factotum/libressl-portable
1
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
libressl-portable/patches/asm/0001-Revert-Add-endbr64-where-needed-by-inspection.-Passe.patch
Brent Cook 98a5122bc6 portable asm generator patches for portable 
This reverses changes from the 3.8.x upstream that are causing issues on
a few different platforms, including macOS x64, Windows, and older
Linux/FreeBSD.
2023-10-30 22:55:29 -05:00

1056 lines
28 KiB
Diff
Raw Permalink Blame History

From 5f7fafff280e6e17d7fafea20fda99762d74581f Mon Sep 17 00:00:00 2001
From: Brent Cook <busterb@gmail.com>
Date: Mon, 30 Oct 2023 21:59:32 -0500
Subject: [PATCH 1/4] Revert "Add endbr64 where needed by inspection. Passes
regresson tests."
This reverts commit e2118101444d3cf3cab87297b363cecd5357ae89.
---
src/lib/libcrypto/aes/asm/aes-x86_64.pl | 13 -------------
src/lib/libcrypto/aes/asm/aesni-sha1-x86_64.pl | 4 ----
src/lib/libcrypto/aes/asm/aesni-x86_64.pl | 15 ---------------
src/lib/libcrypto/aes/asm/bsaes-x86_64.pl | 14 --------------
src/lib/libcrypto/aes/asm/vpaes-x86_64.pl | 14 --------------
src/lib/libcrypto/bn/arch/amd64/bignum_add.S | 1 -
src/lib/libcrypto/bn/arch/amd64/bignum_cmadd.S | 1 -
src/lib/libcrypto/bn/arch/amd64/bignum_cmul.S | 1 -
src/lib/libcrypto/bn/arch/amd64/bignum_mul.S | 1 -
.../libcrypto/bn/arch/amd64/bignum_mul_4_8_alt.S | 1 -
.../libcrypto/bn/arch/amd64/bignum_mul_8_16_alt.S | 1 -
src/lib/libcrypto/bn/arch/amd64/bignum_sqr.S | 1 -
.../libcrypto/bn/arch/amd64/bignum_sqr_4_8_alt.S | 1 -
.../libcrypto/bn/arch/amd64/bignum_sqr_8_16_alt.S | 1 -
src/lib/libcrypto/bn/arch/amd64/bignum_sub.S | 1 -
src/lib/libcrypto/bn/arch/amd64/word_clz.S | 1 -
src/lib/libcrypto/bn/asm/modexp512-x86_64.pl | 5 -----
src/lib/libcrypto/bn/asm/x86_64-mont.pl | 3 ---
src/lib/libcrypto/bn/asm/x86_64-mont5.pl | 5 -----
src/lib/libcrypto/camellia/asm/cmll-x86_64.pl | 8 --------
src/lib/libcrypto/md5/asm/md5-x86_64.pl | 1 -
src/lib/libcrypto/modes/asm/ghash-x86_64.pl | 4 ----
src/lib/libcrypto/rc4/asm/rc4-md5-x86_64.pl | 6 ++----
src/lib/libcrypto/rc4/asm/rc4-x86_64.pl | 7 ++-----
src/lib/libcrypto/sha/asm/sha1-x86_64.pl | 4 ----
src/lib/libcrypto/sha/asm/sha512-x86_64.pl | 1 -
src/lib/libcrypto/whrlpool/asm/wp-x86_64.pl | 1 -
src/lib/libcrypto/x86_64cpuid.pl | 2 --
28 files changed, 4 insertions(+), 114 deletions(-)
diff --git a/src/lib/libcrypto/aes/asm/aes-x86_64.pl b/src/lib/libcrypto/aes/asm/aes-x86_64.pl
index 78ba20ca5..b7399b552 100755
--- a/src/lib/libcrypto/aes/asm/aes-x86_64.pl
+++ b/src/lib/libcrypto/aes/asm/aes-x86_64.pl
@@ -318,7 +318,6 @@ $code.=<<___;
.type _x86_64_AES_encrypt,\@abi-omnipotent
.align 16
_x86_64_AES_encrypt:
- endbr64
xor 0($key),$s0 # xor with key
xor 4($key),$s1
xor 8($key),$s2
@@ -549,7 +548,6 @@ $code.=<<___;
.type _x86_64_AES_encrypt_compact,\@abi-omnipotent
.align 16
_x86_64_AES_encrypt_compact:
- endbr64
lea 128($sbox),$inp # size optimization
mov 0-128($inp),$acc1 # prefetch Te4
mov 32-128($inp),$acc2
@@ -595,7 +593,6 @@ $code.=<<___;
.hidden asm_AES_encrypt
asm_AES_encrypt:
AES_encrypt:
- endbr64
push %rbx
push %rbp
push %r12
@@ -887,7 +884,6 @@ $code.=<<___;
.type _x86_64_AES_decrypt,\@abi-omnipotent
.align 16
_x86_64_AES_decrypt:
- endbr64
xor 0($key),$s0 # xor with key
xor 4($key),$s1
xor 8($key),$s2
@@ -1142,7 +1138,6 @@ $code.=<<___;
.type _x86_64_AES_decrypt_compact,\@abi-omnipotent
.align 16
_x86_64_AES_decrypt_compact:
- endbr64
lea 128($sbox),$inp # size optimization
mov 0-128($inp),$acc1 # prefetch Td4
mov 32-128($inp),$acc2
@@ -1197,7 +1192,6 @@ $code.=<<___;
.hidden asm_AES_decrypt
asm_AES_decrypt:
AES_decrypt:
- endbr64
push %rbx
push %rbp
push %r12
@@ -1297,7 +1291,6 @@ $code.=<<___;
.type AES_set_encrypt_key,\@function,3
.align 16
AES_set_encrypt_key:
- endbr64
push %rbx
push %rbp
push %r12 # redundant, but allows to share
@@ -1323,7 +1316,6 @@ AES_set_encrypt_key:
.type _x86_64_AES_set_encrypt_key,\@abi-omnipotent
.align 16
_x86_64_AES_set_encrypt_key:
- endbr64
mov %esi,%ecx # %ecx=bits
mov %rdi,%rsi # %rsi=userKey
mov %rdx,%rdi # %rdi=key
@@ -1569,7 +1561,6 @@ $code.=<<___;
.type AES_set_decrypt_key,\@function,3
.align 16
AES_set_decrypt_key:
- endbr64
push %rbx
push %rbp
push %r12
@@ -1669,7 +1660,6 @@ $code.=<<___;
.hidden asm_AES_cbc_encrypt
asm_AES_cbc_encrypt:
AES_cbc_encrypt:
- endbr64
cmp \$0,%rdx # check length
je .Lcbc_epilogue
pushfq
@@ -2561,7 +2551,6 @@ $code.=<<___;
.type block_se_handler,\@abi-omnipotent
.align 16
block_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2620,7 +2609,6 @@ block_se_handler:
.type key_se_handler,\@abi-omnipotent
.align 16
key_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2678,7 +2666,6 @@ key_se_handler:
.type cbc_se_handler,\@abi-omnipotent
.align 16
cbc_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/aes/asm/aesni-sha1-x86_64.pl b/src/lib/libcrypto/aes/asm/aesni-sha1-x86_64.pl
index 879d16793..bafa906a0 100644
--- a/src/lib/libcrypto/aes/asm/aesni-sha1-x86_64.pl
+++ b/src/lib/libcrypto/aes/asm/aesni-sha1-x86_64.pl
@@ -89,7 +89,6 @@ $code.=<<___;
.type aesni_cbc_sha1_enc,\@abi-omnipotent
.align 16
aesni_cbc_sha1_enc:
- endbr64
# caller should check for SSSE3 and AES-NI bits
mov OPENSSL_ia32cap_P+0(%rip),%r10d
mov OPENSSL_ia32cap_P+4(%rip),%r11d
@@ -133,7 +132,6 @@ $code.=<<___;
.type aesni_cbc_sha1_enc_ssse3,\@function,6
.align 16
aesni_cbc_sha1_enc_ssse3:
- endbr64
mov `($win64?56:8)`(%rsp),$inp # load 7th argument
#shr \$6,$len # debugging artefact
#jz .Lepilogue_ssse3 # debugging artefact
@@ -652,7 +650,6 @@ $code.=<<___;
.type aesni_cbc_sha1_enc_avx,\@function,6
.align 16
aesni_cbc_sha1_enc_avx:
- endbr64
mov `($win64?56:8)`(%rsp),$inp # load 7th argument
#shr \$6,$len # debugging artefact
#jz .Lepilogue_avx # debugging artefact
@@ -1103,7 +1100,6 @@ $code.=<<___;
.type ssse3_handler,\@abi-omnipotent
.align 16
ssse3_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/aes/asm/aesni-x86_64.pl b/src/lib/libcrypto/aes/asm/aesni-x86_64.pl
index 07d40a84a..e662fbc7c 100644
--- a/src/lib/libcrypto/aes/asm/aesni-x86_64.pl
+++ b/src/lib/libcrypto/aes/asm/aesni-x86_64.pl
@@ -242,7 +242,6 @@ $code.=<<___;
.type ${PREFIX}_encrypt,\@abi-omnipotent
.align 16
${PREFIX}_encrypt:
- endbr64
movups ($inp),$inout0 # load input
mov 240($key),$rounds # key->rounds
___
@@ -256,7 +255,6 @@ $code.=<<___;
.type ${PREFIX}_decrypt,\@abi-omnipotent
.align 16
${PREFIX}_decrypt:
- endbr64
movups ($inp),$inout0 # load input
mov 240($key),$rounds # key->rounds
___
@@ -286,7 +284,6 @@ $code.=<<___;
.type _aesni_${dir}rypt3,\@abi-omnipotent
.align 16
_aesni_${dir}rypt3:
- endbr64
$movkey ($key),$rndkey0
shr \$1,$rounds
$movkey 16($key),$rndkey1
@@ -331,7 +328,6 @@ $code.=<<___;
.type _aesni_${dir}rypt4,\@abi-omnipotent
.align 16
_aesni_${dir}rypt4:
- endbr64
$movkey ($key),$rndkey0
shr \$1,$rounds
$movkey 16($key),$rndkey1
@@ -377,7 +373,6 @@ $code.=<<___;
.type _aesni_${dir}rypt6,\@abi-omnipotent
.align 16
_aesni_${dir}rypt6:
- endbr64
$movkey ($key),$rndkey0
shr \$1,$rounds
$movkey 16($key),$rndkey1
@@ -442,7 +437,6 @@ $code.=<<___;
.type _aesni_${dir}rypt8,\@abi-omnipotent
.align 16
_aesni_${dir}rypt8:
- endbr64
$movkey ($key),$rndkey0
shr \$1,$rounds
$movkey 16($key),$rndkey1
@@ -531,7 +525,6 @@ $code.=<<___;
.type aesni_ecb_encrypt,\@function,5
.align 16
aesni_ecb_encrypt:
- endbr64
and \$-16,$len
jz .Lecb_ret
@@ -837,7 +830,6 @@ $code.=<<___;
.type aesni_ccm64_encrypt_blocks,\@function,6
.align 16
aesni_ccm64_encrypt_blocks:
- endbr64
___
$code.=<<___ if ($win64);
lea -0x58(%rsp),%rsp
@@ -2487,7 +2479,6 @@ $code.=<<___;
.type ${PREFIX}_set_decrypt_key,\@abi-omnipotent
.align 16
${PREFIX}_set_decrypt_key:
- endbr64
sub \$8,%rsp
call __aesni_set_encrypt_key
shl \$4,$bits # rounds-1 after _aesni_set_encrypt_key
@@ -2538,7 +2529,6 @@ $code.=<<___;
.type ${PREFIX}_set_encrypt_key,\@abi-omnipotent
.align 16
${PREFIX}_set_encrypt_key:
- endbr64
__aesni_set_encrypt_key:
sub \$8,%rsp
mov \$-1,%rax
@@ -2760,7 +2750,6 @@ $code.=<<___ if ($PREFIX eq "aesni");
.type ecb_se_handler,\@abi-omnipotent
.align 16
ecb_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2780,7 +2769,6 @@ ecb_se_handler:
.type ccm64_se_handler,\@abi-omnipotent
.align 16
ccm64_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2822,7 +2810,6 @@ ccm64_se_handler:
.type ctr32_se_handler,\@abi-omnipotent
.align 16
ctr32_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2858,7 +2845,6 @@ ctr32_se_handler:
.type xts_se_handler,\@abi-omnipotent
.align 16
xts_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
@@ -2900,7 +2886,6 @@ $code.=<<___;
.type cbc_se_handler,\@abi-omnipotent
.align 16
cbc_se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/aes/asm/bsaes-x86_64.pl b/src/lib/libcrypto/aes/asm/bsaes-x86_64.pl
index 7098ba27f..20e9e1f71 100644
--- a/src/lib/libcrypto/aes/asm/bsaes-x86_64.pl
+++ b/src/lib/libcrypto/aes/asm/bsaes-x86_64.pl
@@ -813,7 +813,6 @@ $code.=<<___;
.type _bsaes_encrypt8,\@abi-omnipotent
.align 64
_bsaes_encrypt8:
- endbr64
lea .LBS0(%rip), $const # constants table
movdqa ($key), @XMM[9] # round 0 key
@@ -878,7 +877,6 @@ $code.=<<___;
.type _bsaes_decrypt8,\@abi-omnipotent
.align 64
_bsaes_decrypt8:
- endbr64
lea .LBS0(%rip), $const # constants table
movdqa ($key), @XMM[9] # round 0 key
@@ -970,7 +968,6 @@ $code.=<<___;
