Re: [PATCH 1/4] i2c: rockchip: fix i2c stop condition

[Sascha Hauer <sha@pengutronix.de>]

On Tue, Sep 12, 2023 at 07:45:36AM +0200, Gerald Loacker wrote:
> 
> 
> Am 08.09.2023 um 15:53 schrieb Sascha Hauer:
> > On Fri, Sep 08, 2023 at 12:16:46PM +0200, Gerald Loacker wrote:
> >> The i2c bus gets disabled without sending a stop condition. This violates
> >> i2c timing on the clock line. Fix this and include all related functions
> >> (rk_i2c_send_start_bit, rk_i2c_send_stop_bit, rk_i2c_disable) onto the same
> >> level.
> >>
> >> Signed-off-by: Gerald Loacker <gerald.loacker@wolfvision.net>
> >> ---
> >>  drivers/i2c/busses/i2c-rockchip.c | 21 ++++++++-------------
> >>  1 file changed, 8 insertions(+), 13 deletions(-)
> >>
> >> @@ -387,6 +376,11 @@ static int rockchip_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs,
> >>  		struct i2c_msg *msg = &msgs[i];
> >>  
> >>  		dev_dbg(dev, "i2c_xfer: chip=0x%x, len=0x%x\n", msg->addr, msg->len);
> >> +
> >> +		ret = rk_i2c_send_start_bit(i2c);
> >> +		if (ret)
> >> +			break;
> >> +
> >>  		if (msg->flags & I2C_M_RD) {
> >>  			ret = rk_i2c_read(i2c, msg->addr, 0, 0, msg->buf,
> >>  					  msg->len);
> >> @@ -394,6 +388,9 @@ static int rockchip_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs,
> >>  			ret = rk_i2c_write(i2c, msg->addr, 0, 0, msg->buf,
> >>  					   msg->len);
> >>  		}
> >> +
> >> +		rk_i2c_send_stop_bit(i2c);
> >> +
> >>  		if (ret) {
> >>  			dev_dbg(dev, "i2c_write: error sending: %pe\n",
> >>  				ERR_PTR(ret));
> >> @@ -402,9 +399,7 @@ static int rockchip_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs,
> >>  		}
> >>  	}
> >>  
> >> -	rk_i2c_send_stop_bit(i2c);
> > 
> > I believe this is wrong. A stop bit should only be sent after the last
> > message, not after all messages.
> > 
> 
> Your're right. This was just because the repeated start does not work
> consitently in our case. I'll come up with another approach.

Might be worth to look at the kernel driver. It has a comment about
this:

	/*
	 * The HW is actually not capable of REPEATED START. But we can
	 * get the intended effect by resetting its internal state
	 * and issuing an ordinary START.
	 */
	ctrl = i2c_readl(i2c, REG_CON) & REG_CON_TUNING_MASK;
	i2c_writel(i2c, ctrl, REG_CON);

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |