Diskless Windows??

[Brian Gregor]

Is it possible to use LinuxBIOS to boot Windows 2000 diskless?

Thanks,

Brian

[Hendricks David W.]

Adam Sulmicki would be the guy to ask about that... He did ADLO a while 
back to boot Win2K. It's quite a bit messier than booting Linux, I gather.

On Mon, 3 May 2004, Brian Gregor wrote:
> Is it possible to use LinuxBIOS to boot Windows 2000 diskless?
>
> Thanks,
>
> Brian
> _______________________________________________
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

[Adam Sulmicki]

I don't feel that my work is that relevant here.

Yes it allows you to boot windows but there's a bit of way to go from
there to having netbooted windows. In theory you could add appropriate
software to have boot block send over network, it starts and then what?

In other words windows would have some suport for getting rest of itself
over network, just like Linux does.... which I do not believe windows can.

or you could try to use VMware from linux.

Having said this all I think someone else did some work in that area,
transparent forwarding of disk I/O or some such..either way it would be
proprietary binary only stuff..

Dunno if it helps.

On Mon, 3 May 2004, Hendricks David W. wrote:
> Adam Sulmicki would be the guy to ask about that... He did ADLO a while
> back to boot Win2K. It's quite a bit messier than booting Linux, I gather.
>
> On Mon, 3 May 2004, Brian Gregor wrote:
> > Is it possible to use LinuxBIOS to boot Windows 2000 diskless?
> >
> > Thanks,
> >
> > Brian
> > _______________________________________________
> > Linuxbios mailing list
> > [e-mail]
> > http://www.clustermatic.org/mailman/listinfo/linuxbios
>
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

[Oleg Goldshmidt]

Adam Sulmicki  writes:
> Having said this all I think someone else did some work in that area,
> transparent forwarding of disk I/O or some such..either way it would be
> proprietary binary only stuff..

I am not sure what the OP meant by "diskless". If the question is
about booting Windows from a networked disk on a computer without a
local disk, then it is possible booting off an iSCSI disk. Check

http://www.haifa.il.ibm.com/projects/storage/iboot/

Disclosure: I work at IBM's Haifa Labs - iBOOT is rather routine for us.

LinuxBIOS is not a part of this, though. Pity.

To be quite fair, Cisco are also working hard:

http://www.internetnews.com/infra/article.php/1490471

[ron minnich]

the way I would want to do diskless windows is under vmware, I would never 
trust it otherwise.

ron

Oleg Goldshmidt  writes:
> Adam Sulmicki  writes:
> > Having said this all I think someone else did some work in that area,
> > transparent forwarding of disk I/O or some such..either way it would be
> > proprietary binary only stuff..
>
> I am not sure what the OP meant by "diskless". If the question is
> about booting Windows from a networked disk on a computer without a
> local disk, then it is possible booting off an iSCSI disk. Check
>
> http://www.haifa.il.ibm.com/projects/storage/iboot/
>
> Disclosure: I work at IBM's Haifa Labs - iBOOT is rather routine for us.

My general impression with this is that you still need iscsi drivers
on the OS side to make this work.  And that you have to be quite careful
at the transition point on the OS side.  Not knowing how windows works
this is where I get leery of these kinds of things.

It sounds like from the description there is no way to do windows diskless
without emulating a disk.

I have serious issues with iSCSI because the only implementation I have
seen was tremendously complex and nasty looking.  Which is a real downside
when compared to something simple like nbd.

How well does iBOOT work on the linux side.  I guess if you can get a kernel
and ramdisk into memory with lilo or something like that using the legacy
BIOS calls it should work ok.  If you have an iSCSI driver for Linux.

Eric

[Oleg Goldshmidt]

[e-mail] (Eric W. Biederman) writes:
> It sounds like from the description there is no way to do windows diskless
> without emulating a disk.

I am not sure I understand what you mean. Isn't that what iSCSI is
about: everything above that thinks there is a local SCSI disk...
> I have serious issues with iSCSI because the only implementation I have
> seen was tremendously complex and nasty looking.  Which is a real downside
> when compared to something simple like nbd.

Well, I don't know what implementations you have seen. There are
undoubtedly some ugly ones. That's an implementation issue, though:
the basic idea looks quite clean to me. Not that there are no issues...
> How well does iBOOT work on the linux side.

Better ;-)
> If you have an iSCSI driver for Linux.

We do.

Oleg Goldshmidt  writes:
> [e-mail] (Eric W. Biederman) writes:
> > It sounds like from the description there is no way to do windows diskless
> > without emulating a disk.
>
> I am not sure I understand what you mean. Isn't that what iSCSI is
> about: everything above that thinks there is a local SCSI disk...

There are some efficiencies to be gained by not having a remote disk,
and instead having a remote filesystem.  From everything I have
heard that does not sound like it has been implemented for windows
with NT kernels.
> > I have serious issues with iSCSI because the only implementation I have
> > seen was tremendously complex and nasty looking.  Which is a real downside
> > when compared to something simple like nbd.
>
> Well, I don't know what implementations you have seen. There are
> undoubtedly some ugly ones. That's an implementation issue, though:
> the basic idea looks quite clean to me. Not that there are no issues...

nbd is 600 likes of obviously correct code, which can put full disk
speed onto the wire.  Nor have I seen any issues with it.  That
makes tough competition for iSCSI in modest deployment locations.

I still need to read the spec to see how sane iSCSI looks on the wire.
> > How well does iBOOT work on the linux side.
>
> Better ;-)
> > If you have an iSCSI driver for Linux.
>
> We do.

Cool.  If iSCSI looks sane I will suggest to the etherboot guys that
we implementing booting over it :)

Eric

[Adam Sulmicki]

On 6 May 2004, Oleg Goldshmidt wrote:
> I am not sure what the OP meant by "diskless". If the question is
> about booting Windows from a networked disk on a computer without a
> local disk, then it is possible booting off an iSCSI disk. Check
>
> http://www.haifa.il.ibm.com/projects/storage/iboot/

seems to me like this solution basically redirect BIOS I/O interrupts.

What happens once Windows boots up and attempt to access hardware directly
without BIOS calls in between? Is Windows smart enough to work by default
with hdd (for example : what about swap?)

[Oleg Goldshmidt]

Adam Sulmicki  writes:
> seems to me like this solution basically redirect BIOS I/O interrupts.
>
> What happens once Windows boots up and attempt to access hardware directly
> without BIOS calls in between? Is Windows smart enough to work by default
> with hdd (for example : what about swap?)

You have to solve the problem at BIOS time and at OS time,
obviously. Google for "Windows iSCSI" to get an idea of iSCSI support
in Windows. Booting is more complicated for a variety of reasons.

Oleg Goldshmidt  writes:
> Adam Sulmicki  writes:
> > seems to me like this solution basically redirect BIOS I/O interrupts.
> >
> > What happens once Windows boots up and attempt to access hardware directly
> > without BIOS calls in between? Is Windows smart enough to work by default
> > with hdd (for example : what about swap?)
>
> You have to solve the problem at BIOS time and at OS time,
> obviously. Google for "Windows iSCSI" to get an idea of iSCSI support
> in Windows. Booting is more complicated for a variety of reasons.

OK.  With iSCSI you really are doing a disked based boot, and
have the same issues you have booting from fiber channel or any
other disk that does not have an OS based driver.  The rough edges
are different but the basic problems are the same.

That make sense.

Eric

V1 2 superio chips of the same type

[Richard Smith]

How do I specify 2 superios in V1 that are the same chip only different 
pnp bases using nsuperio?

V1 2 superio chips of the same type

[ron minnich]

On Thu, 6 May 2004, Richard Smith wrote:
> How do I specify 2 superios in V1 that are the same chip only different 
> pnp bases using nsuperio?

here's how it should work (example, two same part at 0x370 and 0x3f0):

nsuperio SMC/fdc37b72x com1={1} port=0x370
nsuperio SMC/fdc37b72x com1={1} port=0x3f0

V1 2 superio chips of the same type

[Richard Smith]

ron minnich wrote:
> On Thu, 6 May 2004, Richard Smith wrote:
> > How do I specify 2 superios in V1 that are the same chip only different 
> > pnp bases using nsuperio?
>
> here's how it should work (example, two same part at 0x370 and 0x3f0):
>
> nsuperio SMC/fdc37b72x com1={1} port=0x370
> nsuperio SMC/fdc37b72x com1={1} port=0x3f0

