Hi, I'd suggest verifying that it's _just_ a kernel update that messes with it. And once you've verified that, what about doing some kernel source version bisecting to narrow down when the relevant change went in that's caused your regression? Adrian