The last time I looked at actually using composite TVBs for, for
example, reassembly, I thought the problem was that the actual backing
store (for the TVBs that are populated from a file) is re-used. IOW
each TVB handed to epan for dissection is using the same buffer.
This was deep down in the depths somewhere; can't remember exactly where.
On 10/29/2012 02:18 PM, Zadik, Maayan wrote:
Hi Bill,
thanks for the reference :).
so here is where i'm at:
i have data that when too big gets fragmented, then my dissector
re-assembles it.
every data chunk in the assembled frame is wrapped with its own header
(basically containing the chunk's size)
so when i send the data to my mac dissector, i need to collect only the
data without the headers in a single TVB.
my implementation is as below:
{
tvbuff_t* total_payload = tvb_new_composite();
/* Loop over data elements */
for (j = 0; j < num_of_data_chunks; j++)
{
data_size = tvb_get_ntohs(tvb, *plen);
*plen += 2;
// get data chunk
val_ptr = tvb_get_ptr(tvb, *plen, data_size);
tvb_temp = tvb_new_real_data(val_ptr, data_size, data_size);
*plen += data_size;
/* Compose the final buffer with the temp buffer (add the curent data
chunk to the total data buffer) */
tvb_composite_append(total_payload, tvb_temp);
}
/* Finalize the composed buffer */
tvb_composite_finalize(total_payload);
DISSECTOR_ASSERT(total_payload != NULL);
call_dissector_only(my_mac_lte_handle, total_payload, pinfo, tree);
}
the problem occurs when i try to create a subset of this TVB:
tvbuff_t *srb_tvb = tvb_new_subset(tvb, offset, rlc_data_length,
rlc_data_length);
call_dissector_only(my_rlc_lte_handle, srb_tvb, pinfo, tree);
when printing the the original TVB's and the subset's params i see the
following:
tvb_addr 1st
byte offset from base tvb length
so, you see when i try to create a subset, instead of the data being
0x0b i'm getting 0xaa and i don't know why.
can you spot what i'm doing wrong?
Thanks,
maayan
________________________________________
From: wireshark-dev-bounces@xxxxxxxxxxxxx
[wireshark-dev-bounces@xxxxxxxxxxxxx] on behalf of Bill Meier
[wmeier@xxxxxxxxxxx]
Sent: Monday, October 29, 2012 6:06 PM
To: Developer support list for Wireshark
Subject: Re: [Wireshark-dev] Appending 2 TVBs in Wireshark 1.8.2
On 10/28/2012 12:20 PM, Zadik, Maayan wrote:
> Hi all,
>
> How do I create a single TVB from several data chunks?
>
> Is there a way to append two TVBs?
> I know the composite doesn't work. I tried it.
>
>
> Regards,
> Maayan
>
>
>
Well: you can create your own tvb using tvb_new_real_data() pointing to
a buffer you've created containing the data you've copied from the two
tvb's.
That being said, the composite TVB code should work. (I fixed the code
some time ago and the code is part 1.8).
(I think composite TVbs are even being used someplace in 1.80).
What problems did you have using a composite TVB ?
Providing the essentials of code you used would help us to see what's
happening.
Also, see epan/tvbtest.c for an example of code which creates composite
tvb's.
(Ignore the comment about "Composite tvbuffs don't work at the
moment...". AFAIK composite TVBs should work. I apparently forgot to
remove the comment when I did the work a while back).
Bill
___________________________________________________________________________
Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives: http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe
___________________________________________________________________________
Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives: http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe
