Ethernet, IP, ICMP, UDP, TCP Mini-Stack
This is not, as the title implies, a fully functional TCP/IP stack. It does however reply to most incomming TCP/IP connections and other layer requests (including ICMP ECHO requests - Ping). This function should be passed raw ethernet frames, and will generate a response packet that will convince other stacks on a network that this machine can be connected to(SYN/ACK in response to SYN requests). For example, if one telnets to a machine running this function, a connection will be successfully made, and any data sent will be sent back (and printed on the machine running this code.) This code is part of a metwprl operating system Brandon Burr and Martin McCormick (myself), are writing. We cannot release the full code. However I feel that this may be beneficial to others writing network functions. (including TCP/UDP/IP checksum code). NOTE: This should not be taken seriously as a TCP/IP stack, as that it remembers NO state information. It can be used (as this is) as a lower level protocol handler for your own userlevel protocol).
AI
KI-Zusammenfassung: This codebase represents a historical implementation of the logic described in the metadata. Our preservation engine analyzes the structure to provide context for modern developers.
Quellcode
-----BEGIN ETHERNET.H----- /* NETWORK.H - The network header file Written by Martin McCormick ([email protected]) (c) 2004 for the Surf Operating System */ /* because of loopback references, need structure prototypes (weird huh?) */ struct iface; struct sPacketBuf; struct sPacket { unsigned int iLen; unsigned char * pBuf; unsigned int iFlags; /* 32 bit */ /* -Bits 0 through 7 - CHECKSUM FLAGS (1<<0) = ETHERNET checksum checked (1<<1) = ENCAPSULATION checksum checked (1<<2) = IP checksum checked (1<<3) = TCP/UDP checksum checked (1<<4) = ETHERNET checksum correct (1<<5) = ENCAPSULATION checksum correct (1<<6) = IP checksum correct (1<<7) = TCP/UDP checksum correct -Bits 8 through 15 - USER LEVEL FLAGS -Bits 16 through 23 - HANDLER FLAGS (if any set, driver will not return packet as 'on buffer') (1<<16) = Packet is activated for proccessing. (1<<17) = Packet is on TX queue. (1<<18) = Packet has been sent and freed. -Bits 24 through 32 - DRIVER FLAGS (1<<24) = Packet synched. */ }; enum PacketChecksumFlags { CHECKSUM_FLAGS = (0xff<<0), ETH_SUM_CHECKED = (1<<0), ENC_SUM_CHECKED = (1<<1), IP_SUM_CHECKED = (1<<2), TRANS_SUM_CHECKED = (1<<3), ETH_SUM_CORRECT = (1<<4), ENC_SUM_CORRECT = (1<<5), IP_SUM_CORRECT = (1<<6), TRANS_SUM_CORRECT = (1<<7) }; enum PacketHandlerFlags { HANDLER_FLAGS = (0xff<<16), PACKET_ACTIVE = (1<<16), PACKET_ON_TX = (1<<17), PACKET_SENT = (1<<18) }; enum PacketDriverFlags { DRIVER_FLAGS = (0xff<<24), PACKET_SYNCHED = (1<<24), PACKET_LOST = (1<<25) }; /* SnagPackets(struct iface * interface, struct sPacketBuf ** packets, unsigned int * count); SnagPackets receives a given number of packets (count) iface - the interface to receive on (should be owner of function pointer) packets - a pointer to the head of an array of packet pointers count - takes in max number to receive, output is actual count */ typedef int(* FuncSnag) ( /* parameters */ struct iface * interface, struct sPacketBuf ** packets, unsigned int * count ); /* SendPackets(struct iface * interface, struct sPacketBuf ** packets, unsigned int * count); SendPackets sends a given number of packets (count) iface - the interface to send upon (should be owner of function pointer) packets - a pointer to the head of an array of pointers to outgoing packets count - takes in number to send, output is actual sent */ typedef int(* FuncSend) ( /* parameters */ struct iface * interface, struct sPacketBuf ** packets, unsigned int * count ); /* Setting(struct