Re: [E1000-devel] Problem with igb i210 and multiple NICs on one board

[Fujinaka, T. <tod...@in...>]

I finally found the datasheet for the blade you’re using. I would strongly suggest checking your dmesg very carefully and you’re very likely to have to use a newer kernel than you’re using now. In general, I suggest using a kernel at least as new as the hardware. Debian may have backported the necessary patches for your hardware, but that’s for Debian to determine.

Todd Fujinaka
Software Application Engineer
Networking Division (ND)
Intel Corporation
tod...@in...
(503) 712-4565

From: Todor Ivanov [mailto:ti...@xp...]
Sent: Thursday, October 13, 2016 7:50 AM
To: Fujinaka, Todd <tod...@in...>
Cc: e10...@li...
Subject: Re: [E1000-devel] Problem with igb i210 and multiple NICs on one board

   Thank you very much for the support and guidelines. I already contacted the vendor as you suggested.

Regards,
Todor

Best regards,
Todor Ivanov
Senior System Administrator

Xpeqt Ltd
2, Samokovsko Shose Blvd
Sofia 1138, Bulgaria
Mobile: +359 887 383 600

On Thu, Oct 13, 2016 at 5:47 PM, Fujinaka, Todd <tod...@in...<mailto:tod...@in...>> wrote:
igb-5.3.5.4 is the newest. 4.1.2 is from January of 2013. We also don’t support Debian directly, which is why I asked you to try a live DVD.

Multiple cards with the multiple outcomes points to hardware failure. Or if you mean multiple controllers on the motherboard, same thing.

You could file a bug on sourceforge and attach the full logs, but I’m still thinking this is a hardware failure.

Todd Fujinaka
Software Application Engineer
Networking Division (ND)
Intel Corporation
tod...@in...<mailto:tod...@in...>
(503) 712-4565

From: Todor Ivanov [mailto:ti...@xp...<mailto:ti...@xp...>]
Sent: Thursday, October 13, 2016 7:19 AM
To: Fujinaka, Todd <tod...@in...<mailto:tod...@in...>>
Cc: e10...@li...<mailto:e10...@li...>
Subject: Re: [E1000-devel] Problem with igb i210 and multiple NICs on one board

It is not a cabling issue, as I did tests with same cable on both ports and one of them is 100% working. We are using this setup with multiple boards.

Changing the kernel will be a heavy task, as we have this highly customized distribution to fit our specific production needs:  Debian Squeeze with kernel 2.6.32

According to the driver: This driver supports kernel versions 2.6.30 or newer.

Before requesting help from the vendor, I hoped a developer could quick check the e1000_i210 source to see if something could be related to the specific problem we are experiencing. As I said two identical ethernet cards on same board behave differently. From what I'm seeing the vendor Adlink is suggesting the same driver, but older version: igb-4.1.2, which is behaving in the same way as the newest 3.5.3.5.



Todor




On Thu, Oct 13, 2016 at 5:00 PM, Fujinaka, Todd <tod...@in...<mailto:tod...@in...>> wrote:
I would try a newer kernel first, maybe a live DVD. Not sure where you got yours. Also unsure of what OS you're running.

Sounds like a hardware issue as the ports are coming up but not finding link. I'd try new cables first and then contact Adlink.

Todd Fujinaka
Software Application Engineer
Networking Division (ND)
Intel Corporation
tod...@in...<mailto:tod...@in...>
(503) 712-4565

-----Original Message-----
From: Todor Ivanov [mailto:ti...@xp...<mailto:ti...@xp...>]
Sent: Thursday, October 13, 2016 3:35 AM
To: e10...@li...<mailto:e10...@li...>
Subject: [E1000-devel] Problem with igb i210 and multiple NICs on one board

       Hi, All,

       We bough a new integrated PC from Adlink ePCI-3510BLG, which has three physical i210 (8086:1533) interfaces with two ethernet RJ45 terminations.

We are facing issues getting the second network card to work correctly. The problem is that we see three network interfaces and only one of them works.
Our kernel is: Linux 2.6.32-5-686 i686 GNU/Linux

We tried drivers: igb-4.1.2 (Adlink website) (5.2.9.4/5.2.15/igb-5.3.5.3<http://5.2.9.4/5.2.15/igb-5.3.5.3> Intel website)

lspci | grep Ethernet

03:00.0 Ethernet controller: Intel Corporation Device 1533 (rev 03)
04:00.0 Ethernet controller: Intel Corporation Device 1533 (rev 03)
07:00.0 Ethernet controller: Intel Corporation Device 1533 (rev 03)

Drivers starts normally:

