The network link goes down if I copy larger files over the network, then comes up again. No problems with gentoo-sources-3.0.6. If I don't copy or move any data, the network is stable (e.g. normal internet surfing), so it's not a dhcpcd issue Reproducible: Always Steps to Reproduce: 1. copy ~500MB over the network Actual Results: the link goes down every couple of seconds (about 3 times per minute) Expected Results: stable network connection 03:00.0 Ethernet controller: JMicron Technology Corp. JMC250 PCI Express Gigabit Ethernet Controller (rev 05) Subsystem: CLEVO/KAPOK Computer Device 5102 Kernel driver in use: jme Kernel modules: jme Network share is mounted via nfs
Created attachment 290519 [details] /var/log/messages while copying a 1.5GB file over the network
Created attachment 290521 [details] kernel config 3.0.7
I cannot reproduce the problem any more after ~4 hours uptime.
...so, this is a bug or it isn't? ie, did you resolve it by finding actual link (ie, physical / other networking hardware) issues? Also, (if this does end up being a valid bug) the subject mentions a regression; was this issue present in an earlier kernel? Do you know when it was fixed?
This is happening to me too. Form some kernel version, maybe 3.0.0 it stopps network trafic on my eth0. I suspected nfsd, did nfsd debugging and once was happy after I set bigger amount of RAM memmory to be kept free. But next day it happend again. It is unpredictable. I am backing up the problematic machin using NFS over 1 Gbit network. It is autotnegotiated to full duplex, speeds are very good. But after few gigabytes it freezes and I must call ifconfig eth0 down, wait a little and call ifconfig eh0 up and works again. 05:02.0 Ethernet controller [0200]: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller [11ab:4320] (rev 14) Subsystem: ASUSTeK Computer Inc. Marvell 88E8001 Gigabit Ethernet Controller (Asus) [1043:811a] Kernel driver in use: skge Kernel modules: skge It does not depends on a NIC chip type on the other machine and it not depends on distro version too. I tried there fresh Fedora, Centos 5.7 and 6.0 and Ubuntu. With Fedora it maybe crashed earlier.
Created attachment 290597 [details] Kernel config
(In reply to comment #4) > ...so, this is a bug or it isn't? ie, did you resolve it by finding actual > link (ie, physical / other networking hardware) issues? > > Also, (if this does end up being a valid bug) the subject mentions a > regression; was this issue present in an earlier kernel? Do you know when it > was fixed? Sorry, let me explain myself: I had a similar issue here: https://bugs.gentoo.org/show_bug.cgi?id=373109 but this time, the system doesn't freeze :) Unfortunately there is no changelog on kernel.org available to see, if they patched anything on the driver. As David pointed out, I had this with any file I copied over the network yesterday. Then didn't copy anything without reboot (did other stuff) and it was stable. This morning it works fine as well. So yes, I cannot reproduce it.
Sporadic problems suck. Hopefully the kernel guys can help you get to the bottom of this.
Created attachment 290947 [details] /var/log/messages happened again, nearly straight after starting copying. Nothing special happened, another netbook on the same network didn't loose the connection.
FYI happens with gentoo-sources-3.1.0 as well
Can one of you guys identify the last working kernel version? It would be helpful to do a git-bisect between that one and the first non working kernel.
