linux-kernel - Re: [PATCH] mmc: sdhci: use udelay instead of mdelay

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, 30 May 2016 10:00:42 +0200
From:	Arnd Bergmann <arnd@...db.de>
To:	"Baranowska, BeataX" <beatax.baranowska@...el.com>
Cc:	"Hunter, Adrian" <adrian.hunter@...el.com>,
	Ulf Hansson <ulf.hansson@...aro.org>,
	"linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"Dong, Chuanxiao" <chuanxiao.dong@...el.com>,
	"Jarosz, SebastianX" <sebastianx.jarosz@...el.com>
Subject: Re: [PATCH] mmc: sdhci: use udelay instead of mdelay

On Monday, May 30, 2016 7:55:55 AM CEST Baranowska, BeataX wrote:
> From: Chuanxiao Dong <chuanxiao.dong@...el.com>
> 
> This patch will use udelay instead of mdelay when waiting for
> SDHCI hardware to be stable. udelay can help to reduce the waiting
> time when is in critical region which is protected by spinlock.
> 
> With this patch, __sdhci_set_ios only take a few microseconds to be
> done.
> 
> Signed-off-by: Chuanxiao Dong <chuanxiao.dong@...el.com>
> ---
> drivers/mmc/host/sdhci.c | 18 +++++++++---------
>  1 file changed, 9 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
> index e010ea4eb6f5..56d2c7567d97 100644
> --- a/drivers/mmc/host/sdhci.c
> +++ b/drivers/mmc/host/sdhci.c
> @@ -173,8 +173,8 @@ void sdhci_reset(struct sdhci_host *host, u8 mask)
>                         sdhci_runtime_pm_bus_off(host);
>         }
> 
> -       /* Wait max 100 ms */
> -       timeout = 100;
> +       /* Wait max 10000 ms */
> +       timeout = 10000;
> 
>         /* hw clears the bit when it's done */
>         while (sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask) {
> @@ -185,7 +185,7 @@ void sdhci_reset(struct sdhci_host *host, u8 mask)
>                         return;
>                 }
>                 timeout--;
> -               mdelay(1);
> +               udelay(10);
>         }
>  }
>  EXPORT_SYMBOL_GPL(sdhci_reset);

This can significantly increase the timeout length. I think you should
instead use time_before() to see how many jiffies have passed since
the start.

However, the real question is why the reset function gets called under
a spinlock in the first place. Can you try to rearrange the code so
it doesn't need the lock at all and you can just use msleep() instead?

	Arnd