Originally Posted by dura91
Like I stated in my other thread the problem is now solved for me with 260.19.21. Perhaps our problems are different or the 260.19.21 correction is incomplete?
Perhaps there are differences, this problem definitely occurs for me with 260.19.21, on several different kernels.
The only thing I saw disappear in recent releases was the "os_raise_smp_barrier(), invalid context!" errors, those stopped with 260.19.12. "Yield CPU while atomic" still happens with latest driver and wrecks havoc when it does.