successful boot report
der Mouse
"SmartAlpha Server A10". The console reports it as a "Digital AlphaPC
164LX".
I spent a little while poking around the port-alpha supported hardware
pages trying to figure out whether this was supported; I never did find
anything I recognized as representing it. But I tried it anyway, and
it runs fine. NetBSD says it's a "Digital AlphaPC 164LX 533 MHz, s/n",
so clearly _something_ recognizes it (though the odd trailing "s/n"
makes me suspect there's a serial-number field somewhere which isn't
working). Comparing this against the list of CPU support options in
the kernel config makes me think it's one of the EB164 systems.
Perhaps it could help reassure future users if the supported-hardware
list listed it in a more recognizable way?
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mo...@rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-...@muc.de
Dustin Marquess
> it runs fine. NetBSD says it's a "Digital AlphaPC 164LX 533 MHz, s/n",
> so clearly _something_ recognizes it (though the odd trailing "s/n"
> makes me suspect there's a serial-number field somewhere which isn't
> working). Comparing this against the list of CPU support options in
Off the top of my head, I believe the s/n is just a SRM variable that
can be set to anything.
-Dustin
der Mouse
Thank you! But, really, I can't claim much credit; the only difficult
part was getting NetBSD on the machine, and even that wasn't very hard.
The CD drive didn't work in the sense that I couldn't figure out how to
boot from it; I swapped in another drive I had handy and had the same
problem. I put a second disk on my PWS 600au and installed onto it,
then moved it to the A10, and it all Just Worked.
> There was two versions of the 164, the LX and SX if memory serves.
> Only one of them was 'complete' enough to run VMS/Tru64, the other
> was WinNT and Linux only.
Well, this one calls itself an LX, and in the console, the os_type
parameter claims to support VMS and Unix (and one other which I don't
remember in detail but I do remember thinking, wait, how does that
differ from the "Unix" choice?). And it does indeed run NetBSD. So I
suspect the LX is the `complete' one.
..I'd rather be coding ASM!
Well done!
There was two versions of the 164, the LX and SX if memory serves. Only
one of them was 'complete' enough to run VMS/Tru64, the other was WinNT
and Linux only.
On Sun, 3 Jul 2011, der Mouse wrote:
> I was recently given an Alpha which, according to the box, is a
> "SmartAlpha Server A10". The console reports it as a "Digital AlphaPC
> 164LX".
>
> I spent a little while poking around the port-alpha supported hardware
> pages trying to figure out whether this was supported; I never did find
> anything I recognized as representing it. But I tried it anyway, and
> it runs fine. NetBSD says it's a "Digital AlphaPC 164LX 533 MHz, s/n",
> so clearly _something_ recognizes it (though the odd trailing "s/n"
> makes me suspect there's a serial-number field somewhere which isn't
> working). Comparing this against the list of CPU support options in
> the kernel config makes me think it's one of the EB164 systems.
>
> Perhaps it could help reassure future users if the supported-hardware
> list listed it in a more recognizable way?
>
> /~\ The ASCII Mouse
> \ / Ribbon Campaign
> X Against HTML mo...@rodents-montreal.org
> / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
>
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
der Mouse
>> dqa.0.0.11.0
> This should be the DQA0 CDROM device.
Looks to me as though this is just the controller.
Actually, when I did that "show config" there was no CD drive at all
connected. But I just now tried it with a drive connected and nothing
more showed up, neither in SRM nor in NetBSD. And it's not just a duff
drive, either; I put it on a peecee and both the BIOS and NetBSD see it
just fine. As an additional test, I put a real disk drive on there,
with the same cable I used for the CD drive, and it showed up as DQA0,
so it's not just that I happen to have a fried cmdide0. Perhaps it's
just SRM taking its marbles and going home because it's not a real
honest-to-DEC[%] CD drive.
