diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 5b27f70..20d51b4 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -43,7 +43,7 @@ jobs:
       - name: Replay
         run: uftrace replay --data=tracedir --output-fields=tid | tee ci.snap
       - name: Compare to snapshot
-        run: diff ci.snap out.snap
+        run: diff -u out.snap ci.snap
 
   rust:
     name: Rust
@@ -70,4 +70,4 @@ jobs:
       - name: Replay
         run: uftrace replay --data=tracedir --output-fields=tid | tee ci.snap
       - name: Compare to snapshot
-        run: diff ci.snap out.snap
+        run: diff -u out.snap ci.snap
diff --git a/examples/rust/out.snap b/examples/rust/out.snap
index eccf366..2b26afc 100644
--- a/examples/rust/out.snap
+++ b/examples/rust/out.snap
@@ -9,3 +9,13 @@
  [     1] |     } /* rftrace_rs_test::test3 */
  [     1] |   } /* rftrace_rs_test::test2 */
  [     1] | } /* rftrace_rs_test::test1 */
+ [     1] | rftrace_frontend::frontend::dump_full_uftrace() {
+ [     1] |   rftrace_frontend::frontend::dump_traces() {
+ [     1] |     rftrace_frontend::frontend::disable();
+
+uftrace stopped tracing with remaining functions
+================================================
+task: 1
+[1] rftrace_frontend::frontend::dump_traces
+[0] rftrace_frontend::frontend::dump_full_uftrace
+
diff --git a/rftrace/src/backend.rs b/rftrace/src/backend.rs
index bb77366..2b61aab 100644
--- a/rftrace/src/backend.rs
+++ b/rftrace/src/backend.rs
@@ -1,4 +1,4 @@
-use core::arch::asm;
+use core::arch::naked_asm;
 use core::arch::x86_64::_rdtsc;
 use core::slice;
 use core::sync::atomic::{AtomicBool, AtomicU64, AtomicUsize, Ordering};
@@ -87,7 +87,7 @@ pub unsafe extern "C" fn mcount() {
     // mcounts ret addr is directly at rsp
 
     // based on https://github.com/namhyung/uftrace/blob/master/arch/x86_64/mcount.S
-    asm!(
+    naked_asm!(
         // if ENABLED.load(Ordering::Relaxed) {
         //     return;
         // }
@@ -132,7 +132,6 @@ pub unsafe extern "C" fn mcount() {
         "pop rax",
         "ret",
         // TODO: ENABLED = sym ENABLED,
-        options(noreturn),
     );
 }
 
@@ -395,7 +394,7 @@ pub unsafe extern "C" fn mcount_return_trampoline() {
         RSP +0      rax
     */
 
-    asm!(
+    naked_asm!(
         prologue!(),
         // always backup return registers
         "mov [rsp + 8], rdx",
@@ -416,7 +415,6 @@ pub unsafe extern "C" fn mcount_return_trampoline() {
         restore_sse2!(),
         epilogue!(),
         "ret",
-        options(noreturn),
     );
 }