Re: [PATCH 1/2] KVM: async kvm_destroy_vm for vfio devices

[Jason Gunthorpe]

On Thu, Jan 12, 2023 at 12:21:17PM -0500, Matthew Rosato wrote:

> So should I work up a v2 that does symbol gets for kvm_get_kvm_safe
> and kvm_put_kvm from vfio_main and drop kvm_put_kvm_async?  Or is
> the patch Yi is working on changing things such that will also
> address the deadlock issue?

I don't think Yi's part will help

> +361,22 @@ static int vfio_device_first_open(struct vfio_device
> *device, if (ret) goto err_module_put;
>  
> +       if (kvm && !vfio_kvm_get(kvm)) {

Do call it kvm_get_safe though

> +               ret = -ENOENT;
> +               goto err_unuse_iommu;
> +       }
>         device->kvm = kvm;
>         if (device->ops->open_device) {
>                 ret = device->ops->open_device(device);
>                 if (ret)
> -                       goto err_unuse_iommu;
> +                       goto err_put_kvm;
>         }
>         return 0;
>  
> -err_unuse_iommu:
> +err_put_kvm:
> +       vfio_put_kvm(kvm);
>         device->kvm = NULL;
> +err_unuse_iommu:
>         if (iommufd)
>                 vfio_iommufd_unbind(device);
>         else
> @@ -465,6 +471,9 @@ static int vfio_device_fops_release(struct inode *inode, struct file *filep)
>  
>         vfio_device_group_close(device);
>  
> +       if (device->open_count == 0 && device->group->kvm)
> +               vfio_kvm_put(device->group->kvm);
> +

No, you can't touch group->kvm without holding the group lock,
that is the whole point of the problem..

This has to be device->kvm

Jason