[U-Boot] ColdFire I2C implementing I2C idle [PATCH] -- resent in git format

[Heiko Schocher]

Hello Joakim,

Joakim Tjernlund wrote:
>> Hello Michael,
>>
>> Thanks for posting your patches in plain text.
>>
>> Michael Durrant wrote:
>>> drivers_i2c_fsl_i2c.patch
>>>    - need to set I2C to be idle according to the MCF5282 user's manual
>>>
>>>    If I2SR[IBB] is set when the I2C bus module is enabled,
>>>    execute the following code sequence before proceeding with
>>>    normal initialization code. This issues a STOP command to the
>>>    slave device, placing it in idle state as if it were just
>>>    power-cycled on.
>>>
>>>    I2CR = 0x0
>>>    I2CR = 0xA
>>>    dummy read of I2DR
>>>    I2SR = 0x0
>>>    I2CR = 0x0
>>>
>>> Signed-off-by: David Wu <davidwu at arcturusnetworks.com>
>>> Signed-off-by: Michael Durrant <mdurrant at arcturusnetworks.com>
>>> ---
>>>  drivers/i2c/fsl_i2c.c |   13 +++++++++++++
>>>  1 files changed, 13 insertions(+), 0 deletions(-)
>>>
>>> diff --git a/drivers/i2c/fsl_i2c.c b/drivers/i2c/fsl_i2c.c
>>> index 2241990..bce750c 100644
>>> --- a/drivers/i2c/fsl_i2c.c
>>> +++ b/drivers/i2c/fsl_i2c.c
>>> @@ -251,12 +251,25 @@ i2c_init(int speed, int slaveadd)
>>>  #endif
>>>  }
>>>
>>> +static __inline__ int i2c_set_idle(void)
>>> +{
>>> +    writeb(0, &i2c_dev[i2c_bus_num]->cr);
>>> +    writeb(I2C_CR_MEN | I2C_CR_MSTA, &i2c_dev[i2c_bus_num]->cr);
>>> +    readb(&i2c_dev[i2c_bus_num]->dr);
>>> +    writeb(0, &i2c_dev[i2c_bus_num]->sr);
>>> +    writeb(0, &i2c_dev[i2c_bus_num]->cr);
>>> +    writeb(I2C_CR_MEN, &i2c_dev[i2c_bus_num]->cr);
>>> +}
>>> +
>>>  static int
>>>  i2c_wait4bus(void)
>>>  {
>>>      unsigned long long timeval = get_ticks();
>>>      const unsigned long long timeout = usec2ticks(CONFIG_I2C_MBB_TIMEOUT);
>>>
>>> +    if (readb(&i2c_dev[i2c_bus_num]->sr) & I2C_SR_MBB)
>>> +        i2c_set_idle();
>>> +
>>>      while (readb(&i2c_dev[i2c_bus_num]->sr) & I2C_SR_MBB) {
>>>          if ((get_ticks() - timeval) > timeout)
>>>              return -1;
>> Hmm.. I can;t find this for example in the MPC8360ERM.pdf, which
>> uses also this driver ... So I vote for activating this only,
>> if this driver is used for  __M68K__ ...
>>
>> Also, you write, that this sequence is necessary before normal
>> initialization code, but i2c_wait4bus() is called from i2c_read()
>> and i2c_write(), so the I2C bus is long ago initialized ...
>> or what do you mean with "normal initialization code" ?
>>
>> Also, whats with multimaster systems? Your code maybe cuts a running
>> data transmission. The MPC8360ERM.pdf manual says "check the MBB bit,
>> if the bus is free, before switching to master mode", thats what
>> actual code did ... so, I want this only activated, if __M68K__
>> is defined ...
> 
> All true. This cannot go in as is. What you need is a I2C reset sequence
> as the above isn't enough in the general case. Michael, have a look at the
> kernel driver, it has some fixup code you could borrow.

Michael: Maybe you take a look in u-boot:board/keymile/common/common.c
i2c_init_board(), there is a I2C reset sequence for 83xx based boards
from keymile, which use this i2c driver.

Maybe its time to move this code in the i2c driver?

bye
Heiko
-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany