[pulseaudio-discuss] [PATCH 3/3] protocol-native: Fail if trying to push unaligned memblock into queue

[David Henningsson]

On 2014-10-17 12:06, Arun Raghavan wrote:
> On Thu, 2014-10-16 at 11:41 +0200, David Henningsson wrote:
>> We will just ignore the memblock if this happens. We already have
>> a check for this in the client library, so this one is just for
>> security reasons.
>>
>> Signed-off-by: David Henningsson <david.henningsson at canonical.com>
>> ---
>>   src/pulsecore/protocol-native.c | 7 +++++++
>>   1 file changed, 7 insertions(+)
>>
>> diff --git a/src/pulsecore/protocol-native.c b/src/pulsecore/protocol-native.c
>> index 6ec65d6..2ef1854 100644
>> --- a/src/pulsecore/protocol-native.c
>> +++ b/src/pulsecore/protocol-native.c
>> @@ -4922,6 +4922,13 @@ static void pstream_memblock_callback(pa_pstream *p, uint32_t channel, int64_t o
>>       if (playback_stream_isinstance(stream)) {
>>           playback_stream *ps = PLAYBACK_STREAM(stream);
>>
>> +        size_t frame_size = pa_frame_size(&ps->sink_input->sample_spec);
>> +        if (chunk->index % frame_size != 0 || chunk->length % frame_size != 0) {
>> +            pa_log_warn("Client sent non-aligned memblock: index %d, length %d, frame size: %d",
>> +                        (int) chunk->index, (int) chunk->length, (int) frame_size);
>> +            return;
>> +        }
>> +
>
> Do we need to release chunk->memblock if it exists here?

No, unless you see something I don't? There is no call to 
pa_memblock_acquire above the added lines, so we should not call 
pa_memblock_release.

Okay to push this series, given that I drop _se in the first patch?

-- 
David Henningsson, Canonical Ltd.
https://launchpad.net/~diwic