[PATCH v3] i2c: imx: when being a target, mark the last read as processed

[Wolfram Sang]

From: Corey Minyard <minyard at acm.org>

When being a target, NAK from the controller means that all bytes have
been transferred. So, the last byte needs also to be marked as
'processed'. Otherwise index registers of backends may not increase.

Signed-off-by: Corey Minyard <minyard at acm.org>
Tested-by: Andrew Manley <andrew.manley at sealingtech.com>
Reviewed-by: Andrew Manley <andrew.manley at sealingtech.com>
Reviewed-by: Oleksij Rempel <o.rempel at pengutronix.de>
[wsa: fixed comment and commit message to properly describe the case]
Signed-off-by: Wolfram Sang <wsa+renesas at sang-engineering.com>
---

Changes since v2:
* updated commit message and comment

In the stalled discussion[1], it seems I couldn't make my suggestions
clear. So, here are the changes how I meant them. I hope this can be
agreed on.

[1] http://patchwork.ozlabs.org/project/linux-i2c/patch/20211112133956.655179-3-minyard@acm.org/

 drivers/i2c/busses/i2c-imx.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c
index 88a053987403..60e813137f84 100644
--- a/drivers/i2c/busses/i2c-imx.c
+++ b/drivers/i2c/busses/i2c-imx.c
@@ -803,6 +803,11 @@ static irqreturn_t i2c_imx_slave_handle(struct imx_i2c_struct *i2c_imx,
 		ctl &= ~I2CR_MTX;
 		imx_i2c_write_reg(ctl, i2c_imx, IMX_I2C_I2CR);
 		imx_i2c_read_reg(i2c_imx, IMX_I2C_I2DR);
+
+		/* flag the last byte as processed */
+		i2c_imx_slave_event(i2c_imx,
+				    I2C_SLAVE_READ_PROCESSED, &value);
+
 		i2c_imx_slave_finish_op(i2c_imx);
 		return IRQ_HANDLED;
 	}
-- 
2.43.0