Problem with E1000 module or Intel 82541PI network card

[Gijs]

Hello List,

I'm having problems with "Intel Corporation 82541PI Gigabit Ethernet
Controller (rev 05)". Most of the time it works ok, but sometimes the
connection just crashes. Now the last time I saw about 760 million
errors appearing on the interface. When I checked /var/log/messages, I
came across the following stacktrace:

NETDEV WATCHDOG: eth1: transmit timed out
------------[ cut here ]------------
WARNING: at net/sched/sch_generic.c:222 dev_watchdog+0x92/0xe3() (Not
tainted)
Modules linked in: appletalk it87 hwmon_vid fuse sunrpc ipt_REJECT
ipt_ULOG xt_limit ipt_LOG xt_tcpudp xt_state ipt_MASQUERADE iptable_nat
nf_nat nf_conntrack_ipv4 nf_conntrack iptable_filter ip_tables x_tables
ipv6 cpufreq_ondemand powernow_k8 dm_multipath snd_hda_intel
snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device
snd_pcm_oss snd_mixer_oss snd_pcm snd_timer i2c_nforce2 snd_page_alloc
e1000 snd_hwdep i2c_core k8temp usb_storage snd parport_pc hwmon parport
forcedeth soundcore pcspkr floppy sg pata_amd dm_snapshot dm_zero
dm_mirror dm_log dm_mod pata_acpi ata_generic sata_nv libata sd_mod
scsi_mod ext3 jbd mbcache uhci_hcd ohci_hcd ehci_hcd [last unloaded:
scsi_wait_scan]
Pid: 0, comm: swapper Not tainted 2.6.26.8-57.fc8 #1
 [<c0428892>] warn_on_slowpath+0x47/0x72
 [<c041f96d>] ? try_to_wake_up+0x1b3/0x1bd
 [<c041f982>] ? default_wake_function+0xb/0xd
 [<c0438d94>] ? autoremove_wake_function+0xf/0x33
 [<c041d786>] ? __wake_up_common+0x2f/0x5a
 [<c041f4f1>] ? __wake_up+0x31/0x3b
 [<c043633e>] ? insert_work+0x4a/0x50
 [<c04365f0>] ? __queue_work+0x28/0x2d
 [<c043664f>] ? queue_work+0x36/0x40
 [<c0436668>] ? schedule_work+0xf/0x11
 [<c05d0935>] dev_watchdog+0x92/0xe3
 [<c04306d6>] run_timer_softirq+0x124/0x18c
 [<c05d08a3>] ? dev_watchdog+0x0/0xe3
 [<c05d08a3>] ? dev_watchdog+0x0/0xe3
 [<c042cdcf>] __do_softirq+0x6d/0xdf
 [<c0406e5e>] do_softirq+0x6c/0xa9
 [<c042cd2f>] irq_exit+0x38/0x6b
 [<c041494c>] smp_apic_timer_interrupt+0x6e/0x7e
 [<c040378d>] ? default_idle+0x0/0x58
 [<c0405654>] apic_timer_interrupt+0x28/0x30
 [<c040378d>] ? default_idle+0x0/0x58
 [<c0419653>] ? native_safe_halt+0x5/0x7
 [<c04037bd>] default_idle+0x30/0x58
 [<c04036cc>] cpu_idle+0xae/0xce
 [<c06217d6>] rest_init+0x4e/0x50
 =======================
---[ end trace 0c5cc5d9805576cc ]---
e1000: eth1: e1000_watchdog: NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: RX/TX
e1000: eth1: e1000_clean_tx_irq: Detected Tx Unit Hang
  Tx Queue             <0>
  TDH                  <4000fc>
  TDT                  <400054>
  next_to_use          <54>
  next_to_clean        <f3>
buffer_info[next_to_clean]
  time_stamp           <5822fb>
  next_to_watch        <f3>
  jiffies              <582d9d>
  next_to_watch.status <0>

I'm running Fedora 8, kernel version 2.6.26.8-57.fc8. Does anyone else
have a similar problem with this type of card/module?

Regards, Gijs