Multiport Gigabit Card?

[pgwipeout]

Hi, I tried the following card:

   HP NC360T PCI Express Dual Port Gigabit Server Adapter

It uses the 82571EB chipset. I am using the Debian unstable Kernel and, unfortunately, the kernel panics during boot when the card is plugged in. However, I found a very interesting project that provides a free firmware for quad-cards with the BCM5719 chip (it replaces the blobs in the card itself):

https://github.com/meklort/bcm5719-fw

Also, these cards look MUCH more promising to me as they have just one chip instead of four like most Intel NICs and are rated at 4W maximum (instead of 10W that the Intel quad cards seem to use). I am planning to get one and see how well this works.

Edit - could this be related to the following?:

https://forum.pine64.org/showthread.php?tid=8374

Good Morning,

It may be the hardware issue, but do note there is an issue with the rk3399 pcie controller that is currently unmitigated.
See the LKML thread here : https://lore.kernel.org/linux-pci/CAMdYz...gmail.com/
Also see this for additional information : https://lkml.org/lkml/2020/4/6/320

TLDR: We found the rk3399 throws either a synchronous error or a SError when a pcie device sends an unknown message.
The error type is determined by which cpu cluster handles the message.
We hijacked the arm64 error handling and processed it ourselves, and that corrects the issue, but it's not a good fix.
In the end, it was determined that significant changes to how arm64 handles pcie errors in the linux kernel need to happen.

Thanks a lot @pgwipeout - I had overlooked this.

I applied the hack and also enabled PCIe gen2 mode via the already discussed link-speed change in the dts and ... the kernel boots now with the PCIe card inserted!

Even without the taskset-call all NICs are now recognized (I removed the Mac addresses from the output):

Code:

root@devuan:~# lspci
00:00.0 PCI bridge: Fuzhou Rockchip Electronics Co., Ltd Device 0100
01:00.0 Ethernet controller: Broadcom Limited NetXtreme BCM5719 Gigabit Ethernet PCIe (rev 01)
01:00.1 Ethernet controller: Broadcom Limited NetXtreme BCM5719 Gigabit Ethernet PCIe (rev 01)
01:00.2 Ethernet controller: Broadcom Limited NetXtreme BCM5719 Gigabit Ethernet PCIe (rev 01)
01:00.3 Ethernet controller: Broadcom Limited NetXtreme BCM5719 Gigabit Ethernet PCIe (rev 01)

root@devuan:~# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
   link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
   inet 127.0.0.1/8 scope host lo
      valid_lft forever preferred_lft forever
   inet6 ::1/128 scope host
      valid_lft forever preferred_lft forever
2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
   link/ether ***************** brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
   link/ether ***************** brd ff:ff:ff:ff:ff:ff
4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
   link/ether ***************** brd ff:ff:ff:ff:ff:ff
5: eth3: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
   link/ether ***************** brd ff:ff:ff:ff:ff:ff
6: eth4: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
   link/ether ***************** brd ff:ff:ff:ff:ff:ff

Power consumption is at 5.3 Watts with one of the four links active. This sounds quite promising, I will now try the other Dualport card I bought previously.

Update: I tried the Intel Dualport-NIC (on HP NC360T), however, the ports don't get recognized automatically.

I wonder if it was the change to PCIe gen 2 that made it work correctly?
Do you notice if it is running at native speed or downgraded under lspci -vvv?