iface * interface, int iSetting, void * param); Setting() sets (or receives) network card information. iface - interface to use iSetting - a constant reffering to the type of setting (see note below) param - Used for I/O operations on settings */ typedef int(* FuncSetting) ( /* parameters */ struct iface * interface, unsigned int iSetting, void * param ); /* iface - The main interface structure used for network interfaces */ struct iface { unsigned int id; bool bEnabled; void * driver_struct; /* used internally by driver */ /* standard io functions */ FuncSetting Setting; FuncSnag SnagPackets; FuncSend SendPackets; unsigned char MacAddress[6]; unsigned char MacMask[6]; }; /* The Setting technique is as follows: unsigned int iSetting = 32 16 0 (bit) |XXXX XXXX|XXXX XXXX|YYYY YYYY|YYYY YYYY| X -> General Command Type (number from 0x0 to 0xFFFF) Y -> Specific Command Type (number from 0x0 to 0xFFFF) NOTE: Only one Gereral type per command. Specific commands are OR able (for more than one) */ enum GeneralCommands { maskGeneralCommands = (0xffff<<16), IfaceDisable = (0x00<<16), IfaceEnable = (0x01<<16), IfaceReset = (0x02<<16), IfaceQueryState = (0x03<<16) }; enum SpecificCommands { /* for enable/disable/receive: */ maskSpecificCommands = (0xffff), SpecificallyAll = (1<<0), SpecificallyReceive = (1<<1), SpecificallyTransmit = (1<<2), SpecificallyStatistics = (1<<3), SpecificallyInterrupts = (1<<4) }; ----- END ETHERNET.H ----- ----- BEGIN PACKETRESPONDER.H ----- #ifndef _PRESP_H #define _PRESP_H #include <blibc_common.h> struct etherhdr { unsigned char ether_dest[6]; /*Destination MAC address. */ unsigned char ether_src[6]; /*Source MAC address. */ unsigned short ether_type; /*Protocol type. */ }; enum etherTypes { etherTypeIP = 0x0008, etherTypeARP = 0x0608 }; struct iphdr { unsigned char ip_hl:4, /*Header length. */ ip_v:4; /*Version. */ unsigned char ip_tos; /*Type of service. */ short ip_len; /*Total length. */ unsigned short ip_id; /*Identification. */ short ip_off; /*Fragment offset field. */ unsigned char ip_ttl; /*Time to live. */ unsigned char ip_p; /*Protocol. */ unsigned short ip_sum; /*Checksum. */ unsigned int ip_src; /*Source IP address. */ unsigned int ip_dest; /*Destination IP address. */ }; enum ipTypes { ipTypeICMP = 0x01, ipTypeUDP = 0x11, ipTypeTCP = 0x06 }; struct icmphdr { unsigned char icmp_type; /*Type of message. */ unsigned char icmp_code; /*Type sub code. */ unsigned short icmp_cksum; /*Ones complement header checksum. */ unsigned short icmp_id; /* identifier */ unsigned short icmp_seq; /* sequence number */ char icmp_data[1]; /* data */ }; struct udphdr { unsigned short udp_sport; /*Source port */ unsigned short udp_dport; /*Destination port */ unsigned short udp_ulen; /*UDP length */ unsigned short udp_sum; /*UDP checksum */ }; struct tcphdr { unsigned short tcp_sport; /* Source port. */ unsigned short tcp_dport; /* Destination port. */ unsigned int tcp_seq; /* Sequence number of first octet in this segment. */ unsigned int tcp_ack; /* Expected sequence number of next octet. */ unsigned char tcp_x2:4, /* Unused. */ tcp_off:4; /* Data offset. */ unsigned char tcp_flags; /* Control flags. */ unsigned short tcp_win; /* Number of acceptable octects. */ unsigned short tcp_sum; /* 96 byte pseudo header checksum. */ unsigned short tcp_urp; /* Urgent data pointer. */ }; enum tcpFlags { tcpFin = (1<<0), tcpSyn = (1<<1), tcpReset = (1<<2), tcpPush = (1<<3), tcpAck = (1<<4), tcpUrgent = (1<<5), tcpEcho = (1<<6), tcpCWR = (1<<7) }; unsigned int createResponse(struct sPacket * pPacket); #endif ----- END PACKETRESPONDER.H ----- ----- BEGIN PACKETRESPONDER.C ----- #include "ethernet.h" #include "PacketResponder.h" #define errBufferUnderrun 1 #define errBadEtherType 