Re: challenge s boots linux

Re: challenge s boots linux

[Rob Lembree]

hmm, I had the same problem on my Indy.  I found that 
IRIX had stripped the major and minor numbers from the
device files in the tar distribution -- the answer was to
untar using Linux, not IRIX (gtar doesn't help), or to 
manually recreate the dev files.

Try that, and I bet it fixes it.

-- 

Rob Lembree                Strategic Software Organization

Silicon Graphics, Inc.                  http://www.sgi.com
One Cabot Road                             rob@lembree.com
Hudson, MA 01749                           lembree@sgi.com
Phone: 978-567-2141                      FAX: 978-567-2341

PGP: 1F EE F8 58 30 F1 B1 20  C5 4F 12 21 AD 0D 6B 29

challenge s boots linux

[Richard Hartensveld]

Hi,

I have successfully booted linux on a challenge S. but the problem is
that the kernel wants to write his output to the console, which isn't
mij serial port unfortunatelly.

Does anyone have a kernel for me from which i can continue my quest of
installing linux on a challenge s ? :)

Below you'll find my kernel booting information for whom is interested.

Regards,

Richard Hartensveld

Command Monitor.  Type "exit" to return to the menu.
>> boot bootp():/vmlinux
72912+9440+3024+331696+23768d+3644+5808 entry: 0x8ff9a950
Setting $netaddr to 192.168.6.31 (from server perron-null)
Obtaining /vmlinux from server perron-null
PROMLIB: SGI ARCS firmware Version 1 Revision 10
PROMLIB: Total free ram 131379200 bytes (128300K,125MB)
ARCH: SGI-IP22
CPU: MIPS-R5000 FPU<MIPS-R5000FPC> ICACHE DCACHE SCACHE
Loading R4000 MMU routines.
CPU revision is: 00002310
Primary instruction cache 32kb, linesize 32 bytes)
Primary data cache 32kb, linesize 32 bytes)
Linux version 2.1.100 (root@alex3.med.iacnet.com) (gcc version 2.7.2)
#29 Thu Ju
l 9 22:19:39 EDT 1998
MC: SGI memory controller Revision 3
R4600/R5000 SCACHE size 512K, linesize 32 bytes.
calculating r4koff... 000dbd5e(900446)
zs0: console input
zs0: console I/O
Calibrating delay loop... 179.81 BogoMIPS
Memory: 125208k/261732k available (1216k kernel code, 3456k data)
Swansea University Computer Society NET3.039 for Linux 2.1
NET3: Unix domain sockets 0.16 for Linux NET3.038.
Swansea University Computer Society TCP/IP for NET3.037
IP Protocols: ICMP, UDP, TCP
Checking for 'wait' instruction...  available.
POSIX conformance testing by UNIFIX
Starting kswapd v 1.5
Adv: about to run setup()
initialize_kbd: Keyboard reset failed, no ACK
SGI Zilog8530 serial driver version 1.00
Keyboard timeout
Keyboard timeout
tty00 at 0xbfbd9838 (irq = 21) is a Zilog8530
tty01 at 0xbfbd9830 (irq = 21) is a Zilog8530
Ramdisk driver initialized : 16 ramdisks of 4096K size
loop: registered device at major 7
WD93: Driver version 1.25 compiled on Jul  7 1998 at 16:59:57
 debug_flags=0x00
wd33c93-0: chip=WD33c93B/13 no_sync=0xff no_dma=0scsi0 : SGI WD93
scsi : 1 host.
 sending SDTR 0103013f0csync_xfer=2c  Vendor: SGI       Model: IBM
DORS-32160
 Rev: W80D
  Type:   Direct-Access                      ANSI SCSI revision: 02

 Rev: W80D
  Type:   Direct-Access                      ANSI SCSI revision: 02
Detected scsi disk sda at scsi0, channel 0, id 1, lun 0
 sending SDTR 0103013f0csync_xfer=2c  Vendor: IBM       Model:
DCAS-34330
 Rev: S65A
  Type:   Direct-Access                      ANSI SCSI revision: 02
Detected scsi disk sdb at scsi0, channel 0, id 6, lun 0
scsi : detected 2 SCSI disks total.
SCSI device sda: hdwr sector= 512 bytes. Sectors= 4197405 [2049 MB] [2.0

GB]
SCSI device sdb: hdwr sector= 512 bytes. Sectors= 8467200 [4134 MB] [4.1

GB]
sgiseeq.c: David S. Miller (dm@engr.sgi.com)
eth0: SGI Seeq8003 08:00:69:0a:41:7b
Sending BOOTP requests.... OK
IP-Config: Got BOOTP answer from 192.168.6.1, my address is 192.168.6.31

Partition check:
 sda: sda1 sda2 sda3 sda4
 sdb: sdb1 sdb2 sdb3 sdb4
Looking up port of RPC 100003/2 on 192.168.6.1
Looking up port of RPC 100005/1 on 192.168.6.1
VFS: Mounted root (nfs filesystem).
Adv: done running setup()
Freeing unused kernel memory: 44k freed
Warning: unable to open an initial console.

