Re: [Qemu-devel] [PATCH v2] virtio: add check for descriptor's mapped address

[Laszlo Ersek]

On 09/19/16 19:16, Laurent Vivier wrote:
> 
> 
> On 19/09/2016 18:59, Laszlo Ersek wrote:
>> On 09/19/16 17:55, Stefan Hajnoczi wrote:
>>> On Mon, Sep 19, 2016 at 02:36:11PM +0530, P J P wrote:
>>>> From: Prasad J Pandit <address@hidden>
>>>>
>>>> virtio back end uses set of buffers to facilitate I/O operations.
>>>> If its size is too large, 'cpu_physical_memory_map' could return
>>>> a null address. This would result in a null dereference while
>>>> un-mapping descriptors. Add check to avoid it.
>>>>
>>>> Reported-by: Qinghao Tang <address@hidden>
>>>> Signed-off-by: Prasad J Pandit <address@hidden>
>>>> ---
>>>>  hw/virtio/virtio.c | 5 +++++
>>>>  1 file changed, 5 insertions(+)
>>>>
>>>> Update per:
>>>>   -> https://lists.gnu.org/archive/html/qemu-devel/2016-09/msg03889.html
>>>>
>>>> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
>>>> index 15ee3a7..311dd0b 100644
>>>> --- a/hw/virtio/virtio.c
>>>> +++ b/hw/virtio/virtio.c
>>>> @@ -472,6 +472,11 @@ static void virtqueue_map_desc(unsigned int 
>>>> *p_num_sg, hwaddr *addr, struct iove
>>>>          }
>>>>  
>>>>          iov[num_sg].iov_base = cpu_physical_memory_map(pa, &len, 
>>>> is_write);
>>>> +        if (!iov[num_sg].iov_base) {
>>>> +            error_report("virtio: bogus descriptor or out of resources");
>>>> +            exit(EXIT_FAILURE);
>>>
>>> exit(1) is used everywhere else in the file (including just a few lines
>>> above in the same function).  Please use exit(1) for consistency.
>>
>> Laurent's pending series
>>
>> [Qemu-devel] [PATCH 00/26] trivial: use exit(EXIT_SUCCESS) and
>>                            exit(EXIT_FAILURE)
>>
>> specifically his patch
>>
>> [Qemu-devel] [PATCH 11/26] pci, virtio: use exit(EXIT_SUCCESS) and
>>                            exit(EXIT_FAILURE)
>>
>> converts the argument of the exit() that you name to EXIT_FAILURE.
>>
>> So using EXIT_FAILURE in Prasad's patch is actually what will uphold
>> consistency.
>>
>>> Looks fine otherwise.
>>>
> 
> Laszlo, it seems Peter (and some others) would prefer to use exit(1)
> instead of exit(EXIT_FAILURE), so I don't think my series will be applied.

Thanks for the information, Laurent!

Prasad, can you please send a v3 then, to address Stefan's feedback?

Thank you
Laszlo