On Wed, Jun 02, 2010 at 01:15:34PM +0300, Michael S. Tsirkin wrote: > One of the issues I see with the current patch is that > it uses the mlock rlimit to do this pinning. So this wastes the rlimit > for an app that did mlockall already, and also consumes > this resource transparently, so an app might call mlock > on a small buffer and be surprised that it fails. > > Using mmu notifiers might help? MMU notifiers are problematic because they are designed for situations where we can do demand paging. The invalidate_range_start and invalidate_range_end functions are not only called on munmap, they also run when mprotect is called (in which case we don't want to tear down iommu mappings). So what may happen with mmu notifiers is that we accidentially tear down iommu mappings. With demand-paging this is no problem because the io-ptes could be re-faulted. But that does not work here. Joerg -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
