| Message ID | 20201214092048.v5.1.Iec3430c7d3c2a29262695edef7b82a14aaa567e5@changeid |
|---|---|
| State | New |
| Headers | show |
| Series | [v5,1/2] mmc: sdhci-msm: Warn about overclocking SD/MMC | expand |
On 14/12/20 7:21 pm, Douglas Anderson wrote: > As talked about in commit 5e4b7e82d497 ("clk: qcom: gcc-sdm845: Use > floor ops for sdcc clks"), most clocks handled by the Qualcomm clock > drivers are rounded _up_ by default instead of down. We should make > sure SD/MMC clocks are always rounded down in the clock drivers. > Let's add a warning in the Qualcomm SDHCI driver to help catch the > problem. > > This would have saved a bunch of time [1]. > > NOTE: this doesn't actually fix any problems, it just makes it obvious > to devs that there is a problem and that should be an indication to > fix the clock driver. > > [1] http://lore.kernel.org/r/20201210102234.1.I096779f219625148900fc984dd0084ed1ba87c7f@changeid > > Suggested-by: Stephen Boyd <swboyd@chromium.org> > Signed-off-by: Douglas Anderson <dianders@chromium.org> > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> > --- > > (no changes since v4) > > Changes in v4: > - Emphasize in the commit message that this itself doesn't fix anything. > > Changes in v3: > - Proper printf format code. > > Changes in v2: > - Store rate in unsigned long, not unsigned int. > - Reuse the clk_get_rate() in the later print. > > drivers/mmc/host/sdhci-msm.c | 15 +++++++++++++-- > 1 file changed, 13 insertions(+), 2 deletions(-) > > diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c > index 3451eb325513..50beb407dbe9 100644 > --- a/drivers/mmc/host/sdhci-msm.c > +++ b/drivers/mmc/host/sdhci-msm.c > @@ -353,6 +353,7 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, > struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host); > struct mmc_ios curr_ios = host->mmc->ios; > struct clk *core_clk = msm_host->bulk_clks[0].clk; > + unsigned long achieved_rate; > int rc; > > clock = msm_get_clock_rate_for_bus_mode(host, clock); > @@ -363,10 +364,20 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, > curr_ios.timing); > return; > } > + > + /* > + * Qualcomm clock drivers by default round clock _up_ if they can't > + * make the requested rate. This is not good for SD. Yell if we > + * encounter it. > + */ > + achieved_rate = clk_get_rate(core_clk); > + if (achieved_rate > clock) > + pr_warn("%s: Card appears overclocked; req %u Hz, actual %lu Hz\n", > + mmc_hostname(host->mmc), clock, achieved_rate); > + > msm_host->clk_rate = clock; > pr_debug("%s: Setting clock at rate %lu at timing %d\n", > - mmc_hostname(host->mmc), clk_get_rate(core_clk), > - curr_ios.timing); > + mmc_hostname(host->mmc), achieved_rate, curr_ios.timing); > } > > /* Platform specific tuning */ >
On 14/12/20 7:21 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. > > Suggested-by: Adrian Hunter <adrian.hunter@intel.com> > Signed-off-by: Douglas Anderson <dianders@chromium.org> > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Adrian Hunter <adrian.hunter@intel.com> > --- > > Changes in v5: > - Remove unused clock parameter. > - Add a comment that we're stashing the requested rate. > > Changes in v4: > - ("mmc: sdhci-msm: Actually set the actual clock") new for v4. > > drivers/mmc/host/sdhci-msm.c | 35 ++++++++++++++++------------------- > 1 file changed, 16 insertions(+), 19 deletions(-) > > diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c > index 50beb407dbe9..f5669dc858d0 100644 > --- a/drivers/mmc/host/sdhci-msm.c > +++ b/drivers/mmc/host/sdhci-msm.c > @@ -328,8 +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, > - unsigned int clock) > +static unsigned int msm_get_clock_mult_for_bus_mode(struct sdhci_host *host) > { > struct mmc_ios ios = host->mmc->ios; > /* > @@ -342,8 +341,8 @@ 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 +353,16 @@ 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); > + 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 +372,14 @@ 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; > + > + /* Stash the rate we requested to use in sdhci_msm_runtime_resume() */ > + msm_host->clk_rate = desired_rate; > > - msm_host->clk_rate = clock; > pr_debug("%s: Setting clock at rate %lu at timing %d\n", > mmc_hostname(host->mmc), achieved_rate, curr_ios.timing); > } > @@ -1756,13 +1760,6 @@ 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 +1782,7 @@ 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; > } > >
