Add initial support for QUIC, improve logging capabilities.

Add TRACE logging mode

args.c

@@ -32,10 +32,11 @@ struct config_t config = {
 #endif
 
 #ifdef DEBUG
-	.verbose = true,
+	.verbose = 1,
 #else
-	.verbose = false,
+	.verbose = 0,
 #endif
+
 	.domains_str = defaul_snistr,
 	.domains_strlen = sizeof(defaul_snistr),
 
@@ -53,13 +54,15 @@ struct config_t config = {
 #define OPT_FRAG_SNI_REVERSE	12
 #define OPT_FRAG_SNI_FAKED	13
 #define OPT_FK_WINSIZE		14
+#define OPT_TRACE		15
+#define OPT_QUIC_DROP		16
 #define OPT_SEG2DELAY 		5
 #define OPT_THREADS 		6
 #define OPT_SILENT 		7
 #define OPT_NO_GSO 		8
 #define OPT_QUEUE_NUM		9
 
-#define OPT_MAX OPT_FRAG_SNI_FAKED
+#define OPT_MAX OPT_QUIC_DROP
 
 static struct option long_opt[] = {
 	{"help",		0, 0, 'h'},
@@ -73,9 +76,11 @@ static struct option long_opt[] = {
 	{"frag-sni-reverse",	1, 0, OPT_FRAG_SNI_REVERSE},
 	{"frag-sni-faked",	1, 0, OPT_FRAG_SNI_FAKED},
 	{"fk-winsize",		1, 0, OPT_FK_WINSIZE},
+	{"quic-drop",		0, 0, OPT_QUIC_DROP},
 	{"seg2delay",		1, 0, OPT_SEG2DELAY},
 	{"threads",		1, 0, OPT_THREADS},
 	{"silent",		0, 0, OPT_SILENT},
+	{"trace",		0, 0, OPT_TRACE},
 	{"no-gso",		0, 0, OPT_NO_GSO},
 	{"queue-num",		1, 0, OPT_QUEUE_NUM},
 	{0,0,0,0}
@@ -120,9 +125,11 @@ void print_usage(const char *argv0) {
 	printf("\t--frag-sni-reverse={0|1}\n");
 	printf("\t--frag-sni-faked={0|1}\n");
 	printf("\t--fk-winsize=<winsize>\n");
+	printf("\t--quic-drop\n");
 	printf("\t--seg2delay=<delay>\n");
 	printf("\t--threads=<threads number>\n");
 	printf("\t--silent\n");
+	printf("\t--trace\n");
 	printf("\t--no-gso\n");
 	printf("\n");
 }
@@ -140,12 +147,18 @@ int parse_args(int argc, char *argv[]) {
 		case 'v':
 			print_version();
 			goto stop_exec;
+		case OPT_TRACE:
+			config.verbose = 2;
+			break;
 		case OPT_SILENT:
 			config.verbose = 0;
 			break;
 		case OPT_NO_GSO:
 			config.use_gso = 0;
 			break;
+		case OPT_QUIC_DROP:
+			config.quic_drop = 1;
+			break;
 		case OPT_SNI_DOMAINS:
 			if (!strcmp(optarg, "all")) {
 				config.all_domains = 1;
@@ -329,7 +342,12 @@ void print_welcome() {
 		printf("GSO is enabled\n");
 	}
 
+	if (config.quic_drop) {
+		printf("All QUIC packets will be dropped\n");
+	}
+
 	if (config.all_domains) {
 		printf("All Client Hello will be targetted by youtubeUnblock!\n");
 	}
+
 }

config.h

@@ -25,7 +25,11 @@ struct config_t {
 	unsigned char faking_ttl;
 	int fake_sni;
 	unsigned int fake_sni_seq_len;
+#define VERBOSE_INFO	0
+#define VERBOSE_DEBUG	1
+#define VERBOSE_TRACE	2
 	int verbose;
+	int quic_drop;
 	/* In milliseconds */
 	unsigned int seg2_delay;
 	const char *domains_str;

logging.h

@@ -0,0 +1,39 @@
+#ifndef LOGGING_H
+#define LOGGING_H
+#include "config.h"
+
+#define LOG_LEVEL (config.verbose)
+
+#ifdef KERNEL_SPACE
+#include <linux/printk.h>
+#define printf pr_info
+#define perror pr_err
+#define lgerror(msg, ret, ...) __extension__ ({		\
+	printf(msg ": %d\n", ##__VA_ARGS__, ret);	\
+})
+#else
+#include <stdio.h> // IWYU pragma: export
+#include <errno.h>
+#define lgerror(msg, ret, ...) __extension__ ({			\
+	errno = -(ret);						\
+	printf(msg ": %s\n", ##__VA_ARGS__, strerror(errno));	\
+})
+#endif /* PROGRAM_SPACE */
+
+
+#define lgdebugmsg(msg, ...) \
+(LOG_LEVEL >= VERBOSE_DEBUG ? printf(msg "\n", ##__VA_ARGS__) : 0)
+
+#define lgtracemsg(msg, ...) \
+(LOG_LEVEL >= VERBOSE_TRACE ? printf(msg "\n", ##__VA_ARGS__) : 0)
+
+#define lgtrace_start(msg, ...) \
+(LOG_LEVEL >= VERBOSE_TRACE ? printf("[TRACE] " msg " ( ", ##__VA_ARGS__) : 0)
+
+#define lgtrace_addp(msg, ...) \
+(LOG_LEVEL >= VERBOSE_TRACE ? printf(msg", ", ##__VA_ARGS__) : 0)
+
+#define lgtrace_end() \
+(LOG_LEVEL >= VERBOSE_TRACE ? printf(") \n") : 0)
+
+#endif /* LOGGING_H */

mangle.c

@@ -2,22 +2,50 @@
 #include "types.h" // IWYU pragma: keep
 #include "mangle.h"
 #include "config.h"
+#include "utils.h"
+#include "quic.h"
+#include "logging.h"
 
-#ifdef KERNEL_SPACE
+#ifndef KERNEL_SCOPE
-#include <linux/ip.h>
-
-#else 
 #include <stdlib.h>
-#include <libnetfilter_queue/libnetfilter_queue_ipv4.h>
-#include <libnetfilter_queue/libnetfilter_queue_tcp.h>
 #endif
 
-
 int process_packet(const uint8_t *raw_payload, uint32_t raw_payload_len) {
 	if (raw_payload_len > MAX_PACKET_SIZE) {
 		return PKT_ACCEPT;
 	}
 
+	const struct iphdr *iph;
+	uint32_t iph_len;
+	const uint8_t *ip_payload;
+	uint32_t ip_payload_len;
+
+	int ret;
+
+	ret = ip4_payload_split((uint8_t *)raw_payload, raw_payload_len,
+			 (struct iphdr **)&iph, &iph_len, 
+			 (uint8_t **)&ip_payload, &ip_payload_len);
+
+
+	if (ret < 0) 
+		goto accept;
+	
+	switch (iph->protocol) {
+	case IPPROTO_TCP:
+		return process_tcp4_packet(raw_payload, raw_payload_len);
+	case IPPROTO_UDP:
+		return process_udp4_packet(raw_payload, raw_payload_len);
+	default:
+		goto accept;
+	}
+	
+accept:
+	return PKT_ACCEPT;
+drop:
+	return PKT_DROP;
+}
+
+int process_tcp4_packet(const uint8_t *raw_payload, uint32_t raw_payload_len) {
 	const struct iphdr *iph;
 	uint32_t iph_len;
 	const struct tcphdr *tcph;
@@ -36,8 +64,7 @@ int process_packet(const uint8_t *raw_payload, uint32_t raw_payload_len) {
 	struct tls_verdict vrd = analyze_tls_data(data, dlen);
 
 	if (vrd.target_sni) {
-		if (config.verbose)
+		lgdebugmsg("Target SNI detected: %.*s", vrd.sni_len, data + vrd.sni_offset);
-			printf("Target SNI detected: %.*s\n", vrd.sni_len, data + vrd.sni_offset);
 
 		uint8_t payload[MAX_PACKET_SIZE];
 		uint32_t payload_len = raw_payload_len;
@@ -63,7 +90,7 @@ int process_packet(const uint8_t *raw_payload, uint32_t raw_payload_len) {
 
 		
 		if (dlen > 1480 && config.verbose) {
-			printf("WARNING! Client Hello packet is too big and may cause issues!\n");
+			lgdebugmsg("WARNING! Client Hello packet is too big and may cause issues!");
 		}
 
 		if (config.fake_sni) {
@@ -126,6 +153,79 @@ drop:
 	return PKT_DROP;
 }
 
+int process_udp4_packet(const uint8_t *pkt, uint32_t pktlen) {
+	const struct iphdr *iph;
+	uint32_t iph_len;
+	const struct udphdr *udph;
+	const uint8_t *data;
+	uint32_t dlen;
+
+	int ret = udp4_payload_split((uint8_t *)pkt, pktlen,
+			      (struct iphdr **)&iph, &iph_len, 
+			      (struct udphdr **)&udph,
+			      (uint8_t **)&data, &dlen);
+
+	lgtrace_start("Got udp packet");
+
+	if (ret < 0) {
+		lgtrace_addp("undefined");
+		goto accept;
+	}
+
+	if (dlen > 10 && config.verbose >= VERBOSE_TRACE) {
+		printf("UDP payload start: [ ");
+		for (int i = 0; i < 10; i++) {
+			printf("%02x ", data[i]);
+		}
+		printf("], ");
+	}
+
+	lgtrace_addp("QUIC probe");
+	const struct quic_lhdr *qch;
+	uint32_t qch_len;
+	struct quic_cids qci;
+	uint8_t *quic_raw_payload;
+	uint32_t quic_raw_plen;
+	ret = quic_parse_data((uint8_t *)data, dlen, 
+		 (struct quic_lhdr **)&qch, &qch_len, &qci, 
+		 &quic_raw_payload, &quic_raw_plen);
+
+	if (ret < 0) {
+		lgtrace_addp("undefined type");
+		goto accept;
+	}
+
+	lgtrace_addp("QUIC detected");
+	uint8_t qtype = qch->type;
+
+	if (config.quic_drop) {
+		goto drop;
+	}
+
+	if (qch->version == QUIC_V1)
+		qtype = quic_convtype_v1(qtype);
+	else if (qch->version == QUIC_V2) 
+		qtype = quic_convtype_v2(qtype);
+
+	if (qtype != QUIC_INITIAL_TYPE) {
+		lgtrace_addp("quic message type: %d", qtype);
+		goto accept;
+	}
+	
+	lgtrace_addp("quic initial message");
+
+accept:
+	lgtrace_addp("accepted");
+	lgtrace_end();
+
+	return PKT_ACCEPT;
+drop:
+	lgtrace_addp("dropped");
+	lgtrace_end();
+
+	return PKT_DROP;
+}
+
 int send_ip4_frags(const uint8_t *packet, uint32_t pktlen, const uint32_t *poses, uint32_t poses_sz, uint32_t dvs) {
 	if (poses_sz == 0) {
 		if (config.seg2_delay && ((dvs > 0) ^ config.frag_sni_reverse)) {
@@ -150,7 +250,7 @@ int send_ip4_frags(const uint8_t *packet, uint32_t pktlen, const uint32_t *poses
 		int ret;
 
 		if (dvs > poses[0]) {
-			printf("send_frags: Recursive dvs(%d) is more than poses0(%d)\n", dvs, poses[0]);
+			lgerror("send_frags: Recursive dvs(%d) is more than poses0(%d)", -EINVAL, dvs, poses[0]);
 			return -EINVAL;
 		}
 
@@ -158,8 +258,7 @@ int send_ip4_frags(const uint8_t *packet, uint32_t pktlen, const uint32_t *poses
 			frag1, &f1len, frag2, &f2len);
 
 		if (ret < 0) {
-			lgerror("send_frags: frag", ret);
+			lgerror("send_frags: frag: with context packet with size %d, position: %d, recursive dvs: %d", ret, pktlen, poses[0], dvs);
-			printf("Error context: packet with size %d, position: %d, recursive dvs: %d\n", pktlen, poses[0], dvs);
 			return ret;
 		}
 
@@ -214,7 +313,7 @@ int send_tcp4_frags(const uint8_t *packet, uint32_t pktlen, const uint32_t *pose
 		int ret;
 
 		if (dvs > poses[0]) {
-			printf("send_frags: Recursive dvs(%d) is more than poses0(%d)\n", dvs, poses[0]);
+			lgerror("send_frags: Recursive dvs(%d) is more than poses0(%d)", -EINVAL, dvs, poses[0]);
 			return -EINVAL;
 		}
 
@@ -222,8 +321,7 @@ int send_tcp4_frags(const uint8_t *packet, uint32_t pktlen, const uint32_t *pose
 			frag1, &f1len, frag2, &f2len);
 
 		if (ret < 0) {
-			lgerror("send_frags: frag", ret);
+			lgerror("send_frags: frag: with context packet with size %d, position: %d, recursive dvs: %d", ret, pktlen, poses[0], dvs);
-			printf("Error context: packet with size %d, position: %d, recursive dvs: %d\n", pktlen, poses[0], dvs);
 			return ret;
 		}
 
@@ -330,269 +428,7 @@ int post_fake_sni(const struct iphdr *iph, unsigned int iph_len,
 	return 0;
 }
 
-void tcp4_set_checksum(struct tcphdr *tcph, struct iphdr *iph) 
-{
-#ifdef KERNEL_SPACE
-	uint32_t tcp_packet_len = ntohs(iph->tot_len) - (iph->ihl << 2);
-	tcph->check = 0;
-	tcph->check = csum_tcpudp_magic(
-		iph->saddr, iph->daddr, tcp_packet_len,
-		IPPROTO_TCP, 
-		csum_partial(tcph, tcp_packet_len, 0));
-#else
-	nfq_tcp_compute_checksum_ipv4(tcph, iph);
-#endif
-}
 
-void ip4_set_checksum(struct iphdr *iph) 
-{
-#ifdef KERNEL_SPACE
-	iph->check = 0;
-	iph->check = ip_fast_csum(iph, iph->ihl);
-#else
-	nfq_ip_set_checksum(iph);
-#endif
-}
-
-
-int ip4_payload_split(uint8_t *pkt, uint32_t buflen,
-		       struct iphdr **iph, uint32_t *iph_len, 
-		       uint8_t **payload, uint32_t *plen) {
-	if (pkt == NULL || buflen < sizeof(struct iphdr)) {
-		lgerror("ip4_payload_split: pkt|buflen", -EINVAL);
-		return -EINVAL;
-	}
-
-	struct iphdr *hdr = (struct iphdr *)pkt;
-	if (hdr->version != IPVERSION) {
-		lgerror("ip4_payload_split: ipversion", -EINVAL);
-		return -EINVAL;
-	}
-
-	uint32_t hdr_len = hdr->ihl * 4;
-	uint32_t pktlen = ntohs(hdr->tot_len);
-	if (buflen < pktlen || hdr_len > pktlen) {
-		lgerror("ip4_payload_split: buflen cmp pktlen", -EINVAL);
-		return -EINVAL;
-	}
-
-	if (iph) 
-		*iph = hdr;
-	if (iph_len)
-		*iph_len = hdr_len;
-	if (payload)
-		*payload = pkt + hdr_len;
-	if (plen)
-		*plen = pktlen - hdr_len;
-
-	return 0;
-}
-
-int tcp4_payload_split(uint8_t *pkt, uint32_t buflen,
-		       struct iphdr **iph, uint32_t *iph_len,
-		       struct tcphdr **tcph, uint32_t *tcph_len,
-		       uint8_t **payload, uint32_t *plen) {
-	struct iphdr *hdr;
-	uint32_t hdr_len;
-	struct tcphdr *thdr;
-	uint32_t thdr_len;
-	
-	uint8_t *tcph_pl;
-	uint32_t tcph_plen;
-
-	if (ip4_payload_split(pkt, buflen, &hdr, &hdr_len, 
-			&tcph_pl, &tcph_plen)){
-		return -EINVAL;
-	}
-
-
-	if (
-		hdr->protocol != IPPROTO_TCP || 
-		tcph_plen < sizeof(struct tcphdr)) {
-		return -EINVAL;
-	}
-
-
-	thdr = (struct tcphdr *)(tcph_pl);
-	thdr_len = thdr->doff * 4;
-
-	if (thdr_len > tcph_plen) {
-		return -EINVAL;
-	}
-
-	if (iph) *iph = hdr;
-	if (iph_len) *iph_len = hdr_len;
-	if (tcph) *tcph = thdr;
-	if (tcph_len) *tcph_len = thdr_len;
-	if (payload) *payload = tcph_pl + thdr_len;
-	if (plen) *plen = tcph_plen - thdr_len;
-
-	return 0;
-}
-
-// split packet to two ipv4 fragments.
-int ip4_frag(const uint8_t *pkt, uint32_t buflen, uint32_t payload_offset, 
-			uint8_t *frag1, uint32_t *f1len, 
-			uint8_t *frag2, uint32_t *f2len) {
-	
-	struct iphdr *hdr;
-	const uint8_t *payload;
-	uint32_t plen;
-	uint32_t hdr_len;
-	int ret;
-
-	if (!frag1 || !f1len || !frag2 || !f2len)
-		return -EINVAL;
-
-	if ((ret = ip4_payload_split(
-		(uint8_t *)pkt, buflen, 
-		&hdr, &hdr_len, (uint8_t **)&payload, &plen)) < 0) {
-		lgerror("ipv4_frag: TCP Header extract error", ret);
-		return -EINVAL;
-	}
-
-	if (plen <= payload_offset) {
-		return -EINVAL;
-	}
-
-	if (payload_offset & ((1 << 3) - 1)) {
-		lgerror("ipv4_frag: Payload offset MUST be a multiply of 8!", -EINVAL);
-
-		return -EINVAL;
-	}
-
-	uint32_t f1_plen = payload_offset;
-	uint32_t f1_dlen = f1_plen + hdr_len;
-
-	uint32_t f2_plen = plen - payload_offset;
-	uint32_t f2_dlen = f2_plen + hdr_len;
-
-	if (*f1len < f1_dlen || *f2len < f2_dlen) {
-		return -ENOMEM;
-	}
-	*f1len = f1_dlen;
-	*f2len = f2_dlen;
-
-	memcpy(frag1, hdr, hdr_len);
-	memcpy(frag2, hdr, hdr_len);
-
-	memcpy(frag1 + hdr_len, payload, f1_plen);
-	memcpy(frag2 + hdr_len, payload + payload_offset, f2_plen);
-
-	struct iphdr *f1_hdr = (void *)frag1;
-	struct iphdr *f2_hdr = (void *)frag2;
-
-	uint16_t f1_frag_off = ntohs(f1_hdr->frag_off);
-	uint16_t f2_frag_off = ntohs(f2_hdr->frag_off);
-
-	f1_frag_off &= IP_OFFMASK;
-	f1_frag_off |= IP_MF;
-	
-	if ((f2_frag_off & ~IP_OFFMASK) == IP_MF) {
-		f2_frag_off &= IP_OFFMASK;
-		f2_frag_off |= IP_MF;
-	} else {
-		f2_frag_off &= IP_OFFMASK;
-	}
-	
-	f2_frag_off += (uint16_t)payload_offset / 8;
-
-	f1_hdr->frag_off = htons(f1_frag_off);
-	f1_hdr->tot_len = htons(f1_dlen);
-
-	f2_hdr->frag_off = htons(f2_frag_off);
-	f2_hdr->tot_len = htons(f2_dlen);
-
-
-	if (config.verbose)
-		printf("Packet split in portion %u %u\n", f1_plen, f2_plen);
-
-	ip4_set_checksum(f1_hdr);
-	ip4_set_checksum(f2_hdr);
-
-	return 0;
-}
-
-// split packet to two tcp-on-ipv4 segments.
-int tcp4_frag(const uint8_t *pkt, uint32_t buflen, uint32_t payload_offset, 
-			uint8_t *seg1, uint32_t *s1len, 
-			uint8_t *seg2, uint32_t *s2len) {
-
-	struct iphdr *hdr;
-	uint32_t hdr_len;
-	struct tcphdr *tcph;
-	uint32_t tcph_len;
-	uint32_t plen;
-	const uint8_t *payload;
-	int ret;
-
-	if (!seg1 || !s1len || !seg2 || !s2len)
-		return -EINVAL;
-
-	if ((ret = tcp4_payload_split((uint8_t *)pkt, buflen,
-				&hdr, &hdr_len,
-				&tcph, &tcph_len,
-				(uint8_t **)&payload, &plen)) < 0) {
-		lgerror("tcp4_frag: tcp4_payload_split", ret);
-
-		return -EINVAL;
-	}
-
-
-	if (
-		ntohs(hdr->frag_off) & IP_MF || 
-		ntohs(hdr->frag_off) & IP_OFFMASK) {
-		printf("tcp4_frag: frag value: %d\n",
-			ntohs(hdr->frag_off));
-		lgerror("tcp4_frag: ip fragmentation is set", -EINVAL);
-		return -EINVAL;
-	}
-
-
-	if (plen <= payload_offset) {
-		return -EINVAL;
-	}
-
-	uint32_t s1_plen = payload_offset;
-	uint32_t s1_dlen = s1_plen + hdr_len + tcph_len;
-
-	uint32_t s2_plen = plen - payload_offset;
-	uint32_t s2_dlen = s2_plen + hdr_len + tcph_len;
-
-	if (*s1len < s1_dlen || *s2len < s2_dlen) 
-		return -ENOMEM;
-
-	*s1len = s1_dlen;
-	*s2len = s2_dlen;
-
-	memcpy(seg1, hdr, hdr_len);
-	memcpy(seg2, hdr, hdr_len);
-
-	memcpy(seg1 + hdr_len, tcph, tcph_len);
-	memcpy(seg2 + hdr_len, tcph, tcph_len);
-
-	memcpy(seg1 + hdr_len + tcph_len, payload, s1_plen);
-	memcpy(seg2 + hdr_len + tcph_len, payload + payload_offset, s2_plen);
-
-	struct iphdr *s1_hdr = (void *)seg1;
-	struct iphdr *s2_hdr = (void *)seg2;
-
-	struct tcphdr *s1_tcph = (void *)(seg1 + hdr_len);
-	struct tcphdr *s2_tcph = (void *)(seg2 + hdr_len);
-
-	s1_hdr->tot_len = htons(s1_dlen);
-	s2_hdr->tot_len = htons(s2_dlen);
-
-	s2_tcph->seq = htonl(ntohl(s2_tcph->seq) + payload_offset);
-
-	if (config.verbose)
-		printf("Packet split in portion %u %u\n", s1_plen, s2_plen);
-
-	tcp4_set_checksum(s1_tcph, s1_hdr);
-	tcp4_set_checksum(s2_tcph, s2_hdr);
-
-	return 0;
-}
 
