2.6.34 fixes laptop resume

My laptop had still been occasionally hanging on resume, but looking back over my blog, I realise that it may well have been because I re-enabled desktop effects.  So I'm not exactly sure why now, resume is completely fixed by 2.6.34.  There are still considerable improvements and fixes going into the kernel Radeon DRM, so it may have been fixed by that.  Or, it was the new asynchronous suspend & resume code that made its way into the kernel in 2.6.34, a project originated by Rafael Wysocki.  Either way, I am extremely happy.

Desktop effects occasionally borks resume?

After reverting the laptop to x86, I wasn't happy with the bugs in mesa-7.5, especially since they were solved in 7.6, so I keyworded X.org, Mesa and their dependencies and took a deep breath.  Not such a deep breath, as I had recently done the same thing on the desktop to no ill effect.  So the laptop was cruising with xorg-server-1.7 and mesa-7.7, and everything seemed fine.  However, an old problem had returned, or possibly a new one had appeared.  Either way, the symptom was that the laptop would occasionally hang after resuming from sleep.  Wasn't happening all the time, but often enough to be very annoying, as the only way out was by magic sysrq.  On a hunch, I disabled KDE-4's desktop effects and it hasn't happened since.  Doesn't sound like a fun problem to debug if it means rebooting every time it occurs, so I think I'll leave this one to the pros.

Wireless LED: solved by module parameter

At some point, my wireless LED stopped working again, after I thought I had solved it by enabling the extra kernel options mentioned in a previous post.  It was no big deal, so I ignored it.  Then, whilst searching for something else wireless related, I happened upon this laptop matrix about RF switches!  There I learnt that passing led=1 as a parameter to the ipw2200 module will enable the LED.  Hallelujah.  Why the LED worked in the past before I ever knew about this remains a mystery.