* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
@ 1998-11-04 1:08 ` David S. Miller
1998-11-04 1:16 ` Stephan van Hienen
` (4 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: David S. Miller @ 1998-11-04 1:08 UTC (permalink / raw)
To: ultralinux
Date: Tue, 3 Nov 1998 12:29:08 +0100 (CET)
From: Stephan van Hienen <ddx@cable.a2000.nl>
no go with the sun happy meal on eth0 (on the scsi/ethernet pci board)
it's connected to a cable modem at 10mbit/sec half duplex
and it gives ethernet time outs
All I can say at this point is that you must have some very strange
ethernet switch/hub because every single switch/hub I test on here at
Cobalt with both 10 and 100baseT & both full and half duplex all work
correctly with the happy meal.
Try another hub/switch/network if you can, I bet it works.
Then we can figure out what is so strange about the one where you have
the problems.
Later,
David S. Miller
davem@dm.cobaltmicro.com
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
1998-11-04 1:08 ` David S. Miller
@ 1998-11-04 1:16 ` Stephan van Hienen
1998-11-04 10:37 ` Malcolm Beattie
` (3 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Stephan van Hienen @ 1998-11-04 1:16 UTC (permalink / raw)
To: ultralinux
On Tue, 3 Nov 1998, David S. Miller wrote:
> Date: Tue, 3 Nov 1998 12:29:08 +0100 (CET)
> From: Stephan van Hienen <ddx@cable.a2000.nl>
>
> no go with the sun happy meal on eth0 (on the scsi/ethernet pci board)
> it's connected to a cable modem at 10mbit/sec half duplex
> and it gives ethernet time outs
>
> All I can say at this point is that you must have some very strange
> ethernet switch/hub because every single switch/hub I test on here at
> Cobalt with both 10 and 100baseT & both full and half duplex all work
> correctly with the happy meal.
>
> Try another hub/switch/network if you can, I bet it works.
>
> Then we can figure out what is so strange about the one where you have
> the problems.
the card is directly connected to a LANcity cable modem
the only thing i can think of is that the ethernet adres is changed when i
boot the 2.1.125 kernel
i will look this thing out this week
maybe (i hope) it's this
this because the cablemodem only responds to the ethernet adress that has
been programmed in the eeprom (or something) inside the cablemodem
it's only strange that 2.1.112 works ok
and that 2.1.125 doesn't :(
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
1998-11-04 1:08 ` David S. Miller
1998-11-04 1:16 ` Stephan van Hienen
@ 1998-11-04 10:37 ` Malcolm Beattie
1998-11-04 10:41 ` David S. Miller
` (2 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Malcolm Beattie @ 1998-11-04 10:37 UTC (permalink / raw)
To: ultralinux
David S. Miller writes:
> Date: Tue, 3 Nov 1998 12:29:08 +0100 (CET)
> From: Stephan van Hienen <ddx@cable.a2000.nl>
>
> no go with the sun happy meal on eth0 (on the scsi/ethernet pci board)
> it's connected to a cable modem at 10mbit/sec half duplex
> and it gives ethernet time outs
>
> All I can say at this point is that you must have some very strange
> ethernet switch/hub because every single switch/hub I test on here at
> Cobalt with both 10 and 100baseT & both full and half duplex all work
> correctly with the happy meal.
I didn't see the beginning of this thread but the message above
caught my eye. When I tried to install on my ultra5 here (for
details see my "Mixed news: new IP 1.0.9 on Ultra5" message) I had
what sound like a very similar problem with the hme on the SunSwift
(scsi/ethernet) board. It was detected as eth0 successfully (the
onboard hme was detected as eth1) but it didn't seem to be able to
receive packets properly. tcpdump from the remote side showed it
sending out arp queries and showed the corect reply being sent back
but the ultra just kept sending out arp queries as if it hadn't
received anything. The card itself works fine with Solaris. I ended
up removing the card completely after which an installation went
fine with the on-board hme (apart from later non-network problems
that I mention in the other message).
--Malcolm
--
Malcolm Beattie <mbeattie@sable.ox.ac.uk>
Unix Systems Programmer
Oxford University Computing Services
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
` (2 preceding siblings ...)