I checked the kernel config, and all the entries are there to see
cmdide0 -> atabus0 -> atapibus0 -> cd0. I conjecture (with no real
grounds, save that it's the first explanation that comes to mind for
NetBSD's failure to see it) that the cmdide is not ATAPI-compatible.
[%] A phrase I saw in off-list email prompted by this thread and liked
enough to steal.
..I'd rather be coding ASM!
Hi Michael,
> DVA0 would normally be the floppy disk. Any optical media would normally
> be DQ[AB]0. Network devices I think would be E*.
What's DSSI come up as on Alpha? I've only ever used it on VAX4000/m300
with a HSD+BA350 and HSV+BA356.
Al.
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
--
Julian Coleman
> DVA0 would normally be the floppy disk. Any optical media would
> normally be DQ[AB]0. Network devices I think would be E*.
Yep - my PC164LX has:
>>> show dev
dka0.0.0.7.0 DKA0 FUJITSU MAP3367NP 0108
dka100.0.0.7.0 DKA100 FUJITSU MAP3367NP 0108
dqa0.0.0.11.0 DKA0 RICOH DV D/CDRW MP9060 1.80
dva0.0.0.0.0 DVA0
ewa0.0.0.9.0 EWA0 00-40-05-A5-F5-CA
pka0.7.0.7.0 PKA0 SCSI Bus ID 7
The SCSI card is a Symbios Logic 53c875, the network card a DEC 21140A,
the CD/DVD is ATAPI, and the disks are at ID 0 and 1:
esiop0 at pci0 dev 7 function 0: Symbios Logic 53c875 (ultra-wide scsi)
tlp0 at pci0 dev 9 function 0: DECchip 21140A Ethernet, pass 2.2
cd0 at atapibus0 drive 0: <RICOH DVD/CDRW MP9060, , 1.80> cdrom removable
sd0 at scsibus0 target 0 lun 0: <FUJITSU, MAP3367NP, 0108> disk fixed
sd1 at scsibus0 target 1 lun 0: <FUJITSU, MAP3367NP, 0108> disk fixed
I have had an Adaptec 39160 SCSI card in (note that a 29160 does not work),
but that lost against a lightning strike. The IDE controller is the same
(horrible) chipset that also appears in some Sun machines (U5, U10, etc.):
cmdide0: CMD Technology PCI0646 (rev. 0x01)
Thanks,
J
--
My other computer also runs NetBSD / Sailing at Newbiggin
http://www.netbsd.org/ / http://www.newbigginsailingclub.org/
..I'd rather be coding ASM!
> No; both of the CD drives (actually, I think one might have been a DVD
> drive) were IDE (or ATAPI, or whatever it's properly called). Asking
> the console to "show dev" listed only two devices (DVA0 and PKA0, I
> think they were), but only the same two devices even when I had no
> CD/DVD drive at all connected.
Right. ATAPI I should bother to pay more attention to, my ES40 and the
DS20 at home both have it. Usually DKxx is scsi disks. EWxx is etherwebs,
and PKxx I "think" (Please someone correct me) is the ATA controller
pretending to be scsi? CD's for me are DQxx.
DVxx for me is the floppy drive.
Here's fun:
http://h21007.www2.hp.com/portal/download/files/unprot/linux/redhat/ig/s1-alpha-srm.html
>> Another option for future reference is to set up a MOP (maint
>> operation) boot for VMS or *NIX.
>
> Possibly. Good point; I'd forgotten about netbooting.
It's handy. All VAX/Alpha/SPIM/InfoServer/DECServer's all talk MOP.
> DVA0 might be the network; I tried booting from it, and, whatever it
> was, the console said it couldn't open it. But of course that could
> just mean no MOP server responded; it did take long enough that it's
> reasonable it was a network no-response timeout. I think I have some
> kind of MOP setup at least partially set up, because at one point I had
> a MicroVAX-II running diskless, and I think the only network boot its
> ROMs are capable of is MOP. I do see a /tftpboot/mop containing a
> bunch of .SYS files with MAC-address names. I don't see any mopd
> running now, though, so I probably didn't leave it set to start at boot
> on the server.
Nifty. Other than the correct image with appropriate MAC in the fname,
you're merstly good to roll with mopd running :)
Nice to see a uVAX-II still being loved and used in anger.
Al.
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
--
Dave McGuire
>> What's DSSI come up as on Alpha? I've only ever used it on
>> VAX4000/m300 with a HSD+BA350 and HSV+BA356.
>
> shows DIA.
I've had several AS2100 and AS4100 systems with DSSI. I think it was
only available on the higher-end systems, not the desktops.
-Dave
--
Dave McGuire
Port Charlotte, FL
..I'd rather be coding ASM!
> The SCSI card is a Symbios Logic 53c875, the network card a DEC 21140A,
> the CD/DVD is ATAPI, and the disks are at ID 0 and 1:
It's 'spelt' "DEC 21140A" but is pronounced "Tulip" :>
> I have had an Adaptec 39160 SCSI card in (note that a 29160 does not work),
> but that lost against a lightning strike. The IDE controller is the same
> (horrible) chipset that also appears in some Sun machines (U5, U10, etc.):
>
> cmdide0: CMD Technology PCI0646 (rev. 0x01)
Also B&W smurf-tower Mac's. They have a DMA bug. There's some sofware
work-arounds that prevent the chain locking up.
You might be able to use the u160 cotnroller inside nbsd. I had a ds20
driving a yum-cha brand SATA controller after a bit of kernel fiddling.
You can't boot of it, but heck.. load the kernel off some old scsi disk or
MOP and go gang-busters from there. Also with with u160's eg the i2o dev I
have, I had to first set it up in a PeeCee and set all the disks to jbod
and save the config to the card because I couldn't drive the bios on the
alpha and the rotten thing wouldn't come up in 'dumb mode' and present the
drives. First had to go make separate vdisks for each vol.
Al.
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
--
Michael L. Hitch
> PCI Bus
> Bus 00 Slot 05: 905010B7
This would be the 3COM network card. SRM is only going to know about
"real" DEC Tulip boards for network booting.
> Bus 00 Slot 11: CMD PCI0646 IDE
> dqa.0.0.11.0
This should be the DQA0 CDROM device. I suspect it may have to be a
CDROM device that could present 512 byte sectors. I don't remember if I
ever tried booting with the CDROM (now a DVD drive) on my PC164, so I
don't know if the SRM is smart enough to deal with the 2048 byte blocks.
> So I'd say the Ethernet _does_ show up in "show config", just not in a
> very useful form (presumably because SRM doesn't know the 3c905).
Correct.
> cmdide0: primary channel interrupting at isa irq 14
> cmdide0: secondary channel interrupting at isa irq 15
>
> which seem a little odd (a PCI device interrupting at an ISA IRQ?), but
> presumably when everything's in one big ASIC you get things like that.
There's some mention about this in the source (but not much more), and
even the ES40 does that. I do have problems attempting to use the DVD
drive on the ES40 under NetBSD (gets ata timeouts), but the DVD drive in
my PC164 works quite well.
The SRM shows an optical device, and it should show up on the 'show
device' command. It also should show as cd0: during autoconfigure.
Mike
--
Michael L. Hitch mhi...@montana.edu
Computer Consultant
Information Technology Center
Montana State University Bozeman, MT USA
der Mouse
> Does a 'show config' command show any network device?
Nothing I recognize as one at first sight, but see below. This is a
ten-finger copy; whitespace in particular might be a little off.
