Re: [PATCH V3 05/10] vfio: Allow null group for noiommu without containers

[Jacob Pan <jacob.pan@linux.microsoft.com>]

Hi Jason,

On Thu, 9 Apr 2026 15:44:16 -0300
Jason Gunthorpe <jgg@nvidia.com> wrote:

> On Thu, Apr 02, 2026 at 10:11:41PM -0700, Jacob Pan wrote:
> 
> > @@ -598,6 +604,14 @@ static struct vfio_group
> > *vfio_noiommu_group_alloc(struct device *dev, struct vfio_group
> > *group; int ret;
> >  
> > +	/*
> > +	 * With noiommu enabled under cdev interface only, there
> > is no need to
> > +	 * create a vfio_group if the group based containers are
> > not enabled.
> > +	 * The cdev interface is exclusively used for iommufd.
> > +	 */
> > +	if (vfio_null_group_allowed())
> > +		return NULL;  
> 
> Sashiko wondered if this causes a null pointer deref without pointing
> at a specific case. Claude found a case:
> 
> @@ -696,6 +696,9 @@ void vfio_device_remove_group(struct vfio_device
> *device) struct vfio_group *group = device->group;
>         struct iommu_group *iommu_group;
>  
> +       if (!group)
> +               return;
> +
>         if (group->type == VFIO_NO_IOMMU || group->type ==
> VFIO_EMULATED_IOMMU) iommu_group_remove_device(device->dev);
> 
> Happens during error unwind in __vfio_register_dev()
will do.  

> It also points out that the hunks are weirdly split between this patch
> and "vfio: Enable cdev noiommu mode under iommufd" so that things are
> broken at this point.
> 
> I think you should pull the vfio_device_has_group() and related into
> this patch.
> 
makes sense, will regroup.