[PATCH] sdhci-s3c: fix incorrect spinlock usage after merge

[PATCH] sdhci-s3c: fix incorrect spinlock usage after merge

[Marek Szyprowski]

In the commit f522886e202a34a2191dd5d471b3c4d46410a9a0 a merge conflict
in the sdhci-s3c driver been fixed. However the fix used incorrect
spinlock operation - it cause a race with sdhci interrupt service. The
correct way to solve it is to use spin_lock_irqsave/irqrestore() calls.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
---
 drivers/mmc/host/sdhci-s3c.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c
index 71ad416..735d431 100644
--- a/drivers/mmc/host/sdhci-s3c.c
+++ b/drivers/mmc/host/sdhci-s3c.c
@@ -241,8 +241,10 @@ static struct sdhci_ops sdhci_s3c_ops = {
 static void sdhci_s3c_notify_change(struct platform_device *dev, int state)
 {
 	struct sdhci_host *host = platform_get_drvdata(dev);
+	unsigned long flags;
+
 	if (host) {
-		spin_lock(&host->lock);
+		spin_lock_irqsave(&host->lock, flags);
 		if (state) {
 			dev_dbg(&dev->dev, "card inserted.\n");
 			host->flags &= ~SDHCI_DEVICE_DEAD;
@@ -253,7 +255,7 @@ static void sdhci_s3c_notify_change(struct platform_device *dev, int state)
 			host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
 		}
 		tasklet_schedule(&host->card_tasklet);
-		spin_unlock(&host->lock);
+		spin_unlock_irqrestore(&host->lock, flags);
 	}
 }
 
-- 
1.7.1.569.g6f426

[PATCH] sdhci-s3c: fix incorrect spinlock usage after merge

[Chris Ball]

On Mon, Sep 20, 2010 at 03:03:42PM +0200, Marek Szyprowski wrote:
> In the commit f522886e202a34a2191dd5d471b3c4d46410a9a0 a merge conflict
> in the sdhci-s3c driver been fixed. However the fix used incorrect
> spinlock operation - it cause a race with sdhci interrupt service. The
> correct way to solve it is to use spin_lock_irqsave/irqrestore() calls.

Thanks, applied to mmc-next with:
Signed-off-by: Chris Ball <cjb@laptop.org>

Andrew, not sure how best to get this upstream -- do you want to send
this up to Linus (for -rc5) via your queue in -mm?

-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

[PATCH] sdhci-s3c: fix incorrect spinlock usage after merge

[Andrew Morton]

On Tue, 21 Sep 2010 02:17:11 +0100 Chris Ball <cjb@laptop.org> wrote:

> On Mon, Sep 20, 2010 at 03:03:42PM +0200, Marek Szyprowski wrote:
> > In the commit f522886e202a34a2191dd5d471b3c4d46410a9a0 a merge conflict
> > in the sdhci-s3c driver been fixed. However the fix used incorrect
> > spinlock operation - it cause a race with sdhci interrupt service. The
> > correct way to solve it is to use spin_lock_irqsave/irqrestore() calls.
> 
> Thanks, applied to mmc-next with:
> Signed-off-by: Chris Ball <cjb@laptop.org>
> 
> Andrew, not sure how best to get this upstream -- do you want to send
> this up to Linus (for -rc5) via your queue in -mm?

Ordinarily I'd expect you to send it to Linus.

I tend to merge important-looking bugfixes in my tree in case
maintainers lose them (happens regularly).  I'll then start spamming
the maintainer with the patch.  If the maintainer remains dead then I
might merge it myself, or I might tag it for -stable backporting and
try again in the next kernel.  

If the maintainer can't be bothered setting up a pull request then he
can just send an acked-by and explicitly ask me to merge it up and
I'll then add it to my next approximately-weekly Linus patchbombing.

[PATCH] sdhci-s3c: fix incorrect spinlock usage after merge

[Chris Ball]

Hi Ben,

On Mon, Sep 20, 2010 at 03:03:42PM +0200, Marek Szyprowski wrote:
> In the commit f522886e202a34a2191dd5d471b3c4d46410a9a0 a merge conflict
> in the sdhci-s3c driver been fixed. However the fix used incorrect
> spinlock operation - it cause a race with sdhci interrupt service. The
> correct way to solve it is to use spin_lock_irqsave/irqrestore() calls.
> 
> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
> ---
>  drivers/mmc/host/sdhci-s3c.c |    6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c
> index 71ad416..735d431 100644
> --- a/drivers/mmc/host/sdhci-s3c.c
> +++ b/drivers/mmc/host/sdhci-s3c.c
> @@ -241,8 +241,10 @@ static struct sdhci_ops sdhci_s3c_ops = {
>  static void sdhci_s3c_notify_change(struct platform_device *dev, int state)
>  {
>  	struct sdhci_host *host = platform_get_drvdata(dev);
> +	unsigned long flags;
> +
>  	if (host) {
> -		spin_lock(&host->lock);
> +		spin_lock_irqsave(&host->lock, flags);
>  		if (state) {
>  			dev_dbg(&dev->dev, "card inserted.\n");
>  			host->flags &= ~SDHCI_DEVICE_DEAD;
> @@ -253,7 +255,7 @@ static void sdhci_s3c_notify_change(struct platform_device *dev, int state)
>  			host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
>  		}
>  		tasklet_schedule(&host->card_tasklet);
> -		spin_unlock(&host->lock);
> +		spin_unlock_irqrestore(&host->lock, flags);
>  	}
>  }
>  

Could I have your ACK on this before I send it to Linus, please?

Thanks,

-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child