On Wed, Nov 6, 2019 at 12:37 AM Phong Tran <tranmanphong@xxxxxxxxx> wrote: > > In context of USB disconnect, the delaywork trigger and calling > appledisplay_bl_get_brightness() and the msgdata was freed. > > add the checking return value of usb_control_msg() and only update the > data while the retval is valid. > > Reported-by: syzbot+495dab1f175edc9c2f13@xxxxxxxxxxxxxxxxxxxxxxxxx > Reported-and-tested-by: > syzbot+495dab1f175edc9c2f13@xxxxxxxxxxxxxxxxxxxxxxxxx > > https://groups.google.com/d/msg/syzkaller-bugs/dRmkh2UYusY/l2a6Mg3FAQAJ Hi Phong, FYI, when testing patches with the usb-fuzzer instance, you need to provide the same kernel commit id as the one where the bug was triggered. Please see here for details: > > Signed-off-by: Phong Tran <tranmanphong@xxxxxxxxx> > --- > drivers/usb/misc/appledisplay.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/usb/misc/appledisplay.c b/drivers/usb/misc/appledisplay.c > index ac92725458b5..3e3dfa5a3954 100644 > --- a/drivers/usb/misc/appledisplay.c > +++ b/drivers/usb/misc/appledisplay.c > @@ -164,7 +164,8 @@ static int appledisplay_bl_get_brightness(struct backlight_device *bd) > 0, > pdata->msgdata, 2, > ACD_USB_TIMEOUT); > - brightness = pdata->msgdata[1]; > + if (retval >= 0) > + brightness = pdata->msgdata[1]; > mutex_unlock(&pdata->sysfslock); > > if (retval < 0) > -- > 2.20.1 >
