qt1070: Why IRQF_TRIGGER_NONE?

Hi,
I was testing qt1070 driver with my Visstrim SM10 board when I found
out IRQs where not triggered properly.

I had to come up with the following patch in order to solve it:

diff --git a/drivers/input/keyboard/qt1070.c b/drivers/input/keyboard/qt1070.c
index 0b7b2f8..e8b6ae3 100644
--- a/drivers/input/keyboard/qt1070.c
+++ b/drivers/input/keyboard/qt1070.c

@@ -201,7 +201,7 @@ static int __devinit qt1070_probe(struct i2c_client *client,
 	msleep(QT1070_RESET_TIME);

 	err = request_threaded_irq(client->irq, NULL, qt1070_interrupt,
-		IRQF_TRIGGER_NONE, client->dev.driver->name, data);
+		IRQF_TRIGGER_FALLING, client->dev.driver->name, data);
 	if (err) {
 		dev_err(&client->dev, "fail to request irq\n");
 		goto err_free_mem;

According to the datasheet (p10)  "The CHANGE line is active low and
signals when there is a change in state in the Detection or
Input status bytes. It is cleared (allowed to float high) when the
host reads the status bytes." As I understand, this means that we have
to monitor falling edges here.

Is there any reason to register the IRQ as IRQF_TRIGGER_NONE? Would
you accept the above patch instead, provided I send it in the required
format?

Regards.
-- 
Javier Martin
Vista Silicon S.L.
CDTUC - FASE C - Oficina S-345
Avda de los Castros s/n
39005- Santander. Cantabria. Spain
+34 942 25 32 60
www.vista-silicon.com

On 5/2/2012 4:35 PM, javier Martin wrote:

I add Voice Shen to the mail loop, who is writing above code.

Voice, will that IRQF_TRIGGER_XXXX changes impact other platforms?

Best Regards,
Josh Wu

Regards.

Hi All,
  Some information as following,

  According to the datasheet of qt1070, we can use IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW (I think this is the best) for IRQ flag. However, the IRQ line is a GPIO of a SOC. Some SOC can detect the level change of the GPIO, while can not distinguish the falling or rising. So, the IRQ flag depends on the trigger mode of GPIO line.

  Maybe use the "flags" element in "struct i2c_board_info" to pass the IRQ flag, or add another element named "irqflags" into "struct i2c_board_info". I think this will be better, but I am not sure whether this is a good solution.

  BTW, I am reworking the driver for DT (device tree) support.

Best Regards,
Voice Shen

-----Original Message-----
From: Josh Wu [mailto:josh.wu@atmel.com] 
Sent: Thursday, May 03, 2012 9:41 申波
To: javier Martin; Shen, Voice
Cc: linux-input@vger.kernel.org; Wolfram Sang; Jean Delvare; Axel Lin; Dmitry Torokhov
Subject: Re: qt1070: Why IRQF_TRIGGER_NONE?

On 5/2/2012 4:35 PM, javier Martin wrote:

I add Voice Shen to the mail loop, who is writing above code.

Voice, will that IRQF_TRIGGER_XXXX changes impact other platforms?

Best Regards,
Josh Wu

Regards.

On Thu, 3 May 2012 05:15:14 +0000, Shen, Voice wrote:

Hi All,
  Some information as following,

  According to the datasheet of qt1070, we can use IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW (I think this is the best) for IRQ flag. However, the IRQ line is a GPIO of a SOC. Some SOC can detect the level change of the GPIO, while can not distinguish the falling or rising. So, the IRQ flag depends on the trigger mode of GPIO line.

  Maybe use the "flags" element in "struct i2c_board_info" to pass the IRQ flag, or add another element named "irqflags" into "struct i2c_board_info". I think this will be better, but I am not sure whether this is a good solution.

i2c_board_info.flags is for I2C client flags, please do not abuse it
for IRQ information.

I have no objection to an irq_flag member being added, however I
remember past discussions where people argued whether it was the right
thing to do or whether the IRQ mode was best set by platform
initialization code. Part of that discussion was archived here:
http://marc.info/?t=128743170300002&r=1&w=2
Said discussion did not result in any code being merged as I don't
think we came to an agreement. Feel free to restart the discussion with
the interested people.

-- 
Jean Delvare

Hi Jean,
  Thank you for your information. Although the irq_flag can not be added into "struct i2c_board_info" till now, I will try to find other solution.

  Thanks again.

Hi Javier,
  As to the IRQ flag depends on SOC. We try to find other solution. Please describe you issue in detail. And what's the mode does your SOC support.

Best Regards,
Voice Shen
-----Original Message-----
From: Jean Delvare [mailto:khali@linux-fr.org] 
Sent: Thursday, May 03, 2012 14:15 申波
To: Shen, Voice
Cc: Wu, Josh; javier Martin; linux-input@vger.kernel.org; Wolfram Sang; Axel Lin; Dmitry Torokhov
Subject: Re: qt1070: Why IRQF_TRIGGER_NONE?

