Re: [patch 2/9] mempool gfp flag

[Andrew Morton <akpm@osdl.org>]

Nick Piggin <nickpiggin@yahoo.com.au> wrote:
>
>  The first is that mempool_alloc can possibly get stuck in __alloc_pages
>  when they should opt to fail, and take an element from their reserved pool.
> 
>  The second is that it will happily eat emergency PF_MEMALLOC reserves
>  instead of going to their reserved pools.
> 
>  Fix the first by passing __GFP_NORETRY in the allocation calls in
>  mempool_alloc. Fix the second by introducing a __GFP_MEMPOOL flag
>  which directs the page allocator not to allocate from the reserve
>  pool.
> 
> 
>  Index: linux-2.6/include/linux/gfp.h
>  ===================================================================
>  --- linux-2.6.orig/include/linux/gfp.h	2005-04-12 22:26:10.000000000 +1000
>  +++ linux-2.6/include/linux/gfp.h	2005-04-12 22:26:11.000000000 +1000
>  @@ -38,14 +38,16 @@ struct vm_area_struct;
>   #define __GFP_NO_GROW	0x2000u	/* Slab internal usage */
>   #define __GFP_COMP	0x4000u	/* Add compound page metadata */
>   #define __GFP_ZERO	0x8000u	/* Return zeroed page on success */
>  +#define __GFP_MEMPOOL	0x10000u/* Mempool allocation */

I think I'll rename this to "__GFP_NOMEMALLOC".  Things other then mempool
might want to use this.