On Thu, 2015-05-07 at 10:39 +0200, Jiri Kosina wrote:
> On Tue, 5 May 2015, Srinivas Pandruvada wrote:
>
> > When CONFIG_DEBUG_LOCK_ALLOC is defined, mutex magic is compared and
> > warned for (l->magic != l), here l is the address of mutex pointer.
> > In hid-sensor-hub as part of hsdev creation, a per hsdev mutex is
> > initialized during MFD cell creation. This hsdev, which contains, mutex
> > is part of platform data for the a cell. But platform_data is copied
> > in platform_device_add_data() in platform.c. This copy will copy the
> > whole hsdev structure including mutex. But once copied the magic
> > will no longer match. So when client driver call
> > sensor_hub_input_attr_get_raw_value, this will trigger mutex warning.
> > So to avoid this use mutex pointer, which points to the original
> > mutex structure, and use this. This will be same even after copy.
> >
> > Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > ---
> > drivers/hid/hid-sensor-hub.c | 5 +++--
> > include/linux/hid-sensor-hub.h | 2 ++
> > 2 files changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/hid/hid-sensor-hub.c b/drivers/hid/hid-sensor-hub.c
> > index 5eb338d..2ee6a3f8 100644
> > --- a/drivers/hid/hid-sensor-hub.c
> > +++ b/drivers/hid/hid-sensor-hub.c
> > @@ -294,7 +294,7 @@ int sensor_hub_input_attr_get_raw_value(struct hid_sensor_hub_device *hsdev,
> > if (!report)
> > return -EINVAL;
> >
> > - mutex_lock(&hsdev->mutex);
> > + mutex_lock(hsdev->mutex_ptr);
> > if (flag == SENSOR_HUB_SYNC) {
> > memset(&hsdev->pending, 0, sizeof(hsdev->pending));
> > init_completion(&hsdev->pending.ready);
> > @@ -328,7 +328,7 @@ int sensor_hub_input_attr_get_raw_value(struct hid_sensor_hub_device *hsdev,
> > kfree(hsdev->pending.raw_data);
> > hsdev->pending.status = false;
> > }
> > - mutex_unlock(&hsdev->mutex);
> > + mutex_unlock(hsdev->mutex_ptr);
> >
> > return ret_val;
> > }
> > @@ -668,6 +668,7 @@ static int sensor_hub_probe(struct hid_device *hdev,
> > hsdev->product_id = hdev->product;
> > hsdev->usage = collection->usage;
> > mutex_init(&hsdev->mutex);
> > + hsdev->mutex_ptr = &hsdev->mutex;
> > hsdev->start_collection_index = i;
> > if (last_hsdev)
> > last_hsdev->end_collection_index = i;
> > diff --git a/include/linux/hid-sensor-hub.h b/include/linux/hid-sensor-hub.h
> > index d48e91f..6f27c6e 100644
> > --- a/include/linux/hid-sensor-hub.h
> > +++ b/include/linux/hid-sensor-hub.h
> > @@ -75,6 +75,7 @@ struct sensor_hub_pending {
> > * @start_collection_index: Starting index for a phy type collection
> > * @end_collection_index: Last index for a phy type collection
> > * @mutex: synchronizing mutex.
> > + * @mutex_ptr: Pointer to the above synchronizing mutex.
> > * @pending: Holds information of pending sync read request.
> > */
> > struct hid_sensor_hub_device {
> > @@ -85,6 +86,7 @@ struct hid_sensor_hub_device {
> > int start_collection_index;
> > int end_collection_index;
> > struct mutex mutex;
> > + struct mutex *mutex_ptr;
>
> This is quite ugly, isn't it?
>
> Is there any reason why you can't just have a pointer in the struct, and
> allocate the mutex dynamically at the same time you are allocating the
> struct?
I can. Just wanted to avoid another dynamic allocation.
I will resubmit.
Thanks,
Srinivas
> --
> Jiri Kosina
> SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
