debug sporadic vimto hang

[lmb]

ci: debug sporadic vimto hang

There is a case where one of the CI tests hangs for 10 minutes an is then 
killed by the go test harness' timeout. From the dumped goroutines it seems 
like the runner inside the VM is stuck writing to stdout:

goroutine 82 [syscall, 9 minutes]: syscall.Syscall(0x4a4b1c?, 0x708c80?,
0x80000000000?, 0x7ffff80000000000?)
/opt/hostedtoolcache/go/1.21.9/x64/src/syscall/syscall_linux.go:69 +0x25 
syscall.write(0xc000014120?, {0xc0002340c0?, 0x4f0996?, 0xc002048680?})
/opt/hostedtoolcache/go/1.21.9/x64/src/syscall/zsyscall_linux_amd64.go:949
+0x3b syscall.Write(...)
/opt/hostedtoolcache/go/1.21.9/x64/src/syscall/syscall_unix.go:209 
internal/poll.ignoringEINTRIO(...)
/opt/hostedtoolcache/go/1.21.9/x64/src/internal/poll/fd_unix.go:736 
internal/poll.(*FD).Write(0xc000014120, {0xc0002340c0, 0x1c, 0xc0})
/opt/hostedtoolcache/go/1.21.9/x64/src/internal/poll/fd_unix.go:380 +0x35f 
os.(*File).write(...)
/opt/hostedtoolcache/go/1.21.9/x64/src/os/file_posix.go:46 
os.(*File).Write(0xc000040028, {0xc0002340c0?, 0x1c, 0xc0018e4e28:225 +0x97 
testing.(*matcher).fullName(0xc0023b6150, 0xc0023b6150, {0xc00228ec0b?,
0x776c00?})
/opt/hostedtoolcache/go/1.21.9/x64/src/testing/match.go:90 +0x106 
testing.(*T).Run(0xc00217eb60, {0xc00228ec0b?, 0x7fe098330108?},
0xc0001d2a80)
/opt/hostedtoolcache/go/1.21.9/x64/src/testing/testing.go:1639 +0x2db 
github.com/cilium/ebpf/internal/testutils.Files(0xc00217eb60, {0xc0022ae000,
0x51c, 0x2e?}, 0xc00220d320)
/home/runner/work/ebpf/ebpf/internal/testutils/glob.go:25 +0x91 
github.com/cilium/ebpf.TestLibBPFCompat(0xc00217eb60)
/home/runner/work/ebpf/ebpf/elf_reader_test.go:954 +0x14e 
testing.tRunner(0xc00217eb60, 0x792768)
/opt/hostedtoolcache/go/1.21.9/x64/src/testing/testing.go:1595 +0xff created
by testing.(*T).Run in goroutine 1
/opt/hostedtoolcache/go/1.21.9/x64/src/testing/testing.go:1648 +0x3ad

Note that writing the goroutine dump to stderr does work, so not all I/O is
broken. Generate and collect a core dump of the Go test the next time this 
happens.

Signed-off-by: Lorenz Bauer <lmb@isovalent.com>