.type _bsaes_key_convert,\@abi-omnipotent
.align 16
_bsaes_key_convert:
- endbr64
lea .Lmasks(%rip), $const
movdqu ($inp), %xmm7 # load round 0 key
lea 0x10($inp), $inp
@@ -1060,7 +1057,6 @@ $code.=<<___;
.type bsaes_enc_key_convert,\@function,2
.align 16
bsaes_enc_key_convert:
- endbr64
mov 240($inp),%r10d # pass rounds
mov $inp,%rcx # pass key
mov $out,%rax # pass key schedule
@@ -1075,7 +1071,6 @@ bsaes_enc_key_convert:
.align 16
bsaes_encrypt_128:
.Lenc128_loop:
- endbr64
movdqu 0x00($inp), @XMM[0] # load input
movdqu 0x10($inp), @XMM[1]
movdqu 0x20($inp), @XMM[2]
@@ -1108,7 +1103,6 @@ bsaes_encrypt_128:
.type bsaes_dec_key_convert,\@function,2
.align 16
bsaes_dec_key_convert:
- endbr64
mov 240($inp),%r10d # pass rounds
mov $inp,%rcx # pass key
mov $out,%rax # pass key schedule
@@ -1123,7 +1117,6 @@ bsaes_dec_key_convert:
.type bsaes_decrypt_128,\@function,4
.align 16
bsaes_decrypt_128:
- endbr64
.Ldec128_loop:
movdqu 0x00($inp), @XMM[0] # load input
movdqu 0x10($inp), @XMM[1]
@@ -1169,7 +1162,6 @@ $code.=<<___;
.type bsaes_ecb_encrypt_blocks,\@abi-omnipotent
.align 16
bsaes_ecb_encrypt_blocks:
- endbr64
mov %rsp, %rax
.Lecb_enc_prologue:
push %rbp
@@ -1371,7 +1363,6 @@ $code.=<<___;
.type bsaes_ecb_decrypt_blocks,\@abi-omnipotent
.align 16
bsaes_ecb_decrypt_blocks:
- endbr64
mov %rsp, %rax
.Lecb_dec_prologue:
push %rbp
@@ -1577,7 +1568,6 @@ $code.=<<___;
.type bsaes_cbc_encrypt,\@abi-omnipotent
.align 16
bsaes_cbc_encrypt:
- endbr64
___
$code.=<<___ if ($win64);
mov 48(%rsp),$arg6 # pull direction flag
@@ -1865,7 +1855,6 @@ $code.=<<___;
.type bsaes_ctr32_encrypt_blocks,\@abi-omnipotent
.align 16
bsaes_ctr32_encrypt_blocks:
- endbr64
mov %rsp, %rax
.Lctr_enc_prologue:
push %rbp
@@ -2107,7 +2096,6 @@ $code.=<<___;
.type bsaes_xts_encrypt,\@abi-omnipotent
.align 16
bsaes_xts_encrypt:
- endbr64
mov %rsp, %rax
.Lxts_enc_prologue:
push %rbp
@@ -2489,7 +2477,6 @@ $code.=<<___;
.type bsaes_xts_decrypt,\@abi-omnipotent
.align 16
bsaes_xts_decrypt:
- endbr64
mov %rsp, %rax
.Lxts_dec_prologue:
push %rbp
@@ -2966,7 +2953,6 @@ $code.=<<___;
.type se_handler,\@abi-omnipotent
.align 16
se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/aes/asm/vpaes-x86_64.pl b/src/lib/libcrypto/aes/asm/vpaes-x86_64.pl
index 8ff8d8602..3ffb1a303 100644
--- a/src/lib/libcrypto/aes/asm/vpaes-x86_64.pl
+++ b/src/lib/libcrypto/aes/asm/vpaes-x86_64.pl
@@ -82,7 +82,6 @@ $code.=<<___;
.type _vpaes_encrypt_core,\@abi-omnipotent
.align 16
_vpaes_encrypt_core:
- endbr64
mov %rdx, %r9
mov \$16, %r11
mov 240(%rdx),%eax
@@ -173,7 +172,6 @@ _vpaes_encrypt_core:
.type _vpaes_decrypt_core,\@abi-omnipotent
.align 16
_vpaes_decrypt_core:
- endbr64
mov %rdx, %r9 # load key
mov 240(%rdx),%eax
movdqa %xmm9, %xmm1
@@ -281,7 +279,6 @@ _vpaes_decrypt_core:
.type _vpaes_schedule_core,\@abi-omnipotent
.align 16
_vpaes_schedule_core:
- endbr64
# rdi = key
# rsi = size in bits
# rdx = buffer
@@ -467,7 +464,6 @@ _vpaes_schedule_core:
.type _vpaes_schedule_192_smear,\@abi-omnipotent
.align 16
_vpaes_schedule_192_smear:
- endbr64
pshufd \$0x80, %xmm6, %xmm0 # d c 0 0 -> c 0 0 0
