libbpf-tools: add tcpdrop to trace TCP packet drops

libbpf-tools/Makefile

@@ -86,6 +86,7 @@ APPS = \
 	tcptracer \
 	tcpconnect \
 	tcpconnlat \
+	tcpdrop \
 	tcplife \
 	tcppktlat \
 	tcprtt \

libbpf-tools/tcpdrop.bpf.c

@@ -0,0 +1,186 @@
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include <bpf/bpf_tracing.h>
+#include <bpf/bpf_endian.h>
+
+#ifndef AF_INET
+#define AF_INET 2
+#endif
+#ifndef AF_INET6
+#define AF_INET6 10
+#endif
+#ifndef TASK_COMM_LEN
+#define TASK_COMM_LEN 16
+#endif
+#ifndef ETH_P_IP
+#define ETH_P_IP 0x0800
+#endif
+#ifndef ETH_P_IPV6
+#define ETH_P_IPV6 0x86dd
+#endif
+
+struct event {
+    u64 timestamp;
+    u32 pid;
+    u32 drop_reason;
+    u32 ip_version; // 4 for IPv4, 6 for IPv6
+    union {
+        u32 saddr_v4;
+        unsigned __int128 saddr_v6;
+    };
+    union {
+        u32 daddr_v4;
+        unsigned __int128 daddr_v6;
+    };
+    u16 sport;
+    u16 dport;
+    u8 state;
+    u8 tcpflags;
+    char comm[TASK_COMM_LEN];
+    u32 stack_id;
+};
+
+struct {
+    __uint(type, BPF_MAP_TYPE_RINGBUF);
+    __uint(max_entries, 512);
+} events SEC(".maps");
+
+#define MAX_STACK_DEPTH 15
+struct {
+    __uint(type, BPF_MAP_TYPE_STACK_TRACE);
+    __uint(max_entries, 512);
+    __uint(key_size, sizeof(u32));
+    __uint(value_size, MAX_STACK_DEPTH * sizeof(u64));
+} stack_traces SEC(".maps");
+
+char ipv4_only = 0;
+char ipv6_only = 0;
+__u32 netns_id = 0;
+
+SEC("tracepoint/skb/kfree_skb")
+int tp__skb_free_skb(struct trace_event_raw_kfree_skb *args)
+{
+    if (args->reason <= SKB_DROP_REASON_NOT_SPECIFIED) {
+        return 0;
+    }
+
+    if (bpf_ringbuf_query(&events, BPF_RB_AVAIL_DATA) >= 511) {
+        bpf_printk("Ring buffer is almost full\n");
+        return 0;
+    }
+
+    u64 pid_tgid = bpf_get_current_pid_tgid();
+    u32 pid = pid_tgid >> 32;
+
+    struct sk_buff *skb = args->skbaddr;
+    if (!skb) {
+        return 0;
+    }
+    struct sock *sk = NULL;
+    bpf_core_read(&sk, sizeof(sk), &skb->sk);
+
+    // Get packet headers
+    void *head;
+    u16 network_header, transport_header;
+    if (bpf_core_read(&head, sizeof(head), &skb->head) ||
+        bpf_core_read(&network_header, sizeof(network_header), &skb->network_header) ||
+        bpf_core_read(&transport_header, sizeof(transport_header), &skb->transport_header)) {
+        bpf_printk("Failed to read skb headers\n");
+        return 0;
+    }
+
+    // Check protocol and filter by IP family
+    u16 protocol = args->protocol;
+    if (protocol != ETH_P_IP && protocol != ETH_P_IPV6) {
+        bpf_printk("Unsupported protocol: %u\n", protocol);
+        return 0;
+    }
+    if (ipv4_only && protocol != ETH_P_IP) {
+        return 0;
+    }
+    if (ipv6_only && protocol != ETH_P_IPV6) {
+        return 0;
+    }
+
+    // Filter by network namespace
+    if (netns_id && sk) {
+        struct net *net = NULL;
+        bpf_core_read(&net, sizeof(net), &sk->__sk_common.skc_net.net);
+        if (net) {
+            u32 inum;
+            bpf_core_read(&inum, sizeof(inum), &net->ns.inum);
+            if (inum != netns_id) {
+                bpf_printk("Skipping packet from different netns: %u != %u\n", inum, netns_id);
+                return 0;
+            }
+        }
+    }
+
+    struct event *event = bpf_ringbuf_reserve(&events, sizeof(*event), 0);
+    if (!event) {
+        return 0;
+    }
+
+    event->timestamp = bpf_ktime_get_ns();
+    event->pid = pid;
+    event->drop_reason = args->reason;
+    bpf_get_current_comm(&event->comm, sizeof(event->comm));
+    event->stack_id = bpf_get_stackid(args, &stack_traces, 0);
+    event->state = 127;
+    if (sk) {
+        u8 state;
+        if (!bpf_core_read(&state, sizeof(state), &sk->__sk_common.skc_state)) {
+            event->state = state;
+        }
+    }
+
+    if (protocol == ETH_P_IP) {
+        struct iphdr ip;
+        if (bpf_core_read(&ip, sizeof(ip), head + network_header)) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        if (ip.protocol != IPPROTO_TCP) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        struct tcphdr tcp;
+        if (bpf_core_read(&tcp, sizeof(tcp), head + transport_header)) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        event->ip_version = 4;
+        event->saddr_v4 = ip.saddr;
+        event->daddr_v4 = ip.daddr;
+        event->sport = bpf_ntohs(tcp.source);
+        event->dport = bpf_ntohs(tcp.dest);
+        event->tcpflags = ((u8 *)&tcp)[13];
+    } else {
+        struct ipv6hdr ip6;
+        if (bpf_core_read(&ip6, sizeof(ip6), head + network_header)) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        if (ip6.nexthdr != IPPROTO_TCP) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        struct tcphdr tcp;
+        if (bpf_core_read(&tcp, sizeof(tcp), head + transport_header)) {
+            bpf_ringbuf_discard(event, 0);
+            return 0;
+        }
+        event->ip_version = 6;
+        bpf_core_read(&event->saddr_v6, sizeof(event->saddr_v6), &ip6.saddr);
+        bpf_core_read(&event->daddr_v6, sizeof(event->daddr_v6), &ip6.daddr);
+        event->sport = bpf_ntohs(tcp.source);
+        event->dport = bpf_ntohs(tcp.dest);
+        event->tcpflags = ((u8 *)&tcp)[13];
+    }
+
+    bpf_ringbuf_submit(event, 0);
+    return 0;
+}
+
+char _license[] SEC("license") = "Dual BSD/GPL";