

I'm not sure why it is discarding local symbols or what, but whatever I
do, I can't seem to debug it. (I've checked the actual video.o objdumb
.text.exit section, not much help. I've checked the actual code itself,
not much help.) It brings this error message up before the final, get
all object files and put them into vmlinux stage.

Someone told me it might be the util-linux package, which I have version
2.11r-10 of. (It came from RPM binary form) I'm not sure if this is true
or what.

Thanks for all your help so far.

- James Alton
- [e-mail]

On Wed, 2002-11-06 at 20:15, Ronald G. Minnich wrote:
> send me your output from the compile.
>
> ron

[Adam Sulmicki]

> drivers/video/video.o(.data+0x1bd4): undefined reference to `local
> symbols in discarded section .text.exit'
> make: *** [vmlinux] Error 1
> Someone told me it might be the util-linux package, which I have version
> 2.11r-10 of. (It came from RPM binary form) I'm not sure if this is true
> or what.

this may or may not help:

http://marc.theaimsgroup.com/?l=linux-kernel&w=2&r=1&s=.text.exi...

[Bob Drzyzgula]

FWIW, I got precisely this same error yesterday afternoon,
trying to compile for a MS7308E. This was on a fresh
and unpatched install of RH 7.3, with 2.4.19 freshly
unpacked and linuxbios pulled out of CVS in mid-afternoon
US/Eastern. I got to that point just as I was leaving for
the day, I was hoping that maybe today would be better. :-)

--Bob

On Thu, Nov 07, 2002 at 01:23:36AM -0800, James Alton wrote:
>     -o vmlinux
>
> drivers/video/video.o(.data+0x1bd4): undefined reference to `local
> symbols in discarded section .text.exit'
> make: *** [vmlinux] Error 1

[Andrew Ip]

You can try our LinuxBIOS ready kernel.
ftp://ftp.cwlinux.com/pub/downloads/kernel/2.4.19/kernel-source-linuxbios-2.4.19-CWLINUX_4.i386.rpm

It has integrated with sis patch, kexec plue others patches.

-Andrew

On Thu, Nov 07, 2002 at 01:23:36AM -0800, James Alton wrote:
> ron,
>
> Ok. I am going to try 2.4.19 this time round. Updating
> freebios/linuxbios as I type this with cvs update -P.
> tar -xvf linux-2.4.19.tar (linux kernel is now installed at
>
> /usr/src/linux-2.4.19)
> 2. mv linux-2.4.19 linux   (to satisfy all the scripts that require
> this)
> 3. cd linux; cp
> /home/jalton/freebios/src/kernel_patches/linux-2.4.19-sis.patch .; cp
> /home/jalton/freebios/src/kernel_patches/config-2.4.19-sis .config;
> 4. (still in /usr/src/linux) patch -p1 < linux-2.4.19-sis.patch
> 5. (patches scroll by, no errors listed)
> 6. make menuconfig (I enter this, don't change anything, and say save
> changes. To satisfy autoconf makefile stuffs)
> 7. make dep && make bzImage
>
> a few minutes later...
>
> make[1]: Leaving directory `/usr/src/linux/arch/i386/lib'
> ld -m elf_i386 -T /usr/src/linux/arch/i386/vmlinux.lds -e stext
> arch/i386/kernel/head.o arch/i386/kernel/init_task.o init/main.o
> init/version.o init/do_mounts.o \
>         --start-group \
>         arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o
> mm/mm.o fs/fs.o ipc/ipc.o \
>          drivers/char/char.o drivers/block/block.o drivers/misc/misc.o
> drivers/net/net.o drivers/media/media.o drivers/ide/idedriver.o
> drivers/cdrom/driver.o drivers/sound/sounddrivers.o drivers/pci/driver.o
> drivers/video/video.o \
>         net/network.o \
>         /usr/src/linux/arch/i386/lib/lib.a /usr/src/linux/lib/lib.a
> /usr/src/linux/arch/i386/lib/lib.a \
>         --end-group \
>         -o vmlinux
> drivers/video/video.o(.data+0x1bd4): undefined reference to `local
> symbols in discarded section .text.exit'
> make: *** [vmlinux] Error 1
>
> I'm not sure why it is discarding local symbols or what, but whatever I
> do, I can't seem to debug it. (I've checked the actual video.o objdumb
> .text.exit section, not much help. I've checked the actual code itself,
> not much help.) It brings this error message up before the final, get
> all object files and put them into vmlinux stage.
>
> Someone told me it might be the util-linux package, which I have version
> 2.11r-10 of. (It came from RPM binary form) I'm not sure if this is true
> or what.
>
> Thanks for all your help so far.
> James Alton
> [e-mail]
>
> On Wed, 2002-11-06 at 20:15, Ronald G. Minnich wrote:
> > send me your output from the compile.
> >
> > ron
>
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

[Bob Drzyzgula]

This one built OK for me. I'll try to figure out what
was wrong with the other kernel directory; I had
pretty much just followed Antony Stone's HOWTO at

http://www.acl.lanl.gov/linuxbios/developer/portguides/M810LR/index.html

but I could have slipped up.