1998-11-04 10:37 ` Malcolm Beattie
@ 1998-11-04 10:41 ` David S. Miller
1998-11-04 10:55 ` Stephan van Hienen
1998-11-04 12:29 ` Malcolm Beattie
5 siblings, 0 replies; 7+ messages in thread
From: David S. Miller @ 1998-11-04 10:41 UTC (permalink / raw)
To: ultralinux
Date: Wed, 4 Nov 1998 10:37:54 +0000 (GMT)
From: Malcolm Beattie <mbeattie@sable.ox.ac.uk>
It was detected as eth0 successfully (the onboard hme was detected
as eth1) but it didn't seem to be able to receive packets
properly. tcpdump from the remote side showed it sending out arp
queries and showed the corect reply being sent back but the ultra
just kept sending out arp queries as if it hadn't received
anything. The card itself works fine with Solaris. I ended up
removing the card completely after which an installation went fine
with the on-board hme (apart from later non-network problems that I
mention in the other message).
There were some problems with scsi/eth PCI cards on Ultra5's earlier,
when was the most recent kernel you have tried with the card in the
machine? We believe this particular to be fixed by now.
Later,
David S. Miller
davem@dm.cobaltmicro.com
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
` (3 preceding siblings ...)
1998-11-04 10:41 ` David S. Miller
@ 1998-11-04 10:55 ` Stephan van Hienen
1998-11-04 12:29 ` Malcolm Beattie
5 siblings, 0 replies; 7+ messages in thread
From: Stephan van Hienen @ 1998-11-04 10:55 UTC (permalink / raw)
To: ultralinux
On Wed, 4 Nov 1998, David S. Miller wrote:
> Date: Wed, 4 Nov 1998 10:37:54 +0000 (GMT)
> From: Malcolm Beattie <mbeattie@sable.ox.ac.uk>
>
> It was detected as eth0 successfully (the onboard hme was detected
> as eth1) but it didn't seem to be able to receive packets
> properly. tcpdump from the remote side showed it sending out arp
> queries and showed the corect reply being sent back but the ultra
> just kept sending out arp queries as if it hadn't received
> anything. The card itself works fine with Solaris. I ended up
> removing the card completely after which an installation went fine
> with the on-board hme (apart from later non-network problems that I
> mention in the other message).
this looks like my problem
>
> There were some problems with scsi/eth PCI cards on Ultra5's earlier,
> when was the most recent kernel you have tried with the card in the
> machine? We believe this particular to be fixed by now.
kernel 2.1.125 (from rpm's)
>
> Later,
> David S. Miller
> davem@dm.cobaltmicro.com
>
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: Still no go with 10mb on 100mb card
1998-11-03 11:29 Still no go with 10mb on 100mb card Stephan van Hienen
` (4 preceding siblings ...)
1998-11-04 10:55 ` Stephan van Hienen
@ 1998-11-04 12:29 ` Malcolm Beattie
5 siblings, 0 replies; 7+ messages in thread
From: Malcolm Beattie @ 1998-11-04 12:29 UTC (permalink / raw)
To: ultralinux
David S. Miller writes:
> Date: Wed, 4 Nov 1998 10:37:54 +0000 (GMT)
> From: Malcolm Beattie <mbeattie@sable.ox.ac.uk>
>
> It was detected as eth0 successfully (the onboard hme was detected
> as eth1) but it didn't seem to be able to receive packets
> properly. tcpdump from the remote side showed it sending out arp
> queries and showed the corect reply being sent back but the ultra
> just kept sending out arp queries as if it hadn't received
> anything. The card itself works fine with Solaris. I ended up
> removing the card completely after which an installation went fine
> with the on-board hme (apart from later non-network problems that I
> mention in the other message).
>
> There were some problems with scsi/eth PCI cards on Ultra5's earlier,
> when was the most recent kernel you have tried with the card in the
> machine? We believe this particular to be fixed by now.
This was with Jakub's UP 1.0.9 update last Friday (30 Oct). For more
details, see my "Mixed news" message. I'm about to follow that up
right now with details of my attempt today with Jakub's latest
update that he put there on Monday.
--Malcolm
--
Malcolm Beattie <mbeattie@sable.ox.ac.uk>
Unix Systems Programmer
Oxford University Computing Services
^ permalink raw reply [flat|nested] 7+ messages in thread