[Simh] Ultrix network issues

Mark Pizzolato Mark at infocomm.com
Thu Nov 15 20:44:00 EST 2018


On Thursday, November 15, 2018 at 4:31 PM, Douglas Hall wrote:
> On Thu, 15 Nov 2018, at 23:47, Mark Pizzolato wrote:
> 
> > Thanks for the SHOW VERSION output.
> >
> > Meanwhile the more important detail will be things like SHOW ETHERNET
> > And SHOW XQ STATS, and just SHOW XQ (after the OS has exhibited the
> > problem AND after a different OS doesn't exhibit the problem.
> 
> sim> show ethernet
> ETH devices:
>  eth0	eth0                                 (No description available)
>  eth1	eth1                                 (No description available)
>  eth2	eth2                                 (No description available)
>  eth3	tap:tapN                             (Integrated Tun/Tap support)
>  eth4	nat:{optional-nat-parameters}        (Integrated NAT (SLiRP) support)
>  eth5	udp:sourceport:remotehost:remoteport (Integrated UDP bridge
> support)
> libpcap version 1.8.1
> Open ETH Devices:
>  XQ     eth1 (No description available)
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Packets Sent:            38
>   Packets Received:        51
>   Asynch Interrupts:       Enabled
>   Interrupt Latency:       0 uSec
>   Read Queue: Count:       0
>   Read Queue: High:        2
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 1A:AD:FA:BE:CE:19) or (ether dst FF:FF:FF:FF:FF:FF) or
> (ether dst CF:00:00:00:00:00)))
> 
> sim> show xq stats
> XQ Ethernet statistics:
>   Recv:          55
>   Dropped:       0
>   Xmit:          36
>   Xmit Fail:     0
>   Runts:         0
>   Oversize:      0
>   SW Reset:      1
>   Setup:         2
>   Loopback:      0
>   Recv Overrun:  0
>   ReadQ count:   0
>   ReadQ high:    2
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Packets Sent:            39
>   Packets Received:        55
>   Asynch Interrupts:       Enabled
>   Interrupt Latency:       0 uSec
>   Read Queue: Count:       0
>   Read Queue: High:        2
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 1A:AD:FA:BE:CE:19) or (ether dst FF:FF:FF:FF:FF:FF) or
> (ether dst CF:00:00:00:00:00)))
> 
> sim> show xq
> XQ	address=20001920-2000192F, vector=3F8*, BR4,
> MAC=1A:AD:FA:BE:CE:19
> 	type=DELQA-T, mode=DEQNA, polling=disabled, sanity=OFF
> 	throttle=disabled, DEQNALock=OFF, leds=(ON,ON,ON)
> 	attached to eth1
> 
> 
> After problem manifests itself:
> 
> Remote Console Command from 198.18.0.50> show xq
> XQ	address=20001920-2000192F, vector=3F8*, BR4,
> MAC=1A:AD:FA:BE:CE:19
> 	type=DELQA-T, mode=DEQNA, polling=disabled, sanity=OFF
> 	throttle=disabled, DEQNALock=OFF, leds=(ON,ON,ON)
> 	attached to eth1
> 
> sim> show xq stats
> XQ Ethernet statistics:
>   Recv:          422
>   Dropped:       0
>   Xmit:          255
>   Xmit Fail:     0
>   Runts:         0
>   Oversize:      0
>   SW Reset:      1
>   Setup:         2
>   Loopback:      0
>   Recv Overrun:  0
>   ReadQ count:   0
>   ReadQ high:    8
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Jumbo Fragmented:        7
>   Packets Sent:            258
>   Packets Received:        422
>   Asynch Interrupts:       Enabled
>   Interrupt Latency:       0 uSec
>   Read Queue: Count:       0
>   Read Queue: High:        8
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 1A:AD:FA:BE:CE:19) or (ether dst FF:FF:FF:FF:FF:FF) or
> (ether dst CF:00:00:00:00:00)))
> 
> 
> With NetBSD:
> 
> sim> show ethernet
> ETH devices:
>  eth0	eth0                                 (No description available)
>  eth1	eth1                                 (No description available)
>  eth2	eth2                                 (No description available)
>  eth3	tap:tapN                             (Integrated Tun/Tap support)
>  eth4	nat:{optional-nat-parameters}        (Integrated NAT (SLiRP) support)
>  eth5	udp:sourceport:remotehost:remoteport (Integrated UDP bridge
> support)
> libpcap version 1.8.1
> Open ETH Devices:
>  XQ     eth1 (No description available)
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Packets Sent:            3
>   Asynch Interrupts:       Disabled
>   Read Queue: Count:       0
>   Read Queue: High:        0
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 00:00:00:00:00:00) or (ether multicast)))
> 
> sim> show xq stats
> XQ Ethernet statistics:
>   Recv:          0
>   Dropped:       0
>   Xmit:          0
>   Xmit Fail:     0
>   Runts:         0
>   Oversize:      0
>   SW Reset:      1
>   Setup:         1
>   Loopback:      0
>   Recv Overrun:  0
>   ReadQ count:   0
>   ReadQ high:    0
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Packets Sent:            3
>   Asynch Interrupts:       Disabled
>   Read Queue: Count:       0
>   Read Queue: High:        0
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 00:00:00:00:00:00) or (ether multicast)))
> 
> sim> show xq
> XQ	address=20001920-2000192F, vector=3F4*, BR4,
> MAC=08:00:2B:23:AA:69
> 	type=DELQA-T, polling=disabled, sanity=OFF
> 	throttle=disabled, DEQNALock=OFF, leds=(ON,ON,ON)
> 	attached to eth1
> 
> 
> After transferring file that failed on Ultrix:
> 
> sim> show xq stats
> XQ Ethernet statistics:
>   Recv:          5985
>   Dropped:       0
>   Xmit:          3572
>   Xmit Fail:     0
>   Runts:         0
>   Oversize:      0
>   SW Reset:      1
>   Setup:         6
>   Loopback:      0
>   Recv Overrun:  0
>   ReadQ count:   0
>   ReadQ high:    24
> Ethernet Device:
>   Name:                    eth1
>   Reflections:             0
>   Self Loopbacks Sent:     2
>   Self Loopbacks Rcvd:     0
>   Host NIC Address:        1A:DE:AD:BE:EF:19
>   Jumbo Fragmented:        1867
>   Packets Sent:            3575
>   Packets Received:        5985
>   Asynch Interrupts:       Enabled
>   Interrupt Latency:       0 uSec
>   Read Queue: Count:       0
>   Read Queue: High:        24
>   Read Queue: Loss:        0
>   Peak Write Queue Size:   1
>   BPF Filter: (((ether dst 08:00:2B:23:AA:69) or (ether multicast)))
> 
> 
> sim> show xq
> XQ	address=20001920-2000192F, vector=3F4*, BR4,
> MAC=08:00:2B:23:AA:69
> 	type=DELQA-T, polling=disabled, sanity=OFF
> 	throttle=disabled, DEQNALock=OFF, leds=(ON,ON,ON)
> 	attached to eth1
> 
> 
> 
> > In your configuration file for BOTH the working and the non-working cases
> > please try variations of:
> >
> >       SET XQ TYPE=DEQNA
> >       SET XQ TYPE=DELQA
> >       SET XQ TYPE=DELQA-T
> 
> I already tried each of those on Ultrix before writing my first email to the list. It
> made no difference. I noted that in the output of show xq the mode always
> defaults to DEQNA, not sure if that's relevant. Do you still want me to try test
> each of those configurations on the netbsd sim and/or see the output from the
> XQ commands for each variation of the ethernet controller under Ultrix?

Well, if you leave the TYPE=DELQA-T and both VMS and NetBSD end up using it 
In DELQA-T mode, while Ultrix is using it in DEQNA mode, then we've got a 
difference.  The programming mode of the DELQA-T is substantially different
from the DEQNA mode.  There may be buffer descriptor issues, but those buffer 
descriptors should have been the same with your previous host environment.

Additionally, if you force the device to be a DEQNA, then VMS and NetBSD won't 
have a choice about which mode to run it in, and you might see problems there 
as well.

-  Mark

> >
> > This change would occur in your configuration file before you attach the
> > device.
> >
> > If you still get everything but Ultrix working for each of these cases, then
> > the only difference is the potential timing regarding how the Ultrix driver
> > handles the buffer descriptor lists vs how the other operating systems do
> > it.




More information about the Simh mailing list