i2c: Remove setting for 1 second timeout from adapter drivers

PATCH RESEND In review Jarkko Nikula 2 participants 3 messages Dec 2, 2015 – Dec 4, 2015

Subsystems (3)

I2C subsystem host drivers I2C MV64XXX Marvell and Allwinner driver Freescale mpc I2C driver

lore

Share & tools

lore

Jarkko Nikula

author

2015-12-02 11:02 UTC #0 lore raw

PATCH RESEND 1/1 i2c: Remove setting for 1 second timeout from adapter drivers

I2C adapter drivers that are using 1 second timeout can leave the timeout unset and let the i2c-core.c: i2c_register_adapter() to set it instead.

Signed-off-by Jarkko Nikula <[email protected]>

---
 drivers/i2c/busses/i2c-ibm_iic.c      | 1 -
 drivers/i2c/busses/i2c-iop3xx.c       | 1 -
 drivers/i2c/busses/i2c-mpc.c          | 1 -
 drivers/i2c/busses/i2c-mv64xxx.c      | 5 -----
 drivers/i2c/busses/i2c-pca-platform.c | 1 -
 5 files changed, 9 deletions(-)

drivers/i2c/busses/i2c-ibm_iic.c +0 -1

index ab492301581a..5dc0bb7e5dbc 100644
@@ -748,7 +748,6 @@ static int iic_probe(struct platform_device *ofdev)
 	i2c_set_adapdata(adap, dev);
 	adap->class = I2C_CLASS_HWMON | I2C_CLASS_SPD;
 	adap->algo = &iic_algo;
-	adap->timeout = HZ;
 
 	ret = i2c_add_adapter(adap);
 	if (ret  < 0) {

drivers/i2c/busses/i2c-iop3xx.c +0 -1

index 72d6161cf77c..c4ed8281bec2 100644
@@ -479,7 +479,6 @@ iop3xx_i2c_probe(struct platform_device *pdev)
 	/*
 	 * Default values...should these come in from board code?
 	 */
-	new_adapter->timeout = HZ;
 	new_adapter->algo = &iop3xx_i2c_algo;
 
 	init_waitqueue_head(&adapter_data->waitq);

drivers/i2c/busses/i2c-mpc.c +0 -1

index 48ecffecc0ed..b4bb64ffe3cd 100644
@@ -640,7 +640,6 @@ static const struct i2c_algorithm mpc_algo = {
 static struct i2c_adapter mpc_ops = {
 	.owner = THIS_MODULE,
 	.algo = &mpc_algo,
-	.timeout = HZ,
 };
 
 static const struct of_device_id mpc_i2c_of_match[];

drivers/i2c/busses/i2c-mv64xxx.c +0 -5

index 43207f52e5a3..eaef7763f5fc 100644
@@ -839,11 +839,6 @@ mv64xxx_of_config(struct mv64xxx_i2c_data *drv_data,
 		reset_control_deassert(drv_data->rstc);
 	}
 
-	/* Its not yet defined how timeouts will be specified in device tree.
-	 * So hard code the value to 1 second.
-	 */
-	drv_data->adapter.timeout = HZ;
-
 	device = of_match_device(mv64xxx_i2c_of_match_table, dev);
 	if (!device)
 		return -ENODEV;

drivers/i2c/busses/i2c-pca-platform.c +0 -1

index 3bd2e7d06e4b..2b930df0e6b5 100644
@@ -183,7 +183,6 @@ static int i2c_pca_pf_probe(struct platform_device *pdev)
 		i2c->algo_data.i2c_clock = platform_data->i2c_clock_speed;
 		i2c->gpio = platform_data->gpio;
 	} else {
-		i2c->adap.timeout = HZ;
 		i2c->algo_data.i2c_clock = 59000;
 		i2c->gpio = -1;
 	}

Signature · 1 line

2.6.2

Wolfram Sang

2015-12-03 20:52 UTC #1 lore raw

Quoted 4 lines · Jarkko Nikula

I2C adapter drivers that are using 1 second timeout can leave the timeout unset and let the i2c-core.c: i2c_register_adapter() to set it instead.
Signed-off-by Jarkko Nikula <[email protected]>

My take on this is that I prefer the explicit (and thus visible) assignment. What the core does was intended as a fallback to prevent strange things happening with a 0 value.

Jarkko Nikula

author

2015-12-04 08:01 UTC #2 lore raw

Quoted 10 lines · Wolfram Sang

Quoted 4 lines · Jarkko Nikula
I2C adapter drivers that are using 1 second timeout can leave the timeout unset and let the i2c-core.c: i2c_register_adapter() to set it instead.
Signed-off-by Jarkko Nikula <[email protected]>
My take on this is that I prefer the explicit (and thus visible) assignment. What the core does was intended as a fallback to prevent strange things happening with a 0 value.

Fair enough. It's hard to say if some of these really needs to be 1 second instead of being stetson guessed or copy-pasted. Now if we ever change the fall back value from HZ to HZ/10 or similar we might get a regression.

Signature · 1 line

Jarkko