src/cmd/5l/noop.c                   | 23 ++++++++++++++++++-----
 src/pkg/runtime/pprof/pprof_test.go | 25 +++++++++++++++++++++++++
 src/pkg/runtime/traceback_arm.c     |  2 +-

diff --git a/src/cmd/5l/noop.c b/src/cmd/5l/noop.c
index fb70599b514951ec60533faea0f6cf9339c59fa4..305ed684ee4d2c0a09aaad78060293abaad5a5ec 100644
--- a/src/cmd/5l/noop.c
+++ b/src/cmd/5l/noop.c
@@ -472,14 +472,27 @@ 				p->to.type = D_REG;
 				p->to.reg = REGSP;
 				p->spadj = -8;
 	
+				/* Keep saved LR at 0(SP) after SP change. */
+				/* MOVW 0(SP), REGTMP; MOVW REGTMP, -8!(SP) */
+				/* TODO: Remove SP adjustments; see issue 6699. */
+				q1->as = AMOVW;
+				q1->from.type = D_OREG;
+				q1->from.reg = REGSP;
+				q1->from.offset = 0;
+				q1->reg = NREG;
+				q1->to.type = D_REG;
+				q1->to.reg = REGTMP;
+
 				/* SUB $8,SP */
-				q1->as = ASUB;
-				q1->from.type = D_CONST;
-				q1->from.offset = 8;
-				q1->from.reg = NREG;
+				q1 = appendp(q1);
+				q1->as = AMOVW;
+				q1->from.type = D_REG;
+				q1->from.reg = REGTMP;
 				q1->reg = NREG;
-				q1->to.type = D_REG;
+				q1->to.type = D_OREG;
 				q1->to.reg = REGSP;
+				q1->to.offset = -8;
+				q1->scond |= C_WBIT;
 				q1->spadj = 8;
 	
 				break;
diff --git a/src/pkg/runtime/pprof/pprof_test.go b/src/pkg/runtime/pprof/pprof_test.go
index f1fc5faec689d050fc7124a8f0621801f36af4f8..eb76b93c44cfe2e584f28d49eec470f7f550fd4f 100644
--- a/src/pkg/runtime/pprof/pprof_test.go
+++ b/src/pkg/runtime/pprof/pprof_test.go
@@ -8,6 +8,7 @@ import (
 	"bytes"
 	"fmt"
 	"hash/crc32"
+	"math/big"
 	"os/exec"
 	"regexp"
 	"runtime"
@@ -123,6 +124,10 @@ 			}
 		}
 	})
 
+	if len(need) == 0 {
+		return
+	}
+
 	var total uintptr
 	for i, name := range need {
 		total += have[i]
@@ -235,6 +240,26 @@ 				t.Fatalf("found profile entry for runtime.gogo:\n%s", buf.String())
 			}
 		})
 	}
+}
+
+// Test that profiling of division operations is okay, especially on ARM. See issue 6681.
+func TestMathBigDivide(t *testing.T) {
+	testCPUProfile(t, nil, func() {
+		t := time.After(5 * time.Second)
+		pi := new(big.Int)
+		for {
+			for i := 0; i < 100; i++ {
+				n := big.NewInt(2646693125139304345)
+				d := big.NewInt(842468587426513207)
+				pi.Div(n, d)
+			}
+			select {
+			case <-t:
+				return
+			default:
+			}
+		}
+	})
 }
 
 // Operating systems that are expected to fail the tests. See issue 6047.
diff --git a/src/pkg/runtime/traceback_arm.c b/src/pkg/runtime/traceback_arm.c
index 02586f036bba6dad9d60a27fab63b8b14ea5a510..341aa20588010d7a3a0574dce69732a431084e7b 100644
--- a/src/pkg/runtime/traceback_arm.c
+++ b/src/pkg/runtime/traceback_arm.c
@@ -84,7 +84,7 @@ 		if(runtime·topofstack(f)) {
 			frame.lr = 0;
 			flr = nil;
 		} else {
-			if(frame.lr == 0)
+			if((n == 0 && frame.sp < frame.fp) || frame.lr == 0)
 				frame.lr = *(uintptr*)frame.sp;
 			flr = runtime·findfunc(frame.lr);
 			if(flr == nil) {