powerpc: Rearrange cfi_offset calls

This patch rearranges cfi_offset() calls after the last store
so as to avoid extra DW_CFA_advance opcodes in unwind information.

7 files changed, 44 insertions, 34 deletions

diff --git a/ChangeLog b/ChangeLog
index 1064bc8c8e..440b0212d2 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,13 @@
+2016-03-11  Rajalakshmi Srinivasaraghavan  <raji@linux.vnet.ibm.com>
+
+	* sysdeps/powerpc/powerpc32/power4/memcmp.S (memcmp): Rearrange
+	cfi_offset calls.
+	* sysdeps/powerpc/powerpc32/power6/memcpy.S (memcpy): Likewise.
+	* sysdeps/powerpc/powerpc32/power7/memcmp.S (memcmp): Likewise.
+	* sysdeps/powerpc/powerpc64/power4/memcmp.S (memcmp): Likewise.
+	* sysdeps/powerpc/powerpc64/power7/memcmp.S (memcmp): Likewise.
+	* sysdeps/powerpc/powerpc64/power7/strstr.S (strstr): Likewise.
+
 2016-03-10  Carlos O'Donell  <carlos@redhat.com>
 
 	* timezone/checktab.awk: Removed.
diff --git a/sysdeps/powerpc/powerpc32/power4/memcmp.S b/sysdeps/powerpc/powerpc32/power4/memcmp.S
index 602a7957ad..9cb116e232 100644
--- a/sysdeps/powerpc/powerpc32/power4/memcmp.S
+++ b/sysdeps/powerpc/powerpc32/power4/memcmp.S
@@ -54,8 +54,8 @@ EALIGN (memcmp, 4, 0)
 	stwu	1, -64(r1)
 	cfi_adjust_cfa_offset(64)
 	stw	rWORD8, 48(r1)
-	cfi_offset(rWORD8, (48-64))
 	stw	rWORD7, 44(r1)
+	cfi_offset(rWORD8, (48-64))
 	cfi_offset(rWORD7, (44-64))
 	bne	L(unaligned)
 /* At this point we know both strings have the same alignment and the
@@ -747,18 +747,18 @@ L(unaligned):
    the actual start of rSTR2.  */
 	clrrwi	rSTR2, rSTR2, 2
 	stw	rWORD2_SHIFT, 28(r1)
-	cfi_offset(rWORD2_SHIFT, (28-64))
 /* Compute the left/right shift counts for the unaligned rSTR2,
    compensating for the logical (W aligned) start of rSTR1.  */
 	clrlwi	rSHL, rWORD8_SHIFT, 30
 	clrrwi	rSTR1, rSTR1, 2
 	stw	rWORD4_SHIFT, 24(r1)
-	cfi_offset(rWORD4_SHIFT, (24-64))
 	slwi	rSHL, rSHL, 3
 	cmplw	cr5, rWORD8_SHIFT, rSTR2
 	add	rN, rN, r12
 	slwi	rWORD6, r12, 3
 	stw	rWORD6_SHIFT, 20(r1)
+	cfi_offset(rWORD2_SHIFT, (28-64))
+	cfi_offset(rWORD4_SHIFT, (24-64))
 	cfi_offset(rWORD6_SHIFT, (20-64))
 	subfic	rSHR, rSHL, 32
 	srwi	r0, rN, 4	/* Divide by 16 */
@@ -852,15 +852,15 @@ L(duPs4):
 	.align	4
 L(Wunaligned):
 	stw	rWORD8_SHIFT, 32(r1)
-	cfi_offset(rWORD8_SHIFT, (32-64))
 	clrrwi	rSTR2, rSTR2, 2
 	stw	rWORD2_SHIFT, 28(r1)
-	cfi_offset(rWORD2_SHIFT, (28-64))
 	srwi	r0, rN, 4	/* Divide by 16 */
 	stw	rWORD4_SHIFT, 24(r1)
-	cfi_offset(rWORD4_SHIFT, (24-64))
 	andi.	r12, rN, 12	/* Get the W remainder */
 	stw	rWORD6_SHIFT, 20(r1)
