We were missing a transfer delay in one execution path leading to hangs and the bus timeout was too low leading to errors under stress tests.
Acked-by: Srinidhi Kasagar <[email protected]> Signed-off-by: Sundar R Iyer <[email protected]> Signed-off-by: Linus Walleij <[email protected]> --- drivers/i2c/busses/i2c-nomadik.c | 8 ++++++-- 1 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/i2c/busses/i2c-nomadik.c b/drivers/i2c/busses/i2c-nomadik.c index e461de7..c9fffd0 100644 --- a/drivers/i2c/busses/i2c-nomadik.c +++ b/drivers/i2c/busses/i2c-nomadik.c @@ -103,6 +103,9 @@ /* maximum threshold value */ #define MAX_I2C_FIFO_THRESHOLD 15 +/* per-transfer delay, required for the hardware to stabilize */ +#define I2C_DELAY 150 + enum i2c_status { I2C_NOP, I2C_ON_GOING, @@ -118,7 +121,7 @@ enum i2c_operation { }; /* controller response timeout in ms */ -#define I2C_TIMEOUT_MS 500 +#define I2C_TIMEOUT_MS 2000 /** * struct i2c_nmk_client - client specific data @@ -267,6 +270,7 @@ static int init_hw(struct nmk_i2c_dev *dev) clk_disable(dev->clk); + udelay(I2C_DELAY); return 0; } @@ -599,7 +603,7 @@ static int nmk_i2c_xfer(struct i2c_adapter *i2c_adap, clk_disable(dev->clk); return status; } - mdelay(1); + udelay(I2C_DELAY); } clk_disable(dev->clk); -- 1.6.3.3 -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
