[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

NIC problem



Hi all,

Finally, a _real_ Linux problem!  It's taken me months to get one!

The box is a bridging firewall, with a custom compiled kernel.  I've
done this configuration loads of times, but not with these NICs.

Here's some pertinent info:
00:0a.0 Ethernet controller: National Semiconductor Corporation DP83815
(MacPhyter) Ethernet Controller
00:0b.0 Ethernet controller: Silicon Integrated Systems [SiS] SiS900
10/100 Ethernet (rev 02)

The natsemi driver is built as a module, and the SiS900 driver is
compiled into the kernel.

The problem is that clients are finding WAN access very slow.  Here's
why - we're dropping packets on eth0:
br0       Link encap:Ethernet  HWaddr 00:06:4F:03:FE:73  
          inet addr:82.68.135.113  Bcast:82.68.135.119
Mask:255.255.255.248
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:139468 errors:0 dropped:0 overruns:0 frame:0
          TX packets:7293 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:11348234 (10.8 MiB)  TX bytes:527446 (515.0 KiB)

eth0      Link encap:Ethernet  HWaddr 00:06:4F:03:FE:73  
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:607499 errors:0 dropped:0 overruns:0 frame:0
          TX packets:440599 errors:0 dropped:41853 overruns:0 carrier:0
          collisions:0 txqueuelen:100 
          RX bytes:325194062 (310.1 MiB)  TX bytes:102535968 (97.7 MiB)
          Interrupt:5 Base address:0x8000 

eth1      Link encap:Ethernet  HWaddr 00:09:5B:0A:64:B2  
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:480767 errors:0 dropped:0 overruns:0 frame:0
          TX packets:589029 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100 
          RX bytes:116528647 (111.1 MiB)  TX bytes:323089149 (308.1 MiB)
          Interrupt:12 Base address:0x7000 

hotbox:~# tail /var/log/messages
Nov 21 03:05:53 hotbox kernel: NETDEV WATCHDOG: eth0: transmit timed out
Nov 21 03:05:53 hotbox kernel: eth0: Transmit timeout, status 00000004
00000000 
Nov 21 03:08:01 hotbox kernel: NETDEV WATCHDOG: eth0: transmit timed out
Nov 21 03:08:01 hotbox kernel: eth0: Transmit timeout, status 00000004
00000000 
Nov 21 03:08:09 hotbox kernel: NETDEV WATCHDOG: eth0: transmit timed out
Nov 21 03:08:09 hotbox kernel: eth0: Transmit timeout, status 00000004
00000000 
Nov 21 03:09:25 hotbox kernel: NETDEV WATCHDOG: eth0: transmit timed out
Nov 21 03:09:25 hotbox kernel: eth0: Transmit timeout, status 00000004
00000000 
Nov 21 03:10:01 hotbox kernel: NETDEV WATCHDOG: eth0: transmit timed out
Nov 21 03:10:01 hotbox kernel: eth0: Transmit timeout, status 00000004
00000000 
hotbox:~# 

I'm ashamed to say that I don't know how to see which card is eth0.  I'm
not on site, and I'd rather not go there just for this.  I'm guessing
it's the one with the lower PCI id (the natsemi)?

APIC is compiled into the kernel, and I've heard that removing it
might help.  Anyone have any advice?  (Running 2.4.18.)

A



Reply to: