Search code examples
pythonscapypcaptcpdump

How to generate packet having sequence number as `first:last` with Scapy?


I am trying to generate pcap file from a tcpdump output, how can I generate those packets having sequence number as first:last?

Here is what my tcpdump input looks like:

tcpdump: listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes
1509471560.944080 MAC1 > MAC2, ethertype IPv4 (0x0800), length 74: (tos 0x0, ttl 64, id 23237, offset 0, flags [DF], proto TCP (6), length 60)
    IP1.port > IP2.port: Flags [S], cksum 0x6d2f (incorrect -> 0x0b4a), seq 1127096708, win 65535, options [mss 1460,sackOK,TS val 817985 ecr 0,nop,wscale 6], length 0
1509471561.042855 MAC2 > MAC1, ethertype IPv4 (0x0800), length 58: (tos 0x0, ttl 64, id 3107, offset 0, flags [none], proto TCP (6), length 44)
    IP2.port > IP1.port: Flags [S.], cksum 0x85d8 (correct), seq 449984001, ack 1127096709, win 65535, options [mss 1460], length 0
1509471561.044008 MAC1 > MAC2, ethertype IPv4 (0x0800), length 54: (tos 0x0, ttl 64, id 23238, offset 0, flags [DF], proto TCP (6), length 40)
    IP1.port > IP2.port: Flags [.], cksum 0x6d1b (incorrect -> 0x9d95), seq 1, ack 1, win 65535, length 0
1509471561.046607 MAC1 > MAC2, ethertype IPv4 (0x0800), length 191: (tos 0x0, ttl 64, id 23239, offset 0, flags [DF], proto TCP (6), length 177)
    IP1.port > IP2.port: Flags [P.], cksum 0x6da4 (incorrect -> 0x98df), seq 1:138, ack 1, win 65535, length 137
1509471914.089046 MAC1 > MAC2, ethertype IPv4 (0x0800), length 82: (tos 0x0, ttl 64, id 54304, offset 0, flags [DF], proto UDP (17), length 68)

Following is the code I have prepared to process the TCP packets:

from scapy.all import *
import secrets

def generatePcapfromText(inputtxt,output):
    with open (inputtxt,encoding='cp850') as input:
        framenum=0
        for line in input:
            if line[0].isdigit(): # line one
                framenum += 1
                frametime=float(line[:16])
                
                srcmac= line[18:34]
                dstmac= line[38:54]
                ethertype = int(line[line.find('(')+1:line.find(')')], 16)
                frameLen=int(line[line.find('length')+7:line.find(': (')])
                frameTos=int(line[line.find('tos')+4:line.find(', ttl')],16)
                frameTtl=int(line[line.find('ttl')+4:line.find(', id')])
                frameId=int(line[line.find('id')+3:line.find(', offset')])
                frameOffset=line[line.find('offset')+7:line.find(', flags')]
                frameFlags=line[line.find('[')+1:line.find(']')]
                protocol = line[line.find('proto')+6:line.rfind('(')-1]
                ipLen = int(line[line.rfind('length')+6:line.rfind(')')])
                
                if frameFlags == "none":
                    frameFlags = ""

                ether = Ether(dst=dstmac, src=srcmac, type=ethertype) 

            elif len(line)>5:
                if line[5].isdigit(): # line two
                    srcinfo = line[4:line.find ( '>' )]
                    dstinfo = line[line.find ( '>' ) + 2:line.find ( ':' )]
                    ipsrc = srcinfo[:srcinfo.rfind ( '.' )]
                    ipdst = dstinfo[:dstinfo.rfind ( '.' )]
                    srcport = int(srcinfo[srcinfo.rfind ( '.' ) + 1:])
                    dstport = int(dstinfo[dstinfo.rfind ( '.' ) + 1:])
                    
                    ip = ether/IP(src=ipsrc, dst=ipdst, len=frameLen, tos=frameTos, ttl=frameTtl, id=frameId, flags=frameFlags, proto=protocol.lower())
                   
                    if protocol == "TCP":

                        frameFlag = line[line.find ( '[' ) + 1:line.find ( ']' )]
                        frameFlag=frameFlag.replace(".","A")
                            
                        cksum = int(line[line.find ( 'cksum' ) + 6:line.find ( '(' )],16)
                        
                        if ", ack" in line:      
                            seq_n = line[line.find ( ', seq' ) + 6:line.find ( ', ack' )]
                            ack_n = int(line[line.find ( 'ack' ) + 4:line.find ( ', win' )])
                        else:
                            seq_n = line[line.find ( ', seq' ) + 6:line.find ( ', win' )]
                            ack_n = 0
                            
                        if "options" in line:
                            win = int(line[line.find ( 'win' ) + 4:line.find ( ', options' )])
                            options= line[line.find ( 'options' ) + 8:line.find ( ', length' )]
                        else:    
                            win = int(line[line.find ( 'win' ) + 4:line.find ( ', length' )])
                            options="[]"
                            
                        pktlen = int(line[line.find ( ', length' ) + 9:])
                        
                        if ":" in seq_n: 
                            # ???
                        else:   
                            pkt = ip / TCP(sport=srcport, dport=dstport , flags=frameFlag, seq=int(seq_n), ack=ack_n, chksum=cksum, window=win) / secrets.token_hex(pktlen)                  

                    pkt.time = frametime
                                                  
                    wrpcap(output, pkt, append=True)

As TCP in Scapy need an integer for the sequence number I cannot pass it first:last as the sequence number, so It needs some modification which I am not familiar with as It is my first time working with Scapy. I have mark where this modification should be done via #??? in the code above.

For my purpose it is important that the packets have the same timestamp as the tcpdump input, so I have set the packet timestamp via pkt.time=timestamp.

PS: You can find history behind this question here.


Solution

  • According to the tcpdump manpage

    The notation is first:last which means sequence numbers first up to but not including last

    The first:last notation doesn't actually exist within the packet, therefore Scapy won't understand it. Tcpdump gives you this additional information based on the analysis of the TCP stream (and possible regrouping of the packets)

    You should also note that your sequence numbers are probably relative, which means they don't really mean anything else than the order.

    Using the first part of the number as a sequence number will probably be enough for your needs:

    seq = int(seq_n.split(":")[0])