Hi,

2012/2/29 Måns Rullgård <[email protected]>:
> Vitor Sessak <[email protected]> writes:
>
>> On 02/29/2012 04:28 PM, Janne Grunau wrote:
>>> On 2012-02-26 09:52:44 +0100, Vitor Sessak wrote:
>>>> ---
>>>>   libavcodec/ra144dec.c |    2 ++
>>>>   libavcodec/ra288.c    |    2 ++
>>>>   libavcodec/sipr.c     |    2 ++
>>>>   libavcodec/twinvq.c   |    2 ++
>>>>   4 files changed, 8 insertions(+), 0 deletions(-)
>>>
>>> Why?
>>>
>>> Have you proofed that each of the decoder can't overread?
>>
>> Of course I did. I concede didn't do it with the AMRNB in my first
>> patch. I was almost sure I saw the check when I reviewed it, but I was
>> wrong.
>
> [...]
>
>>> I would say the decoders are not important enough and speed penalty
>>> for audio doesn't matter enough to disable the safe bitstream reader.
>>
>> How hard is it to check a single constant value correctly? What is the
>> use of the safe bitstream reader if the check is done right?
>
> There's much more to it than that.  Almost anything using
> variable-length codes will need more than a simple packet size check, or
> a damaged/malicious bitstream may cause over-reads.

This is the concern that I have also... We really want almost-academic
sort of proof that the decoder can not possibly ever consume more than
X bits of data from /dev/random per single decoding iteration before
unsetting the safe bitstream reader flag.

Ronald
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to