On Tuesday 24 August 2004 00:19, you wrote:
> +       if (likely(next == tss->io_bitmap_owner)) {

Probably a stupid question, but what's stopping the tss->io_bitmap_owner from 
being killed, and then a new thread_struct being kmalloc()'ed in the exact 
same place as the old one? I realize it's highly unlikely, I'm just wondering 
if it's possible at all.

I guess clearing tss->io_bitmap_owner whenever we kfree() the bitmap owner's 
thread_struct would plug that up.

-Ryan