At 08:31 10/13/2008 -0700, Ronciak, John wrote: >David, > >First thanks for the report. It looks like you have a number of >different things going on. > >First is that the only interface reporting that it has link up >is eth6 which is one of the igb ports. All of the rest show "no >link" including the Tigon3 NIC. Not getting link is most likely >something to do with the switch or cabling. Have you tried >plugging any of the ports in back to back to another system? >Link errors almost never are driver related.
Only 'eth6' and 'eth7' are presently cabled. The staff at the data center misunderstood instructions. However I see no reason why other ports not being connected should make any difference. The two ports are connected to a 48-port line card in a Cisco 6509-E switch. A second system with 82571EB ports (e1000e) is racked right next to this one and has no problems. Let me emphasize that the 82575s are working perfectly fine when Fedora with the latest Fedora kernel are in use. You do not appear to have read the original message completely. >Second, is that all of the interfaces (at least the Intel ones) >are reporting having to use legacy interrupts. So something is >wrong with how you are compiling your kernels or your system is >configured in the BIOS to do this. Since this is showing up on >multiple NICs this is most likely a system or kernel thing. It only works with the pure Fedora kernel, which we did not compile. Note that it fails to work with the RHEL 5 kernel, which we also did not compile. A noticeable difference is that the RHEL 5 kernel configures the 82575's with MSI-X interrupts. As you see if you read the original message, we have compiled quite a sample of Kernel.org kernels and tried them. In every case the instructions that came with the driver were carefully observed. The 'igb' driver was built both with and without DCA support--this makes no difference. You have assumed we did not work this carefully. Let me be absolutely clear that we have. 'e1000e' drivers compiled by us on a similar system work perfectly. >Third, is you x16 port(s) are most likely graphics port(s) that >may or may not be able to be used for things other than a >graphics card. This is system dependent. Please look to the >BIOS to see if the port configuration can be changed. Most do >allow this but you need to change the settings. This is a HP DL160 G5, which is a 1U rack mount server. The two x16 ports are intended for network and storage devices, not for graphics. Nobody puts fancy GPUs in 1U rack systems unless they are to be used as numeric co-processors. The BIOS has no setting other than a PCI 1 versus PCI 2 indication. It has been tried both ways--makes no difference. >Forth, I'm a bit confused as to why you are trying to use an >ioatdma driver that didn't come with the kernel. What ioatdma >driver are you trying to use? The only ones (ioatdma and dca) >that should be used are the ones that come with the particular >kernel. Are you saying that these don't compile either? Please >give us more information about how this is failing including the >config files for F9 that you are using. This works for us >without us doing anything to enable it. Everything compiles fine except for 'ioatdma.ko' under 2.6.26 and 2.6.27 kernels. It's obvious that the internal APIs have changed, since the error is for a structure member 'ack' that no longer exists. We've tried every conceivable combination (in the true mathematical sense) of hand-compiled versus Fedora 9 and RHEL 5.2 distribution drivers. Nothing works except the pure Fedora 9 drivers, the one case where MSI interrupts are disabled. It is clear that the driver is immature and has flaws. This Bugzilla reinforces the perception: https://bugzilla.redhat.com/show_bug.cgi?id=452289 David > >Cheers, >John >----------------------------------------------------------- >"Those who would give up essential Liberty, to purchase a little >temporary Safety, deserve neither Liberty nor Safety.", Benjamin >Franklin 1755 > > >>-----Original Message----- >>From: David Lawless [mailto:[EMAIL PROTECTED] >>Sent: Sunday, October 12, 2008 8:49 PM >>To: e1000-devel@lists.sourceforge.net >>Subject: [E1000-devel] IGB driver 1.2.44.9 problem >> >>We're trying to evaluate an Intel Pro VT Quad and are having >>major problems coaxing it to work on an HP DL160 G5 (Seaburg >>5400 chipset). RAM is 16GB and cpus are two E5430s. BIOS >>is latest 8/22/08 revision. >> >>Have tried many combinations: >> RHEL 5.2 (2.6.18-92.1.13.el5) >> Fedora 9 w/ Kernel.org (2.6.24.7) >> Fedora 9 w/ Kernel.org (2.6.25.18) >> Fedora 9 w/ Kernel.org (2.6.27) >> Fedora 9 (2.6.26.5-45.fc9.x86_64) >>With both the native device drivers and the Intel site >>device drivers. >> >>In most cases the 'igb' devices seem to come up, but 'ethtool' >shows >> - >> Speed: 1000Mb/s >> Duplex: Full >> - >> Link detected: no >> - >>and they do not function. >> >>It does work under Fedora 9 (2.6.26.5-45.fc9.x86_64), but in >>that case MSI interrupts are completely disabled and it falls >>back to PCI-APIC interrupts. MSI-X interrupts are indicated >for >>RHEL 5.2 and the Kernel.org kernel, though as I said it does >not >>actually work. Couldn't successfully compile 'ioatdma.ko' for >>2.6.26 and 2.6.27 kernels. Under those versions the results >are >>the same when using the native 'ioatdma.ko' and 'dca.ko'. >> >>In addition to the Quad-VT card, we have an Intel 10GbE dual >CX4 >>card in the other slot. In every case we're seeing the 'ixgbe' > >>driver claim that the bus width is x1 even though the slot is >>x16. >> >>Any help would be appreciated. >> >>Thanks, >> >>David >> ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ E1000-devel mailing list E1000-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/e1000-devel