--Bob

On Thu, Nov 07, 2002 at 09:27:32PM +0800, Andrew Ip wrote:
> You can try our LinuxBIOS ready kernel.
> ftp://ftp.cwlinux.com/pub/downloads/kernel/2.4.19/kernel-source-linuxbios-2.4.19-CWLINUX_4.i386.rpm
>
> It has integrated with sis patch, kexec plue others patches.
>
> -Andrew

[Ronald G. Minnich]

note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
problems, send you gcc version too.

ron

[Bob Drzyzgula]

Sorry, stock RH 7.3:

% gcc -v
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs
gcc version 2.96 20000731 (Red Hat Linux 7.3 2.96-110)

On Thu, Nov 07, 2002 at 09:22:01AM -0700, Ronald G. Minnich wrote:
> note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
> problems, send you gcc version too.
>
> ron

[Stefan Reinauer]

* Ronald G. Minnich  [021107 17:22]:
> note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
> problems, send you gcc version too.

Is there a list of known problems with gcc 3.2 for LinuxBIOS somewhere?

Stefan

[Ronald G. Minnich]

On Thu, 7 Nov 2002, Stefan Reinauer wrote:
> Ronald G. Minnich  [021107 17:22]:
> > note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
> > problems, send you gcc version too.
>
> Is there a list of known problems with gcc 3.2 for LinuxBIOS somewhere?

good idea. The current big one is that the MP table stopped working. Not 
sure why.

ron

[James Alton]

Stefan,
Every linux kernel out that I know of is not gcc 3.2 safe. Which means
it will not work. It's a linux kernel problem, not a linux bios problem.
I just didn't check my version when I did the build. *Do not use gcc 3.2
and try to compile a linux kernel* 99.999% chance it will not work.

- James Alton
- [e-mail]

On Thu, 2002-11-07 at 08:40, Stefan Reinauer wrote:
> Ronald G. Minnich  [021107 17:22]:
> > note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
> > problems, send you gcc version too.
>
> Is there a list of known problems with gcc 3.2 for LinuxBIOS somewhere?
>
> Stefan

[Stefan Reinauer]

Hi James,
> Every linux kernel out that I know of is not gcc 3.2 safe. Which means
> it will not work. It's a linux kernel problem, not a linux bios problem.
> I just didn't check my version when I did the build. *Do not use gcc 3.2
> and try to compile a linux kernel* 99.999% chance it will not work.

I cannot agree. There are several distributions out there now which are
completely based on gcc 3.2 and there has been a pretty reliable set of
patches floating around to fix the kernels 2.4 and newer to work when
compiled with gcc 3.2. You will have no success with 2.2 kernels, but
they are outdated anyways. gcc 3.2 is proven to produce a lot better
code than gcc 2.95.x on many platforms. Main problem when "porting"
applications to gcc 3.2 is the different behaviour of the gcc's in some
details, i.e. packing of structs and unions (which likely bites us with
the mptables). That's not a bug, but a feature. gcc reorganizes your
structs so they are aligned/packed in a proper way (faster, and needed
on some platforms, such as alpha). LinuxBIOS should be ported to fit
these new assumptions, as gcc 3.2 will be the default gcc in a very
short while. We have almost no machines running with gcc 2.95.x here
anymore.

Stefan

[Ronald G. Minnich]

On Thu, 7 Nov 2002, Stefan Reinauer wrote:
> code than gcc 2.95.x on many platforms. Main problem when "porting"
> applications to gcc 3.2 is the different behaviour of the gcc's in some
> details, i.e. packing of structs and unions (which likely bites us with
> the mptables). That's not a bug, but a feature. gcc reorganizes your
> structs so they are aligned/packed in a proper way (faster, and needed
> on some platforms, such as alpha).

from my point of view, given the original goals of C, that is a bug.

But it is what we have. How do we tell gcc not to reorder our structs?

ron

[Steve M. Gehlbach]

> But it is what we have. How do we tell gcc not to reorder our structs?
>
> ron

"__attribute__ ((packed))" (
http://gcc.gnu.org/onlinedocs/gcc-3.2/gcc/Variable-Attributes.html )

Or is this a different issue?

-Steve

[Ronald G. Minnich]

On Thu, 7 Nov 2002, Steve M. Gehlbach wrote:
> "__attribute__ ((packed))" (
> http://gcc.gnu.org/onlinedocs/gcc-3.2/gcc/Variable-Attributes.html )
>
> Or is this a different issue?

it is a different issue.

ron

[James Alton]

ron,
Oh geez. I don't do a lot of compiling on this system (it has redhat 8.0
with all the defaults) so I never realized it had gcc 3.2. Yes, I
realize it breaks stuff, I just didn't check for that. Thanks.

P.S. Haven't compiled yet, but I'm sure going to a stable gcc will work.

- James Alton
- [e-mail]

On Thu, 2002-11-07 at 08:22, Ronald G. Minnich wrote:
> note that gcc 3.2 is breaking stuff everywhere. From now on, with compile 
> problems, send you gcc version too.
>
> ron