E100 Wake On Lan fails for kernels > 3.5.7 There is now no stable kernel for < sys-kernel/gentoo-sources-3.8.13. Hardened sources (sys-kernel/hardened-sources) is the only kernel with multiple stable versions. One would expect sys-kernel/vanilla-sources to also have several versions for working around regressions and bugs, and there area no stable < sys-kernel/vanilla-sources-3.7.10! See Kernel Bug Filed: https://bugzilla.kernel.org/show_bug.cgi?id=56181 Reproducible: Always Expected Results: Should be multiple versions available for working around bug regressions for users! Likely, sys-kernel/gentoo-sources should keep a series of versions of kernel source available, such as "=sys-kernel/gentoo-sources-3.0, =sys-kernel/gentoo-sources-3.2, =sys-kernel/gentoo-sources-3.4, ..." And then when moving forward, exponentially keep only varying versions such as, "=sys-kernel/gentoo-sources-3.0, =sys-kernel/gentoo-sources-3.4, =sys-kernel/gentoo-sources-3.8, ..."
The e100 module is for Intel PCI Network cards. Likely somebody messed with the e100 driver source code breaking the driver sometime immediately kernel-3.5.7. I troubleshooted and filed the bug a couple of months ago and have seen no activity since. So I expect to be running an older kernel for that particular x86 platform from now on. Regressions are quite common, especially for older hardware as more people move to newer hardware. I'm also seeing other distributions such as Fedora pushing newer kernels as stable lately. Seems like a trend to push the latest and greatest. But deleting or masking all older stable versions probably isn't a great idea for legacy hardware! Even if there is a security vulnerability, submit a patch and keep at least a few older stable kernels for legacy hardware.
SOme information on your ethernet card would help. lspci -vvv
# lspci -vv 00:10.0 Ethernet controller: Intel Corporation 82557/8/9/0/1 Ethernet Pro 100 (rev 08) Subsystem: Intel Corporation EtherExpress PRO/100+ Management Adapter Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx- Latency: 64 (2000ns min, 14000ns max), Cache Line Size: 32 bytes Interrupt: pin A routed to IRQ 16 Region 0: Memory at febfe000 (32-bit, non-prefetchable) [size=4K] Region 1: I/O ports at ef00 [size=64] Region 2: Memory at fea00000 (32-bit, non-prefetchable) [size=1M] Expansion ROM at fe900000 [disabled] [size=1M] Capabilities: [dc] Power Management version 2 Flags: PMEClk- DSI+ D1+ D2+ AuxCurrent=0mA PME(D0+,D1+,D2+,D3hot+,D3cold+) Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=2 PME- Kernel driver in use: e100 Completely forgot, however likely irrelevant as it's a known issue they tend to break WOL with kernel patches. For this specific instance and after grepping through the patches after 3.5.7, sure looks like one of the e100 patches disabled WOL!
Can you test with kernel vanilla kernel 3.10.0?
Has there been any specific change with Wake On Lan functions upstream, or are you just suggesting that a newer version might have changed something? Note, this is an upstream bug and has been filed here and showing no activity since being filed. (Likely, everybody has had the money to upgrade to amd64 hardware and tossed out all their legacy x86 hardware. :-/) https://bugzilla.kernel.org/show_bug.cgi?id=56181 Also note, my box with that Intel PCI E100 device is temporarily down, although I do also have a laptop with an Intel E100 network device. Since Wake On Lan worked in kernel-3.5.7, I'll rule-out any possible hardware issues as it worked up until a certain kernel version (easily reproduced) and rely on the fact also this Wake On Lan function code is routinely broken every year or two. A tested an x86 laptop with this e100 device, but it doesn't display the broken Wake On Lan functions as it seems this bug is chipset version dependant, or particular to only certain e100 devices. (e100 driver/module can work on a number of different Intel network cards including 32 or 64 bit platforms, of which laptop versions are well modified and have different implementations, apart from also being a different chipset! Not only this, but low level routines are more obviously sensitive to minute or minimal hardware differences.) I won't have time to fix my other legacy box displaying this problem for several months (likely until October or November or a little later), as I need to buy several new minimal costing parts, all a while I have far more important things to complete here with my house in the Arctic preparing for winter. If I had much more time, I would have read and fixed the C code myself, as there are only at most two or three patches to the code to eliminate or target the problem code along with a name and email address of the person whom broke the feature. ;-)
Mike, I see you're the lead kernel maintainer Gentoo, so I'll safely presume you understand all the fore mentioned items I mentioned, and can see I'm somewhat competent here. ;-) I think a version of Ubuntu/Mint also marked kernel-3.5.7 as an old stable version as well, but this info is from a recent news article. (I'll set a reminder on this bug as one of my winter tasks.)
I have set two email reminders, October 1 and November 1, 2013 roughly about the time I'll have more time to allot to deal with this. Maybe I'll be have some sporadic time to acquire some of the required parts from Digikey. As far as labor, I usually have to perform most tasks as most people are not competent these days. As such, it will be a task requiring room amidst major drywalling and other major house projects. Not a place for sensitive computer equipment to be strewn about. :-/ )
The reason I asked was to take it upstream if it wasn't fixed. I see now you already have an upstream bug. We can revisit in oct/nov when you have more time.
We are tracking upstream, please post more details there instead of here; thank you for reporting and good luck, let's hope they can get this fixed.
Seems like my laptop based on a later Intel mainboard chipset can now use linux-3.12.13-gentoo with wakeonlan feature. However an older 440BX Intel mainboard chipset using an e100 network card still has no wakeonlan fuction with the later kernels. Still, only linux-3.5.7-gentoo is usable (or stable) with wakeonlan and e100.
