Re: Regression after commit 19809c2da28a ("mm, vmalloc: use __GFP_HIGHMEM implicitly")

[Michal Hocko]

On Sun 11-02-18 15:51:07, Matthew Wilcox wrote:
> On Sun, Feb 11, 2018 at 04:05:15AM -0800, Matthew Wilcox wrote:
> > On Sun, Feb 11, 2018 at 03:28:08AM -0800, Matthew Wilcox wrote:
> > > Now, longer-term, perhaps we should do the following:
> > > 
> > > #ifdef CONFIG_ZONE_DMA32
> > > #define OPT_ZONE_DMA32	ZONE_DMA32
> > > #elif defined(CONFIG_64BIT)
> > > #define OPT_ZONE_DMA	OPT_ZONE_DMA
> > > #else
> > > #define OPT_ZONE_DMA32 ZONE_NORMAL
> > > #endif
> > 
> > For consistent / coherent memory, we have an allocation function.
> > But we don't have an allocation function for streaming memory, which is
> > what these drivers want.  They also flush the DMA memory and then access
> > the memory through a different virtual mapping, which I'm not sure is
> > going to work well on virtually-indexed caches like SPARC and PA-RISC
> > (maybe not MIPS either?)
> 
> Perhaps I (and a number of other people ...) have misunderstood the
> semantics of GFP_DMA32.  Perhaps GFP_DMA32 is not "allocate memory below
> 4GB", perhaps it's "allocate memory which can be mapped below 4GB".

Well, GFP_DMA32 is clearly under-documented. But I _believe_ the
intention was to really return a physical memory within 32b address
range.

> Machines with an IOMMU can use ZONE_NORMAL.  Machines with no IOMMU can
> choose to allocate memory with a physical address below 4GB.

This would be something for the higher level allocator I think. The page
allocator is largely unaware of IOMMU or any remapping and that is good
IMHO.

> After all, it has 'DMA' right there in the name.

The name is misnomer following GFP_DMA which is arguably a better fit.
GFP_MEM32 would be a better name.

Btw. I believe the GFP_VMALLOC32 shows that our GFP_DM32 needs some
love. The user shouldn't really care about lowmem zones layout.
GFP_DMA32 should simply use the appropriate zone regardless the arch
specific details.

-- 
Michal Hocko
SUSE Labs