diff --git a/Makefile b/Makefile
index dc5c6fd..ad17b56 100644
--- a/Makefile
+++ b/Makefile
@@ -1,10 +1,10 @@
 DIRS = $(sort $(wildcard src/*))
 
 all:
-	for i in $(DIRS); do echo $$i; make -s -C $$i; done
-t: all
-#	for i in $(DIRS); do echo $$i; $$i/t; done
-b: all
-	for i in $(DIRS); do echo $$i; $$i/b; done
+	for i in $(DIRS); do printf "%-20s" $$i; make -s -C $$i; done
+t:
+	@for i in $(DIRS); do printf "%-20s" $$i; make -s -C $$i 2>/dev/null; done
+b: t
+	@for i in $(DIRS); do echo $$i; $$i/b; done
 clean:
 	for i in $(DIRS); do make -s -C $$i clean; done
diff --git a/Makefile.inc b/Makefile.inc
index 4613e08..f0ebc00 100644
--- a/Makefile.inc
+++ b/Makefile.inc
@@ -9,7 +9,7 @@ OBJS = $(SRCS:.c=.o)
 AR=ar
 RANLIB=ranlib
 
-CFLAGS += -g -D_POSIX_C_SOURCE=200809L -std=c99 -pipe -Wall
+CFLAGS += -g -D_POSIX_C_SOURCE=200809L -std=c99 -pipe -Wall -Wno-unused-function
 LDFLAGS += -g -lpthread -lrt -lm
 
 -include $(TROOT)/config.mak
diff --git a/common/b.c b/common/b.c
index 91f137e..07b299c 100644
--- a/common/b.c
+++ b/common/b.c
@@ -16,7 +16,7 @@
 static int verbose = 1;
 
 void error__(const char *n, int l, const char *s, ...) {
-	fprintf(stderr, "use error in tests only\n");
+	dprintf(1, "use error in tests only\n");
 }
 
 static int N;
@@ -32,7 +32,7 @@ static unsigned long long tic() {
 	struct timespec ts;
 
 	if (clock_gettime(CLOCK_REALTIME, &ts) < 0) {
-		fprintf(stderr, "bench: clock_gettime failed: %s\n", strerror(errno));
+		dprintf(1, "bench: clock_gettime failed: %s\n", strerror(errno));
 		return 0;
 	}
 	return ts.tv_sec*SEC + ts.tv_nsec;
@@ -98,17 +98,17 @@ void vmstats() {
 		}
 	}
 	if (f) fclose(f);
-	fprintf(stderr, " %7zu virt %7zu res %7zu dirty", vm_size, vm_rss, vm_priv_dirty);
+	dprintf(1, " %7zu virt %7zu res %7zu dirty", vm_size, vm_rss, vm_priv_dirty);
 }
 
 void stats() {
-	if (dt/N > 100)
-		fprintf(stderr, "%10d N %10llu ns/op   ", N, dt/N);
+	if (dt/N >= 100)
+		dprintf(1, "%10d N %10llu ns/op   ", N, dt/N);
 	else
-		fprintf(stderr, "%10d N %13.2f ns/op", N, (double)dt/N);
+		dprintf(1, "%10d N %13.2f ns/op", N, (double)dt/N);
 	if (verbose)
 		vmstats();
-	fputc('\n', stderr);
+	dprintf(1, "\n");
 }
 
 static void run(const char *name, void (*f)(int)) {
@@ -116,23 +116,23 @@ static void run(const char *name, void (*f)(int)) {
 	if (p) {
 		int s;
 		if (p<0 || wait(&s)<0 || !WIFEXITED(s) || WEXITSTATUS(s))
-			fprintf(stderr, "benchmark %s failed\n", name);
+			dprintf(1, "benchmark %s failed\n", name);
 		return;
 	}
-	fprintf(stderr, "%-32s", name);
+	dprintf(1, "%-32s", name);
 	for (N=1; ; N=nextN()) {
 		// TODO: fork at each iteration and pass N,dt..?
 		reset_timer();
 		start_timer();
 		f(N);
 		stop_timer();
-//		fprintf(stderr, "%10d%12llu next: %d\n", N, dt, nextN());
+//		dprintf(1, "%10d%12llu next: %d\n", N, dt, nextN());
 		if (dt >= SEC/2 || N >= MAXN) {
 			stats();
 			exit(0);
 		}
 		if (N <= 0) {
-			fprintf(stderr, "bench: fatal: N <= 0\n");
+			dprintf(1, "bench: fatal: N <= 0\n");
 			exit(1);
 		}
 	}
diff --git a/common/t.c b/common/t.c
index ecaaf7f..2373b09 100644
--- a/common/t.c
+++ b/common/t.c
@@ -29,10 +29,12 @@ void reset_timer() { errtimer(); }
 void error__(const char *n, int l, const char *s, ...) {
 	va_list ap;
 
+	if (failed == 0 && nfailed == 0)
+		dprintf(1, "FAIL\n", n);
 	failed = 1;
-	fprintf(stderr, " ERROR %s %s:%d: ", name, n, l);
+	dprintf(1, " ERROR %s %s:%d: ", name, n, l);
 	va_start(ap, s);
-	vfprintf(stderr, s, ap);
+	vdprintf(1, s, ap);
 	va_end(ap);
 }
 
@@ -44,7 +46,7 @@ static void run(const char *n, void (*f)()) {
 	failed = 0;
 	name = n;
 	if (verbose)
-		fprintf(stderr, "running %s:\n", name);
+		dprintf(1, "running %s:\n", name);
 
 	pid = fork();
 	if (pid == 0) {
@@ -66,21 +68,21 @@ static void run(const char *n, void (*f)()) {
 
 	if (failed) {
 		nfailed++;
-		fprintf(stderr, "FAILED %s\n", name);
+		dprintf(1, "FAILED %s\n", name);
 	} else if (verbose)
-		fprintf(stderr, "PASSED %s\n", name);
+		dprintf(1, "PASSED %s\n", name);
 }
 
 static int summary() {
 	if (nfailed)
-		fprintf(stderr, "FAIL:%d (all: %d)\n", nfailed, count);
+		dprintf(1, "FAIL (%d out of %d tests)\n", nfailed, count);
 	else
-		fprintf(stderr, "PASS (all: %d)\n", count);
+		dprintf(1, "ok (%d tests)\n", count);
 	return !!nfailed;
 }
 
 static void usage() {
-	fprintf(stderr, "usage: ./t [-vs]\n");
+	dprintf(1, "usage: ./t [-vs]\n");
 	exit(1);
 }
 
diff --git a/src/udiv/udiv.c b/src/udiv/udiv.c
index 945e144..42010ee 100644
--- a/src/udiv/udiv.c
+++ b/src/udiv/udiv.c
@@ -166,7 +166,7 @@ void test_udiv()
 	}
 }
 
-void bench_div(int N)
+void bench_udiv(int N)
 {
 	int i;
 	volatile uint64_t r = 0;
@@ -176,7 +176,7 @@ void bench_div(int N)
 		r += ((uint64_t)i<<32) / d;
 }
 
-void bench_mod(int N)
+void bench_umod(int N)
 {
 	int i;
 	volatile uint64_t r = 0;