[opensuse] Network card driver problem
Hi, ever since I moved my system to an msi 945 based motherboard I have had an irritating intermittent problem with my Ethernet. It works fine, including after power off power on most of the time until it decides not to. I power on my pc and suddenly eth0, although showing up as working does not connect to the router (have tried restarting router etc & there's a vista machine with same giglan type that does not have problem). I've tried various things like starting up with lan disabled in the bios and then restarting with it enabled again, changing cables etc etc. A few weeks ago I recompiled the kernel with an internal lan driver and I had no more problems. I upgraded my kernel and all worked fine until a power fail brought the problem back again. I reinstalled kernel and eth0 worked again for two power cycles then this morning same problem, which was easy to fix, I just reinstalled the kernel and rebooted. The key symptom of failure is when ethtool shows Speed: 100Mb/s it will connect properly but if it shows Speed: 1000Mb/s the system just sits trying to connect to the router but the leds don't indicate any network activity. I've tried to set the speed to 100 but that does not happen, in fact whether eth0 connects or not ethtool cannot change any settings on lan anyway. It seems to me like the driver thinks it is working with eth0 but in fact it doesn't. Please help, I'm stumped. Chip is RTL8111. hwinfo --netcard 08: PCI 300.0: 0200 Ethernet controller [Created at pci.310] UDI: /org/freedesktop/Hal/devices/pci_10ec_8168 Unique ID: rBUF.tMmV8WnWhuE Parent ID: qTvu.4chw4Z9dkyF SysFS ID: /devices/pci0000:00/0000:00:1c.1/0000:03:00.0 SysFS BusID: 0000:03:00.0 Hardware Class: network Model: "Realtek RTL8111/8168B PCI Express Gigabit Ethernet controller" Vendor: pci 0x10ec "Realtek Semiconductor Co., Ltd." Device: pci 0x8168 "RTL8111/8168B PCI Express Gigabit Ethernet controller" SubVendor: pci 0x1462 "Micro-Star International Co., Ltd." SubDevice: pci 0x507c Revision: 0x02 Driver: "r8169" Driver Modules: "r8169" Device File: eth0 I/O Ports: 0xd800-0xd8ff (rw) Memory Range: 0xefeff000-0xefefffff (rw,prefetchable) Memory Range: 0xefee0000-0xefeeffff (rw,prefetchable) Memory Range: 0xcfcf0000-0xcfcfffff (ro,prefetchable,disabled) IRQ: 220 (11764 events) HW Address: 00:1d:92:82:05:c1 Link detected: yes Module Alias: "pci:v000010ECd00008168sv00001462sd0000507Cbc02sc00i00" Driver Info #0: Driver Status: r8169 is active Driver Activation Cmd: "modprobe r8169" Config Status: cfg=no, avail=yes, need=no, active=unknown Attached to: #19 (PCI bridge) Thanks in advance. Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
Hi, ever since I moved my system to an msi 945 based motherboard I have had an irritating intermittent problem with my Ethernet. It works fine, including after power off power on most of the time until it decides not to. I power on my pc and suddenly eth0, although showing up as working does not connect to the router (have tried restarting router etc & there's a vista machine with same giglan type that does not have problem). I've tried various things like starting up with lan disabled in the bios and then restarting with it enabled again, changing cables etc etc. A few weeks ago I recompiled the kernel with an internal lan driver and I had no more problems. I upgraded my kernel and all worked fine until a power fail brought the problem back again. I reinstalled kernel and eth0 worked again for two power cycles then this morning same problem, which was easy to fix, I just reinstalled the kernel and rebooted. The key symptom of failure is when ethtool shows Speed: 100Mb/s it will connect properly but if it shows Speed: 1000Mb/s the system just sits trying to connect to the router but the leds don't indicate any network activity. I've tried to set the speed to 100 but that does not happen, in fact whether eth0 connects or not ethtool cannot change any settings on lan anyway. It seems to me like the driver thinks it is working with eth0 but in fact it doesn't. Please help, I'm stumped.
Any error messages from "dmesg"? /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Dave Plater wrote:
Hi, ever since I moved my system to an msi 945 based motherboard I have had an irritating intermittent problem with my Ethernet. It works fine, including after power off power on most of the time until it decides not to. I power on my pc and suddenly eth0, although showing up as working does not connect to the router (have tried restarting router etc & there's a vista machine with same giglan type that does not have problem). I've tried various things like starting up with lan disabled in the bios and then restarting with it enabled again, changing cables etc etc. A few weeks ago I recompiled the kernel with an internal lan driver and I had no more problems. I upgraded my kernel and all worked fine until a power fail brought the problem back again. I reinstalled kernel and eth0 worked again for two power cycles then this morning same problem, which was easy to fix, I just reinstalled the kernel and rebooted. The key symptom of failure is when ethtool shows Speed: 100Mb/s it will connect properly but if it shows Speed: 1000Mb/s the system just sits trying to connect to the router but the leds don't indicate any network activity. I've tried to set the speed to 100 but that does not happen, in fact whether eth0 connects or not ethtool cannot change any settings on lan anyway. It seems to me like the driver thinks it is working with eth0 but in fact it doesn't. Please help, I'm stumped.
Any error messages from "dmesg"?
/Per Jessen, Zürich
I picked this up from the failure this morning in messages, the NETDEV WATCHDOG: part didn't occur when eth0 started working again. Jun 8 09:09:38 Arbuthnot ifup: eth0 Jun 8 09:09:38 Arbuthnot ntpd[3599]: Listening on interface #3 eth0, fe80::21d:92ff:fe82:5c1#123 Enabled Jun 8 09:09:38 Arbuthnot ntpd[3599]: Listening on interface #6 eth0, 10.0.0.54#123 Enabled Jun 8 09:09:50 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:09:51 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:10:56 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:10:57 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:12:14 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:12:15 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:13:20 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:13:21 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:14:32 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:14:33 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:15:44 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:15:45 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:16:56 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:16:57 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:18:08 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out Jun 8 09:18:08 Arbuthnot ifdown: eth0 device: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev ff) Jun 8 09:18:09 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:19:32 Arbuthnot ntpd[2423]: Listening on interface #3 eth0, fe80::21d:92ff:fe82:5c1#123 Enabled Jun 8 09:19:32 Arbuthnot ntpd[2423]: Listening on interface #6 eth0, 10.0.0.54#123 Enabled Jun 8 09:19:37 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:19:37 Arbuthnot kernel: r8169: eth0: link up Jun 8 09:19:41 Arbuthnot kernel: eth0: no IPv6 routers present Jun 8 09:19:50 Arbuthnot ifup: eth0 device: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 02) Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
I picked this up from the failure this morning in messages, the NETDEV WATCHDOG: part didn't occur when eth0 started working again. Jun 8 09:09:38 Arbuthnot ifup: eth0 Jun 8 09:09:38 Arbuthnot ntpd[3599]: Listening on interface #3 eth0, fe80::21d:92ff:fe82:5c1#123 Enabled Jun 8 09:09:38 Arbuthnot ntpd[3599]: Listening on interface #6 eth0, 10.0.0.54#123 Enabled Jun 8 09:09:50 Arbuthnot kernel: NETDEV WATCHDOG: eth0: transmit timed out
Any chance this might be a cable problem? Have you tried another one? /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Any chance this might be a cable problem? Have you tried another one?
I've tried changing cables, sometimes booting with the cable out fixes it after an ifup. The only thing that fixes it consistently, is a kernel reinstall. I don't know if this is connected but I filed a bug ( Bug 393461) a while ago about yast2-qt update that stopped the runlevel editor from working. The bug eventually went but returned this morning. When I reinstalled the kernel to start my network again it also fixed that problem. What would the kernel reinstall do to make eth0 work again. I've tried depmod and SuSEconfig. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
Per Jessen wrote:
Any chance this might be a cable problem? Have you tried another one?
I've tried changing cables, sometimes booting with the cable out fixes it after an ifup. The only thing that fixes it consistently, is a kernel reinstall.
OK, so what happens when you re-install the kernel - during the reboot, the network driver module is reloaded. When you see the problem, try stopping the network device (ifdown eth0), then unload and reload the associated module (r8169 or whatever it is). Then start the network device again (ifup eth0).
What would the kernel reinstall do to make eth0 work again.
More importantly, what makes it stop working? /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Dave Plater wrote:
Per Jessen wrote:
Any chance this might be a cable problem? Have you tried another one?
I've tried changing cables, sometimes booting with the cable out fixes it after an ifup. The only thing that fixes it consistently, is a kernel reinstall.
OK, so what happens when you re-install the kernel - during the reboot, the network driver module is reloaded. When you see the problem, try stopping the network device (ifdown eth0), then unload and reload the associated module (r8169 or whatever it is). Then start the network device again (ifup eth0).
