1. Changes how can_desegment works so that can_desegment is
only != 0 for whichever dissector is running immediately on
top of whoever offers the can_desegment service.
Thus DCERPC needs no special handling to see if it can trust
can_desegment (which is currently only available ontop of TCP
and not ontop of tcp->nbss->smb).
2. Changes fragment reassembly of transaction smb to only show
the defragmented packet for the transaction smb holding the
first fragment.
To see why, test it with a transaction SMB containing a ~60kb
PDU or larger. The old behaviour had approximately quadratic
behaviour regarding runtime for dissecting such PDUs.
(example: NetShareEnum is a command which can grow really really
large if the number of shares and comments are large)
svn path=/trunk/; revision=4296
"smb_saved_info_t" in the table of requests whose replies have been
found, don't look it up in the table of requests whose replies have not
been found - if the request in question has no reply in the capture,
that may find some later frame in the same conversation with the same
MID, and we don't need that information anyway - the only reason we
*need* that structure is to save information in it for use when
processing its reply, and we already did that the first time we
processed the request. (The information for the later frame may be bad,
e.g. having a null "extra_info" pointer, or having one that points to
information for another request.)
Arrange that we don't use the pointer to the "smb_saved_info_t" when
processing a request except to save information if the request hasn't
already been processed, as that pointer may not be valid if the request
has already been processed, as per the above.
svn path=/trunk/; revision=4292
Add some checks for null tvbuff arguments.
When dissecting transaction setup, parameters, and data when we couldn't
dissect it as a pipe or mailslot transaction, use the reported length of
the supplied tvbuff, not the actual length, as the amount of data
present.
svn path=/trunk/; revision=4291
don't have the frame number of the request, which we use as the ID of
the transaction being reassembled. (If we're reassembling a reply,
should we not use the frame number of the reply instead? We used to
have a hash table to keep track of that, so we might just be able to
bring it back....)
svn path=/trunk/; revision=4274
Add a few small functions to reassemble.c to cope with protocols
where the total length of defragmented PDUs are specified in the
first fragment (all previous uses of reassembly has been for
PDUs where the last fragment is signalled by a flag in the
header for the last fragment).
Add a few small functions to reassemble.c to abort-and-delete
defragmentation of PDUs and also detect IF a PDU is currently
being defragmented. (Useful for PDUs where the "unique"
identifier is rather ununique, or may be reused often enough so
it can be a problem for Ethereal.)
Change where NT Cancel presents its Cancelation-to output, and
makes the three trans secondary requests also output similar
information.
svn path=/trunk/; revision=4255
#defines for SMB commands with ones that use the names from the SNIA
CIFS spec.
Use those #define values rather than hardcoded values in various places
that check for specific commands.
svn path=/trunk/; revision=4244
routines used for that.
Rename some named pipe functions as per the SNIA CIFS spec.
Label the "number of files moved" field of the reply to a Move SMB as
such, rather than as an unspecified "Count".
svn path=/trunk/; revision=4229
obviates the need to have a protocol tree item for "MSRPC-over-SMB", as
the setup words for it are just standard TransactNmPipe setup words
(0x26 is the TransactNmPipe function code, and the next setup word is
the FID for the pipe in question.)
Pass to the pipe dissector tvbuffs for setup-words-plus-pipe (which is
the data for the pipe protocol) and parameters-plus-data (which is the
data for the protocol running atop the pipe protocol); use the former
for the top-level protocol tree item for the pipe protocol, and the
latter for the top-level protocol tree item for the LANMAN protocol.
svn path=/trunk/; revision=4223
mailslot-based logon protocols just be regular register routines,
detected by the script that generates the "register.c" file, rather than
special stuff known to the SMB dissector.
svn path=/trunk/; revision=4222
"dissect_pipe_smb()", a tvbuff containing the setup words and the
pipe/mailslot pathname, as those are arguably the part of the packet
that contains the "mailslot protocol" and the "pipe protocol", as
opposed to the protocol running atop mailslots or pipes.
Pass a setup tvbuff to "dissect_pipe_smb()" for it to pass on to the
MSRPC-over-named-pipe dissector, and have the setup tvbuff passed to it
and "dissect_mailslot_smb()" contain *only* the setup words; don't
extract anything other than the setup words from it.
Declare "register_proto_smb_mailslot()" in "packet-smb-mailslot.h"
rather than "packet-smb.c", and declare "register_proto_smb_pipe()" in
"packet-smb-pipe.h" rather than "packet-smb.c".
Add a protocol for MSRPC-over-named-pipes.
Move the stuff to handle the FID in the setup words of
MSRPC-over-named-pipe transactions out of the SMB Transaction dissector
into the MSRPC dissector. Add a routine to "packet-smb.c", callable
from outside "packet-smb.c", to put an "smb.fid" field into the protocol
tree, and to add ", FID: XXXX" to the Info column, for use by the
MSRPC-over-named-pipe dissector; use it in the SMB dissector as well, in
all the places where we put a FID into the protocol tree.
Move the stuff to check whether the LANMAN protocol is enabled, and to
set "pinfo->current_proto" to "LANMAN" if it is, into the LANMAN
API-over-named-pipe dissector out of the named pipe protocol dissector.
If we didn't dissect a Transaction request or reply as a named pipe or
mailslot message, put any setup words, parameters, and data it has into
the protocol tree as separate items.
Don't put a "Response in" item into the protocol tree for an NT Cancel
request, as there are no responses to NT Cancel requests.
svn path=/trunk/; revision=4221
display the returned FID in the Info column for NT Create And X
replies;
display the setup words, and treat the second word as a FID in
Transaction requests presumed to contain DCE RPC-over-SMB.
Add the FID to the Info column for other open/create replies while we're
at it.
svn path=/trunk/; revision=4219
structure, so that it can be updated by subdissectors; this way the
updates affect the structure immediately, and don't get lost if the
subdissector later throws an exception.
Use "tvb_reported_length()" to check for an interim mailslot reply;
"tvb_length()" could give the wrong answer if a short snapshot length
was given in the capture.
svn path=/trunk/; revision=4218
"smb_saved_info_t". Put all the information needed to dissect NT
Transaction replies, Transaction2 replies, or Transaction replies into
separate data structures, allocated separately, and put a pointer to
that data structure in the "void *" in question.
Use the return value of "dissect_pipe_smb()" and
"dissect_mailslot_smb()" to control whether to display as data the stuff
those routines were asked to dissect.
If we've seen a request before, but its "smb_saved_info_t" isn't in the
"matched" hash table, look in the "unmatched" hash table - perhaps we
haven't seen the reply yet.
svn path=/trunk/; revision=4216
SMB FILE SHARING PROTOCOL EXTENSIONS, SMB File Sharing Protocol
Extensions Version 2.0, Document Version 3.3".
Fix the test for the "connectionless mode" to test the correct bit.
svn path=/trunk/; revision=4215
and replies, just save a structure holding that information that can't
be derived from the contents of one of the SMBs.
Don't save anything at all for NT Cancel requests - they have the same
TID/PID/MID/UID as the SMB being cancelled, and you want the information
for that request used when dissecting the NT Cancel (so it gets the
number of the frame containing the request being cancelled) and when
dissecting the reply to the request being cancelled.
Get rid of an unused routine.
svn path=/trunk/; revision=4213
Get rid of "Response to" stuff in the LANMAN dissector, as that's now
done in the SMB dissector.
Add a routine for dissecting unknown SMBs (gets the word and byte
counts, and just adds text entries for the word and byte parameters, if
any), and replace null pointers in the dissector table with pointers to
that routine. Get rid of the check for a null dissector pointer.
svn path=/trunk/; revision=4212
piece of information in the reply, as a file can have more than one
stream; show all of them.
Don't use the "File Name" field for stream names.
svn path=/trunk/; revision=4211
"#if 0" out an unused routine, pending determination of whether there's
any place that would use it.
Fix some typos.
Display the NT create options in hex, as they appear to be a bitmask.
svn path=/trunk/; revision=4189
Add some checks of the return value of
"get_unicode_or_ascii_string_tvb()" - if a null terminator is missing,
it might well run past the end of the byte parameters - and add some
code to keep the byte count updated so that the right byte count is
passed to "get_unicode_or_ascii_string_tvb()".
svn path=/trunk/; revision=4180
common subroutine.
Label the "total data length" field in Write Raw and Write Multiplexed
requests as such.
Dissect the IPC State/Device State field of Open and X requests and NT
Create and X requests ass per the stuff on page 67 of
ftp://ftp.microsoft.com/developr/drg/CIFS/dosextp.txt
Make a variable that doesn't need to be static not static.
svn path=/trunk/; revision=4179
Get rid of a bunch of stuff for which said tvbuffication removes the
need.
When dissecting byte parameters, make sure you don't consume more bytes
than the byte count, and handle captures where the last string in the
byte parameters area isn't properly null-terminated (I think I've seen
that in packets from various versions of Windows NT).
Make various bitfields given as decimal in SMB specs decimal.
svn path=/trunk/; revision=4172
Do more sanity checking on DOS dates and times.
Convert a bunch of "proto_tree_add_XXX" calls to "proto_tree_add_item"
calls.
Put the word and byte data for untvbuffified SMBs under a subtree just
as is done for tvbuffified SMBs.
Get rid of some no-longer-used routines.
Fix some displays in untvbuffified SMBs to resemble the way those fields
are displayed for tvbuffified SMBs.
Display timesouts as seconds and milliseconds when they're in units of
milliseconds.
svn path=/trunk/; revision=4157
Make it possible for a non-tvbuffied dissector for an andX SMB to call a
tvbuffified dissector for the andX command, and fix the non-tvbuffified
dissectors in question to do so.
svn path=/trunk/; revision=4154
When converting DOS-date years to "struct tm" years, add 1980 and
subtract 1900, to make it clearer what the conversion involves (DOS-date
years are (year - 1980); "struct tm" years are (year - 1900)).
svn path=/trunk/; revision=4146
Fix up Info column to put "Request" or "Response" *after* the name of
the request.
Give the Negotiate Protocol request its full name.
svn path=/trunk/; revision=4139
Get rid of the "unknown-0xXX" entries in the "value_string" table for
SMB command codes - they make it much more painful to select one of them
in the filter-editing dialog box.
svn path=/trunk/; revision=3985
tvbuffified heuristic-dissector interface, but have it immediately turn
its arguments into an old-style buffer pointer and offset.
Register the SMB dissector as a heuristic NetBIOS dissector, and have
"dissect_netbios_payload()" just try the heuristics, as it no longer has
to call the SMB dissector explicitly.
svn path=/trunk/; revision=3973
packets.
Make a "dissect_netbios_payload()" routine, called from the
NetBIOS-over-802.2 (NBF), NetBIOS-over-IPX, and NetBIOS-over-TCP
dissectors. Take Todd Sabin's changes to add a heuristic dissector list
to the NBSS dissector, and apply them to "dissect_netbios_payload()"
instead. Make the SMB dissector heuristic, returning FALSE if it
doesn't see 0xFF S M B at the beginning of the packet, and have
"dissect_netbios_payload()" first try the heuristic dissector list, then
try the SMB dissector if no other heuristic dissector claims the packet,
then just dissect the payload as data.
From Todd Sabin: have the DCE/RPC dissector register as a heuristic
dissector for NetBIOS.
svn path=/trunk/; revision=3969
"WordCount > 0".
Always put the byte count field into the protocol tree, regardless of
whether WordCount is 0 - it's not one of the word parameters counted by
WordCount, so it's present even if WordCount is 0.
