Status RM200

Status RM200

[Florian Lohoff]

[-- Attachment #1: Type: text/plain, Size: 317 bytes --]


Hi,
i guess the RM200 parts are untested for at least a year (possibly
even more). Does anyone work on this or does know a working
checkout date ?

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
Nine nineth on september the 9th              Welcome to the new billenium

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Status RM200

[Karsten Merker]

On Mon, Nov 26, 2001 at 08:45:09PM +0100, Florian Lohoff wrote:

> i guess the RM200 parts are untested for at least a year (possibly
> even more). Does anyone work on this or does know a working
> checkout date ?

Ralf has at least made the RM200 support compile again while we
were driving home from Oldenburg :-).
I do not know if it really works though - wrong endianess...

Greetings,
Karsten
-- 
#include <standard_disclaimer>
Nach Paragraph 28 Abs. 3 Bundesdatenschutzgesetz widerspreche ich der Nutzung
oder Uebermittlung meiner Daten fuer Werbezwecke oder fuer die Markt- oder
Meinungsforschung.

Re: Status RM200

[Florian Lohoff]

[-- Attachment #1: Type: text/plain, Size: 689 bytes --]

On Mon, Nov 26, 2001 at 09:34:50PM +0100, Karsten Merker wrote:
> Ralf has at least made the RM200 support compile again while we
> were driving home from Oldenburg :-).
> I do not know if it really works though - wrong endianess...

How can the RM200 port be wrong endianess - The RM200 is bi-endian
thus any endianess would be ok (As long as the port does not assume
a specific endianess except the prom stuff).

AFAIK the RM200 was supported little endian as it was ported to the
Windows NT delivered machines.

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
Nine nineth on september the 9th              Welcome to the new billenium

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Status RM200

[Martin Schulze]

Florian Lohoff wrote:
> On Mon, Nov 26, 2001 at 09:34:50PM +0100, Karsten Merker wrote:
> > Ralf has at least made the RM200 support compile again while we
> > were driving home from Oldenburg :-).
> > I do not know if it really works though - wrong endianess...
> 
> How can the RM200 port be wrong endianess - The RM200 is bi-endian
> thus any endianess would be ok (As long as the port does not assume
> a specific endianess except the prom stuff).

As I remember, you can't switch to "the right" endianess without a support
drivers f*ckup disk - which hasn't appeared on the stage yet.

Regards,

	Joey

-- 
Computers are not intelligent.  They only think they are.

Re: Status RM200

[Florian Lohoff]

[-- Attachment #1: Type: text/plain, Size: 472 bytes --]

On Mon, Nov 26, 2001 at 11:35:48PM +0100, Martin Schulze wrote:
> 
> As I remember, you can't switch to "the right" endianess without a support
> drivers f*ckup disk - which hasn't appeared on the stage yet.
> 

If you have the PartNo of the machine i might be able to get some
system boot-floppies.

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
Nine nineth on september the 9th              Welcome to the new billenium

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Status RM200

[Karsten Merker]

On Mon, Nov 26, 2001 at 11:35:48PM +0100, Martin Schulze wrote:

> > How can the RM200 port be wrong endianess - The RM200 is bi-endian
> > thus any endianess would be ok (As long as the port does not assume
> > a specific endianess except the prom stuff).

Unfortunately the firmware functions are different between little- and
big-endian firmware and there are quite some parts in the RM200 support
which currently do not work (and some even do not compile) on a big-endian
machine due to missing (correct) definitions. Another problem regarding 
the big endian firmware is that nobody seems to have documentation about
it, not even a function vector table.

> As I remember, you can't switch to "the right" endianess without a support
> drivers f*ckup disk - which hasn't appeared on the stage yet.

Right - Ralf had asked around for some disks to convert his LE RM200C to
BE firmware for further development,  but nobody had yet been able to
come up with some. Additionally, AFAICS these disks are different for
different machine types and possibly even for different CPU types
(RM200 EISA vs. RM200C PCI and R4k vs R5k). If anybody knows more about
these issues, any help is welcome.

Greetings,
Karsten
-- 
#include <standard_disclaimer>
Nach Paragraph 28 Abs. 3 Bundesdatenschutzgesetz widerspreche ich der Nutzung
oder Uebermittlung meiner Daten fuer Werbezwecke oder fuer die Markt- oder
Meinungsforschung.

