Initial state: Linux nerff 2.6.17-gentoo-r4 #1 PREEMPT Sun Jul 23 13:14:18 NOVST 2006 i686 AMD Athlon(tm) 64 Processor 3200+ GNU/Linux gcc version 4.1.1 (Gentoo 4.1.1) nerff ~ # modinfo sky2 filename: /lib/modules/2.6.17-gentoo-r4/kernel/drivers/net/sky2.ko description: Marvell Yukon 2 Gigabit Ethernet driver author: Stephen Hemminger <shemminger@osdl.org> license: GPL version: 1.5 vermagic: 2.6.17-gentoo-r4 preempt mod_unload K8 REGPARM gcc-4.1 depends: alias: pci:v00001148d00009000sv*sd*bc*sc*i* alias: pci:v00001148d00009E00sv*sd*bc*sc*i* alias: pci:v00001186d00004B00sv*sd*bc*sc*i* alias: pci:v000011ABd00004340sv*sd*bc*sc*i* alias: pci:v000011ABd00004341sv*sd*bc*sc*i* alias: pci:v000011ABd00004342sv*sd*bc*sc*i* alias: pci:v000011ABd00004343sv*sd*bc*sc*i* alias: pci:v000011ABd00004344sv*sd*bc*sc*i* alias: pci:v000011ABd00004345sv*sd*bc*sc*i* alias: pci:v000011ABd00004346sv*sd*bc*sc*i* alias: pci:v000011ABd00004347sv*sd*bc*sc*i* alias: pci:v000011ABd00004350sv*sd*bc*sc*i* alias: pci:v000011ABd00004351sv*sd*bc*sc*i* alias: pci:v000011ABd00004352sv*sd*bc*sc*i* alias: pci:v000011ABd00004360sv*sd*bc*sc*i* alias: pci:v000011ABd00004361sv*sd*bc*sc*i* alias: pci:v000011ABd00004362sv*sd*bc*sc*i* alias: pci:v000011ABd00004363sv*sd*bc*sc*i* alias: pci:v000011ABd00004364sv*sd*bc*sc*i* srcversion: 4692B864E577184E726261F parm: idle_timeout:Idle timeout workaround for lost interrupts (ms) (int) parm: disable_msi:Disable Message Signaled Interrupt (MSI) (int) parm: copybreak:Receive copy threshold (int) parm: debug:Debug level (0=none,...,16=all) (int) 03:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8053 PCI-E Gigabit Ethernet Controller (rev 15) Subsystem: Micro-Star International Co., Ltd. Marvell 88E8053 Gigabit Ethernet Controller (MSI) Flags: bus master, fast devsel, latency 0, IRQ 90 Memory at fdbfc000 (64-bit, non-prefetchable) [size=16K] I/O ports at 9c00 [size=256] [virtual] Expansion ROM at fda00000 [disabled] [size=128K] Capabilities: [48] Power Management version 2 Capabilities: [50] Vital Product Data Capabilities: [5c] Message Signalled Interrupts: 64bit+ Queue=0/1 Enable+ Capabilities: [e0] Express Legacy Endpoint IRQ 0 Capabilities: [100] Advanced Error Reporting nerff ~ # cat /proc/interrupts CPU0 0: 34812214 IO-APIC-edge timer 1: 60 IO-APIC-edge i8042 7: 1 IO-APIC-edge parport0 8: 2 IO-APIC-edge rtc 9: 0 IO-APIC-level acpi 15: 323689 IO-APIC-edge ide1 50: 2772154 IO-APIC-level nvidia 58: 52572 IO-APIC-level au8830 66: 3594728 IO-APIC-level eth0 74: 2 IO-APIC-level ohci1394 90: 2832261 PCI-MSI sky2 217: 259375 IO-APIC-level libata 225: 2244 IO-APIC-level ehci_hcd:usb1 233: 130450 IO-APIC-level ohci_hcd:usb2 NMI: 0 LOC: 34813251 ERR: 0 MIS: 0 Problem description: /*Probably this bugreport is duplicate*/ I'm using onboard Marvell 88E8053 Gigabit Ethernet Controller (MSI) with sky2. Sometimes, network is going down and I need to run "rmmod sky2 && modprobe sky2" to continue working. Steps to reproduce: Reproduceable: sometimes. Network is going down when I try to copy about 1Gb from my Linux box over network.I can reproduce it on large files (e.g. movie files) and directories with small files. Problem was discovered when network interface just stop working on files copying from SAMBA(net-fs/samba-3.0.22-r3). I can try to reproduce it with another initial conditions if it will be nesessary. (Sorry for my poor english)
Created attachment 93207 [details] My emerge --info output
*** This bug has been marked as a duplicate of 137504 ***
Reopening to keep one sky2 bug per user
Sorry, I can't reproduce this bug on my home box right now. I run copying from Linux box to Windows box directory with movies (about 72G) and copying do not stop. I break this process about 50% of all files. But - i found some strange thing: test interface is eth1, and TX bytes value goes back from about 3G tranferred files! Please look at my ifconfig output nerff ~ # ifconfig eth0 Link encap:Ethernet HWaddr 00:13:D3:9A:38:9C inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:10340 errors:0 dropped:0 overruns:0 frame:0 TX packets:10679 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:7628579 (7.2 Mb) TX bytes:1785652 (1.7 Mb) Interrupt:10 Base address:0xe000 eth1 Link encap:Ethernet HWaddr 00:13:D3:9A:33:69 inet addr:192.168.200.1 Bcast:192.168.200.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:7180558 errors:0 dropped:0 overruns:0 frame:0 TX packets:13733112 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:494644241 (471.7 Mb) TX bytes:3151262129 (3005.2 Mb) Interrupt:10 ^^^^^^^^^^ lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:13622 errors:0 dropped:0 overruns:0 frame:0 TX packets:13622 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:6241189 (5.9 Mb) TX bytes:6241189 (5.9 Mb) nerff ~ # ifconfig eth0 Link encap:Ethernet HWaddr 00:13:D3:9A:38:9C inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:10353 errors:0 dropped:0 overruns:0 frame:0 TX packets:10689 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:7629619 (7.2 Mb) TX bytes:1787162 (1.7 Mb) Interrupt:10 Base address:0xe000 eth1 Link encap:Ethernet HWaddr 00:13:D3:9A:33:69 inet addr:192.168.200.1 Bcast:192.168.200.