Messages in this thread | | | Date | Tue, 27 Oct 2009 18:11:11 +0900 | Subject | Re: [PATCH] x86: adjust GFP mask handling for coherent allocations | From | FUJITA Tomonori <> |
| |
On Tue, 27 Oct 2009 08:58:14 +0000 "Jan Beulich" <JBeulich@novell.com> wrote:
> >Well, in the first place, we don't need the #ifdef in Jan's patch. We > >can always use DMA_BIT_MASK(24) for the fallback device. > > But this one I don't agree with - the purpose of the patch is to not have > a 24-bit (or 32-bit) mask here unconditionally: It would result in GFP_DMA > to be forced on for the allocation (which the patch specifically eliminates),
If a driver doesn't want to GFP_DMA, it should set up the dma_coherent_mask of the device and pass it. In fact, it should do. A driver that uses the fallback device is broken.
Why can't you fix drivers that use the fallback instead of adding another hack to the common place?
| |