On Thu, Oct 09, 2014 at 05:21:58PM -0700, Andrew Duggan wrote:

On 10/09/2014 10:28 AM, Dmitry Torokhov wrote:

quoted

On Thu, Oct 09, 2014 at 09:52:46AM -0700, Dmitry Torokhov wrote:

quoted

On Thu, Oct 09, 2014 at 11:34:26AM +0200, Nicole Faerber wrote:

quoted

Hi!
Just installed the just released 3.17 kernel and found a bad behavior of
the new Synaptics driver on my Thinkpad Yoga which has the new Synaptics
clickpad enabled touchpad:

psmouse serio1: synaptics: Touchpad model: 1, fw: 8.1, id: 0x1e2b1,
caps: 0xd002a3/0x940300/0x12f800, board id: 2911, fw id: 2560

The issue is that now a button release is only issued after the finger
has completely left the touchpad and not when releasing the physical
button. Is this physical button now called FORCEPAD? Anyway, this is
pretty annoying. Double clicking become a real pain.
I did comment out the new

if (SYN_CAP_FORCEPAD(priv->ext_cap_0c)) {
...

and everything is back to normal again, i.e. when I do release the pad
physical button but keep the finger on the pad, the button release event
is issued properly.

Hmm, the forcepad code should only activate if the devoice do4es not
have physical buttons at all. Let me see what's the diffference in
capabilities between your and mine touchpads...

OK, so your extended caps are 0x12f800 while on my forcepad they are
0x12e800. The forcepad bit is supposed to be bit 15, so it is set for
both our devices, but bit 12 (counting from 0) is different.

Andrew, Chris, could you please tell us what bit 12 indicates? In fact,
if you could share the updated description for all currently defined
capability bits that would be awesome.

Thanks!

Hmm, looks like I got incorrect information about the ForcePad
capabilities bit and unfortunately there does not seem to be a
capabilities bit for ForcePad on PS/2. Too bad that wasn't caught
before 3.17 was released. Bit 12 is for "uniform clickpad" which
means that the whole clickpad moves when you press it as opposed to
it being hinged at the top. That makes sense that a ForcePad would
not have that capability.

Argh, that is unfortunate. Well, I guess we are back to DMI checks.

Thanks.

-- 
Dmitry

Input: synaptics - gate forcepad support by DMI check

From: Dmitry Torokhov <dmitry.torokhov@gmail.com>

Unfortunately, ForcePad capability is not actually exported over PS/2, so
we have to resort to DMI checks.

Cc: stable@vger.kernel.org
Reported-by: Nicole Faerber <redacted>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/input/mouse/synaptics.c |   22 +++++++++++++++++++++-
 drivers/input/mouse/synaptics.h |    8 ++------
 2 files changed, 23 insertions(+), 7 deletions(-)

diff --git a/drivers/input/mouse/synaptics.c b/drivers/input/mouse/synaptics.c
index 6394d9b..9031a0a 100644
--- a/drivers/input/mouse/synaptics.c
+++ b/drivers/input/mouse/synaptics.c

@@ -607,6 +607,8 @@ static void synaptics_parse_agm(const unsigned char buf[],
 	priv->agm_pending = true;
 }
 
+static bool is_forcepad;
+
 static int synaptics_parse_hw_state(const unsigned char buf[],
 				    struct synaptics_data *priv,
 				    struct synaptics_hw_state *hw)

@@ -636,7 +638,7 @@ static int synaptics_parse_hw_state(const unsigned char buf[],
 		hw->left  = (buf[0] & 0x01) ? 1 : 0;
 		hw->right = (buf[0] & 0x02) ? 1 : 0;
 
-		if (SYN_CAP_FORCEPAD(priv->ext_cap_0c)) {
+		if (is_forcepad) {
 			/*
 			 * ForcePads, like Clickpads, use middle button
 			 * bits to report primary button clicks.

@@ -1667,11 +1669,29 @@ static const struct dmi_system_id __initconst cr48_dmi_table[] = {
 	{ }
 };
 
+static const struct dmi_system_id forcepad_dmi_table[] __initconst = {
+#if defined(CONFIG_DMI) && defined(CONFIG_X86)
+	{
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook Folio 1040 G1"),
+		},
+	},
+#endif
+	{ }
+};
+
 void __init synaptics_module_init(void)
 {
 	impaired_toshiba_kbc = dmi_check_system(toshiba_dmi_table);
 	broken_olpc_ec = dmi_check_system(olpc_dmi_table);
 	cr48_profile_sensor = dmi_check_system(cr48_dmi_table);
+
+	/*
+	 * Unfortunately ForcePad capability is not exported over PS/2,
+	 * so we have to resort to checking DMI.
+	 */
+	is_forcepad = dmi_check_system(forcepad_dmi_table);
 }
 
 static int __synaptics_init(struct psmouse *psmouse, bool absolute_mode)

diff --git a/drivers/input/mouse/synaptics.h b/drivers/input/mouse/synaptics.h
index fb2e076..1bd01f2 100644
--- a/drivers/input/mouse/synaptics.h
+++ b/drivers/input/mouse/synaptics.h

@@ -77,12 +77,9 @@
  *					for noise.
  * 2	0x08	image sensor		image sensor tracks 5 fingers, but only
  *					reports 2.
+ * 2	0x01	uniform clickpad	whole clickpad moves instead of being
+ *					hinged at the top.
  * 2	0x20	report min		query 0x0f gives min coord reported
- * 2	0x80	forcepad		forcepad is a variant of clickpad that
- *					does not have physical buttons but rather
- *					uses pressure above certain threshold to
- *					report primary clicks. Forcepads also have
- *					clickpad bit set.
  */
 #define SYN_CAP_CLICKPAD(ex0c)		((ex0c) & 0x100000) /* 1-button ClickPad */
 #define SYN_CAP_CLICKPAD2BTN(ex0c)	((ex0c) & 0x000100) /* 2-button ClickPad */

@@ -91,7 +88,6 @@
 #define SYN_CAP_ADV_GESTURE(ex0c)	((ex0c) & 0x080000)
 #define SYN_CAP_REDUCED_FILTERING(ex0c)	((ex0c) & 0x000400)
 #define SYN_CAP_IMAGE_SENSOR(ex0c)	((ex0c) & 0x000800)
-#define SYN_CAP_FORCEPAD(ex0c)		((ex0c) & 0x008000)
 
 /* synaptics modes query bits */
 #define SYN_MODE_ABSOLUTE(m)		((m) & (1 << 7))