On Fri, Nov 01, 2013 at 12:01:26PM -0600, Greg Edwards wrote: > >> Is it expected the same pages might be mapped in multiple memslots? I > >> noticed the gfn overlap check in __kvm_set_memory_region(). > >> > >> It appears pfn_to_dma_pte() is behaving as expected, given half the huge > >> page is still mapped. Do I have that correct? If so, then we really > >> can't rely on iommu_iova_to_phys() alone to determine if its safe to > >> unpin a page in kvm_iommu_put_pages(). > >> > >> Ideas on how to best handle this condition? > > > > iommu_unmap should grab lpage_level bits from the virtual address > > (should fix the BUG), and should return correct number of freed pfns in > > case of large ptes (should fix the leak). Will send a patch shortly. > > Thanks, Marcelo. This patch also fixes the BUG: > > http://www.spinics.net/lists/kvm/msg97784.html Was using an old tree, without leak bug fixes from present upstream. -- 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
