On Mon, 2015-06-29 at 10:27 +0800, Axel Lin wrote:
2015-06-29 10:02 GMT+08:00 Haojian Zhuang haojian.zhuang@linaro.org:
On Sun, 2015-06-28 at 23:00 +0800, Axel Lin wrote:
Hi Haojian, In sound/soc/codecs/88pm860x-codec.c (pm860x_set_dai_sysclk): When dir != PM860X_CLK_DIR_OUT, current code set pm860x->dir = PM860X_CLK_DIR_IN; and then return -EINVAL. This looks confusion. Since there is no comment for that, I'm wondering if the return -EINVAL is intentional or not.
Regards, Axel
Hi Axel,
Since we didn't support slave mode in our solution, I just return -EINVAL for PM860X_CLK_DIR_IN. I'm sorry to add the confused code. Could you help to format a patch to add the comments?
The main confusion part is current code actually *changes* pm860x->dir setting when return -EINVAL in pm860x_set_dai_sysclk. (And the pm860x->dir is used in pm860x_pcm_set_dai_fmt.) so I'm wondering if it make sense to just return -EINVAL without setting pm860x->dir = PM860X_CLK_DIR_IN; in pm860x_set_dai_sysclk.
It's fine to return -EINVAL directly without setting pm680x->dir.
Regards Haojian