[Q] ultrapenguin-1.1 on UE4000

[Q] ultrapenguin-1.1 on UE4000

[ebrower]

I've managed access to an Ultra Enterprise 4000 for an unknown
duration-- a day, a week, a month-- basically until it is needed.

I MAY be suffering from the same memory-slot problem as described
by Svante Sormark earlier today on the list.

I've attempted to install UltraPenguin-1.1 from Jakub's CD image,
but to no avail.  I'll be happy to put some spare-time work into
getting this puppy running Linux, but I'll probably need some
divine intervention regarding the following boot-sequence error
message.  Here's the entire boot sequence, with the error and banner
info at the end.
  
I MAY be suffering from the same memory-slot problem as described
by Svante Sormark earlier today on the list.
To make a long story short, I get the following:

	Partition check:
	request_module[block-major-11]: Root fs not mounted
	VFS: Cannot open root device 0b:00
	Kernel panic: VFS: Unable to mount root fs on 0b:00
	Press L1-A to return to the boot prom
	
At times, the boot gets one stage further and lists the partitions
found on the disks.

Here's the works:
-----------------

{0} ok boot cdrom
Boot device: /sbus@3,0/SUNW,fas@3,8800000/sd@6,0:f  File and args: 
SILO 

Welcome to UltraPenguin Linux!

You have just booted to the SILO prompt on a SUN4U 8-slot Sun Enterprise 
4000/5000.

You probably just need to hit "Enter" and you will have the install
loaded soon.

boot: 
Uncompressing image...

root child is f002cf50
PROMLIB: Sun IEEE Boot Prom 3.2.16 1998/06/08 16:58
Linux version 2.1.126 (root@manka) (gcc version egcs-2.92.07 19980915 (gcc2 
ss-980609 experimental)) #1 Wed Nov 4 20:30:31 1998
ARCH: SUN4U
Ethernet address: 08:00:20:xx:xx:xx
CENTRAL: found central PROM node f005df90.
FHC Version[1] PartID[fa0] Manufacturer[3e]
Calibrating delay loop... 494.80 BogoMIPS
initrd overwritten (0xfffff80000990000 < 0xfffff80000d2a358) - disabling it.
Memory: 1022912k available (1304k kernel code, 11968k data, 128k init) 
[fffff80000000000,fffff8003fcf8000]
POSIX conformance testing by UNIFIX
PCI: Probing for controllers.
PCI: No PCI bus detected
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9c600000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[000000003fcf20a0] ... ENABLED
sbus0: Clock 25.0 MHz
dma0: HME DVMA gate array 
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9cc00000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[000000003fcf21c0] ... ENABLED
sbus1: Clock 25.0 MHz
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9ce00000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[000000003fcf2280] ... ENABLED
sbus2: Clock 25.0 MHz
dma1: HME DVMA gate array 
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9d400000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[000000003fcf23a0] ... ENABLED
sbus3: Clock 25.0 MHz
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9d600000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[000000003fcf2460] ... ENABLED
sbus4: Clock 25.0 MHz
dma2: HME DVMA gate array 
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
Starting kswapd v 1.5 
Sparc Zilog8530 serial driver version 1.31
tty00 at 0xfff66004 (irq = 12,39) is a Zilog8530
tty01 at 0xfff66000 (irq = 12,39) is a Zilog8530
tty02 at 0xffd9e004 (irq = 12,39) is a Zilog8530
tty03 at 0xffd9e000 (irq = 12,39) is a Zilog8530
keyboard: not present
Console: ttyS0 (Zilog8530)
Sun Mouse-Systems mouse driver version 1.00
RAM disk driver initialized:  16 RAM disks of 4096K size
md driver 0.36.6 MAX_MD_DEV=4, MAX_REAL=8
esp0: IRQ 7,db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp0: FAST chip is fasHME (family\x10, version=5)
esp1: IRQ 7,1db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp1: FAST chip is fasHME (family\x10, version=5)
esp2: IRQ 7,2db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp2: FAST chip is fasHME (family\x10, version=5)
ESP: Total of 3 ESP hosts found, 3 actually in use.
scsi0 : Sparc ESP366-HME
scsi1 : Sparc ESP366-HME
scsi2 : Sparc ESP366-HME
scsi : 3 hosts.
scsi : detected total.
Partition check:
request_module[block-major-11]: Root fs not mounted
VFS: Cannot open root device 0b:00
Kernel panic: VFS: Unable to mount root fs on 0b:00
Press L1-A to return to the boot prom

Type  'go' to resume
{0} ok 
{0} ok banner

8-slot Sun Enterprise 4000/5000, No Keyboard
OpenBoot 3.2.16, 1024 MB memory installed, Serial #xxxxxxxx.
Ethernet address x:x:xx:xx:xx:xx, Host ID: xxxxxxxx.

Re: [Q] ultrapenguin-1.1 on UE4000

[David S. Miller]

Ugh, are all of your disks hung off of the fibre channel cards?
Crap... Jakub, you know why it fails if this is true.

One quick hack would be to hook up a spare SCSI disk to one of the
normal scsi controllers on the machine, or else wait for a fix from
either myself or Jakub.

Later,
David S. Miller
davem@dm.cobaltmicro.com

Re: [Q] ultrapenguin-1.1 on UE4000

[Jakub Jelinek]

> 
> I've managed access to an Ultra Enterprise 4000 for an unknown
> duration-- a day, a week, a month-- basically until it is needed.
> 
> I MAY be suffering from the same memory-slot problem as described
> by Svante Sormark earlier today on the list.

Yes, you suffer from the same problem, but that's not the whole story.
Indeed I'm now hacking SILO to overcome that problem.
What's strange is that no SCSI devices has been found, also, the number of
found SBUSes seems to be weird (I always thought every SBUS board has two
SBUSes).

If you have Solaris on the box, can you please send me
prtconf -pv | mail -s "E4000" jj@ultra.linux.cz

Thanks.

Cheers,
    Jakub
___________________________________________________________________
Jakub Jelinek | jj@sunsite.mff.cuni.cz | http://sunsite.mff.cuni.cz
Administrator of SunSITE Czech Republic, MFF, Charles University
___________________________________________________________________
Ultralinux - first 64bit OS to take full power of the UltraSparc
Linux version 2.1.128 on a sparc64 machine (498.80 BogoMips).
___________________________________________________________________

Re: [Q] ultrapenguin-1.1 on UE4000

[Jakub Jelinek]

> 
> 
> Ugh, are all of your disks hung off of the fibre channel cards?

Don't worry. At least the CD is apparently on some ESP, but I think E?000
boxes don't have FC-AL internal disks, only E?500 boxes do.
I've started already looking into SOC+ manual, plus have read most of the
FC-AL, FC-PLDA and other stuff, but still don't have enough knowledge to
write the socal and fcal drivers without access to the physical hardware.
I should be getting a E3500 in 3 month, but even there I doubt it will have
any FC-AL disks in it, as it is pretty costy (7000$ to buy necessary boards
and one single FC-AL 9GB disk :( ).

> Crap... Jakub, you know why it fails if this is true.
> 
> One quick hack would be to hook up a spare SCSI disk to one of the
> normal scsi controllers on the machine, or else wait for a fix from
> either myself or Jakub.

Cheers,
    Jakub
___________________________________________________________________
Jakub Jelinek | jj@sunsite.mff.cuni.cz | http://sunsite.mff.cuni.cz
Administrator of SunSITE Czech Republic, MFF, Charles University
___________________________________________________________________
Ultralinux - first 64bit OS to take full power of the UltraSparc
Linux version 2.1.128 on a sparc64 machine (498.80 BogoMips).
___________________________________________________________________

Re: [Q] ultrapenguin-1.1 on UE4000

[ebrower]

[DaveM]> Ugh, are all of your disks hung off of the fibre channel cards?

None.  To keep things simple, I've only attached one external drive
to the SCSI on system board two. Other than the CDROM and TAPE, there 
are no peripherals.

I decided not to muck with using the A1000 array, as it probably needs
a non-existent driver for the SBus Wide Intelligent Ultra SCSI Differential 
Host Adapter.
 
[Jakub]> If you have Solaris on the box, can you please send me
[Jakub]> prtconf -pv | mail -s "E4000" jj@ultra.linux.cz

I don't, but I'll slap it on there now.  When an updated SILO is
available, I'll be in line to give it a shot.

E

Re: [Q] ultrapenguin-1.1 on UE4000

[Matthew Jacob]

I have a SOC+ card in a sun4m and a wad of FC disks. If I had time, I'd
write the SOC+ driver- but I really don't have much time. I have the
interest in doing a FC-4 (FCP)- but my current jobs and contracts make it
difficult to find time to even finish the Qlogic FC driver for linux
which has been 'ready to go except for Linux 2.1 && sparc/alpha
crosscheck' for months now.



On Wed, 18 Nov 1998, Jakub Jelinek wrote:

> > 
> > 
> > Ugh, are all of your disks hung off of the fibre channel cards?
> 
> Don't worry. At least the CD is apparently on some ESP, but I think E?000
> boxes don't have FC-AL internal disks, only E?500 boxes do.
> I've started already looking into SOC+ manual, plus have read most of the
> FC-AL, FC-PLDA and other stuff, but still don't have enough knowledge to
> write the socal and fcal drivers without access to the physical hardware.
> I should be getting a E3500 in 3 month, but even there I doubt it will have
> any FC-AL disks in it, as it is pretty costy (7000$ to buy necessary boards
> and one single FC-AL 9GB disk :( ).
> 
> > Crap... Jakub, you know why it fails if this is true.
> > 
> > One quick hack would be to hook up a spare SCSI disk to one of the
> > normal scsi controllers on the machine, or else wait for a fix from
> > either myself or Jakub.
> 
> Cheers,
>     Jakub
> ___________________________________________________________________
> Jakub Jelinek | jj@sunsite.mff.cuni.cz | http://sunsite.mff.cuni.cz
> Administrator of SunSITE Czech Republic, MFF, Charles University
> ___________________________________________________________________
> Ultralinux - first 64bit OS to take full power of the UltraSparc
> Linux version 2.1.128 on a sparc64 machine (498.80 BogoMips).
> ___________________________________________________________________
> 

Re: [Q] ultrapenguin-1.1 on UE4000

[David S. Miller]

   From: Jakub Jelinek <jj@sunsite.ms.mff.cuni.cz>
   Date: 	Wed, 18 Nov 1998 14:31:39 +0100 (MET)

   What's strange is that no SCSI devices has been found, also, the
   number of found SBUSes seems to be weird (I always thought every
   SBUS board has two SBUSes).

Look closely at his prtconf (he sent it to me privately, and I believe
to you too Jakub).  There are 5 types of Ex000 board types I know of:

1) 'cpu'		Ultra CPUs and memory simms
2) 'dual-sbus'		Two SBUS's, embedded SOC and FAS/HME
3) 'upa-sbus'		One UPA (for Creator3D etc.) slot and one SBUS
4) 'dual-sbus-soc+'	Same as dual-sbus except uses SOC+
5) 'pci'		PSYCHO based PCI board

One of his boards is a upa-sbus one, thus the odd number of SBUS's.
This still doesn't explain why his SCSI disk was not detected.  All of
his FAS scsi adapters were detected, and the driver did not print any
errors while probing the SCSI bus each time.. hmmm..

Eric, could you send us the output of "probe-scsi-all" from the PROM
prompt on this box?

