Re: [PATCH] input synaptics-rmi4: Use put_device() and device_type.release() to free storage.

On 02/11/2014 05:59 PM, Courtney Cavin wrote:

On Wed, Feb 12, 2014 at 12:13:30AM +0100, Christopher Heiny wrote:

quoted

For rmi_sensor and rmi_function device_types, use put_device() and
the assocated device_type.release() function to clean up related
structures and storage in the correct and safe order.

Signed-off-by: Christopher Heiny <redacted>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: Benjamin Tissoires <redacted>
Cc: Linux Walleij <redacted>
Cc: David Herrmann <redacted>
Cc: Jiri Kosina <redacted>
Cc: Courtney Cavin <redacted>

I'm not a huge fan of you taking my patches, re-formatting them and
sending them as your own.  More out of principle then actually caring
about ownership.  You at least cc'd me on this one....

Sorry - no slight was intended at all!  I wasn't sure what the protocol 
was for picking up an idea from someone else's patch and building on 
that idea, so I just went with the CC.  I definitely prefer to attribute 
sources correctly - if you could clarify what should be done (beyond the 
CC) to acknowledge the author of the original patch, I'd appreciate it.

quoted

---

  drivers/input/rmi4/rmi_bus.c    | 65 +++++++++++++++--------------------------
  drivers/input/rmi4/rmi_driver.c | 11 ++-----
  2 files changed, 25 insertions(+), 51 deletions(-)

diff --git a/drivers/input/rmi4/rmi_bus.c b/drivers/input/rmi4/rmi_bus.c
index 96a76e7..1b9ad80 100644
--- a/drivers/input/rmi4/rmi_bus.c
+++ b/drivers/input/rmi4/rmi_bus.c

[...]

quoted

@@ -185,6 +153,23 @@ static void rmi_function_teardown_debugfs(struct rmi_function *fn)
  }

  #endif
+static void rmi_release_function(struct device *dev)
+{
+	struct rmi_function *fn = to_rmi_function(dev);
+	rmi_function_teardown_debugfs(fn);
+	kfree(fn->irq_mask);

If you are going to do this, then you need to remove the other call to
free this mask in rmi_free_function_list().

Okidoki.

quoted

+	kfree(fn);
+}

-Courtney