Re: challenge s boots linux

[Richard Hartensveld]

Rob Lembree wrote:

> hmm, I had the same problem on my Indy.  I found that
> IRIX had stripped the major and minor numbers from the
> device files in the tar distribution -- the answer was to
> untar using Linux, not IRIX (gtar doesn't help), or to
> manually recreate the dev files.

I've untarred the hardhat distr. on my linux machine and that's the
machine that plays nfs server for the
challenge.

Should the default kernel support serial terminal consoles?? My goal is
to install linux onto the challenge
(which is headless) using a serial terminal connected to it, is this
possible by default or should i cross-compile my own kernel on another
platform?.

Regards,

Richard Hartensveld
The Netherlands

Re: challenge s boots linux

[Alex deVries]

On Tue, 22 Sep 1998, Richard Hartensveld wrote:
> Should the default kernel support serial terminal consoles?? My goal is
> to install linux onto the challenge
> (which is headless) using a serial terminal connected to it, is this
> possible by default or should i cross-compile my own kernel on another
> platform?.

Rich,

	First, I wonder how much of the kernel will actually work on the
Challenge S; I'm told there are SCSI controller differences.

	The install kernel you're using (and oh god, is it ever a bad one)
is broken in that serial console won't work.  Yes, I should have looked
into it way back when.  There's two things you can do:

- rebuild a kernel by cross compiling in the correct options (this is
reasonably easy to do)
- wait for me to have my machine moved (finally) and online and I'll whip
up a modern install kernel for you and the others with serial console

Just to reiterate, yes, I *will* be fixing the install.  The initrd stuff
is going to have to make it in, even if it kills me. And I'll be working
on porting Rawhide 1.0 too.

- Alex

Re: challenge s boots linux

[Thomas Bogendoerfer]

On Tue, Sep 22, 1998 at 08:24:20PM -0400, Alex deVries wrote:
> 	First, I wonder how much of the kernel will actually work on the
> Challenge S; I'm told there are SCSI controller differences.

looking linus.linux.sgi.com, which is AFAIK a Challenge S, it seems the
Challenge has a wd33c93 and a wd33c95. The wd33c93 is what we already support.
As long as the harddisk are attached to it, it should work. The wd33c95
is a rather strange chip, and from the datasheet I have here, it looks like
a lot of work to write a driver for it, because besides the kernel code,
you have to write sequence code for the chip, too.

> 	The install kernel you're using (and oh god, is it ever a bad one)
> is broken in that serial console won't work.  Yes, I should have looked
> into it way back when.

When I did the new console code, I first made the serial console working.
As far as I remember I had to do some ugly hacks, but I'll try to clean
it up and check it in.

Thomas.

-- 
See, you not only have to be a good coder to create a system like Linux,
you have to be a sneaky bastard too ;-)
                   [Linus Torvalds in <4rikft$7g5@linux.cs.Helsinki.FI>]

Re: challenge s boots linux

[Richard Hartensveld]

Thomas Bogendoerfer wrote:

> On Tue, Sep 22, 1998 at 08:24:20PM -0400, Alex deVries wrote:
> >       First, I wonder how much of the kernel will actually work on the
> > Challenge S; I'm told there are SCSI controller differences.
>
> looking linus.linux.sgi.com, which is AFAIK a Challenge S, it seems the
> Challenge has a wd33c93 and a wd33c95. The wd33c93 is what we already support.
> As long as the harddisk are attached to it, it should work. The wd33c95
> is a rather strange chip, and from the datasheet I have here, it looks like
> a lot of work to write a driver for it, because besides the kernel code,
> you have to write sequence code for the chip, too.
>
> >       The install kernel you're using (and oh god, is it ever a bad one)
> > is broken in that serial console won't work.  Yes, I should have looked
> > into it way back when.
>
> When I did the new console code, I first made the serial console working.
> As far as I remember I had to do some ugly hacks, but I'll try to clean
> it up and check it in.
>

I'm going to build the cross compiler right now and get myself involved in
linux/sgi porting, i've got
a reasonable amount of sgi's overhere to test on and maybe even get linux to boot
a kernel on a non-indy machine.

