Incomplete Source Releases

Apparently some of the companies upon whom we've put legal pressure for GPL compliance still don't comply. That is sad, and we won't tolerate this behaviour. The sources need to contain the tool for creating the firmware image, and they need to compile ;)

Fujitsu-Siemens and Sitecom are still lacking the firmware build tool. We're threatening Fujitsu-Siemens with enforcing the contract penalty set forth in our out-of-court settlement. Sitecom is threatened with enforcing the penalty stated in the preliminary injunction.

I don't do this for fun, and I would feel much better if I hadn't to threaten anybody with anything. Apparently even under such threat, those companies find themselves unable to comply with the GPL. Why can't they just make everybody happy and release those missing pieces.... *sigh*

I've heard rumors that Belkin and Asus sources don't compile. As my time is very limited (esp. considered the large number of cases): Please report to me if you have problems with the respective source releases. I am very happy to act on your behalf. After all, I'm doing this mostly for you users. There aren't any valuable modifications in those firmware sources that I need to integrate... all I want to achieve is enabling the users/customers of those WLAN-AP's to be able to exert their GPL-granted right to modify the firmware and to run modified versions of the firmware.

Doing lots of benchmarks / tuning / profiling lately

During the last weeks I've been working on tuning/benchmarking/profiling the Sun V20z dual Opteron boxes for high-speed packet filtering purpose.

Some of my findings:

• i386 kernels give you higher pps than x86_64 (because sk_buff is smaller)
• e1000 are way faster than tg3 boards (could be hardware or driver issue)
• Intel PRO/1000MT Quad e1000 boards suck (apparently problems with the onboard PCI-X bridge)
• Connection Tracking performance is not that bad...
• ip_tables performance sucks, even if the ruleset is empty ?!?
• 2.4.x has slightly worse results than 2.6.x if you use IRQ affinity, but really sucks if you don't, since the kernel doesn't balance IRQ's by itself (and irqbalance daemon only balances every 10 seconds)
• You can route up to 1Mpps at 64bytes packet size
• ip_conntrack and iptable_filter at suck at least 300kpps, giving 700kpps as a result

Expect a more detailed report within the next weeks.

Sun V20z is a Newisys 2100

I just discovered that the Sun V20z dual Opteron systems are actually developed and produced by Newisys.

Newisys apparently is a extremely pro-Linux company. Not just for marketing purpose, but they mean it. They release all drivers (IPMI, jnet, ...) under the GPL, even actively contribute them back to the free software community.

They're even looking into running LinuxBIOS on their boxes... While LinuxBIOS is actually an improvement, I'd rather like to see OpenFirmware. What is the point of putting Linux there? OpenFirmware provides you with whatever you need, even device-drivers written in forth / f-code. Well...