 #define TLS_CONTENT_TYPE_HANDSHAKE 0x16
 #define TLS_HANDSHAKE_TYPE_CLIENT_HELLO 0x01
@@ -807,8 +643,13 @@ int fail4_packet(uint8_t *payload, uint32_t plen) {
 	}
 
 	if (config.faking_strategy == FAKE_STRAT_ACK_SEQ) {
+#ifdef KERNEL_SCOPE
+		tcph->seq = 124;
+		tcph->ack_seq = 124;
+#else
 		tcph->seq = random();
 		tcph->ack_seq = random();
+#endif
 	} else if (config.faking_strategy == FAKE_STRAT_TTL) {
 		iph->ttl = config.faking_ttl;
 	}

mangle.h

@@ -3,30 +3,6 @@
 
 #include "types.h"
 
-#ifdef KERNEL_SPACE
-#include <linux/string.h>
-#include <linux/stddef.h>
-#include <linux/net.h>
-#include <linux/in.h>
-#include <linux/ip.h>
-#include <linux/tcp.h>
-
-#include <asm/byteorder.h>
-
-/* from <netinet/ip.h> */
-#define	IP_RF 0x8000			/* reserved fragment flag */
-#define	IP_DF 0x4000			/* dont fragment flag */
-#define	IP_MF 0x2000			/* more fragments flag */
-#define	IP_OFFMASK 0x1fff		/* mask for fragmenting bits */
-#else
-#define USER_SPACE
-#include <stdint.h>
-#include <string.h>
-#include <arpa/inet.h>
-#include <netinet/ip.h>
-#include <netinet/tcp.h>
-#endif
-
 /**
  * Result of analyze_tls_data function
  */
@@ -42,42 +18,6 @@ struct tls_verdict {
  */
 struct tls_verdict analyze_tls_data(const uint8_t *data, uint32_t dlen);
 
-/**
- * Splits the packet to two IP fragments on position payload_offset.
- * payload_offset indicates the position relatively to start of IP payload
- * (start of transport header)
- */
-int ip4_frag(const uint8_t *pkt, uint32_t pktlen, 
-			uint32_t payload_offset, 
-			uint8_t *frag1, uint32_t *f1len, 
-			uint8_t *frag2, uint32_t *f2len);
-
-/**
- * Splits the packet to two TCP segments on position payload_offset
- * payload_offset indicates the position relatively to start of TCP payload.
- */
-int tcp4_frag(const uint8_t *pkt, uint32_t pktlen, 
-			uint32_t payload_offset, 
-			uint8_t *seg1, uint32_t *s1len, 
-			uint8_t *seg2, uint32_t *s2len);
-
-/**
- * Splits the raw packet payload to ip header and ip payload.
- */
-int ip4_payload_split(uint8_t *pkt, uint32_t buflen,
-		       struct iphdr **iph, uint32_t *iph_len, 
-		       uint8_t **payload, uint32_t *plen);
-
-/**
- * Splits the raw packet payload to ip header, tcp header and tcp payload.
- */
-int tcp4_payload_split(uint8_t *pkt, uint32_t buflen,
-		       struct iphdr **iph, uint32_t *iph_len,
-		       struct tcphdr **tcph, uint32_t *tcph_len,
-		       uint8_t **payload, uint32_t *plen);
-
-void tcp4_set_checksum(struct tcphdr *tcph, struct iphdr *iph);
-void ip4_set_checksum(struct iphdr *iph);
 
 /**
  * Generates fake client hello message
@@ -100,6 +40,20 @@ int fail4_packet(uint8_t *payload, uint32_t plen);
  */
 int process_packet(const uint8_t *packet, uint32_t packet_len);
 
+
+/**
+ * Processe the TCP packet.
+ * Returns verdict.
+ */
+int process_tcp4_packet(const uint8_t *raw_payload, uint32_t raw_payload_len);
+
+
+/**
+ * Processes the UDP packet.
+ * Returns verdict.
+ */
+int process_udp4_packet(const uint8_t *pkt, uint32_t pktlen);
+
 /**
  * Sends fake client hello.
  */

quic.c

@@ -1,9 +1,6 @@
 #include "quic.h"
+#include "logging.h"
 
-static const uint32_t supported_versions[] = {
-	1,		// version 1, RFC 9000
-	0x6b3343cf,	// version 2, RFC 9369
-};
 
 /**
  * Packet number.
@@ -45,18 +42,24 @@ int quic_parse_data(uint8_t *raw_payload, uint32_t raw_payload_len,
 	struct quic_lhdr *nqch = (struct quic_lhdr *)raw_payload;
 	uint32_t left_len = raw_payload_len - sizeof(struct quic_lhdr);
 	uint8_t *cur_rawptr = raw_payload + sizeof(struct quic_lhdr);
-	if (!nqch->fixed)
+	if (!nqch->fixed) {
+		lgtrace_addp("quic fixed uset");
 		return -EPROTO;
+	}
 
 	uint8_t found = 0;
 	for (uint8_t i = 0; i < sizeof(supported_versions); i++) {
-		if (nqch->version == supported_versions[i]) {
+		if (ntohl(nqch->version) == supported_versions[i]) {
 			found = 1;
 		}
 	}
 
-	if (!found) 
+	if (!found) {
+		lgtrace_addp("quic version undefined %d", ntohl(nqch->version));
 		return -EPROTO;
+	}
+
+	lgtrace_addp("quic version valid %d", ntohl(nqch->version));
 
 	if (left_len < 2) goto invalid_packet;
 	struct quic_cids nqci = {0};

quic.h

@@ -2,6 +2,7 @@
 #define QUIC_H
 #include "types.h"
 
+
 /**
 * @macro
 *
@@ -22,15 +23,30 @@
  "\x0d\xed\xe3\xde\xf7\x00\xa6\xdb\x81\x93\x81\xbe\x6e\x26\x9d\xcb" \
  "\xf9\xbd\x2e\xd9"
 