Richard

Re: challenge s boots linux

[Alex deVries]

On Wed, 23 Sep 1998, Richard Hartensveld wrote:
> I'm going to build the cross compiler right now and get myself involved in
> linux/sgi porting, i've got

There are actually cross compiler RPMs that may save you some trouble...

Let us know how the Challenge S works!

- Alex

Re: challenge s boots linux

[Richard Masoner]

> > 	First, I wonder how much of the kernel will actually work on the
> > Challenge S; I'm told there are SCSI controller differences.
> 
> looking linus.linux.sgi.com, which is AFAIK a Challenge S, it seems the
> Challenge has a wd33c93 and a wd33c95.

Correct.  Challenge S has the same wd33c93 SCSI chip on it that the
Indy has.  Additionally, it has the wd33c95 chip for more SCSI buses
in place of of video and audio that's on the Indy.

--
Richard F. Masoner                         Digi International
richardm@cd.com                             1602 Newton Drive
http://www.cd.com/                    Champaign, IL 61822 USA 

Re: challenge s boots linux

[ralf]

On Wed, Sep 23, 1998 at 09:22:41PM +0200, Thomas Bogendoerfer wrote:

> On Tue, Sep 22, 1998 at 08:24:20PM -0400, Alex deVries wrote:
> > 	First, I wonder how much of the kernel will actually work on the
> > Challenge S; I'm told there are SCSI controller differences.
> 
> looking linus.linux.sgi.com, which is AFAIK a Challenge S, it seems the
> Challenge has a wd33c93 and a wd33c95. The wd33c93 is what we already support.
> As long as the harddisk are attached to it, it should work. The wd33c95
> is a rather strange chip, and from the datasheet I have here, it looks like
> a lot of work to write a driver for it, because besides the kernel code,
> you have to write sequence code for the chip, too.

Did I missunderstand the 95's documentation - I thought we can abuse the
95 as a 93, thereby saving alot of driver work for now?

  Ralf

Re: challenge s boots linux

[Richard Hartensveld]

ralf@uni-koblenz.de wrote:

> On Wed, Sep 23, 1998 at 09:22:41PM +0200, Thomas Bogendoerfer wrote:
>
> > On Tue, Sep 22, 1998 at 08:24:20PM -0400, Alex deVries wrote:
> > >     First, I wonder how much of the kernel will actually work on the
> > > Challenge S; I'm told there are SCSI controller differences.
> >
> > looking linus.linux.sgi.com, which is AFAIK a Challenge S, it seems the
> > Challenge has a wd33c93 and a wd33c95. The wd33c93 is what we already support.
> > As long as the harddisk are attached to it, it should work. The wd33c95
> > is a rather strange chip, and from the datasheet I have here, it looks like
> > a lot of work to write a driver for it, because besides the kernel code,
> > you have to write sequence code for the chip, too.
>
> Did I missunderstand the 95's documentation - I thought we can abuse the
> 95 as a 93, thereby saving alot of driver work for now?
>

My experience is that the wd93 driver doens't work with the wd95 controller, maybe
someone else hassuccessfully tried this?

Richard

Re: challenge s boots linux

[ralf]

On Thu, Sep 24, 1998 at 02:17:08AM +0200, Richard Hartensveld wrote:

> > Did I missunderstand the 95's documentation - I thought we can abuse the
> > 95 as a 93, thereby saving alot of driver work for now?
> >
> 
> My experience is that the wd93 driver doens't work with the wd95 controller, maybe
> someone else hassuccessfully tried this?

It's not supposed to work without any changes at all.

  Ralf

Re: challenge s boots linux

[Jeremy John Welling]

What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?  
What is the status of the newport Xserver? What is the best place to get
specific hardware info?  Has anyone started on the Indy XZ graphics?  
Thanks!
Jeremy Welling
Detroit

Re: challenge s boots linux

[Robin Humble]

>Warning: unable to open an initial console.

