Hi Andrew, > aargh. That's a totally different crash :( Isn't is exactly what would happen in the case I described, i.e. adapters freed but still in the list? The freed area could be overwritten randomly, causing crashes at different locations each time we slightly change the code? > Were you using this patch? I think we would spare a significant amount of time if Miles could test Tony's patch to nvidiafb - unless he already tried and I am not aware? Miles? Thanks, -- Jean Delvare