From: Geert Uytterhoeven <geert@linux-m68k.org> To: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Cc: Ulf Hansson <ulf.hansson@linaro.org>, linux-mmc <linux-mmc@vger.kernel.org>, Simon <horms@verge.net.au>, Magnus <magnus.damm@gmail.com>, Linux-SH <linux-sh@vger.kernel.org>, Laurent <laurent.pinchart@ideasonboard.com>, kobayashi <keita.kobayashi.ym@renesas.com> Subject: Re: [PATCH 3/3 v3] mmc: sh_mmcif: calculate best clock with parent clock Date: Tue, 21 Apr 2015 10:31:21 +0000 [thread overview] Message-ID: <CAMuHMdUM4Cizij6q+6o9mADF_UjkzpZSvsUCdG+Y=PHo=A_qLw@mail.gmail.com> (raw) In-Reply-To: <87r3rdubmv.wl%kuninori.morimoto.gx@renesas.com> Hi Morimoto-san, On Tue, Apr 21, 2015 at 10:31 AM, Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> wrote: > From: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> > > MMCIF IP on R-Car series has parent clock which can be set > several rate, and it was not implemented on old SH-Mobile series > (= SH-Mobile series parent clock was fixed rate) > R-Car series MMCIF can use more high speed access if it setup > parent clock. This patch adds parent clock setup method, > and r8a7790/r8a7791 can use it. > > renesas,mmcif already contain renesas,mmcif-r8a7790/r8a7791 on > compatible string. So there is no update for binding Document. > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> > Tested-by: Keita Kobayashi <keita.kobayashi.ym@renesas.com> > diff --git a/drivers/mmc/host/sh_mmcif.c b/drivers/mmc/host/sh_mmcif.c > index 5282c5b..6ae836b 100644 > --- a/drivers/mmc/host/sh_mmcif.c > +++ b/drivers/mmc/host/sh_mmcif.c > @@ -224,6 +225,12 @@ enum mmcif_wait_for { > MMCIF_WAIT_FOR_STOP, > }; > > +struct sh_mmcif_parent_clk { > + unsigned int max; /* if exist: R-Car has MMC CKCR on CPG */ > + unsigned int min; /* if exist: R-Car has MMC CKCR on CPG */ > + u32 clkdiv_map; /* see CE_CLK_CTRL::CLKDIV */ > +}; > + > struct sh_mmcif_host { > struct mmc_host *mmc; > struct mmc_request *mrq; > @@ -249,6 +256,7 @@ struct sh_mmcif_host { > bool ccs_enable; /* Command Completion Signal support */ > bool clk_ctrl2_enable; > struct mutex thread_lock; > + const struct sh_mmcif_parent_clk *parent_clk; > > /* DMA support */ > struct dma_chan *chan_rx; > @@ -257,8 +265,16 @@ struct sh_mmcif_host { > bool dma_active; > }; > > +static const struct sh_mmcif_parent_clk mmcif_gen2_parent_clk = { > + .max = 97500000, > + .min = 12187500, > + .clkdiv_map = 0x3ff, Shouldn't this come from private data in the CPG clock driver, which supplies the parent clock? Then the mmcif driver will be independent from the parent clock. > +}; > + > static const struct of_device_id mmcif_of_match[] = { > { .compatible = "renesas,sh-mmcif" }, > + { .compatible = "renesas,mmcif-r8a7790", .data = &mmcif_gen2_parent_clk}, > + { .compatible = "renesas,mmcif-r8a7791", .data = &mmcif_gen2_parent_clk}, > { } > }; > MODULE_DEVICE_TABLE(of, mmcif_of_match); > @@ -490,12 +506,51 @@ static void sh_mmcif_clock_control(struct sh_mmcif_host *host, unsigned int clk) > > if (!clk) > return; > - if (sup_pclk && clk = host->clk) > + > + if (host->parent_clk) { > + const struct sh_mmcif_parent_clk *pclk = host->parent_clk; > + unsigned int parent_freq, clkdiv, myclk, diff_min, diff; > + int i, j; > + > + /* FIXME */ > + if (pclk->max != pclk->min * (pclk->max / pclk->min)) { > + dev_err(&host->pd->dev, "not assumed parent clk\n"); > + return; > + } Why? > + parent_freq = 0; > + clkdiv = 0; > + diff_min = ~0; > + for (i = pclk->max / pclk->min; i > 0; i--) { > + for (j = fls(pclk->clkdiv_map); j >= 0; j--) { > + if (!