Re: DELQA (qe) breaking RLV12/RL02(rl) functionality

[Scott G Taylor <bub7734%yahoo.com@localhost>]

----- Original Message ----
> From: Johnny Billquist <bqt%softjar.se@localhost>
> To: Scott G Taylor <bub7734%yahoo.com@localhost>
> Cc: port-vax%netbsd.org@localhost
> Sent: Thu, February 3, 2011 6:36:31 AM
> Subject: Re: DELQA (qe) breaking RLV12/RL02(rl) functionality
> 
> On 02/03/11 03:28, Scott G Taylor wrote:
> > I'm finding that I can't have  both an RLV12 (RL02) controller and a DELQA
> > (non-turbo)
> >  installed in my KA655 machine at the same time when running netbsd 5.1.
> > 
> > When both are installed, the DELQA runs fine.  However, access to  the rl02
> > drives
> > results in blocks of zero (NULL) data... if it  doesn't crash the kernel.
> > 
> > Removing the DELQA board results is a  perfectly working RL02 controller and
> > drives.  I actually went  through three RLV12 (M8061) boards before 
>accidentally
> > discovering that  the boards weren't actually bad.  I just can't have both 
>the
> > RL02  drives and the DELQA in at the same time.
> > 
> > I have moved boards  around, as well as swapped in multiple boards of each
> > type with no  success.
> > 
> > Before I spout the stripped kernel config, dmesg,  etc., is there anything
> > that I don't know about this controller  combination and NetBSD 5.1?
> > Is anyone running RL02/RLV12(M8061) and  DELQA(M7516-YM) boards
> > together on their QBUS VAX with NetBSD, and  actually using the RL02?
> 
> That sounds very strange, and that kind of combo  was not unusual in the past 
>(although not with NetBSD I would guess).
> A few  things comes to mind:
> I assume that the bus chain was kept intact as you  added/removed cards and 
>fooled around. You know that depending on which  backplane you have, the bus 
>might not run in a straight line.
> Also, make sure  that you don't have any conflicts for interrupt vectors.
> The RLV12 sets the  interrupt vector with jumpers. I can't remember for sure 
>how the DELQA do, but I  seem to remember it being fixed by a jumper or two on 
>that one as well.
> 
> I  don't know how good the RL device driver is, but there might be problems 
> in  
>there as well. I doubt it has been tested or exercised much...
> But if it  works when the DELQA is removed, that would suggest the driver 
>should be  ok.
> 
> Feel free to post the dmesg. Maybe something will stand  out.
> 
> My main suspect might be the interrupt vector, but I would have  expected 
> some 
>error or warning message if we have a  conflict...
> 
>     Johnny

Hei Johnny,

I've duplicated the problem now in simh's vax KA655 emulation.  This would 
suggest there is indeed
some conflict between the qe and rlc drivers.  I won't have access to the real 
hardware for another 4-5 hours
when I get home from work.

In simh, disabling the qe device allows the rlc controller to function 
seemingly 
perfectly for RL02 drives
just as it would work well on the real KA655 when the DELQA is removed.  I 
further note that utilizing
the qe device in SIMH while the rlc controller is enabled has also produced a 
crash.   Simh tends
not to result in a kernel crash as easily as the real VAX does, however.  This 
could just be the result
of the GENERIC kernel on SIMH vs a stripped down kernel on the real vax.

When only one device or the other is installed on SIMH or a real KA655, the 
installed device appears
to function fine.  Just can't have both configured in the kernel.

As Patrick mentioned in a separate email in this thread, the combination works 
fine in VMS.  I don't
believe there to be a condition or configuration issue with the physical 
hardware.

With the above stated, your assumptions are correct: the bus chain is correct 
(mine's a serpentine Q22-type
BA23), and the interrupt vectors are not at issue (the DELQA vectoring is soft; 
the RLV12 is by jumper).

While I could stick my neck out and state that each driver works well, they 
apparently don't work
together.  I don't know which is the problem, or if both are.  But it seems 
clear to me that they
can't both be used.

My testing today on simh utilized the GENERIC kernel (maxusers 8) on a clean 
install.  My real vax KA655 testing
was performed with a stripped down kernel (maxusers 32).  The maxusers was 
bumped up on the real vax in case
this was a buffer/kernel structure size issue.