PCI Bus
Bus 00 Slot 05: 905010B7
Bus 00 Slot 07: 051B102B/0300102B
Bus 00 Slot 08: Intel 82378
Bridge to Bus 1, ISA
Bus 00 Slot 09: Adaptec AIC-7880
pka0.7.0.9.0 SCSI Bus ID 7
dka100.1.0.9.0 IBM-PSG ST39103LC
Bus 00 Slot 11: CMD PCI0646 IDE
dqa.0.0.11.0
ISA
Slot Device Name Type Enabled BaseAddr IRQ DMA
0
0 MOUSE Embedded Yes 60 12
0 KBD Embedded Yes 60 1
0 COM1 Embedded Yes 3f8 4
0 COM2 Embedded Yes 2f8 3
0 LPT1 Embedded Yes 3bc 7
0 FLOPPY Embedded Yes 3f0 6 2
> Or anything that looks like and IDE controller?
Well, yes. :)
> If the 164LX is anything like my PC164, it might need an ISACFG command
> if either the IDE or network device is an ISA device. I recall playing
> with that command a bit on my PC164, but I don't remember what I was
> trying to do at the time.
Excerpting from the NetBSD boot messages, the PCI enumeration above
matches up well:
ex0 at pci0 dev 5 function 0: 3Com 3c905-TX 10/100 Ethernet (rev. 0x0)
...
vga0 at pci0 dev 7 function 0: Matrox MGA Millennium II 2164W (rev. 0x00)
...
sio0 at pci0 dev 8 function 0: Intel 82378ZB System I/O (rev. 0x43)
ahc0 at pci0 dev 9 function 0: Adaptec 2940 Ultra SCSI adapter
...
cmdide0 at pci0 dev 11 function 0
cmdide0: CMD Technology PCI0646 (rev. 0x01)
...
isa0 at sio0
So I'd say the Ethernet _does_ show up in "show config", just not in a
very useful form (presumably because SRM doesn't know the 3c905).
The only things that attach "at isa" are
lpt0 at isa0 port 0x3bc-0x3bf irq 7
com0 at isa0 port 0x3f8-0x3ff irq 4: ...
com1 at isa0 port 0x2f8-0x2ff irq 3: ...
pckbc0 at isa0 port 0x60-0x64
attimer0 at isa0 port 0x40-0x43: AT Timer
pcppi0 at isa0 port 0x61
fxc0 at isa0 port 03f0-0x3f7 irq 6 drq 2
mcclock0 at isa0 port 0x70-0x71: ...
There _are_ two other lines
cmdide0: primary channel interrupting at isa irq 14
cmdide0: secondary channel interrupting at isa irq 15
which seem a little odd (a PCI device interrupting at an ISA IRQ?), but
presumably when everything's in one big ASIC you get things like that.
Someone asked me offlist what SRM rev the machine has. I'm answering
here in case anyone else cares. On reset, it calls itself
Digital AlphaPC 164LX 533 MHz Console V5.8-1, Jun 21 2000 11:28:32
"show version" prints "V5.8-1 Jun 21 2000 11:28:32" - just like the
last 40% or so of that message, except for the comma. And "show"
includes a variable "version" which has that content. I hope that's
revision information enough.
[uridium]
>> [...] at one point I had a MicroVAX-II running diskless, [...]
> Nice to see a uVAX-II still being loved and used in anger.
Well, it _has_ been a year or two since I turned it on. I keep wanting
to, though....
Michael L. Hitch
> What's DSSI come up as on Alpha? I've only ever used it on VAX4000/m300 with
> a HSD+BA350 and HSV+BA356.
I don't think I've ever seen an Alpha with DSSI, but a google search
shows DIA.
--
Michael L. Hitch mhi...@montana.edu
Computer Consultant
Information Technology Center
Montana State University Bozeman, MT USA
--
der Mouse
> If the CDROM is scsi,
No; both of the CD drives (actually, I think one might have been a DVD
drive) were IDE (or ATAPI, or whatever it's properly called). Asking
the console to "show dev" listed only two devices (DVA0 and PKA0, I
think they were), but only the same two devices even when I had no
CD/DVD drive at all connected.
> Another option for future reference is to set up a MOP (maint
> operation) boot for VMS or *NIX.
Possibly. Good point; I'd forgotten about netbooting.
