Eyal Lebedinsky | 4 Apr 10:25 2004
Picon

make strangeness

I ended up with a makefile that contains something like this
(call it test.mak):

xxx xxx:
	echo $(MAKE_VERSION)
         echo -n " " ld -o xxx   > make.crf
         echo -n " " a.o         >>make.crf
         echo -n " " b.o         >>make.crf
         -rm xxx
         touch xxx

Naturally, in my makfile the two 'xxx's are actually different
variables that ended up with the same value, and the final
touch was actually executing make.crf. When making I get:

$ rm xxx ; make -f test.mak
echo netbsd-20020909
netbsd-20020909
echo netbsd-20020909
netbsd-20020909
echo -n " " ld -o xxx   > make.crf
echo -n " " ld -o xxx   > make.crf
echo -n " " a.o         >>make.crf
echo -n " " a.o         >>make.crf
echo -n " " b.o         >>make.crf
echo -n " " b.o         >>make.crf
rm xxx
rm: xxx: No such file or directory
*** Error code 1 (ignored)
rm xxx
(Continue reading)

Todd Vierling | 9 Apr 04:31 2004

Free Personal DECstation 5000/20 (Atlanta, GA, USA)

It's time to clean up my attic.  I'm getting somewhat depressed seeing so
much NetBSD-capable hardware lying dormant.  So:

I have a barebones Personal DECstation 5000/20 (actually a 5000/25 named on
the cover but with a 5000/20 CPU module) available for pick-up in the
Atlanta, GA, USA area, or just the cost of shipping it.  Specs details:

* V2.0m BIOS
* 20MHz (welcome to the early MAXINE days :)
* 40MB memory installed:  8MB onboard and two 16MB DEC SIMMs installed
* No HD included (uses any SCSI-1)
* No keyboard included, but KB and video ports have been verified to work
* Missing HD sled (I had a drive loose with standoffs to prop it up), but
  has connection cables

===== dmesg =====

Personal DECstation 5000/20 (MAXINE)
total memory = 40960 KB
avail memory = 34992 KB
mainbus0 (root)
cpu0 at mainbus0: MIPS R3000A CPU (0x230) Rev. 3.0 with MIPS R3010 FPC Rev. 3.0
cpu0: 64KB/4B direct-mapped Instruction cache, 64 TLB entries
cpu0: 64KB/4B direct-mapped write-through Data cache
tc0 at mainbus0: 12.5 MHz clock
ioasic0 at tc0 slot 3 offset 0x0
le0 at ioasic0 offset 0xc0000: address 08:00:2b:2a:fc:3b
le0: 32 receive buffers, 8 transmit buffers
scc0 at ioasic0 offset 0x100000: console
mcclock0 at ioasic0 offset 0x200000: mc146818 or compatible
(Continue reading)

Eyal Lebedinsky | 11 Apr 05:04 2004
Picon

afio problem

I am running a 1.6 kernel (a full 1.6 install actually). I needed afio so
I grabbed the 1.6.2 pkgsrc (first time afio was included) and built it.
No problems here, but when testing it I get errors:

$ find -x . | afio -ov /tmp/test1.afio
. -- okay
CVS -- okay
CVS/Root -- okay
CVS/Repository -- okay
[rest of list trimmed - all okay]

$ ls -l /tmp/test1.afio
-rw-rw-r--  1 eyal  wheel  558080 Apr 11 12:51 /tmp/test1.afio

$ file /tmp/test1.afio
/tmp/test1.afio: ASCII cpio archive (pre-SVR4 or odc)

$ afio -tv /tmp/test1.afio
drwxr-xr-x  1 root     wsrc              Apr 11 12:36:18 2004 .
afio: "/tmp/test1.afio" [offset 540k+0]: Fatal error:
afio: "/tmp/test1.afio": Premature input EOF

I verified that the archive is bad by testing it on another system
(same errors).

I built the same afio package on my pc532 and it works just fine there.

Anyone got afio working here? Maybe someone with 1.6.2 installed from
the distro?

(Continue reading)

Mauricio | 15 Apr 04:58 2004

My dec 5000/25 went boink on me today

	My Dec 5000/25 run netbsd 1.6.1 and is my dhcp/dns server. 
It has been a quite happy machine for a while.  In fact it has been 
running without a hitch since I had done some upgrading ("Peter C. 
Wallace" <pcw <at> mesanet.com>, do email me back with an address so I can 
pay for thge memory and the CPU.  Pretty please?) in it.  But, today, 
I went to do some stuff in it and could not ssh to it.  So, I decided 
to use the console.  This is what I got:

db> help
break           dmesg           p               show            watch
c               dwatch          print           sifting         write
call            examine         ps              step            x
callout         kill            reboot          sync
continue        machine         s               trace
d               match           search          until
delete          next            set             w
db> show
all             buf             map             page            uvmexp
arptab          event           ncache          pool            vnode
breaks          malloc          object          registers       watches
db> show all
callout procs
db> quit
No such command
db> reboot
syncing disks... panic: lockmgr: locking against myself
Begin traceback...
pid -2144880964 not found
End traceback...