((1 << j) & pclk->clkdiv_map)) > + continue; > + > + myclk = ((pclk->min * i) / (1 << (j + 1))); > + diff = (myclk > clk) ? myclk - clk : clk - myclk; > + > + if (diff <= diff_min) { > + parent_freq = pclk->min * i; > + clkdiv = j; > + diff_min = diff; > + } > + } > + } Shouldn't the above be handled by the CPG clock driver, through a clk API? > + dev_dbg(&host->pd->dev, "clk %d/%d (%d, %x)\n", "%u" (all four) > + (parent_freq / (1 << (clkdiv + 1))), clk, > + parent_freq, clkdiv); > + > + clk_set_rate(host->hclk, parent_freq); Note that the last parameter of clk_set_rate() is "unsigned long", while parent_freq is "unsigned int". > + sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, > + CLK_CLEAR & (clkdiv << 16)); > + } else if (sup_pclk && clk = host->clk) { > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_SUP_PCLK); > - else > + } else { > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR & > ((fls(DIV_ROUND_UP(host->clk, > clk) - 1) - 1) << 16)); > + } > > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_ENABLE); > } > @@ -982,10 +1037,23 @@ static int sh_mmcif_clk_update(struct sh_mmcif_host *host) > { > int ret = clk_prepare_enable(host->hclk); > > - if (!ret) { > + if (ret) > + return ret; > + > + if (!host->parent_clk) { > host->clk = clk_get_rate(host->hclk); > host->mmc->f_max = host->clk / 2; > host->mmc->f_min = host->clk / 512; > + } else { > + const struct sh_mmcif_parent_clk *pclk = host->parent_clk; > + > + host->clk = clk_get_rate(host->hclk); clk_get_rate() returns "unsigned long", while "host->clk" is "unsigned int". > + host->mmc->f_max = pclk->max / (1 << ffs(pclk->clkdiv_map)); > + host->mmc->f_min = pclk->min / (1 << fls(pclk->clkdiv_map)); > + > + dev_dbg(&host->pd->dev, "parent clk %d/%d, %d/%d\n", "%u" (all four) > + pclk->max, pclk->min, > + host->mmc->f_max, host->mmc->f_min); > } > > return ret; Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
WARNING: multiple messages have this Message-ID (diff)
From: Geert Uytterhoeven <geert@linux-m68k.org> To: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Cc: Ulf Hansson <ulf.hansson@linaro.org>, linux-mmc <linux-mmc@vger.kernel.org>, Simon <horms@verge.net.au>, Magnus <magnus.damm@gmail.com>, Linux-SH <linux-sh@vger.kernel.org>, Laurent <laurent.pinchart@ideasonboard.com>, kobayashi <keita.kobayashi.ym@renesas.com> Subject: Re: [PATCH 3/3 v3] mmc: sh_mmcif: calculate best clock with parent clock Date: Tue, 21 Apr 2015 12:31:21 +0200 [thread overview] Message-ID: <CAMuHMdUM4Cizij6q+6o9mADF_UjkzpZSvsUCdG+Y=PHo=A_qLw@mail.gmail.com> (raw) In-Reply-To: <87r3rdubmv.wl%kuninori.morimoto.gx@renesas.com> Hi Morimoto-san, On Tue, Apr 21, 2015 at 10:31 AM, Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> wrote: > From: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> > > MMCIF IP on R-Car series has parent clock which can be set > several rate, and it was not implemented on old SH-Mobile series > (= SH-Mobile series parent clock was fixed rate) > R-Car series MMCIF can use more high speed access if it setup > parent clock. This patch adds parent clock setup method, > and r8a7790/r8a7791 can use it. > > renesas,mmcif already contain renesas,mmcif-r8a7790/r8a7791 on > compatible string. So there is no update for