diff --git a/crypto/fipsmodule/bn/asm/armv8-mont.pl b/crypto/fipsmodule/bn/asm/armv8-mont.pl
index 0fd43c34bb..29334ceb05 100644
--- a/crypto/fipsmodule/bn/asm/armv8-mont.pl
+++ b/crypto/fipsmodule/bn/asm/armv8-mont.pl
@@ -76,6 +76,7 @@
 .type	bn_mul_mont,%function
 .align	5
 bn_mul_mont:
+.cfi_startproc
 	AARCH64_SIGN_LINK_REGISTER
 	tst	$num,#7
 	b.eq	__bn_sqr8x_mont
@@ -83,10 +84,20 @@
 	b.eq	__bn_mul4x_mont
 .Lmul_mont:
 	stp	x29,x30,[sp,#-64]!
+.cfi_def_cfa_offset 64
+.cfi_offset x29, -64
+.cfi_offset x30, -56
 	add	x29,sp,#0
+.cfi_def_cfa x29, 64
 	stp	x19,x20,[sp,#16]
+.cfi_offset x19, -48
+.cfi_offset x20, -40
 	stp	x21,x22,[sp,#32]
+.cfi_offset x21, -32
+.cfi_offset x22, -24
 	stp	x23,x24,[sp,#48]
+.cfi_offset x23, -16
+.cfi_offset x24, -8
 
 	ldr	$m0,[$bp],#8		// bp[0]
 	sub	$tp,sp,$num,lsl#3
@@ -269,13 +280,23 @@
 	str	$nj,[$rp,#-8]
 
 	ldp	x19,x20,[x29,#16]
+.cfi_restore x19
+.cfi_restore x20
 	mov	sp,x29
+.cfi_def_cfa sp, 64
 	ldp	x21,x22,[x29,#32]
+.cfi_restore x21
+.cfi_restore x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore x23
+.cfi_restore x24
 	ldr	x29,[sp],#64
+.cfi_restore x29
+.cfi_def_cfa_offset 0
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	bn_mul_mont,.-bn_mul_mont
 ___
 {
@@ -292,18 +313,33 @@
 .type	__bn_sqr8x_mont,%function
 .align	5
 __bn_sqr8x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_sqr8x_mont is jumped to
 	// only from bn_mul_mont which has already signed the return address.
 	cmp	$ap,$bp
 	b.ne	__bn_mul4x_mont
 .Lsqr8x_mont:
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset 128
+.cfi_offset x29, -128
+.cfi_offset x30, -120
 	add	x29,sp,#0
+.cfi_def_cfa x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset x19, -112
+.cfi_offset x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset x21, -96
+.cfi_offset x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset x23, -80
+.cfi_offset x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset x25, -64
+.cfi_offset x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset x27, -48
+.cfi_offset x28, -40
 	stp	$rp,$np,[sp,#96]	// offload rp and np
 
 	ldp	$a0,$a1,[$ap,#8*0]
@@ -1007,6 +1043,7 @@
 .Lsqr8x8_post_condition:
 	adc	$carry,xzr,xzr
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore x30
 	// $acc0-7,$carry hold result, $a0-7 hold modulus
 	subs	$a0,$acc0,$a0
 	ldr	$ap,[x29,#96]		// pull rp
@@ -1043,16 +1080,29 @@
 
 .Lsqr8x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore x19
+.cfi_restore x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore x21
+.cfi_restore x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore x23
+.cfi_restore x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore x25
+.cfi_restore x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore x27
+.cfi_restore x28
 	ldr	x29,[sp],#128
+.cfi_restore x29
+.cfi_def_cfa_offset 0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	__bn_sqr8x_mont,.-__bn_sqr8x_mont
 ___
 }
@@ -1075,16 +1125,31 @@
 .type	__bn_mul4x_mont,%function
 .align	5
 __bn_mul4x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_mul4x_mont is jumped to
 	// only from bn_mul_mont or __bn_mul8x_mont which have already signed the
 	// return address.
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x30, -120
+.cfi_offset	x29, -128
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 
 	sub	$tp,sp,$num,lsl#3
 	lsl	$num,$num,#3
@@ -1485,6 +1550,7 @@
 	// $acc0-3,$carry hold result, $m0-7 hold modulus
 	subs	$a0,$acc0,$m0
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	sbcs	$a1,$acc1,$m1
 	 stp	xzr,xzr,[sp,#8*0]
 	sbcs	$a2,$acc2,$m2
@@ -1504,16 +1570,29 @@
 
