Bug 44572
| Summary: | EPIA onboard VIA VT6103 network if hangs, vr driver problem | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Base System | Reporter: | Teemu Kuusijärvi <tpk> | ||||
| Component: | i386 | Assignee: | silby | ||||
| Status: | Closed FIXED | ||||||
| Severity: | Affects Only Me | ||||||
| Priority: | Normal | ||||||
| Version: | Unspecified | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
> >Description: > After a couple of minutes the network connection hangs. Nothing goes through it. Ifconfig claims that the vr0 is UP and active but 'ifconfig vr0 up' corrects the error. > > Same thing happens when I download and upload simultaneously via ftp. > Single upload or download seems to be working. It reminds me my problem with ste I try to write down in http://docs.freebsd.org/cgi/getmsg.cgi?fetch=126424+0+archive/2002/freebsd-net/20021027.freebsd-net Does you try to determine, RX or TX or both does not work? -- @BABOLO http://links.ru/ Hi! I have the Via Epia too, and unfortunately exactly the same problems with the vr0 device (FreeBSD 4.7-RELEASE). The D-Link 530TX NIC based on the same chipset (Via RhineII) seems to have had this problem, I don't know if it's fixed yet. Linux or Windows seems to work fine with this hardware, so it must be some sort of driveproblem. The problem again: Under heavy load the card somehow disable itself. A simple 'ifconfig up' fixes the problem. Running in 10baseT makes the problem show up less often. Too bad really, the Via Epia is very nifty. I don't want to spoil the whole thing by running linux ;-) regards /johan -- Johan Östensson <johan@ostensson.org> ICQ: 255645 Tfn: +46(0)73 6548283 <johos384@student.liu.se> http://www.ostensson.org Responsible Changed From-To: freebsd-bugs->silby I own all the other vr PRs, why not this one? Hello, I'd like to add that I'm also experiencing this bug running 4.7-RELEASE. I verified that the "ping" test given here causes the lockup. I also have the VT6102 card. -mark http://mark.stosberg.com/ I'd like to add some more data points to my previous report. - I was able to reproduce the error by running this ping command in both directions with another host on my local network: ping -f -q -n -s 10000 host_name Then, even after I stopped flood-pinging the other host, it appeared that net connectivity was lost-- I couldn't execute normal ping commands any more my machine to the outside world. - I ran tcpdump while the network card appeared hung to see what was happening. I observed data flowing both in and out. Here are some sample lines (my computer is "asana"): #### 21:15:27.072814 192.168.1.100 > asana.summersault.com: (frag 47804:1480@1480+) 21:15:27.074056 192.168.1.100 > asana.summersault.com: icmp: echo request (frag 47804:1480@0+) 5273 packets received by filter 4427 packets dropped by kernel ##### # later on... ##### 21:16:04.612632 192.168.1.100 > asana.summersault.com: (frag 51560:1480@1480+) 21:16:04.613884 192.168.1.100 > asana.summersault.com: icmp: echo request (frag 51560:1480@0+) 21:16:04.614051 asana.summersault.com > 192.168.1.100: icmp: echo reply (frag 47759:1480@0+) 21:16:04.614067 asana.summersault.com > 192.168.1.100: (frag 47759:1480@1480+) 533 packets received by filter 0 packets dropped by kernel ##### - rebooting the interface seems to fix it, as illustrated here: su-2.05b# ping summersault.com ping: cannot resolve summersault.com: Host name lookup failure su-2.05b# ifconfig vr0 up su-2.05b# ping summersault.com PING summersault.com (208.10.44.140): 56 data bytes 64 bytes from 208.10.44.140: icmp_seq=0 ttl=52 time=27.255 ms 64 bytes from 208.10.44.140: icmp_seq=1 ttl=52 time=77.052 ms ^C ###### - Sometimes during this process, ping would return these