Hi, I'd like to report a bug with e1000 driver. I have Intel PRO/1000 MT nic which works fine as long 'rx-checksumming' and 'tx-checksumming' is disabled. newton:~# ethtool -k eth0 Offload parameters for eth0: rx-checksumming: off tx-checksumming: off scatter-gather: on tcp segmentation offload: on If either rx or tx checksumming is enabled and I start copying large file trees over NFS I see following in my logs: Feb 14 13:28:32 newton kernel: NFS: readdir reply truncated! Feb 14 13:28:33 newton kernel: NFS: readdir reply truncated! Feb 14 13:29:13 newton kernel: NFS: READ count doesn't match RPC opaque count. Feb 14 13:29:14 newton kernel: NFS: READ count doesn't match RPC opaque count. Feb 14 13:29:22 newton kernel: nfs: server corfe not responding, still trying Feb 14 13:29:30 newton kernel: NETDEV WATCHDOG: eth0: transmit timed out Feb 14 13:29:33 newton kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex Feb 14 13:29:34 newton kernel: nfs: server corfe OK Feb 14 13:29:39 newton kernel: nfs: server corfe not responding, still trying Feb 14 13:29:40 newton kernel: e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex Feb 14 13:29:41 newton kernel: nfs: server corfe OK newton:~# uname -a Linux newton 2.6.3-rc2newton.4 #1 Thu Feb 12 23:37:34 CET 2004 i686 GNU/Linux thanks, daniel ps: I hope this is the right place to report a nic driver bug. pss: I'm not on the list, so please CC me. - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