2 #define errBadIPHeaderLen 3 #define errBadIpType 4 #define errNotReplyable 5 #define errBadIPLen 6 #define errBadTCPHeaderLen 7 /* icmpReplyCodes: Index = incoming request TYPE, Value = Reply value, 0xff = no reply */ unsigned char icmpReplyCodes[256] = {0xff, 0xff, 0xff, 0xff, 0, 0, 0, 0xff, 0, 10, 0xff, 0xff, 0xff, 14, 0xff, 16, 0xff, 18, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 34, 0xff, 36, 0xff, 38, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-50xff */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-75 */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-10xff0xff */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-125 */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-150xff */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-175 */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-20xff0xff */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-225 */ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* <-250xff */ 0xff, 0xff, 0xff, 0xff, 0xff}; /* <-255 */ unsigned short h2ns(unsigned short ushort) { return ((ushort >> 8) & 0xff) | ((ushort << 8) & 0xff00); } unsigned int h2nsl(unsigned int uint) { return ((uint >> 24) & 0x000000ff) | ((uint >> 8 ) & 0x0000ff00) | ((uint << 8) & 0x00ff0000) | ((uint << 24) & 0xff000000); } /* IP Checksum Function This is used for both the IP header and an ICMP packet */ unsigned short checksum(unsigned short *buffer, unsigned int size) { unsigned long cksum=0; while (size > 1) { cksum += *buffer++; size -= sizeof(unsigned short); } if (size) cksum += *(unsigned char*)buffer; cksum = (cksum >> 16) + (cksum & 0xffff); cksum += (cksum >>16); cksum = (~cksum); cksum = cksum ? cksum : 0xffff; return (unsigned short)cksum; } unsigned short TransChecksum(struct iphdr * ipHeader) { unsigned short * buffer; unsigned int size; unsigned long cksum; cksum = 0; cksum += (ipHeader->ip_src & 0xffff) + ((ipHeader->ip_src>>16) & 0xffff); /* ip_src */ cksum += (ipHeader->ip_dest & 0xffff) + ((ipHeader->ip_dest>>16) & 0xffff); /* ip_dest */ cksum += (unsigned short)((ipHeader->ip_p << 8) & 0xff00); cksum += h2ns(h2ns(ipHeader->ip_len)-(4 * ipHeader->ip_hl)); size = h2ns(ipHeader->ip_len) - (4 * ipHeader->ip_hl); /* previously checked for sanity */ buffer = (unsigned short *)((unsigned char *)ipHeader + (4 * ipHeader->ip_hl)); while (size > 1) { cksum += *buffer++; size -= sizeof(unsigned short); } if (size) cksum += *(unsigned char*)buffer & 0x00ff; cksum = (cksum >>16) + (cksum & 0xffff); cksum += (cksum >>16); cksum = (~cksum); cksum = cksum ? cksum : 0xffff; return (unsigned short)cksum; } unsigned int createResponse(struct sPacket * pPacket) { /* Header pointers */ struct etherhdr * etherHeader; struct iphdr * ipHeader; struct udphdr * udpHeader; struct icmphdr * icmpHeader; struct tcphdr * tcpHeader; /* Length currently left to proccess */ unsigned int iLenUnCasted; unsigned char * pProcessPoint; /* Temp variables for creating response */ int i; unsigned char tempMac[6]; unsigned int tempIP, tempSeq; unsigned short tempPort; pProcessPoint = pPacket->pBuf; iLenUnCasted = pPacket->iLen; /* Look for ethernet header */ if(iLenUnCasted < sizeof(struct etherhdr)) return errBufferUnderrun; etherHeader = (struct etherhdr *)pProcessPoint;/* cast it to ethernet header */ pProcessPoint += sizeof(struct etherhdr); iLenUnCasted -= sizeof(struct etherhdr); for(i=0;i<6;i++) { /* flip source / destination MAC */ tempMac[i] = etherHeader->ether_src[i]; etherHeader->ether_src[i] = etherHeader->ether_dest[i]; etherHeader->ether_dest[i] = tempMac[i]; } switch(etherHeader->ether_type) { case etherTypeIP: { /* look for ip headers */ if(iLenUnCasted < sizeof(struct iphdr)) return errBufferUnderrun; ipHeader = (struct iphdr *)pProcessPoint;/* cast it to ip header */ if((4 * ipHeader->ip_hl) < sizeof(struct iphdr) || (4 * ipHeader->ip_hl) > iLenUnCasted) return errBadIPHeaderLen; if(h2ns(ipHeader->ip_len) < sizeof(struct iphdr) || h2ns(ipHeader->ip_len) > iLenUnCasted) return errBadIPLen; pProcessPoint += (4 * ipHeader->ip_hl); iLenUnCasted = h2ns(ipHeader->ip_len) - (4 * ipHeader->ip_hl); ipHeader->ip_id--; /* just simply decrement to make it different */ ipHeader->ip_off = h2ns(h2ns(ipHeader->ip_off) & (0x3<<13)); /* keep flags same, but zero offset */ tempIP = ipHeader->ip_src; /* flip source / destination IP */ ipHeader->ip_src = ipHeader->ip_dest; ipHeader->ip_dest = tempIP; switch(ipHeader->ip_p) { case ipTypeICMP: { if(iLenUnCasted < sizeof(struct icmphdr)) return errBufferUnderrun; icmpHeader = (struct icmphdr *)pProcessPoint;/* cast it to icmp header */ pProcessPoint += sizeof(struct icmphdr); iLenUnCasted -= sizeof(struct icmphdr); icmpHeader->icmp_type = icmpReplyCodes[icmpHeader->icmp_type]; if(icmpHeader->icmp_type==0xff) return errNotReplyable; /* proccess subsequent data (note: icmp header may be larger) */ for(i=0;i<iLenUnCasted;i++) putch(pProcessPoint[i]); icmpHeader->icmp_cksum = 0; icmpHeader->icmp_cksum = checksum((unsigned short *)icmpHeader, iLenUnCasted + sizeof(struct icmphdr)); /* TODO: fix checksum */ break; } case ipTypeUDP: { if(iLenUnCasted < sizeof(struct udphdr)) return errBufferUnderrun; udpHeader = (struct udphdr *)pProcessPoint;/* cast it to icmp header */ pProcessPoint += sizeof(struct udphdr); iLenUnCasted -= sizeof(struct udphdr); tempPort = udpHeader->udp_sport; /* flip source /destination port */ udpHeader->udp_sport = udpHeader->udp_dport; udpHeader->udp_dport = tempPort; /* pass to proccess (if replyable) */ for(i=0;i<iLenUnCasted;i++) putch(pProcessPoint[i]); /* fix length then checksum */ udpHeader->udp_sum = 0; udpHeader->udp_sum = TransChecksum(ipHeader); break; } case ipTypeTCP: { if(iLenUnCasted < sizeof(struct tcphdr)) return errBufferUnderrun; tcpHeader = (struct tcphdr *)pProcessPoint;/* cast it to icmp header */ if((4 * tcpHeader->tcp_off) < sizeof(struct tcphdr) || (4 * tcpHeader->tcp_off) > iLenUnCasted) return errBadTCPHeaderLen; pProcessPoint += (4 * tcpHeader->tcp_off); iLenUnCasted -= (4 * tcpHeader->tcp_off); /* don't respond to packets that don't need responses. */ if(!iLenUnCasted && !(tcpHeader->tcp_flags & (tcpSyn | tcpFin)) ) return errNotReplyable; /* flip source /destination port */ tempPort = tcpHeader->tcp_sport; tcpHeader->tcp_sport = tcpHeader->tcp_dport; tcpHeader->tcp_dport = tempPort; /* sequence numbers */ tempSeq = tcpHeader->tcp_seq; /* save their seq */ if(tcpHeader->tcp_flags & tcpSyn) /* this is the first packet */ { tcpHeader->tcp_seq = h2nsl(h2nsl(tcpHeader->tcp_seq)+0xff); /* use one based on theirs */ }else{ /* just use what they are ack'ing */ tcpHeader->tcp_seq = tcpHeader->tcp_ack; } tcpHeader->tcp_flags |= tcpAck; /* we are always ack'ing a packet */ tcpHeader->tcp_ack = h2nsl(h2nsl(tempSeq)+1); /* ack their sequence */ /* proccess options */ for(i=0;i<iLenUnCasted;i++) putch(pProcessPoint[i]); /* checksum */ tcpHeader->tcp_sum = 0; tcpHeader->tcp_sum = TransChecksum(ipHeader); //pPacket->iFlags &= ~TRANS_SUM_CHECKED; break; } default: return errBadIpType; } /* END switch(ipHeader->ip_p) */ /* TODO: fix ip len and checksum */ ipHeader->ip_sum = 0; ipHeader->ip_sum = checksum((unsigned short *)ipHeader, sizeof(struct iphdr)); break; } case etherTypeARP: { return 10; break; } default: printf("BADETHER:0x%x",etherHeader->ether_type); return errBadEtherType; } /* END switch(etherHeader->ether_type) */ return 0; } ----- END PACKETRESPONDER.C ----- Upload Upload
Originalkommentare (3)
Wiederhergestellt von der Wayback Machine