Any update on linuxbios for laptops

[Justin Rietz]

I have been following/searching this mailing list, and
haven't been able to find whether or not there has /
will be a successful attempt at linuxbios on a laptop.
 Any news?  Thanks.

-Justin

[Ronald G. Minnich]

On Tue, 22 Feb 2005, Justin Rietz wrote:
> I have been following/searching this mailing list, and
> haven't been able to find whether or not there has /
> will be a successful attempt at linuxbios on a laptop.

we need a cooperative company. IBM *almost* was one, but it did not work 
out. I tried linspire, and there was interest, but not at the high levels.

It's hard.

ron

[Justin Rietz]

Is there a detailed list of what the LinuxBios project
would require from a laptop manufacturer?

-Justin

--- "Ronald G. Minnich"  wrote:
> On Tue, 22 Feb 2005, Justin Rietz wrote:
> > I have been following/searching this mailing list,
>
> and
> > haven't been able to find whether or not there has
>
> /
> > will be a successful attempt at linuxbios on a
>
> laptop.
>
> we need a cooperative company. IBM *almost* was one,
> but it did not work 
> out. I tried linspire, and there was interest, but
> not at the high levels.
>
> It's hard.
>
> ron

[Bari Ari]

Laptop designs route lots of signals through microcontrollers that do 
combined things such as keyboard scan, Flash write enable, battery 
management, power management, lid open/close, power buttons etc.

These are the areas that we would need documentation for.

The chipset docs may all be available but not docs to the power 
management/keyboard scan controller firmware. Their settings are 
typically virtual registers in firmware. The docs for the micros 
describe the general purpose micro itself, but that won't tell you how 
they are using all its gpio pins and ports.

-Bari

Justin Rietz wrote:
> Is there a detailed list of what the LinuxBios project
> would require from a laptop manufacturer?
>
> -Justin
>
> --- "Ronald G. Minnich"  wrote:
> > On Tue, 22 Feb 2005, Justin Rietz wrote:
> > > I have been following/searching this mailing list,
> >
> > and
> > > haven't been able to find whether or not there has
> >
> > /
> > > will be a successful attempt at linuxbios on a
> >
> > laptop.
> >
> > we need a cooperative company. IBM *almost* was one,
> > but it did not work 
> > out. I tried linspire, and there was interest, but
> > not at the high levels.
> >
> > It's hard.
> >
> > ron
>
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

[yhlu]

Any laptop with athlon 64 or Opteron cpu?

And how about chipset on that?

I'd like to get one to play LinuxBIOS with it at home.

[yhlu]

Any one is using 
New! Asus Z80K Athlon 64 Notebook- 15inch SXGA+ w/ATI Radeon 9700
128MB/CDRW+DVD/Reader/Wi-Fi B+G

YH

On Wed, 23 Feb 2005 18:12:03 -0800, yhlu  wrote:
> Any laptop with athlon 64 or Opteron cpu?
>
> And how about chipset on that?
>
> I'd like to get one to play LinuxBIOS with it at home.

[Ronald G. Minnich]

On Wed, 23 Feb 2005, Justin Rietz wrote:
> Is there a detailed list of what the LinuxBios project
> would require from a laptop manufacturer?

I used to write these up monthly for vendors and I don't do it any more. I 
do the writeup and they come back and say "Oh, that's NDA material, 
sorry".

So, no, I don't have a list any more.

ron

Build error on the EPIA-M

[Adam Talbot]

build_dir=via/epia-m/epia-m
Trying to find one of TARGET on line 4:
> loadoptions
>  ^

List of nearby tokens:
===> ERROR: Could not parse file
via/epia-m/Config.lb:0

Any ideas?
-Adam

Build error on the EPIA-M

[Ronald G. Minnich]

it's clear we failed to update that target. I have some epia work to do 
and we will look at this next week.

ron

Build error on the EPIA-M

[Adam Talbot]

I am more then willing to do the updating.  Just need to no why its failing.
-Adam
----- Original Message ----- 
From: "Ronald G. Minnich" 
To: "Adam Talbot" 
Cc: 
Sent: Thursday, February 24, 2005 7:03 AM
Subject: Re: Build error on the EPIA-M
> it's clear we failed to update that target. I have some epia work to do 
> and we will look at this next week.
>
> ron

Build error on the EPIA-M

[Ronald G. Minnich]

On Thu, 24 Feb 2005, Adam Talbot wrote:
> I am more then willing to do the updating.  Just need to no why its failing.

the syntax of config files has changed. Look at a working one in the tyan 
directories to see how they should look.

ron

Build error on the EPIA-M

[Adam Talbot]

Well, its a step in the right direction, I updated the configs and started
the compile... ERROR!!
Any ideas on this one. Am I missing something in my configs?

