[PATCH v3] driver core: take parent lock in async device attach

[PATCH v3] driver core: take parent lock in async device attach

[Johan Hovold]

From: Mario Peter <mario.peter@leica-geosystems.com>

On buses with need_parent_lock set (only USB), probe() must run with the
parent device locked. A synchronous attach after device_add() gets this
from the caller, e.g. usb_set_configuration() holds the udev lock while
adding interfaces. __device_attach_async_helper() runs the probe from an
async worker instead, where that lock isn't held, and only takes
device_lock(dev).

With async probing enabled for the hub driver (e.g. driver_async_probe=*
or driver_async_probe=hub), this races when a multi-TT hub is
configured. For multi-TT hubs, hub_probe() calls
usb_set_interface(hdev, 0, 1) to select the TT-per-port altsetting, and
usb_set_interface() creates the endpoint devices of the interface a
second time:

  usb_set_configuration()           async worker
    [udev locked]
    device_add(intf)
      queues the probe  ----------> __device_attach_async_helper()
                                      [intf locked, udev not locked]
                                      hub_probe()
                                        usb_set_interface(hdev, 0, 1)
    create_intf_ep_devs(intf)             create_intf_ep_devs(intf)

create_intf_ep_devs() checks and sets intf->ep_devs_created without a
lock of its own. If usb_set_configuration() has created ep_81 but not
yet set the flag, usb_set_interface() skips removing the endpoint
devices of the old altsetting and creates ep_81 again. With a
synchronous probe, hub_probe() runs inside device_add() with the udev
lock held, and usb_set_configuration() then finds ep_devs_created
already set.

Single-TT hubs don't call usb_set_interface() from hub_probe() and
don't show this warning, but their probe runs without the udev lock as
well.

On an i.MX8MM board with an onboard USB2514 hub this hit 11 of 100
boots:

  sysfs: cannot create duplicate filename '.../1-1/1-1:1.0/ep_81'
  Call trace:
   sysfs_warn_dup
   usb_create_ep_devs
   create_intf_ep_devs
   usb_set_interface
   hub_probe
   usb_probe_interface
   really_probe
   __device_attach_async_helper
   async_run_entry_fn

It also shows on v7.3-rc6 in QEMU, with dummy_hcd and a raw-gadget
program emulating a multi-TT hub: re-authorizing the hub in a loop,
while a SCHED_FIFO task preempts the writing task at random points for
3 ms, hit the warning in 29 of 6000 iterations.

Take the parent lock in __device_attach_async_helper() as well, like
__driver_attach_async_helper() does, and move
__device_driver_lock/unlock() up for that. With this the warning was
gone in 300 boots on the board and in 6000 iterations in QEMU.

Fixes: 765230b5f084 ("driver-core: add asynchronous probing support for drivers")
Assisted-by: LLM
Signed-off-by: Mario Peter <mario.peter@leica-geosystems.com>
[ johan: switch to async_schedule_dev_nocall() to avoid deadlock if
         async_schedule_dev() runs synchronously when parent lock is
	 held ]
Signed-off-by: Johan Hovold <johan@kernel.org>
---

Changes in v3:
 - fix deadlock if async_schedule_dev() runs synchronously

v2: https://lore.kernel.org/all/20261007131514.1260696-1-mario.peter@leica-geosystems.com/
v1: https://lore.kernel.org/all/20261006132335.4014316-1-mario.peter@leica-geosystems.com/


 drivers/base/dd.c | 85 ++++++++++++++++++++++++++---------------------
 1 file changed, 47 insertions(+), 38 deletions(-)

diff --git a/drivers/base/dd.c b/drivers/base/dd.c
index f6525a7ee8c5..9dc7cce9bd8d 100644
--- a/drivers/base/dd.c
+++ b/drivers/base/dd.c
@@ -1029,6 +1029,38 @@ static int __device_attach_driver(struct device_driver *drv, void *_data)
 	return ret == 0;
 }
 
