SUN GEM on 32bit x86 looses connectivity

[Beezly]

--=-NkQvMlngIVPFcpMrlt1r
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

Hi,

after sorting out the non-existant MAC address problem, I've hit another
road block.

The kernel is 2.4.19-pre1

The GEM card connects fine (although the multiple "Link is up" messages
might be interesting);

sungem.c:v0.96 11/17/01 David S. Miller ([e-mail])
PCI: Found IRQ 5 for device 00:0a.0
PCI: Sharing IRQ 5 with 00:0b.1
eth0: Sun GEM (PCI) 10/100/1000BaseT Ethernet 00:00:00:00:00:00=3D20
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: PCS AutoNEG complete.
eth0: PCS link is now up.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Link is up at 1000 Mbps, full-duplex.

Everything appears to work fine; Ping works fine;

But after a short while (usually around a minute), the connection stops
working.=3D20

ifconfig shows this interesting output;

eth0      Link encap:Ethernet  HWaddr 00:10:5A:41:E6:14 =3D20
          inet addr:10.0.0.12  Bcast:10.0.0.255  Mask:255.255.255.0
          inet6 addr: fe80::210:5aff:fe41:e614/10 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:5244 errors:0 dropped:1 overruns:1 frame:1
          TX packets:5252 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100=3D20
          RX bytes:7833582 (7.4 MiB)  TX bytes:7812928 (7.4 MiB)
          Interrupt:5 Base address:0x8400=3D20

notice the RX - dropped:1 overruns:1 frame:1

I suspect that only RX capability is lost, although I haven't been able
to check this yet. I believe this because if I ping from the box with
the GE card in to another box (which is stable), once the GEM stops
working, TX packets increases, whilst RC packets does not.

I /think/ I can reproduce this problem quicker by doing a ping -f -s
1472 <somehost> from this host, although this is a purely qualitative
judgement.

The switch i am connecting to is an extreme summit 48 - which supports
the 802.1q VLAN protocol, if this has anything to do with the problem.

At the moment, my "workaround" is to ifdown the interface every 30
seconds, rmmod sungem and then modprobe sungem (auto-reconfiguring the
interface). This gets the interface going again, but means I loose
connectivity for about 2 seconds out of 30!

Any help gratefully appreciated,

Beezly

--=-NkQvMlngIVPFcpMrlt1r
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQA8h/zvXu4ZFsMQjPgRAvdXAJ459x9AbRWxegYl1LuSBIbzTf+3CgCfWLDk
zr1poKUKEqPMv2UAgcrU3CI=
=trJr
-----END PGP SIGNATURE-----

--=-NkQvMlngIVPFcpMrlt1r--