string: arm: Add new functionality to prologue/epilogue assembler macros.

This patch adds options for automatic alignment enforcement and for
pushing/popping the lr register to prologue and epilogue assembler macros,
while making the pushing of the ip register optional for PACBTI.

Furthermore, as the use of these macros is independent of PACBTI and
may be used on architectures without the feature, the macros are moved
to a common header.

Improvements are also made to cfi handling. Where absolute cfi offset
calculation is complicated by optional function prologue
parameters (e.g. the pushing of pac-codes to the stack for M-profile
pacbti on function entry and pushing of dummy register when alignment
required), replace the use of .cfi_offset for .cfi_rel_offset,
simplifying cfi calculations by basing offsets on SP rather than the
cfa.

Finally, extensive in-source documentation is added to these macros to
facilitate their use and further development.

Built w/ arm-none-linux-gnueabihf, ran make check-string w/ qemu-arm-static.

3 files changed, 32 insertions, 27 deletions

diff --git a/string/arm/memchr.S b/string/arm/memchr.S
index 83a96ca..125618d 100644
--- a/string/arm/memchr.S
+++ b/string/arm/memchr.S
@@ -36,8 +36,8 @@
 #define CHARTSTMASK(c) 1<<(c*8)
 #endif
 	.thumb
+#include "../asmdefs.h"
 
-#include "../pacbti.h"
 
 @ ---------------------------------------------------------------------------
 	.thumb_func
@@ -74,10 +74,10 @@ __memchr_arm:
 	@ At this point, we are aligned, we know we have at least 8 bytes to work with
 	push	{r4,r5,r6,r7}
 	.cfi_adjust_cfa_offset 16
-	.cfi_offset 4, -(16+PAC_CFI_ADJ)
-	.cfi_offset 5, -(12+PAC_CFI_ADJ)
-	.cfi_offset 6, -(8+PAC_CFI_ADJ)
-	.cfi_offset 7, -(4+PAC_CFI_ADJ)
+	.cfi_rel_offset 4, 0
+	.cfi_rel_offset 5, 4
+	.cfi_rel_offset 6, 8
+	.cfi_rel_offset 7, 12
 	orr	r1, r1, r1, lsl #8	@ expand the match word across to all bytes
 	orr	r1, r1, r1, lsl #16
 	bic	r4, r2, #7	@ Number of double words to work with
@@ -116,16 +116,20 @@ __memchr_arm:
 	bne	21b		@ on r2 flags
 
 40:
+	.cfi_remember_state
 	movs	r0,#0		@ not found
 	epilogue
 
 50:
+	.cfi_restore_state
+	.cfi_remember_state
 	subs	r0,r0,#1	@ found
 	epilogue
 
 60:  @ We're here because the fast path found a hit - now we have to track down exactly which word it was
 	@ r0 points to the start of the double word after the one that was tested
 	@ r5 has the 00/ff pattern for the first word, r6 has the chained value
+	.cfi_restore_state
 	cmp	r5, #0
 	itte	eq
 	moveq	r5, r6		@ the end is in the 2nd word
@@ -144,8 +148,14 @@ __memchr_arm:
 	addeq	r0,r0,#1
 
 61:
+	pop	{r4,r5,r6,r7}
+	.cfi_restore 7
+	.cfi_restore 6
+	.cfi_restore 5
+	.cfi_restore 4
+	.cfi_adjust_cfa_offset -16
 	subs	r0,r0,#1
-	epilogue 4 7
+	epilogue
 	.cfi_endproc
 	.cantunwind
 	.fnend
diff --git a/string/arm/strcmp.S b/string/arm/strcmp.S
index eafb9f6..b01c02e 100644
--- a/string/arm/strcmp.S
+++ b/string/arm/strcmp.S
@@ -13,7 +13,6 @@
    the compares.  */
 
 #include "../asmdefs.h"
-#include "../pacbti.h"
 
 /* Build Options:
    STRCMP_NO_PRECHECK: Don't run a quick pre-check of the first
@@ -106,7 +105,7 @@
 	.cfi_restore 5
 	.cfi_adjust_cfa_offset -16
 	sub	result, result, r1, lsr #24
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 #else
 	/* To use the big-endian trick we'd have to reverse all three words.
 	   that's slower than this approach.  */
@@ -129,7 +128,7 @@
 	.cfi_adjust_cfa_offset -16
 	sub	result, result, r1
 
