[PATCH v2] usb: Avoid use-after-free by flushing endpoints early in usb_set_interface()

[PATCH v2] usb: Avoid use-after-free by flushing endpoints early in usb_set_interface()

[Mathias Nyman]

The steps taken by usb core to set a new interface is very different from
what is done on the xHC host side.

xHC hardware will do everything in one go. One command is used to set up
new endpoints, free old endpoints, check bandwidth, and run the new
endpoints.

All this is done by xHC when usb core asks the hcd to check for
available bandwidth. At this point usb core has not yet flushed the old
endpoints, which will cause use-after-free issues in xhci driver as
queued URBs are cancelled on a re-allocated endpoint.

To resolve this add a call to usb_disable_interface() which will flush
the endpoints before calling usb_hcd_alloc_bandwidth()

Additional checks in xhci driver will also be implemented to gracefully
handle stale URB cancel on freed and re-allocated endpoints

Cc: <stable@vger.kernel.org>
Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
---
  v2: update kerneldoc as well
---
 drivers/usb/core/message.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
index 228672f..bfa5eda 100644
--- a/drivers/usb/core/message.c
+++ b/drivers/usb/core/message.c
@@ -1341,6 +1341,11 @@ void usb_enable_interface(struct usb_device *dev,
  * is submitted that needs that bandwidth.  Some other operating systems
  * allocate bandwidth early, when a configuration is chosen.
  *
+ * xHCI reserves bandwidth and configures the alternate setting in
+ * usb_hcd_alloc_bandwidth(). If it fails the original interface altsetting
+ * may be disabled. Drivers cannot rely on any particular alternate
+ * setting being in effect after a failure.
+ *
  * This call is synchronous, and may not be used in an interrupt context.
  * Also, drivers must not change altsettings while urbs are scheduled for
  * endpoints in that interface; all such urbs must first be completed
@@ -1376,6 +1381,12 @@ int usb_set_interface(struct usb_device *dev, int interface, int alternate)
 			 alternate);
 		return -EINVAL;
 	}
+	/*
+	 * usb3 hosts configure the interface in usb_hcd_alloc_bandwidth,
+	 * including freeing dropped endpoint ring buffers.
+	 * Make sure the interface endpoints are flushed before that
+	 */
+	usb_disable_interface(dev, iface, false);
 
 	/* Make sure we have enough bandwidth for this alternate interface.
 	 * Remove the current alt setting and add the new alt setting.
-- 
2.7.4

[v2] usb: Avoid use-after-free by flushing endpoints early in usb_set_interface()

[Mathias Nyman]

The steps taken by usb core to set a new interface is very different from
what is done on the xHC host side.

xHC hardware will do everything in one go. One command is used to set up
new endpoints, free old endpoints, check bandwidth, and run the new
endpoints.

All this is done by xHC when usb core asks the hcd to check for
available bandwidth. At this point usb core has not yet flushed the old
endpoints, which will cause use-after-free issues in xhci driver as
queued URBs are cancelled on a re-allocated endpoint.

To resolve this add a call to usb_disable_interface() which will flush
the endpoints before calling usb_hcd_alloc_bandwidth()

Additional checks in xhci driver will also be implemented to gracefully
handle stale URB cancel on freed and re-allocated endpoints

Cc: <stable@vger.kernel.org>
Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
---
  v2: update kerneldoc as well
---
 drivers/usb/core/message.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
index 228672f..bfa5eda 100644
--- a/drivers/usb/core/message.c
+++ b/drivers/usb/core/message.c
@@ -1341,6 +1341,11 @@ void usb_enable_interface(struct usb_device *dev,
  * is submitted that needs that bandwidth.  Some other operating systems
  * allocate bandwidth early, when a configuration is chosen.
  *
+ * xHCI reserves bandwidth and configures the alternate setting in
+ * usb_hcd_alloc_bandwidth(). If it fails the original interface altsetting
+ * may be disabled. Drivers cannot rely on any particular alternate
+ * setting being in effect after a failure.
+ *
  * This call is synchronous, and may not be used in an interrupt context.
  * Also, drivers must not change altsettings while urbs are scheduled for
  * endpoints in that interface; all such urbs must first be completed
@@ -1376,6 +1381,12 @@ int usb_set_interface(struct usb_device *dev, int interface, int alternate)
 			 alternate);
 		return -EINVAL;
 	}
+	/*
+	 * usb3 hosts configure the interface in usb_hcd_alloc_bandwidth,
+	 * including freeing dropped endpoint ring buffers.
+	 * Make sure the interface endpoints are flushed before that
+	 */
+	usb_disable_interface(dev, iface, false);
 
 	/* Make sure we have enough bandwidth for this alternate interface.
 	 * Remove the current alt setting and add the new alt setting.

