A negative `size' will bypass FFMIN(). In the subsequent memcpy() call,
`size' will be considered as a large positive value, leading to a buffer
overflow.
Change the type of `size' to unsigned int to avoid buffer overflow, and
simplify overflow checks accordingly. Also change a literal buffer
size to use sizeof, and limit the amount of data copied in another
memcpy call as well.
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Sanity checks like `data + size >= data_end || data + size < data' are
broken, because `data + size < data' assumes pointer overflow, which is
undefined behavior in C. Many compilers such as gcc/clang optimize such
checks away.
Use `size < 0 || size >= data_end - data' instead.
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes sure all incoming packets are read and handled (and reacted
to) while sending an FLV stream over RTMP to a server. If there were
enough incoming data to fill the TCP buffers, this could potentially
make things block at unexpected places. For the upcoming RTMPT support,
we need to consume all incoming data before we can send the next
request.
Signed-off-by: Martin Storsjö <martin@martin.st>
The existing functions defined in intfloat_readwrite.[ch] are
both slow and incorrect (infinities are not handled).
This introduces a new header with fast, inline conversion
functions using direct union punning assuming an IEEE-754
system, an assumption already made throughout the code.
The one use of Intel/Motorola extended 80-bit format is
replaced by simpler code sufficient under the present
constraints (positive normal values).
The old functions are marked deprecated and retained for
compatibility.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Some received packets can have size 0. The return value from
av_malloc(0) may be NULL, which is ok if the size was 0. On
OS X, however, the returned pointer is non-null but leads to
crashes when trying to free it.
Signed-off-by: Martin Storsjö <martin@martin.st>
track timestamp difference as well.
Patch by Sergiy (mail.composeAddress("piratfm","gmail.com"))
Originally committed as revision 20714 to svn://svn.ffmpeg.org/ffmpeg/trunk
difference, so make all read packets store absolute timestamp.
As a consequence, we don't need to track audio/video timestamps separately
any longer in protocol handler.
Originally committed as revision 20685 to svn://svn.ffmpeg.org/ffmpeg/trunk
instead of assuming it should occur right at given position.
This helps finding human-readable error descriptions in RTMP server replies.
Originally committed as revision 20575 to svn://svn.ffmpeg.org/ffmpeg/trunk