On 12/14/2020 10:51 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. > > Suggested-by: Adrian Hunter <adrian.hunter@intel.com> > Signed-off-by: Douglas Anderson <dianders@chromium.org> > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Reviewed-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org> > --- > > Changes in v5: > - Remove unused clock parameter. > - Add a comment that we're stashing the requested rate. > > Changes in v4: > - ("mmc: sdhci-msm: Actually set the actual clock") new for v4. > > drivers/mmc/host/sdhci-msm.c | 35 ++++++++++++++++------------------- > 1 file changed, 16 insertions(+), 19 deletions(-) > > diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c > index 50beb407dbe9..f5669dc858d0 100644 > --- a/drivers/mmc/host/sdhci-msm.c > +++ b/drivers/mmc/host/sdhci-msm.c > @@ -328,8 +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, > - unsigned int clock) > +static unsigned int msm_get_clock_mult_for_bus_mode(struct sdhci_host *host) > { > struct mmc_ios ios = host->mmc->ios; > /* > @@ -342,8 +341,8 @@ 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 +353,16 @@ 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); > + 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 +372,14 @@ 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; > + > + /* Stash the rate we requested to use in sdhci_msm_runtime_resume() */ > + msm_host->clk_rate = desired_rate; > > - msm_host->clk_rate = clock; > pr_debug("%s: Setting clock at rate %lu at timing %d\n", > mmc_hostname(host->mmc), achieved_rate, curr_ios.timing); > } > @@ -1756,13 +1760,6 @@ 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 +1782,7 @@ 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; > } >
Ulf, On Mon, Dec 14, 2020 at 9:23 AM Douglas Anderson <dianders@chromium.org> wrote: > > As talked about in commit 5e4b7e82d497 ("clk: qcom: gcc-sdm845: Use > floor ops for sdcc clks"), most clocks handled by the Qualcomm clock > drivers are rounded _up_ by default instead of down. We should make > sure SD/MMC clocks are always rounded down in the clock drivers. > Let's add a warning in the Qualcomm SDHCI driver to help catch the > problem. > > This would have saved a bunch of time [1]. > > NOTE: this doesn't actually fix any problems, it just makes it obvious > to devs that there is a problem and that should be an indication to > fix the clock driver. > > [1] http://lore.kernel.org/r/20201210102234.1.I096779f219625148900fc984dd0084ed1ba87c7f@changeid > > Suggested-by: Stephen Boyd <swboyd@chromium.org> > Signed-off-by: Douglas Anderson <dianders@chromium.org> > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> > --- > > (no changes since v4) > > Changes in v4: > - Emphasize in the commit message that this itself doesn't fix anything. > > Changes in v3: > - Proper printf format code. > > Changes in v2: > - Store rate in unsigned long, not unsigned int. > - Reuse the clk_get_rate() in the later print. > > drivers/mmc/host/sdhci-msm.c | 15 +++++++++++++-- > 1 file changed, 13 insertions(+), 2 deletions(-) Is there anything you need me to do for this patch and the next one? They are both reviewed / Acked and hopefully have sat around long enough that folks who took a long holiday break had a chance to shout if they were going to, so I think they could land. ;-) Please yell if there's something you need me to do, or feel free to tell me to sit quietly and be patient. Thanks! -Doug
On Mon, 14 Dec 2020 at 18:23, Douglas Anderson <dianders@chromium.org> wrote: > > As talked about in commit 5e4b7e82d497 ("clk: qcom: gcc-sdm845: Use > floor ops for sdcc clks"), most clocks handled by the Qualcomm clock > drivers are rounded _up_ by default instead of down. We should make > sure SD/MMC clocks are always rounded down in the clock drivers. > Let's add a warning in the Qualcomm SDHCI driver to help catch the > problem. > > This would have saved a bunch of time [1]. > > NOTE: this doesn't actually fix any problems, it just makes it obvious > to devs that there is a problem and that should be an indication to > fix the clock