ALPS DualPoint double click bug

Commit Message

Hans de Goede July 30, 2015, 3 p.m. UTC

Hi,

On 30-07-15 16:46, Pali Rohár wrote:
> On Thursday 30 July 2015 16:17:23 Hans de Goede wrote:
>> Hi,
>>
>> On 28-07-15 01:38, Douglas Christman wrote:
>>> On Mon, Jul 27, 2015 at 12:40 PM, Hans de Goede <[e-mail]> wrote:
>>>>
>>>> OK, please let me know soon, I would like to get to the bottom
>>>> of this, and knowing the exact commit causing the problem will
>>>> help a lot.
>>>>
>>>> Regards,
>>>>
>>>> hans
>>>
>>> I've verified that reverting 92bac83d on a clean v4.1 kernel
>>> (b953c0d2) resolves the issue.
>>
>> Thanks,
>>
>> Can you please apply the attached patch on a clean v4.1 kernel,
>> and confirm that that fixes this ?
>>
>> Regards,
>>
>> Hans
>
>>  From ee3d5d5a298b178ae5284b9766ca849665a37670 Mon Sep 17 00:00:00 2001
>> From: Hans de Goede <[e-mail]>
>> Date: Thu, 30 Jul 2015 15:49:16 +0200
>> Subject: [PATCH] alps: Only Dell laptops have separate button bits for v2
>>   dualpoint sticks
>>
>> It turns out that only Dell laptops have the separate button bits for
>> v2 dualpoint sticks and that commit 92bac83dd79e ("Input: alps - non
>> interleaved V2 dualpoint has separate stick button bits") causes
>> regressions on Toshiba laptops.
>>
>> This commit adds a check for Dell laptops to the code for handling these
>> extra button bits, fixing this regression.
>>
>> This patch has been tested on a Dell Latitude D620 to make sure that it
>> does not reintroduce the original problem.
>>
>> Reported-by: Douglas Christman <[e-mail]>
>> Cc: [e-mail]
>> Signed-off-by: Hans de Goede <[e-mail]>
>> ---
>>   drivers/input/mouse/alps.c | 6 ++++--
>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/input/mouse/alps.c b/drivers/input/mouse/alps.c
>> index 113d6f1..889aec1 100644
>> --- a/drivers/input/mouse/alps.c
>> +++ b/drivers/input/mouse/alps.c
>> @@ -20,6 +20,7 @@
>>   #include <linux/input/mt.h>
>>   #include <linux/serio.h>
>>   #include <linux/libps2.h>
>> +#include <linux/dmi.h>
>>
>>   #include "psmouse.h"
>>   #include "alps.h"
>> @@ -251,8 +252,9 @@ static void alps_process_packet_v1_v2(struct psmouse *psmouse)
>>   		return;
>>   	}
>>
>> -	/* Non interleaved V2 dualpoint has separate stick button bits */
>> -	if (priv->proto_version == ALPS_PROTO_V2 &&
>> +	/* Dell non interleaved V2 dualpoint has separate stick button bits */
>> +	if (dmi_name_in_vendors("Dell") &&
>> +	    priv->proto_version == ALPS_PROTO_V2 &&
>>   	    priv->flags == (ALPS_PASS | ALPS_DUALPOINT)) {
>>   		left |= packet[0] & 1;
>>   		right |= packet[0] & 2;
>
> What about introducing new flag ALPS_<something> instead calling
> dmi_name_in_vendors() function every time when we need to process
> packet?

That is a good idea. Douglas can you test the attached version
instead of the previous one please ?

Thanks & Regards,

Hans

Comments

Pali Rohár July 30, 2015, 3:49 p.m. UTC | #1

On Thursday 30 July 2015 17:00:56 Hans de Goede wrote:
> Hi,
> 
> On 30-07-15 16:46, Pali Rohár wrote:
> >What about introducing new flag ALPS_<something> instead calling
> >dmi_name_in_vendors() function every time when we need to process
> >packet?
> 
> That is a good idea. Douglas can you test the attached version
> instead of the previous one please ?
> 
> Thanks & Regards,
> 
> Hans

> @@ -251,9 +253,9 @@ static void alps_process_packet_v1_v2(struct psmouse *psmouse)
>  		return;
>  	}
>  
> -	/* Non interleaved V2 dualpoint has separate stick button bits */
> +	/* Dell non interleaved V2 dualpoint has separate stick button bits */
>  	if (priv->proto_version == ALPS_PROTO_V2 &&
> -	    priv->flags == (ALPS_PASS | ALPS_DUALPOINT)) {
> +	    priv->flags == (ALPS_DELL | ALPS_PASS | ALPS_DUALPOINT)) {