On Thu, 3 May 2012 05:15:14 +0000, Shen, Voice wrote:

Hi All,
  Some information as following,

  According to the datasheet of qt1070, we can use IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW (I think this is the best) for IRQ flag. However, the IRQ line is a GPIO of a SOC. Some SOC can detect the level change of the GPIO, while can not distinguish the falling or rising. So, the IRQ flag depends on the trigger mode of GPIO line.

  Maybe use the "flags" element in "struct i2c_board_info" to pass the IRQ flag, or add another element named "irqflags" into "struct i2c_board_info". I think this will be better, but I am not sure whether this is a good solution.

i2c_board_info.flags is for I2C client flags, please do not abuse it
for IRQ information.

I have no objection to an irq_flag member being added, however I
remember past discussions where people argued whether it was the right
thing to do or whether the IRQ mode was best set by platform
initialization code. Part of that discussion was archived here:
http://marc.info/?t=128743170300002&r=1&w=2
Said discussion did not result in any code being merged as I don't
think we came to an agreement. Feel free to restart the discussion with
the interested people.

-- 
Jean Delvare

Hi all,
thank you for your comments.

On 4 May 2012 04:07, Shen, Voice [off-list ref] wrote:

Hi Jean,
 Thank you for your information. Although the irq_flag can not be added into "struct i2c_board_info" till now, I will try to find other solution.

 Thanks again.

Hi Javier,
 As to the IRQ flag depends on SOC. We try to find other solution. Please describe you issue in detail. And what's the mode does your SOC support.

As I stated, with the current 'IRQF_TRIGGER_NONE' flag, this driver
doesn't work in i.MX27. Furthermore, I don't think it works on any
platform since as you previously pointed according to the datasheet of
qt1070, we can use either IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW.
But I don't know what is the sense of using 'IRQF_TRIGGER_NONE' here.

My SoC, an i.MX27, theoretically supports both RQF_TRIGGER_FALLING and
IRQF_TRIGGER_LOW but I've only tested the former.

If we take a look at a similar device, a pca953x they simply use the
following flags: (IRQF_TRIGGER_LOW | IRQF_ONESHOT)
http://lxr.linux.no/#linux+v3.3.4/drivers/gpio/gpio-pca953x.c#L495

We could do the same here. It's true there can be SoC's which doesn't
support IRQF_TRIGGER_LOW flag, but if this is the case it means they
don't support this HW feature then. So you are not breaking anything.

Best Regards,
Voice Shen
-----Original Message-----
From: Jean Delvare [mailto:khali@linux-fr.org]
Sent: Thursday, May 03, 2012 14:15 申波
To: Shen, Voice
Cc: Wu, Josh; javier Martin; linux-input@vger.kernel.org; Wolfram Sang; Axel Lin; Dmitry Torokhov
Subject: Re: qt1070: Why IRQF_TRIGGER_NONE?

On Thu, 3 May 2012 05:15:14 +0000, Shen, Voice wrote:

quoted

Hi All,
  Some information as following,

  According to the datasheet of qt1070, we can use IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW (I think this is the best) for IRQ flag. However, the IRQ line is a GPIO of a SOC. Some SOC can detect the level change of the GPIO, while can not distinguish the falling or rising. So, the IRQ flag depends on the trigger mode of GPIO line.

  Maybe use the "flags" element in "struct i2c_board_info" to pass the IRQ flag, or add another element named "irqflags" into "struct i2c_board_info". I think this will be better, but I am not sure whether this is a good solution.

i2c_board_info.flags is for I2C client flags, please do not abuse it
for IRQ information.

I have no objection to an irq_flag member being added, however I
remember past discussions where people argued whether it was the right
thing to do or whether the IRQ mode was best set by platform
initialization code. Part of that discussion was archived here:
http://marc.info/?t=128743170300002&r=1&w=2
Said discussion did not result in any code being merged as I don't
think we came to an agreement. Feel free to restart the discussion with
the interested people.

--
Jean Delvare

-- 
Javier Martin
Vista Silicon S.L.
CDTUC - FASE C - Oficina S-345
Avda de los Castros s/n
39005- Santander. Cantabria. Spain
+34 942 25 32 60
www.vista-silicon.com
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Hi Javier,

On Fri, May 04, 2012 at 09:06:16AM +0200, javier Martin wrote:

Hi all,
thank you for your comments.

On 4 May 2012 04:07, Shen, Voice [off-list ref] wrote:

quoted

Hi Jean,
 Thank you for your information. Although the irq_flag can not be added into "struct i2c_board_info" till now, I will try to find other solution.

 Thanks again.

Hi Javier,
 As to the IRQ flag depends on SOC. We try to find other solution. Please describe you issue in detail. And what's the mode does your SOC support.

As I stated, with the current 'IRQF_TRIGGER_NONE' flag, this driver
doesn't work in i.MX27. Furthermore, I don't think it works on any
platform since as you previously pointed according to the datasheet of
qt1070, we can use either IRQF_TRIGGER_FALLING or IRQF_TRIGGER_LOW.
But I don't know what is the sense of using 'IRQF_TRIGGER_NONE' here.

IRQF_TRIGGER_NONE == 0, i.e. "use whatever method board code set up".

-- 
Dmitry