Re: [PATCH] add __GFP_ZERO to GFP_LEVEL_MASK

[Christoph Lameter]

On Tue, 24 Jul 2007, Peter Zijlstra wrote:

> > There is another exception for __GFP_DMA.
> 
> non of the zone specifiers are

__GFP_DMA is handled in a similar way to __GFP_ZERO though. Its explicitly 
listed in BUG_ON() because it can be specified in the gfpflags to kmalloc 
but also set by having created a slab with SLAB_DMA. It is also cleared 
by the & GFP_LEVEL_MASK.
 
> > > Anybody else got a preference?
> > 
> > >  #define __GFP_BITS_MASK ((__force gfp_t)((1 << __GFP_BITS_SHIFT) - 1))
> > >  
> > > -/* if you forget to add the bitmask here kernel will crash, period */
> > > +/*
> > > + * If you forget to add the bitmask here kernel will crash, period!
> > > + *
> > > + * GFP_LEVEL_MASK is used to filter out the flags that are to be passed to the
> > > + * page allocator.
> > > + *
> > 
> > GFP_LEVEL_MASK is also used in mm/vmalloc.c. We need a definition that 
> > goes beyond slab allocators.
> 
> Right, bugger.

Lets get rid of the cryptic sentence there and explain it in a better way. 
GFP_LEVEL_MASK contains the flags that are passed to the page allocator
by derived allocators (such as slab allocators and vmalloc, maybe the 
uncached allocator may use it in the future?).

__get_vm_area_node also relies on GFP_LEVEL_MASK to clear the __GFP_ZERO 
flag. Otherwise the kmalloc_node there would needlessly return zeroed 
memory (or have failed in the past).
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/