On Sat, Jan 28, 2012 at 11:25 AM, Jason Gunthorpe
[off-list ref] wrote:

I know accessing system memory (eg obtained via mmap on
/sys/bus/pci/devices/0000:00:02.0/resource0) has been asked for in the
past, and IIRC, the problem was that some of the common code, (GUP?)
errored on these maps. I don't know if Roland's case is similar.

I think the problem there is that this is done via remap_pfn_range()
or similar, and the mapping has no underlying pages at all.  So we
would need a new interface that gives us different information for
such cases.

This is quite a bit trickier since I don't think the DMA API even has
a way to express getting a "device A" bus address for some memory
that is in a BAR for "device B".  So I'm not trying to address this case
(yet).  First I'd like to deal with as many flavors of page-backed
mappings as I can.

 - R.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>