netxen_nic firmware hang detected

[Marco Barbero <marco.barbero@gmail.com>]

Hi.
HP Proliant with a HP NC375T (qlogic chipset)
Debian Squeeze amd64

# lsmod -k

07:00.0 Ethernet controller: NetXen Incorporated NX3031 Multifunction
1/10-Gigabit Server Adapter (rev 42)
       Subsystem: Hewlett-Packard Company NC375T PCI Express Quad
Port Gigabit Server Adapter
       Kernel driver in use: netxen_nic
07:00.1 Ethernet controller: NetXen Incorporated NX3031 Multifunction
1/10-Gigabit Server Adapter (rev 42)
       Subsystem: Hewlett-Packard Company NC375T PCI Express Quad
Port Gigabit Server Adapter
       Kernel driver in use: netxen_nic
07:00.2 Ethernet controller: NetXen Incorporated NX3031 Multifunction
1/10-Gigabit Server Adapter (rev 42)
       Subsystem: Hewlett-Packard Company NC375T PCI Express Quad
Port Gigabit Server Adapter
       Kernel driver in use: netxen_nic
07:00.3 Ethernet controller: NetXen Incorporated NX3031 Multifunction
1/10-Gigabit Server Adapter (rev 42)
       Subsystem: Hewlett-Packard Company NC375T PCI Express Quad
Port Gigabit Server Adapter
       Kernel driver in use: netxen_nic


# ethtool -i eth4
driver: netxen_nic
version: 4.0.77
firmware-version: 4.0.579
bus-info: 0000:07:00.2


# dpkg -l |grep netxen
ii  firmware-netxen                     0.35~bpo60+1
    Binary firmware for QLogic Intelligent Ethernet (3000 and 3100
Series)

# uname -a
Linux firewall-slave 3.2.12 #1 SMP Wed Mar 21 17:33:10 CET 2012 x86_64 GNU/Linux


I'm experiencing 'firmware hang detected' issue on this quad
sometime it recovers after some seconds
sometime a reboot is needed

I have noticed HP has releases a very new firmware      5.0.0 (4 Jun 2012)
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=329290&prodSeriesId=4038765&prodNameId=4038767&swEnvOID=4004&swLang=8&mode=2&taskId=135&swItem=MTX-9c4312ff7adb4e6482acf4c1f7

Looking at their docs they say:

Important Information
======================
1) The nx_nic driver uses host-based firmware load.  When the user loads the
driver the host-based firmware image in the driver will be the version that
is displayed when checking the firmware version using ethtool.  The burned-in
firmware on the device as well as the host-based firmware version can be
verified by checking dmesg for entries similiar to the following:
nx_nic: Flash Version: Firmware[4.0.516], BIOS[2.1.3]
nx_nic: File FW[nx3fwmn.bin] version[4.0.516:12818]


In short: do I need another host-based firmware image after flashing
new firmware from HP? I have checked logs but I cannot find any
entries about flash version, only info about host based firmware:
Jun 20 01:03:50 firewall-slave kernel: [    8.898253] netxen_nic
0000:07:00.0: loading firmware from phanfw.bin
Jun 20 01:03:50 firewall-slave kernel: [   10.621102] netxen_nic
0000:07:00.0: firmware v4.0.579 [legacy]
Jun 20 01:03:50 firewall-slave kernel: [   10.622516] netxen_nic
0000:07:00.1: firmware v4.0.579 [legacy]
Jun 20 01:03:50 firewall-slave kernel: [   10.623869] netxen_nic
0000:07:00.2: firmware v4.0.579 [legacy]
Jun 20 01:03:50 firewall-slave kernel: [   10.625266] netxen_nic
0000:07:00.3: firmware v4.0.579 [legacy]



> Anyway, have you considered in requesting a firmware update to the BPO guys?

Mmm, let's re-read the HP info:

***
When the user loads the driver the host-based firmware image in the
driver will be the version that is displayed when checking the firmware
version using ethtool.
***

So, does it mean that when the kernel module is loaded (netxen_nic →
phanfw.bin) the embedded flash firmware of the adapter becames the same
that the loaded .bin file? If I read it correctly, this is only a notice
not a warning nor a requirement to fulfill :-?

>> The only problem I see here for using the provided package is that
>> you're not using a rpm, other than that it should be fine.
>
>> Anyway, have you considered in requesting a firmware update to the BPO
>> guys?
>
>
> what is not clear to me is if I need to update both flash based and host
> based firmware
> Or maybe I just need to update host based and don't touch flash at all?

(...)

As I see it (but I can be wrong), what happens is that when you load the
kernel module for the card which calls the firmware file, the embedded
flash firmware is superseded by the uploaded bin file, thus not enabled/
used.

My guess is that the embedded flash firmware is only used when the kernel
module cannot locate nor load the required firmware (*.bin) file so it
has to use then the outdated/stock firmware version.

Does it make any sense? O:-)