The start/stop condtions are set in different places repetedly in the i2c-pnx
driver. Beside in i2c_pnx_start and i2c_pnx_stop the start/stop bit are also
set during the transfer of a i2c message in the master_xmit/rcv calls. This is
wrong since we can't set the start/stop condition during the transaction of a
single message any way. As a matter of fact, the driver will sometimes set both
the start and the stop bits at one time. This can be easily reproduced by
sending a simple read request like e.g
struct i2c_msg msgs[] = {
{ addr, 0, 1, buf },
{ addr, I2C_M_RD, offset, buf }
};
While processing the first message the i2c_pnx_master_xmit will set both the
start_bit and the stop_bit, which will eventually confuse the slave.
Fixed by remove setting start/stop condition from the transmit routines.
Signed-off-by: Luotao Fu <l.fu@pengutronix.de>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
/* We still have something to talk about... */
val = *alg_data->mif.buf++;
- if (alg_data->mif.len == 1) {
- val |= stop_bit;
- if (!alg_data->last)
- val |= start_bit;
- }
-
alg_data->mif.len--;
iowrite32(val, I2C_REG_TX(alg_data));
__func__);
if (alg_data->mif.len == 1) {
- /* Last byte, do not acknowledge next rcv. */
- val |= stop_bit;
- if (!alg_data->last)
- val |= start_bit;
-
/*
* Enable interrupt RFDAIE (data in Rx fifo),
* and disable DRMIE (need data for Tx)