hi,

ah, I see you're wait ahead of me.  looks good to me.

-Chuck


On Wed, Nov 20, 2002 at 10:22:59PM +0100, Manuel Bouyer wrote:
> Hi,
> so here is my proposed fix to this problem.
> There are a few interface changes:
> - amap_extend() gains a flags parameter, can be either M_NOWAIT or M_WAITOK
>   (amap_extend() use malloc)
> - uvm_km_kmemalloc() will pass UVM_KMF_NOWAIT in addition to
>   UVM_KMF_TRYLOCK to uvm_map()
> - uvm_mapent_alloc() gains a flags parameter (PR_NOWAIT or PR_WAITOK)
...