Arch Linux

steve_v

Member

Registered: 2006-02-11

Posts: 80

Dodgy cable? Atheros gigabit controller locks up.

While attempting to transfer a large (~20GiB) file my onboard atheros NIC locks up completely:

[Tue Feb 28 18:05:11 2012] ATL1E 0000:02:00.0: eth0: rx sequence number error (rx=30633) (expect=30218)
[Tue Feb 28 18:06:31 2012] ATL1E 0000:02:00.0: PCI INT A disabled
[Tue Feb 28 18:06:36 2012] ATL1E 0000:02:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
[Tue Feb 28 18:06:36 2012] ATL1E 0000:02:00.0: setting latency timer to 64
[Tue Feb 28 18:06:44 2012] ATL1E 0000:02:00.0: irq 48 for MSI/MSI-X
[Tue Feb 28 18:06:44 2012] ADDRCONF(NETDEV_UP): eth0: link is not ready

Removing & reloading the module appears to work (no errors & nothing unusual in dmesg output), but all I get is "link is not ready" until I reboot the machine.

I suspect this is caused by a dodgy cat5 cable on the other side of the switch, however my PCIE Realtek card does not have a problem. In the same test the transfer completes succesfully with no errors at all.

02:00.0 Ethernet controller: Atheros Communications Inc. AR8121/AR8113/AR8114 Gigabit or Fast Ethernet (rev b0)          
        Subsystem: ASUSTeK Computer Inc. Device 8226                                                                     
        Flags: bus master, fast devsel, latency 0, IRQ 17                                                                
        Memory at fe8c0000 (64-bit, non-prefetchable) [size=256K]                                                        
        I/O ports at bc00 [size=128]                                                                                     
        Capabilities: <access denied>                                                                                    
        Kernel driver in use: ATL1E

04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01)                                                                                                                       
        Subsystem: Realtek Semiconductor Co., Ltd. TEG-ECTX Gigabit PCI-E Adapter [Trendnet]                             
        Flags: bus master, fast devsel, latency 0, IRQ 46                                                                
        I/O ports at d800 [size=256]                                                                                     
        Memory at feaff000 (64-bit, non-prefetchable) [size=4K]                                                          
        Expansion ROM at feac0000 [disabled] [size=128K]                                                                 
        Capabilities: <access denied>                                                                                    
        Kernel driver in use: r8169    

If this is a cable problem I would expect media errors, but i would also expect a reset / retry rather than a dead NIC until reboot.
Has anyone else seen this behaviour?

Cheers,

Steve.