Wireshark-bugs: [Wireshark-bugs] [Bug 10503] TCP sequence number rollover: Wireshark incorrectly
maximalmin@gmail.com
changed
bug 10503
What |
Removed |
Added |
CC |
|
maximalmin@gmail.com
|
Comment # 7
on bug 10503
from maximalmin@gmail.com
I recently encountered what I believe to be the related bug (might be same bug)
where the Sequence number did not roll over, however, the packet analysis
apparently kicked off some other dissector mid stream, and when the next packet
is dissected, it reports the New fragment overlaps old data issue (where as a
manual analysis shows no such issue)
Capture and Analysis both on v1.12.2 (v1.12.2-0-g898fa22 from master-1.12)
The stream has mostly the same data so it's highly compressible.
Packet number 42643 shows the packet using standard TCP dissector
seq 5713245 (rel), len 268, next = 5713513 (abs seq = 00574756)
Packet number 42645 shows the packet using HiSLIP?? dissector, but tcp header:
seq 5713513 (rel), len 268, next = 5713781
Packet number 42647 shows the packet back to standard TCP dissector
seq 5713781 (rel), len 268, next = 5714049
Summary for packet shows the TCP: New fragment overlaps old data (retrans...)
text
Attaching the files as Wireshark-TCP-Sequence-Bug.7z
Screenshot in Wireshark-TCP-Sequence-Bug.png
You are receiving this mail because:
- You are watching all bug changes.