Hi again. Now I'm trying to understand what this condition means and you
probably wanted to write... priv->flags is field and so == comparator is
hard to decode and understood. Now it means that priv->flags must have
set ALPS_DELL, ALPS_PASS and ALPS_DUALPOINT and must not set ALPS_WHEEL,
ALPS_FW_BK_1, ALPS_FW_BK_2, ALPS_FOUR_BUTTONS, ALPS_PS2_INTERLEAVED,
ALPS_BUTTONPAD and all other future flags! With future flags this code
is fragile and can be easy broken in future (by introducing new flags).
Because of "Non interleaved" in description you probably wanted
something like this?

 if (priv->proto_version == ALPS_PROTO_V2 &&
     (priv->flags & (ALPS_DELL | ALPS_PASS | ALPS_DUALPOINT)) &&
     !(priv->flags & ALPS_PS2_INTERLEAVED))

(flags must contains ALPS_DELL, ALPS_PASS, ALPS_DUALPOINT and must not
ALPS_PS2_INTERLEAVED)

Hans de Goede July 31, 2015, 8:12 a.m. UTC | #2

Hi,

On 30-07-15 17:49, Pali Rohár wrote:
> On Thursday 30 July 2015 17:00:56 Hans de Goede wrote:
>> Hi,
>>
>> On 30-07-15 16:46, Pali Rohár wrote:
>>> What about introducing new flag ALPS_<something> instead calling
>>> dmi_name_in_vendors() function every time when we need to process
>>> packet?
>>
>> That is a good idea. Douglas can you test the attached version
>> instead of the previous one please ?
>>
>> Thanks & Regards,
>>
>> Hans
>
>> @@ -251,9 +253,9 @@ static void alps_process_packet_v1_v2(struct psmouse *psmouse)
>>   		return;
>>   	}
>>
>> -	/* Non interleaved V2 dualpoint has separate stick button bits */
>> +	/* Dell non interleaved V2 dualpoint has separate stick button bits */
>>   	if (priv->proto_version == ALPS_PROTO_V2 &&
>> -	    priv->flags == (ALPS_PASS | ALPS_DUALPOINT)) {
>> +	    priv->flags == (ALPS_DELL | ALPS_PASS | ALPS_DUALPOINT)) {
>
> Hi again. Now I'm trying to understand what this condition means and you
> probably wanted to write... priv->flags is field and so == comparator is
> hard to decode and understood.

The == operator was already being used before this patch, and it does the job
just fine IMHO.

Regards,

Hans


  Now it means that priv->flags must have
> set ALPS_DELL, ALPS_PASS and ALPS_DUALPOINT and must not set ALPS_WHEEL,
> ALPS_FW_BK_1, ALPS_FW_BK_2, ALPS_FOUR_BUTTONS, ALPS_PS2_INTERLEAVED,
> ALPS_BUTTONPAD and all other future flags! With future flags this code
> is fragile and can be easy broken in future (by introducing new flags).
> Because of "Non interleaved" in description you probably wanted
> something like this?
>
>   if (priv->proto_version == ALPS_PROTO_V2 &&
>       (priv->flags & (ALPS_DELL | ALPS_PASS | ALPS_DUALPOINT)) &&
>       !(priv->flags & ALPS_PS2_INTERLEAVED))
>
> (flags must contains ALPS_DELL, ALPS_PASS, ALPS_DUALPOINT and must not
> ALPS_PS2_INTERLEAVED)
>
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Douglas Christman July 31, 2015, 9:12 p.m. UTC | #3

On Thu, Jul 30, 2015 at 11:00 AM, Hans de Goede <[e-mail]> wrote:
> Hi,
>
> On 30-07-15 16:46, Pali Rohár wrote:
>> What about introducing new flag ALPS_<something> instead calling
>> dmi_name_in_vendors() function every time when we need to process
>> packet?
>
> That is a good idea. Douglas can you test the attached version
> instead of the previous one please ?
>
> Thanks & Regards,
>
> Hans

This patch works for me.

Doug
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Pali Rohár July 31, 2015, 9:17 p.m. UTC | #4

On Friday 31 July 2015 23:12:56 Douglas Christman wrote:
> On Thu, Jul 30, 2015 at 11:00 AM, Hans de Goede <[e-mail]>
> wrote:
> > Hi,
> > 
> > On 30-07-15 16:46, Pali Rohár wrote:
> >> What about introducing new flag ALPS_<something> instead calling
> >> dmi_name_in_vendors() function every time when we need to process
> >> packet?
> > 
> > That is a good idea. Douglas can you test the attached version
> > instead of the previous one please ?
> > 
> > Thanks & Regards,
> > 
> > Hans
> 
> This patch works for me.
> 
> Doug

Douglas: Thanks for testing!