driver. > > [1] http://lore.kernel.org/r/20201210102234.1.I096779f219625148900fc984dd0084ed1ba87c7f@changeid > > Suggested-by: Stephen Boyd <swboyd@chromium.org> > Signed-off-by: Douglas Anderson <dianders@chromium.org> > Reviewed-by: Stephen Boyd <swboyd@chromium.org> > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Applied for next, thanks! Kind regards Uffe > --- > > (no changes since v4) > > Changes in v4: > - Emphasize in the commit message that this itself doesn't fix anything. > > Changes in v3: > - Proper printf format code. > > Changes in v2: > - Store rate in unsigned long, not unsigned int. > - Reuse the clk_get_rate() in the later print. > > drivers/mmc/host/sdhci-msm.c | 15 +++++++++++++-- > 1 file changed, 13 insertions(+), 2 deletions(-) > > diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c > index 3451eb325513..50beb407dbe9 100644 > --- a/drivers/mmc/host/sdhci-msm.c > +++ b/drivers/mmc/host/sdhci-msm.c > @@ -353,6 +353,7 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, > struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host); > struct mmc_ios curr_ios = host->mmc->ios; > struct clk *core_clk = msm_host->bulk_clks[0].clk; > + unsigned long achieved_rate; > int rc; > > clock = msm_get_clock_rate_for_bus_mode(host, clock); > @@ -363,10 +364,20 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, > curr_ios.timing); > return; > } > + > + /* > + * Qualcomm clock drivers by default round clock _up_ if they can't > + * make the requested rate. This is not good for SD. Yell if we > + * encounter it. > + */ > + achieved_rate = clk_get_rate(core_clk); > + if (achieved_rate > clock) > + pr_warn("%s: Card appears overclocked; req %u Hz, actual %lu Hz\n", > + mmc_hostname(host->mmc), clock, achieved_rate); > + > msm_host->clk_rate = clock; > pr_debug("%s: Setting clock at rate %lu at timing %d\n", > - mmc_hostname(host->mmc), clk_get_rate(core_clk), > - curr_ios.timing); > + mmc_hostname(host->mmc), achieved_rate, curr_ios.timing); > } > > /* Platform specific tuning */ > -- > 2.29.2.576.ga3fc446d84-goog >
diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c index 3451eb325513..50beb407dbe9 100644 --- a/drivers/mmc/host/sdhci-msm.c +++ b/drivers/mmc/host/sdhci-msm.c @@ -353,6 +353,7 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host); struct mmc_ios curr_ios = host->mmc->ios; struct clk *core_clk = msm_host->bulk_clks[0].clk; + unsigned long achieved_rate; int rc; clock = msm_get_clock_rate_for_bus_mode(host, clock); @@ -363,10 +364,20 @@ static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host, curr_ios.timing); return; } + + /* + * Qualcomm clock drivers by default round clock _up_ if they can't + * make the requested rate. This is not good for SD. Yell if we + * encounter it. + */ + achieved_rate = clk_get_rate(core_clk); + if (achieved_rate > clock) + pr_warn("%s: Card appears overclocked; req %u Hz, actual %lu Hz\n", + mmc_hostname(host->mmc), clock, achieved_rate); + msm_host->clk_rate = clock; pr_debug("%s: Setting clock at rate %lu at timing %d\n", - mmc_hostname(host->mmc), clk_get_rate(core_clk), - curr_ios.timing); + mmc_hostname(host->mmc), achieved_rate, curr_ios.timing); } /* Platform specific tuning */
As talked about in commit 5e4b7e82d497 ("clk: qcom: gcc-sdm845: Use floor ops for sdcc clks"), most clocks handled by the Qualcomm clock drivers are rounded _up_ by default instead of down. We should make sure SD/MMC clocks are always rounded down in the clock drivers. Let's add a warning in the Qualcomm SDHCI driver to help catch the problem. This would have saved a bunch of time [1]. NOTE: this doesn't actually fix any problems, it just makes it obvious to devs that there is a problem and that should be an indication to fix the clock driver. [1] http://lore.kernel.org/r/20201210102234.1.I096779f219625148900fc984dd0084ed1ba87c7f@changeid Suggested-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> --- (no changes since v4) Changes in v4: - Emphasize in the commit message that this itself doesn't fix anything. Changes in v3: - Proper printf format code. Changes in v2: - Store rate in unsigned long, not unsigned int. - Reuse the clk_get_rate() in the later print. drivers/mmc/host/sdhci-msm.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)