pxor %xmm0, %xmm6 # -> c+d c 0 0
pshufd \$0xFE, %xmm7, %xmm0 # b a _ _ -> b b b a
@@ -499,7 +495,6 @@ _vpaes_schedule_192_smear:
.type _vpaes_schedule_round,\@abi-omnipotent
.align 16
_vpaes_schedule_round:
- endbr64
# extract rcon from xmm8
pxor %xmm1, %xmm1
palignr \$15, %xmm8, %xmm1
@@ -567,7 +562,6 @@ _vpaes_schedule_low_round:
.type _vpaes_schedule_transform,\@abi-omnipotent
.align 16
_vpaes_schedule_transform:
- endbr64
movdqa %xmm9, %xmm1
pandn %xmm0, %xmm1
psrld \$4, %xmm1
@@ -606,7 +600,6 @@ _vpaes_schedule_transform:
.type _vpaes_schedule_mangle,\@abi-omnipotent
.align 16
_vpaes_schedule_mangle:
- endbr64
movdqa %xmm0, %xmm4 # save xmm0 for later
movdqa .Lk_mc_forward(%rip),%xmm5
test %rcx, %rcx
@@ -680,7 +673,6 @@ _vpaes_schedule_mangle:
.type ${PREFIX}_set_encrypt_key,\@function,3
.align 16
${PREFIX}_set_encrypt_key:
- endbr64
___
$code.=<<___ if ($win64);
lea -0xb8(%rsp),%rsp
@@ -729,7 +721,6 @@ $code.=<<___;
.type ${PREFIX}_set_decrypt_key,\@function,3
.align 16
${PREFIX}_set_decrypt_key:
- endbr64
___
$code.=<<___ if ($win64);
lea -0xb8(%rsp),%rsp
@@ -783,7 +774,6 @@ $code.=<<___;
.type ${PREFIX}_encrypt,\@function,3
.align 16
${PREFIX}_encrypt:
- endbr64
___
$code.=<<___ if ($win64);
lea -0xb8(%rsp),%rsp
@@ -827,7 +817,6 @@ $code.=<<___;
.type ${PREFIX}_decrypt,\@function,3
.align 16
${PREFIX}_decrypt:
- endbr64
___
$code.=<<___ if ($win64);
lea -0xb8(%rsp),%rsp
@@ -877,7 +866,6 @@ $code.=<<___;
.type ${PREFIX}_cbc_encrypt,\@function,6
.align 16
${PREFIX}_cbc_encrypt:
- endbr64
xchg $key,$len
___
($len,$key)=($key,$len);
@@ -961,7 +949,6 @@ $code.=<<___;
.type _vpaes_preheat,\@abi-omnipotent
.align 16
_vpaes_preheat:
- endbr64
lea .Lk_s0F(%rip), %r10
movdqa -0x20(%r10), %xmm10 # .Lk_inv
movdqa -0x10(%r10), %xmm11 # .Lk_inv+16
@@ -1092,7 +1079,6 @@ $code.=<<___;
.type se_handler,\@abi-omnipotent
.align 16
se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_add.S b/src/lib/libcrypto/bn/arch/amd64/bignum_add.S
index 06298ca69..d56fa5e3a 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_add.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_add.S
@@ -49,7 +49,6 @@
S2N_BN_SYMBOL(bignum_add):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_cmadd.S b/src/lib/libcrypto/bn/arch/amd64/bignum_cmadd.S
index 5ad712749..1dc1e5870 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_cmadd.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_cmadd.S
@@ -54,7 +54,6 @@
S2N_BN_SYMBOL(bignum_cmadd):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_cmul.S b/src/lib/libcrypto/bn/arch/amd64/bignum_cmul.S
index 9199c8f48..c1a23ccea 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_cmul.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_cmul.S
@@ -51,7 +51,6 @@
S2N_BN_SYMBOL(bignum_cmul):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_mul.S b/src/lib/libcrypto/bn/arch/amd64/bignum_mul.S
index 2d7ed1909..42ac988a1 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_mul.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_mul.S
@@ -59,7 +59,6 @@