/root/freebios2/src/southbridge/via/vt8235/vt8235.c:523: warning: `enum
chip_pass' declared inside parameter list
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:523: warning: its scope
is only this definition or declaration, which is probably not what you want
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:523: warning: `struct
chip' declared inside parameter list
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:524: error: parameter
`pass' has incomplete type
/root/freebios2/src/southbridge/via/vt8235/vt8235.c: In function
`southbridge_init':
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:527: error:
dereferencing pointer to incomplete type
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:530: error:
`CONF_PASS_PRE_PCI' undeclared (first use in this function)
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:530: error: (Each
undeclared identifier is reported only once
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:530: error: for each
function it appears in.)
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:534: error:
`CONF_PASS_POST_PCI' undeclared (first use in this function)
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:538: warning: implicit
declaration of function `setup_i8259'
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:543: warning: implicit
declaration of function `vga_fixup'
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:549: error:
`CONF_PASS_PRE_BOOT' undeclared (first use in this function)
/root/freebios2/src/southbridge/via/vt8235/vt8235.c: At top level:
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:562: error: unknown
field `enable' specified in initializer
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:562: warning:
initialization from incompatible pointer type
/root/freebios2/src/southbridge/via/vt8235/vt8235.c:102: warning: `nvram_on'
defined but not used
make[1]: *** [vt8235.o] Error 1
make[1]: Leaving directory
`/root/freebios2/targets/via/epia-m/epia-m/normal'
make: *** [normal/linuxbios.rom] Error 1

-Adam Talbot

----- Original Message ----- 
From: "Ronald G. Minnich" 
To: "Adam Talbot" 
Cc: 
Sent: Thursday, February 24, 2005 9:37 AM
Subject: Re: Build error on the EPIA-M
> On Thu, 24 Feb 2005, Adam Talbot wrote:
> > I am more then willing to do the updating.  Just need to no why its

failing.
> the syntax of config files has changed. Look at a working one in the tyan
> directories to see how they should look.
>
> ron

Build error on the EPIA-M

[Ronald G. Minnich]

On Thu, 24 Feb 2005, Adam Talbot wrote:
> Well, its a step in the right direction, I updated the configs and started
> the compile... ERROR!!
> Any ideas on this one. Am I missing something in my configs?

this may be obsolete code, compare against the amd southbridge code.

sorry.

ron

Build error on the EPIA-M

[Adam Talbot]

-LinuxBios Team

Humm, ok been working out some boot issues that I may/will have on the
EPIA-MII.

I think I remember hearing something about booting of the CF card on the
EPIA-MII if I have linuxbios.

If I can boot off the onboard CF device what device does it come up as?
still hde?

Here is the problem I am running into:
On the stock bios I can not boot off the CF, so I boot off a hard disk.
Well, when I add linuxbios, the system will be asking for data before the
hard disk will be ready. My hard disk has a 5~7 sec spin up time. So in turn
I plan to boot off the CF so solve this problem.  OK, fixed the spin up time
problem, but brought up a new problem.  My current kernel config loads the
ide-cs (access to the CF) after software suspend 2. So software suspend does
not work, as it can't read the software suspend file during boot;  ide-cs is
not loaded yet, and I don't know how to change the device loading order in
the kernel.
Any ideas??
-Adam Talbot

----- Original Message ----- 
From: "Ronald G. Minnich" 
To: "Adam Talbot" 
Cc: 
Sent: Thursday, February 24, 2005 9:37 AM
Subject: Re: Build error on the EPIA-M
> On Thu, 24 Feb 2005, Adam Talbot wrote:
> > I am more then willing to do the updating.  Just need to no why its

failing.
> the syntax of config files has changed. Look at a working one in the tyan
> directories to see how they should look.
>
> ron

