From 554ae92194d82663999f9d1fe916f2c3cf2a249b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=C3=96zg=C3=BCr=20Kesim?= <oec@codeblau.de>
Date: Thu, 16 Jan 2020 10:30:49 +0100
Subject: [PATCH] Test added to compare the Scanner vs GDB

gdb_test.go and testdata implement a go test to compare and verify that
we get correct results from the scanner.

testdata/simple.c is a small program that embeds Python and runs the
same code as in runforever.py.  The Makefile compiles this using
python3.7.  Adjustments to the flags might be needed in your environment

gdb_test.go contains only one test, TestSimpleGDB, that
  1. compiles simple.c
  2. runs it multiple times
  3. calls gdb to extract the address of symbol _PyRuntime for each pid
  4. runs our Scanner
  5. compares the results from gdb and our scanner
---
 GetRuntimeAddresses/symbolyze/gdb_test.go     | 106 ++++++++++++++++++
 .../symbolyze/testdata/.gitignore             |   2 +
 .../symbolyze/testdata/Makefile               |  12 ++
 .../symbolyze/testdata/simple.c               |  22 ++++
 4 files changed, 142 insertions(+)
 create mode 100644 GetRuntimeAddresses/symbolyze/gdb_test.go
 create mode 100644 GetRuntimeAddresses/symbolyze/testdata/.gitignore
 create mode 100644 GetRuntimeAddresses/symbolyze/testdata/Makefile
 create mode 100644 GetRuntimeAddresses/symbolyze/testdata/simple.c

diff --git a/GetRuntimeAddresses/symbolyze/gdb_test.go b/GetRuntimeAddresses/symbolyze/gdb_test.go
new file mode 100644
index 0000000..36eac45
--- /dev/null
+++ b/GetRuntimeAddresses/symbolyze/gdb_test.go
@@ -0,0 +1,106 @@
+package symbolyze
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"os/exec"
+	"regexp"
+	"strconv"
+	"testing"
+)
+
+func buildSimple() error {
+	cmd := exec.Command("make")
+	cmd.Dir = "testdata"
+	cmd.Stdout = os.Stdout
+	cmd.Stderr = os.Stderr
+
+	return cmd.Run()
+}
+
+func TestSimpleGDB(t *testing.T) {
+	ourResults := map[int]uint64{}
+	gdbResults := map[int]uint64{}
+
+	err := buildSimple()
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	scanner := NewScanner("_PyRuntime", "*python3*")
+	scanner.OnFound(func(pid int, offset uint64) error {
+		t.Logf("scanner setting pid %d to offset %x", pid, offset)
+		ourResults[pid] = offset
+		return nil
+	})
+
+	for i := 0; i < 5; i++ {
+		simple := exec.Command("testdata/simple")
+		if err := simple.Start(); err != nil {
+			t.Fatal(err)
+		}
+
+		defer func() {
+			simple.Process.Kill()
+			go simple.Wait()
+		}()
+
+		t.Logf("asking Gdb about PID %d", simple.Process.Pid)
+
+		offset, err := extractOffsetWithGdb(simple.Process.Pid, t)
+		if err != nil {
+			t.Fatalf("extractOffsetWithGdb: %v", err)
+		}
+
+		t.Logf("Found offset at 0x%x", offset)
+		gdbResults[simple.Process.Pid] = offset
+	}
+
+	if err = scanner.Run(); err != nil {
+		t.Logf("Scanner failure: %v", err)
+	}
+
+	for pid, off := range gdbResults {
+		if off2, ok := ourResults[pid]; !ok {
+			t.Fatalf("Scanner has no offset for pid %d", pid)
+		} else if off != off2 {
+			t.Fatalf("Scanner has found offset %d while gdb has found offset %d for pid %d", off2, off, pid)
+		}
+	}
+}
+
+var searchRX = regexp.MustCompile(`Symbol "_PyRuntime" is static storage at address 0x([0-9a-fA-F]+)`)
+
+func extractOffsetWithGdb(pid int, t *testing.T) (offset uint64, err error) {
+	cmd := exec.Command("gdb",
+		"--batch",
+		"-p", strconv.Itoa(pid),
+		"-ex", "info address _PyRuntime",
+		"-ex", "quit")
+
+	output, err := cmd.StdoutPipe()
+	if err != nil {
+		return 0, err
+	}
+
+	if err := cmd.Start(); err != nil {
+		return 0, err
+	}
+
+	scanner := bufio.NewScanner(output)
+	for scanner.Scan() {
+		line := scanner.Text()
+		// t.Logf("[gdb]: %v", line)
+		match := searchRX.FindStringSubmatch(line)
+		if len(match) > 1 {
+			return strconv.ParseUint(match[1], 16, 64)
+		}
+	}
+
+	if err := cmd.Wait(); err != nil {
+		return 0, err
+	}
+
+	return 0, fmt.Errorf("symbol not found with gdb")
+}
diff --git a/GetRuntimeAddresses/symbolyze/testdata/.gitignore b/GetRuntimeAddresses/symbolyze/testdata/.gitignore
new file mode 100644
index 0000000..2f17ffb
--- /dev/null
+++ b/GetRuntimeAddresses/symbolyze/testdata/.gitignore
@@ -0,0 +1,2 @@
+simple
+simple.o
diff --git a/GetRuntimeAddresses/symbolyze/testdata/Makefile b/GetRuntimeAddresses/symbolyze/testdata/Makefile
new file mode 100644
index 0000000..0c30885
--- /dev/null
+++ b/GetRuntimeAddresses/symbolyze/testdata/Makefile
@@ -0,0 +1,12 @@
+CC=gcc
+PC=python3.7dm-config
+PIE=-fPIE # this was necessary on my machine
+CFLAGS=$(shell $(PC) --cflags)
+LDFLAGS=$(shell $(PC) --ldflags)
+
+simple: simple.o
+	$(CC) -o $@ $< $(LDFLAGS)
+
+simple.o: simple.c
+	$(CC) -c $(CFLAGS) $(PIE) $<
+
diff --git a/GetRuntimeAddresses/symbolyze/testdata/simple.c b/GetRuntimeAddresses/symbolyze/testdata/simple.c
new file mode 100644
index 0000000..437019f
--- /dev/null
+++ b/GetRuntimeAddresses/symbolyze/testdata/simple.c
@@ -0,0 +1,22 @@
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+
+static char script[] = 	"import time\n"
+			"import os\n"
+			"\n"
+			"print(\"Process ID: {}\".format(os.getpid()))\n"
+			"print(\"Entering infinite loop ...\")\n"
+			"while 1:\n"
+			"    time.sleep(2)\n"
+			"\n";
+
+int
+main(int argc, char *argv[])
+{
+    Py_Initialize();
+    PyRun_SimpleString(script);
+    if (Py_FinalizeEx() < 0) {
+        exit(120);
+    }
+    return 0;
+}