>I have some unpleasant news: it looks like it takes slightly longer with
>the
>fix, but the crash still managed to occur with 3.2 and I lack ESP to check
> for warnings immediately before that. I thought that 8 hours is enough
> testing, heh. It simply took 9 and a half under nigh-ideal conditions
> (stable signal from fixed AP with notebook also steadily in one place).
By the way, what could trigger these crashes in "network event sense"? If this
can be reproduced with Ad-Hoc (something like ""Ad-Hoc peer changes ESSID then
changes it back"), I could easily do some stress-testing with a second Linux-
running notebook.