(Continue reading)

Toru Nishimura | 15 Apr 09:29 2004
Picon

Re: My dec 5000/25 went boink on me today

Mauricio supremedalek <at> hotpop.com asked;

> today, I went to do some stuff in it and could not ssh to it.  So, I decided 
> to use the console.  This is what I got:
> ...
> sd0(asc0:0:0:0): polling command not done
> panic: scsipi_execute_xs
> Begin traceback...
> pid -2144800340 not found
> End traceback...
>
> dumping to dev 19,1 offset 50927
> dump device not ready
>
> What is going on here?  What was it trying to tell me?

The stack contents has been smashed and system got total confusion.

> HD is going  boink?

SCSIPI subsystem needs to have stack memory to run.  Upon the incident
stack has been damaged too much to complete panic dump.

> I eventually turned it off for a bit and then rebooted it. 
> Right now it seems to be happy once more, but I do not know for how 
> long.  Suggestions?

The output from "show registers" and  "machine tlb" might be helpful to see
the last moment the life.

(Continue reading)

Toru Nishimura | 15 Apr 10:09 2004
Picon

Re: My dec 5000/25 went boink on me today

Here is a technical tip.

Due to the distinctive feature of software TLB update scheme kernel stack
needs to be "protected" from TLB miss, that is, kernel stack should never
make TLB miss to use.  Software TLB update processing is likely to
produce a string of TLB miss traversing the kernel date structure to
complete.  Then to make kernel stack "TLB miss free" NetBSD kernel
hardwires the 8KB memory  with 2 pcs of TLB entry on every context switch.
If the hardwiring goes wrong (out of sync), kernel stack gets effectively
destroyed at a moment.

Toru Nishimura/ALKYL Technology

Wai-Sun Chia | 15 Apr 17:03 2004
Picon

Re: My dec 5000/25 went boink on me today

Yeah..looks like HD problem to me also...
I have a uVAXII which had a similar problem, although MFM disks are 
arguably more robust to corruptions than SCSI disks as a reformat would 
take those bad sectors out...

Mauricio wrote:
>     My Dec 5000/25 run netbsd 1.6.1 and is my dhcp/dns server. It has 
> been a quite happy machine for a while.  In fact it has been running 
> without a hitch since I had done some upgrading ("Peter C. Wallace" 
> <pcw <at> mesanet.com>, do email me back with an address so I can pay for 
> thge memory and the CPU.  Pretty please?) in it.  But, today, I went to 
> do some stuff in it and could not ssh to it.  So, I decided to use the 
> console.  This is what I got:
> 
> db> help
> break           dmesg           p               show            watch
> c               dwatch          print           sifting         write
> call            examine         ps              step            x
> callout         kill            reboot          sync
> continue        machine         s               trace
> d               match           search          until
> delete          next            set             w
> db> show
> all             buf             map             page            uvmexp
> arptab          event           ncache          pool            vnode
> breaks          malloc          object          registers       watches
> db> show all
> callout procs
> db> quit
> No such command
(Continue reading)

Anders Gavare | 18 Apr 17:16 2004
Picon
Picon

Re: pmax DEBUG kernel

On Thu, 2 Oct 2003, Anders Gavare wrote:

> Hi,
..
> (I'm in the process of writing a pmax emulator, and need to track down
> some bugs. Having a DEBUG enabled kernel might make this a bit easier.)

Even without access to a DEBUG kernel, I've been able to make some
progress. The emulator now runs well enough to let NetBSD/pmax install
itself onto a harddisk image, and the emulator can even be compiled
inside the emulated NetBSD environment (and run NetBSD :-).

  http://www.mdstud.chalmers.se/~md1gavan/mips64emul/

Perhaps emulation is off topic on this list, but since it is pmax and it
runs NetBSD, it might still be of interest to someone.

Anders

Lord Isildur | 19 Apr 17:23 2004
Picon

Re: pmax DEBUG kernel


i for one think this is pretty cool! i still run the real iron, but it
is good to have emulators too! is this the second freely available real
machine emulator which netbsd actually boots on? (like simh) 

isildur

On Sun, 18 Apr 2004, Anders Gavare wrote:

> On Thu, 2 Oct 2003, Anders Gavare wrote:
> 
> > Hi,
> ..
> > (I'm in the process of writing a pmax emulator, and need to track down
> > some bugs. Having a DEBUG enabled kernel might make this a bit easier.)
> 
> Even without access to a DEBUG kernel, I've been able to make some
> progress. The emulator now runs well enough to let NetBSD/pmax install
> itself onto a harddisk image, and the emulator can even be compiled
> inside the emulated NetBSD environment (and run NetBSD :-).
> 
>   http://www.mdstud.chalmers.se/~md1gavan/mips64emul/
> 
> Perhaps emulation is off topic on this list, but since it is pmax and it
> runs NetBSD, it might still be of interest to someone.
> 
> 
> Anders
> 

(Continue reading)

matthew green | 19 Apr 17:47 2004
Picon

re: pmax DEBUG kernel


   i for one think this is pretty cool! i still run the real iron, but it
   is good to have emulators too! is this the second freely available real
   machine emulator which netbsd actually boots on? (like simh) 

netbsd has booted on "psim" (powerpc) for many many years.


Gmane