Thats kinda what I though but here's the nsuperio.c that gets built

#include <pci.h>
extern struct superio_control superio_NSC_pc87351_control;
struct superio superio_NSC_pc87351= { &superio_NSC_pc87351_control, 
.keyboard=1, .com1={1}, .com2={1}, .floppy=0, .lpt=0, .mouse=1,
.port=0x2f};
extern struct superio_control superio_NSC_pc87351_control;
struct superio superio_NSC_pc87351= { &superio_NSC_pc87351_control, 
.keyboard=1, .com1={1}, .com2={1}, .floppy=0, .lpt=0, .mouse=1, .port=0x2f};
struct superio *all_superio[] = {&superio_NSC_pc87351,
&superio_NSC_pc87351,
};
unsigned long nsuperio = 2;

As you can see you end up with 2 structs that are the same name which 
won't build.

V1 2 superio chips of the same type

[ron minnich]

bug. I screwed up.

I'll try to see how to fix this. It will take me a bit but I think it is 
easy, I did handle this in V2.

ron

V1 2 superio chips of the same type

[Stefan Reinauer]

* ron minnich  [040507 03:32]:
> bug. I screwed up.
>
> I'll try to see how to fix this. It will take me a bit but I think it is 
> easy, I did handle this in V2.

Before getting into backporting a lot, Richard, can your motherboard be
ported to V2?

Stefan

V1 2 superio chips of the same type

[Richard Smith]

Stefan Reinauer wrote:
> ron minnich  [040507 03:32]:
> > bug. I screwed up.
> >
> > I'll try to see how to fix this. It will take me a bit but I think it is 
> > easy, I did handle this in V2.
>
> Before getting into backporting a lot, Richard, can your motherboard be
> ported to V2?

Not without a lot of work.  I'm one of the few 440bx users and no mem 
init code for this chipset has been written in V2 yet.  I need spd 
detection.  I keep waiting for the epia to get this all working and then 
I might look at porting forward.

I also currently use ADLO and LILO.  Eventually I'll move to FILO but 
right now I have my hands, feet and head buried under tyring to get an 
ATI M1 chip up without a bios.

ADLO is not strictly necessary.  I used it to run my Assiliaint 69000 
video bios.  Dosen't work with ATI so once I have an ATI solution I 
won't need ADLO.

V1 2 superio chips of the same type

[ron minnich]

I will try to fix v1 config tool today.

ron

V1 2 superio chips of the same type

[Adam Sulmicki]

> I also currently use ADLO and LILO.

I'm curious, how many people actually use ADLO?  I had been recently
speaking with Kevin Lawton (of BOCHS fame) and he was curious about it. He
also mentioned that BOCHS runs Windows XP (well) those days and was
curious if this success has been transposed into ADLO as well.

V1 2 superio chips of the same type

[Richard Smith]

Adam Sulmicki wrote:
> > I also currently use ADLO and LILO.
>
> I'm curious, how many people actually use ADLO?  I had been recently
> speaking with Kevin Lawton (of BOCHS fame) and he was curious about it. He
> also mentioned that BOCHS runs Windows XP (well) those days and was
> curious if this success has been transposed into ADLO as well.

Does the BIOS of later versions of BOCHS handle video int10 replacement 
correctly?  That would make my life a whole lot easier right now.

V1 2 superio chips of the same type

[Adam Sulmicki]

> > > I also currently use ADLO and LILO.
> >
> > I'm curious, how many people actually use ADLO?  I had been recently
> > speaking with Kevin Lawton (of BOCHS fame) and he was curious about it. He
> > also mentioned that BOCHS runs Windows XP (well) those days and was
> > curious if this success has been transposed into ADLO as well.
>
> Does the BIOS of later versions of BOCHS handle video int10 replacement
> correctly?  That would make my life a whole lot easier right now.

uh, I don't remember it being problem i the first place.

Brian Gregor  writes:
> Is it possible to use LinuxBIOS to boot Windows 2000 diskless?

Is it possible to boot Windows 2000 diskless?
If someone can document that it is likely possible.

There are also games that can be played with System Management
Mode that may allow you to emulate a hardware device and fool
windows it has a local disk, instead of a Ramdisk or something.

Basically on the LinuxBIOS side you have the source so you can
get it to do anything you can code :)

Eric