I have tried kernel 3.9.0-030900rc5-generic, and the computer has now been running for 5 days! (Previously the problem would have surfaced after a day or two). I had some issues initially after booting: it would boot but when trying to login the session would hang whilst running .profile, so I think it's not completely working.
This time though eth2 recovered and the network is still functioning, so that's an improvement but not a complete solution.
I don't know whether that means the problem is fixed or not, so I don't know whether to tag "kernel-bug-exists-upstream" or "kernel-fixed-upstream" (or even "kernel-unable-to-test-upstream" given the initial problems).
Thank you for your suggestions.
I have tried kernel 3.9.0-030900rc5 -generic, and the computer has now been running for 5 days! (Previously the problem would have surfaced after a day or two). I had some issues initially after booting: it would boot but when trying to login the session would hang whilst running .profile, so I think it's not completely working.
I've just had another instance of:
niu 0000:09:00.0: eth2: Transmit timed out, resetting
This time though eth2 recovered and the network is still functioning, so that's an improvement but not a complete solution.
I don't know whether that means the problem is fixed or not, so I don't know whether to tag "kernel- bug-exists- upstream" or "kernel- fixed-upstream" (or even "kernel- unable- to-test- upstream" given the initial problems).
Here's the syslog:
Apr 10 02:58:58 metope2 kernel: [388060.816009] ------------[ cut here ]------------ COD/linux/ net/sched/ sch_generic. c:255 dev_watchdog+ 0x262/0x270( ) netbios_ ns nf_conntrack_ broadcast nf_nat_ftp nf_nat i5400_edac tpm_tis lpc_ich nf_conntrack_ftp edac_core nf_conntrack psmouse dca shpchp i5k_amb joydev serio_raw iptable_filter mac_hid lp ip_tables parport x_tables hid_generic usbhid hid raid10 raid456 async_pq async_xor xor async_memcpy async_raid6_recov e1000e ptp pps_core raid6_pq async_tx niu raid1 raid0 multipath linear -generic #201303311835 53f>] warn_slowpath_ common+ 0x7f/0xc0 636>] warn_slowpath_ fmt+0x46/ 0x50 574>] ? wake_up_ worker+ 0x24/0x30 3f2>] dev_watchdog+ 0x262/0x270 fb0>] ? __queue_ work+0x2a0/ 0x2a0 190>] ? pfifo_fast_ dequeue+ 0xe0/0xe0 3a6>] call_timer_ fn+0x46/ 0x160 e77>] run_timer_ softirq+ 0x267/0x2c0 ad9>] ? read_tsc+0x9/0x20 190>] ? pfifo_fast_ dequeue+ 0xe0/0xe0 0b8>] __do_softirq+ 0xd8/0x270 3b6>] irq_exit+0x96/0xb0 98e>] smp_apic_ timer_interrupt +0x6e/0x99 85d>] apic_timer_ interrupt+ 0x6d/0x80 f08>] ? hrtimer_ start+0x18/ 0x20 1e6>] ? native_ safe_halt+ 0x6/0x10 c35>] default_ idle+0x45/ 0x120 749>] cpu_idle+0xd9/0x120 e85>] start_secondary +0xc3/0xc5
Apr 10 02:58:58 metope2 kernel: [388060.816031] WARNING: at /home/apw/
Apr 10 02:58:58 metope2 kernel: [388060.816037] Hardware name: SUN FIRE X2250
Apr 10 02:58:58 metope2 kernel: [388060.816039] NETDEV WATCHDOG: eth2 (niu): transmit queue 7 timed out
Apr 10 02:58:58 metope2 kernel: [388060.816042] Modules linked in: nfsv3 autofs4 nfsd nfs_acl auth_rpcgss nfs fscache lockd sunrpc tpm_infineon xt_hl ip6t_rt nf_conntrack_ipv6 nf_defrag_ipv6 ast ipt_REJECT xt_LOG ttm xt_limit drm_kms_helper drm xt_tcpudp xt_addrtype coretemp kvm_intel kvm i2c_algo_bit nf_conntrack_ipv4 nf_defrag_ipv4 xt_state sysimgblt sysfillrect ip6table_filter gpio_ich ip6_tables syscopyarea microcode ioatdma nf_conntrack_
Apr 10 02:58:58 metope2 kernel: [388060.816126] Pid: 0, comm: swapper/1 Tainted: G I 3.9.0-030900rc5
Apr 10 02:58:58 metope2 kernel: [388060.816133] Call Trace:
Apr 10 02:58:58 metope2 kernel: [388060.816136] <IRQ> [<ffffffff8105a
Apr 10 02:58:58 metope2 kernel: [388060.816147] [<ffffffff8105a
Apr 10 02:58:58 metope2 kernel: [388060.816156] [<ffffffff81077
Apr 10 02:58:58 metope2 kernel: [388060.816163] [<ffffffff8160f
Apr 10 02:58:58 metope2 kernel: [388060.816169] [<ffffffff81077
Apr 10 02:58:58 metope2 kernel: [388060.816172] [<ffffffff8160f
Apr 10 02:58:58 metope2 kernel: [388060.816180] [<ffffffff8106a
Apr 10 02:58:58 metope2 kernel: [388060.816185] [<ffffffff8106b
Apr 10 02:58:58 metope2 kernel: [388060.816191] [<ffffffff8101b
Apr 10 02:58:58 metope2 kernel: [388060.816198] [<ffffffff8160f
Apr 10 02:58:58 metope2 kernel: [388060.816205] [<ffffffff81063
Apr 10 02:58:58 metope2 kernel: [388060.816209] [<ffffffff81063
Apr 10 02:58:58 metope2 kernel: [388060.816216] [<ffffffff8170e
Apr 10 02:58:58 metope2 kernel: [388060.816221] [<ffffffff8170d
Apr 10 02:58:58 metope2 kernel: [388060.816224] <EOI> [<ffffffff81083
Apr 10 02:58:58 metope2 kernel: [388060.816236] [<ffffffff81045
Apr 10 02:58:58 metope2 kernel: [388060.816243] [<ffffffff8101c
Apr 10 02:58:58 metope2 kernel: [388060.816246] [<ffffffff8101d
Apr 10 02:58:58 metope2 kernel: [388060.816251] [<ffffffff816df
Apr 10 02:58:58 metope2 kernel: [388060.816254] ---[ end trace 4a1e57573edccef8 ]---
Apr 10 02:58:58 metope2 kernel: [388060.816260] niu 0000:09:00.0: eth2: Transmit timed out, resetting