[Linux-kernel-mentees,v2,RESEND] usbhid: Fix slab-out-of-bounds write in hiddev_ioctl_usage()

Commit Message

Peilin Ye July 29, 2020, 11:37 a.m. UTC

`uref->usage_index` is not always being properly checked, causing
hiddev_ioctl_usage() to go out of bounds under some cases. Fix it.

Reported-by: [e-mail]
Link: https://syzkaller.appspot.com/bug?id=f2aebe90b8c56806b050a20b36f51ed6acabe802
Reviewed-by: Dan Carpenter <[e-mail]>
Signed-off-by: Peilin Ye <[e-mail]>
---
Change in v2:
    - Add the same check for the `HIDIOCGUSAGE` case. (Suggested by
      Dan Carpenter <[e-mail]>)

 drivers/hid/usbhid/hiddev.c | 4 ++++
 1 file changed, 4 insertions(+)

Comments

Dan Carpenter July 29, 2020, 2:35 p.m. UTC | #1

On Wed, Jul 29, 2020 at 07:37:12AM -0400, Peilin Ye wrote:
> `uref->usage_index` is not always being properly checked, causing
> hiddev_ioctl_usage() to go out of bounds under some cases. Fix it.
> 
> Reported-by: [e-mail]
> Link: https://syzkaller.appspot.com/bug?id=f2aebe90b8c56806b050a20b36f51ed6acabe802
> Reviewed-by: Dan Carpenter <[e-mail]>
> Signed-off-by: Peilin Ye <[e-mail]>
> ---
> Change in v2:
>     - Add the same check for the `HIDIOCGUSAGE` case. (Suggested by
>       Dan Carpenter <[e-mail]>)

Why are you resending this?

regards,
dan carpenter

Jiri Kosina Aug. 17, 2020, 10:21 a.m. UTC | #2

On Wed, 29 Jul 2020, Peilin Ye wrote:

> `uref->usage_index` is not always being properly checked, causing
> hiddev_ioctl_usage() to go out of bounds under some cases. Fix it.
> 
> Reported-by: [e-mail]
> Link: https://syzkaller.appspot.com/bug?id=f2aebe90b8c56806b050a20b36f51ed6acabe802
> Reviewed-by: Dan Carpenter <[e-mail]>
> Signed-off-by: Peilin Ye <[e-mail]>
> ---
> Change in v2:
>     - Add the same check for the `HIDIOCGUSAGE` case. (Suggested by
>       Dan Carpenter <[e-mail]>)

Applied, thanks.

Peilin Ye Aug. 18, 2020, 10 a.m. UTC | #3

On Mon, Aug 17, 2020 at 12:21:41PM +0200, Jiri Kosina wrote:
> On Wed, 29 Jul 2020, Peilin Ye wrote:
> 
> > `uref->usage_index` is not always being properly checked, causing
> > hiddev_ioctl_usage() to go out of bounds under some cases. Fix it.
> > 
> > Reported-by: [e-mail]
> > Link: https://syzkaller.appspot.com/bug?id=f2aebe90b8c56806b050a20b36f51ed6acabe802
> > Reviewed-by: Dan Carpenter <[e-mail]>
> > Signed-off-by: Peilin Ye <[e-mail]>
> > ---
> > Change in v2:
> >     - Add the same check for the `HIDIOCGUSAGE` case. (Suggested by
> >       Dan Carpenter <[e-mail]>)
> 
> Applied, thanks.

Thank you for reviewing the patch!

Peilin Ye

Patch

diff --git a/drivers/hid/usbhid/hiddev.c b/drivers/hid/usbhid/hiddev.c
index 4140dea693e9..4f97e6c12059 100644
--- a/drivers/hid/usbhid/hiddev.c
+++ b/drivers/hid/usbhid/hiddev.c
@@ -519,12 +519,16 @@  static noinline int hiddev_ioctl_usage(struct hiddev *hiddev, unsigned int cmd,
 
 		switch (cmd) {
 		case HIDIOCGUSAGE:
+			if (uref->usage_index >= field->report_count)
+				goto inval;
 			uref->value = field->value[uref->usage_index];
 			if (copy_to_user(user_arg, uref, sizeof(*uref)))
 				goto fault;
 			goto goodreturn;
 
 		case HIDIOCSUSAGE:
+			if (uref->usage_index >= field->report_count)
+				goto inval;
 			field->value[uref->usage_index] = uref->value;
 			goto goodreturn;