ARM64 assembly pack: make it Windows-friendly.

[openssl.git] / crypto / bn / asm / armv8-mont.pl

diff --git a/crypto/bn/asm/armv8-mont.pl b/crypto/bn/asm/armv8-mont.pl
index 9a8ff6b92f07630199e5947bb1fab73f386ab8fd..c755555d884da06c2f98cb09faa57fccc0f110de 100755 (executable)
--- a/crypto/bn/asm/armv8-mont.pl
+++ b/crypto/bn/asm/armv8-mont.pl
@@ -197,7 +197,7 @@ bn_mul_mont:
        mul     $nlo,$nj,$m1            // np[j]*m1
        adds    $lo1,$lo1,$lo0
        umulh   $nhi,$nj,$m1
-       str     $lo1,[$tp,#-16]         // tp[j-1]
+       stur    $lo1,[$tp,#-16]         // tp[j-1]
        cbnz    $j,.Linner
 
 .Linner_skip:
@@ -253,13 +253,13 @@ bn_mul_mont:
        csel    $nj,$tj,$aj,lo          // did it borrow?
        ldr     $tj,[$tp],#8
        ldr     $aj,[$rp],#8
-       str     xzr,[$tp,#-16]          // wipe tp
-       str     $nj,[$rp,#-16]
+       stur    xzr,[$tp,#-16]          // wipe tp
+       stur    $nj,[$rp,#-16]
        cbnz    $num,.Lcond_copy
 
        csel    $nj,$tj,$aj,lo
-       str     xzr,[$tp,#-8]           // wipe tp
-       str     $nj,[$rp,#-8]
+       stur    xzr,[$tp,#-8]           // wipe tp
+       stur    $nj,[$rp,#-8]
 
        ldp     x19,x20,[x29,#16]
        mov     sp,x29
@@ -287,6 +287,7 @@ __bn_sqr8x_mont:
        cmp     $ap,$bp
        b.ne    __bn_mul4x_mont
 .Lsqr8x_mont:
+       .inst   0xd503233f              // paciasp
        stp     x29,x30,[sp,#-128]!
        add     x29,sp,#0
        stp     x19,x20,[sp,#16]
@@ -595,7 +596,7 @@ __bn_sqr8x_mont:
        ldp     $a4,$a5,[$tp,#8*4]
        ldp     $a6,$a7,[$tp,#8*6]
        adds    $acc0,$acc0,$a0
-       ldr     $n0,[$rp,#-8*8]
+       ldur    $n0,[$rp,#-8*8]
        adcs    $acc1,$acc1,$a1
        ldp     $a0,$a1,[$ap,#8*0]
        adcs    $acc2,$acc2,$a2
@@ -793,7 +794,7 @@ $code.=<<___;
        //adc   $carry,xzr,xzr          // moved below
        cbz     $cnt,.Lsqr8x8_post_condition
 
-       ldr     $n0,[$tp,#-8*8]
+       ldur    $n0,[$tp,#-8*8]
        ldp     $a0,$a1,[$np,#8*0]
        ldp     $a2,$a3,[$np,#8*2]
        ldp     $a4,$a5,[$np,#8*4]
@@ -851,7 +852,7 @@ $code.=<<___;
        ldp     $a6,$a7,[$tp,#8*6]
        cbz     $cnt,.Lsqr8x_tail_break
 
-       ldr     $n0,[$rp,#-8*8]
+       ldur    $n0,[$rp,#-8*8]
        adds    $acc0,$acc0,$a0
        adcs    $acc1,$acc1,$a1
        ldp     $a0,$a1,[$np,#8*0]
@@ -1040,6 +1041,7 @@ $code.=<<___;
        ldp     x25,x26,[x29,#64]
        ldp     x27,x28,[x29,#80]
        ldr     x29,[sp],#128
+       .inst   0xd50323bf              // autiasp
        ret
 .size  __bn_sqr8x_mont,.-__bn_sqr8x_mont
 ___
@@ -1063,6 +1065,7 @@ $code.=<<___;
 .type  __bn_mul4x_mont,%function
 .align 5
 __bn_mul4x_mont:
+       .inst   0xd503233f              // paciasp
        stp     x29,x30,[sp,#-128]!
        add     x29,sp,#0
        stp     x19,x20,[sp,#16]
@@ -1496,6 +1499,7 @@ __bn_mul4x_mont:
        ldp     x25,x26,[x29,#64]
        ldp     x27,x28,[x29,#80]
        ldr     x29,[sp],#128
+       .inst   0xd50323bf              // autiasp
        ret
 .size  __bn_mul4x_mont,.-__bn_mul4x_mont
 ___