binding Document. > > Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> > Tested-by: Keita Kobayashi <keita.kobayashi.ym@renesas.com> > diff --git a/drivers/mmc/host/sh_mmcif.c b/drivers/mmc/host/sh_mmcif.c > index 5282c5b..6ae836b 100644 > --- a/drivers/mmc/host/sh_mmcif.c > +++ b/drivers/mmc/host/sh_mmcif.c > @@ -224,6 +225,12 @@ enum mmcif_wait_for { > MMCIF_WAIT_FOR_STOP, > }; > > +struct sh_mmcif_parent_clk { > + unsigned int max; /* if exist: R-Car has MMC CKCR on CPG */ > + unsigned int min; /* if exist: R-Car has MMC CKCR on CPG */ > + u32 clkdiv_map; /* see CE_CLK_CTRL::CLKDIV */ > +}; > + > struct sh_mmcif_host { > struct mmc_host *mmc; > struct mmc_request *mrq; > @@ -249,6 +256,7 @@ struct sh_mmcif_host { > bool ccs_enable; /* Command Completion Signal support */ > bool clk_ctrl2_enable; > struct mutex thread_lock; > + const struct sh_mmcif_parent_clk *parent_clk; > > /* DMA support */ > struct dma_chan *chan_rx; > @@ -257,8 +265,16 @@ struct sh_mmcif_host { > bool dma_active; > }; > > +static const struct sh_mmcif_parent_clk mmcif_gen2_parent_clk = { > + .max = 97500000, > + .min = 12187500, > + .clkdiv_map = 0x3ff, Shouldn't this come from private data in the CPG clock driver, which supplies the parent clock? Then the mmcif driver will be independent from the parent clock. > +}; > + > static const struct of_device_id mmcif_of_match[] = { > { .compatible = "renesas,sh-mmcif" }, > + { .compatible = "renesas,mmcif-r8a7790", .data = &mmcif_gen2_parent_clk}, > + { .compatible = "renesas,mmcif-r8a7791", .data = &mmcif_gen2_parent_clk}, > { } > }; > MODULE_DEVICE_TABLE(of, mmcif_of_match); > @@ -490,12 +506,51 @@ static void sh_mmcif_clock_control(struct sh_mmcif_host *host, unsigned int clk) > > if (!clk) > return; > - if (sup_pclk && clk == host->clk) > + > + if (host->parent_clk) { > + const struct sh_mmcif_parent_clk *pclk = host->parent_clk; > + unsigned int parent_freq, clkdiv, myclk, diff_min, diff; > + int i, j; > + > + /* FIXME */ > + if (pclk->max != pclk->min * (pclk->max / pclk->min)) { > + dev_err(&host->pd->dev, "not assumed parent clk\n"); > + return; > + } Why? > + parent_freq = 0; > + clkdiv = 0; > + diff_min = ~0; > + for (i = pclk->max / pclk->min; i > 0; i--) { > + for (j = fls(pclk->clkdiv_map); j >= 0; j--) { > + if (!((1 << j) & pclk->clkdiv_map)) > + continue; > + > + myclk = ((pclk->min * i) / (1 << (j + 1))); > + diff = (myclk > clk) ? myclk - clk : clk - myclk; > + > + if (diff <= diff_min) { > + parent_freq = pclk->min * i; > + clkdiv = j; > + diff_min = diff; > + } > + } > + } Shouldn't the above be handled by the CPG clock driver, through a clk API? > + dev_dbg(&host->pd->dev, "clk %d/%d (%d, %x)\n", "%u" (all four) > + (parent_freq / (1 << (clkdiv + 1))), clk, > + parent_freq, clkdiv); > + > + clk_set_rate(host->hclk, parent_freq); Note that the last parameter of clk_set_rate() is "unsigned long", while parent_freq is "unsigned int". > + sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, > + CLK_CLEAR & (clkdiv << 16)); > + } else if (sup_pclk && clk == host->clk) { > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_SUP_PCLK); > - else > + } else { > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR & > ((fls(DIV_ROUND_UP(host->clk, > clk) - 1) - 1) << 16)); > + } > > sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_ENABLE); > } > @@ -982,10 +1037,23 @@ static int sh_mmcif_clk_update(struct sh_mmcif_host *host) > { > int ret = clk_prepare_enable(host->hclk); > > - if (!ret) { > + if (ret) > + return ret; > + > + if (!host->parent_clk) { > host->clk = clk_get_rate(host->hclk); > host->mmc->f_max = host->clk / 2; > host->mmc->f_min = host->clk / 512; > + } else { > + const struct sh_mmcif_parent_clk *pclk = host->parent_clk; > + > + host->clk = clk_get_rate(host->hclk); clk_get_rate() returns "unsigned long", while "host->clk" is "unsigned int". > + host->mmc->f_max = pclk->max / (1 << ffs(pclk->clkdiv_map)); > + host->mmc->f_min = pclk->min / (1 << fls(pclk->clkdiv_map)); > + > + dev_dbg(&host->pd->dev, "parent clk %d/%d, %d/%d\n", "%u" (all four) > + pclk->max, pclk->min, > + host->mmc->f_max, host->mmc->f_min); > } > > return ret; Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
next prev parent reply other threads:[~2015-04-21 10:31 UTC|newest] Thread overview: 95+ messages / expand[flat|nested] mbox.gz Atom feed top [not found] <873840a4ch.wl%kuninori.morimoto.gx@renesas.com> 2015-04-21 3:49 ` mmc: sh_mmcif: add PLL support Kuninori Morimoto 2015-04-21 3:50 ` [PATCH 1/3] mmc: sh_mmcif: move mmcif_of_match to upside Kuninori Morimoto 2015-04-21 3:50 ` [PATCH 2/3] mmc: sh_mmcif: cleanup to use dev instead of &pdev->dev Kuninori Morimoto 2015-04-21 3:51 ` [PATCH 3/3] mmc: sh_mmcif: calculate best clock with PLL Kuninori Morimoto 2015-04-21 7:43 ` Kuninori Morimoto 2015-04-21 7:51 ` Laurent Pinchart 2015-04-21 7:58 ` Kuninori Morimoto 2015-04-21 7:53 ` [PATCH 0/3 v2] mmc: sh_mmcif: add PLL support Kuninori Morimoto 2015-04-21 7:54 ` [PATCH 1/3 v2] mmc: sh_mmcif: move mmcif_of_match to upside Kuninori Morimoto 2015-04-21 7:54 ` [PATCH 2/3 v2] mmc: sh_mmcif: cleanup to use dev instead of &pdev->dev Kuninori Morimoto 2015-04-21 7:55 ` [PATCH 3/3 v2] mmc: sh_mmcif: calculate best clock with parent clock Kuninori Morimoto 2015-04-21 8:23 ` [PATCH 0/3 v2] mmc: sh_mmcif: add PLL support Kuninori Morimoto 2015-04-21 8:26 ` [PATCH 0/3 v3] " Kuninori Morimoto 2015-04-21 8:26 ` [PATCH 1/3 v3] mmc: sh_mmcif: move mmcif_of_match to upside Kuninori Morimoto 2015-04-21 10:07 ` Geert Uytterhoeven 2015-04-21 10:07 ` Geert Uytterhoeven 2015-04-21 8:27 ` [PATCH 2/3 v3] mmc: sh_mmcif: cleanup to use dev instead of &pdev->dev Kuninori Morimoto 2015-04-21 10:07 ` Geert Uytterhoeven 2015-04-21 10:07 ` Geert Uytterhoeven 2015-04-21 8:31 ` [PATCH 3/3 v3] mmc: sh_mmcif: calculate best clock with parent clock Kuninori Morimoto 2015-04-21 8:31 ` Kuninori Morimoto 2015-04-21 10:31 ` Geert Uytterhoeven [this message] 2015-04-21 10:31 ` Geert Uytterhoeven 2015-04-21 13:07 ` Laurent Pinchart 2015-04-21 13:07 ` Laurent Pinchart 2015-04-22 1:05 ` Kuninori Morimoto 2015-05-04 1:04 ` Laurent Pinchart 2015-05-04 1:04 ` Laurent Pinchart 2015-05-11 2:15 ` Kuninori Morimoto 2015-05-11 2:15 ` Kuninori Morimoto 2015-04-22 1:04 ` Kuninori Morimoto 2015-04-22 7:49 ` Geert Uytterhoeven 2015-04-22 7:49 ` Geert Uytterhoeven 2015-04-22 8:18 ` Ulf Hansson 2015-04-22 8:18 ` Ulf Hansson 2015-04-22 8:22 ` Geert Uytterhoeven 2015-04-22 8:22 ` Geert Uytterhoeven 2015-04-22 9:16 ` Kuninori Morimoto 2015-04-23 8:11 ` [PATCH 0/7 v4] mmc: sh_mmcif: add parent clk support Kuninori Morimoto 2015-04-23 8:13 ` [PATCH 1/7 v4] mmc: sh_mmcif: move mmcif_of_match to upside Kuninori Morimoto 2015-04-23 8:14 ` [PATCH 2/7 v4] mmc: sh_mmcif: cleanup to use dev instead of &pdev->dev Kuninori Morimoto 2015-04-23 8:15 ` [PATCH 3/7 v4] mmc: sh_mmcif: remove unnecessary int clk from struct sh_mmcif_host Kuninori Morimoto 2015-04-23 8:15 ` Kuninori Morimoto 2015-04-23 10:01 ` Geert Uytterhoeven 2015-04-23 10:01 ` Geert Uytterhoeven 2015-04-23 8:16 ` [PATCH 4/7 v4] mmc: sh_mmcif: separate sh_mmcif_clk_update() into setup and prepare Kuninori Morimoto 2015-04-23 10:00 ` Geert Uytterhoeven 2015-04-23 10:00 ` Geert Uytterhoeven 2015-04-23 8:17 ` [PATCH 5/7 v4] mmc: sh_mmcif: calculate best clock with parent clock Kuninori Morimoto 2015-04-23 8:17 ` Kuninori Morimoto 2015-05-12 10:22 ` Laurent Pinchart 2015-05-12 10:22 ` Laurent Pinchart 2015-05-13 0:08 ` Kuninori Morimoto 2015-04-23 8:18 ` [PATCH 6/7 v4] ARM: shmobile: r8a7790: add MMCIF parent clock range Kuninori Morimoto 2015-05-07 5:26 ` Simon Horman 2015-05-07 5:26 ` Simon Horman 2015-05-11 2:53 ` Kuninori Morimoto 2015-05-11 5:39 ` Simon Horman 2015-05-11 5:39 ` Simon Horman 2015-04-23 8:18 ` [PATCH 7/7 v4] ARM: shmobile: r8a7791: " Kuninori Morimoto 2015-04-23 10:07 ` [PATCH 0/7 v4] mmc: sh_mmcif: add parent clk support Laurent Pinchart 2015-04-23 10:07 ` Laurent Pinchart 2015-05-05 8:33 ` Ulf Hansson 2015-05-05 8:33 ` Ulf Hansson 2015-05-13 2:16 ` [PATCH 0/3 v5] " Kuninori Morimoto 2015-05-13 2:17 ` [PATCH 1/3 v5] mmc: sh_mmcif: add sh_mmcif_host_to_dev() macro and use it Kuninori Morimoto 2015-05-13 2:17 ` Kuninori Morimoto 2015-05-13 2:18 ` [PATCH 2/3 v5] mmc: sh_mmcif: use sh_mmcif_xxx prefix for all functions Kuninori Morimoto 2015-05-13 2:18 ` [PATCH 3/3 v5] mmc: sh_mmcif: calculate best clock with parent clock Kuninori Morimoto 2015-05-13 2:18 ` Kuninori Morimoto 2015-05-13 7:55 ` Geert Uytterhoeven 2015-05-13 7:55 ` Geert Uytterhoeven 2015-05-13 8:37 ` Kuninori Morimoto 2015-05-13 8:43 ` Geert Uytterhoeven 2015-05-13 8:43 ` Geert Uytterhoeven 2015-05-13 9:27 ` Kuninori Morimoto 2015-05-14 7:20 ` [PATCH 0/3 v6] mmc: sh_mmcif: add parent clk support Kuninori Morimoto 2015-05-14 7:21 ` [PATCH 1/5 v6] mmc: sh_mmcif: add sh_mmcif_host_to_dev() macro and use it Kuninori Morimoto 2015-05-14 7:21 ` Kuninori Morimoto 2015-05-14 7:21 ` [PATCH 2/5 v6] mmc: sh_mmcif: use sh_mmcif_xxx prefix for all functions Kuninori Morimoto 2015-05-14 7:22 ` [PATCH 3/5 v6] mmc: sh_mmcif: calculate best clock with parent clock Kuninori Morimoto 2015-05-14 7:22 ` Kuninori Morimoto 2015-05-14 7:23 ` [PATCH 4/5 v6] ARM: shmobile: r8a7790: add MMCIF max-frequency Kuninori Morimoto 2015-05-14 7:23 ` [PATCH 5/5 v6] ARM: shmobile: r8a7791: " Kuninori Morimoto 2015-05-22 14:01 ` [PATCH 0/3 v6] mmc: sh_mmcif: add parent clk support Ulf Hansson 2015-05-22 14:01 ` Ulf Hansson 2015-05-25 0:24 ` Kuninori Morimoto 2015-05-25 0:50 ` Simon Horman 2015-05-25 0:50 ` Simon Horman 2015-05-25 0:26 ` Simon Horman 2015-05-25 0:26 ` Simon Horman 2015-05-25 0:38 ` Kuninori Morimoto 2015-05-25 8:44 ` Ulf Hansson 2015-05-25 8:44 ` Ulf Hansson 2015-05-25 9:38 ` Kuninori Morimoto
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='CAMuHMdUM4Cizij6q+6o9mADF_UjkzpZSvsUCdG+Y=PHo=A_qLw@mail.gmail.com' \ --to=geert@linux-m68k.org \ --cc=horms@verge.net.au \ --cc=keita.kobayashi.ym@renesas.com \ --cc=kuninori.morimoto.gx@renesas.com \ --cc=laurent.pinchart@ideasonboard.com \ --cc=linux-mmc@vger.kernel.org \ --cc=linux-sh@vger.kernel.org \ --cc=magnus.damm@gmail.com \ --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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.