Re: [PATCH v2] usb: Avoid use-after-free by flushing endpoints early in usb_set_interface()

[Alan Stern]

On Mon, 3 Sep 2018, Mathias Nyman wrote:

> The steps taken by usb core to set a new interface is very different from
> what is done on the xHC host side.
> 
> xHC hardware will do everything in one go. One command is used to set up
> new endpoints, free old endpoints, check bandwidth, and run the new
> endpoints.
> 
> All this is done by xHC when usb core asks the hcd to check for
> available bandwidth. At this point usb core has not yet flushed the old
> endpoints, which will cause use-after-free issues in xhci driver as
> queued URBs are cancelled on a re-allocated endpoint.
> 
> To resolve this add a call to usb_disable_interface() which will flush
> the endpoints before calling usb_hcd_alloc_bandwidth()
> 
> Additional checks in xhci driver will also be implemented to gracefully
> handle stale URB cancel on freed and re-allocated endpoints
> 
> Cc: <stable@vger.kernel.org>
> Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
> Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
> ---
>   v2: update kerneldoc as well
> ---
>  drivers/usb/core/message.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
> index 228672f..bfa5eda 100644
> --- a/drivers/usb/core/message.c
> +++ b/drivers/usb/core/message.c
> @@ -1341,6 +1341,11 @@ void usb_enable_interface(struct usb_device *dev,
>   * is submitted that needs that bandwidth.  Some other operating systems
>   * allocate bandwidth early, when a configuration is chosen.
>   *
> + * xHCI reserves bandwidth and configures the alternate setting in
> + * usb_hcd_alloc_bandwidth(). If it fails the original interface altsetting
> + * may be disabled. Drivers cannot rely on any particular alternate
> + * setting being in effect after a failure.
> + *
>   * This call is synchronous, and may not be used in an interrupt context.
>   * Also, drivers must not change altsettings while urbs are scheduled for
>   * endpoints in that interface; all such urbs must first be completed
> @@ -1376,6 +1381,12 @@ int usb_set_interface(struct usb_device *dev, int interface, int alternate)
>  			 alternate);
>  		return -EINVAL;
>  	}
> +	/*
> +	 * usb3 hosts configure the interface in usb_hcd_alloc_bandwidth,
> +	 * including freeing dropped endpoint ring buffers.
> +	 * Make sure the interface endpoints are flushed before that
> +	 */
> +	usb_disable_interface(dev, iface, false);
>  
>  	/* Make sure we have enough bandwidth for this alternate interface.
>  	 * Remove the current alt setting and add the new alt setting.

Acked-by: Alan Stern <stern@rowland.harvard.edu>

[v2] usb: Avoid use-after-free by flushing endpoints early in usb_set_interface()

[Alan Stern]

On Mon, 3 Sep 2018, Mathias Nyman wrote:

> The steps taken by usb core to set a new interface is very different from
> what is done on the xHC host side.
> 
> xHC hardware will do everything in one go. One command is used to set up
> new endpoints, free old endpoints, check bandwidth, and run the new
> endpoints.
> 
> All this is done by xHC when usb core asks the hcd to check for
> available bandwidth. At this point usb core has not yet flushed the old
> endpoints, which will cause use-after-free issues in xhci driver as
> queued URBs are cancelled on a re-allocated endpoint.
> 
> To resolve this add a call to usb_disable_interface() which will flush
> the endpoints before calling usb_hcd_alloc_bandwidth()
> 
> Additional checks in xhci driver will also be implemented to gracefully
> handle stale URB cancel on freed and re-allocated endpoints
> 
> Cc: <stable@vger.kernel.org>
> Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
> Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
> ---
>   v2: update kerneldoc as well
> ---
>  drivers/usb/core/message.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
> index 228672f..bfa5eda 100644
> --- a/drivers/usb/core/message.c
> +++ b/drivers/usb/core/message.c
> @@ -1341,6 +1341,11 @@ void usb_enable_interface(struct usb_device *dev,
>   * is submitted that needs that bandwidth.  Some other operating systems
>   * allocate bandwidth early, when a configuration is chosen.
>   *
> + * xHCI reserves bandwidth and configures the alternate setting in
> + * usb_hcd_alloc_bandwidth(). If it fails the original interface altsetting
> + * may be disabled. Drivers cannot rely on any particular alternate
> + * setting being in effect after a failure.
> + *
>   * This call is synchronous, and may not be used in an interrupt context.
>   * Also, drivers must not change altsettings while urbs are scheduled for
>   * endpoints in that interface; all such urbs must first be completed
> @@ -1376,6 +1381,12 @@ int usb_set_interface(struct usb_device *dev, int interface, int alternate)
>  			 alternate);
>  		return -EINVAL;
>  	}
> +	/*
> +	 * usb3 hosts configure the interface in usb_hcd_alloc_bandwidth,
> +	 * including freeing dropped endpoint ring buffers.
> +	 * Make sure the interface endpoints are flushed before that
> +	 */
> +	usb_disable_interface(dev, iface, false);
>  
>  	/* Make sure we have enough bandwidth for this alternate interface.
>  	 * Remove the current alt setting and add the new alt setting.

Acked-by: Alan Stern <stern@rowland.harvard.edu>