偶现NETDEV WATCHDOG: eth0: transmit timed out ,求指导

cuiyilong21

运行一段时间出现以下打印，端口被重置
2014-03-09 19:18:49.081 kernel: NETDEV WATCHDOG: eth0: transmit timed out
2014-03-09 19:19:24.080 kernel: NETDEV WATCHDOG: eth0: transmit timed out (repeated 7 times)
2014-03-09 19:19:36.000 kernel: NETDEV WATCHDOG: eth0: transmit timed out (repeated 2 times)
2014-03-09 19:19:36.000 kernel: e1000e use e1000_clean_rx_irq
2014-03-09 19:19:36.024 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
2014-03-09 19:19:46.030 kernel: NETDEV WATCHDOG: eth0: transmit timed out
2014-03-09 19:20:21.031 kernel: NETDEV WATCHDOG: eth0: transmit timed out (repeated 7 times)
2014-03-09 19:20:40.762 kernel: NETDEV WATCHDOG: eth0: transmit timed out (repeated 3 times)
2014-03-09 19:20:40.762 kernel: 0000:06:00.0: eth0: Detected Tx Unit Hang:
2014-03-09 19:20:40.762 kernel:   TDH                  <19>
2014-03-09 19:20:40.762 kernel:   TDT                  <1d>
2014-03-09 19:20:40.762 kernel:   next_to_use          <1d>
2014-03-09 19:20:40.762 kernel:   next_to_clean        <19>
2014-03-09 19:20:40.762 kernel: buffer_info[next_to_clean]:
2014-03-09 19:20:40.762 kernel:   time_stamp           <21e1c64e>
2014-03-09 19:20:40.762 kernel:   next_to_watch        <19>
2014-03-09 19:20:40.762 kernel:   jiffies              <21e1df98>
2014-03-09 19:20:40.762 kernel:   next_to_watch.status <0>
2014-03-09 19:20:40.895 kernel: e1000e use e1000_clean_rx_irq
2014-03-09 19:20:40.963 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
2014-03-09 19:20:41.063 kernel: e1000e use e1000_clean_rx_irq
2014-03-09 19:20:41.090 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None


在网上搜了不少信息，但由于是现网使用环境，不能做什么操作尝试
想复现出来该打印，不知有什么好方法

附一些信息

[root@localhost ~]# ethtool -S eth0
NIC statistics:
     rx_packets: 3175211793
     tx_packets: 3871596970
     rx_bytes: 1490223283937
     tx_bytes: 1284205829103
     rx_broadcast: 118764
     tx_broadcast: 57
     rx_multicast: 3645364
     tx_multicast: 0
     rx_errors: 0
     tx_errors: 0
     tx_dropped: 0
     multicast: 3645364
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_no_buffer_count: 2102
     rx_missed_errors: 78238
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     tx_timeout_count: 23
     tx_restart_queue: 4
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 0
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 0
     rx_flow_control_xoff: 0
     tx_flow_control_xon: 0
     tx_flow_control_xoff: 0
     rx_long_byte_count: 1490223283937
     rx_csum_offload_good: 2390702267
     rx_csum_offload_errors: 2
     rx_header_split: 0
     alloc_rx_buff_failed: 0
     tx_smbus: 0
     rx_smbus: 0
     dropped_smbus: 0
     rx_dma_failed: 0
     tx_dma_failed: 0



[root@localhost /]# ethtool -i eth0
driver: e1000e
version: 0.5.18.3-NAPI
firmware-version: 5.0-2
bus-info: 0000:05:00.0



ethtool -k eth0
Offload parameters for eth0:
Cannot get device udp large send offload settings: Operation not supported
rx-checksumming: on
tx-checksumming: on
scatter-gather: on
tcp segmentation offload: off
udp fragmentation offload: off
generic segmentation offload: off

humjb_1983

发送超时了，就是说DMA在指定时间内没有将数据发送出去，通常是网卡或驱动的问题。

cuiyilong21

回复 2# humjb_1983


    感谢关注！

    驱动版本是09年的
    硬件也已经换过了，应该不是硬件问题

humjb_1983

升级下驱动试试呢~