Build error on the EPIA-M

[Adam Sulmicki]

> If I can boot off the onboard CF device what device does it come up as?
> still hde?

yes, ide.
> Here is the problem I am running into:
> On the stock bios I can not boot off the CF, so I boot off a hard disk.
> Well, when I add linuxbios, the system will be asking for data before the
> hard disk will be ready. My hard disk has a 5~7 sec spin up time. So in turn
> I plan to boot off the CF so solve this problem.  OK, fixed the spin up time
> problem, but brought up a new problem.  My current kernel config loads the
> ide-cs (access to the CF) after software suspend 2. So software suspend does
> not work, as it can't read the software suspend file during boot;  ide-cs is
> not loaded yet, and I don't know how to change the device loading order in
> the kernel.

ide-cs ? that's part of pcmcia, it shouldn't have anything to do with cf 
at all if you are are connecting CF via IDE bus (which is the way you 
should do it, and which should the the way if it is on board cf device).

Build error on the EPIA-M

[Adam Talbot]

-Adam Sulmicki
On the EPIA-MII the CF is part of the PCMCIA and does not connect to the ide
bus at any time... So the only way I can get access to the CF is through
ide-cs. I do have an adaptor on order (CD-IDE) to help with this issue, but
I would much rather use the onboard CF.  The less cables the smaller the
case.
-Adam Talbot

----- Original Message ----- 
From: "Adam Sulmicki" 
To: "Adam Talbot" 
Cc: "Ronald G. Minnich" ; 
Sent: Sunday, February 27, 2005 5:35 AM
Subject: Re: Build error on the EPIA-M
> > If I can boot off the onboard CF device what device does it come up as?
> > still hde?
>
> yes, ide.
> > Here is the problem I am running into:
> > On the stock bios I can not boot off the CF, so I boot off a hard disk.
> > Well, when I add linuxbios, the system will be asking for data before

the
> > hard disk will be ready. My hard disk has a 5~7 sec spin up time. So in

turn
> > I plan to boot off the CF so solve this problem.  OK, fixed the spin up

time
> > problem, but brought up a new problem.  My current kernel config loads

the
> > ide-cs (access to the CF) after software suspend 2. So software suspend

does
> > not work, as it can't read the software suspend file during boot;

ide-cs is
> > not loaded yet, and I don't know how to change the device loading order

in
> > the kernel.
>
> ide-cs ? that's part of pcmcia, it shouldn't have anything to do with cf
> at all if you are are connecting CF via IDE bus (which is the way you
> should do it, and which should the the way if it is on board cf device).
> _______________________________________________
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

Build error on the EPIA-M

[Adam Sulmicki]

On Sun, 27 Feb 2005, Adam Talbot wrote:
> On the EPIA-MII the CF is part of the PCMCIA and does not connect to the ide
> bus at any time... So the only way I can get access to the CF is through
> ide-cs. I do have an adaptor on order (CD-IDE) to help with this issue, but
> I would much rather use the onboard CF.  The less cables the smaller the
> case.

ouh, that's nasty.

yes, via pmcia it is still an ide device..

i guess you can do it either way
    a) pcmcia/ide support in kernel and use that (big!) kernel to
    boot from that device
    b) use ide<cf> adapater and smaller kernel with ide support only.

Build error on the EPIA-M

[Richard Smith]

> i guess you can do it either way
>         a) pcmcia/ide support in kernel and use that (big!) kernel to
>         boot from that device
>         b) use ide<cf> adapater and smaller kernel with ide support only.

I don't think in-kernel will be enough.  pcmica services depend on the
card manager deamon to detect device insertions and register the
device.  cardmgr is user space.

So I think you will have to go early userspace or an initrd and either
pivot root or kexec into another kernel.

When I did my wireless pcmcia net boot setup I had to compile the
pcmcia-cs stuff against uclibc to get it small and remove the glib
dependency.  I was placeing it all in 1Meg of ROM so I even went
further and hacked down cardmgr so that all it would do was register
the device.  Then I built a small initrd with all the tools.  The boot
system then pulled the real kernel down off the network and kexeced
into it.

Build error on the EPIA-M

[Adam Sulmicki]

> > i guess you can do it either way
> >         a) pcmcia/ide support in kernel and use that (big!) kernel to
> >         boot from that device
> >         b) use ide<cf> adapater and smaller kernel with ide support only.
>
> I don't think in-kernel will be enough.  pcmica services depend on the
> card manager deamon to detect device insertions and register the
> device.  cardmgr is user space.

it was my understanding that the reason the pcmcia/cardbus stuff got 
integrated into kernel in the first place was to support booting from such 
devices.
> So I think you will have to go early userspace or an initrd and either
> pivot root or kexec into another kernel.
>
> When I did my wireless pcmcia net boot setup I had to compile the
> pcmcia-cs stuff against uclibc to get it small and remove the glib
> dependency.  I was placeing it all in 1Meg of ROM so I even went
> further and hacked down cardmgr so that all it would do was register
> the device.  Then I built a small initrd with all the tools.  The boot
> system then pulled the real kernel down off the network and kexeced
> into it.

Build error on the EPIA-M

[Richard Smith]

> it was my understanding that the reason the pcmcia/cardbus stuff got
> integrated into kernel in the first place was to support booting from such
> devices.

I believe you are correct but I seem to remember that it will only be
posible via early userspace due to the database lookup cardmgr has to
do.

I just did a test with my adapter.  I compiled bot the bridge driver
and the ide-cs into my 2.6.10 kernel, moved cardmgr to cardmgr.norun
and re-booted.  The bridge was found but the ide-cs device was not. 
Running /sbin/cardmgr.norun by hand enabled the ide-cs like normal. 
So looks to me like its still dependent on cardmgr.

Build error on the EPIA-M

[Adam Talbot]

Cool, thank you for the test.
I think I will fall back and go with the ide-->cf adaptor. That will take a
few days to come in. After that I will be looking to get linuxbios installed
on my EPIA-MII.

----- Original Message ----- 
From: "Richard Smith" 
To: "Adam Sulmicki" 
Cc: "Adam Talbot" ; 
Sent: Sunday, February 27, 2005 6:44 PM
Subject: Re: Build error on the EPIA-M
> > it was my understanding that the reason the pcmcia/cardbus stuff got
> > integrated into kernel in the first place was to support booting from

such
> > devices.
>
> I believe you are correct but I seem to remember that it will only be
> posible via early userspace due to the database lookup cardmgr has to
> do.
>
> I just did a test with my adapter.  I compiled bot the bridge driver
> and the ide-cs into my 2.6.10 kernel, moved cardmgr to cardmgr.norun
> and re-booted.  The bridge was found but the ide-cs device was not.
> Running /sbin/cardmgr.norun by hand enabled the ide-cs like normal.
> So looks to me like its still dependent on cardmgr.

Re[2]: Build error on the EPIA-M

[Alexander Amelkin]

Hi!

Richard Smith:

RS> I don't think in-kernel will be enough.  pcmica services depend on the
RS> card manager deamon to detect device insertions and register the
RS> device.  cardmgr is user space.

No, you're a bit wrong. PCMCIA services do not _depend_ on cardmgr.
They may take advantage of it, but do not depend on it. The only case
when you need cardmgr is for removable (not potentially, but actually)
devices. If you place your CF card into EPIA MII once and for all, you
can easily load all your pcmcia drivers manually (or put them
in-kernel) and have a working PCMCIA IDE.

Re[2]: Build error on the EPIA-M

[Richard Smith]

> RS> I don't think in-kernel will be enough.  pcmica services depend on the
> RS> card manager deamon to detect device insertions and register the
> RS> device.  cardmgr is user space.
>
> No, you're a bit wrong. PCMCIA services do not _depend_ on cardmgr.
> They may take advantage of it, but do not depend on it. The only case
> when you need cardmgr is for removable (not potentially, but actually)
> devices. If you place your CF card into EPIA MII once and for all, you
> can easily load all your pcmcia drivers manually (or put them
> in-kernel) and have a working PCMCIA IDE.

I did a test where I did exactly that. And the IDE device did not
appear until I ran cardmgr  even though both the card bridge driver
and ide-cs were compiled in.

I'm guessing thought that cardmgr just knows the right /proc/ things
to tickle to make the IDE device appear.  Do you happen to know what
those are?