Jason Thorpe wrote:
> Actually, I think it would be kind of nice to generalize the handling 
> of this in UVM, rather than handling it per-platform in the fault code.

Indeed. However, MD code updates vm_ssize only, it doesn't actually
handle stack extending. I'll look into making them all use
uvm_grow() at least - or perhaps it should be done in uvm_fault()
automagically?

> Seems like it would be simple enough to mark a single vm_map_entry in 
> this way, though you'd need some additional information (i.e. the 
> current "depth").

Actually standard stack appears to be mapped in whole at startup,
and individual pages are faulted in as needed. There is no
extending of the map entry AFAICS, only call to uvm_map_protect()
when the stack resource limit is changed. So I will do similar
uvm_map_protect() thing even in the Linux code, I guess.

Jaromir
-- 
Jaromir Dolecek <jdolecek@NetBSD.org>            http://www.NetBSD.cz/
-=- We should be mindful of the potential goal, but as the Buddhist -=-
-=- masters say, ``You may notice during meditation that you        -=-
-=- sometimes levitate or glow.   Do not let this distract you.''   -=-