summaryrefslogtreecommitdiff
path: root/drivers/i2c
diff options
context:
space:
mode:
authorWolfram Sang <wsa+renesas@sang-engineering.com>2018-04-28 22:01:02 +0200
committerWolfram Sang <wsa@the-dreams.de>2018-04-30 11:14:35 +0200
commitb1437dcb973de79d1f9bca55c6056e8b40ff7ba7 (patch)
tree561fb170f41c4b1610e63ca7f4e7977bfa56510b /drivers/i2c
parent27aaa8ad5a5e79d01653523c305fe6af707dce8e (diff)
i2c: rcar: enhance comment to avoid regressions
Give a clear testcase for people wishing to change this code. It is also a reminder for me if people ask about it. Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Diffstat (limited to 'drivers/i2c')
-rw-r--r--drivers/i2c/busses/i2c-rcar.c2
1 files changed, 2 insertions, 0 deletions
diff --git a/drivers/i2c/busses/i2c-rcar.c b/drivers/i2c/busses/i2c-rcar.c
index c6915b835396..9d8d5b91220f 100644
--- a/drivers/i2c/busses/i2c-rcar.c
+++ b/drivers/i2c/busses/i2c-rcar.c
@@ -542,6 +542,8 @@ static void rcar_i2c_irq_recv(struct rcar_i2c_priv *priv, u32 msr)
* If next received data is the _LAST_, go to STOP phase. Might be
* overwritten by REP START when setting up a new msg. Not elegant
* but the only stable sequence for REP START I have found so far.
+ * If you want to change this code, make sure sending one transfer with
+ * four messages (WR-RD-WR-RD) works!
*/
if (priv->pos + 1 >= msg->len)
rcar_i2c_write(priv, ICMCR, RCAR_BUS_PHASE_STOP);