HP Proliant Microserver AMI BIOS MOD

[thielj]

I know, that's the theory about WoL. In practice it doesn't work with e.g. FreeNAS (BSD) or UnRAID (Linux). However, people report to have it working with Windows. Might be a driver issue or BIOS/ACPI related - will need to waste a bit more time on this using a more recent Linux system to find out!

Thanks for the BIOS mod, by the way. I'm currently running the 'Russian version' but will happily update to something of known origin.

[camiloml]

Hmmm on poweroff the linux os instructs ethernet card to go completely off, i have seen that before. One of the shutdown scripts put card in non wol mode before cutting power, it's just a matter to add a new line to the script that shutsdown network to keep the card alive and waiting for the magic paket.

[DJBenson]

Hi

The modifications to stop the linux OS (Ubuntu in my case) do not work. I've added the following;

echo PCE6 > /proc/acpi/wakeup to /etc/rc.local and
NETDOWN=no to /etc/init.d/halt

When the server hibernates, you can wake it again for a short period of time (a minute at most) after which time it does not respond. The strange thing is, the ethernet indicator on the front panel is lit/flashing indicating traffic on the network, and if you send a wake packet to the server, the ethernet indicator flashes faster again proving it is receiving the packets, just doing nothing with them.

Hopefully this can be resolved (either in BIOS or by modding the OS) as I love my MicroServer, just don't like leaving it on 24/7 with the cost of electricity rising as it is.

[camiloml]

No, the mod i talk about is using ifconfig command

[DJBenson]

The command that stops the card from listening is NETDOWN=YES I thought, so having that set to NETDOWN=NO should keep the card alive.

What ifconfig commands are there in relation to WOL?

[camiloml]

Yes, the commands are to instruct the card when does she wake, by traffic, by magic packet, never... And are stored on network card eeprom, but linux normally resets that settings on poweroff command.

I can't remember now the exact script and sintax of the command (on iphone now) will look for them later when over my pc keyboard.

[camiloml]

Ok, my memory is weak sometimes, it was not ifconfig command, but ethtool

In my case was adding ethtool -s eth0 wol g to the script that shuts down the machine just before cutting the power.

That ethtool command means put the card in wake on magic packet mode. I'm running CentOS 5. But i guess Ubuntu is same.

If you want to see which wake modes support your ethernet card simply issue ethtool eth0 and it will show you.

Code:

[root@archivos ~]# ethtool eth0
Settings for eth0:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Advertised auto-negotiation: Yes
Speed: 1000Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 0
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: umbg
Wake-on: g
Current message level: 0x00000007 (7)
Link detected: yes

In case you want to know the wake modes supported for your card simply match the letters to this table

Code:

wol p|u|m|b|a|g|s|d...
              Set Wake-on-LAN options.  Not all devices support this.  The argument to this option is a string of characters specifying which
              options to enable.
              p  Wake on phy activity
              u  Wake on unicast messages
              m  Wake on multicast messages
              b  Wake on broadcast messages
              a  Wake on ARP
              g  Wake on MagicPacket(tm)
              s  Enable SecureOn(tm) password for MagicPacket(tm)
              d  Disable (wake on nothing).  This option clears all previous options.

[DJBenson]

Mmmm yes that is what I already had but it doesn't work on mine.

Are those supported sleep modes from a Microserver or are they just an example - mine only supports g "magic packet".

[thielj]

For reference:

Code:

root@HALCYON:/var/log# uname -a
Linux HALCYON 2.6.37.6-unRAID #3 SMP Wed Jun 29 17:54:44 MDT 2011 i686 AMD Athlon(tm) II Neo N36L Dual-Core Processor AuthenticAMD GNU/Linux

root@HALCYON:~# grep -o 'tg3.*' /var/log/syslog
tg3.c:v3.115 (October 14, 2010)
tg3 0000:02:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
tg3 0000:02:00.0: setting latency timer to 64
tg3 0000:02:00.0: eth0: Tigon3 [partno(BCM95723) rev 5784100] (PCI Express) MAC address 98:4b:e1:xx:xx:xx
tg3 0000:02:00.0: eth0: attached PHY is 5784 (10/100/1000Base-T Ethernet) (WireSpeed[1])
tg3 0000:02:00.0: eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1]
tg3 0000:02:00.0: eth0: dma_rwctrl[76180000] dma_mask[64-bit]
tg3 0000:02:00.0: irq 42 for MSI/MSI-X
tg3 0000:02:00.0: eth0: Link is up at 1000 Mbps, full duplex
tg3 0000:02:00.0: eth0: Flow control is on for TX and on for RX

root@HALCYON:/var/log# ethtool -i eth0
driver: tg3
version: 3.115
firmware-version: 5723-v3.35
bus-info: 0000:02:00.0

root@HALCYON:/var/log# ethtool eth0
Settings for eth0:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: No
Advertised auto-negotiation: Yes
Speed: 1000Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
MDI-X: Unknown
Supports Wake-on: g
Wake-on: g
Current message level: 0x000000ff (255)
Link detected: yes

[thielj]

I followed up some old bug reports on WoL problems with the tg3 drivers but these seems to have been fixed. Also, the tg3 driver in my system is fairly recent (Oct 2010).

More possibly leads....

a) firmware: As far as I understand, tg3 firmware BLOBs have been removed in Debian for license/purity reasons, assuming that these won't effect most NIC functions. Could that explain the differences between WoL behaviour in Linux and Windows?

b) /proc/acpi/wakeup: This only lists S4 (suspend to disk) states - Why isn't there a single S5 (soft off) state wakeup event?

[EDIT: Same with full Ubuntu 11.10 - However, an echo PCE6 into this only changes PCE6 to enabled, not most of the other values as well]

c) cat /sys/power/state - why is this completely empty? Is this just the unRAID kernel or is someone with a stock Ubuntu kernel or similar seeing the same? Docs are available in http://www.mjmwired.net/kernel/Documenta...erface.txt and http://www.mjmwired.net/kernel/Documenta...states.txt

[ECHO: Ubuntu 11.10: has 'disk' state available]