linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Jeffery <andrew@codeconstruct.com.au>
To: Cool Lee <cool_lee@aspeedtech.com>,
	adrian.hunter@intel.com,  ulf.hansson@linaro.org, joel@jms.id.au,
	p.zabel@pengutronix.de,  linux-aspeed@lists.ozlabs.org,
	openbmc@lists.ozlabs.org,  linux-mmc@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	 linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/8] mmc: sdhci-of-aspeed: Add runtime tuning
Date: Wed, 18 Jun 2025 12:01:57 +0930	[thread overview]
Message-ID: <750e6e51e0943e50162851ca8df2befcea3e9bbe.camel@codeconstruct.com.au> (raw)
In-Reply-To: <20250615035803.3752235-3-cool_lee@aspeedtech.com>

On Sun, 2025-06-15 at 11:57 +0800, Cool Lee wrote:
> Add support for runtime tuning in the Aspeed SDHCI driver.
> Using the timing phase register to adjust the clock phase with mmc
> tuning command to find the left and right boundary.
> 
> Signed-off-by: Cool Lee <cool_lee@aspeedtech.com>
> ---
>  drivers/mmc/host/sdhci-of-aspeed.c | 68 ++++++++++++++++++++++++++++++
>  1 file changed, 68 insertions(+)
> 
> diff --git a/drivers/mmc/host/sdhci-of-aspeed.c b/drivers/mmc/host/sdhci-of-aspeed.c
> index 01bc574272eb..5e5ae1894456 100644
> --- a/drivers/mmc/host/sdhci-of-aspeed.c
> +++ b/drivers/mmc/host/sdhci-of-aspeed.c
> @@ -24,6 +24,7 @@
>  #define ASPEED_SDC_PHASE               0xf4
>  #define   ASPEED_SDC_S1_PHASE_IN       GENMASK(25, 21)
>  #define   ASPEED_SDC_S0_PHASE_IN       GENMASK(20, 16)
> +#define   ASPEED_SDC_S0_PHASE_IN_SHIFT  16
>  #define   ASPEED_SDC_S1_PHASE_OUT      GENMASK(15, 11)
>  #define   ASPEED_SDC_S1_PHASE_IN_EN    BIT(10)
>  #define   ASPEED_SDC_S1_PHASE_OUT_EN   GENMASK(9, 8)
> @@ -375,6 +376,72 @@ static void aspeed_sdhci_reset(struct sdhci_host *host, u8 mask)
>         sdhci_reset(host, mask);
>  }
>  
> +static int aspeed_sdhci_execute_tuning(struct sdhci_host *host, u32 opcode)
> +{
> +       struct sdhci_pltfm_host *pltfm_priv;
> +       struct aspeed_sdhci *sdhci;
> +       struct aspeed_sdc *sdc;
> +       struct device *dev;
> +
> +       u32 val, left, right, edge;
> +       u32 window, oldwindow = 0, center;
> +       u32 in_phase, out_phase, enable_mask, inverted = 0;
> +
> +       dev = mmc_dev(host->mmc);
> +       pltfm_priv = sdhci_priv(host);
> +       sdhci = sdhci_pltfm_priv(pltfm_priv);
> +       sdc = sdhci->parent;
> +
> +       out_phase = readl(sdc->regs + ASPEED_SDC_PHASE) & ASPEED_SDC_S0_PHASE_OUT;
> +
> +       enable_mask = ASPEED_SDC_S0_PHASE_OUT_EN | ASPEED_SDC_S0_PHASE_IN_EN;
> +
> +       /*
> +        * There are two window upon clock rising and falling edge.
> +        * Iterate each tap delay to find the valid window and choose the
> +        * bigger one, set the tap delay at the middle of window.
> +        */
> +       for (edge = 0; edge < 2; edge++) {
> +               if (edge == 1)
> +                       inverted = ASPEED_SDHCI_TAP_PARAM_INVERT_CLK;
> +
> +               val = (out_phase | enable_mask | (inverted << ASPEED_SDC_S0_PHASE_IN_SHIFT));
> +
> +               /* find the left boundary */
> +               for (left = 0; left < ASPEED_SDHCI_NR_TAPS + 1; left++) {

Bit of a nit, but maybe `left <= ASPEED_SDHCI_NR_TAPS` rather than + 1?