+/*
+ * __device_driver_lock - acquire locks needed to manipulate dev->drv
+ * @dev: Device we will update driver info for
+ * @parent: Parent device. Needed if the bus requires parent lock
+ *
+ * This function will take the required locks for manipulating dev->drv.
+ * Normally this will just be the @dev lock, but when called for a USB
+ * interface, @parent lock will be held as well.
+ */
+static void __device_driver_lock(struct device *dev, struct device *parent)
+{
+	if (parent && dev->bus->need_parent_lock)
+		device_lock(parent);
+	device_lock(dev);
+}
+
+/*
+ * __device_driver_unlock - release locks needed to manipulate dev->drv
+ * @dev: Device we will update driver info for
+ * @parent: Parent device. Needed if the bus requires parent lock
+ *
+ * This function will release the required locks for manipulating dev->drv.
+ * Normally this will just be the @dev lock, but when called for a
+ * USB interface, @parent lock will be released as well.
+ */
+static void __device_driver_unlock(struct device *dev, struct device *parent)
+{
+	device_unlock(dev);
+	if (parent && dev->bus->need_parent_lock)
+		device_unlock(parent);
+}
+
 static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
 {
 	struct device *dev = _dev;
@@ -1038,7 +1070,7 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
 		.want_async	= true,
 	};
 
-	device_lock(dev);
+	__device_driver_lock(dev, dev->parent);
 
 	/*
 	 * Check if device has already been removed or claimed. This may
@@ -1060,7 +1092,7 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
 	if (dev->parent)
 		pm_runtime_put(dev->parent);
 out_unlock:
-	device_unlock(dev);
+	__device_driver_unlock(dev, dev->parent);
 
 	put_device(dev);
 }
@@ -1117,8 +1149,13 @@ static int __device_attach(struct device *dev, bool allow_async)
 	}
 out_unlock:
 	device_unlock(dev);
-	if (async)
-		async_schedule_dev(__device_attach_async_helper, dev);
+	if (async) {
+		if (!async_schedule_dev_nocall(__device_attach_async_helper, dev)) {
+			put_device(dev);
+			ret = 2;	/* retry synchronously */
+		}
+	}
+
 	return ret;
 }
 
@@ -1145,48 +1182,20 @@ EXPORT_SYMBOL_GPL(device_attach);
 void device_initial_probe(struct device *dev)
 {
 	struct subsys_private *sp = bus_to_subsys(dev->bus);
+	int ret;
 
 	if (!sp)
 		return;
 
-	if (sp->drivers_autoprobe)
-		__device_attach(dev, true);
+	if (sp->drivers_autoprobe) {
+		ret = __device_attach(dev, true);
+		if (ret == 2)
+			__device_attach(dev, false);
+	}
 
 	subsys_put(sp);
 }
 
-/*
- * __device_driver_lock - acquire locks needed to manipulate dev->drv
- * @dev: Device we will update driver info for
- * @parent: Parent device. Needed if the bus requires parent lock
- *
- * This function will take the required locks for manipulating dev->drv.
- * Normally this will just be the @dev lock, but when called for a USB
- * interface, @parent lock will be held as well.
- */
-static void __device_driver_lock(struct device *dev, struct device *parent)
-{
-	if (parent && dev->bus->need_parent_lock)
-		device_lock(parent);
-	device_lock(dev);
-}
-
-/*
- * __device_driver_unlock - release locks needed to manipulate dev->drv
- * @dev: Device we will update driver info for
- * @parent: Parent device. Needed if the bus requires parent lock
- *
- * This function will release the required locks for manipulating dev->drv.
- * Normally this will just be the @dev lock, but when called for a
- * USB interface, @parent lock will be released as well.
- */
-static void __device_driver_unlock(struct device *dev, struct device *parent)
-{
-	device_unlock(dev);
-	if (parent && dev->bus->need_parent_lock)
-		device_unlock(parent);
-}
-
 /**
  * device_driver_attach - attach a specific driver to a specific device
  * @drv: Driver to attach
-- 
2.55.0

