From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.2.1) Gecko/20010901 Description of problem: After a period of time, 25 minutes to 36 hours, tulip.o will stop responding in the 2.4.9-smp-i686 kernel(stock or compiled using source with RH's kernel-2.4.9-i686-smp.config.) TULIP_MWI and TULIP_MMIO configurations do not appear to correct the problem. The problem has been verified with 2.4.9-7 under RH7.2 and 2.4.9-6 under RH7.1. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Update to 2.4.9 2. Wait 3. Actual Results: No traffic is able to get through. /sbin/ifconfig shows eth0 RX incrementing but TX remains the same. Additional info: The network can be restored through ifdown eth0 and ifup eth0 though the tulip module did need to be rmmod'ed at one point. No problems have been experienced with tulip.o under the up-i586 version of 2.4.9 running under 7.1. Due to the length of time to reproduce the error, I have not been able to verify whether the up-i686 version of the kernel causes the same problem(testing system == main workstation.)
There is also the version of tulip we shipped in 2.4.2-2 and 2.4.3-12 called tulip_old.o ; can you see if that works fine in your prefered configuration ? (I'm considering making the tulip_old the default due to your problem and one or two similar reports)
I installed 7.2 on a system with a netgear FA310TX ethernet card (tulip) and after about 10 - 15 minutes (1800 pings) my 7.2 system stopped responding. An ifdown, an rmmod of tulip.o, and an ifup fixed the problem.
I have seen the same problem on a non-SMP i686 kernel, RH 7.2 kernel 2.4.9-13 on a Duron/VIA KT133a system with a Netgear FA310TX card. The tulip driver reports a "Lite-On 82c168 PNIC rev 32". Unfortunately, this problem is hard to reproduce. I have seen it happen only once under heavy network load (copying many messages from a local folder to an IMAP folder on the server in question). As others report, network traffic stops but a "service network restart" got things moving again.
I have verified that tulip_old.o does indeed work--I now have two smp-i686 systems that are both running LNE100TXs and have a two week uptime on both with no network interruptions. Change `alias eth0 tulip' to `alias eth0 tulip_old' in /etc/modules.conf to correct.
found same problem with Intel EtherExpress 1000 (e1000)... :( I use a Dell PowerEdge 1650, with 1G RAM, and two CPU... Kernel : 2.4.7-10smp
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/