linux-kernel - Re: [PATCH] mmc: block: workaround long ioctl busy timeout

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]

Date:   Mon, 13 Feb 2023 14:38:02 +0100
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     Christian Löhle <CLoehle@...erstone.com>
Cc:     "adrian.hunter@...el.com" <adrian.hunter@...el.com>,
        Avri Altman <Avri.Altman@....com>,
        "linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] mmc: block: workaround long ioctl busy timeout

On Mon, 16 Jan 2023 at 15:38, Christian Löhle <CLoehle@...erstone.com> wrote:
>
> The ioctl interface allowed to set cmd_timeout_ms when polling for
> busy on R1B commands. This was often limited by the max hw timeout
> so work around it like for the sanitize command.
>
> Signed-off-by: Christian Loehle <cloehle@...erstone.com>
> ---
>  drivers/mmc/core/block.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
> index 20da7ed43e6d..ba3bc9014179 100644
> --- a/drivers/mmc/core/block.c
> +++ b/drivers/mmc/core/block.c
> @@ -472,6 +472,8 @@ static int __mmc_blk_ioctl_cmd(struct mmc_card *card, struct mmc_blk_data *md,
>         struct scatterlist sg;
>         int err;
>         unsigned int target_part;
> +       unsigned int busy_timeout = MMC_BLK_TIMEOUT_MS;
> +       int poll_prog = false;
>
>         if (!card || !md || !idata)
>                 return -EINVAL;
> @@ -493,6 +495,12 @@ static int __mmc_blk_ioctl_cmd(struct mmc_card *card, struct mmc_blk_data *md,
>         cmd.opcode = idata->ic.opcode;
>         cmd.arg = idata->ic.arg;
>         cmd.flags = idata->ic.flags;
> +       /* R1B flag might be removed here to work around hw, so save it */
> +       poll_prog = (idata->rpmb || (cmd.flags & MMC_RSP_R1B) == MMC_RSP_R1B);
> +       busy_timeout = idata->ic.cmd_timeout_ms ? :
> +               MMC_BLK_TIMEOUT_MS;
> +       if (poll_prog)
> +               mmc_prepare_busy_cmd(card->host, &cmd, busy_timeout);

This may end up using R1B for rpmb commands, we don't want that.

>
>         if (idata->buf_bytes) {
>                 data.sg = &sg;
> @@ -596,7 +604,7 @@ static int __mmc_blk_ioctl_cmd(struct mmc_card *card, struct mmc_blk_data *md,
>         if (idata->ic.postsleep_min_us)
>                 usleep_range(idata->ic.postsleep_min_us, idata->ic.postsleep_max_us);
>
> -       if (idata->rpmb || (cmd.flags & MMC_RSP_R1B) == MMC_RSP_R1B) {
> +       if (poll_prog) {

If we end up using R1B and the host supports HW busy detection, we
should skip polling.

>                 /*
>                  * Ensure RPMB/R1B command has completed by polling CMD13 "Send Status". Here we
>                  * allow to override the default timeout value if a custom timeout is specified.

I believe I understand the problem you are trying to address in the
$subject patch. Let me try to help out, by sending a re-worked patch
to try to cover the things I pointed out above in a more consistent
way.

Kind regards
Uffe