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 #endif
 	.endm
 
@@ -140,14 +139,14 @@ L(strcmp_start_addr):
 #if STRCMP_NO_PRECHECK == 0
 L(fastpath_exit):
 	sub	r0, r2, r3
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 	nop
 #endif
 	.global __strcmp_arm
 	.type __strcmp_arm,%function
 	.align 0
 __strcmp_arm:
-	prologue savepac=HAVE_PAC_LEAF
+	prologue push_ip=HAVE_PAC_LEAF
 #if STRCMP_NO_PRECHECK == 0
 	ldrb	r2, [src1]
 	ldrb	r3, [src2]
@@ -158,17 +157,12 @@ __strcmp_arm:
 #endif
 	strd	r4, r5, [sp, #-16]!
 	.cfi_adjust_cfa_offset 16
-	.cfi_offset 5, -(12+PAC_CFI_ADJ_DEFAULT)
-	.cfi_offset 4, -(16+PAC_CFI_ADJ_DEFAULT)
+	.cfi_rel_offset 4, 0
+	.cfi_rel_offset 5, 4
 	orr	tmp1, src1, src2
 	strd	r6, r7, [sp, #8]
-#if HAVE_PAC_LEAF
-	.cfi_offset 6, -12
-	.cfi_offset 7, -8
-#else
-	.cfi_offset 6, -8
-	.cfi_offset 7, -4
-#endif /* HAVE_PAC_LEAF */
+	.cfi_rel_offset 6, 8
+	.cfi_rel_offset 7, 12
 	mvn	const_m1, #0
 	lsl	r2, tmp1, #29
 	cbz	r2, L(loop_aligned8)
@@ -339,7 +333,7 @@ L(misaligned_exit):
 	.cfi_restore 4
 	.cfi_adjust_cfa_offset -16
 
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 
 #if STRCMP_NO_PRECHECK == 0
 L(aligned_m1):
@@ -391,7 +385,7 @@ L(overlap3):
 	.cfi_restore 7
 	.cfi_adjust_cfa_offset -16
 	neg	result, result
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 6:
 	.cfi_restore_state
 	S2LO	data1, data1, #24
@@ -467,7 +461,7 @@ L(strcmp_done_equal):
 	.cfi_restore 6
 	.cfi_restore 7
 	.cfi_adjust_cfa_offset -16
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 
 L(strcmp_tail):
 	.cfi_restore_state
@@ -491,7 +485,7 @@ L(strcmp_tail):
 	.cfi_restore 7
 	.cfi_adjust_cfa_offset -16
 	sub	result, result, data2, lsr #24
-	epilogue savepac=HAVE_PAC_LEAF
+	epilogue push_ip=HAVE_PAC_LEAF
 
 END (__strcmp_arm)
 
diff --git a/string/arm/strlen-armv6t2.S b/string/arm/strlen-armv6t2.S
index 6e0352d..f06b238 100644
--- a/string/arm/strlen-armv6t2.S
+++ b/string/arm/strlen-armv6t2.S
@@ -14,7 +14,6 @@
  */
 
 #include "../asmdefs.h"
-#include "../pacbti.h"
 
 #ifdef __ARMEB__
 #define S2LO		lsl
@@ -47,7 +46,7 @@
 #define tmp2		r5
 
 ENTRY (__strlen_armv6t2)
-	prologue 4 5 savepac=HAVE_PAC_LEAF
+	prologue 4 5 push_ip=HAVE_PAC_LEAF
 	pld	[srcin, #0]
 	bic	src, srcin, #7
 	mvn	const_m1, #0
@@ -98,6 +97,7 @@ L(start_realigned):
 	beq	L(loop_aligned)
 
 L(null_found):
+	.cfi_remember_state
 	cmp	data1a, #0
 	itt	eq
 	addeq	result, result, #4
@@ -107,9 +107,10 @@ L(null_found):
 #endif
 	clz	data1a, data1a
 	add	result, result, data1a, lsr #3	/* Bits -> Bytes.  */
-	epilogue 4 5 savepac=HAVE_PAC_LEAF
+	epilogue 4 5 push_ip=HAVE_PAC_LEAF
 
 L(misaligned8):
+	.cfi_restore_state
 	ldrd	data1a, data1b, [src]
 	and	tmp2, tmp1, #3
 	rsb	result, tmp1, #0