S2N_BN_SYMBOL(bignum_mul):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_mul_4_8_alt.S b/src/lib/libcrypto/bn/arch/amd64/bignum_mul_4_8_alt.S
index f02b09b28..3b7848b28 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_mul_4_8_alt.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_mul_4_8_alt.S
@@ -72,7 +72,6 @@
adc h, rdx
S2N_BN_SYMBOL(bignum_mul_4_8_alt):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_mul_8_16_alt.S b/src/lib/libcrypto/bn/arch/amd64/bignum_mul_8_16_alt.S
index 97be83e1f..1be37840d 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_mul_8_16_alt.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_mul_8_16_alt.S
@@ -72,7 +72,6 @@
adc h, rdx
S2N_BN_SYMBOL(bignum_mul_8_16_alt):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr.S b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr.S
index c4a0cabf3..2e05b9c17 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr.S
@@ -62,7 +62,6 @@
#define llshort ebp
S2N_BN_SYMBOL(bignum_sqr):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_4_8_alt.S b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_4_8_alt.S
index b228414dc..a635177c6 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_4_8_alt.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_4_8_alt.S
@@ -71,7 +71,6 @@
adc c, 0
S2N_BN_SYMBOL(bignum_sqr_4_8_alt):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_8_16_alt.S b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_8_16_alt.S
index 04efeec7e..f698202d2 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_8_16_alt.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_sqr_8_16_alt.S
@@ -103,7 +103,6 @@
adc c, 0
S2N_BN_SYMBOL(bignum_sqr_8_16_alt):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/bignum_sub.S b/src/lib/libcrypto/bn/arch/amd64/bignum_sub.S
index 11a9bd7ed..f8e1fe35a 100644
--- a/src/lib/libcrypto/bn/arch/amd64/bignum_sub.S
+++ b/src/lib/libcrypto/bn/arch/amd64/bignum_sub.S
@@ -49,7 +49,6 @@
S2N_BN_SYMBOL(bignum_sub):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/arch/amd64/word_clz.S b/src/lib/libcrypto/bn/arch/amd64/word_clz.S
index 464a9d90f..025e98f9c 100644
--- a/src/lib/libcrypto/bn/arch/amd64/word_clz.S
+++ b/src/lib/libcrypto/bn/arch/amd64/word_clz.S
@@ -30,7 +30,6 @@
.text
S2N_BN_SYMBOL(word_clz):
- endbr64
#if WINDOWS_ABI
push rdi
diff --git a/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl b/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
index af78fff54..2e71a7f03 100644
--- a/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
+++ b/src/lib/libcrypto/bn/asm/modexp512-x86_64.pl
@@ -347,7 +347,6 @@ $code.=<<___;
.type MULADD_128x512,\@abi-omnipotent
.align 16
MULADD_128x512:
- endbr64
___
&MULSTEP_512([map("%r$_",(8..15))], "(+8*0)(%rcx)", "%rsi", "%rbp", "%rbx");
$code.=<<___;
@@ -415,7 +414,6 @@ $code.=<<___;
.type mont_reduce,\@abi-omnipotent
.align 16
mont_reduce:
- endbr64
___
my $STACK_DEPTH = 8;
@@ -678,7 +676,6 @@ $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
@@ -1080,7 +1077,6 @@ $code.=<<___;
.type sqr_reduce,\@abi-omnipotent
.align 16
sqr_reduce:
- endbr64
mov (+$pResult_offset+8)(%rsp), %rcx