+	cfi_offset(rWORD8_SHIFT, (32-64))
+	cfi_offset(rWORD2_SHIFT, (28-64))
+	cfi_offset(rWORD4_SHIFT, (24-64))
 	cfi_offset(rWORD6_SHIFT, (20-64))
 	slwi	rSHL, rSHL, 3
 #ifdef __LITTLE_ENDIAN__
diff --git a/sysdeps/powerpc/powerpc32/power6/memcpy.S b/sysdeps/powerpc/powerpc32/power6/memcpy.S
index 6dff0ed6df..ae796a26c1 100644
--- a/sysdeps/powerpc/powerpc32/power6/memcpy.S
+++ b/sysdeps/powerpc/powerpc32/power6/memcpy.S
@@ -46,8 +46,8 @@ EALIGN (memcpy, 5, 0)
     ble-   cr1,L(word_unaligned_short)	/* If move < 32 bytes.  */
     cmplw  cr6,10,11
     stw    31,24(1)
-    cfi_offset(31,(24-32))
     stw    30,20(1)
+    cfi_offset(31,(24-32))
     cfi_offset(30,(20-32))
     mr     30,3
     beq    .L0
diff --git a/sysdeps/powerpc/powerpc32/power7/memcmp.S b/sysdeps/powerpc/powerpc32/power7/memcmp.S
index 9c06a89491..13e8492106 100644
--- a/sysdeps/powerpc/powerpc32/power7/memcmp.S
+++ b/sysdeps/powerpc/powerpc32/power7/memcmp.S
@@ -54,8 +54,8 @@ EALIGN (memcmp, 4, 0)
 	stwu	1, -64(r1)
 	cfi_adjust_cfa_offset(64)
 	stw	rWORD8, 48(r1)
-	cfi_offset(rWORD8, (48-64))
 	stw	rWORD7, 44(r1)
+	cfi_offset(rWORD8, (48-64))
 	cfi_offset(rWORD7, (44-64))
 	bne	L(unaligned)
 /* At this point we know both strings have the same alignment and the
@@ -747,18 +747,18 @@ L(unaligned):
    the actual start of rSTR2.  */
 	clrrwi	rSTR2, rSTR2, 2
 	stw	rWORD2_SHIFT, 28(r1)
-	cfi_offset(rWORD2_SHIFT, (28-64))
 /* Compute the left/right shift counts for the unaligned rSTR2,
    compensating for the logical (W aligned) start of rSTR1.  */
 	clrlwi	rSHL, rWORD8_SHIFT, 30
 	clrrwi	rSTR1, rSTR1, 2
 	stw	rWORD4_SHIFT, 24(r1)
-	cfi_offset(rWORD4_SHIFT, (24-64))
 	slwi	rSHL, rSHL, 3
 	cmplw	cr5, rWORD8_SHIFT, rSTR2
 	add	rN, rN, r12
 	slwi	rWORD6, r12, 3
 	stw	rWORD6_SHIFT, 20(r1)
+	cfi_offset(rWORD2_SHIFT, (28-64))
+	cfi_offset(rWORD4_SHIFT, (24-64))
 	cfi_offset(rWORD6_SHIFT, (20-64))
 	subfic	rSHR, rSHL, 32
 	srwi	r0, rN, 4	/* Divide by 16 */
@@ -852,15 +852,15 @@ L(duPs4):
 	.align	4
 L(Wunaligned):
 	stw	rWORD8_SHIFT, 32(r1)
-	cfi_offset(rWORD8_SHIFT, (32-64))
 	clrrwi	rSTR2, rSTR2, 2
 	stw	rWORD2_SHIFT, 28(r1)
-	cfi_offset(rWORD2_SHIFT, (28-64))
 	srwi	r0, rN, 4	/* Divide by 16 */
 	stw	rWORD4_SHIFT, 24(r1)
-	cfi_offset(rWORD4_SHIFT, (24-64))
 	andi.	r12, rN, 12	/* Get the W remainder */
 	stw	rWORD6_SHIFT, 20(r1)
+	cfi_offset(rWORD8_SHIFT, (32-64))
+	cfi_offset(rWORD2_SHIFT, (28-64))
+	cfi_offset(rWORD4_SHIFT, (24-64))
 	cfi_offset(rWORD6_SHIFT, (20-64))
 	slwi	rSHL, rSHL, 3
 #ifdef __LITTLE_ENDIAN__
