Comment 59 for bug 31709

Lukáš,
I'm very happy you managed to solve it! :)
By the way, that's an "old" workaround; you may just add the kernel parameter "pata_ali.atapi_dma=1" on Lucid, but in my case it just prevents the CPU from turning into a hog, while the ata_generic workaround does not perform very good on my system. The only way to "solve" this issue and let a proper module handle my drives is to turn back to Edgy or just to build a kernel by myself.