vendor: bump tcptracer-bpf

This includes:

* an eBPF object built with a newer kernel (4.14)
* an increased threshold for guessing offsets, which is necessary for
the eBPF tracer to work on Google's Container-Optimized OS (used by
GKE)

vendor/github.com/weaveworks/tcptracer-bpf/pkg/tracer/offsetguess.go

@@ -28,7 +28,7 @@ type tcpTracerStatus C.struct_tcptracer_status_t
 const (
 	// When reading kernel structs at different offsets, don't go over that
 	// limit. This is an arbitrary choice to avoid infinite loops.
-	threshold = 200
+	threshold = 400
 
 	// The source port is much further away in the inet sock.
 	thresholdInetSock = 2000

vendor/github.com/weaveworks/tcptracer-bpf/tcptracer-bpf.c

@@ -2,6 +2,7 @@
 
 #pragma clang diagnostic push
 #pragma clang diagnostic ignored "-Wgnu-variable-sized-type-not-at-end"
+#pragma clang diagnostic ignored "-Waddress-of-packed-member"
 #include <linux/ptrace.h>
 #pragma clang diagnostic pop
 #include <linux/version.h>
@@ -12,6 +13,7 @@
 #pragma clang diagnostic push
 #pragma clang diagnostic ignored "-Wtautological-compare"
 #pragma clang diagnostic ignored "-Wgnu-variable-sized-type-not-at-end"
+#pragma clang diagnostic ignored "-Wenum-conversion"
 #include <net/sock.h>
 #pragma clang diagnostic pop
 #include <net/inet_sock.h>