2.6.2
it’s exciting to see old friends again. andrew owen made it up this weekend to interview for some jobs, and it was wonderful to have him here. san antonio is not meeting his cold weather requirements. we hope he can find work up here.
the DeRonne family also made it up for the weekend, which meant the past several days were filled with chronic overeating, drinking and revelry. kevin and I finally got to watch “breaking away,” a 70s film about a boy’s romance with bicycling the summer after his high school graduation.
saturday night we made ‘lit crit lasagne’ - also known as american style lasagna; it was as tasty as always. we even made a vegetarian version with zucchini, mushrooms and extra green pepper. i am slightly inclined to modify the recipe just a hair: less cream cheese mixture, more green peppers, more tomato sauce. it’s okay if the meat sauce seems very runny when it goes in; the lasagne tends to dry up while baking, and it would make sense to cover it with tin foil after 30 minutes. i would like to try putting the cheese mixture on first, below the red sauce, to see if it can mix more (maybe i’ll even try sauce/cheese/sauce for a red oreo cookie appearance). 5 noodles suffice for one noodle layer in a 9×13 pan but it might be fun to put more in anyway (8 or 9). this time, we quadrupled the recipe: one vegetarian double recipe, one beef, and ended up using just 2 pans. 12 people were easily served.
what would a pleasant weekend be without a kernel re-compile? since 2.6.2 came out, i’ve been itching to recompile. more precisely, i’ve been itching to grab the latest from prism54.org for my netgear card. several obstacles presented themselves.
- dma didn’t want to work when i started using 2.6.2. this has always been a sore spot for me. i never want to apply all the bug fixes for various pci/ide drivers, but at the same time don’t want to keep re-compiling my kernel incessantly. for starters, i learned that makefiles are awesome, and on a gigahertz machine it doesn’t really mean more than about a minute wait for linux to recompile just one area of the kernel, link it all together and end up with a brand new compressed image. for finishers, it became clear that i’ve been relying on the “set dma by default” type option in the kernel, when really I could have just run my hdparm startup script and be done with my dma worries.
- the first time i tried alsa i couldn’t get it to work. i ended up reverting to oss drivers which, despite working, took a lot of coaxing for initialization. i’m not sure what the cause was, but the only way i could reliably listen to music was to open up xmms and adjust the levels, open up aumix and adjust the levels, and then use the proprietary keys on the thinkpad to adjust the levels. invariably, only after i had exercised all three volume adjustments did anything become audible.
- i had never bothered to script the initialization of my wifi card on startup. it’s so easy to be lazy, and i had two aliases to bring up the card after logging in. now, having spent the time to put one line in a startup script that already existed, i’m delighted with the results and amazed i didn’t go looking for the correct location earlier.
- when I had apm support compiled into the kernel, I would have to avoid closing the lid for fear of a certain crash. this has something to do with my wifi card. now with acpi, i can monitor the temp in addition to just the battery, and NOTHING happens when i close the lid. there is room for improvement, but the previous situation was far worse.
