On Friday 29 May 2015 07:29 PM, Rob Herring wrote:
On Thu, May 28, 2015 at 8:03 AM, Vaibhav Hiremath
<vaibhav.hiremath@xxxxxxxxxx> wrote:
From: Rob Herring <robh@xxxxxxxxxx>
This probably should still be Leilei, but...
Ok, Since I am taking forward from your sign-off, did not change it.
Anyway will fix in next version.
Since there is some problematic i2c slave devices on some
platforms such as dkb (sometimes), it will drop down sda
and make i2c bus hang, at that time, it need to config
scl/sda into gpio to simulate "stop" sequence to recover
i2c bus, so add this interface.
Signed-off-by: Leilei Shang <shangll@xxxxxxxxxxx>
Signed-off-by: Rob Herring <robh@xxxxxxxxxx>
[vaibhav.hiremath@xxxxxxxxxx: Updated Changelog]
Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@xxxxxxxxxx>
Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@xxxxxxxxxx>
---
drivers/i2c/busses/i2c-pxa.c | 90 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 90 insertions(+)
diff --git a/drivers/i2c/busses/i2c-pxa.c b/drivers/i2c/busses/i2c-pxa.c
index 8ca5552..eb09071 100644
--- a/drivers/i2c/busses/i2c-pxa.c
+++ b/drivers/i2c/busses/i2c-pxa.c
@@ -37,6 +37,8 @@
#include <linux/slab.h>
#include <linux/io.h>
#include <linux/i2c/pxa-i2c.h>
+#include <linux/of_gpio.h>
+#include <linux/pinctrl/consumer.h>
#include <asm/irq.h>
@@ -177,6 +179,9 @@ struct pxa_i2c {
bool highmode_enter;
unsigned int ilcr;
unsigned int iwcr;
+ struct pinctrl *pinctrl;
+ struct pinctrl_state *pin_i2c;
+ struct pinctrl_state *pin_gpio;
};
#define _IBMR(i2c) ((i2c)->reg_ibmr)
@@ -269,6 +274,62 @@ static void i2c_pxa_show_state(struct pxa_i2c *i2c, int lno, const char *fname)
#define show_state(i2c) i2c_pxa_show_state(i2c, __LINE__, __func__)
+static void i2c_bus_reset(struct pxa_i2c *i2c)
There's a generic mechanism in i2c_generic_gpio_recovery we should use
here. It appears to be similar, but not exactly the same. The pinctrl
part should probably be done by gpio driver.
Good point.
As you mentioned, they are not exactly same.
But we can achieve exactly what we wanted using prepare & unprepare
callbacks.
Generate stop signal in unprepare() callback should suffice our need.
But I am not quite sure about pinctrl handling by gpio driver.
I was tracing the calls from gpio_request_one()anf gpio_free()
but it seems they are not bringing back pins to default state.
correct me if I am wrong.
Thanks,
Vaibhav
--
To unsubscribe from this list: send the line "unsubscribe linux-i2c" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