Fix a "val_to_str()" call.
svn path=/trunk/; revision=3966
value being non-zero, not on whether the error code is zero. Don't
bother passing the error code to dissectors for particular SMBs, as they
don't need to use it.
In "get_unicode_or_ascii_string()", when aligning to an even boundary,
align to an even boundary in the SMB message, not in the packet as a
whole - there's no guarantee that there are an even number of bytes in
the frame before the SMB message.
In the Info column, mark the packet as a request or response based on
the request/response bit in the Flags field, not on the matched port -
for NBIPX, the source and destination ports (IPX sockets) may be the
same, so you may not be able to determine whether it's a request or a
response based on that.
svn path=/trunk/; revision=3965
being written, but the 2 bytes of data length and one byte of buffer
type preceding that data; use the data length (which doesn't count
itself or the buffer type byte), rather than the byte count, to
determine how much data is being written.
svn path=/trunk/; revision=3917
of protocol-id-plus-datum pairs, so that multiple protocols can attach
information to the same conversation.
Dissectors that attach information to a conversation should not assume
that if they find a conversation it has one of its data attached to it;
the conversation might've been created by another dissector.
svn path=/trunk/; revision=3901
value.
Put in a comment noting what may be *another* bug in some versions of
Windows when constructing Session Setup and X requests.
svn path=/trunk/; revision=3871
depending on the setting of the "Strings are Unicode" bit in the SMB.
Correctly handle Unicode strings in Session Setup and X and Tree Connect
and X messages.
Always display the Word Count and Byte Count fields of a Session Setup
and X message, regardless of whether we recognize the word count value
as one we can handle or not.
Correctly handle Session Setup and X messages if extended security
exchanges are being used.
Decode the (known) bits of the Action field in an Session Setup and X
message, and the (known) bits of the optional flags field of a Tree
Connect and X message. Show the optional flags field as a 16-bit hex
quantity.
svn path=/trunk/; revision=3868
matching request, or for responses where we don't have the
pathname/transaction code of the matching request, indicate the SMB
opcode of the transaction, rather than just saying it's a response to a
generic message.
svn path=/trunk/; revision=3841
don't know the path name, don't give up, just show the parameters and
data, as is done with transactions that aren't mailslot browser
transactions or LANMAN RAP pipe transactions.
svn path=/trunk/; revision=3840
sense; instead, it's a "multiplex ID" used when there's more than one
request *currently* in flight, to distinguish replies.
This means that the MID and PID don't uniquely identify a request in
a conversation.
Therefore, we have to use some other value to distinguish between
requests with the same MID and PID.
Add a mechanism to do so.
svn path=/trunk/; revision=3829
places) dissector tvbuffified, from Ronnie Sahlberg and me.
Additional "are we past the end of the buffer" checks added, so that we
don't hand random junk to the transaction and transact2 dissectors.
svn path=/trunk/; revision=3824
that rather than passing another copy of that flag to dissectors of
particular messages.
Pass that structure to the pipe subdissector by making "pi.private"
point to it, rather than by passing it as an explicit argument.
Change more of the
if (dirn == 1) {
...
}
if (dirn == 0) {
...
}
stuff to
if (dirn == 1) {
...
} else {
...
}
and then, as per the first paragraph, check the "request" flag in the
"smb_info" structure rather than checking a "dirn" flag.
Set "last_transact2_command" to -1 in the "smb_request_val" structures
for TRANSACTION requests, as it doesn't apply to those requests.
As "dissect_transact_params()" doesn't do any work if the "TransactName"
argument is null, don't bother calling it for a reply if we don't have an
"smb_request_val" for the corresponding request, as that means we can't
find out the value to pass as the "TransactName" argument.
svn path=/trunk/; revision=3822
(This fixes an incorrect string for TRANS2_GET_DFS_REFERRAL, which has
the code 0x10 according to the current SNIA CIFS draft spec; I've seen
those in packet captures.)
Create entries in the transaction hash tables only for requests, not for
replies; this means a reply might not have an entry in the table, if the
request didn't appear in the capture, so handle that case.
