Re: [PATCH] iommu: Split iommu_unmaps

2013-11-11 Thread Alex Williamson
On Thu, 2013-11-07 at 16:37 +, David Woodhouse wrote: > On Fri, 2013-05-24 at 11:14 -0600, Alex Williamson wrote: > > iommu_map splits requests into pages that the iommu driver reports > > that it can handle. The iommu_unmap path does not do the same. This > > can cause problems not only from

Re: [PATCH] iommu: Split iommu_unmaps

2013-11-07 Thread David Woodhouse
On Fri, 2013-05-24 at 11:14 -0600, Alex Williamson wrote: > iommu_map splits requests into pages that the iommu driver reports > that it can handle. The iommu_unmap path does not do the same. This > can cause problems not only from callers that might expect the same > behavior as the map path, bu

Re: [RESEND PATCH] iommu: Split iommu_unmaps

2013-06-20 Thread Joerg Roedel
On Mon, Jun 17, 2013 at 07:57:34PM -0600, Alex Williamson wrote: > iommu_map splits requests into pages that the iommu driver reports > that it can handle. The iommu_unmap path does not do the same. This > can cause problems not only from callers that might expect the same > behavior as the map p

[RESEND PATCH] iommu: Split iommu_unmaps

2013-06-17 Thread Alex Williamson
iommu_map splits requests into pages that the iommu driver reports that it can handle. The iommu_unmap path does not do the same. This can cause problems not only from callers that might expect the same behavior as the map path, but even from the failure path of iommu_map, should it fail at a poi

Re: [PATCH] iommu: Split iommu_unmaps

2013-06-05 Thread Alex Williamson
Joerg, Any comments on this? I need this for vfio hugepage support, otherwise we risk getting a map failure that results in a BUG_ON from iommu_unmap_page in amd_iommu. I can take it in through my vfio tree to keep the dependencies together if you want to provide an ack. Thanks, Alex On Fri,

[PATCH] iommu: Split iommu_unmaps

2013-05-24 Thread Alex Williamson
iommu_map splits requests into pages that the iommu driver reports that it can handle. The iommu_unmap path does not do the same. This can cause problems not only from callers that might expect the same behavior as the map path, but even from the failure path of iommu_map, should it fail at a poi