 .Lmul4x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	__bn_mul4x_mont,.-__bn_mul4x_mont
 ___
 }
diff --git a/crypto/fipsmodule/bn/asm/bn-armv8.pl b/crypto/fipsmodule/bn/asm/bn-armv8.pl
index 5aed8df15f..c9380a4b05 100755
--- a/crypto/fipsmodule/bn/asm/bn-armv8.pl
+++ b/crypto/fipsmodule/bn/asm/bn-armv8.pl
@@ -42,6 +42,7 @@
 .globl	bn_add_words
 .align	4
 bn_add_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Clear the carry flag.
 	cmn	xzr, xzr
@@ -72,6 +73,7 @@
 .Ladd_exit:
 	cset	x0, cs
 	ret
+.cfi_endproc
 .size	bn_add_words,.-bn_add_words
 
 // BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
@@ -80,6 +82,7 @@
 .globl	bn_sub_words
 .align	4
 bn_sub_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Set the carry flag. Arm's borrow bit is flipped from the carry flag,
 	# so we want C = 1 here.
@@ -111,7 +114,8 @@
 .Lsub_exit:
 	cset x0, cc
 	ret
-size	bn_sub_words,.-bn_sub_words
+.cfi_endproc
+.size   bn_sub_words,.-bn_sub_words
 ____
 
 print $code;
diff --git a/generated-src/ios-aarch64/crypto/fipsmodule/armv8-mont.S b/generated-src/ios-aarch64/crypto/fipsmodule/armv8-mont.S
index cf798a357d..c787294226 100644
--- a/generated-src/ios-aarch64/crypto/fipsmodule/armv8-mont.S
+++ b/generated-src/ios-aarch64/crypto/fipsmodule/armv8-mont.S
@@ -13,6 +13,7 @@
 
 .align	5
 _bn_mul_mont:
+.cfi_startproc
 	AARCH64_SIGN_LINK_REGISTER
 	tst	x5,#7
 	b.eq	__bn_sqr8x_mont
@@ -20,10 +21,20 @@ _bn_mul_mont:
 	b.eq	__bn_mul4x_mont
 Lmul_mont:
 	stp	x29,x30,[sp,#-64]!
+.cfi_def_cfa_offset	64
+.cfi_offset	x29, -64
+.cfi_offset	x30, -56
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 64
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -48
+.cfi_offset	x20, -40
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -32
+.cfi_offset	x22, -24
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -16
+.cfi_offset	x24, -8
 
 	ldr	x9,[x2],#8		// bp[0]
 	sub	x22,sp,x5,lsl#3
@@ -206,29 +217,54 @@ Lcond_copy:
 	str	x14,[x0,#-8]
 
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
+.cfi_def_cfa	sp, 64
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldr	x29,[sp],#64
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 
 .align	5
 __bn_sqr8x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_sqr8x_mont is jumped to
 	// only from bn_mul_mont which has already signed the return address.
 	cmp	x1,x2
 	b.ne	__bn_mul4x_mont
 Lsqr8x_mont:
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x29, -128
+.cfi_offset	x30, -120
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 	stp	x0,x3,[sp,#96]	// offload rp and np
 
 	ldp	x6,x7,[x1,#8*0]
@@ -929,6 +965,7 @@ Lsqr4x_cond_copy:
 Lsqr8x8_post_condition:
 	adc	x28,xzr,xzr
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	// x19-7,x28 hold result, x6-7 hold modulus
 	subs	x6,x19,x6
 	ldr	x1,[x29,#96]		// pull rp
@@ -965,30 +1002,58 @@ Lsqr8x8_post_condition:
 
 Lsqr8x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 
 .align	5
 __bn_mul4x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_mul4x_mont is jumped to
 	// only from bn_mul_mont or __bn_mul8x_mont which have already signed the
 	// return address.
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x30, -120
+.cfi_offset	x29, -128
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 
 	sub	x26,sp,x5,lsl#3
 	lsl	x5,x5,#3
@@ -1389,6 +1454,7 @@ Lmul4x4_post_condition:
 	// x19-3,x0 hold result, x14-7 hold modulus
 	subs	x6,x19,x14
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	sbcs	x7,x20,x15
 	stp	xzr,xzr,[sp,#8*0]
 	sbcs	x8,x21,x16
@@ -1408,16 +1474,29 @@ Lmul4x4_post_condition:
 
