Hi Ben,
I'm have a concern about the current change_page_attr() patch. It
gives the impression that the GART memory always gets mapped
write-combined. This is not always true. For example, on IA-64,
attribute-aliasing issues are prevented by requiring IA-64 platforms
to use coherent AGP DMA. In other words, the AGP memory will be
mapped writeback, as usual.
As long as change_page_attr() is used for AGP-related stuff only,
there is probably no real issue with the patch in its current form
(its simply a no-op on most non-x86 platforms). However, I'm a bit
worried that someone might start to use it for other things, such that
change_page_attr() could no longer be a no-op on those platforms.
Since the DMA coherency issue is an AGP specific issue, perhaps just
renaming the macro to agp_change_page_attr() would take care of my
concern. What do you think?
--david
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jun 15 2002 - 22:00:28 EST