>>> On 14.11.18 at 12:57, <roger....@citrix.com> wrote: > Bridges are not behind an IOMMU, and are already special cased and > silently skipped in amd_iommu_add_device. Apply the same special > casing when updating page tables.
But bridges also don't issue I/O on their own if I'm not mistaken. So what I'm missing here is a word on the benefit of this change. I also question the "silently" in your wording, seeing the AMD_IOMMU_DEBUG() there. The code change itself looks fine to me, albeit personally I'd prefer if it fully matched the other conditional (i.e. if you flipped the operands of && ). Of course the special casing of the hardware domain here is somewhat odd anyway. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