 Lmul4x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 .byte	77,111,110,116,103,111,109,101,114,121,32,77,117,108,116,105,112,108,105,99,97,116,105,111,110,32,102,111,114,32,65,82,77,118,56,44,32,67,82,89,80,84,79,71,65,77,83,32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115,115,108,46,111,114,103,62,0
 .align	2
diff --git a/generated-src/ios-aarch64/crypto/fipsmodule/bn-armv8.S b/generated-src/ios-aarch64/crypto/fipsmodule/bn-armv8.S
index 5e3471a04e..a98baff57e 100644
--- a/generated-src/ios-aarch64/crypto/fipsmodule/bn-armv8.S
+++ b/generated-src/ios-aarch64/crypto/fipsmodule/bn-armv8.S
@@ -15,6 +15,7 @@
 .private_extern	_bn_add_words
 .align	4
 _bn_add_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Clear the carry flag.
 	cmn	xzr, xzr
@@ -45,6 +46,7 @@ Ladd_tail:
 Ladd_exit:
 	cset	x0, cs
 	ret
+.cfi_endproc
 
 
 // BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
@@ -54,6 +56,7 @@ Ladd_exit:
 .private_extern	_bn_sub_words
 .align	4
 _bn_sub_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Set the carry flag. Arm's borrow bit is flipped from the carry flag,
 	# so we want C = 1 here.
@@ -85,5 +88,6 @@ Lsub_tail:
 Lsub_exit:
 	cset	x0, cc
 	ret
+.cfi_endproc
 
 #endif  // !OPENSSL_NO_ASM && defined(OPENSSL_AARCH64) && defined(__APPLE__)
diff --git a/generated-src/linux-aarch64/crypto/fipsmodule/armv8-mont.S b/generated-src/linux-aarch64/crypto/fipsmodule/armv8-mont.S
index 13f045c73f..0003e70fae 100644
--- a/generated-src/linux-aarch64/crypto/fipsmodule/armv8-mont.S
+++ b/generated-src/linux-aarch64/crypto/fipsmodule/armv8-mont.S
@@ -13,6 +13,7 @@
 .type	bn_mul_mont,%function
 .align	5
 bn_mul_mont:
+.cfi_startproc
 	AARCH64_SIGN_LINK_REGISTER
 	tst	x5,#7
 	b.eq	__bn_sqr8x_mont
@@ -20,10 +21,20 @@ bn_mul_mont:
 	b.eq	__bn_mul4x_mont
 .Lmul_mont:
 	stp	x29,x30,[sp,#-64]!
+.cfi_def_cfa_offset	64
+.cfi_offset	x29, -64
+.cfi_offset	x30, -56
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 64
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -48
+.cfi_offset	x20, -40
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -32
+.cfi_offset	x22, -24
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -16
+.cfi_offset	x24, -8
 
 	ldr	x9,[x2],#8		// bp[0]
 	sub	x22,sp,x5,lsl#3
@@ -206,29 +217,54 @@ bn_mul_mont:
 	str	x14,[x0,#-8]
 
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
+.cfi_def_cfa	sp, 64
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldr	x29,[sp],#64
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	bn_mul_mont,.-bn_mul_mont
 .type	__bn_sqr8x_mont,%function
 .align	5
 __bn_sqr8x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_sqr8x_mont is jumped to
 	// only from bn_mul_mont which has already signed the return address.
 	cmp	x1,x2
 	b.ne	__bn_mul4x_mont
 .Lsqr8x_mont:
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x29, -128
+.cfi_offset	x30, -120
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 	stp	x0,x3,[sp,#96]	// offload rp and np
 
 	ldp	x6,x7,[x1,#8*0]
@@ -929,6 +965,7 @@ __bn_sqr8x_mont:
 .Lsqr8x8_post_condition:
 	adc	x28,xzr,xzr
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	// x19-7,x28 hold result, x6-7 hold modulus
 	subs	x6,x19,x6
 	ldr	x1,[x29,#96]		// pull rp
@@ -965,30 +1002,58 @@ __bn_sqr8x_mont:
 
