-= MVME 2300 =-

-= MVME 2300 =-

[Lionel TAILLURAT]

Hi,

I have some problems using Linux with the MVME 2300.

I'm using the debian potato for Prep (http://www.debian.org/~porter/).
The installation and boot are ok. The problems are  :

1- Universe VME driver.

I have used these drivers :
http://lisa2.physik.uni-bonn.de/~hannappe/software/index.html
http://www.vmelinux.org/

- The first could be loaded but when I'm looking at /dev/vme_stat the
informations are bad (fffffff ...)
- The second don't load and return me an error with PCI_ID

So I had decided to try the MVME2600 kernel image of Gabriel Paubert, this
one don't boot (PC keyb error, there is no keyboard controller on the
mvme2300! ).

Seeing that, I try to compile a new kernel using the kernel patch created by
Gabriel Paubert, and the second second problem appears !

2- SCSI drivers.

I'm using a Technobox 2044 PMC (ncr53c875) as SCSI device.
The kernel provided by Debian see well my SCSI device.

When I want to compile a new kernel using the NCR53c8xx driver, this one
don't  seem to see my SCSI device ( the scsi driver is compiled in the
kernel and not as a module).

The only message I have is :
scsi : 0 hosts.
scsi : detected total.

And I should have that :
...
ncr53c8xx: at PCI bus 0, device 16, function 0
ncr53c8xx: 53c875 detected
ncr53c875-0: rev=0x04, base=0x2ffeee00, io_port=0xfffde00, irq=25
ncr53c875-0: NCR clock is 40218KHz, 40218KHz
ncr53c875-0: ID 7, Fast-20, Parity Checking
ncr53c875-0: on-chip RAM at 0x2ffed000
ncr53c875-0: restart (scsi reset).
ncr53c875-0: Downloading SCSI SCRIPTS.
scsi0 : ncr53c8xx - version 3.2a-2
scsi : 1 host.
  Vendor: IBM       Model: DNES-309170       Rev: SAH0
  Type:   Direct-Access                      ANSI SCSI revision: 03
Detected scsi disk sda at scsi0, channel 0, id 0, lun 0
  Vendor: IBM       Model: DNES-309170       Rev: SAH0
  Type:   Direct-Access                      ANSI SCSI revision: 03
Detected scsi disk sdb at scsi0, channel 0, id 1, lun 0
ncr53c875-0-<0,0>: tagged command queue depth set to 4
ncr53c875-0-<1,0>: tagged command queue depth set to 4
scsi : detected 2 SCSI disks total.
ncr53c875-0-<0,*>: FAST-5 SCSI 5.0 MB/s (200 ns, offset 16)
SCSI device sda: hdwr sector= 512 bytes. Sectors= 17916240 [8748 MB] [8.7
GB]
ncr53c875-0-<1,*>: FAST-5 SCSI 5.0 MB/s (200 ns, offset 16)
SCSI device sdb: hdwr sector= 512 bytes. Sectors= 17916240 [8748 MB] [8.7
GB]
...

I had tested differents configurations of the scsi drivers and low SCSI
NCR53c8xx driver : No way !


How can I do to make my MVME 2300 work with Linux ?

Thank's in advance.
-------------------------------------------------------
Lionel TAILLURAT
TECHNICAL SERVICE

ECRIN AUTOMATISMES -industrial computer's-
Rue du pré-de-l'Horme
B.P 15
38921 CROLLES CEDEX
FRANCE
Tel :(33) 4 76 92 20 01
Fax :(33) 4 76 08 03 18
Email: Lionel.TAILLURAT@ecrin.com
---------------------------------------------------------

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: -= MVME 2300 =-

[Ben Sapp]

Lionel TAILLURAT wrote:
> So I had decided to try the MVME2600 kernel image of Gabriel Paubert, this
> one don't boot (PC keyb error, there is no keyboard controller on the
> mvme2300! ).
>
> Seeing that, I try to compile a new kernel using the kernel patch created by
> Gabriel Paubert, and the second second problem appears !
>

I am using Gabriel's patches for the MVME2301  with out any problems.
You might try specifying "console=/dev/ttyS0"  when you get to the
bootlooader prompt.(remove the double quotes)

> 2- SCSI drivers.
>

I don't understand, the MVME 2300 series does not have a SCSI
controller.  Why do you want a SCSI driver?

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: -= MVME 2300 =-

[Gabriel Paubert]

On Wed, 12 Jul 2000, Lionel TAILLURAT wrote:

>
> Hi,
>
> I have some problems using Linux with the MVME 2300.
>
> I'm using the debian potato for Prep (http://www.debian.org/~porter/).
> The installation and boot are ok. The problems are  :
>
> 1- Universe VME driver.
>
> I have used these drivers :
> http://lisa2.physik.uni-bonn.de/~hannappe/software/index.html
> http://www.vmelinux.org/
>
> - The first could be loaded but when I'm looking at /dev/vme_stat the
> informations are bad (fffffff ...)
> - The second don't load and return me an error with PCI_ID
>
> So I had decided to try the MVME2600 kernel image of Gabriel Paubert, this
> one don't boot (PC keyb error, there is no keyboard controller on the
> mvme2300! ).