diff --git a/sysdeps/powerpc/powerpc64/power4/memcmp.S b/sysdeps/powerpc/powerpc64/power4/memcmp.S
index c1a77c64b0..65c659660d 100644
--- a/sysdeps/powerpc/powerpc64/power4/memcmp.S
+++ b/sysdeps/powerpc/powerpc64/power4/memcmp.S
@@ -52,8 +52,8 @@ EALIGN (memcmp, 4, 0)
    byte loop.  */
 	blt	cr1, L(bytealigned)
 	std	rWORD8, -8(r1)
-	cfi_offset(rWORD8, -8)
 	std	rWORD7, -16(r1)
+	cfi_offset(rWORD8, -8)
 	cfi_offset(rWORD7, -16)
 	bne	L(unaligned)
 /* At this point we know both strings have the same alignment and the
@@ -728,18 +728,18 @@ L(unaligned):
    the actual start of rSTR2.  */
 	clrrdi	rSTR2, rSTR2, 3
 	std	rWORD2_SHIFT, -48(r1)
-	cfi_offset(rWORD2_SHIFT, -48)
 /* Compute the left/right shift counts for the unaligned rSTR2,
    compensating for the logical (DW aligned) start of rSTR1.  */
 	clrldi	rSHL, rWORD8_SHIFT, 61
 	clrrdi	rSTR1, rSTR1, 3
 	std	rWORD4_SHIFT, -56(r1)
-	cfi_offset(rWORD4_SHIFT, -56)
 	sldi	rSHL, rSHL, 3
 	cmpld	cr5, rWORD8_SHIFT, rSTR2
 	add	rN, rN, r12
 	sldi	rWORD6, r12, 3
 	std	rWORD6_SHIFT, -64(r1)
+	cfi_offset(rWORD2_SHIFT, -48)
+	cfi_offset(rWORD4_SHIFT, -56)
 	cfi_offset(rWORD6_SHIFT, -64)
 	subfic	rSHR, rSHL, 64
 	srdi	r0, rN, 5	/* Divide by 32 */
@@ -833,15 +833,15 @@ L(duPs4):
 	.align	4
 L(DWunaligned):
 	std	rWORD8_SHIFT, -40(r1)
-	cfi_offset(rWORD8_SHIFT, -40)
 	clrrdi	rSTR2, rSTR2, 3
 	std	rWORD2_SHIFT, -48(r1)
-	cfi_offset(rWORD2_SHIFT, -48)
 	srdi	r0, rN, 5	/* Divide by 32 */
 	std	rWORD4_SHIFT, -56(r1)
-	cfi_offset(rWORD4_SHIFT, -56)
 	andi.	r12, rN, 24	/* Get the DW remainder */
 	std	rWORD6_SHIFT, -64(r1)
+	cfi_offset(rWORD8_SHIFT, -40)
+	cfi_offset(rWORD2_SHIFT, -48)
+	cfi_offset(rWORD4_SHIFT, -56)
 	cfi_offset(rWORD6_SHIFT, -64)
 	sldi	rSHL, rSHL, 3
 #ifdef __LITTLE_ENDIAN__
diff --git a/sysdeps/powerpc/powerpc64/power7/memcmp.S b/sysdeps/powerpc/powerpc64/power7/memcmp.S
index 4be29008c7..881c7d5838 100644
--- a/sysdeps/powerpc/powerpc64/power7/memcmp.S
+++ b/sysdeps/powerpc/powerpc64/power7/memcmp.S
@@ -82,17 +82,17 @@ EALIGN (memcmp, 4, 0)
    byte loop.  */
 	blt	cr1, L(bytealigned)
 	std	rWORD8, rWORD8SAVE(r1)
-	cfi_offset(rWORD8, rWORD8SAVE)
 	std	rWORD7, rWORD7SAVE(r1)
-	cfi_offset(rWORD7, rWORD7SAVE)
 	std	rOFF8, rOFF8SAVE(r1)
-	cfi_offset(rWORD7, rOFF8SAVE)
 	std	rOFF16, rOFF16SAVE(r1)
-	cfi_offset(rWORD7, rOFF16SAVE)
 	std	rOFF24, rOFF24SAVE(r1)
-	cfi_offset(rWORD7, rOFF24SAVE)
 	std	rOFF32, rOFF32SAVE(r1)
-	cfi_offset(rWORD7, rOFF32SAVE)
+	cfi_offset(rWORD8, rWORD8SAVE)
+	cfi_offset(rWORD7, rWORD7SAVE)
+	cfi_offset(rOFF8, rOFF8SAVE)
+	cfi_offset(rOFF16, rOFF16SAVE)
