On Sat, Sep 24, 2016 at 06:10:37PM +0200, Paolo Bonzini wrote: > > > On 24/09/2016 00:05, Michael S. Tsirkin wrote: > >> e events... > >> > Checking PATCH 18/19: hw/i386: Introduce AMD IOMMU... > >> > ERROR: struct MemoryRegionIOMMUOps should normally be const > >> > #1527: FILE: hw/i386/amd_iommu.h:280: > >> > + MemoryRegionIOMMUOps iommu_ops; > > > > False positive. > > Not quite, there's no reason for the ops to be inside AMDVIState rather > than a (static) global. > > Paolo
I guess but it's a logic refactoring, not merely a style one. I feel we can do it on top, tool shouldn't require this.
