[U-Boot] [PATCH V3 05/11] I2C: mxc_i2c: address failure with mx35 processor

[Stefano Babic <sbabic@denx.de>]

On 01/21/2011 07:36 AM, Heiko Schocher wrote:
>> +static int wait_complete(void)
>> +{
>> +	int timeout = I2C_MAX_TIMEOUT;
>> +
>> +	while ((!(readw(I2C_BASE + I2SR) & I2SR_ICF)) && (--timeout)) {
>> +		writew(0, I2C_BASE + I2SR);
>> +		udelay(1);
>> +	}
>> +	udelay(200);
> 
> Why is this delay necessary? Why exactly 200? Is this documented
> somewhere in the doc?

Rather I do not have a clear explanation. In the manual there is a hint
regarding a delay that SW must introduce after setting the RSTA bit and
before writing data into the I2DR register. Really this delay should be
very short, but I checked removing the udelay() or decreasing the value
to some uSec and it does not work.

This is the output with debugging after removing the udelay call:

i2c_read chip: 0x08 addr: 0x0007 alen: 1 len: 3
i2c_addr:chip address cycle fail(a1)
i2c_addr failed
i2c_read chip: 0x08 addr: 0x001e alen: 1 len: 3
i2c_addr:chip address cycle fail(a1)
i2c_addr failed

This happens in the i2_read() call, and it is the function where the
RSTA bit is set. I can only presume the two things are related.
I checked then with the driver provided by Freescale in the LTIB, and
also in this driver a delay is set after checking transfer is completed.

Stefano

-- 
=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================