Follow-up...
Looking closer at that LKML item, that was a test involving fuzzing the
kernel (i.e. feeding it deliberate garbage). This is related to things
like producing a deliberately bogus USB configuration or sending
impossible data on endpoints that don't exist. The pvrusb2 driver was
It might be related. I have some digging to do, and it's been quite a
while since I've seen any issues in that area of logic.
However that said, one the hardest areas of that driver was implementing
a proper tear-down after a hot-unplug event. That sort of thing is
going to inherently be a r
I will do what I can. Kind of surprising to have something like this
happen now after so many years of stability. There must have been some
kind of environmental change that set this up.
-Mike
On Sat, 16 Dec 2017, Hans Verkuil wrote:
> Hi Mike,
>
> Can you take a look?
>
> Regards,
>
Hi Mike,
Can you take a look?
Regards,
Hans
On 16/12/17 00:02, Oleksandr Ostrenko wrote:
> Dear all,
>
> Unplugging the TV tuner (WinTV HVR-1900) from USB port causes a NULL pointer
> dereference in pvr2_v4l2_internal_check:
>
> [ 2128.129776] usb 1-1: USB disconnect, device number 6
Dear all,
Unplugging the TV tuner (WinTV HVR-1900) from USB port causes a NULL
pointer dereference in pvr2_v4l2_internal_check:
[ 2128.129776] usb 1-1: USB disconnect, device number 6
[ 2128.129987] pvrusb2: Device being rendered inoperable
[ 2128.130055] BUG: unable to handle kernel NULL poin