Make the "last_transact2_command" field of a "smb_request_val" structure
an "int", so it can be given the value -1, which is different from all
the valid 16-bit unsigned values, to indicate that we couldn't get the
transaction code from the request (e.g., because it's too short).
Show the first Setup word in a TRANSACT2 request as the transaction
code, as that's what it is.
"dirn" is a Boolean, so
if (dirn == 1) {
...
}
if (dirn == 0) {
...
}
is equivalent to
if (dirn == 1) {
...
} else {
...
}
and the latter is a bit clearer, so use it.
Distinguish between a TRANSACTION or TRANSACT2 reply where we didn't see
the request and one where we saw the request but didn't see the request
path for TRANSACTION or the request code for TRANSACT2.
Use "g_strdup()" rather than "g_malloc()" followed by "strcpy()".
svn path=/trunk/; revision=3819
"last_data_descrip" fields of an "smb_request_val" structure to null
when you allocate it, so that, when the hash table is cleaned out, we
don't try handing random junk to "g_free()".
svn path=/trunk/; revision=3811
as the MID and the conversation index; the current CIFS spec from the SNIA
Web site says
The multiplex ID (Mid) is used along with the Pid to allow
multiplexing the single client and server connection among the
client's multiple processes, threads, and requests per thread.
Clients may have many outstanding requests (up to the negotiated
number, MaxMpxCount) at one time. Servers MAY respond to
requests in any order, but a response message MUST always
contain the same Mid and Pid values as the corresponding request
message. The client MUST NOT have multiple outstanding requests
to a server with the same Mid and Pid.
(although, in practice, at least as I remember from working on the
NetApp CIFS server N years ago, Windows clients tend to use the same PID
in all requests, so only the MID acts as a transaction ID).
When initializing the dissector, free up all the data attached to
"smb_request_val" structures in the request hash table before destroying
the hash table and the structures in question.
svn path=/trunk/; revision=3809
Show the parameters, data, and padding in transact/transact2 SMBs as
hex, not text; it's usually binary.
"dissect_transact_params()" returns immediately if TransactName is null,
so we don't need to check whether it's null when copying it; just use
"g_strdup()" to copy it.
svn path=/trunk/; revision=3808
packet; this is far from a complete set of checks - the right way to
make this dissector safe is to tvbuffify it - but it's sufficient to
eliminate most cases where my regression tests bogusly reported that the
packet was dissected differently due to different stuff being past the
end of the packet.
svn path=/trunk/; revision=3807
Move the declaration of routines exported from "packet-smb-mailslot.c"
into a "packet-smb-mailslot.h" header file, and have modules that import
those routines include "packet-smb-mailslot.h" rather than declaring the
routines themselves; do the same for routines exported from
"packet-smb-pipe.c". Make routines not exported static, and make
routines that return a true/false return value "gboolean" rather than
"guint32".
svn path=/trunk/; revision=3146
with not terminating their arrays because they knew the limits of the
value used to look up strings in the value_string array, but the
dfilter_expr_dlg does not know these limits and must rely on the terminating
{0, NULL} record.
Also, in SNA fixed a bug in which a field should have been defined as FT_UINT8
but was defined as FT_BOOLEAN.
In WTP, fixed a value string which had duplicate keys.
svn path=/trunk/; revision=2817
protocols, in addition to adding structures to the list of filterable
fields. Give it an extra argument that specifies a "short name" for the
protocol, for use in such places as
pinfo->current_proto;
the dialog box for constructing filters;
the preferences tab for the protocol;
and so on (although we're not yet using it in all those places).
Make the preference name that appears in the preferences file and the
command line for the DIAMETER protocol "diameter", not "Diameter"; the
convention is that the name in question be all-lower-case.
Make some routines and variables that aren't exported static.
Update a comment in the ICP dissector to make it clear that the
dissector won't see fragments other than the first fragment of a
fragmented datagram.
svn path=/trunk/; revision=2810
Make sure that if _gtime is null, a bad format message returned.
