Re: [PATCH v3 3/5] ASoC: qcom: Add support for lpass hdmi driver
On 04/09/2020 12:21, Srinivasa Rao Mandadapu wrote:
Before going into detail review, I see real issue in the overall approach here to add new interface to exiting lpass!!
Intention of struct lpass_variant is to address differences between SoCs or different lpass versions. But you should not duplicate this and use it for addressing differences between each lpass interfaces! All the dai related register offsets should still go in to this structure and driver should be able to know which dai its talking to based on snd_soc_dai_driver id and select correct register offset.
Do You suggest to use separate structure like struct lpass_hdmi_interface in lpass_data?
No, you should just put every thing lpass_variant as you did but no "id" thing in that.
--srini
Thanks for your reply Srinivas!!!
On 9/4/2020 4:54 PM, Srinivas Kandagatla wrote:
On 04/09/2020 12:21, Srinivasa Rao Mandadapu wrote:
Before going into detail review, I see real issue in the overall approach here to add new interface to exiting lpass!!
Intention of struct lpass_variant is to address differences between SoCs or different lpass versions. But you should not duplicate this and use it for addressing differences between each lpass interfaces! All the dai related register offsets should still go in to this structure and driver should be able to know which dai its talking to based on snd_soc_dai_driver id and select correct register offset.
Do You suggest to use separate structure like struct lpass_hdmi_interface in lpass_data?
No, you should just put every thing lpass_variant as you did but no "id" thing in that.
Okay. will change accordingly.
--srini
participants (2)
-
Srinivas Kandagatla -
Srinivasa Rao Mandadapu