Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1664567 > unrolled thread

Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit Juergen Gross <jgross@suse.com> - 2017-06-13 09:50 +0200
    Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit Stefano Stabellini <sstabellini@kernel.org> - 2017-06-14 03:00 +0200
      Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit Boris Ostrovsky <boris.ostrovsky@oracle.com> - 2017-06-14 16:00 +0200
        Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit Stefano Stabellini <sstabellini@kernel.org> - 2017-06-14 21:20 +0200

#1664567 — Re: [PATCH v3 14/18] xen/pvcalls: disconnect and module_exit

On 02/06/17 21:31, Stefano Stabellini wrote:
> Implement backend_disconnect. Call pvcalls_back_release_active on active
> sockets and pvcalls_back_release_passive on passive sockets.
> 
> Implement module_exit by calling backend_disconnect on frontend
> connections.
> 
> Signed-off-by: Stefano Stabellini <stefano@aporeto.com>
> CC: boris.ostrovsky@oracle.com
> CC: jgross@suse.com
> ---
>  drivers/xen/pvcalls-back.c | 49 ++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 49 insertions(+)
> 
> diff --git a/drivers/xen/pvcalls-back.c b/drivers/xen/pvcalls-back.c
> index b541887..6afe7a0 100644
> --- a/drivers/xen/pvcalls-back.c
> +++ b/drivers/xen/pvcalls-back.c
> @@ -800,6 +800,38 @@ static int backend_connect(struct xenbus_device *dev)
>  
>  static int backend_disconnect(struct xenbus_device *dev)
>  {
> +	struct pvcalls_fedata *priv;
> +	struct sock_mapping *map, *n;
> +	struct sockpass_mapping *mappass;
> +	struct radix_tree_iter iter;
> +	void **slot;
> +
> +
> +	priv = dev_get_drvdata(&dev->dev);
> +
> +	list_for_each_entry_safe(map, n, &priv->socket_mappings, list) {
> +		pvcalls_back_release_active(dev, priv, map);
> +	}

You can drop the {}

> +
> +	radix_tree_for_each_slot(slot, &priv->socketpass_mappings, &iter, 0) {
> +		mappass = radix_tree_deref_slot(slot);
> +		if (!mappass || radix_tree_exception(mappass)) {

This looks fishy.

You might call radix_tree_deref_retry(NULL). Right now this is okay,
but you depend on the radix tree internals here.

> +			if (radix_tree_deref_retry(mappass)) {
> +				slot = radix_tree_iter_retry(&iter);
> +				continue;

The continue; statement is pointless here.

> +			}
> +		} else
> +			pvcalls_back_release_passive(dev, priv, mappass);
> +	}
> +
> +	xenbus_unmap_ring_vfree(dev, (void *)priv->sring);

Drop the cast.

> +	unbind_from_irqhandler(priv->irq, dev);
> +
> +	list_del(&priv->list);
> +	destroy_workqueue(priv->wq);
> +	kfree(priv);
> +	dev_set_drvdata(&dev->dev, NULL);
> +
>  	return 0;
>  }
>  
> @@ -993,3 +1025,20 @@ static int __init pvcalls_back_init(void)
>  	return 0;
>  }
>  module_init(pvcalls_back_init);
> +
> +static void __exit pvcalls_back_fin(void)
> +{
> +	struct pvcalls_fedata *priv, *npriv;
> +
> +	down(&pvcalls_back_global.frontends_lock);
> +	list_for_each_entry_safe(priv, npriv, &pvcalls_back_global.frontends,
> +				 list) {
> +		backend_disconnect(priv->dev);
> +	}
> +	up(&pvcalls_back_global.frontends_lock);
> +
> +	xenbus_unregister_driver(&pvcalls_back_driver);
> +	memset(&pvcalls_back_global, 0, sizeof(pvcalls_back_global));

Why?


Juergen

> +}
> +
> +module_exit(pvcalls_back_fin);
> 

[toc] | [next] | [standalone]

#1665338

On Tue, 13 Jun 2017, Juergen Gross wrote:
> On 02/06/17 21:31, Stefano Stabellini wrote:
> > Implement backend_disconnect. Call pvcalls_back_release_active on active
> > sockets and pvcalls_back_release_passive on passive sockets.
> > 
> > Implement module_exit by calling backend_disconnect on frontend
> > connections.
> > 
> > Signed-off-by: Stefano Stabellini <stefano@aporeto.com>
> > CC: boris.ostrovsky@oracle.com
> > CC: jgross@suse.com
> > ---
> >  drivers/xen/pvcalls-back.c | 49 ++++++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 49 insertions(+)
> > 
> > diff --git a/drivers/xen/pvcalls-back.c b/drivers/xen/pvcalls-back.c
> > index b541887..6afe7a0 100644
> > --- a/drivers/xen/pvcalls-back.c
> > +++ b/drivers/xen/pvcalls-back.c
> > @@ -800,6 +800,38 @@ static int backend_connect(struct xenbus_device *dev)
> >  
> >  static int backend_disconnect(struct xenbus_device *dev)
> >  {
> > +	struct pvcalls_fedata *priv;
> > +	struct sock_mapping *map, *n;
> > +	struct sockpass_mapping *mappass;
> > +	struct radix_tree_iter iter;
> > +	void **slot;
> > +
> > +
> > +	priv = dev_get_drvdata(&dev->dev);
> > +
> > +	list_for_each_entry_safe(map, n, &priv->socket_mappings, list) {
> > +		pvcalls_back_release_active(dev, priv, map);
> > +	}
> 
> You can drop the {}

OK


> > +
> > +	radix_tree_for_each_slot(slot, &priv->socketpass_mappings, &iter, 0) {
> > +		mappass = radix_tree_deref_slot(slot);
> > +		if (!mappass || radix_tree_exception(mappass)) {
> 
> This looks fishy.
> 
> You might call radix_tree_deref_retry(NULL). Right now this is okay,
> but you depend on the radix tree internals here.

I'll avoid id


> > +			if (radix_tree_deref_retry(mappass)) {
> > +				slot = radix_tree_iter_retry(&iter);
> > +				continue;
> 
> The continue; statement is pointless here.

I'll remove it


> > +			}
> > +		} else
> > +			pvcalls_back_release_passive(dev, priv, mappass);
> > +	}
> > +
> > +	xenbus_unmap_ring_vfree(dev, (void *)priv->sring);
> 
> Drop the cast.

OK


> > +	unbind_from_irqhandler(priv->irq, dev);
> > +
> > +	list_del(&priv->list);
> > +	destroy_workqueue(priv->wq);
> > +	kfree(priv);
> > +	dev_set_drvdata(&dev->dev, NULL);
> > +
> >  	return 0;
> >  }
> >  
> > @@ -993,3 +1025,20 @@ static int __init pvcalls_back_init(void)
> >  	return 0;
> >  }
> >  module_init(pvcalls_back_init);
> > +
> > +static void __exit pvcalls_back_fin(void)
> > +{
> > +	struct pvcalls_fedata *priv, *npriv;
> > +
> > +	down(&pvcalls_back_global.frontends_lock);
> > +	list_for_each_entry_safe(priv, npriv, &pvcalls_back_global.frontends,
> > +				 list) {
> > +		backend_disconnect(priv->dev);
> > +	}
> > +	up(&pvcalls_back_global.frontends_lock);
> > +
> > +	xenbus_unregister_driver(&pvcalls_back_driver);
> > +	memset(&pvcalls_back_global, 0, sizeof(pvcalls_back_global));
> 
> Why?

legacy code, I'll remove it


> Juergen
> 
> > +}
> > +
> > +module_exit(pvcalls_back_fin);
> > 
> 

[toc] | [prev] | [next] | [standalone]

#1665861

>>>  static int backend_disconnect(struct xenbus_device *dev)
>>>  {
>>> +	struct pvcalls_fedata *priv;
>>> +	struct sock_mapping *map, *n;
>>> +	struct sockpass_mapping *mappass;
>>> +	struct radix_tree_iter iter;
>>> +	void **slot;
>>> +
>>> +
>>> +	priv = dev_get_drvdata(&dev->dev);

Can you also rename priv to something else (like fedata)? And in other
routines too.

-boris

[toc] | [prev] | [next] | [standalone]

#1666161

On Wed, 14 Jun 2017, Boris Ostrovsky wrote:
> >>>  static int backend_disconnect(struct xenbus_device *dev)
> >>>  {
> >>> +	struct pvcalls_fedata *priv;
> >>> +	struct sock_mapping *map, *n;
> >>> +	struct sockpass_mapping *mappass;
> >>> +	struct radix_tree_iter iter;
> >>> +	void **slot;
> >>> +
> >>> +
> >>> +	priv = dev_get_drvdata(&dev->dev);
> 
> Can you also rename priv to something else (like fedata)? And in other
> routines too.

Yes, done

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web