[Linaro-mm-sig] [RFC] ARM DMA mapping TODO, v1
thellstrom at vmware.com
Fri Apr 29 05:50:12 UTC 2011
On 04/29/2011 02:29 AM, Benjamin Herrenschmidt wrote:
> On Thu, 2011-04-28 at 15:37 -0400, Jerome Glisse wrote:
>> Jesse also pointed out space exhaustion inside the iommu and i believe
>> this should also be considered. This is why i believe the dma_* api is
>> not well suited. In DRM/TTM we use pci_dma_mapping* and we also play
>> with with page set_page*_uc|wc|wb.
> Which are yet another set of completely x86-centric APIs that have not
> been thought in the context of other architectures and are probably
> mostly unimplementables on half of them :-)
I've been doing some thinking over the years on how we could extend that
functionality to other architectures. The reason we need those is
because some x86 processors (early AMDs and, I think VIA c3) dislike
multiple mappings of the same pages with conflicting caching attributes.
What we really want to be able to do is to unmap pages from the linear
kernel map, to avoid having to transition the linear kernel map every
time we change other mappings.
The reason we need to do this in the first place is that AGP and modern
GPUs has a fast mode where snooping is turned off.
However, we should be able to construct a completely generic api around
these operations, and for architectures that don't support them we need
a) Whether we want to support them anyway (IIRC the problem with PPC is
that the linear kernel map has huge tlb entries that are very
inefficient to break up?)
b) Whether they are needed at all on the particular architecture. The
Intel x86 spec is, (according to AMD), supposed to forbid conflicting
caching attributes, but the Intel graphics guys use them for GEM. PPC
appears not to need it.
c) If neither of the above applies, we might be able to either use
explicit cache flushes (which will require a TTM cache sync API), or
require the device to use snooping mode. The architecture may also
perhaps have a pool of write-combined pages that we can use. This should
be indicated by defines in the api header.
> Linaro-mm-sig mailing list
> Linaro-mm-sig at lists.linaro.org
More information about the Linaro-mm-sig