Re: Using qsort(3) in the kernel

[Aleksey Cheusov <cheusov%tut.by@localhost>]

 >> Our implementation of qsort() uses recursion. That is IMHO a dangerous
 >> thing to do in the kernel.

> Another good reason to prefer mergesort is that it is a real time
> algorithm -- no degenerated cases.

Mergesort and Heapsort are completely different.  Qsort can also be
adapted to have O(n*log(n)) complexity in a worst case, but heapsort
is much easier to implement than that variant of qsort.

-- 
Best regards, Aleksey Cheusov.