Most recent kernel where this bug did not occur: Distribution: Gentoo Steps to reproduce: After going to standby this card will start dropping packets intermittantly. Then it will go back to normal. Cycle. The obvious workaround for the time being is to write a script to remove the module before suspending and modprobe it afterwards. I'm not sure when this bug began as it has been here since I've had the card, but it definitely exists everywhere in between v2.6.15-v2.6.16-rc4. hdparm -vvv: 01:06.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8169 Gigabit Ethernet (rev 10) Subsystem: Realtek Semiconductor Co., Ltd. RTL-8169 Gigabit Ethernet Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- Latency: 32 (8000ns min, 16000ns max), Cache Line Size 10 Interrupt: pin A routed to IRQ 20 Region 0: I/O ports at c000 [size=256] Region 1: Memory at e7000000 (32-bit, non-prefetchable) [size=256] [virtual] Expansion ROM at 30000000 [disabled] [size=128K] Capabilities: [dc] Power Management version 2 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0-,D1+,D2+,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=0 DScale=0 PME- I will attach the config.
Created attachment 7470 [details] kernel-config
Can you try 2.6.16-rc5 ? A few r8169 patches went in. If it does not work better, please send: - complete dmesg from boot; - lspci -vx; - compiler revision (should appear in the very first lines of dmesg); - lsmod; - /proc/interrupts after resume; - ifconfig after resume. -- Ueimor
It appears to be fixed! I didn't test to see if this was reproducable but I will reopen if it comes back up. Thanks for the quick response.
Very sorry it still exists :( After further testing the problem has popped back up.
Created attachment 7499 [details] dmesg output
Created attachment 7500 [details] Requested ifconfig
Created attachment 7501 [details] Requested interrupts
Created attachment 7502 [details] Requested lsmod
Created attachment 7503 [details] Requested lspci-vx
Is it still reproducible without NETCONSOLE and PREEMPT ? -- Ueimor
Recompiled with 'no preemption', loaded the netconsole module the first time, this time was a failure. Rebooted without the netconsole module and this time it appears to be a sucess.
I just had a crash on this computer and wasn't able to get a panic, due to netconsole begin off, due to this bug. Anything I can do to help further?
Olrich: [...] > Anything I can do to help further? Does passing 'pci=routeirq' as a command line option when the kernel boots make a difference ? -- Ueimor
Francois: [...] > Does passing 'pci=routeirq' as a command line option when the kernel boots make > a difference ? Btw, please add an updated lspci -vvv to verify that PME is set. -- Ueimor
Please reopen this bug if: - it is still present in kernel 2.6.19-rc5 and - you can provide the requested information.