diff --git a/sys/arm64/arm64/vfp.c b/sys/arm64/arm64/vfp.c
--- a/sys/arm64/arm64/vfp.c
+++ b/sys/arm64/arm64/vfp.c
@@ -199,6 +199,32 @@
 	critical_exit();
 }
 
+/*
+ * Update the VFP state for a forked process or new thread. The PCB will
+ * have been copied from the old thread.
+ */
+void
+vfp_new_thread(struct thread *newtd, struct thread *oldtd, bool fork)
+{
+	struct pcb *newpcb;
+
+	newpcb = newtd->td_pcb;
+
+	/* Kernel threads start with clean VFP */
+	if ((oldtd->td_pflags & TDP_KTHREAD) != 0) {
+		newpcb->pcb_fpflags &=
+		    ~(PCB_FP_STARTED | PCB_FP_KERN | PCB_FP_NOSAVE);
+	} else {
+		MPASS((newpcb->pcb_fpflags & (PCB_FP_KERN|PCB_FP_NOSAVE)) == 0);
+		if (!fork) {
+			newpcb->pcb_fpflags &= ~PCB_FP_STARTED;
+		}
+	}
+
+	newpcb->pcb_fpusaved = &newpcb->pcb_fpustate;
+	newpcb->pcb_vfpcpu = UINT_MAX;
+}
+
 /*
  * Reset the FP state to avoid leaking state from the parent process across
  * execve() (and to ensure that we get a consistent floating point environment
diff --git a/sys/arm64/arm64/vm_machdep.c b/sys/arm64/arm64/vm_machdep.c
--- a/sys/arm64/arm64/vm_machdep.c
+++ b/sys/arm64/arm64/vm_machdep.c
@@ -109,8 +109,8 @@
 	td2->td_pcb->pcb_x[20] = (uintptr_t)td2;
 	td2->td_pcb->pcb_lr = (uintptr_t)fork_trampoline;
 	td2->td_pcb->pcb_sp = (uintptr_t)td2->td_frame;
-	td2->td_pcb->pcb_fpusaved = &td2->td_pcb->pcb_fpustate;
-	td2->td_pcb->pcb_vfpcpu = UINT_MAX;
+
+	vfp_new_thread(td2, td1, true);
 
 	/* Setup to release spin count in fork_exit(). */
 	td2->td_md.md_spinlock_count = 1;
@@ -187,9 +187,9 @@
 	td->td_pcb->pcb_x[20] = (uintptr_t)td;
 	td->td_pcb->pcb_lr = (uintptr_t)fork_trampoline;
 	td->td_pcb->pcb_sp = (uintptr_t)td->td_frame;
-	td->td_pcb->pcb_fpflags &= ~(PCB_FP_STARTED | PCB_FP_KERN | PCB_FP_NOSAVE);
-	td->td_pcb->pcb_fpusaved = &td->td_pcb->pcb_fpustate;
-	td->td_pcb->pcb_vfpcpu = UINT_MAX;
+
+	/* Update VFP state for the new thread */
+	vfp_new_thread(td, td0, false);
 
 	/* Setup to release spin count in fork_exit(). */
 	td->td_md.md_spinlock_count = 1;
diff --git a/sys/arm64/include/vfp.h b/sys/arm64/include/vfp.h
--- a/sys/arm64/include/vfp.h
+++ b/sys/arm64/include/vfp.h
@@ -68,6 +68,7 @@
 
 void	vfp_init(void);
 void	vfp_discard(struct thread *);
+void	vfp_new_thread(struct thread *, struct thread *, bool);
 void	vfp_reset_state(struct thread *, struct pcb *);
 void	vfp_restore_state(void);
 void	vfp_save_state(struct thread *, struct pcb *);