From: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
To: Douglas Anderson <dianders@chromium.org>,
Ulf Hansson <ulf.hansson@linaro.org>,
Adrian Hunter <adrian.hunter@intel.com>
Cc: Stephen Boyd <swboyd@chromium.org>,
Taniya Das <tdas@codeaurora.org>, Andy Gross <agross@kernel.org>,
Bjorn Andersson <bjorn.andersson@linaro.org>,
linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-mmc@vger.kernel.org
Subject: Re: [PATCH v4 2/2] mmc: sdhci-msm: Actually set the actual clock
Date: Mon, 14 Dec 2020 18:13:51 +0530 [thread overview]
Message-ID: <e817284c-1ae9-7d3f-5195-7313651ef9da@codeaurora.org> (raw)
In-Reply-To: <20201211091150.v4.2.I7564620993acd4baa63fa0e3925ca879a86d3ee3@changeid>
On 12/11/2020 10:42 PM, Douglas Anderson wrote:
> The MSM SDHCI driver always set the "actual_clock" field to 0. It had
> a comment about it not being needed because we weren't using the
> standard SDHCI divider mechanism and we'd just fallback to
> "host->clock". However, it's still better to provide the actual
> clock. Why?
>
> 1. It will make timeout calculations slightly better. On one system I
> have, the eMMC requets 200 MHz (for HS400-ES) but actually gets 192
> MHz. These are close, but why not get the more accurate one.
>
> 2. If things are seriously off in the clock driver and it's missing
> rates or picking the wrong rate (maybe it's rounding up instead of
> down), this will make it much more obvious what's going on.
>
> NOTE: we have to be a little careful here because the "actual_clock"
> field shouldn't include the multiplier that sdhci-msm needs
> internally.
>
> Signed-off-by: Douglas Anderson <dianders@chromium.org>
> ---
>
> Changes in v4:
> - ("mmc: sdhci-msm: Actually set the actual clock") new for v4.
>
> drivers/mmc/host/sdhci-msm.c | 32 ++++++++++++++------------------
> 1 file changed, 14 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c
> index 50beb407dbe9..08a3960001ad 100644
> --- a/drivers/mmc/host/sdhci-msm.c
> +++ b/drivers/mmc/host/sdhci-msm.c
> @@ -328,7 +328,7 @@ static void sdhci_msm_v5_variant_writel_relaxed(u32 val,
> writel_relaxed(val, host->ioaddr + offset);
> }
>
> -static unsigned int msm_get_clock_rate_for_bus_mode(struct sdhci_host *host,
> +static unsigned int msm_get_clock_mult_for_bus_mode(struct sdhci_host *host,
> unsigned int clock)
nit: clock variable not being used anymore. We can drop it.
> {
> struct mmc_ios ios = host->mmc->ios;
> @@ -342,8 +342,8 @@ static unsigned int msm_get_clock_rate_for_bus_mode(struct sdhci_host *host,
> ios.timing == MMC_TIMING_MMC_DDR52 ||
> ios.timing == MMC_TIMING_MMC_HS400 ||
> host->flags & SDHCI_HS400_TUNING)
> - clock *= 2;
> - return clock;
> + return 2;
> + return 1;
> }
>
> static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host,
> @@ -354,14 +354,16 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host,
> struct mmc_ios curr_ios = host->mmc->ios;
> struct clk *core_clk = msm_host->bulk_clks[0].clk;
> unsigned long achieved_rate;
> + unsigned int desired_rate;
> + unsigned int mult;
> int rc;
>
> - clock = msm_get_clock_rate_for_bus_mode(host, clock);
> - rc = dev_pm_opp_set_rate(mmc_dev(host->mmc), clock);
> + mult = msm_get_clock_mult_for_bus_mode(host, clock);
> + desired_rate = clock * mult;
> + rc = dev_pm_opp_set_rate(mmc_dev(host->mmc), desired_rate);
> if (rc) {
> pr_err("%s: Failed to set clock at rate %u at timing %d\n",
> - mmc_hostname(host->mmc), clock,
> - curr_ios.timing);
> + mmc_hostname(host->mmc), desired_rate, curr_ios.timing);
> return;
> }
>
> @@ -371,11 +373,12 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host,
> * encounter it.
> */
> achieved_rate = clk_get_rate(core_clk);
> - if (achieved_rate > clock)
> + if (achieved_rate > desired_rate)
> pr_warn("%s: Card appears overclocked; req %u Hz, actual %lu Hz\n",
> - mmc_hostname(host->mmc), clock, achieved_rate);
> + mmc_hostname(host->mmc), desired_rate, achieved_rate);
> + host->mmc->actual_clock = achieved_rate / mult;
>
> - msm_host->clk_rate = clock;
> + msm_host->clk_rate = desired_rate;
Can you set msm_host->clk_rate also to achieved_rate?
At few places in this driver, host->clock is being used where
achieved_rate should be used ideally.
I will replace those instances with 'msm_host->clk_rate' in a separate
patch once this change merged.
> pr_debug("%s: Setting clock at rate %lu at timing %d\n",
> mmc_hostname(host->mmc), achieved_rate, curr_ios.timing);
> }
> @@ -1756,13 +1759,6 @@ static unsigned int sdhci_msm_get_min_clock(struct sdhci_host *host)
> static void __sdhci_msm_set_clock(struct sdhci_host *host, unsigned int clock)
> {
> u16 clk;
> - /*
> - * Keep actual_clock as zero -
> - * - since there is no divider used so no need of having actual_clock.
> - * - MSM controller uses SDCLK for data timeout calculation. If
> - * actual_clock is zero, host->clock is taken for calculation.
> - */
> - host->mmc->actual_clock = 0;
>
> sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
>
> @@ -1785,7 +1781,7 @@ static void sdhci_msm_set_clock(struct sdhci_host *host, unsigned int clock)
> struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
>
> if (!clock) {
> - msm_host->clk_rate = clock;
> + host->mmc->actual_clock = msm_host->clk_rate = 0;
> goto out;
> }
>
next prev parent reply other threads:[~2020-12-14 12:45 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-11 17:12 [PATCH v4 1/2] mmc: sdhci-msm: Warn about overclocking SD/MMC Douglas Anderson
2020-12-11 17:12 ` [PATCH v4 2/2] mmc: sdhci-msm: Actually set the actual clock Douglas Anderson
2020-12-14 12:43 ` Veerabhadrarao Badiganti [this message]
2020-12-14 17:22 ` Doug Anderson
2020-12-14 16:42 ` Bjorn Andersson
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=e817284c-1ae9-7d3f-5195-7313651ef9da@codeaurora.org \
--to=vbadigan@codeaurora.org \
--cc=adrian.hunter@intel.com \
--cc=agross@kernel.org \
--cc=bjorn.andersson@linaro.org \
--cc=dianders@chromium.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mmc@vger.kernel.org \
--cc=swboyd@chromium.org \
--cc=tdas@codeaurora.org \
--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).