> +                       in_phase = val | (left << ASPEED_SDC_S0_PHASE_IN_SHIFT);
> +                       writel(in_phase, sdc->regs + ASPEED_SDC_PHASE);
> +
> +                       if (!mmc_send_tuning(host->mmc, opcode, NULL))
> +                               break;
> +               }
> +
> +               /* find the right boundary */
> +               for (right = left + 1; right < ASPEED_SDHCI_NR_TAPS + 1; right++) {

<= again here if you agree.

> +                       in_phase = val | (right << ASPEED_SDC_S0_PHASE_IN_SHIFT);
> +                       writel(in_phase, sdc->regs + ASPEED_SDC_PHASE);
> +
> +                       if (mmc_send_tuning(host->mmc, opcode, NULL))
> +                               break;
> +               }
> +
> +               window = right - left;
> +               dev_info(dev, "tuning window = %d\n", window);

I think this should be dev_dbg() rather than dev_info(). Tuning data is
not something that should normally be printed. I'd also print the
values of left and right, for reference.

> +
> +               if (window > oldwindow) {
> +                       oldwindow = window;
> +                       center = (((right - 1) + left) / 2) | inverted;
> +               }
> +       }
> +
> +       val = (out_phase | enable_mask | (center << ASPEED_SDC_S0_PHASE_IN_SHIFT));
> +       writel(val, sdc->regs + ASPEED_SDC_PHASE);
> +
> +       dev_info(dev, "tuning result=%x\n", val);

dev_dbg() again.

Cheers,

Andrew

  reply	other threads:[~2025-06-18  2:34 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-15  3:57 [PATCH 0/8] Aspeed SDHCI driver workaround and auto tune Cool Lee
2025-06-15  3:57 ` [PATCH 1/8] mmc: sdhci-of-aspeed: Fix sdhci software reset can't be cleared issue Cool Lee
2025-06-16 13:22   ` Philipp Zabel
2025-06-18  1:34     ` Cool Lee
2025-06-18  2:14   ` Andrew Jeffery
2025-06-19  6:53     ` Cool Lee
2025-06-20  7:43       ` Andrew Jeffery
2025-06-21  8:29         ` Cool Lee
2025-06-15  3:57 ` [PATCH 2/8] mmc: sdhci-of-aspeed: Add runtime tuning Cool Lee
2025-06-18  2:31   ` Andrew Jeffery [this message]
2025-06-19  6:57     ` Cool Lee
2025-06-15  3:57 ` [PATCH 3/8] mmc: sdhci-of-aspeed: Patch HOST_CONTROL2 register missing after top reset Cool Lee
2025-06-18  2:32   ` Andrew Jeffery
2025-06-19  6:57     ` Cool Lee
2025-06-15  3:57 ` [PATCH 4/8] mmc: sdhci-of-aspeed: Get max clockk by using default api Cool Lee
2025-06-18  2:39   ` Andrew Jeffery
2025-06-20  8:18     ` Cool Lee
2025-06-15  3:58 ` [PATCH 5/8] mmc: sdhci-of-aspeed: Fix null pointer Cool Lee
2025-06-18  2:49   ` Andrew Jeffery
2025-06-20  8:18     ` Cool Lee
2025-06-15  3:58 ` [PATCH 6/8] mmc: sdhci-of-aspeed: Add output timing phase tuning Cool Lee
2025-06-18  2:51   ` Andrew Jeffery
2025-06-20  8:19     ` Cool Lee
2025-06-15  3:58 ` [PATCH 7/8] mmc: sdhci-of-aspeed: Remove timing phase Cool Lee
2025-06-18  2:56   ` Andrew Jeffery
2025-06-20 10:23     ` Cool Lee
2025-06-24 23:31       ` Andrew Jeffery
2025-06-25  0:22         ` Cool Lee
2025-06-25  0:23           ` Andrew Jeffery
2025-06-15  3:58 ` [PATCH 8/8] mmc: sdhci-of-aspeed: Add sdr50 support Cool Lee
2025-06-18  3:06   ` Andrew Jeffery

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=750e6e51e0943e50162851ca8df2befcea3e9bbe.camel@codeconstruct.com.au \
    --to=andrew@codeconstruct.com.au \
    --cc=adrian.hunter@intel.com \
    --cc=cool_lee@aspeedtech.com \
    --cc=joel@jms.id.au \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-aspeed@lists.ozlabs.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mmc@vger.kernel.org \
    --cc=openbmc@lists.ozlabs.org \
    --cc=p.zabel@pengutronix.de \
    --cc=ulf.hansson@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).