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:8339258 errors:0 dropped:0 overruns:0 frame:0 TX packets:15951136 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:574303510 (547.6 Mb) TX bytes:2141482572 (2042.2 Mb) Interrupt:10 ^^^^^^^^^^^ lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:13622 errors:0 dropped:0 overruns:0 frame:0 TX packets:13622 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:6241189 (5.9 Mb) TX bytes:6241189 (5.9 Mb) nerff ~ # ifconfig eth0 Link encap:Ethernet HWaddr 00:13:D3:9A:38:9C inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:10368 errors:0 dropped:0 overruns:0 frame:0 TX packets:10703 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:7630857 (7.2 Mb) TX bytes:1789480 (1.7 Mb) Interrupt:10 Base address:0xe000 eth1 Link encap:Ethernet HWaddr 00:13:D3:9A:33:69 inet addr:192.168.200.1 Bcast:192.168.200.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:9730557 errors:0 dropped:0 overruns:0 frame:0 TX packets:18611718 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:669921266 (638.8 Mb) TX bytes:1787593359 (1704.7 Mb) Interrupt:10 ^^^^^^^^^^ lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:13664 errors:0 dropped:0 overruns:0 frame:0 TX packets:13664 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:6247991 (5.9 Mb) TX bytes:6247991 (5.9 Mb)
How soon were those ifconfigs invoked after each other? It is normal for the TX bytes counter to wrap around to 0 after it reaches a certain value. It's hard to say if that is what is happening with no idea of timing.
(In reply to comment #5) > How soon were those ifconfigs invoked after each other? > The time interval was... about 10 - 30 seconds. I think I know how to reproduce this bug and I will run such test today. Initial conditions should include more than 2-3 copy sessions.
Any news? gentoo-sources-2.6.17-r6 includes a new sky2 which may fix any problems you are still seeing.
Thanks, I try to reproduce bug on gentoo-sources-2.6.17-r6 today on my home Linux box. Problem with gentoo-sources-2.6.17-r4 still reproduceable (sometimes). To reproduce: Try to copy data by portions (1-3G). At 3th or 4th session driver just stop all traffic on interface. I can't reproduce this problem if I try to copy files in single session.
looks like fixed sky2 gentoo-sources-2.6.17-r6 nerff ~ # modinfo sky2 filename: /lib/modules/2.6.17-gentoo-r6/kernel/drivers/net/sky2.ko description: Marvell Yukon 2 Gigabit Ethernet driver author: Stephen Hemminger <shemminger@osdl.org> license: GPL version: 1.6 vermagic: 2.6.17-gentoo-r6 preempt mod_unload K8 REGPARM gcc-4.1 WFM using sky2 1.6 driver. Network speed about 10000 Kb/s (+|- 350). All works just fine, thanks. P.S. I can't run valid load-testing, sorry. It's home box, not server.
OK - thanks for letting us know.
reopen I had run to backup all windows box on samba share (about 40G). While data restoring from winbox, networking is unusable again. Sorry for poor english :(
Please reproduce with 2.6.17-r8
(In reply to comment #12) > Please reproduce with 2.6.17-r8 > Hello, exactly the same problem here. Kernel is 2.6.17-r8. The NIC hangs when under load. Disabling "CONFIG_X86_UP_APIC" in the kernel .config fixed the problem for me. Hope this is helpful. Ludek Stepan
Please retest with gentoo-sources-2.6.18-r1
Add this patch as well: http://marc.theaimsgroup.com/?l=linux-netdev&m=116138920228073&q=raw
Ok, I'll run some tests this week.
Please reopen if this is still a problem
sorry for my absence :( reopen for: filename: /lib/modules/2.6.19-gentoo-r2/kernel/drivers/net/sky2.ko description: Marvell Yukon 2 Gigabit Ethernet driver author: Stephen Hemminger <shemminger@osdl.org> license: GPL version: 1.10 vermagic: 2.6.19-gentoo-r2 mod_unload K8 REGPARM Reproduced on data copying from DVD-9 disk (files ~ 1G).
See also http://bugzilla.kernel.org/show_bug.cgi?id=6839 workaround (Comment #12) and work patch (Comment #55) here also reload sky2 module is workaround Why it still exists in stable hardened ??? Bug appear here too :( kernel: 2.6.18-hardened Ethernet controller: Marvell Technology Group Ltd. 88E8053 PCI-E Gigabit Ethernet Controller
Should be better in gentoo-sources-2.6.20-r2, please reopen if that is not the case
but what about hardened kernel ?
File a new bug with them if you have confirmed that the newer driver does solve your problem and they can then consider backporting it
sorry but I can't check it now we change ethernet card to intel at this server but when I try different version of hardened sources 2 months ago - problem exists I try 2.6.18-hardened and 2.6.19-hardened-r5
