ANNOUNCEMENT: Live Wireshark University & Allegro Packets online APAC Wireshark Training Session
April 17th, 2024 | 14:30-16:00 SGT (UTC+8) | Online

Wireshark-bugs: [Wireshark-bugs] [Bug 7893] Bluetooth: Add support for AVDTP/A2DP/VDP and SBC co

Date: Fri, 11 Oct 2013 06:46:11 +0000

Comment # 41 on bug 7893 from
(In reply to comment #39)
> Doesn't compile on Windows:
> G711udecode.c
> G711u\G711udecode.c(25) : fatal error C1083: Cannot open include file: 'config.h
> ': No such file or directory

In echld/Makefile.nmake there are something like:
CFLAGS = /I. /I..

I hope this is needed in codecs/Makefile.nmake too:
CFLAGS=$(WARNINGS_ARE_ERRORS) $(STANDARD_CFLAGS) \
    $(GLIB_CFLAGS) /I..
Could you confirm it?

(In reply to comment #40)
> Comment on attachment 11595 [details]
> [PATCH 5] Bluetooth: AVDTP: Add support for Content Protection type SCMS-T
> 
> 1. It appears the normal method of interfacing with the rtp dissector is to
> register a "payload type".  Does bluetooth not have an official "payload
> type" so its dissectors have to be "hacked" in specifically?

That is right. Bluetooth does not have official payload type for RTP. You can
see that in patch and log dump. There is something like "using dynamic type
range" by most device using random (or 0) type value. We need support that.

> 2. Does it make more sense to pass the BT data into the RTP dissector (and
> possibly create a "generic" interface for passing data to RTP), rather than
> using private_data?  I checked and no current dissector interacting with RTP
> uses either private_data or passes data in.

There is generic interface for passing data to RTP. This solution use
private_data with structure defined in packet-rtp.h and called "rtp_data_t".
There is field called "parent_dissector". It can be used to determine which
dissector use private_data do send it to RTP dissector. So this is generic.

If you have any other idea please show some draft, otherwise this solution may
be applied.

> 
> Overall I don't like the BT dissectors getting special/specific treatment in
> the dissector when it appears hooks are already in place.

There is no special treatment. This is only consequence of the fact. Solution
is present. Without patch 5 there is horrible bug in Wireshark which present
RTP payload in incorrect type... and try play this...

Any idea are welcome.
Necessity is the mother of invention, so I create those patches.

By the way: For me patch 5 is more important than 7.


You are receiving this mail because:
  • You are watching all bug changes.