 .Lsqr8x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	__bn_sqr8x_mont,.-__bn_sqr8x_mont
 .type	__bn_mul4x_mont,%function
 .align	5
 __bn_mul4x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_mul4x_mont is jumped to
 	// only from bn_mul_mont or __bn_mul8x_mont which have already signed the
 	// return address.
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x30, -120
+.cfi_offset	x29, -128
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 
 	sub	x26,sp,x5,lsl#3
 	lsl	x5,x5,#3
@@ -1389,6 +1454,7 @@ __bn_mul4x_mont:
 	// x19-3,x0 hold result, x14-7 hold modulus
 	subs	x6,x19,x14
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	sbcs	x7,x20,x15
 	stp	xzr,xzr,[sp,#8*0]
 	sbcs	x8,x21,x16
@@ -1408,16 +1474,29 @@ __bn_mul4x_mont:
 
 .Lmul4x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 .size	__bn_mul4x_mont,.-__bn_mul4x_mont
 .byte	77,111,110,116,103,111,109,101,114,121,32,77,117,108,116,105,112,108,105,99,97,116,105,111,110,32,102,111,114,32,65,82,77,118,56,44,32,67,82,89,80,84,79,71,65,77,83,32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115,115,108,46,111,114,103,62,0
 .align	2
diff --git a/generated-src/linux-aarch64/crypto/fipsmodule/bn-armv8.S b/generated-src/linux-aarch64/crypto/fipsmodule/bn-armv8.S
index 2b8823ab36..3431e36555 100644
--- a/generated-src/linux-aarch64/crypto/fipsmodule/bn-armv8.S
+++ b/generated-src/linux-aarch64/crypto/fipsmodule/bn-armv8.S
@@ -15,6 +15,7 @@
 .hidden	bn_add_words
 .align	4
 bn_add_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Clear the carry flag.
 	cmn	xzr, xzr
@@ -45,6 +46,7 @@ bn_add_words:
 .Ladd_exit:
 	cset	x0, cs
 	ret
+.cfi_endproc
 .size	bn_add_words,.-bn_add_words
 
 // BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
@@ -54,6 +56,7 @@ bn_add_words:
 .hidden	bn_sub_words
 .align	4
 bn_sub_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Set the carry flag. Arm's borrow bit is flipped from the carry flag,
 	# so we want C = 1 here.
@@ -85,5 +88,6 @@ bn_sub_words:
 .Lsub_exit:
 	cset	x0, cc
 	ret
+.cfi_endproc
 .size	bn_sub_words,.-bn_sub_words
 #endif  // !OPENSSL_NO_ASM && defined(OPENSSL_AARCH64) && defined(__ELF__)
diff --git a/generated-src/win-aarch64/crypto/fipsmodule/armv8-mont.S b/generated-src/win-aarch64/crypto/fipsmodule/armv8-mont.S
index dcce02c95d..2640109611 100644
--- a/generated-src/win-aarch64/crypto/fipsmodule/armv8-mont.S
+++ b/generated-src/win-aarch64/crypto/fipsmodule/armv8-mont.S
@@ -15,6 +15,7 @@
 .endef
 .align	5
 bn_mul_mont:
+.cfi_startproc
 	AARCH64_SIGN_LINK_REGISTER
 	tst	x5,#7
 	b.eq	__bn_sqr8x_mont
@@ -22,10 +23,20 @@ bn_mul_mont:
 	b.eq	__bn_mul4x_mont
 Lmul_mont:
 	stp	x29,x30,[sp,#-64]!
+.cfi_def_cfa_offset	64
+.cfi_offset	x29, -64
+.cfi_offset	x30, -56
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 64
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -48
+.cfi_offset	x20, -40
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -32
+.cfi_offset	x22, -24
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -16
+.cfi_offset	x24, -8
 
 	ldr	x9,[x2],#8		// bp[0]
 	sub	x22,sp,x5,lsl#3
@@ -208,31 +219,56 @@ Lcond_copy:
 	str	x14,[x0,#-8]
 
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
+.cfi_def_cfa	sp, 64
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldr	x29,[sp],#64
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 .def __bn_sqr8x_mont
    .type 32
 .endef
 .align	5
 __bn_sqr8x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_sqr8x_mont is jumped to
 	// only from bn_mul_mont which has already signed the return address.
 	cmp	x1,x2
 	b.ne	__bn_mul4x_mont
 Lsqr8x_mont:
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x29, -128
+.cfi_offset	x30, -120
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 	stp	x0,x3,[sp,#96]	// offload rp and np
 
 	ldp	x6,x7,[x1,#8*0]
@@ -933,6 +969,7 @@ Lsqr4x_cond_copy:
 Lsqr8x8_post_condition:
 	adc	x28,xzr,xzr
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	// x19-7,x28 hold result, x6-7 hold modulus
 	subs	x6,x19,x6
 	ldr	x1,[x29,#96]		// pull rp
@@ -969,32 +1006,60 @@ Lsqr8x8_post_condition:
 
 Lsqr8x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 .def __bn_mul4x_mont
    .type 32
 .endef
 .align	5
 __bn_mul4x_mont:
+.cfi_startproc
 	// Not adding AARCH64_SIGN_LINK_REGISTER here because __bn_mul4x_mont is jumped to
 	// only from bn_mul_mont or __bn_mul8x_mont which have already signed the
 	// return address.
 	stp	x29,x30,[sp,#-128]!
+.cfi_def_cfa_offset	128
+.cfi_offset	x30, -120
+.cfi_offset	x29, -128
 	add	x29,sp,#0
+.cfi_def_cfa	x29, 128
 	stp	x19,x20,[sp,#16]
+.cfi_offset	x19, -112
+.cfi_offset	x20, -104
 	stp	x21,x22,[sp,#32]
+.cfi_offset	x21, -96
+.cfi_offset	x22, -88
 	stp	x23,x24,[sp,#48]
+.cfi_offset	x23, -80
+.cfi_offset	x24, -72
 	stp	x25,x26,[sp,#64]
+.cfi_offset	x25, -64
+.cfi_offset	x26, -56
 	stp	x27,x28,[sp,#80]
+.cfi_offset	x27, -48
+.cfi_offset	x28, -40
 
 	sub	x26,sp,x5,lsl#3
 	lsl	x5,x5,#3
@@ -1395,6 +1460,7 @@ Lmul4x4_post_condition:
 	// x19-3,x0 hold result, x14-7 hold modulus
 	subs	x6,x19,x14
 	ldr	x30,[x29,#8]		// pull return address
+.cfi_restore	x30
 	sbcs	x7,x20,x15
 	stp	xzr,xzr,[sp,#8*0]
 	sbcs	x8,x21,x16
@@ -1414,16 +1480,29 @@ Lmul4x4_post_condition:
 
 Lmul4x_done:
 	ldp	x19,x20,[x29,#16]
+.cfi_restore	x19
+.cfi_restore	x20
 	mov	sp,x29
 	ldp	x21,x22,[x29,#32]
+.cfi_restore	x21
+.cfi_restore	x22
 	mov	x0,#1
 	ldp	x23,x24,[x29,#48]
+.cfi_restore	x23
+.cfi_restore	x24
 	ldp	x25,x26,[x29,#64]
+.cfi_restore	x25
+.cfi_restore	x26
 	ldp	x27,x28,[x29,#80]
+.cfi_restore	x27
+.cfi_restore	x28
 	ldr	x29,[sp],#128
+.cfi_restore	x29
+.cfi_def_cfa_offset	0
 	// x30 is popped earlier
 	AARCH64_VALIDATE_LINK_REGISTER
 	ret
+.cfi_endproc
 
 .byte	77,111,110,116,103,111,109,101,114,121,32,77,117,108,116,105,112,108,105,99,97,116,105,111,110,32,102,111,114,32,65,82,77,118,56,44,32,67,82,89,80,84,79,71,65,77,83,32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115,115,108,46,111,114,103,62,0
 .align	2
diff --git a/generated-src/win-aarch64/crypto/fipsmodule/bn-armv8.S b/generated-src/win-aarch64/crypto/fipsmodule/bn-armv8.S
index af97080007..c72c0db415 100644
--- a/generated-src/win-aarch64/crypto/fipsmodule/bn-armv8.S
+++ b/generated-src/win-aarch64/crypto/fipsmodule/bn-armv8.S
@@ -15,6 +15,7 @@
 
 .align	4
 bn_add_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Clear the carry flag.
 	cmn	xzr, xzr
@@ -45,6 +46,7 @@ Ladd_tail:
 Ladd_exit:
 	cset	x0, cs
 	ret
+.cfi_endproc
 
 
 // BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
@@ -54,6 +56,7 @@ Ladd_exit:
 
 .align	4
 bn_sub_words:
+.cfi_startproc
 	AARCH64_VALID_CALL_TARGET
 	# Set the carry flag. Arm's borrow bit is flipped from the carry flag,
 	# so we want C = 1 here.
@@ -85,5 +88,6 @@ Lsub_tail:
 Lsub_exit:
 	cset	x0, cc
 	ret
+.cfi_endproc
 
 #endif  // !OPENSSL_NO_ASM && defined(OPENSSL_AARCH64) && defined(_WIN32)