- if offset is 0, tvb_length is the same as tvb_length_remaining, just faster.
Replace
- col_append_fstr() with faster col_append_str()
- col_add_str() with col_set_str()
when it's safe
svn path=/trunk/; revision=23252
scsi sense information potentially followed by iscsi event data.
this is used by a target to tell an initiator (among other things) that the lun configuration has changed and now is a good time to rescan the target for lun changes.
svn path=/trunk/; revision=20531
and also (if reassembly is disabled) only dissect the initial (offset==0) data pdu.
dissect_scsi_payload() does not yet use this parameter.
now that we have both data offset and expected data length/bidir expected data length and also the read/write flags availabe we have what we need to reassemble data in/out pdus (modulo overflow/underflow but those are so rare we can worry about them later).
ndmp: ndmp conceptually always has a data in and a data out phase and never fragment the data into smaller pdu's os that dissector always report offset as 0.
svn path=/trunk/; revision=19511
pass conversation form the transports up to the scsi layer
add tracking of conversation specific info to scsi osd
add tracking of conversation+lun specific info to scsi osd
for scsi osd add tracking of PARTITIONS and display in which frame they were created/removed
svn path=/trunk/; revision=19505
so that the two scsi transports FCP and ISCSI can provide the expected data transfer lengths to SCSI to allow SCSI reassembly.
NDMP does not really need these hints since for NDMP (and also iscsi-lite) there is conceptually always both data in and data out phases and there is never any fragmentation.
svn path=/trunk/; revision=19493
account for this extra reserved byte in the ahs length so that the reconstructed cdb has the correct length and does not contain one extra byte at the end
svn path=/trunk/; revision=19387
iscsi: when iscsi transfers a cdb that is alrger than 16 bytes, the first 16 bytes are transferred in the normal place in the header and ther remainder of the cdb is transported inside the AHS.
reassemble these cdb into a proper tvb before passing it to the scsi dissector
svn path=/trunk/; revision=19376
it is absolutely amazing that none of the iscsi implementors and users of wireshark had noticed this breakage and reported it. they apparently do not use wireshark.
svn path=/trunk/; revision=19362
and the weak heuristics often cause wireshark to mistake some segment containing read/write data to be iscsi.
make the heuristics to check that a packet really is iscsi much stronger
svn path=/trunk/; revision=18523
update the comment in packet-scsi.c to reflect that it is the transport now that is responsible to track itl and itlq data
make scsi tapable
svn path=/trunk/; revision=17974
structures for scsi.
we no longer need the scsi_task_id structure passed by pinfo->private_data so get rid of it.
we no longer need the (broken by design) scsi_task_data hash table since this has been replaced byt hte itl and itlq structures and tracking
svn path=/trunk/; revision=17952
remove the two fields opcode and devtype from the scsi_task_data structure since these are also part of the itlq and itl structures
svn path=/trunk/; revision=17949
and get rid of some breakage in the design
let the scsi transport keep track of itl (initiator, target, lun) matching
and let it pass a itl structure to scsi that is persistent across packets.
let scsi use this itl structure to track device type for a specific itl instead of the (must have been) broken hashtable.
update both iscsi and fc to track the itl structure for scsi and schange the scsi signature to accept itl as a parameter.
more to come.
svn path=/trunk/; revision=17942
this finally allows us to have scsi.time for scsi transactions but we need to cleanup and refactor the other three scsi entrypoints before we should implement scsi srt to make the other three also take an exchange data structure as parameter from their transports (and get rid of the pinfo->private_data )
svn path=/trunk/; revision=17838
while reusing this structure for multi-purpose would make it berret suited to move the struct from packet-fc.h to packet-scsi.h
we would have to include packet-scsi.h for everyone that includes packet-fc.h and that are a few dependencies.
(would have to include it even in packet-llc.c)
so leave it as a fc structure for now.
svn path=/trunk/; revision=17833
use the pointer to iscsi_Session as conversation index for the exchange tracking for now, it works just as well and will be replaced with better exchange tracking shortly anyway
svn path=/trunk/; revision=17831
I know nothing about iscsi , but try the attached patch. It fix the
wrong desegmentation and the ethereal bug isn't triggered anymore.
svn path=/trunk/; revision=17445
I've done more than a day to change the timestamp resolution from microseconds to nanoseconds. As I really don't want to loose those changes, I'm going to check in the changes I've done so far. Hopefully someone else will give me a helping hand with the things left ...
What's done: I've changed the timestamp resolution from usec to nsec in almost any place in the sources. I've changed parts of the implementation in nstime.s/.h and a lot of places elsewhere.
As I don't understand the editcap source (well, I'm maybe just too tired right now), hopefully someone else might be able to fix this soon.
Doing all those changes, we get native nanosecond timestamp resolution in Ethereal. After fixing all the remaining issues, I'll take a look how to display this in a convenient way...
As I've also changed the wiretap timestamp resolution from usec to nsec we might want to change the wiretap version number...
svn path=/trunk/; revision=15520
If we encounter a frame we have already seen (i.e. visited==1) but we can not fing the ISCSI command inside the _matched table, this just means there were no response pdu nor any data pdu with the S bit set.
So, then just pick the cdata structure up from the _unmatched table instead.
The SCSI CDB dissector really really want a real CDATA structure passed to it.
svn path=/trunk/; revision=15291
making pointers to byte data be "guint8 *" rather than "char *",
and making buffers holding byte data arrays of "guint8" rather
than arrays of "char";
making pointers to text strings "char *" rather than "guchar *";
appropriately casting pointers (cast to "guint8 *" when passing
to routines expecting "guint8 *" or when assigning to "guint8
*");
making port-number preferences "guint";
making enum preferences "gint";
making hf_ variables "int".
Clean up white space.
svn path=/trunk/; revision=14884
in a simple approach, I've replaced all g_assert() and g_assert_not_reached() calls by their exception throwing counterparts DISSECTOR_ASSERT() and DISSECTOR_ASSERT_NOT_REACHED()
this will replace application crash by showing a dissector bug, which is the desired behaviour
there were some g_assert calls in the protocol registering functions, which might not be acting as expected now, but to be able to simply search for g_assert in the future I've replaced that calls too
one g_assert remained, the one when someone throws an unknown exception "into" packet_frame.c, but IMHO this one should remain.
svn path=/trunk/; revision=14608
so that we block any heuristic dissectors from attempring and mistaking this conversation from being it.
In particular: DCERPC often mistakes iscsi DataOut PDUs for being DCERPC.
svn path=/trunk/; revision=14243
Read the LUN value into a local variable always and use that to pass to the scsi dissector since we can always use this value from the iscsi heade for COMMAND pdus (but not data/in/out)
svn path=/trunk/; revision=13791
the iscsi layer will not hold the LUN value (it is reserved)
so we need to remember if from the initial Command PDU.
make the LUN reporting work for errorrecoverylevel==0 targets
svn path=/trunk/; revision=13563
on scsi.lun and prettify the summary line a bit.
ndmp still needs some work to track luns between commands
and fcp needs verification it works for volumesetaddressing.
svn path=/trunk/; revision=13420
1) Added a setup_frame parameter to conversation_t
2) Used the conversation_t next to maintain a list of conversations with the
same src/dest tuple but different setup_frame number.
3) Changed the signature of find_conversation() and conversation_new() to pass
in the frame number.
4) Adjusted packet-sdp to select RTP conversation if both m=audio and m=image
are present, and T.38 conversation if only m=image is present. I expect that
RTP/T.38 dissecting to be better, but I don't have a way to generate T.38
packets.
svn path=/trunk/; revision=13243