Alphaserver 1000A 5/400 does not detect serial ports

[Johan Zeylon]

I am running Digital UNIX 3.2 on four AlphaServer 1000A. Three are model 5/500 and one is a 5/400. I can't get the 5/400 to detect the serial ports.
The software configurations of the four machines are identical but there are some firmware differences:
(I am comparing two machines since the other two are running mission critical software and can not be tampered with.)

When i examine the "messages" file i see that the 5/400 does not initialize the ace0 and ace1 devices.

I have two theories:
1: The old firmware needs updating
2: The 5/400 needa a different configuration file

Any good ideas?

The 5/500 (detects serial ports) has the following firmware revisions:
SRM: V5.5-75, OSF PALcode: V1.23-6

The 5/400 (does not detect serial ports) has the following firmware revisions:
SRM: 4.9-164, OSF PALcode: V1.21-6

The 5/500 messages file looks like this:
------------------------------------------
Jan 21 08:40:56 o1pda vmunix: Alpha boot: available memory from 0x9ec000 to 0xffce000
Jan 21 08:40:56 o1pda vmunix: Digital UNIX V3.2G (Rev. 62); Thu Oct 1 13:39:06 MET DST 2009
Jan 21 08:40:57 o1pda vmunix: physical memory = 256.00 megabytes.
Jan 21 08:40:57 o1pda vmunix: available memory = 246.05 megabytes.
Jan 21 08:40:57 o1pda vmunix: using 975 buffers containing 7.61 megabytes of memory
Jan 21 08:40:57 o1pda vmunix: Firmware revision: 5.5
Jan 21 08:40:57 o1pda vmunix: PALcode: OSF version 1.23
Jan 21 08:40:57 o1pda vmunix: AlphaServer 1000A 5/500
Jan 21 08:40:57 o1pda vmunix: Mem_size = 266346496, consmem = 255
Jan 21 08:40:57 o1pda vmunix: pci0 at nexus
Jan 21 08:40:57 o1pda vmunix: eisa0 at pci0
Jan 21 08:40:57 o1pda vmunix: ace0 at eisa0
Jan 21 08:40:57 o1pda vmunix: ace1 at eisa0
Jan 21 08:40:57 o1pda vmunix: lp0 at eisa0
Jan 21 08:40:57 o1pda vmunix: fdi0 at eisa0
Jan 21 08:40:57 o1pda vmunix: fd0 at fdi0 unit 0
Jan 21 08:40:57 o1pda vmunix: pci2000 at pci0 slot 8
...


The 5/400 messages file looks like this:
------------------------------------------
Jan 21 10:07:22 o1pdb vmunix: Alpha boot: available memory from 0xee6000 to 0x1ffee000
Jan 21 10:07:22 o1pdb vmunix: Digital UNIX V3.2G (Rev. 62); Mon Feb 4 15:24:14 MET 2008
Jan 21 10:07:22 o1pdb vmunix: physical memory = 512.00 megabytes.
Jan 21 10:07:22 o1pdb vmunix: available memory = 497.22 megabytes.
Jan 21 10:07:22 o1pdb vmunix: using 1959 buffers containing 15.30 megabytes of memory
Jan 21 10:07:22 o1pdb vmunix: Firmware revision: 4.9
Jan 21 10:07:22 o1pdb vmunix: PALcode: OSF version 1.21
Jan 21 10:07:22 o1pdb vmunix: AlphaServer 1000A 5/400
Jan 21 10:07:22 o1pdb vmunix: Mem_size = 534945792, consmem = 235
Jan 21 10:07:22 o1pdb vmunix: pci0 at nexus
Jan 21 10:07:22 o1pdb vmunix: eisa0 at pci0
Jan 21 10:07:22 o1pdb vmunix: pci2000 at pci0 slot 8
...