___
&SQR_512("%rsp+$tmp16_offset+8", "%rcx", [map("%r$_",(10..15,8..9))], "%rbx", "%rbp", "%rsi", "%rdi");
@@ -1110,7 +1106,6 @@ $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 6f5ab331e..cae7309d5 100755
--- a/src/lib/libcrypto/bn/asm/x86_64-mont.pl
+++ b/src/lib/libcrypto/bn/asm/x86_64-mont.pl
@@ -63,7 +63,6 @@ $code=<<___;
.type bn_mul_mont,\@function,6
.align 16
bn_mul_mont:
- endbr64
test \$3,${num}d
jnz .Lmul_enter
cmp \$8,${num}d
@@ -279,7 +278,6 @@ $code.=<<___;
.align 16
bn_mul4x_mont:
.Lmul4x_enter:
- endbr64
push %rbx
push %rbp
push %r12
@@ -707,7 +705,6 @@ $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 3b3325a6c..7b9c6df27 100755
--- a/src/lib/libcrypto/bn/asm/x86_64-mont5.pl
+++ b/src/lib/libcrypto/bn/asm/x86_64-mont5.pl
@@ -57,7 +57,6 @@ $code=<<___;
.type bn_mul_mont_gather5,\@function,6
.align 64
bn_mul_mont_gather5:
- endbr64
test \$3,${num}d
jnz .Lmul_enter
cmp \$8,${num}d
@@ -388,7 +387,6 @@ $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
@@ -927,7 +925,6 @@ $code.=<<___;
.type bn_scatter5,\@abi-omnipotent
.align 16
bn_scatter5:
- endbr64
cmp \$0, $num
jz .Lscatter_epilogue
lea ($tbl,$idx,8),$tbl
@@ -946,7 +943,6 @@ 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
@@ -1057,7 +1053,6 @@ $code.=<<___;
.type mul_handler,\@abi-omnipotent
.align 16
mul_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/camellia/asm/cmll-x86_64.pl b/src/lib/libcrypto/camellia/asm/cmll-x86_64.pl
index 3ceed3e89..586e5d6e9 100644
--- a/src/lib/libcrypto/camellia/asm/cmll-x86_64.pl
+++ b/src/lib/libcrypto/camellia/asm/cmll-x86_64.pl
@@ -116,7 +116,6 @@ $code=<<___;
.type Camellia_EncryptBlock,\@abi-omnipotent
.align 16
Camellia_EncryptBlock:
- endbr64
movl \$128,%eax
subl $arg0d,%eax
movl \$3,$arg0d
@@ -129,7 +128,6 @@ Camellia_EncryptBlock:
.align 16
.Lenc_rounds:
Camellia_EncryptBlock_Rounds:
- endbr64
push %rbx
push %rbp
push %r13
@@ -178,7 +176,6 @@ Camellia_EncryptBlock_Rounds:
.type _x86_64_Camellia_encrypt,\@abi-omnipotent
.align 16
_x86_64_Camellia_encrypt:
- endbr64
xor 0($key),@S[1]
xor 4($key),@S[0] # ^=key[0-3]
xor 8($key),@S[3]
@@ -229,7 +226,6 @@ $code.=<<___;
.type Camellia_DecryptBlock,\@abi-omnipotent
.align 16
Camellia_DecryptBlock:
- endbr64
movl \$128,%eax
subl $arg0d,%eax
movl \$3,$arg0d
@@ -242,7 +238,6 @@ Camellia_DecryptBlock:
.align 16
.Ldec_rounds:
Camellia_DecryptBlock_Rounds:
- endbr64
push %rbx
push %rbp
push %r13
@@ -291,7 +286,6 @@ Camellia_DecryptBlock_Rounds:
.type _x86_64_Camellia_decrypt,\@abi-omnipotent
.align 16
_x86_64_Camellia_decrypt:
- endbr64
xor 0($key),@S[1]
xor 4($key),@S[0] # ^=key[0-3]
xor 8($key),@S[3]
@@ -406,7 +400,6 @@ $code.=<<___;
.type Camellia_Ekeygen,\@function,3
.align 16
Camellia_Ekeygen:
- endbr64
push %rbx
push %rbp
push %r13
@@ -637,7 +630,6 @@ $code.=<<___;
.type Camellia_cbc_encrypt,\@function,6
.align 16
Camellia_cbc_encrypt:
- endbr64
cmp \$0,%rdx
je .Lcbc_abort
push %rbx
diff --git a/src/lib/libcrypto/md5/asm/md5-x86_64.pl b/src/lib/libcrypto/md5/asm/md5-x86_64.pl
index 06d69094f..c902a1b53 100755
--- a/src/lib/libcrypto/md5/asm/md5-x86_64.pl