I got this message for (it seemed) a variety of reasons. One was that it
couldn't find the nfs server, but in the end it was that I was
booting off an O2 and the permissions on the dev/ directory were all
screwed from SGIs tar. Re-building dev to match a x86/Linux machine
fixed it.
So it's a long shot, but check that your dev directory in the nfs
mounted distribution looks like:

crw-------   1 root     root       4,   0 Sep 13 20:05 console
crw-rw-rw-   1 root     root       1,   3 Jan  1  1980 null
brw-r-----   1 root     disk       1,   1 Jan  1  1980 ram
crw-------   1 root     root       4,   0 Sep  3  1995 systty
crw-------   1 root     root       4,   1 Sep 13 20:05 tty1
crw-------   1 root     root       4,   2 Sep 13 20:05 tty2
crw-------   1 root     root       4,   3 Sep 13 20:05 tty3
crw-------   1 root     root       4,   4 Sep 13 20:05 tty4
crw-------   1 root     root       4,   5 Sep 13 20:05 tty5

+
There's not a line that goes here that  +         Robin Humble
rhymes with anything, anything,      /     rjh@pixel.maths.monash.edu.au
anything. - Camper van Beethoven  + http://www.maths.monash.edu.au/~rjh/

Re: challenge s boots linux

[Robin Humble]

>Warning: unable to open an initial console.

I got this message for (it seemed) a variety of reasons. One was that it
couldn't find the nfs server, but in the end it was that I was
booting off an O2 and the permissions on the dev/ directory were all
screwed from SGIs tar. Re-building dev to match a x86/Linux machine
fixed it.
So it's a long shot, but check that your dev directory in the nfs
mounted distribution looks like:

crw-------   1 root     root       4,   0 Sep 13 20:05 console
crw-rw-rw-   1 root     root       1,   3 Jan  1  1980 null
brw-r-----   1 root     disk       1,   1 Jan  1  1980 ram
crw-------   1 root     root       4,   0 Sep  3  1995 systty
crw-------   1 root     root       4,   1 Sep 13 20:05 tty1
crw-------   1 root     root       4,   2 Sep 13 20:05 tty2
crw-------   1 root     root       4,   3 Sep 13 20:05 tty3
crw-------   1 root     root       4,   4 Sep 13 20:05 tty4
crw-------   1 root     root       4,   5 Sep 13 20:05 tty5

+
There's not a line that goes here that  +         Robin Humble
rhymes with anything, anything,      /     rjh@pixel.maths.monash.edu.au
anything. - Camper van Beethoven  + http://www.maths.monash.edu.au/~rjh/

Re: challenge s boots linux

[Thomas Bogendoerfer]

On Thu, Sep 24, 1998 at 01:59:38AM +0200, ralf@uni-koblenz.de wrote:
> > As long as the harddisk are attached to it, it should work. The wd33c95
> > is a rather strange chip, and from the datasheet I have here, it looks like
> > a lot of work to write a driver for it, because besides the kernel code,
> > you have to write sequence code for the chip, too.
> 
> Did I missunderstand the 95's documentation - I thought we can abuse the
> 95 as a 93, thereby saving alot of driver work for now?

The documentation isn't really clear about that, but I believe you need
sequencer code to make the 95 acting like a 93. And I couldn't find such
a thing, yet.

Thomas.

-- 
See, you not only have to be a good coder to create a system like Linux,
you have to be a sneaky bastard too ;-)
                   [Linus Torvalds in <4rikft$7g5@linux.cs.Helsinki.FI>]

Re: challenge s boots linux

[Richard Hartensveld]

Jeremy John Welling wrote:

> What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?
> What is the status of the newport Xserver? What is the best place to get
> specific hardware info?  Has anyone started on the Indy XZ graphics?

Well i've got a indigo2 r4400 with a defective motherboard, when i'm able  to
get this fixed (they're quite expensive) i'll be happy to volunteer in the
development.

Richard

Re: challenge s boots linux

[Alan Cox]

> What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?  
> What is the status of the newport Xserver? What is the best place to get

Im still learning about the internals of the Xserver and building an
"fb" driver layer to glue to the accelerated stuff. I've been dicing up
afb so that it generates newport_something() commands for each area
after having done its clipping. Hopefully they'll approximately match
the real newport 8)

Re: challenge s boots linux

[Alan Cox]

> What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?  
> What is the status of the newport Xserver? What is the best place to get

Im still learning about the internals of the Xserver and building an
"fb" driver layer to glue to the accelerated stuff. I've been dicing up
afb so that it generates newport_something() commands for each area
after having done its clipping. Hopefully they'll approximately match
the real newport 8)

Re: challenge s boots linux

[Alex deVries]

On Tue, 22 Sep 1998, Rob Lembree wrote:
> hmm, I had the same problem on my Indy.  I found that 
> IRIX had stripped the major and minor numbers from the
> device files in the tar distribution -- the answer was to
> untar using Linux, not IRIX (gtar doesn't help), or to 
> manually recreate the dev files.
> Try that, and I bet it fixes it.

