Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
In-Reply-To: <1654696929-20205-3-git-send-email-quic_srivasam@quicinc.com>
References: <1654696929-20205-1-git-send-email-quic_srivasam@quicinc.com> <1654696929-20205-3-git-send-email-quic_srivasam@quicinc.com>
From:   Stephen Boyd <swboyd@chromium.org>
User-Agent: alot/0.10
Date:   Wed, 8 Jun 2022 12:29:04 -0700
Message-ID: <CAE-0n53EY1eKqnVLhU__e7t63BbVoKz++6aijOpEw0k5Cxa8-w@mail.gmail.com>
Subject: Re: [PATCH v3 2/2] ASoC: qcom: soundwire: Add software clock gating
 requirement check
To:     Srinivasa Rao Mandadapu <quic_srivasam@quicinc.com>,
        agross@kernel.org, alsa-devel@alsa-project.org,
        bgoswami@quicinc.com, bjorn.andersson@linaro.org,
        broonie@kernel.org, devicetree@vger.kernel.org,
        judyhsiao@chromium.org, lgirdwood@gmail.com,
        linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org,
        perex@perex.cz, quic_plai@quicinc.com, quic_rohkumar@quicinc.com,
        robh+dt@kernel.org, srinivas.kandagatla@linaro.org, tiwai@suse.com,
        vkoul@kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Quoting Srinivasa Rao Mandadapu (2022-06-08 07:02:09)
> Validate software clock gating required or not and do software
> clock gating on hclk if soundwire is operational and keep it
> running by adding flag in private dat structure.
> This is to avoid conflict between older architectures,
> where software clock gating is not required and on latest
> architectues, where software clock gating is mandatory.

This talks about software clock gating but the code is getting a reset
and asserting it. Is that because the power on reset value of the clock
gating is to have hardware clock gating disabled, but some earlier code
is enabling hardware clock gating?

>
> Signed-off-by: Srinivasa Rao Mandadapu <quic_srivasam@quicinc.com>
> Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> ---
>  drivers/soundwire/qcom.c | 27 ++++++++++++++++++++-------
>  1 file changed, 20 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
> index 38c3bf5..ebd7479 100644
> --- a/drivers/soundwire/qcom.c
> +++ b/drivers/soundwire/qcom.c
> @@ -659,7 +665,8 @@ static int qcom_swrm_init(struct qcom_swrm_ctrl *ctrl)
>         val = FIELD_PREP(SWRM_MCP_FRAME_CTRL_BANK_ROW_CTRL_BMSK, ctrl->rows_index);
>         val |= FIELD_PREP(SWRM_MCP_FRAME_CTRL_BANK_COL_CTRL_BMSK, ctrl->cols_index);
>
> -       reset_control_reset(ctrl->audio_cgcr);
> +       if (ctrl->audio_cgcr)
> +               reset_control_reset(ctrl->audio_cgcr);
>
>         ctrl->reg_write(ctrl, SWRM_MCP_FRAME_CTRL_BANK_ADDR(0), val);
>
> @@ -1494,7 +1506,8 @@ static int __maybe_unused swrm_runtime_resume(struct device *dev)
>                 qcom_swrm_get_device_status(ctrl);
>                 sdw_handle_slave_status(&ctrl->bus, ctrl->status);
>         } else {
> -               reset_control_reset(ctrl->audio_cgcr);
> +               if (ctrl->audio_cgcr)
> +                       reset_control_reset(ctrl->audio_cgcr);

reset_control_reset() is a no-op if the pointer is NULL so the if
condition is not necessary in the above two statements.