What would the kernel reinstall do to make eth0 work again.
More importantly, what makes it stop working?
/Per Jessen, Zürich
I tried ifdown unload module reload module then ifup and it didn't work. Then I compiled the driver into the kernel and it seemed to fix it and no more problems for a few weeks. I upgraded the kernel last week, without recompiling it and I had no problem until the power fail. That's when I discovered reinstalling the kernel consistently fixes the problem. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
I tried ifdown unload module reload module then ifup and it didn't work. Then I compiled the driver into the kernel and it seemed to fix it and no more problems for a few weeks. I upgraded the kernel last week, without recompiling it and I had no problem until the power fail. That's when I discovered reinstalling the kernel consistently fixes the problem.
OK, probably totally no connection but just a thought, Yesterday I had a power failure while I was out. The server stayed up, because of its UPS, but a thin client went down and would not restart. PXE complained about a cable being disconnected, even though the status LEDs were on. Changed cables, ports on switches, nothing. Then thought the onboard card must be US, tried to disable it from the BIOS but there was no setting showing. Thought I might as well try clearing the BIOS by removing the battery and trying again later. All back to normal. As I said just a thought. DC -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Cotton wrote:
Dave Plater wrote:
I tried ifdown unload module reload module then ifup and it didn't work. Then I compiled the driver into the kernel and it seemed to fix it and no more problems for a few weeks. I upgraded the kernel last week, without recompiling it and I had no problem until the power fail. That's when I discovered reinstalling the kernel consistently fixes the problem.
OK, probably totally no connection but just a thought, Yesterday I had a power failure while I was out. The server stayed up, because of its UPS, but a thin client went down and would not restart. PXE complained about a cable being disconnected, even though the status LEDs were on. Changed cables, ports on switches, nothing. Then thought the onboard card must be US, tried to disable it from the BIOS but there was no setting showing. Thought I might as well try clearing the BIOS by removing the battery and trying again later.
All back to normal. As I said just a thought.
DC
Mines worse, it thinks theres a link even when the cables disconnected. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
I tried ifdown unload module reload module then ifup and it didn't work. Then I compiled the driver into the kernel and it seemed to fix it and no more problems for a few weeks. I upgraded the kernel last week, without recompiling it and I had no problem until the power fail. That's when I discovered reinstalling the kernel consistently fixes the problem. Regards Dave P
Well, to be honst, there isn't much to go on then - the problem is caused by a power failure and fixed by re-installing the kernel. Both involve a reboot, from which the system comes up in two different states. If that's all there is, I think you'll have to look at why the two reboots are different. The power failure causes a cold-boot - how do you reboot after you've re-installed the kernel? Try resetting the box or even removing power. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Well, to be honst, there isn't much to go on then - the problem is caused by a power failure and fixed by re-installing the kernel. Both involve a reboot, from which the system comes up in two different states. If that's all there is, I think you'll have to look at why the two reboots are different. The power failure causes a cold-boot - how do you reboot after you've re-installed the kernel? Try resetting the box or even removing power.
/Per Jessen, Zürich
The power failure simply made an old problem return. Anyway I tried a few more things and couldn't get the network up so I recompiled the driver into the kernel, went out for a beer and came back finished the install_modules and install. So until the next kernel upgrade hopefully it will give no more problems. By the way I proved that reinstalling the kernel doesn't fix it. Its as if starting my eth0 is like starting a car with bad spark plugs on a cold morning. If you're lucky it starts. This is the first time I have had eth0 up since the last message. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
I googled NETDEV WATCHDOG and found a thread on kerneltrap.org with the same driver and the problem seemed to be as slippery as mine. I think there is a driver issue that no one has actually put a finger on yet anyway I don't have the problem after compiling the driver into the kernel so it must be a driver loading problem in my case. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen wrote:
Dave Plater wrote:
I tried ifdown unload module reload module then ifup and it didn't work. Then I compiled the driver into the kernel and it seemed to fix it and no more problems for a few weeks. I upgraded the kernel last week, without recompiling it and I had no problem until the power fail. That's when I discovered reinstalling the kernel consistently fixes the problem. Regards Dave P
Well, to be honst, there isn't much to go on then - the problem is caused by a power failure and fixed by re-installing the kernel. Both involve a reboot, from which the system comes up in two different states. If that's all there is, I think you'll have to look at why the two reboots are different. The power failure causes a cold-boot - how do you reboot after you've re-installed the kernel? Try resetting the box or even removing power.