The configurations I have are for 2600 with a PMC graphics board, remove
CONFIG_VT from the setup and it will use a serial console.

BTW use PrePboot for these boards (CONFIG_PREPBOOT) since it gives a nicer
PCI layout to access the VMEbus...

Please use the patches mvme2600.*.bz2...

> Seeing that, I try to compile a new kernel using the kernel patch created by
> Gabriel Paubert, and the second second problem appears !

My patches include a VME driver, which I consider of course superior to
the other ones, sincee it  allows you to load modules for board specific
drivers (the universe driver exports service to allow you to catch
interrupts, perform DMA etc...). The other drivers I saw did not (last
time I looked) were too limited in interrupt processing for my
applications.

>
> 2- SCSI drivers.
>
> I'm using a Technobox 2044 PMC (ncr53c875) as SCSI device.
> The kernel provided by Debian see well my SCSI device.
>
> When I want to compile a new kernel using the NCR53c8xx driver, this one
> don't  seem to see my SCSI device ( the scsi driver is compiled in the
> kernel and not as a module).
>
> The only message I have is :
> scsi : 0 hosts.
> scsi : detected total.
>
> And I should have that :
> ...
> ncr53c8xx: at PCI bus 0, device 16, function 0
> ncr53c8xx: 53c875 detected
> ncr53c875-0: rev=0x04, base=0x2ffeee00, io_port=0xfffde00, irq=25
> ncr53c875-0: NCR clock is 40218KHz, 40218KHz
> ncr53c875-0: ID 7, Fast-20, Parity Checking
> ncr53c875-0: on-chip RAM at 0x2ffed000
> ncr53c875-0: restart (scsi reset).
> ncr53c875-0: Downloading SCSI SCRIPTS.
> scsi0 : ncr53c8xx - version 3.2a-2
> scsi : 1 host.
>   Vendor: IBM       Model: DNES-309170       Rev: SAH0
>   Type:   Direct-Access                      ANSI SCSI revision: 03
> Detected scsi disk sda at scsi0, channel 0, id 0, lun 0
>   Vendor: IBM       Model: DNES-309170       Rev: SAH0
>   Type:   Direct-Access                      ANSI SCSI revision: 03
> Detected scsi disk sdb at scsi0, channel 0, id 1, lun 0
> ncr53c875-0-<0,0>: tagged command queue depth set to 4
> ncr53c875-0-<1,0>: tagged command queue depth set to 4
> scsi : detected 2 SCSI disks total.
> ncr53c875-0-<0,*>: FAST-5 SCSI 5.0 MB/s (200 ns, offset 16)
> SCSI device sda: hdwr sector= 512 bytes. Sectors= 17916240 [8748 MB] [8.7
> GB]
> ncr53c875-0-<1,*>: FAST-5 SCSI 5.0 MB/s (200 ns, offset 16)
> SCSI device sdb: hdwr sector= 512 bytes. Sectors= 17916240 [8748 MB] [8.7
> GB]
> ...
>
> I had tested differents configurations of the scsi drivers and low SCSI
> NCR53c8xx driver : No way !


My systems work with an SCSI disk (the basic 2600 NCR53c825).
Did you try to directly boot one of the images I have on my ftp site ?

(If it works, you've got a configuration problem.)

	Gabriel


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: -= MVME 2300 =-

[Crusman]

>The configurations I have are for 2600 with a PMC graphics board, remove
>CONFIG_VT from the setup and it will use a serial console.
>
>BTW use PrePboot for these boards (CONFIG_PREPBOOT) since it gives a nicer
>PCI layout to access the VMEbus...
>
>Please use the patches mvme2600.*.bz2...

I have tested the patches but the new kernel don't see the scsi device...so
I can't try it anyway !


>My systems work with an SCSI disk (the basic 2600 NCR53c825).
>Did you try to directly boot one of the images I have on my ftp site ?
>
>(If it works, you've got a configuration problem.)
>
> Gabriel

Yes, I have tested your kernel, but there is a message ( PC keyb error... or
something like this) wich lock the kernel during the boot ! It's probably
due to the Virtual Terminal support in the kernel !
I always use the "console=ttyS0,9600" kernel option.

In short :

- I can't use your kernel because of the PC_key error !
- I can't patch & build a new kernel because this new kernel don't see my
SCSI PMC !

What can I do ?

Lionel

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/