linux-kernel - Re: [PATCH] mmc: core: Improve fallback to speed modes if eMMC HS200 fails

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <CAPDyKFod3LdfRRgVfdASqgSBAXqJnGkdKMM9qKnAYcfUQptF_g@mail.gmail.com>
Date:   Mon, 14 Mar 2022 11:54:06 +0100
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     Heiner Kallweit <hkallweit1@...il.com>
Cc:     linux-mmc@...r.kernel.org, Adrian Hunter <adrian.hunter@...el.com>,
        Wolfram Sang <wsa+renesas@...g-engineering.com>,
        Yann Gautier <yann.gautier@...s.st.com>,
        linux-kernel@...r.kernel.org, Ulf Hansson <ulf.hansson@...aro.org>
Subject: Re: [PATCH] mmc: core: Improve fallback to speed modes if eMMC HS200 fails

On Thu, 3 Mar 2022 at 17:45, Ulf Hansson <ulf.hansson@...aro.org> wrote:
>
> In the error path of mmc_select_hs200() we are trying our best to restore
> the card/host into a valid state. This makes sense, especially if we
> encounter a simple switch error (-EBADMSG). However, rather than then
> continue with using the legacy speed mode, let's try the other better speed
> modes first. Additionally, let's update the card->mmc_avail_type to avoid
> us from trying a broken HS200 mode again.
>
> In an Amlogic S905W based TV box where the switch to HS200 mode fails for
> the eMMC, this allows us to use the eMMC in DDR mode in favor of the legacy
> mode, which greatly improves the performance.
>
> Suggested-by: Heiner Kallweit <hkallweit1@...il.com>
> Signed-off-by: Ulf Hansson <ulf.hansson@...aro.org>

Heiner, does this solve your problems? Can you perhaps run some tests
at your side to see how it works for you?

Kind regards
Uffe

> ---
>  drivers/mmc/core/mmc.c | 16 +++++++++++++---
>  1 file changed, 13 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
> index 141f851c9f58..6e7db45cb8e0 100644
> --- a/drivers/mmc/core/mmc.c
> +++ b/drivers/mmc/core/mmc.c
> @@ -1518,13 +1518,23 @@ static int mmc_select_timing(struct mmc_card *card)
>         if (!mmc_can_ext_csd(card))
>                 goto bus_speed;
>
> -       if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS400ES)
> +       if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS400ES) {
>                 err = mmc_select_hs400es(card);
> -       else if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS200)
> +               goto out;
> +       }
> +
> +       if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS200) {
>                 err = mmc_select_hs200(card);
> -       else if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS)
> +               if (err == -EBADMSG)
> +                       card->mmc_avail_type &= ~EXT_CSD_CARD_TYPE_HS200;
> +               else
> +                       goto out;
> +       }
> +
> +       if (card->mmc_avail_type & EXT_CSD_CARD_TYPE_HS)
>                 err = mmc_select_hs(card);
>
> +out:
>         if (err && err != -EBADMSG)
>                 return err;
>
> --
> 2.25.1
>