Re: [PATCH 3/4] remoteproc: qcom: q6v5-mss: Add support for SDM630/636/660 MSS remoteproc

From: Stephan Gerhold <stephan@gerhold.net>
To: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Cc: Andy Gross <agross@kernel.org>,
	Bjorn Andersson <bjorn.andersson@linaro.org>,
	Konrad Dybcio <konrad.dybcio@somainline.org>,
	Mathieu Poirier <mathieu.poirier@linaro.org>,
	Rob Herring <robh+dt@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	linux-arm-msm@vger.kernel.org, linux-remoteproc@vger.kernel.org,
	devicetree@vger.kernel.org, Konrad Dybcio <konradybcio@gmail.com>
Subject: Re: [PATCH 3/4] remoteproc: qcom: q6v5-mss: Add support for SDM630/636/660 MSS remoteproc
Date: Sun, 3 Jul 2022 12:13:51 +0200	[thread overview]
Message-ID: <YsFrtn0X6ywJh6d3@gerhold.net> (raw)
In-Reply-To: <20220514000108.3070363-4-dmitry.baryshkov@linaro.org>

On Sat, May 14, 2022 at 03:01:07AM +0300, Dmitry Baryshkov wrote:
> From: Konrad Dybcio <konradybcio@gmail.com>
> 
> This adds support for sdm630/636/660 modem subsystem
> remote processor.
> 
> Signed-off-by: Konrad Dybcio <konradybcio@gmail.com>
> [DB: fixed commit message, removed note about modem restarting regularly]
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>

I was only looking at this by coincidence recently but since it doesn't
seem to be applied yet(?) some comments below.

> ---
>  drivers/remoteproc/qcom_q6v5_mss.c | 111 +++++++++++++++++++++++++++++
>  1 file changed, 111 insertions(+)
> 
> diff --git a/drivers/remoteproc/qcom_q6v5_mss.c b/drivers/remoteproc/qcom_q6v5_mss.c
> index af217de75e4d..7a4cca30db8a 100644
> --- a/drivers/remoteproc/qcom_q6v5_mss.c
> +++ b/drivers/remoteproc/qcom_q6v5_mss.c
[...]
> @@ -754,6 +759,79 @@ static int q6v5proc_reset(struct q6v5 *qproc)
>  			val |= readl(qproc->reg_base + mem_pwr_ctl);
>  			udelay(1);
>  		}
> +		/* Remove word line clamp */
> +		val = readl(qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +		val &= ~QDSP6v56_CLAMP_WL;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +	} else if (qproc->version == MSS_SDM660) {
> +		int mem_pwr_ctl;
> +
> +		/* Override the ACC value if required */
> +		writel(QDSP6SS_ACC_OVERRIDE_VAL,
> +		       qproc->reg_base + QDSP6SS_STRAP_ACC);
> +
> +		/* Assert resets, stop core */
> +		val = readl(qproc->reg_base + QDSP6SS_RESET_REG);
> +		val |= Q6SS_CORE_ARES | Q6SS_BUS_ARES_ENABLE | Q6SS_STOP_CORE;
> +		writel(val, qproc->reg_base + QDSP6SS_RESET_REG);
> +
> +		/* BHS require xo cbcr to be enabled */
> +		val = readl(qproc->reg_base + QDSP6SS_XO_CBCR);
> +		val |= 1;

val |= Q6SS_CBCR_CLKEN; (like in the existing 8996 code)

> +		writel(val, qproc->reg_base + QDSP6SS_XO_CBCR);

I would also expect this to wait for !Q6SS_CBCR_CLKOFF like on all the
other SoCs. Pretty sure downstream does that for all of them in
q6v55_branch_clk_enable().

> +
> +		/* Enable power block headswitch and wait for it to stabilize */
> +		val = readl(qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +		val |= QDSP6v56_BHS_ON;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +		mb();
> +		udelay(1);
> +
> +		for (i = BHS_CHECK_MAX_LOOPS; i > 0; i--) {
> +			if (readl_relaxed(qproc->reg_base + QDSP6V62SS_BHS_STATUS)
> +			    & QDSP6v55_BHS_EN_REST_ACK)
> +				break;
> +			udelay(1);
> +		}

This looks like readl_poll_timeout().

> +		if (!i) {
> +			pr_err("%s: BHS_EN_REST_ACK not set!\n", __func__);

dev_err()

> +			return -ETIMEDOUT;
> +		}
> +
> +		/* Put LDO in bypass mode */
> +		val |= QDSP6v56_LDO_BYP;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +
> +		/* Remove QMC_MEM clamp */
> +		val &= ~QDSP6v56_CLAMP_QMC_MEM;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +
> +		/* Deassert QDSP6 compiler memory clamp */
> +		val = readl(qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +		val &= ~QDSP6v56_CLAMP_QMC_MEM;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +

Why does this clear QDSP6v56_CLAMP_QMC_MEM twice?

> +		/* Deassert memory peripheral sleep and L2 memory standby */
> +		val |= Q6SS_L2DATA_STBY_N | Q6SS_SLP_RET_N;
> +		writel(val, qproc->reg_base + QDSP6SS_PWR_CTL_REG);
> +
> +		/* Turn on L1, L2, ETB and JU memories 1 at a time */
> +		mem_pwr_ctl = QDSP6V6SS_MEM_PWR_CTL;
> +		i = 29;
> +
> +		val = readl(qproc->reg_base + mem_pwr_ctl);
> +		for (; i >= 0; i--) {
> +			val |= BIT(i);
> +			writel(val, qproc->reg_base + mem_pwr_ctl);
> +			/*
> +			 * Read back value to ensure the write is done then
> +			 * wait for 1us for both memory peripheral and data
> +			 * array to turn on.
> +			 */
> +			val |= readl(qproc->reg_base + mem_pwr_ctl);
> +			udelay(1);
> +		}
> +
>  		/* Remove word line clamp */
>  		val = readl(qproc->reg_base + QDSP6SS_PWR_CTL_REG);
>  		val &= ~QDSP6v56_CLAMP_WL;

All in all this looks almost exactly like the existing code for
MSS_MSM8996/8. Wouldn't it be cleaner to just add an if statement for
the QDSP6V62SS_BHS_STATUS readl_poll_timeout() to the existing code?

Thanks,
Stephan