On Thu, Feb 13, 2014 at 11:10:29AM -0800, Dmitry Torokhov wrote:

On Thursday, February 13, 2014 10:56:25 AM Christopher Heiny wrote:

quoted

On 02/12/2014 10:36 PM, Dmitry Torokhov wrote:

quoted

On Wed, Feb 05, 2014 at 05:36:09PM -0800, Christopher Heiny wrote:

quoted

quoted

On 02/05/2014 05:09 PM, Dmitry Torokhov wrote:

quoted

quoted

On Tue, Feb 04, 2014 at 03:08:12PM -0800, Christopher Heiny wrote:

quoted

quoted

quoted

quoted

On 01/23/2014 04:00 PM, Courtney Cavin wrote:

quoted

quoted

quoted

quoted

Cc: Christopher Heiny<redacted>
Cc: Dmitry Torokhov<dmitry.torokhov@gmail.com>
Signed-off-by: Courtney Cavin<redacted>
---

 drivers/input/rmi4/rmi_bus.c    |  4 ++--
 drivers/input/rmi4/rmi_bus.h    |  2 +-
 drivers/input/rmi4/rmi_driver.c | 17 ++++++++++++-----
 drivers/input/rmi4/rmi_f11.c    |  4 +++-
 4 files changed, 18 insertions(+), 9 deletions(-)

diff --git a/drivers/input/rmi4/rmi_bus.c

b/drivers/input/rmi4/rmi_bus.c
index 96a76e7..8a939f3 100644

--- a/drivers/input/rmi4/rmi_bus.c
+++ b/drivers/input/rmi4/rmi_bus.c

@@ -37,7 +37,7 @@ static void rmi_release_device(struct

device *dev)

 	kfree(rmi_dev);
 
 }

-struct device_type rmi_device_type = {
+static struct device_type rmi_device_type = {

 	.name		= "rmi_sensor",
 	.release	= rmi_release_device,
 
 };

This struct is used by diagnostic modules to identify sensor
devices, so it cannot be static.

Then we need to declare it somewhere or provide an accessor function.

Currently it's in a header not included in the patches.  We'll move
it to rmi_bus.h.

Hmm, we do have rmi_is_physical_device() to identify whether it is a
sensor or a function, so I believe we should mark all structures static
to avoid anyone poking at them.

I was poking around in the dependent code late last night and came to
the same conclusion.  I'll send a micropatch later today to get it out
of the way.

No need, I untangled relevant bits from the one Courtney sent.

Thanks.

Here it is by the way.

Input: synaptics-rmi4 - make device types module-private

From: Courtney Cavin <redacted>

Nobody outside of RMI bus core should be accessing rmi_device_type,
rmi_function_type or rmi_physical_driver structures so let's limit their
scope.

We provide rmi_is_physical_device() and rmi_is_function_device() to allow
determining whether one is dealing with sensor or function device.

Signed-off-by: Courtney Cavin <redacted>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/input/rmi4/rmi_bus.c    |    4 ++--
 drivers/input/rmi4/rmi_driver.c |    2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/input/rmi4/rmi_bus.c b/drivers/input/rmi4/rmi_bus.c
index 7efe7ed..6e0454a 100644
--- a/drivers/input/rmi4/rmi_bus.c
+++ b/drivers/input/rmi4/rmi_bus.c

@@ -38,7 +38,7 @@ static void rmi_release_device(struct device *dev)
 	kfree(rmi_dev);
 }
 
-struct device_type rmi_device_type = {
+static struct device_type rmi_device_type = {
 	.name		= "rmi_sensor",
 	.release	= rmi_release_device,
 };

@@ -154,7 +154,7 @@ static void rmi_release_function(struct device *dev)
 	kfree(fn);
 }
 
-struct device_type rmi_function_type = {
+static struct device_type rmi_function_type = {
 	.name		= "rmi_function",
 	.release	= rmi_release_function,
 };

diff --git a/drivers/input/rmi4/rmi_driver.c b/drivers/input/rmi4/rmi_driver.c
index 788343a..4406a7f 100644
--- a/drivers/input/rmi4/rmi_driver.c
+++ b/drivers/input/rmi4/rmi_driver.c

@@ -946,7 +946,7 @@ err_free_mem:
 	return retval < 0 ? retval : 0;
 }
 
-struct rmi_driver rmi_physical_driver = {
+static struct rmi_driver rmi_physical_driver = {
 	.driver = {
 		.owner	= THIS_MODULE,
 		.name	= "rmi_physical",

-- 
Dmitry