[ 4649.365204] Intel(R) Gigabit Ethernet Network Driver - version 4.1.2 [ 4649.369385] Copyright (c) 2007-2012 Intel Corporation.
[ 4649.373538] igb 0000:03:00.0: PCI INT A -> GSI 21 (level, low) -> IRQ 21 [ 4649.377788] igb 0000:03:00.0: setting latency timer to 64 [ 4649.378126] igb: 0000:03:00.0: igb_check_options: VMDq option is not supported.
[ 4649.382623] igb 0000:03:00.0: irq 31 for MSI/MSI-X [ 4649.382625] igb 0000:03:00.0: irq 32 for MSI/MSI-X [ 4649.438254] igb 0000:03:00.0: Intel(R) Gigabit Ethernet Network Connection [ 4649.442769] igb 0000:03:00.0: eth0: (PCIe:2.5GT/s:Width x1) [ 4649.443006] igb 0000:03:00.0: eth0: MAC: 00:30:64:16:33:35 [ 4649.451885] igb 0000:03:00.0: eth0: PBA No: 000400-000 [ 4649.460493] igb 0000:03:00.0: LRO is disabled [ 4649.464750] igb 0000:03:00.0: Using MSI-X interrupts. 1 rx queue(s), 1 tx queue(s) [ 4649.469207] igb 0000:04:00.0: PCI INT A -> GSI 22 (level, low) -> IRQ 22 [ 4649.473695] igb 0000:04:00.0: setting latency timer to 64 [ 4649.474037] igb: 0000:04:00.0: igb_check_options: VMDq option is not supported.
[ 4649.478624] igb 0000:04:00.0: irq 33 for MSI/MSI-X [ 4649.478625] igb 0000:04:00.0: irq 34 for MSI/MSI-X [ 4649.516885] igb 0000:04:00.0: Intel(R) Gigabit Ethernet Network Connection [ 4649.521582] igb 0000:04:00.0: eth1: (PCIe:2.5GT/s:Width x1) [ 4649.521744] igb 0000:04:00.0: eth1: MAC: 00:30:64:16:33:36 [ 4649.530902] igb 0000:04:00.0: eth1: PBA No: 000400-000 [ 4649.540338] igb 0000:04:00.0: LRO is disabled [ 4649.544804] igb 0000:04:00.0: Using MSI-X interrupts. 1 rx queue(s), 1 tx queue(s) [ 4649.549409] igb 0000:07:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18 [ 4649.554068] igb 0000:07:00.0: setting latency timer to 64 [ 4649.554439] igb: 0000:07:00.0: igb_check_options: VMDq option is not supported.
[ 4649.559170] igb 0000:07:00.0: irq 35 for MSI/MSI-X [ 4649.559171] igb 0000:07:00.0: irq 36 for MSI/MSI-X [ 4649.596770] igb 0000:07:00.0: Intel(R) Gigabit Ethernet Network Connection [ 4649.601623] igb 0000:07:00.0: eth2: (PCIe:2.5GT/s:Width x1) [ 4649.601786] igb 0000:07:00.0: eth2: MAC: 00:30:64:16:33:37 [ 4649.611346] igb 0000:07:00.0: eth2: PBA No: 000400-000 [ 4649.620185] igb 0000:07:00.0: LRO is disabled [ 4649.624866] igb 0000:07:00.0: Using MSI-X interrupts. 1 rx queue(s), 1 tx queue(s) [ 4651.172103] ADDRCONF(NETDEV_UP): eth0: link is not ready [ 4651.990442] ADDRCONF(NETDEV_UP): eth1: link is not ready [ 4652.825065] ADDRCONF(NETDEV_UP): eth2: link is not ready [ 4655.549405] igb: eth2 NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: None
[ 4655.554425] ADDRCONF(NETDEV_CHANGE): eth2: link becomes ready


But only one link is up (We have cables in both. Also tried them in a row)

eth0      Link encap:Ethernet  HWaddr 00:30:64:16:33:35
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

eth1      Link encap:Ethernet  HWaddr 00:30:64:16:33:36
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

eth2      Link encap:Ethernet  HWaddr 00:30:64:16:33:37
          inet addr:10.35.91.197  Bcast:10.35.95.255  Mask:255.255.240.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:47540 errors:0 dropped:0 overruns:0 frame:0
          TX packets:1490 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:4603113 (4.3 MiB)  TX bytes:164743 (160.8 KiB)


root@XPEQTSIM:/mnt/source/igb-4.1.2-fromAdlink#<mailto:root@XPEQTSIM:/mnt/source/igb-4.1.2-fromAdlink> ip link
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
14: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN qlen 1000
    link/ether 00:30:64:16:33:35 brd ff:ff:ff:ff:ff:ff
15: eth1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN qlen 1000
    link/ether 00:30:64:16:33:36 brd ff:ff:ff:ff:ff:ff
16: eth2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen
1000
    link/ether 00:30:64:16:33:37 brd ff:ff:ff:ff:ff:ff


The drivers would not compile unless we put the following code into /usr/src/linux-headers-2.6.32-5-common/include/linux/pm.h :

#ifdef CONFIG_PM_RUNTIME
#define SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) \
        .runtime_suspend = suspend_fn, \
        .runtime_resume = resume_fn, \
        .runtime_idle = idle_fn,
#else
#define SET_RUNTIME_PM_OPS(suspend_fn, resume_fn, idle_fn) #endif

I played with bios, but there are not network related options.
Do you know where the problem might be?

Best regards,
Todor Ivanov
Senior System Administrator

Xpeqt Ltd
2, Samokovsko Shose Blvd
Sofia 1138, Bulgaria
Mobile: *+359 887 383 600<tel:%2B359%20887%20383%20600>*