Also noticed that I am going to have to do something about Unicode strings soon and the SMBopenX dissect is slightly wrong ... Oh well, it is the Xmas break soon :-) No rest for the Wicca'd (please don't interpret that as a statement of my religious affiliation, it is just a cute saying :-)
svn path=/trunk/; revision=2764
can be put, and a pointer to the string for the column, which might or
might not point to that buffer.
Add a routine "col_set_str()", which sets the string for the column to
the string passed to it as an argument; it should only be handed a
static string (a string constant would be ideal). It doesn't do any
copying, so it's faster than "col_add_str()".
Make the routines that append to columns check whether the pointer to
the string for the column points to the buffer for the column and, if
not, copy the string for the column to the buffer for the column so that
you can append to it (so you can use "col_set_str()" and then use
"col_append_str()" or "col_append_fstr()").
Convert a bunch of "col_add_str()" calls that take a string constant as
an argument to "col_set_str()" calls.
Convert some "col_add_fstr()" calls that take a string constant as the
only argument - i.e., the format string doesn't have any "%" slots into
which to put strings for subsequent arguments to "col_set_str()" calls
(those calls are just like "col_add_str()" calls).
Replace an END_OF_FRAME reference in a tvbuffified dissector with a
"tvb_length(tvb)" call.
svn path=/trunk/; revision=2670
the following:
It is now possible to enable/disable a particular protocol decoding
(i.e. the protocol dissector is void or not). When a protocol
is disabled, it is displayed as Data and of course, all linked
sub-protocols are disabled as well.
Disabling a protocol could be interesting:
- in case of buggy dissectors
- in case of wrong heuristics
- for performance reasons
- to decode the data as another protocol (TODO)
Currently (if I am not wrong), all dissectors but NFS can be disabled
(and dissectors that do not register protocols :-)
I do not like the way the RPC sub-dissectors are disabled (in the
sub-dissectors) since this could be done in the RPC dissector itself,
knowing the sub-protocol hfinfo entry (this is why, I've not modified
the NFS one yet).
Two functions are added in proto.c :
gboolean proto_is_protocol_enabled(int n);
void proto_set_decoding(int n, gboolean enabled);
and two MACROs which can be used in dissectors:
OLD_CHECK_DISPLAY_AS_DATA(index, pd, offset, fd, tree)
CHECK_DISPLAY_AS_DATA(index, tvb, pinfo, tree)
See also the XXX in proto_dlg.c and proto.c around the new functions.
svn path=/trunk/; revision=2267
a particular type, rather than taking a varargs list, along the lines of
the "proto_tree_add_XXX_format()" routines.
Replace most calls to "proto_tree_add_item()" and
"proto_tree_add_item_hidden()" with calls to those routines.
Rename "proto_tree_add_item()" and "proto_tree_add_item_hidden()" to
"proto_tree_add_item_old()" and "proto_tree_add_item_hidden_old()", and
add new "proto_tree_add_item()" and "proto_tree_add_item_hidden()"
routines that don't take the item to be added as an argument - instead,
they fetch the argument from the packet whose tvbuff was handed to them,
from the offset handed to them.
svn path=/trunk/; revision=2031
Add exceptions routines.
Convert proto_tree_add_*() routines to require tvbuff_t* argument.
Convert all dissectors to pass NULL argument ("NullTVB" macro == NULL) as
the tvbuff_t* argument to proto_tree_add_*() routines.
dissect_packet() creates a tvbuff_t, wraps the next dissect call in
a TRY block, will print "Short Frame" on the proto_tree if a BoundsError
exception is caught.
The FDDI dissector is converted to use tvbuff's.
svn path=/trunk/; revision=1939
Don't use a global variable named "gmtime" - some versions of
NetBSD define it in <time.h> - use "_gmtime" instead, so we can
build on those NetBSD systems.
svn path=/trunk/; revision=1375
responses and adds some more stuff.
