linux-kernel - Re: [PATCH] mmc: dw_mmc: rewrite CLKDIV computation

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:	Tue, 26 Mar 2013 15:53:50 -0700
From:	Grant Grundler <grundler@...omium.org>
To:	Chris Ball <cjb@...top.org>
Cc:	Doug Anderson <dianders@...omium.org>,
	Will Newton <will.newton@...il.com>,
	Seungwon Jeon <tgih.jun@...sung.com>,
	Bing Zhao <bzhao@...vell.com>,
	Jaehoon Chung <jh80.chung@...sung.com>,
	Ashok Nagarajan <asnagarajan@...omium.org>,
	Olof Johansson <olof@...om.net>,
	"linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] mmc: dw_mmc: rewrite CLKDIV computation

I've attached the test program I wrote to compare the different
flavors of CLKDIV computation: old (3.4 kernel), current upstream, and
my rewrite.

thanks
grant

On Tue, Mar 26, 2013 at 3:50 PM, Grant Grundler <grundler@...omium.org> wrote:
> Last year Seungwon Jeon (Samsung) fixed a bug in CLKDIV computation.
> But when debugging a related issue (http://crbug.com/221828) I found
> the code unreadable.  This rewrite simplifies the computation and
> explains each step.
>
> Signed-off-by: Grant Grundler <grundler@...omium.org>
> ---
> Tested on Samsung Series 3 Chromebook (exynos 5250 chipset) using
> ChromeOS 3.4 kernel (not 3.9-rc3 which this patch is based against).
>
> I've written a test program to confirm this patch generates the
> same correct values and will share that separately.
>
>  drivers/mmc/host/dw_mmc.c | 22 +++++++++++++++-------
>  1 file changed, 15 insertions(+), 7 deletions(-)
>
>
> diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
> index 9834221..6fdf55b 100644
> --- a/drivers/mmc/host/dw_mmc.c
> +++ b/drivers/mmc/host/dw_mmc.c
> @@ -631,14 +631,22 @@ static void dw_mci_setup_bus(struct dw_mci_slot *slot, bool force_clkinit)
>
>         if (slot->clock != host->current_speed || force_clkinit) {
>                 div = host->bus_hz / slot->clock;
> -               if (host->bus_hz % slot->clock && host->bus_hz > slot->clock)
> -                       /*
> -                        * move the + 1 after the divide to prevent
> -                        * over-clocking the card.
> +               if (host->bus_hz > slot->clock) {
> +                       /* don't overclock due to integer math losses */
> +                       if ((div * slot->clock) < host->bus_hz)
> +                               div++;
> +
> +                       /* don't overclock due to resolution of HW */
> +                       if (div & 1)
> +                               div++;
> +
> +                       /* See 6.2.3 CLKDIV in "Mobile Storage Host Databook"
> +                        * Look for dwc_mobile_storage_db.pdf from Synopsys.
> +                        * CLKDIV value 0 means divisor 1, val 1 -> 2, ...
>                          */
> -                       div += 1;
> -
> -               div = (host->bus_hz != slot->clock) ? DIV_ROUND_UP(div, 2) : 0;
> +                       div /= 2;
> +               } else
> +                       div = 0;        /* use bus_hz */
>
>                 dev_info(&slot->mmc->class_dev,
>                          "Bus speed (slot %d) = %dHz (slot req %dHz, actual %dHZ"
> --
> 1.8.1.3
>

View attachment "test_dw_mmc.c" of type "text/x-csrc" (2894 bytes)