2014-06-10 14:57:16

by Roman Peniaev

Subject: [PATCH] mmc: mxs: fix card detection in case of 'broken-cd' flag set

In case of reboot my olinuxino imx23 board does not see
mmc card any more. mmc_rescan is being called by delayed
work in loop, but mxs_mmc_get_cd always returns 0, so we
will never pass the card detection check and will not do
further card inition.

This patch is just an attempt to partially revert the patch
a91fe279ae of Sascha Hauer, where it is claimed that upper
layer will handle broken card detection using the polling
logic and MMC_CAP_NEEDS_POLL capability, but seems it is not
true, because upper logic still expects 1 from 'get_cd'.

So, here we always return 1 (card present) in case of
MMC_CAP_NEEDS_POLL capability set.

Signed-off-by: Roman Pen <[email protected]>
CC: Chris Ball <[email protected]>
CC: Sascha Hauer <[email protected]>
CC: Fabio Estevam <[email protected]>
CC: Shawn Guo <[email protected]>
CC: Ulf Hansson <[email protected]>
CC: [email protected]
CC: [email protected]
---
drivers/mmc/host/mxs-mmc.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
index 073e871..9187ce1 100644
--- a/drivers/mmc/host/mxs-mmc.c
+++ b/drivers/mmc/host/mxs-mmc.c
@@ -82,7 +82,8 @@ static int mxs_mmc_get_cd(struct mmc_host *mmc)
if (ret >= 0)
return ret;

- present = !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
+ present = mmc->caps & MMC_CAP_NEEDS_POLL ||
+ !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
BM_SSP_STATUS_CARD_DETECT);

if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
--
1.9.1

2014-07-07 06:28:13

by Sascha Hauer

Subject: Re: [PATCH] mmc: mxs: fix card detection in case of 'broken-cd' flag set

On Tue, Jun 10, 2014 at 11:57:02PM +0900, Roman Pen wrote:
> In case of reboot my olinuxino imx23 board does not see
> mmc card any more. mmc_rescan is being called by delayed
> work in loop, but mxs_mmc_get_cd always returns 0, so we
> will never pass the card detection check and will not do
> further card inition.
>
> This patch is just an attempt to partially revert the patch
> a91fe279ae of Sascha Hauer, where it is claimed that upper
> layer will handle broken card detection using the polling
> logic and MMC_CAP_NEEDS_POLL capability, but seems it is not
> true, because upper logic still expects 1 from 'get_cd'.
>
> So, here we always return 1 (card present) in case of
> MMC_CAP_NEEDS_POLL capability set.
>
> Signed-off-by: Roman Pen <[email protected]>
> CC: Chris Ball <[email protected]>
> CC: Sascha Hauer <[email protected]>
> CC: Fabio Estevam <[email protected]>
> CC: Shawn Guo <[email protected]>
> CC: Ulf Hansson <[email protected]>
> CC: [email protected]
> CC: [email protected]

I still didn't find the time to take a proper look at it. Since this
fixes a bug for you:

Acked-by: Sascha Hauer <[email protected]>

Sascha

> ---
> drivers/mmc/host/mxs-mmc.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
> index 073e871..9187ce1 100644
> --- a/drivers/mmc/host/mxs-mmc.c
> +++ b/drivers/mmc/host/mxs-mmc.c
> @@ -82,7 +82,8 @@ static int mxs_mmc_get_cd(struct mmc_host *mmc)
> if (ret >= 0)
> return ret;
>
> - present = !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
> + present = mmc->caps & MMC_CAP_NEEDS_POLL ||
> + !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
> BM_SSP_STATUS_CARD_DETECT);
>
> if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
> --
> 1.9.1
>
>

--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |

2014-07-08 12:41:38

by Ulf Hansson

Subject: Re: [PATCH] mmc: mxs: fix card detection in case of 'broken-cd' flag set

On 10 June 2014 16:57, Roman Pen <[email protected]> wrote:
> In case of reboot my olinuxino imx23 board does not see
> mmc card any more. mmc_rescan is being called by delayed
> work in loop, but mxs_mmc_get_cd always returns 0, so we
> will never pass the card detection check and will not do
> further card inition.
>
> This patch is just an attempt to partially revert the patch
> a91fe279ae of Sascha Hauer, where it is claimed that upper
> layer will handle broken card detection using the polling
> logic and MMC_CAP_NEEDS_POLL capability, but seems it is not
> true, because upper logic still expects 1 from 'get_cd'.
>
> So, here we always return 1 (card present) in case of
> MMC_CAP_NEEDS_POLL capability set.
>
> Signed-off-by: Roman Pen <[email protected]>
> CC: Chris Ball <[email protected]>
> CC: Sascha Hauer <[email protected]>
> CC: Fabio Estevam <[email protected]>
> CC: Shawn Guo <[email protected]>
> CC: Ulf Hansson <[email protected]>
> CC: [email protected]
> CC: [email protected]

Thanks! Applied for next.

Kind regards
Uffe

> ---
> drivers/mmc/host/mxs-mmc.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
> index 073e871..9187ce1 100644
> --- a/drivers/mmc/host/mxs-mmc.c
> +++ b/drivers/mmc/host/mxs-mmc.c
> @@ -82,7 +82,8 @@ static int mxs_mmc_get_cd(struct mmc_host *mmc)
> if (ret >= 0)
> return ret;
>
> - present = !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
> + present = mmc->caps & MMC_CAP_NEEDS_POLL ||
> + !(readl(ssp->base + HW_SSP_STATUS(ssp)) &
> BM_SSP_STATUS_CARD_DETECT);
>
> if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
> --
> 1.9.1
>