DVA0 might be the network; I tried booting from it, and, whatever it
was, the console said it couldn't open it. But of course that could
just mean no MOP server responded; it did take long enough that it's
reasonable it was a network no-response timeout. I think I have some
kind of MOP setup at least partially set up, because at one point I had
a MicroVAX-II running diskless, and I think the only network boot its
ROMs are capable of is MOP. I do see a /tftpboot/mop containing a
bunch of .SYS files with MAC-address names. I don't see any mopd
running now, though, so I probably didn't leave it set to start at boot
on the server.
/~\ The ASCII Mouse
..I'd rather be coding ASM!
> I've had several AS2100 and AS4100 systems with DSSI. I think it was only
> available on the higher-end systems, not the desktops.
There was definitely turbo-channel DSSI cards about and I "think" pci
cards with it on.
Just curious.. I have one stashed away I've never used. I should try and
find another pelican.
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
--
..I'd rather be coding ASM!
G'day,
> Thank you! But, really, I can't claim much credit; the only difficult
> part was getting NetBSD on the machine, and even that wasn't very hard.
> The CD drive didn't work in the sense that I couldn't figure out how to
> boot from it; I swapped in another drive I had handy and had the same
> problem. I put a second disk on my PWS 600au and installed onto it,
> then moved it to the A10, and it all Just Worked.
If the CDROM is scsi, check if it's got the 512byte block jumper
set. Sometimes this is called "sector size". Might've been the problem.
Another option for future reference is to set up a MOP (maint operation)
boot for VMS or *NIX. Both NetBSD and TheoLinux come with mopd
pre-installed. Very useful for VAX/Alpha's.
> Well, this one calls itself an LX, and in the console, the os_type
> parameter claims to support VMS and Unix (and one other which I don't
> remember in detail but I do remember thinking, wait, how does that
> differ from the "Unix" choice?). And it does indeed run NetBSD. So I
> suspect the LX is the `complete' one.
Nifty :)
Have fun!
Regards,
Al
--
--
Al Boyanich
adb -w -P "world> " -k /dev/meta/galaxy/ksyms /dev/god/brain
--
der Mouse
> normally be DQ[AB]0. Network devices I think would be E*.
Then I guess the network that's in there isn't SRM-compatible.
> Does a 'show config' command show any network device? Or anything
> that looks like and IDE controller? I don't seem to have a show
> config from my PC164 anywhere handy to see what it shows.
I don't have that handy; I'll be turning that machine on this evening
and collecting various information.
> If the 164LX is anything like my PC164, it might need an ISACFG
> command if either the IDE or network device is an ISA device.
I _think_ NetBSD sees them as attached via PCI, but that's pretty fuzzy
memory; I'll check that too this evening.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mo...@rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
--
Anders Magnusson
>> There was two versions of the 164, the LX and SX if memory serves.
>> Only one of them was 'complete' enough to run VMS/Tru64, the other
>> was WinNT and Linux only.
> Well, this one calls itself an LX, and in the console, the os_type
> parameter claims to support VMS and Unix (and one other which I don't
> remember in detail but I do remember thinking, wait, how does that
> differ from the "Unix" choice?). And it does indeed run NetBSD. So I
> suspect the LX is the `complete' one.
Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
2005, 2006
The NetBSD Foundation, Inc. All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
The Regents of the University of California. All rights reserved.
NetBSD 4.0_BETA2 (SPEEDY) #3: Sat Jan 30 09:31:42 MET 2027
l...@speedy.ludd.ltu.se:/usr/src/sys/arch/alpha/compile/SPEEDY
Digital AlphaPC 164LX 533 MHz, s/n
8192 byte page size, 1 processor.
total memory = 512 MB
(1896 KB reserved for PROM, 510 MB used by NetBSD)
avail memory = 497 MB
mainbus0 (root)
...