Later,
David S. Miller
davem@dm.cobaltmicro.com

Re: [Q] ultrapenguin-1.1 on UE4000

[ebrower]

> Eric, could you send us the output of "probe-scsi-all" from the PROM
> prompt on this box?

{0} ok probe-scsi-all
/sbus@b,0/SUNW,fas@3,8800000

/sbus@7,0/SUNW,fas@3,8800000

/sbus@3,0/SUNW,fas@3,8800000
Target 0 
  Unit 0   Disk     SEAGATE ST32430N SUN2.1G050800521699
                    Copyright (c) 1995 Seagate
                    All rights reserved     
Target 5 
  Unit 0   Removable Tape     EXABYTE EXB-8505SMBANSH20098            
                                     1
                              6085290  
Target 6 
  Unit 0   Removable Read Only device     TOSHIBA XM5701TASUN12XCD2395708/27/9
                                          7

Re: [Q] ultrapenguin-1.1 on UE4000

[David S. Miller]

   Date: Thu, 19 Nov 1998 10:15:53 -0800 (PST)
   From: ebrower@usa.net

   > Eric, could you send us the output of "probe-scsi-all" from the PROM
   > prompt on this box?

   {0} ok probe-scsi-all

 ...

Ok thanks, now grab the tftpboot image from:

ftp.cobaltnet.com:/pub/users/davem/for_ebrower/e4000test.img

and tftpboot that and send me the console output, it may be lengthy.

If it works and detects your disks, it will be a step in the right
direction, but this image is not an install image, just a kernel by
itself so it will die when it tries to mount root even if the device
probing works :-)

Let me know how it goes.

Later,
David S. Miller
davem@dm.cobaltmicro.com

Re: [Q] ultrapenguin-1.1 on UE4000

[ebrower]

> Ok thanks, now grab the tftpboot image from:
> 
> ftp.cobaltnet.com:/pub/users/davem/for_ebrower/e4000test.img
> 
> and tftpboot that and send me the console output, it may be lengthy.
> 
> If it works and detects your disks, it will be a step in the right
> direction, but this image is not an install image, just a kernel by
> itself so it will die when it tries to mount root even if the device
> probing works :-)
> 
> Let me know how it goes.

Here's the output from the kernel:

E

---
{0} ok boot net
Boot device: /sbus@3,0/SUNW,hme@3,8c00000  File and args: 
root child is f002cf50
Found CPU 0 (nodeð06f708,mid=0)
Found CPU 1 (nodeð06fac8,mid=1)
Found CPU 2 (nodeð14f6e4,mid=4)
Found CPU 3 (nodeð14faa4,mid=5)
Found CPU 4 (nodeð22f6e4,mid=8)
Found CPU 5 (nodeð22faa4,mid=9)
Found CPU 6 (nodeð30f6e4,mid\x12)
Found CPU 7 (nodeð30faa4,mid\x13)
PROMLIB: Sun IEEE Boot Prom 3.2.16 1998/06/08 16:58
Linux version 2.1.129 (davem@pizda.cobaltnet.com) (gcc driver version 2.7.2 
snapshot 970621 executing gcc version 2.7.2) #4 SMP Thu Nov 19 12:41:20 PST 1998
ARCH: SUN4U
Ethernet address: 08:00:20:9a:71:d5
Found CPU 0 (nodeð06f708,mid=0)
Found CPU 1 (nodeð06fac8,mid=1)
Found CPU 2 (nodeð14f6e4,mid=4)
Found CPU 3 (nodeð14faa4,mid=5)
Found CPU 4 (nodeð22f6e4,mid=8)
Found CPU 5 (nodeð22faa4,mid=9)
Found CPU 6 (nodeð30f6e4,mid\x12)
Found CPU 7 (nodeð30faa4,mid\x13)
Found 8 CPU prom device tree node(s).
CENTRAL: found central PROM node f005df90.
FHC Version[1] PartID[fa0] Manufacturer[3e]
Calibrating delay loop... 494.80 BogoMIPS
Memory: 1022408k available (1616k kernel code, 12000k data, 160k init) 
[fffff80000000000,fffff8003fcd8000]
POSIX conformance testing by UNIFIX
Entering UltraSMPenguin Mode...
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Calibrating delay loop... 494.80 BogoMIPS
Total of 8 processors activated (3958.37 BogoMIPS).
PCI: Probing for controllers.
PCI: No PCI bus detected
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9c600000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[0000000000d78400] ... ENABLED
sbus0: Clock 25.0 MHz
dma0: HME DVMA gate array 
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9cc00000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[0000000000d78520] ... ENABLED
sbus1: Clock 25.0 MHz
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9ce00000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[0000000000d785e0] ... ENABLED
sbus2: Clock 25.0 MHz
dma1: HME DVMA gate array 
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9d400000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[0000000000d78700] ... ENABLED
sbus3: Clock 25.0 MHz
IOMMU(SBUS): IMPL[0] VERS[0] SYSIO mapped at fffff9d600000000
IOMMU: Streaming Buffer IMPL[0] REV[0] FlushFLAG[0000000000d787c0] ... ENABLED
sbus4: Clock 25.0 MHz
dma2: HME DVMA gate array 
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
Starting kswapd v 1.5 
Sparc Zilog8530 serial driver version 1.32
tty00 at 0xfff66004 (irq = 12,339) is a Zilog8530
tty01 at 0xfff66000 (irq = 12,339) is a Zilog8530
tty02 at 0xffd9e004 (irq = 12,339) is a Zilog8530
tty03 at 0xffd9e000 (irq = 12,339) is a Zilog8530
keyboard: not present
Console: ttyS0 (Zilog8530)
Sun Mouse-Systems mouse driver version 1.00
RAM disk driver initialized:  16 RAM disks of 4096K size
md driver 0.36.6 MAX_MD_DEV=4, MAX_REAL=8
esp0: IRQ 7,db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp0: FAST chip is fasHME (family\x10, version=5)
esp1: IRQ 7,1db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp1: FAST chip is fasHME (family\x10, version=5)
esp2: IRQ 7,2db SCSI ID 7  Clock 40 MHz CCF=8 Time-Out 167 NCR53C9XF(espfast) 
detected
esp2: FAST chip is fasHME (family\x10, version=5)
ESP: Total of 3 ESP hosts found, 3 actually in use.
scsi0 : Sparc ESP366-HME
scsi1 : Sparc ESP366-HME
scsi2 : Sparc ESP366-HME
scsi : 3 hosts.
scsi : detected total.
sunhme.c:v1.2 10/Oct/96 David S. Miller (davem@caipfs.rutgers.edu)
eth0: HAPPY MEAL (SBUS) 10/100baseT Ethernet 08:00:20:9a:71:d5 
eth1: HAPPY MEAL (SBUS) 10/100baseT Ethernet 08:00:20:9a:71:d5 
eth2: HAPPY MEAL (SBUS) 10/100baseT Ethernet 08:00:20:9a:71:d5 
Partition check:
VFS: Cannot open root device 00:00
Kernel panic: VFS: Unable to mount root fs on 00:00
Press L1-A to return to the boot prom

Re: [Q] ultrapenguin-1.1 on UE4000

[David S. Miller]

   Date: Thu, 19 Nov 1998 12:27:36 -0800 (PST)
   From: ebrower@usa.net

   > Let me know how it goes.

   Here's the output from the kernel:

 ...

   Total of 8 processors activated (3958.37 BogoMIPS).

Fuckin' A!  ~4000 bogomip machine :-)

Looks like the scsi layer or driver is doing something very stupid on
this box.  I'll try to enable some clever debugging in a test kernel
and put up another image tomorrow.  I hope you'll still be able to
test things on this machine.

Thanks for all the help so far.

Later,
David S. Miller
davem@dm.cobaltmicro.com