2019-02-09 00:15:53

by Miguel Ojeda

[permalink] [raw]
Subject: [PATCH] auxdisplay: ht16k33: fix potential user-after-free on module unload

On module unload/remove, we need to ensure that work does not run
after we have freed resources. Concretely, cancel_delayed_work()
may return while the callback function is still running.

From kernel/workqueue.c:

The work callback function may still be running on return,
unless it returns true and the work doesn't re-arm itself.
Explicitly flush or use cancel_delayed_work_sync() to wait on it.

Link: https://lore.kernel.org/lkml/[email protected]/
Reported-by: Sven Van Asbroeck <[email protected]>
Signed-off-by: Miguel Ojeda <[email protected]>
---
drivers/auxdisplay/ht16k33.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/auxdisplay/ht16k33.c b/drivers/auxdisplay/ht16k33.c
index a43276c76fc6..21393ec3b9a4 100644
--- a/drivers/auxdisplay/ht16k33.c
+++ b/drivers/auxdisplay/ht16k33.c
@@ -509,7 +509,7 @@ static int ht16k33_remove(struct i2c_client *client)
struct ht16k33_priv *priv = i2c_get_clientdata(client);
struct ht16k33_fbdev *fbdev = &priv->fbdev;

- cancel_delayed_work(&fbdev->work);
+ cancel_delayed_work_sync(&fbdev->work);
unregister_framebuffer(fbdev->info);
framebuffer_release(fbdev->info);
free_page((unsigned long) fbdev->buffer);
--
2.17.1



2019-02-09 16:53:19

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: [PATCH] auxdisplay: ht16k33: fix potential user-after-free on module unload

On Sat, Feb 09, 2019 at 01:15:22AM +0100, Miguel Ojeda wrote:
> On module unload/remove, we need to ensure that work does not run
> after we have freed resources. Concretely, cancel_delayed_work()
> may return while the callback function is still running.
>
> From kernel/workqueue.c:
>
> The work callback function may still be running on return,
> unless it returns true and the work doesn't re-arm itself.
> Explicitly flush or use cancel_delayed_work_sync() to wait on it.
>
> Link: https://lore.kernel.org/lkml/[email protected]/
> Reported-by: Sven Van Asbroeck <[email protected]>
> Signed-off-by: Miguel Ojeda <[email protected]>

Reviewed-by: Dmitry Torokhov <[email protected]>

> ---
> drivers/auxdisplay/ht16k33.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/auxdisplay/ht16k33.c b/drivers/auxdisplay/ht16k33.c
> index a43276c76fc6..21393ec3b9a4 100644
> --- a/drivers/auxdisplay/ht16k33.c
> +++ b/drivers/auxdisplay/ht16k33.c
> @@ -509,7 +509,7 @@ static int ht16k33_remove(struct i2c_client *client)
> struct ht16k33_priv *priv = i2c_get_clientdata(client);
> struct ht16k33_fbdev *fbdev = &priv->fbdev;
>
> - cancel_delayed_work(&fbdev->work);
> + cancel_delayed_work_sync(&fbdev->work);
> unregister_framebuffer(fbdev->info);
> framebuffer_release(fbdev->info);
> free_page((unsigned long) fbdev->buffer);
> --
> 2.17.1
>

--
Dmitry

2019-02-10 22:06:10

by Sven Van Asbroeck

[permalink] [raw]
Subject: Re: [PATCH] auxdisplay: ht16k33: fix potential user-after-free on module unload

On Fri, Feb 8, 2019 at 7:15 PM Miguel Ojeda
<[email protected]> wrote:
>
> On module unload/remove, we need to ensure that work does not run
> after we have freed resources. Concretely, cancel_delayed_work()
> may return while the callback function is still running.
>
> From kernel/workqueue.c:
>
> The work callback function may still be running on return,
> unless it returns true and the work doesn't re-arm itself.
> Explicitly flush or use cancel_delayed_work_sync() to wait on it.
>
> Link: https://lore.kernel.org/lkml/[email protected]/
> Reported-by: Sven Van Asbroeck <[email protected]>
> Signed-off-by: Miguel Ojeda <[email protected]>

Yes, this should do the trick.

Reviewed-by: Sven Van Asbroeck <[email protected]>

2019-02-11 07:26:42

by Robin van der Gracht

[permalink] [raw]
Subject: Re: [PATCH] auxdisplay: ht16k33: fix potential user-after-free on module unload

On Sat, 9 Feb 2019 01:15:22 +0100
Miguel Ojeda <[email protected]> wrote:

> On module unload/remove, we need to ensure that work does not run
> after we have freed resources. Concretely, cancel_delayed_work()
> may return while the callback function is still running.
>
> From kernel/workqueue.c:
>
> The work callback function may still be running on return,
> unless it returns true and the work doesn't re-arm itself.
> Explicitly flush or use cancel_delayed_work_sync() to wait on it.
>
> Link: https://lore.kernel.org/lkml/[email protected]/
> Reported-by: Sven Van Asbroeck <[email protected]>
> Signed-off-by: Miguel Ojeda <[email protected]>
> ---
> drivers/auxdisplay/ht16k33.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/auxdisplay/ht16k33.c b/drivers/auxdisplay/ht16k33.c
> index a43276c76fc6..21393ec3b9a4 100644
> --- a/drivers/auxdisplay/ht16k33.c
> +++ b/drivers/auxdisplay/ht16k33.c
> @@ -509,7 +509,7 @@ static int ht16k33_remove(struct i2c_client *client)
> struct ht16k33_priv *priv = i2c_get_clientdata(client);
> struct ht16k33_fbdev *fbdev = &priv->fbdev;
>
> - cancel_delayed_work(&fbdev->work);
> + cancel_delayed_work_sync(&fbdev->work);
> unregister_framebuffer(fbdev->info);
> framebuffer_release(fbdev->info);
> free_page((unsigned long) fbdev->buffer);

Looks good

Acked-by: Robin van der Gracht <[email protected]>

2019-02-14 10:13:03

by Miguel Ojeda

[permalink] [raw]
Subject: Re: [PATCH] auxdisplay: ht16k33: fix potential user-after-free on module unload

On Mon, Feb 11, 2019 at 8:26 AM Robin van der Gracht <[email protected]> wrote:
>
> Looks good
>
> Acked-by: Robin van der Gracht <[email protected]>

Thanks all! I will send this for -rc7.

Cheers,
Miguel