2009-10-31 16:03:35

by Rodolfo Giometti

Subject: [PATCH] pps: locking scheme fix up for PPS_GETPARAMS.

Signed-off-by: Rodolfo Giometti <[email protected]>
Tested-by: Reg Clemens <[email protected]>
---
drivers/pps/pps.c | 11 ++++++++---
1 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/drivers/pps/pps.c b/drivers/pps/pps.c
index fea17e7..ca5183b 100644
--- a/drivers/pps/pps.c
+++ b/drivers/pps/pps.c
@@ -71,9 +71,14 @@ static long pps_cdev_ioctl(struct file *file,
case PPS_GETPARAMS:
pr_debug("PPS_GETPARAMS: source %d\n", pps->id);

- /* Return current parameters */
- err = copy_to_user(uarg, &pps->params,
- sizeof(struct pps_kparams));
+ spin_lock_irq(&pps->lock);
+
+ /* Get the current parameters */
+ params = pps->params;
+
+ spin_unlock_irq(&pps->lock);
+
+ err = copy_to_user(uarg, &params, sizeof(struct pps_kparams));
if (err)
return -EFAULT;

--
1.6.3.3

2009-11-03 01:54:14

by Andrew Morton

Subject: Re: [PATCH] pps: locking scheme fix up for PPS_GETPARAMS.

On Sat, 31 Oct 2009 17:03:23 +0100
Rodolfo Giometti <[email protected]> wrote:

> Signed-off-by: Rodolfo Giometti <[email protected]>
> Tested-by: Reg Clemens <[email protected]>

Please don't send unchangelogged patches.

>
> diff --git a/drivers/pps/pps.c b/drivers/pps/pps.c
> index fea17e7..ca5183b 100644
> --- a/drivers/pps/pps.c
> +++ b/drivers/pps/pps.c
> @@ -71,9 +71,14 @@ static long pps_cdev_ioctl(struct file *file,
> case PPS_GETPARAMS:
> pr_debug("PPS_GETPARAMS: source %d\n", pps->id);
>
> - /* Return current parameters */
> - err = copy_to_user(uarg, &pps->params,
> - sizeof(struct pps_kparams));
> + spin_lock_irq(&pps->lock);
> +
> + /* Get the current parameters */
> + params = pps->params;
> +
> + spin_unlock_irq(&pps->lock);
> +
> + err = copy_to_user(uarg, &params, sizeof(struct pps_kparams));
> if (err)
> return -EFAULT;
>

OK, I can see what the patch does and I can guess what sort of
situations would trigger it. But that's really not good enough.

Put yourself in the position of someone who is hitting a PPS bug and
wants to work out if your patch might fix it. Because the patch fails
to describe the user-visible symptoms (ie: the bug) then that person is
in the dark.

Also, someone (ie: me) needs to decide if this fix is to be backported
into earlier kernels. With no description of the end-user impact, how
can I possibly do that?