messages, which may or may not be of interest: 556 bytes from 192.168.1.100: Frag reassembly time exceeded ### ping: sendto: No buffer space available ###### - I also tried setting the card to work at 10baseT, by adding this to an ifconfig statement for the interface: media 10baseT/UTP With the configuration, the two way ping floods don't immediately lock up the interface. However, it seems that with enough time, it can still happen. Hello! You have both reported problems with the vr-driver, I have made a patch that might solve this problem. Could you please try this patch and report the result to me. I could reproduce your problem and this patch fixes the problem for me! It have been tested on 4.7-RELEASE with the latest vr-driver from the stable-branch. /thn -- --------------------------------------------------------------- Svensk Aktuell Elektronik AB Thomas Nyström Box 10 Phone: +46 8 35 92 85 S-191 21 Sollentuna Fax: +46 8 59 47 45 36 Sweden Email: thn@saeab.se --------------------------------------------------------------- State Changed From-To: open->closed Thomas's patch is believed to have fixed this problem. I just wanted to thank Thomas for submitting his patch. We were having the same problem with the vr-driver on a 2.9 OpenBSD machine. The if_vr.c code on 2.9 was too old for the patch that Thomas submitted, but I was able to alter the patch enough to make it work. If anybody else out there still on OpenBSD < 3.4 needs this problem fixed as well, then they can upgrade to 3.4 (first release that contains the updates required). If upgrade is not possible for some reason, then email me for the patch. (Since this is the freebsd site, not openbsd, I wont post it here). Thanks again Thomas for the fix! Brett. brettbirkett@yahoo.com.au |
After a couple of minutes the network connection hangs. Nothing goes through it. Ifconfig claims that the vr0 is UP and active but 'ifconfig vr0 up' corrects the error. Same thing happens when I download and upload simultaneously via ftp. Single upload or download seems to be working. WindowsXP works O.K. with this mboard. FBSD 4.7 and Intel EE Pro/100+ works without any problems. 100Mbps switch I'm using is 8 port Lantech Mini Switch 800. There is no difference if I use autoselect or define if options manually. ifconfig -a: ----------- vr0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 inet X.X.X.X netmask 0xfffffff8 broadcast X.X.X.X inet6 X::X:X:X:X%vr0 prefixlen 64 scopeid 0x1 ether 00:40:63:c1:15:56 media: Ethernet 100baseTX <full-duplex> status: active fxp0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> mtu 1500 ether 00:d0:b7:70:1a:33 media: Ethernet autoselect (none) status: no carrier <CLIP> dmesg: ----- FreeBSD 4.7-RELEASE #0: Wed Oct 9 15:08:34 GMT 2002 root@builder.freebsdmall.com:/usr/obj/usr/src/sys/GENERIC Timecounter "i8254" frequency 1193182 Hz CPU: VIA C3 Samuel 2 (800.03-MHz 686-class CPU) Origin = "CentaurHauls" Id = 0x678 Stepping = 8 Features=0x803035<FPU,DE,TSC,MSR,MTRR,PGE,MMX> real memory = 528416768 (516032K bytes) <CLIP> vr0: <VIA VT6102 Rhine II 10/100BaseTX> port 0xe800-0xe8ff mem 0xd6100000-0xd61000ff irq 10 at device 18.0 on pci0 vr0: Ethernet address: 00:40:63:c1:15:56 miibus0: <MII bus> on vr0 ukphy0: <Generic IEEE 802.3u media interface> on miibus0 ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto fxp0: <Intel Pro 10/100B/100+ Ethernet> port 0xec00-0xec3f mem 0xd6000000-0xd60fffff,0xd6101000-0xd6101fff irq 11 at device 20.0 on pci0 fxp0: Ethernet address 00:d0:b7:70:1a:33 inphy0: <i82555 10/100 media interface> on miibus1 inphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto <CLIP> Fix: ifconfig vr0 up How-To-Repeat: ping -f -q -n -s 10000 <ip> or FTP download+upload at the same time.