Re: [PATCH v3] driver core: take parent lock in async device attach

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] driver-core: UAF on dev->parent in async device attach
--

commit c5f34060c20ab88f32e94e3a0ebf364f523925ee
Author: Mario Peter <mario.peter@leica-geosystems.com>

    driver core: take parent lock in async device attach
    
    On buses with need_parent_lock set (only USB), probe() must run with the
    parent device locked. A synchronous attach after device_add() gets this
    from the caller, e.g. usb_set_configuration() holds the udev lock while
    adding interfaces.

> diff --git a/drivers/base/dd.c b/drivers/base/dd.c
> index f6525a7ee8c54..9dc7cce9bd8d2 100644
> --- a/drivers/base/dd.c
> +++ b/drivers/base/dd.c
[ ... ]
> @@ -1029,6 +1029,38 @@ static int __device_attach_driver(struct device_driver *drv, void *_data)
>  	return ret == 0;
>  }
>  
> +/*
> + * __device_driver_lock - acquire locks needed to manipulate dev->drv
> + * @dev: Device we will update driver info for
> + * @parent: Parent device. Needed if the bus requires parent lock
> + *
> + * This function will take the required locks for manipulating dev->drv.
> + * Normally this will just be the @dev lock, but when called for a USB
> + * interface, @parent lock will be held as well.
> + */
> +static void __device_driver_lock(struct device *dev, struct device *parent)
> +{
> +	if (parent && dev->bus->need_parent_lock)
> +		device_lock(parent);

[Severity: High]
Could this acquire a mutex within freed parent memory?

If this is called with a dangling parent pointer from
__device_attach_async_helper(), this will dereference the freed memory.

[ ... ]
> @@ -1038,7 +1070,7 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
>  		.want_async	= true,
>  	};
>  
> -	device_lock(dev);
> +	__device_driver_lock(dev, dev->parent);

[Severity: High]
Can this cause a use-after-free on dev->parent?

In the device removal path, device_del() sets dev->p->dead to true via
kill_device() and later drops the child's reference to dev->parent via
put_device(dev->parent).

If the parent device is also being removed and its final reference is dropped,
the parent could be freed. 

If this async worker runs delayed, taking the lock before checking
dev->p->dead could pass a dangling dev->parent pointer. Prior to this patch,
device_lock(dev) was called, and dev->parent was never accessed if
dev->p->dead was true.

>  
>  	/*
>  	 * Check if device has already been removed or claimed. This may
> @@ -1060,7 +1092,7 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
>  	if (dev->parent)
>  		pm_runtime_put(dev->parent);
>  out_unlock:
> -	device_unlock(dev);
> +	__device_driver_unlock(dev, dev->parent);

[Severity: High]
If dev->parent was freed before the dev->p->dead check above, wouldn't this
unlock path also dereference the same dangling pointer?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261013140408.1226841-1-johan@kernel.org?part=1

Re: [PATCH v3] driver core: take parent lock in async device attach

[PETER Mario]

Hi Johan,

On 10/13/26 16:04, Johan Hovold wrote:
> @@ -1038,7 +1070,7 @@ static void __device_attach_async_helper(void *_dev, async_cookie_t cookie)
>                 .want_async     = true,
>         };
> 
> -       device_lock(dev);
> +       __device_driver_lock(dev, dev->parent);

Thanks for picking this up and for fixing the deadlock.

The use-after-free that the AI review flags here is real, I can
trigger it in QEMU. With a debug delay before this line,
driver_async_probe=usb,hub, and dummy_hcd unbound while the probe of
the emulated hub's usb_device waited, KASAN reports a
slab-use-after-free in __mutex_lock() called from
__device_attach_async_helper(), on the root hub. An interface pins its
usb_device, but a usb_device doesn't pin its parent hub, and dev drops
its reference to the parent in device_del(). v2 has the same problem.

Mario