The following exhibits the general problem with accessing the RL02 drive when 
the DELQA is 

installed:

# dd if=/netbsd of=/dev/rrl0c bs=512x20 conv=sync
label: 0
rl0: no disk labelrl0: size 20480 sectors
280+1 records in
281+0 records out
2877440 bytes transferred in 3.650 secs (788339 bytes/sec)

[[ The above indicates to problems writing to the device]]

# dd if=/dev/rrl0c bs=512x20 conv=sync | cmp - /netbsd
label: 0
rl0: no disk labelrl0: size 20480 sectors
stdin /netbsd differ: char 1, line 1

[[ Reading from the RL02 fails to compare what it just wrote ]]

# dd if=/dev/rrl0c bs=512
label: 0
rl0: no disk labelrl0: size 20480 sectors

^C116+0 records in
115+0 records out
58880 bytes transferred in 6.630 secs (8880 bytes/sec)

[[ Generally no errors reported reading from the drive...]]

# dd if=/dev/rrl0c bs=512 count=10 | wc
label: 0
rl0: no disk labelrl0: size 20480 sectors
10+0 records in
10+0 records out
5120 bytes transferred in 0.040 secs (128000 bytes/sec)
       0       1    5120

[[Only NULLs being read from the drive, in spite of having just written /netbsd 
directly to it]]

Now, on multiple occasions, reading or writing from the drive has resulted in a 
crash
of the netbsd kernel, but I've only hit that once (out of a few attempts) in 
SIMH.


I don't believe that dmesg or kernel config messages are useful at this point.
But here's the GENERIC boot-up under simh (which, to its credit is identical to 
the real KA655
with the exception of added TK50 and the ts0 devices, as well as 512MB memory
vs only 32MB on my real vax):

>> NetBSD/vax boot [1.11 Sat Nov  6 19:40:01 UTC 2010] <<
>> Press any key to abort autoboot 5
Press '?' for help
> boot -s
2403628+308444 [194288+184371]=0x2f2c58
Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005,
    2006, 2007, 2008, 2009, 2010
    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 5.1 (GENERIC) #0: Sat Nov  6 19:48:36 UTC 2010
        
builds%b8.netbsd.org@localhost:/home/builds/ab/netbsd-5-1-RELEASE/vax/201011061943Z-obj/home/builds/ab/netbsd-5-1-RELEASE/src/sys/arch/vax/compile/GENERIC

MicroVAX 3800/3900
total memory = 511 MB
avail memory = 495 MB
mainbus0 (root)
cpu0 at mainbus0: KA655, CVAX microcode rev 6 Firmware rev 83
lance at mainbus0 not configured
uba0 at mainbus0: Q22
dz1 at uba0 csr 160100 vec 304 ipl 17
mtc0 at uba0 csr 174500 vec 774 ipl 17
mscpbus0 at mtc0: version 5 model 3
mscpbus0: DMA burst size set to 4
mt0 at mscpbus0 drive 0: TK50
mt1 at mscpbus0 drive 1: TK50
mt2 at mscpbus0 drive 2: TK50
mt3 at mscpbus0 drive 3: TK50
uda0 at uba0 csr 172150 vec 770 ipl 17
mscpbus1 at uda0: version 3 model 3
mscpbus1: DMA burst size set to 4
ra0 at mscpbus1 drive 0: RRD40
ra1 at mscpbus1 drive 1: RA82
ra2 at mscpbus1 drive 2: RA82
ra3 at mscpbus1 drive 3: RA82
qe0 at uba0 csr 174440 vec 764 ipl 17: delqa, hardware address 08:00:2b:11:22:33
rlc0 at uba0 csr 174400 vec 160 ipl 17
rl0 at rlc0 drive 0: RL02, drive ready
rl1 at rlc0 drive 1: RL01, drive not loaded
rl2 at rlc0 drive 2: RL01, drive not loaded
rl3 at rlc0 drive 3: RL01, drive not loaded
ts0 at uba0 csr 172520 vec 224 ipl 17: TS11
ts0: rev 0, extended features enabled, transport offline
Kernelized RAIDframe activated
ra0label: 0
: no disk label: size 1331200 sectors
ra1: size 4194685 sectors
ra2: attempt to bring on line failed:  unit offline (not mounted) (code 3, 
subcode 1)
ra3: attempt to bring on line failed:  unit offline (not mounted) (code 3, 
subcode 1)
label: 63616c61
rl0: no disk labelrl0: size 20480 sectors
boot device: ra1
root on ra1a dumps on ra1b
root file system type: ffs
WARNING: clock lost 277 days
WARNING: using filesystem time
WARNING: CHECK AND RESET THE DATE!
Enter pathname of shell or RETURN for /bin/sh: 

At this point i'm avoiding installation of both devices.  However, that defeats 
the purpose of my wanting
to install NetBSD on this vax: to be able to image RL02 devices and transfer 
them over the network
to another machine for archival purposes :)

Takk,
-scott