I will have to decode NTcreateAndX requests and responses soon as well
as the MSRPC stuff ...
svn path=/trunk/; revision=1356
Now should be decoding the names of lots more LanMan API request. These
were culled from Samba. Would be good to go through and give names to the
fields as well.
Will soon decode the response structures returned and then will look at
ways to specify that built-in routines should be called to decode an element.
I also need some captures with UNICODE in them. Anyone got any? Someone
sent in a patch for UNICODE handling, but I did not realize what it was and
now the code has diverged so far it is hard to apply the patch ...
Send captures to rsharpe@ns.aus.com./
svn path=/trunk/; revision=1334
from Guy, plus a few more of my own.
Also added in basic response decoding where we don't know what it is ...
Got more to do, as well as decoding returned data ... Thinking about that
now, and will have a data-drived approach.
I need some way to specify that an internal routine be called for some types
of data where we know what type it is, in the case of Server Types for
example ...
svn path=/trunk/; revision=1294
Fixed up some bugs to do with NetShareEnum. There is still a persistent bug
left that looks like an alignment problem. Damn documentation does not talk
about the need to align the response structures for a NetServerEnum2 on SHORT
boundaries, but it sure looks like they should be so.
svn path=/trunk/; revision=1235
UNICODE strings in transact SMBs.
Added decode of NetShareEnum transact request. Will have to clean that all
up and use the decode engine when I get it done.
Still more fix ups to be done, but the book is calling, and I have to write
some stuff after an interview with LinuxCare.
svn path=/trunk/; revision=1113
I still have compile warnings, but I am too tired to chase them down.
Have also fixed a number of problems.
Next thing to add is a general engine that can decode Transact messages
as they can be decode from the descriptors in the Parameters area, and I
can feed in a list of names where we know what a structure looks like,
otherwise we use made up names.
svn path=/trunk/; revision=1079
(remove commas following the last member of an enum, make all bit fields
"guint32" - GCC lets you get away with that, but at least some other
compilers don't).
svn path=/trunk/; revision=1052
routines, which are called before a dissection pass is made over all the
packets in a capture - the "init" routine would clear out any state
information that needs to be initialized before such a dissection pass.
Make the NCP, SMB, AFS, and ONC RPC dissectors register their "init"
routines with that mechanism, have the code that reads in a capture file
call the routine that calls all registered "init" routines rather than
calling a wired-in set of "init" routines, and also have the code that
runs a filtering or colorizing pass over all the packets call that
routine, as a filtering or colorizing pass is a dissection pass.
Have the ONC RPC "init" routine zero out the table of RPC calls, so that
it completely erases any state from the previous dissection pass (so
that, for example, if you run a filtering pass, it doesn't mark any
non-duplicate packets as duplicates because it remembers them from the
previous pass).
svn path=/trunk/; revision=1050
dynamically-assigned "ett_" integer values, assigned by
"proto_register_subtree_array()"; this:
obviates the need to update "packet.h" whenever you add a new
subtree type - you only have to add a call to
"proto_register_subtree_array()" to a "register" routine and an
array of pointers to "ett_", if they're not already there, and
add a pointer to the new "ett_" variable to the array, if they
are there;
would allow run-time-loaded dissectors to allocate subtree types
when they're loaded.
svn path=/trunk/; revision=1043
There was a core dump because of a coding oversight. Should be fixed
now Gilbert.
Should now handle names of form \<somepipe>
Will screw up if there is no leading slash, but in a non-fatal way, I
think.
svn path=/trunk/; revision=1041
much more complete decoding of browse messages.
They are now shown in the parent tree as well.
I still have problems with:
1. Times
2. Election criteral
I also haven't been able to see a BecomeBackup request, nor a
GetBackupListResp with more than one browser ... Should run a Windows NT
server or another Samba on my network.
I am also not sure of there are any more message types. Damn! Old, and
wrong MS documents!
svn path=/trunk/; revision=1025
MS Windows Browser messages. Can decode host announcements now.