/Per Jessen, Zürich
I think Dave needs a new network card.... At least for testing. -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David C. Rankin wrote:
Per Jessen wrote:
Well, to be honst, there isn't much to go on then - the problem is caused by a power failure and fixed by re-installing the kernel. Both involve a reboot, from which the system comes up in two different states. If that's all there is, I think you'll have to look at why the two reboots are different. The power failure causes a cold-boot - how do you reboot after you've re-installed the kernel? Try resetting the box or even removing power.
/Per Jessen, Zürich
I think Dave needs a new network card.... At least for testing.
I have had this bug for a few months before emailing the list. The only cure I have is to unload the driver and compile the driver into the kernel. This makes it rock solid. There must be a problem when the driver loads, maybe timing. Anyway my nic works every time now. I'll trouble shoot a bit more next kernel update I do. Regards Dave P -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Dave Plater wrote:
Hi, ever since I moved my system to an msi 945 based motherboard I have had an irritating intermittent problem with my Ethernet. It works fine, including after power off power on most of the time until it decides not to. I power on my pc and suddenly eth0, although showing up as working does not connect to the router (have tried restarting router etc & there's a vista machine with same giglan type that does not have problem). I've tried various things like starting up with lan disabled in the bios and then restarting with it enabled again, changing cables etc etc. A few weeks ago I recompiled the kernel with an internal lan driver and I had no more problems. I upgraded my kernel and all worked fine until a power fail brought the problem back again. I reinstalled kernel and eth0 worked again for two power cycles then this morning same problem, which was easy to fix, I just reinstalled the kernel and rebooted. The key symptom of failure is when ethtool shows Speed: 100Mb/s it will connect properly but if it shows Speed: 1000Mb/s the system just sits trying to connect to the router but the leds don't indicate any network activity. I've tried to set the speed to 100 but that does not happen, in fact whether eth0 connects or not ethtool cannot change any settings on lan anyway. It seems to me like the driver thinks it is working with eth0 but in fact it doesn't. Please help, I'm stumped.
Chip is RTL8111. hwinfo --netcard 08: PCI 300.0: 0200 Ethernet controller [Created at pci.310] UDI: /org/freedesktop/Hal/devices/pci_10ec_8168 Unique ID: rBUF.tMmV8WnWhuE Parent ID: qTvu.4chw4Z9dkyF SysFS ID: /devices/pci0000:00/0000:00:1c.1/0000:03:00.0 SysFS BusID: 0000:03:00.0 Hardware Class: network Model: "Realtek RTL8111/8168B PCI Express Gigabit Ethernet controller" Vendor: pci 0x10ec "Realtek Semiconductor Co., Ltd." Device: pci 0x8168 "RTL8111/8168B PCI Express Gigabit Ethernet controller" SubVendor: pci 0x1462 "Micro-Star International Co., Ltd." SubDevice: pci 0x507c Revision: 0x02 Driver: "r8169" Driver Modules: "r8169" Device File: eth0 I/O Ports: 0xd800-0xd8ff (rw) Memory Range: 0xefeff000-0xefefffff (rw,prefetchable) Memory Range: 0xefee0000-0xefeeffff (rw,prefetchable) Memory Range: 0xcfcf0000-0xcfcfffff (ro,prefetchable,disabled) IRQ: 220 (11764 events) HW Address: 00:1d:92:82:05:c1 Link detected: yes Module Alias: "pci:v000010ECd00008168sv00001462sd0000507Cbc02sc00i00" Driver Info #0: Driver Status: r8169 is active Driver Activation Cmd: "modprobe r8169" Config Status: cfg=no, avail=yes, need=no, active=unknown Attached to: #19 (PCI bridge)
Thanks in advance. Dave P
Hmm.... Sounds familiar. See: [opensuse] kernel: NETDEV WATCHDOG: eth0: transmit timed out? (Date: Sun, 21 Oct 2007 23:55:20 -0500). Your not going to like the answer. (The thread references a SuSE 10.0 server) Answer: restarting the network cured the problem temporarily, and the problem went away a week or so later.. I wrote it off to gremlins.. The same box is still running without problems. -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (5)
-
Dave Cotton -
Dave Plater -
Dave Plater -
David C. Rankin -
Per Jessen