-#define QUIC_INITIAL_TYPE_V1 0x00
+#define QUIC_INITIAL_TYPE	0
-#define QUIC_0_RTT_TYPE_V1 0x01
+#define QUIC_0_RTT_TYPE		1
-#define QUIC_HANDSHAKE_TYPE_V1 0x02
+#define QUIC_HANDSHAKE_TYPE	2
-#define QUIC_RETRY_TYPE_V1 0x03
+#define QUIC_RETRY_TYPE		3
+
+#define QUIC_INITIAL_TYPE_V1	0b00
+#define QUIC_0_RTT_TYPE_V1	0b01
+#define QUIC_HANDSHAKE_TYPE_V1	0b10
+#define QUIC_RETRY_TYPE_V1	0b11
+#define quic_convtype_v1(type) (type)
 
 #define QUIC_INITIAL_TYPE_V2	0b01
 #define QUIC_0_RTT_TYPE_V2	0b10
 #define QUIC_HANDSHAKE_TYPE_V2	0b11
 #define QUIC_RETRY_TYPE_V2	0b00
+#define quic_convtype_v2(type) (((type) + 1) & 0b11)
+
+#define QUIC_V1	1		// RFC 9000
+#define QUIC_V2	0x6b3343cf	// RFC 9369
+
+static const uint32_t supported_versions[] = {
+	QUIC_V1,
+	QUIC_V2,
+};
 
 /**
  * Quic Large Header
@@ -51,9 +67,8 @@ struct quic_lhdr {
 #else
 #error "Undefined endian"
 #endif
-
 	uint32_t version;
-};
+}__attribute__((packed));
 
 /**
  * Quic Large Header Ids 

types.h

@@ -15,22 +15,33 @@ typedef __i8	int8_t;
 typedef __i16	int16_t;
 typedef __i32	int32_t;
 typedef __i64	int64_t;
-
-#include <linux/printk.h>
-#define printf pr_info
-#define perror pr_err
-#define lgerror(msg, ret) (pr_err(msg ": %d\n", ret))
-
 #else /* USERSPACE_SCOPE */
 
 #include <errno.h>  // IWYU pragma: export
 #include <stdint.h> // IWYU pragma: export
 #include <string.h> // IWYU pragma: export
 
-#include <stdio.h> // IWYU pragma: export
-#define lgerror(msg, ret) __extension__ ({errno = -ret; perror(msg);})
-
-
 #endif /* SCOPES */
 
+// Network specific structures
+#ifdef KERNEL_SPACE
+#include <linux/stddef.h> // IWYU pragma: export
+#include <linux/net.h> // IWYU pragma: export
+#include <linux/in.h> // IWYU pragma: export
+#include <linux/ip.h> // IWYU pragma: export
+#include <linux/tcp.h> // IWYU pragma: export
+
+/* from <netinet/ip.h> */
+#define	IP_RF 0x8000			/* reserved fragment flag */
+#define	IP_DF 0x4000			/* dont fragment flag */
+#define	IP_MF 0x2000			/* more fragments flag */
+#define	IP_OFFMASK 0x1fff		/* mask for fragmenting bits */
+#else
+#define USER_SPACE
+#include <arpa/inet.h>		// IWYU pragma: export
+#include <netinet/ip.h>		// IWYU pragma: export
+#include <netinet/tcp.h>	// IWYU pragma: export
+#include <netinet/udp.h>	// IWYU pragma: export
+#endif
+
 #endif /* TYPES_H */

uspace.mk

@@ -22,7 +22,7 @@ export CC CCLD LD CFLAGS LDFLAGS LIBNFNETLINK_CFLAGS LIBNFNETLINK_LIBS LIBMNL_CF
 
 APP:=$(BUILD_DIR)/youtubeUnblock
 
-SRCS := youtubeUnblock.c mangle.c args.c
+SRCS := youtubeUnblock.c mangle.c args.c utils.c quic.c
 OBJS := $(SRCS:%.c=$(BUILD_DIR)/%.o)
 
 LIBNFNETLINK := $(DEPSDIR)/lib/libnfnetlink.a

utils.c

@@ -0,0 +1,312 @@
+#include "utils.h"
+#include "logging.h"
+
+#ifdef KERNEL_SPACE
+#include <linux/ip.h>
+#else 
+#include <stdlib.h>
+#include <libnetfilter_queue/libnetfilter_queue_ipv4.h>
+#include <libnetfilter_queue/libnetfilter_queue_tcp.h>
+#endif
+
+
+void tcp4_set_checksum(struct tcphdr *tcph, struct iphdr *iph) 
+{
+#ifdef KERNEL_SPACE
+	uint32_t tcp_packet_len = ntohs(iph->tot_len) - (iph->ihl << 2);
+	tcph->check = 0;
+	tcph->check = csum_tcpudp_magic(
+		iph->saddr, iph->daddr, tcp_packet_len,
+		IPPROTO_TCP, 
+		csum_partial(tcph, tcp_packet_len, 0));
+#else
+	nfq_tcp_compute_checksum_ipv4(tcph, iph);
+#endif
+}
+
+void ip4_set_checksum(struct iphdr *iph) 
+{
+#ifdef KERNEL_SPACE
+	iph->check = 0;
+	iph->check = ip_fast_csum(iph, iph->ihl);
+#else
+	nfq_ip_set_checksum(iph);
+#endif
+}
+
+
+int ip4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len, 
+		       uint8_t **payload, uint32_t *plen) {
+	if (pkt == NULL || buflen < sizeof(struct iphdr)) {
+		lgerror("ip4_payload_split: pkt|buflen", -EINVAL);
+		return -EINVAL;
+	}
+
+	struct iphdr *hdr = (struct iphdr *)pkt;
+	if (hdr->version != IPVERSION) {
+		lgerror("ip4_payload_split: ipversion", -EINVAL);
+		return -EINVAL;
+	}
+
+	uint32_t hdr_len = hdr->ihl * 4;
+	uint32_t pktlen = ntohs(hdr->tot_len);
+	if (buflen < pktlen || hdr_len > pktlen) {
+		lgerror("ip4_payload_split: buflen cmp pktlen", -EINVAL);
+		return -EINVAL;
+	}
+
+	if (iph) 
+		*iph = hdr;
+	if (iph_len)
+		*iph_len = hdr_len;
+	if (payload)
+		*payload = pkt + hdr_len;
+	if (plen)
+		*plen = pktlen - hdr_len;
+
+	return 0;
+}
+
+int tcp4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len,
+		       struct tcphdr **tcph, uint32_t *tcph_len,
+		       uint8_t **payload, uint32_t *plen) {
+	struct iphdr *hdr;
+	uint32_t hdr_len;
+	struct tcphdr *thdr;
+	uint32_t thdr_len;
+	
+	uint8_t *tcph_pl;
+	uint32_t tcph_plen;
+
+	if (ip4_payload_split(pkt, buflen, &hdr, &hdr_len, 
+			&tcph_pl, &tcph_plen)){
+		return -EINVAL;
+	}
+
+
+	if (
+		hdr->protocol != IPPROTO_TCP || 
+		tcph_plen < sizeof(struct tcphdr)) {
+		return -EINVAL;
+	}
+
+
+	thdr = (struct tcphdr *)(tcph_pl);
+	thdr_len = thdr->doff * 4;
+
+	if (thdr_len > tcph_plen) {
+		return -EINVAL;
+	}
+
+	if (iph) *iph = hdr;
+	if (iph_len) *iph_len = hdr_len;
+	if (tcph) *tcph = thdr;
+	if (tcph_len) *tcph_len = thdr_len;
+	if (payload) *payload = tcph_pl + thdr_len;
+	if (plen) *plen = tcph_plen - thdr_len;
+
+	return 0;
+}
+
+int udp4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len,
+		       struct udphdr **udph,
+		       uint8_t **payload, uint32_t *plen) {
+	struct iphdr *hdr;
+	uint32_t hdr_len;
+	struct udphdr *uhdr;
+	uint32_t uhdr_len;
+	
+	uint8_t *ip_ph;
+	uint32_t ip_phlen;
+
+	if (ip4_payload_split(pkt, buflen, &hdr, &hdr_len, 
+			&ip_ph, &ip_phlen)){
+		return -EINVAL;
+	}
+
+
+	if (
+		hdr->protocol != IPPROTO_UDP || 
+		ip_phlen < sizeof(struct udphdr)) {
+		return -EINVAL;
+	}
+
+
+	uhdr = (struct udphdr *)(ip_ph);
+	if (uhdr->len != 0 && ntohs(uhdr->len) != ip_phlen) {
+		return -EINVAL;
+	}
+
+	if (iph) *iph = hdr;
+	if (iph_len) *iph_len = hdr_len;
+	if (udph) *udph = uhdr;
+	if (payload) *payload = ip_ph + sizeof(struct udphdr);
+	if (plen) *plen = ip_phlen - sizeof(struct udphdr);
+
+	return 0;
+}
+
+// split packet to two ipv4 fragments.
+int ip4_frag(const uint8_t *pkt, uint32_t buflen, uint32_t payload_offset, 
+			uint8_t *frag1, uint32_t *f1len, 
+			uint8_t *frag2, uint32_t *f2len) {
+	
+	struct iphdr *hdr;
+	const uint8_t *payload;
+	uint32_t plen;
+	uint32_t hdr_len;
+	int ret;
+
+	if (!frag1 || !f1len || !frag2 || !f2len)
+		return -EINVAL;
+
+	if ((ret = ip4_payload_split(
+		(uint8_t *)pkt, buflen, 
+		&hdr, &hdr_len, (uint8_t **)&payload, &plen)) < 0) {
+		lgerror("ipv4_frag: TCP Header extract error", ret);
+		return -EINVAL;
+	}
+
+	if (plen <= payload_offset) {
+		return -EINVAL;
+	}
+
+	if (payload_offset & ((1 << 3) - 1)) {
+		lgerror("ipv4_frag: Payload offset MUST be a multiply of 8!", -EINVAL);
+
+		return -EINVAL;
+	}
+
+	uint32_t f1_plen = payload_offset;
+	uint32_t f1_dlen = f1_plen + hdr_len;
+
+	uint32_t f2_plen = plen - payload_offset;
+	uint32_t f2_dlen = f2_plen + hdr_len;
+
+	if (*f1len < f1_dlen || *f2len < f2_dlen) {
+		return -ENOMEM;
+	}
+	*f1len = f1_dlen;
+	*f2len = f2_dlen;
+
+	memcpy(frag1, hdr, hdr_len);
+	memcpy(frag2, hdr, hdr_len);
+
+	memcpy(frag1 + hdr_len, payload, f1_plen);
+	memcpy(frag2 + hdr_len, payload + payload_offset, f2_plen);
+
+	struct iphdr *f1_hdr = (void *)frag1;
+	struct iphdr *f2_hdr = (void *)frag2;
+
+	uint16_t f1_frag_off = ntohs(f1_hdr->frag_off);
+	uint16_t f2_frag_off = ntohs(f2_hdr->frag_off);
+
+	f1_frag_off &= IP_OFFMASK;
+	f1_frag_off |= IP_MF;
+	
+	if ((f2_frag_off & ~IP_OFFMASK) == IP_MF) {
+		f2_frag_off &= IP_OFFMASK;
+		f2_frag_off |= IP_MF;
+	} else {
+		f2_frag_off &= IP_OFFMASK;
+	}
+	
+	f2_frag_off += (uint16_t)payload_offset / 8;
+
+	f1_hdr->frag_off = htons(f1_frag_off);
+	f1_hdr->tot_len = htons(f1_dlen);
+
+	f2_hdr->frag_off = htons(f2_frag_off);
+	f2_hdr->tot_len = htons(f2_dlen);
+
+
+	lgdebugmsg("Packet split in portion %u %u", f1_plen, f2_plen);
+
+	ip4_set_checksum(f1_hdr);
+	ip4_set_checksum(f2_hdr);
+
+	return 0;
+}
+
+// split packet to two tcp-on-ipv4 segments.
+int tcp4_frag(const uint8_t *pkt, uint32_t buflen, uint32_t payload_offset, 
+			uint8_t *seg1, uint32_t *s1len, 
+			uint8_t *seg2, uint32_t *s2len) {
+
+	struct iphdr *hdr;
+	uint32_t hdr_len;
+	struct tcphdr *tcph;
+	uint32_t tcph_len;
+	uint32_t plen;
+	const uint8_t *payload;
+	int ret;
+
+	if (!seg1 || !s1len || !seg2 || !s2len)
+		return -EINVAL;
+
+	if ((ret = tcp4_payload_split((uint8_t *)pkt, buflen,
+				&hdr, &hdr_len,
+				&tcph, &tcph_len,
+				(uint8_t **)&payload, &plen)) < 0) {
+		lgerror("tcp4_frag: tcp4_payload_split", ret);
+
+		return -EINVAL;
+	}
+
+
+	if (
+		ntohs(hdr->frag_off) & IP_MF || 
+		ntohs(hdr->frag_off) & IP_OFFMASK) {
+		lgdebugmsg("tcp4_frag: frag value: %d",
+			ntohs(hdr->frag_off));
+		lgerror("tcp4_frag: ip fragmentation is set", -EINVAL);
+		return -EINVAL;
+	}
+
+
+	if (plen <= payload_offset) {
+		return -EINVAL;
+	}
+
+	uint32_t s1_plen = payload_offset;
+	uint32_t s1_dlen = s1_plen + hdr_len + tcph_len;
+
+	uint32_t s2_plen = plen - payload_offset;
+	uint32_t s2_dlen = s2_plen + hdr_len + tcph_len;
+
+	if (*s1len < s1_dlen || *s2len < s2_dlen) 
+		return -ENOMEM;
+
+	*s1len = s1_dlen;
+	*s2len = s2_dlen;
+
+	memcpy(seg1, hdr, hdr_len);
+	memcpy(seg2, hdr, hdr_len);
+
+	memcpy(seg1 + hdr_len, tcph, tcph_len);
+	memcpy(seg2 + hdr_len, tcph, tcph_len);
+
+	memcpy(seg1 + hdr_len + tcph_len, payload, s1_plen);
+	memcpy(seg2 + hdr_len + tcph_len, payload + payload_offset, s2_plen);
+
+	struct iphdr *s1_hdr = (void *)seg1;
+	struct iphdr *s2_hdr = (void *)seg2;
+
+	struct tcphdr *s1_tcph = (void *)(seg1 + hdr_len);
+	struct tcphdr *s2_tcph = (void *)(seg2 + hdr_len);
+
+	s1_hdr->tot_len = htons(s1_dlen);
+	s2_hdr->tot_len = htons(s2_dlen);
+
+	s2_tcph->seq = htonl(ntohl(s2_tcph->seq) + payload_offset);
+
+	lgdebugmsg("Packet split in portion %u %u", s1_plen, s2_plen);
+
+	tcp4_set_checksum(s1_tcph, s1_hdr);
+	tcp4_set_checksum(s2_tcph, s2_hdr);
+
+	return 0;
+}

