diff mbox

[07/12] i2c: pxa: Reset i2c controller on timeout in interrupt and pio mode

Message ID 1432818224-17070-8-git-send-email-vaibhav.hiremath@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

Vaibhav Hiremath May 28, 2015, 1:03 p.m. UTC 
In case of timeout during msg xfer assert reset to
i2c controller for both interrupt and PIO mode of operation.

Signed-off-by: Jett.Zhou <jtzhou@marvell.com>
[vaibhav.hiremath@linaro.org: Split & merge patches into logical changes
and update the Changelog]
Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>

Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
---
 drivers/i2c/busses/i2c-pxa.c | 8 ++++++++
 1 file changed, 8 insertions(+)

Comments

Robert Jarzmik May 29, 2015, 9:13 p.m. UTC | #1 
Vaibhav Hiremath <vaibhav.hiremath@linaro.org> writes:

> In case of timeout during msg xfer assert reset to
> i2c controller for both interrupt and PIO mode of operation.
>
> Signed-off-by: Jett.Zhou <jtzhou@marvell.com>
> [vaibhav.hiremath@linaro.org: Split & merge patches into logical changes
> and update the Changelog]
> Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
>
> Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
I have the same comment as before.
I don't like a reset in the transfer path, especially in normal busy phases for
slow I2C devices. A quirk as before.

Cheers.

--
Robert
Vaibhav Hiremath May 29, 2015, 9:19 p.m. UTC | #2 
On Saturday 30 May 2015 02:43 AM, Robert Jarzmik wrote:
> Vaibhav Hiremath <vaibhav.hiremath@linaro.org> writes:
>
>> In case of timeout during msg xfer assert reset to
>> i2c controller for both interrupt and PIO mode of operation.
>>
>> Signed-off-by: Jett.Zhou <jtzhou@marvell.com>
>> [vaibhav.hiremath@linaro.org: Split & merge patches into logical changes
>> and update the Changelog]
>> Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
>>
>> Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
> I have the same comment as before.
> I don't like a reset in the transfer path, especially in normal busy phases for
> slow I2C devices. A quirk as before.
>

Note that this assertion of reset in in case of timeout error.

Timeout error may be due to various reasons, and this is recovery
mechanism.

Thanks,
Vaibhav
diff mbox

Patch

diff --git a/drivers/i2c/busses/i2c-pxa.c b/drivers/i2c/busses/i2c-pxa.c
index 2777d5c..3c6ebb5 100644
--- a/drivers/i2c/busses/i2c-pxa.c
+++ b/drivers/i2c/busses/i2c-pxa.c
@@ -197,6 +197,8 @@  struct pxa_i2c {
  */
 #define I2C_PXA_SLAVE_ADDR      0x1
 
+static void i2c_pxa_reset(struct pxa_i2c *i2c);
+
 #ifdef DEBUG
 
 struct bits {
@@ -846,6 +848,9 @@  out:
 		ret = I2C_RETRY;
 	}
 
+	if (ret < 0)
+		i2c_pxa_reset(i2c);
+
 	return ret;
 }
 
@@ -912,6 +917,9 @@  static int i2c_pxa_do_xfer(struct pxa_i2c *i2c, struct i2c_msg *msg, int num)
 	}
 
  out:
+	if (ret < 0)
+		i2c_pxa_reset(i2c);
+
 	return ret;
 }