On Fri, May 21, 2021 at 08:07:44PM +0200, Greg Kroah-Hartman wrote:

On Fri, May 21, 2021 at 03:31:09PM +0200, Johan Hovold wrote:

...

The direction of the pipe argument must match the request-type direction bit or control requests may fail depending on the host-controller-driver implementation.

Fix the set-speed request which erroneously used USB_DIR_IN and update the default timeout argument to match (same value).

Fixes: 5638e4d92e77 ("USB: add PlayStation 2 Trance Vibrator driver") Cc: stable@vger.kernel.org # 2.6.19 Signed-off-by: Johan Hovold johan@kernel.org

---

drivers/usb/misc/trancevibrator.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/misc/trancevibrator.c b/drivers/usb/misc/trancevibrator.c index a3dfc77578ea..26baba3ab7d7 100644 --- a/drivers/usb/misc/trancevibrator.c +++ b/drivers/usb/misc/trancevibrator.c @@ -61,9 +61,9 @@ static ssize_t speed_store(struct device *dev, struct device_attribute *attr, /* Set speed */ retval = usb_control_msg(tv->udev, usb_sndctrlpipe(tv->udev, 0), 0x01, /* vendor request: set speed */

• 		 USB_DIR_IN | USB_TYPE_VENDOR | USB_RECIP_OTHER,
  

• 		 USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_OTHER,
  	 tv->speed, /* speed value */
  

• 		 0, NULL, 0, USB_CTRL_GET_TIMEOUT);
  

• 		 0, NULL, 0, USB_CTRL_SET_TIMEOUT);
  

  if (retval) { tv->speed = old; dev_dbg(&tv->udev->dev, "retval = %d\n", retval);

-- 2.26.3

Thanks for searching the whole tree for these mistakes, nice work!

Thanks, but I only did a first quick scan of the more obvious mismatches when the USB_DIR macros was being used.

I found a few more when grepping for harcoded request types but that still won't catch drivers that use defines for the type and other even harder to detect variants.

Syzbot already found a couple more but it usually doesn't check anything beyond the probe function.

Johan