Hans: Can you update also documentation in Documentation/input/alps.txt 
to match with this your change?

Hans de Goede Aug. 1, 2015, 6:48 a.m. UTC | #5

Hi,

On 31-07-15 23:17, Pali Rohár wrote:
> On Friday 31 July 2015 23:12:56 Douglas Christman wrote:
>> On Thu, Jul 30, 2015 at 11:00 AM, Hans de Goede <[e-mail]>
>> wrote:
>>> Hi,
>>>
>>> On 30-07-15 16:46, Pali Rohár wrote:
>>>> What about introducing new flag ALPS_<something> instead calling
>>>> dmi_name_in_vendors() function every time when we need to process
>>>> packet?
>>>
>>> That is a good idea. Douglas can you test the attached version
>>> instead of the previous one please ?
>>>
>>> Thanks & Regards,
>>>
>>> Hans
>>
>> This patch works for me.
>>
>> Doug
>
> Douglas: Thanks for testing!

Ack, thanks for testing :)

> Hans: Can you update also documentation in Documentation/input/alps.txt
> to match with this your change?

Good point, I've fixed this up and submitted the patch fixing
this upstream.

Regards,

Hans
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

From 93a1852f32ab37b2cbcdb7e797f1ad59eb6bbef4 Mon Sep 17 00:00:00 2001
From: Hans de Goede <[e-mail]>
Date: Thu, 30 Jul 2015 15:49:16 +0200
Subject: [PATCH] alps: Only Dell laptops have separate button bits for v2
 dualpoint sticks

It turns out that only Dell laptops have the separate button bits for
v2 dualpoint sticks and that commit 92bac83dd79e ("Input: alps - non
interleaved V2 dualpoint has separate stick button bits") causes
regressions on Toshiba laptops.

This commit adds a check for Dell laptops to the code for handling these
extra button bits, fixing this regression.

This patch has been tested on a Dell Latitude D620 to make sure that it
does not reintroduce the original problem.

Reported-by: Douglas Christman <[e-mail]>
Cc: [e-mail]
Signed-off-by: Hans de Goede <[e-mail]>
---
 drivers/input/mouse/alps.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/input/mouse/alps.c b/drivers/input/mouse/alps.c
index 113d6f1..4d24686 100644
--- a/drivers/input/mouse/alps.c
+++ b/drivers/input/mouse/alps.c
@@ -20,6 +20,7 @@ 
 #include <linux/input/mt.h>
 #include <linux/serio.h>
 #include <linux/libps2.h>
+#include <linux/dmi.h>
 
 #include "psmouse.h"
 #include "alps.h"
@@ -99,6 +100,7 @@  static const struct alps_nibble_commands alps_v6_nibble_commands[] = {
 #define ALPS_FOUR_BUTTONS	0x40	/* 4 direction button present */
 #define ALPS_PS2_INTERLEAVED	0x80	/* 3-byte PS/2 packet interleaved with
 					   6-byte ALPS packet */
+#define ALPS_DELL		0x100	/* device is a Dell laptop */
 #define ALPS_BUTTONPAD		0x200	/* device is a clickpad */
 
 static const struct alps_model_info alps_model_data[] = {
@@ -251,9 +253,9 @@  static void alps_process_packet_v1_v2(struct psmouse *psmouse)
 		return;
 	}
 
-	/* Non interleaved V2 dualpoint has separate stick button bits */
+	/* Dell non interleaved V2 dualpoint has separate stick button bits */
 	if (priv->proto_version == ALPS_PROTO_V2 &&
-	    priv->flags == (ALPS_PASS | ALPS_DUALPOINT)) {
+	    priv->flags == (ALPS_DELL | ALPS_PASS | ALPS_DUALPOINT)) {
 		left |= packet[0] & 1;
 		right |= packet[0] & 2;
 		middle |= packet[0] & 4;
@@ -2550,6 +2552,8 @@  static int alps_set_protocol(struct psmouse *psmouse,
 	priv->byte0 = protocol->byte0;
 	priv->mask0 = protocol->mask0;
 	priv->flags = protocol->flags;
+	if (dmi_name_in_vendors("Dell"))
+		priv->flags |= ALPS_DELL;
 
 	priv->x_max = 2000;
 	priv->y_max = 1400;
-- 
2.4.3