There is emem_header_t struct, so IMHO is better to use struct instead of
global variables.
Split emem_alloc() into two functions: emem_alloc_chunk and emem_alloc_glib.
Current allocator is available in emem_header_t.memory_alloc
It should be easier to develop new memory allocator and maintain current.
It might also give some speedup (we don't need to check mem->debug_use_chunks
every time :)
From me:
Initialize emem_header_t.trees. Some indentation cleanup.
svn path=/trunk/; revision=30624
WPS encodes Bool attributes as single-octet values. Selected Registrar
attribute is parsed correctly (packet-wps.c), but many of the other Bool
attributes are marked incorrectly to use two-octet fields. This results in
showing incorrect octets for some of the attributes and incorrectly marking
frames invalid if a Bool attribute is included as the last attribute in a WPS
IE.
svn path=/trunk/; revision=30622
- Added standard copyright boiler plate
- Made all local functions static
- Added several checks for valid pointers
- Fixed some variable shadowing
- Fixed some printf formats (%p)
- Removed unused variables
svn path=/trunk/; revision=30613
attmepting to malloc memory to ensure we've been asked to decode a reasonable
number of bits (and thus are asking for a reasonable number of bytes from
malloc()).
That should fix the fuzz failure in
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=4138
Add a comment to new_octet_aligned_subset() asking why we're resetting the
length to tvb_length()--that means we'll never throw an exception.
svn path=/trunk/; revision=30611
g_hash_tables that we then try to free in ros_reinit(). Instead, g_malloc()
the conversation structure and free it (along with the hash tables) in the
reinit function.
svn path=/trunk/; revision=30601
- it contains pointers to a couple malloc()'d addresses
- it is inserted in the fragment table (the contents of which are
g_free()'d in free_all_fragments())
Instead, do like fragment_key_copy() and use a g_slice or g_chunk, depending
on the glib version.
svn path=/trunk/; revision=30599
IEEE 802.11k added a new Action frame category: Public Action. I'm proposing
some initial code to recognize this and to add a placeholder for switch
statement that can be used for processing vendor specific Public Action frames
(they are defined in IEEE 802.11w).
svn path=/trunk/; revision=30593
It would be nice to have dissection of the Fibre-Channel FCP
"Sequence retransmission request" (SRR) request.
This is like an FC ELS request, but it has FC type FCP, so it's
a little strange. It seemed like the best place to put it is in
packet-fcp.c but a slight hook is needed in packet-fc.c to
recognize that packet-fcp is the correct dissector.
svn path=/trunk/; revision=30587
could fix cross-compilation problems, according to Jeff Morriss, and
"make dist" seems to build a ustar archive after we do that, so give
that a try. (I don't have any cross-compilers handy to test whether it
fixes cross-compilation or not.)
svn path=/trunk/; revision=30586