-- Ragge
Michael L. Hitch
> DVA0 might be the network; I tried booting from it, and, whatever it
> was, the console said it couldn't open it. But of course that could
DVA0 would normally be the floppy disk. Any optical media would
normally be DQ[AB]0. Network devices I think would be E*.
Does a 'show config' command show any network device? Or anything that
looks like and IDE controller? I don't seem to have a show config from my
PC164 anywhere handy to see what it shows.
If the 164LX is anything like my PC164, it might need an ISACFG command
if either the IDE or network device is an ISA device. I recall playing
with that command a bit on my PC164, but I don't remember what I was
trying to do at the time.
Mike
--
Michael L. Hitch mhi...@montana.edu
Computer Consultant
Information Technology Center
Montana State University Bozeman, MT USA
--
KJ Seefried
>
>> I've had several AS2100 and AS4100 systems with DSSI. I think it
>> was only available on the higher-end systems, not the desktops.
>
> cards with it on.
>
> Just curious.. I have one stashed away I've never used. I should try
> and find another pelican.
>
called a KFPSA. They don't seem to be particularly rare.
KJ
Staffan Thomén
>
> There was two versions of the 164, the LX and SX if memory serves.
> Only one of them was 'complete' enough to run VMS/Tru64, the other
> was WinNT and Linux only.
Heh, sleepy, Al? Surely you remember my machine. There are in fact three
EB164 versions, the SX, one without letters, and LX. As far as I've been able
to find out, the SX only supports lower clock cpus, while the unnamed goes up
to 500MHz and the LX can handle higher ones.. I have a fuzzy memory about there
being differences in cache size as well.
Staffan
--
Staffan Thom�n - ADB3 455F 10D5 86D1 78D6 048D 11BB D66E 7C7E 2EF8
..I'd rather be coding ASM!
Hi Duck! Yes.. rather. I'm down in Melbourne. Why do I always think you
have the SX? :)
Al.
On Thu, 7 Jul 2011, Staffan Thomén wrote:
> On Mon, Jul 04, 2011 at 08:16:41AM +0300, ..I'd rather be coding ASM! wrote:
>>
>> There was two versions of the 164, the LX and SX if memory serves.
>> Only one of them was 'complete' enough to run VMS/Tru64, the other
>> was WinNT and Linux only.
>
> Heh, sleepy, Al? Surely you remember my machine. There are in fact three
> EB164 versions, the SX, one without letters, and LX. As far as I've been able
> to find out, the SX only supports lower clock cpus, while the unnamed goes up
> to 500MHz and the LX can handle higher ones.. I have a fuzzy memory about there
> being differences in cache size as well.
>
> Staffan
>
>
--
Michael Kronsteiner
> On Mon, Jul 04, 2011 at 08:16:41AM +0300, ..I'd rather be coding ASM! wrote:
> >
> > There was two versions of the 164, the LX and SX if memory serves.
> > Only one of them was 'complete' enough to run VMS/Tru64, the other
> > was WinNT and Linux only.
>
> Heh, sleepy, Al? Surely you remember my machine. There are in fact three
> EB164 versions, the SX, one without letters, and LX. As far as I've been able
> to find out, the SX only supports lower clock cpus, while the unnamed goes up
> to 500MHz and the LX can handle higher ones.. I have a fuzzy memory about there
> being differences in cache size as well.
The sx boards i own all have a 533mhz cpu installed
Michael Kronsteiner
>
> Hmh, how unexpected. My information is only from what I've gleaned off the
> intarwebs while looking for docs/firmware for my alpha (since it has no suffix
> I get a lot of hits for the others too...), so I stand corrected.
>
alphalinux.org (?) says there are 2 cpus on the sx boards... 400 and
533. seems i am lucky :)
http://www.alphalinux.org/faq/FAQ-5.html
(quote)
The low end of the EV5 generation is the AlphaPC 164SX. It uses the
21164PC variant of the Alpha microprocessor, at 400 or 533 MHz. This
board is targeted at the NT Workstation market, but it will run Linux
just fine.
(/quote)