+	cfi_offset(rOFF24, rOFF24SAVE)
+	cfi_offset(rOFF32, rOFF32SAVE)
 
 	li	rOFF8,8
 	li	rOFF16,16
@@ -601,18 +601,18 @@ L(unaligned):
    the actual start of rSTR2.  */
 	clrrdi	rSTR2, rSTR2, 3
 	std	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
-	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
 /* Compute the left/right shift counts for the unaligned rSTR2,
    compensating for the logical (DW aligned) start of rSTR1.  */
 	clrldi	rSHL, rWORD8_SHIFT, 61
 	clrrdi	rSTR1, rSTR1, 3
 	std	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
-	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
 	sldi	rSHL, rSHL, 3
 	cmpld	cr5, rWORD8_SHIFT, rSTR2
 	add	rN, rN, r12
 	sldi	rWORD6, r12, 3
 	std	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
+	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
 	cfi_offset(rWORD6_SHIFT, rWORD6SHIFTSAVE)
 	subfic	rSHR, rSHL, 64
 	srdi	r0, rN, 5	/* Divide by 32 */
@@ -689,15 +689,15 @@ L(duPs4):
 	.align	4
 L(DWunaligned):
 	std	rWORD8_SHIFT, rWORD8SHIFTSAVE(r1)
-	cfi_offset(rWORD8_SHIFT, rWORD8SHIFTSAVE)
 	clrrdi	rSTR2, rSTR2, 3
 	std	rWORD2_SHIFT, rWORD2SHIFTSAVE(r1)
-	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
 	srdi	r0, rN, 5	/* Divide by 32 */
 	std	rWORD4_SHIFT, rWORD4SHIFTSAVE(r1)
-	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
 	andi.	r12, rN, 24	/* Get the DW remainder */
 	std	rWORD6_SHIFT, rWORD6SHIFTSAVE(r1)
+	cfi_offset(rWORD8_SHIFT, rWORD8SHIFTSAVE)
+	cfi_offset(rWORD2_SHIFT, rWORD2SHIFTSAVE)
+	cfi_offset(rWORD4_SHIFT, rWORD4SHIFTSAVE)
 	cfi_offset(rWORD6_SHIFT, rWORD6SHIFTSAVE)
 	sldi	rSHL, rSHL, 3
 	LD	rWORD6, 0, rSTR2
diff --git a/sysdeps/powerpc/powerpc64/power7/strstr.S b/sysdeps/powerpc/powerpc64/power7/strstr.S
index fefac1c9e5..0e18193e91 100644
--- a/sysdeps/powerpc/powerpc64/power7/strstr.S
+++ b/sysdeps/powerpc/powerpc64/power7/strstr.S
@@ -59,14 +59,14 @@ EALIGN (strstr, 4, 0)
 	CALL_MCOUNT 2
 	mflr	r0			/* Load link register LR to r0.  */
 	std	r31, -8(r1)		/* Save callers register r31.  */
-	cfi_offset(r31, -8)
 	std	r30, -16(r1)		/* Save callers register r30.  */
-	cfi_offset(r30, -16)
 	std	r29, -24(r1)		/* Save callers register r29.  */
-	cfi_offset(r29, -24)
 	std	r28, -32(r1)		/* Save callers register r28.  */
-	cfi_offset(r28, -32)
 	std	r0, 16(r1)		/* Store the link register.  */
+	cfi_offset(r31, -8)
+	cfi_offset(r30, -16)
+	cfi_offset(r28, -32)
+	cfi_offset(r29, -24)
 	cfi_offset(lr, 16)
 	stdu	r1, -FRAMESIZE(r1)	/* Create the stack frame.  */
 	cfi_adjust_cfa_offset(FRAMESIZE)