Still need to decode more. Also need to break the new code out.
I also have the Browse tree at the wrong location. Can I get at the
parent of the tree somewhere, or do I have to pass it in as a variable?
svn path=/trunk/; revision=1024
Replace "add_to_conversation()" with:
"conversation_new()", which creates a new conversation, given
source and destination addresses and ports, and returns a
pointer to the structure for the conversation;
"find_conversation()", which tries to find a conversation for
given source and destination addresses and ports, and returns a
pointer to the structure for the conversation if found, and a
null pointer if not found.
Add a private data pointer field to the conversation structure, and have
"conversation_new()" take an argument that specifies what to set that
pointer to; that lets clients of the conversation code hang arbitrary
data off the conversation (e.g., a hash table of protocol requests and
replies, in case the protocol is a request/reply protocol wherein the
reply doesn't say what type of request it's a reply to, and you need
that information to dissect the reply).
svn path=/trunk/; revision=920
structure to "dl_src"/"dl_dst", "net_src"/"net_dst", and "src"/"dst"
addresses, where an address is an address type, an address length in
bytes, and a pointer to that many bytes.
"dl_{src,dst}" are the link-layer source/destination; "net_{src,dst}"
are the network-layer source/destination; "{src,dst}" are the
source/destination from the highest of those two layers that we have in
the packet.
Add a port type to "packet_info" as well, specifying whether it's a TCP
or UDP port.
Don't set the address and port columns in the dissector functions; just
set the address and port members of the "packet_info" structure. Set
the columns in "fill_in_columns()"; this means that if we're showing
COL_{DEF,RES,UNRES}_SRC" or "COL_{DEF,RES,UNRES}_DST", we only generate
the string from "src" or "dst", we don't generate a string for the
link-layer address and then overwrite it with a string for the
network-layer address (generating those strings costs CPU).
Add support for "conversations", where a "conversation" is (at present)
a source and destination address and a source and destination port. (In
the future, we may support "conversations" above the transport layer,
e.g. a TFTP conversation, where the first packet goes from the client to
the TFTP server port, but the reply comes back from a different port,
and all subsequent packets go between the client address/port and the
server address/new port, or an NFS conversation, which might include
lock manager, status monitor, and mount packets, as well as NFS
packets.)
Currently, all we support is a call that takes the source and
destination address/port pairs, looks them up in a hash table, and:
if nothing is found, creates a new entry in the hash table, and
assigns it a unique 32-bit conversation ID, and returns that
conversation ID;
if an entry is found, returns its conversation ID.
Use that in the SMB and AFS code to keep track of individual SMB or AFS
conversations. We need to match up requests and replies, as, for
certain replies, the operation code for the request to which it's a
reply doesn't show up in the reply - you have to find the request with a
matching transaction ID. Transaction IDs are per-conversation, so the
hash table for requests should include a conversation ID and transaction
ID as the key.
This allows SMB and AFS decoders to handle IPv4 or IPv6 addresses
transparently (and should allow the SMB decoder to handle NetBIOS atop
other protocols as well, if the source and destination address and port
values in the "packet_info" structure are set appropriately).
In the "Follow TCP Connection" code, check to make sure that the
addresses are IPv4 addressses; ultimately, that code should be changed
to use the conversation code instead, which will let it handle IPv6
transparently.
svn path=/trunk/; revision=909
Is this file generated code ?
If not, please :
- get rid of compilation warnings
- put the # of preprocessor macros at the first column
svn path=/trunk/; revision=863
Added decoder for transact SMBs.
Changed things a lot for state keeping.
Next thing is to decode transacts and transact2s more.
svn path=/trunk/; revision=860
Also added first pass of state keeping. I am using glib's hash
functions.
Modelled after packet-ncp.c.
We will need to standardize the <proto>_init_protocol functions called in
file.c at some stage ...
I will have a couple of more goes at the state keeping before I am finished.
At the moment, the infrastructure is there but I do nothing with it.
svn path=/trunk/; revision=798