linux-kernel - Re: [PATCH v5 1/6] mmc: core: Add helper for parsing clock phase properties

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAPDyKFrceNPNz9+88p+mzbYEo-ZqWOwTBWaqycxPr3MQEFtbaA@mail.gmail.com>
Date:   Mon, 14 Dec 2020 16:48:54 +0100
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     Andrew Jeffery <andrew@...id.au>
Cc:     "linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
        Rob Herring <robh+dt@...nel.org>,
        Joel Stanley <joel@....id.au>,
        Adrian Hunter <adrian.hunter@...el.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        DTML <devicetree@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        linux-aspeed <linux-aspeed@...ts.ozlabs.org>,
        ryan_chen@...eedtech.com
Subject: Re: [PATCH v5 1/6] mmc: core: Add helper for parsing clock phase properties

On Tue, 8 Dec 2020 at 02:26, Andrew Jeffery <andrew@...id.au> wrote:
>
> Drivers for MMC hosts that accept phase corrections can take advantage
> of the helper by embedding a mmc_clk_phase_map_t object in their
> private data and invoking mmc_of_parse_clk_phase() to extract phase
> parameters. It is the responsibility of the host driver to translate and
> apply the extracted values to hardware as required.
>
> Signed-off-by: Andrew Jeffery <andrew@...id.au>
> ---
>  drivers/mmc/core/host.c  | 44 ++++++++++++++++++++++++++++++++++++++++
>  include/linux/mmc/host.h | 17 ++++++++++++++++
>  2 files changed, 61 insertions(+)
>
> diff --git a/drivers/mmc/core/host.c b/drivers/mmc/core/host.c
> index 96b2ca1f1b06..b1697f00c4b5 100644
> --- a/drivers/mmc/core/host.c
> +++ b/drivers/mmc/core/host.c
> @@ -163,6 +163,50 @@ static void mmc_retune_timer(struct timer_list *t)
>         mmc_retune_needed(host);
>  }
>
> +static void mmc_of_parse_timing_phase(struct device *dev, const char *prop,
> +                                     struct mmc_clk_phase *phase)
> +{
> +       int degrees[2] = {0};
> +       int rc;
> +
> +       rc = device_property_read_u32_array(dev, prop, degrees, 2);
> +       phase->valid = !rc;
> +       if (phase->valid) {
> +               phase->in_deg = degrees[0];
> +               phase->out_deg = degrees[1];
> +       }
> +}
> +
> +void
> +mmc_of_parse_clk_phase(struct mmc_host *host, mmc_clk_phase_map_t map)

Would you mind to change to pass a "struct mmc_clk_phase_map *map" to this?

See more comments below.

> +{
> +       struct device *dev = host->parent;
> +
> +       mmc_of_parse_timing_phase(dev, "clk-phase-legacy",
> +                                 &map[MMC_TIMING_LEGACY]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs",
> +                                 &map[MMC_TIMING_MMC_HS]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-sd-hs",
> +                                 &map[MMC_TIMING_SD_HS]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr12",
> +                                 &map[MMC_TIMING_UHS_SDR12]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr25",
> +                                 &map[MMC_TIMING_UHS_SDR25]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr50",
> +                                 &map[MMC_TIMING_UHS_SDR50]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr104",
> +                                 &map[MMC_TIMING_UHS_SDR104]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-uhs-ddr50",
> +                                 &map[MMC_TIMING_UHS_DDR50]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-mmc-ddr52",
> +                                 &map[MMC_TIMING_MMC_DDR52]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs200",
> +                                 &map[MMC_TIMING_MMC_HS200]);
> +       mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs400",
> +                                 &map[MMC_TIMING_MMC_HS400]);
> +}
> +EXPORT_SYMBOL(mmc_of_parse_clk_phase);
> +
>  /**
>   *     mmc_of_parse() - parse host's device-tree node
>   *     @host: host whose node should be parsed.
> diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
> index 01bba36545c5..bc4731c9738f 100644
> --- a/include/linux/mmc/host.h
> +++ b/include/linux/mmc/host.h
> @@ -79,6 +79,22 @@ struct mmc_ios {
>         bool enhanced_strobe;                   /* hs400es selection */
>  };
>
> +struct mmc_clk_phase {
> +       bool valid;
> +       u16 in_deg;
> +       u16 out_deg;
> +};
> +
> +/*
> + * Define a type to map between bus timings and phase correction values. To
> + * avoid bloat in struct mmc_host we leave it to the host driver to define the
> + * phase map object in its private data if it supports phase correction.
> + * However, mmc_of_parse_clk_phase() is provided by the mmc core and needs the
> + * provided array to be correctly sized, so typedef an appropriately sized
> + * array to minimise the chance that the wrong size object is passed.
> + */
> +typedef struct mmc_clk_phase mmc_clk_phase_map_t[MMC_TIMING_MMC_HS400 + 1];
> +

Nitpick: I would appreciate if we could avoid using "typedefs", as I
think they in many cases makes the code harder to read. How about
doing this instead?

#define MMC_NUM_CLK_PHASES (MMC_TIMING_MMC_HS400 + 1)

struct mmc_clk_phase_map {
        struct mmc_clk_phase phase[MMC_NUM_CLK_PHASES];
};

[...]

Kind regards
Uffe