You can also just recreate the block devices from within Irix.

- Alex

Re: challenge s boots linux

[Alex deVries]

On Thu, 24 Sep 1998, Jeremy John Welling wrote:
> What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?  
> What is the status of the newport Xserver? What is the best place to get
> specific hardware info?  Has anyone started on the Indy XZ graphics?  

This is an interesting question, and I for one would much rather see
things like the graphics for Newport and audio working on the INdy before
we start off on other things like the Indigo2.

But, i've been thinking for awhile that an Indigo2 port would be the next
box to target... i suppose there's no harm in looking at both machines
concurrently.

Getting docs on the Indigo2 would be quite important for a successful
port, though.  That would require SGI's help, and I can see them being a
lot more open to it if we got the Indy completely supported first.

- Alex

Re: challenge s boots linux

[William J. Earl]

Alex deVries writes:
...
 > But, i've been thinking for awhile that an Indigo2 port would be the next
 > box to target... i suppose there's no harm in looking at both machines
 > concurrently.
 > 
 > Getting docs on the Indigo2 would be quite important for a successful
 > port, though.  That would require SGI's help, and I can see them being a
 > lot more open to it if we got the Indy completely supported first.
...

     Except for the EISA slots, Indigo2 is pretty close to Indy.  They run
the same IRIX kernel binary.  The graphics are more of an issue, since most
Indigo2 systems use something other than Newport ("XL") graphics.  It might
be possible for Linux to get by using the very basic microcode downloaded by
the PROM for XZ and other microcoded graphics cards, but I have not had time
to research the interface presented.  

     There is general willingness to release information, although releasing
Indigo2-specific information would require a new authorization.  The main
problem is finding it in a form useful to people outside SGI.  (That is not
true for all hardware, but it is for quite a few of our older systems.)
The easiest path would be if someone inside SGI cobbled up basic sample
drivers for the various graphics cards, but so far there are no volunteers.
(The reason it would be the easiest path is that SGI employees can use the
IRIX source as a guide.)

Re: challenge s boots linux

[Richard Hartensveld]

William J. Earl wrote:

> Alex deVries writes:
> ...
>  > But, i've been thinking for awhile that an Indigo2 port would be the next
>  > box to target... i suppose there's no harm in looking at both machines
>  > concurrently.
>  >
>  > Getting docs on the Indigo2 would be quite important for a successful
>  > port, though.  That would require SGI's help, and I can see them being a
>  > lot more open to it if we got the Indy completely supported first.
> ...
>
>      Except for the EISA slots, Indigo2 is pretty close to Indy.  They run
> the same IRIX kernel binary.  The graphics are more of an issue, since most
> Indigo2 systems use something other than Newport ("XL") graphics.  It might
> be possible for Linux to get by using the very basic microcode downloaded by
> the PROM for XZ and other microcoded graphics cards, but I have not had time
> to research the interface presented.
>

If i'll get my indigo2 with extreme gfx back to work, i'll see how far i can
come with a kernel that usesa serial terminal for it's output. (no gfx card
needed there).

But first get my crosscompiler to work properly ;]

Richard

Cool

[Jeremy John Welling]

I agree a complete working port for Indy should be the first priority.  I
haven't been following this port till recently.  But now I'm running it on
my Indy.  Very cool.  I haven't seen too many congrats to the core team on
this list, but good job people!  Thanks for the hard work.

Jeremy Welling

On Thu, 24 Sep 1998, Alex deVries wrote:

> 
> On Thu, 24 Sep 1998, Jeremy John Welling wrote:
> > What is the next architecture sgi/linux will be ported to?  Indigo2 r4kXZ?  
> > What is the status of the newport Xserver? What is the best place to get
> > specific hardware info?  Has anyone started on the Indy XZ graphics?  
> 
> This is an interesting question, and I for one would much rather see
> things like the graphics for Newport and audio working on the INdy before
> we start off on other things like the Indigo2.
> 
> But, i've been thinking for awhile that an Indigo2 port would be the next
> box to target... i suppose there's no harm in looking at both machines
> concurrently.
> 
> Getting docs on the Indigo2 would be quite important for a successful
> port, though.  That would require SGI's help, and I can see them being a
> lot more open to it if we got the Indy completely supported first.
> 
> - Alex
> 
>