Add endbr64 where needed by inspection. Passes regresson tests.

ok jsing, and kind of tb an earlier version

3 files changed, 13 insertions, 0 deletions

diff --git a/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl b/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
index 2e71a7f03d..af78fff541 100644
--- a/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
+++ b/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
@@ -347,6 +347,7 @@ $code.=<<___;
 .type   MULADD_128x512,\@abi-omnipotent
 .align  16
 MULADD_128x512:
+        endbr64
 ___
         &MULSTEP_512([map("%r$_",(8..15))], "(+8*0)(%rcx)", "%rsi", "%rbp", "%rbx");
 $code.=<<___;
@@ -414,6 +415,7 @@ $code.=<<___;
 .type   mont_reduce,\@abi-omnipotent
 .align  16
 mont_reduce:
+        endbr64
 ___
 
 my $STACK_DEPTH         =  8;
@@ -676,6 +678,7 @@ $code.=<<___;
 .type   mont_mul_a3b,\@abi-omnipotent
 .align  16
 mont_mul_a3b:
+        endbr64
         #
         # multiply tmp = src1 * src2
         # For multiply: dst = rcx, src1 = rdi, src2 = rsi
@@ -1077,6 +1080,7 @@ $code.=<<___;
 .type   sqr_reduce,\@abi-omnipotent
 .align  16
 sqr_reduce:
+        endbr64
          mov    (+$pResult_offset+8)(%rsp), %rcx
 ___
         &SQR_512("%rsp+$tmp16_offset+8", "%rcx", [map("%r$_",(10..15,8..9))], "%rbx", "%rbp", "%rsi", "%rdi");
@@ -1106,6 +1110,7 @@ $code.=<<___;
 .globl  mod_exp_512
 .type   mod_exp_512,\@function,4
 mod_exp_512:
+        endbr64
          push   %rbp
          push   %rbx
          push   %r12
diff --git a/src/lib/libcrypto/bn/asm/x86_64-mont.pl b/src/lib/libcrypto/bn/asm/x86_64-mont.pl
index cae7309d5b..6f5ab331e2 100755
--- a/src/lib/libcrypto/bn/asm/x86_64-mont.pl
+++ b/src/lib/libcrypto/bn/asm/x86_64-mont.pl
@@ -63,6 +63,7 @@ $code=<<___;
 .type   bn_mul_mont,\@function,6
 .align  16
 bn_mul_mont:
+        endbr64
         test    \$3,${num}d
         jnz     .Lmul_enter
         cmp     \$8,${num}d
@@ -278,6 +279,7 @@ $code.=<<___;
 .align  16
 bn_mul4x_mont:
 .Lmul4x_enter:
+        endbr64
         push    %rbx
         push    %rbp
         push    %r12
@@ -705,6 +707,7 @@ $code.=<<___;
 .align  16
 bn_sqr4x_mont:
 .Lsqr4x_enter:
+        endbr64
         push    %rbx
         push    %rbp
         push    %r12
diff --git a/src/lib/libcrypto/bn/asm/x86_64-mont5.pl b/src/lib/libcrypto/bn/asm/x86_64-mont5.pl
index 7b9c6df273..3b3325a6cc 100755
--- a/src/lib/libcrypto/bn/asm/x86_64-mont5.pl
+++ b/src/lib/libcrypto/bn/asm/x86_64-mont5.pl
@@ -57,6 +57,7 @@ $code=<<___;
 .type   bn_mul_mont_gather5,\@function,6
 .align  64
 bn_mul_mont_gather5:
+        endbr64
         test    \$3,${num}d
         jnz     .Lmul_enter
         cmp     \$8,${num}d
@@ -387,6 +388,7 @@ $code.=<<___;
 .type   bn_mul4x_mont_gather5,\@function,6
 .align  16
 bn_mul4x_mont_gather5:
+        endbr64
 .Lmul4x_enter:
         mov     ${num}d,${num}d
         movd    `($win64?56:8)`(%rsp),%xmm5     # load 7th argument
@@ -925,6 +927,7 @@ $code.=<<___;
 .type   bn_scatter5,\@abi-omnipotent
 .align  16
 bn_scatter5:
+        endbr64
         cmp     \$0, $num
         jz      .Lscatter_epilogue
         lea     ($tbl,$idx,8),$tbl
@@ -943,6 +946,7 @@ bn_scatter5:
 .type   bn_gather5,\@abi-omnipotent
 .align  16
 bn_gather5:
+        endbr64
 .LSEH_begin_bn_gather5:                 # Win64 thing, but harmless in other cases
         # I can't trust assembler to use specific encoding:-(
         .byte   0x4c,0x8d,0x14,0x24                     # lea    (%rsp),%r10
@@ -1053,6 +1057,7 @@ $code.=<<___;
 .type   mul_handler,\@abi-omnipotent
 .align  16
 mul_handler:
+        endbr64
         push    %rsi
         push    %rdi
         push    %rbx