On Thu, Oct 12, 2023 at 05:11:09PM +0800, Yi Liu wrote: > On 2023/10/11 00:58, Jason Gunthorpe wrote: > > On Thu, Sep 21, 2023 at 12:51:22AM -0700, Yi Liu wrote: > > > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > > > index 660dc1931dc9..12e12e5563e6 100644 > > > --- a/include/linux/iommu.h > > > +++ b/include/linux/iommu.h > > > @@ -14,6 +14,7 @@ > > > #include <linux/err.h> > > > #include <linux/of.h> > > > #include <uapi/linux/iommu.h> > > > +#include <uapi/linux/iommufd.h> > > > > Oh we should definately avoid doing that! > > Maybe this is a good moment to start a new header file exclusively for > > iommu drivers and core subsystem to include? > > > > include/linux/iommu-driver.h > > > > ? > > > > Put iommu_copy_user_data() and struct iommu_user_data in there > > > > Avoid this include in this file. > > sure. btw. seems all the user of this API and structure are in the > drivers/iommu directory. can we just putting them in > drivers/iommu/iommu-priv.h? iommu-priv.h should be private to the core iommu code, and we sort of extended it to iommufd as well. iommu-driver.h would be "private" to the core and all the drivers only. As include ../.. is often frown on at large scale it is probably better to be in include/linux > Just one concern. There are other paths (like cache_invalidate of > this series and Nic's set_dev_data) uses this struct as well. I'm > a bit worrying if it is good to put type here as type is meaningful > for the domain_alloc_user path. There is always a type though? I haven't got that far in the series yet to see.. Jason