The kernel configuration file looks like this:
----------------------------------------------
ident "O1PDA"

options UERF
options OSF
options _LMF_
options BIN_COMPAT
options COMPAT_43
options MACH
options MACH_IPC_TCACHE
options MACH_IPC_WWA
options MACH_IPC_XXXHACK
options BUFCACHE_STATS
options INOCACHE_STATS
options STAT_TIME
options VAGUE_STATS
options UFS
options NFS
options NFS_SERVER
options STRKINFO
options STREAMS
options LDTTY
options RPTY
options INET
options UIPC
options SYSV_COFF
options QUOTA
options LABELS
options SL
options SNMPINFO
options DLI
options BSD_TTY
options LAT
options DLB
options PROCFS
#
# Standard options.
#
options UNIX_LOCKS
options SER_COMPAT
options RT_PREEMPT
options RT_SCHED
options RT_SCHED_RQ
options RT_PML
options RT_TIMER
options RT_SEM
options RT_CSEM
options RT_IPC
#
makeoptions CDEBUGOPTS="-g3"
makeoptions CCOMPRESS="-compress"
makeoptions PROFOPTS="-DPROFILING -DPROFTYPE=4"
#
# Max number of processors in the system (DO NOT CHANGE)
#
processors 16

#
# Special options (see configuring the kernel chapter
# in the Guide to System Administration)
#
timezone 0 dst 0
dfldsiz 134217728
maxdsiz 1073741824
dflssiz 2097152
maxssiz 33554432
cpu "DEC1000A_5"
maxusers 32
machine alpha

config vmunix swap generic

bus pci0 at nexus?
callout after_c "../bin/mkdata pci"
bus eisa0 at pci0 slot 7
callout after_c "../bin/mkdata eisa"
controller ace0 at eisa0 slot 9 vector aceintr
controller ace1 at eisa0 slot 10 vector aceintr
controller lp0 at eisa0 slot 11 vector lpintr
controller fdi0 at eisa0 slot 12 vector fdintr
device disk fd0 at fdi0 drive 0
controller gpc0 at eisa0 slot 0 vector gpcintr
bus pci2000 at pci0 slot 8
bus isp0 at pci2000 slot 0 vector ispintr
controller scsi0 at isp0 slot 0
device disk rz0 at scsi0 drive 0
device disk rz1 at scsi0 drive 8
device disk rz2 at scsi0 drive 16
device disk rz3 at scsi0 drive 24
device disk rz4 at scsi0 drive 32
device tape tz5 at scsi0 drive 40
controller tu0 at pci2000 slot 1 vector tuintr
controller tu1 at pci2000 slot 2 vector tuintr
controller tu2 at pci2000 slot 3 vector tuintr
controller tu3 at pci2000 slot 4 vector tuintr
controller tga0 at pci0 slot 12 vector tgaintr


scs_sysid 1
pseudo-device sysv_hab
pseudo-device svid_three_hab
pseudo-device svr_four_hab
pseudo-device soe_two_hab
pseudo-device rt_hab
pseudo-device ether
pseudo-device loop
#
# Max number of processors in the system (DO NOT CHANGE)
#
pseudo-device cpus 16
pseudo-device ws

[Martin Moore · Accepted solution]

Wow, that's really ancient! It's been a long time, but I think this may be caused by the EISA configuration data in NVRAM becoming corrupted. If so, what you need to do is re-run the ECU (EISA configuration utility). If you're also upgrading the firmware, do that before re-running the ECU.

Martin

[Rob Leadbeater]

> Digital UNIX 3.2
> AlphaServer 1000A
> mission critical software

Some things never cease to amaze me !

If the firmware upgrade doesn't solve things, I'll guess you might want to try booting from the generic kernel and rebuild the kernel... That said, 3.2 is way before my time...

Cheers,

Rob

[Johan Zeylon]

I had trouble locating a ECU disk. We bought the machine used and no documentation or software was included. The main machines were bought pre-configured many years ago and the ECU disks are missing.
I finally found a copy of the ECU disk on the web (Ehem...) and ran it.
The EISA conifuration was obviously faulty since this solved my problem. I think that the problem might have been that this machine had some EISA boards in it in its previous life and they were simply removed before we bought it.

[Johan Zeylon]

See my comment above