Re: Status RM200

[Ralf Baechle]

On Tue, Nov 27, 2001 at 07:18:10AM +0100, Karsten Merker wrote:

> > > How can the RM200 port be wrong endianess - The RM200 is bi-endian
> > > thus any endianess would be ok (As long as the port does not assume
> > > a specific endianess except the prom stuff).
> 
> Unfortunately the firmware functions are different between little- and
> big-endian firmware and there are quite some parts in the RM200 support
> which currently do not work (and some even do not compile) on a big-endian
> machine due to missing (correct) definitions. Another problem regarding 
> the big endian firmware is that nobody seems to have documentation about
> it, not even a function vector table.

As I understand the big endian firmware is just yet another ARC firmware
implementation with some support for old MIPS firmware style vector tables.

  Ralf

Re: Status RM200

[Florian Lohoff]

[-- Attachment #1: Type: text/plain, Size: 1644 bytes --]

On Tue, Nov 27, 2001 at 07:18:10AM +0100, Karsten Merker wrote:
> On Mon, Nov 26, 2001 at 11:35:48PM +0100, Martin Schulze wrote:
> 
> > > How can the RM200 port be wrong endianess - The RM200 is bi-endian
> > > thus any endianess would be ok (As long as the port does not assume
> > > a specific endianess except the prom stuff).
> 
> Unfortunately the firmware functions are different between little- and
> big-endian firmware and there are quite some parts in the RM200 support
> which currently do not work (and some even do not compile) on a big-endian
> machine due to missing (correct) definitions. Another problem regarding 
> the big endian firmware is that nobody seems to have documentation about
> it, not even a function vector table.

I compiled a kernel yesterday - The only thing comlaining is a
header file for some definitions.

> > As I remember, you can't switch to "the right" endianess without a support
> > drivers f*ckup disk - which hasn't appeared on the stage yet.
> 
> Right - Ralf had asked around for some disks to convert his LE RM200C to
> BE firmware for further development,  but nobody had yet been able to
> come up with some. Additionally, AFAICS these disks are different for
> different machine types and possibly even for different CPU types
> (RM200 EISA vs. RM200C PCI and R4k vs R5k). If anybody knows more about
> these issues, any help is welcome.

They are - The disk type can be determined by the PartNo.

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
Nine nineth on september the 9th              Welcome to the new billenium

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Status RM200

[Ralf Baechle]

On Mon, Nov 26, 2001 at 08:45:09PM +0100, Florian Lohoff wrote:

> i guess the RM200 parts are untested for at least a year (possibly
> even more). Does anyone work on this or does know a working
> checkout date ?

I'm only maintaining in the sense of making sure the default configuration
still compiles.  The last kernel I know is working is 2.1.73 which needs
some patches that are not in the standard CVS.  Even if the RM200C is
broken, it's a relativly sane system and so fixing should be fairly easy.

  Ralf

Re: Status RM200

[Florian Lohoff]

[-- Attachment #1: Type: text/plain, Size: 573 bytes --]

On Tue, Nov 27, 2001 at 05:19:50PM +1100, Ralf Baechle wrote:
> I'm only maintaining in the sense of making sure the default configuration
> still compiles.  The last kernel I know is working is 2.1.73 which needs
> some patches that are not in the standard CVS.  Even if the RM200C is
> broken, it's a relativly sane system and so fixing should be fairly easy.

BTW: Do you still have those patches ?

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
Nine nineth on september the 9th              Welcome to the new billenium

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Status RM200

[Ralf Baechle]

On Tue, Nov 27, 2001 at 05:39:47PM +0100, Florian Lohoff wrote:

> On Tue, Nov 27, 2001 at 05:19:50PM +1100, Ralf Baechle wrote:
> > I'm only maintaining in the sense of making sure the default configuration
> > still compiles.  The last kernel I know is working is 2.1.73 which needs
> > some patches that are not in the standard CVS.  Even if the RM200C is
> > broken, it's a relativly sane system and so fixing should be fairly easy.
> 
> BTW: Do you still have those patches ?

Yes, on a machine that is on the other side of the globe ...  As I remember
those patches were mostly stuff like endianess fixes to the NCR driver
which is looking very different today.  Nothing that would still apply
these days.

  Ralf

Re: o200 oops

[Ralf Baechle]

On Sun, Dec 02, 2001 at 09:56:07PM -0500, nick@snowman.net wrote:

> unknown ioctl: 00005382

Request to users of the 64-bit kernel - if you get a kernel message like
this one please tell me drop me a mail.  Would be great if you can
actually find the symbolic name of the ioctl triggering this message.

  Ralf