utils.h

@@ -0,0 +1,51 @@
+#ifndef UTILS_H
+#define UTILS_H
+
+#include "types.h"
+
+/**
+ * Splits the packet to two IP fragments on position payload_offset.
+ * payload_offset indicates the position relatively to start of IP payload
+ * (start of transport header)
+ */
+int ip4_frag(const uint8_t *pkt, uint32_t pktlen, 
+			uint32_t payload_offset, 
+			uint8_t *frag1, uint32_t *f1len, 
+			uint8_t *frag2, uint32_t *f2len);
+
+/**
+ * Splits the packet to two TCP segments on position payload_offset
+ * payload_offset indicates the position relatively to start of TCP payload.
+ */
+int tcp4_frag(const uint8_t *pkt, uint32_t pktlen, 
+			uint32_t payload_offset, 
+			uint8_t *seg1, uint32_t *s1len, 
+			uint8_t *seg2, uint32_t *s2len);
+
+/**
+ * Splits the raw packet payload to ip header and ip payload.
+ */
+int ip4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len, 
+		       uint8_t **payload, uint32_t *plen);
+
+/**
+ * Splits the raw packet payload to ip header, tcp header and tcp payload.
+ */
+int tcp4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len,
+		       struct tcphdr **tcph, uint32_t *tcph_len,
+		       uint8_t **payload, uint32_t *plen);
+
+/**
+ * Splits the raw packet payload to ip header, udp header and udp payload.
+ */
+int udp4_payload_split(uint8_t *pkt, uint32_t buflen,
+		       struct iphdr **iph, uint32_t *iph_len,
+		       struct udphdr **udph,
+		       uint8_t **payload, uint32_t *plen);
+
+void tcp4_set_checksum(struct tcphdr *tcph, struct iphdr *iph);
+void ip4_set_checksum(struct iphdr *iph);
+
+#endif /* UTILS_H */

youtubeUnblock.c

@@ -31,6 +31,7 @@
 #include "config.h"
 #include "mangle.h"
 #include "args.h"
+#include "utils.h"
 
 pthread_mutex_t rawsocket_lock;
 int rawsocket = -2;