crypto/sha/asm/sha*-x86_64.pl: comply with Win64 ABI.

(cherry picked from commit 006784378db1a8957fcf587ee1deaa5cf2ae76c7)

diff --git a/crypto/sha/asm/sha1-x86_64.pl b/crypto/sha/asm/sha1-x86_64.pl
index ff94ac5f18b39edcca8a9fea997772c0a2688011..2c89b1feea36e08b9815f68ead0c11fc432763d5 100755 (executable)
--- a/crypto/sha/asm/sha1-x86_64.pl
+++ b/crypto/sha/asm/sha1-x86_64.pl
@@ -806,7 +806,7 @@ $code.=<<___;
        mov     %rdi,$ctx       # reassigned argument
        mov     %rsi,$inp       # reassigned argument
        mov     %rdx,$num       # reassigned argument
-       vzeroall
+       vzeroupper
 
        shl     \$6,$num
        add     $inp,$num
@@ -1096,7 +1096,7 @@ ___
        &Xtail_avx(\&body_20_39);
 
 $code.=<<___;
-       vzeroall
+       vzeroupper
 
        add     0($ctx),$A                      # update context
        add     4($ctx),@T[0]

diff --git a/crypto/sha/asm/sha512-x86_64.pl b/crypto/sha/asm/sha512-x86_64.pl
index d1d20914a59068961f79c0e8cf76c47f1f2c1d46..8070d09c94f2d59406b19de79bac122f8c957bf4 100755 (executable)
--- a/crypto/sha/asm/sha512-x86_64.pl
+++ b/crypto/sha/asm/sha512-x86_64.pl
@@ -948,7 +948,7 @@ ___
 $code.=<<___;
 .Lprologue_xop:
 
-       vzeroall
+       vzeroupper
        mov     $SZ*0($ctx),$A
        mov     $SZ*1($ctx),$B
        mov     $SZ*2($ctx),$C
@@ -1260,7 +1260,7 @@ $code.=<<___;
        jb      .Lloop_xop
 
        mov     $_rsp,%rsi
-       vzeroall
+       vzeroupper
 ___
 $code.=<<___ if ($win64);
        movaps  16*$SZ+32(%rsp),%xmm6
@@ -1324,7 +1324,7 @@ ___
 $code.=<<___;
 .Lprologue_avx:
 
-       vzeroall
+       vzeroupper
        mov     $SZ*0($ctx),$A
        mov     $SZ*1($ctx),$B
        mov     $SZ*2($ctx),$C
@@ -1568,7 +1568,7 @@ $code.=<<___;
        jb      .Lloop_avx
 
        mov     $_rsp,%rsi
-       vzeroall
+       vzeroupper
 ___
 $code.=<<___ if ($win64);
        movaps  16*$SZ+32(%rsp),%xmm6
@@ -1676,7 +1676,7 @@ ___
 $code.=<<___;
 .Lprologue_avx2:
 
-       vzeroall
+       vzeroupper
        sub     \$-16*$SZ,$inp          # inp++, size optimization
        mov     $SZ*0($ctx),$A
        mov     $inp,%r12               # borrow $T1
@@ -1942,7 +1942,7 @@ $code.=<<___;
 .Ldone_avx2:
        lea     ($Tbl),%rsp
        mov     $_rsp,%rsi
-       vzeroall
+       vzeroupper
 ___
 $code.=<<___ if ($win64);
        movaps  16*$SZ+32(%rsp),%xmm6