Nathan Bryant wrote: > Nathan Bryant wrote > >> - I haven't seen those oopses again either; they may have been caused >> by i810_configure_clocking being called twice during the >> initialization due to a merging goof on my part... > > > Ok, I spoke too soon. That piece of code couldn't cause the problem, > because of the surrounding if (clocking==... > > So there may be some VM/buffer related problem lurking under the covers > still. Originally the oops popped up in kswapd, for me, but I can't > trigger it again. > Well, your second version of the file had the merge done right (my code didn't include S/PDIF support or PM support, so those parts were different, but the parts that were the same as my code were done correctly). I'm attaching a patch that bumps the code from your 0.05b to a unified 0.06 and I'm also placing the 0.06 i810_audio.c.gz file on my web site in the same place that I put the 0.05 version. If people could please test this and report problems back, I would like to get this one off my plate (aka, I don't want to hear any more about artsd not working ever again so I want testers to tell me that it's fixed ;-) -- Doug Ledford http://people.redhat.com/dledford Please check my web site for aic7xxx updates/answers before e-mailing me about problems