jump to navigation

Attansic L1 giganet driver on ASUS P5B-E November 30, 2006

Posted by TSAI HONG-BIN in Linux.
add a comment

So, it’s always hurt to port Linux to hardwares that are new to market. ASUS P5B-E is one of the best examples. Its onboard LAN chip, Attansic L1, is not supported in Fedora Core 6 by default. http://www.hogchain.net/attansic/attansic.html here is a website offering links to official driver download and kernel patches from other contributors.

Something worth of mentioning is that on kernel 2.6.18, a little modification to driver source code is required. in at_main.c, all skb_shinfo(skb)->tso_size should be changed to skb_shinfo(skb)->gso_size. Done. Remember to install kernel-devel rpm with corresponding kernel arch.

NOTE: there is a update of this driver issue on kernel 2.6.19, please refer to https://furseal.wordpress.com/2007/03/13/attansic-l1-giganet-driver-on-asus-p5b-e-kernel-2618/

NOTE: update to kernel 2.6.20 https://furseal.wordpress.com/2007/04/09/attansic-l1-giganet-driver-on-asus-p5b-e-kernel2620/

Advertisements

Micro$oft: crap maker November 23, 2006

Posted by TSAI HONG-BIN in Misc.
add a comment

Why M$ can always piss me off so easily so much, I’m not a man of anger at all…

Hibernation? No, that’s “Play dead, Oops, it becomes true” feature of Windows’ Power Management. Have you ever failed to resume from hibernation and let the reboot process kill all opened documents, saved or not, with helplessness? I DO! every other time. Now Vista is suggesting people to use HORM (Hibernation once/ Resume many) and it sets Power off process to hibernate by default. Maybe this tuning makes people forget how slow it boots from S5, for a very short period. Then its users shall start to suffer from the “true dead” instead of hibernation.

Don’t let the stupid Ad confuse you, Microsoft won’t be able to boot off quickly and stably. If you put expectation on its resume, you’ll soon be disappointed. Why doesn’t Micro$oft hibernate itself and see what will happen? It’s such a terrible waste to put one’s effort on solving problems made by M$.

By the way, there is another crap thing about M$ Hibernation, it cannot be saved on another partition. Do you see what’s wrong? Windows boot loader is able to recognize different disks and partitions on the box and put the boot process to a specified boot sector, but not capable to load a hiberfil.sys anywhere else than C:\? Does this sound reasonable? I don’t want to criticize M$’ engineers because I’m an engineer too. But making fool of all Windows users in such an obvious way is against the morality of engineering. No wonder why they need to sue so many company. Because they just cannot bear to see engineers smarter, imaginative, more innovative and considerate than they are.