diff --git a/asm/inline.rs b/asm/inline.rs
index a7898c57..5887bafb 100644
--- a/asm/inline.rs
+++ b/asm/inline.rs
@@ -55,6 +55,7 @@ pub unsafe fn __delay(cyc: u32) {
     // The loop will normally take 3 to 4 CPU cycles per iteration, but superscalar cores
     // (eg. Cortex-M7) can potentially do it in 2, so we use that as the lower bound, since delaying
     // for more cycles is okay.
+    // Add 1 to prevent an integer underflow which would cause a long freeze
     let real_cyc = 1 + cyc / 2;
     asm!(
         // Use local labels to avoid R_ARM_THM_JUMP8 relocations which fail on thumbv6m.