+++ b/src/lib/libcrypto/md5/asm/md5-x86_64.pl
@@ -128,7 +128,6 @@ $code .= <<EOF;
.globl md5_block_asm_data_order
.type md5_block_asm_data_order,\@function,3
md5_block_asm_data_order:
- endbr64
push %rbp
push %rbx
push %r12
diff --git a/src/lib/libcrypto/modes/asm/ghash-x86_64.pl b/src/lib/libcrypto/modes/asm/ghash-x86_64.pl
index 9ce0c3814..71d0822ac 100644
--- a/src/lib/libcrypto/modes/asm/ghash-x86_64.pl
+++ b/src/lib/libcrypto/modes/asm/ghash-x86_64.pl
@@ -412,7 +412,6 @@ $code.=<<___;
.type gcm_init_clmul,\@abi-omnipotent
.align 16
gcm_init_clmul:
- endbr64
movdqu ($Xip),$Hkey
pshufd \$0b01001110,$Hkey,$Hkey # dword swap
@@ -450,7 +449,6 @@ $code.=<<___;
.type gcm_gmult_clmul,\@abi-omnipotent
.align 16
gcm_gmult_clmul:
- endbr64
movdqu ($Xip),$Xi
movdqa .Lbswap_mask(%rip),$T3
movdqu ($Htbl),$Hkey
@@ -478,7 +476,6 @@ $code.=<<___;
.type gcm_ghash_clmul,\@abi-omnipotent
.align 16
gcm_ghash_clmul:
- endbr64
___
$code.=<<___ if ($win64);
.LSEH_begin_gcm_ghash_clmul:
@@ -689,7 +686,6 @@ $code.=<<___;
.type se_handler,\@abi-omnipotent
.align 16
se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/rc4/asm/rc4-md5-x86_64.pl b/src/lib/libcrypto/rc4/asm/rc4-md5-x86_64.pl
index 3190e6a8e..c65a2c751 100644
--- a/src/lib/libcrypto/rc4/asm/rc4-md5-x86_64.pl
+++ b/src/lib/libcrypto/rc4/asm/rc4-md5-x86_64.pl
@@ -38,7 +38,7 @@ my ($rc4,$md5)=(1,1); # what to generate?
my $D="#" if (!$md5); # if set to "#", MD5 is stitched into RC4(),
# but its result is discarded. Idea here is
# to be able to use 'openssl speed rc4' for
- # benchmarking the stitched subroutine...
+ # benchmarking the stitched subroutine...
my $flavour = shift;
my $output = shift;
@@ -109,7 +109,6 @@ $code.=<<___;
.globl $func
.type $func,\@function,$nargs
$func:
- endbr64
cmp \$0,$len
je .Labort
push %rbx
@@ -405,7 +404,7 @@ $code.=<<___ if ($rc4 && (!$md5 || $D));
and \$63,$len # remaining bytes
jnz .Loop1
jmp .Ldone
-
+
.align 16
.Loop1:
add $TX[0]#b,$YY#b
@@ -454,7 +453,6 @@ $code.=<<___;
.type RC4_set_key,\@function,3
.align 16
RC4_set_key:
- endbr64
lea 8($dat),$dat
lea ($inp,$len),$inp
neg $len
diff --git a/src/lib/libcrypto/rc4/asm/rc4-x86_64.pl b/src/lib/libcrypto/rc4/asm/rc4-x86_64.pl
index 0472acce8..f678daaac 100755
--- a/src/lib/libcrypto/rc4/asm/rc4-x86_64.pl
+++ b/src/lib/libcrypto/rc4/asm/rc4-x86_64.pl
@@ -41,7 +41,7 @@
# April 2005
#
-# P4 EM64T core appears to be "allergic" to 64-bit inc/dec. Replacing
+# P4 EM64T core appears to be "allergic" to 64-bit inc/dec. Replacing
# those with add/sub results in 50% performance improvement of folded
# loop...
@@ -127,9 +127,7 @@ $code=<<___;
.globl RC4
.type RC4,\@function,4
.align 16
-RC4:
- endbr64
- or $len,$len
+RC4: or $len,$len
jne .Lentry
ret
.Lentry:
@@ -435,7 +433,6 @@ $code.=<<___;
.type RC4_set_key,\@function,3
.align 16
RC4_set_key:
- endbr64
lea 8($dat),$dat
lea ($inp,$len),$inp
neg $len
diff --git a/src/lib/libcrypto/sha/asm/sha1-x86_64.pl b/src/lib/libcrypto/sha/asm/sha1-x86_64.pl
index e15ff47f8..43eee73c4 100755
--- a/src/lib/libcrypto/sha/asm/sha1-x86_64.pl
+++ b/src/lib/libcrypto/sha/asm/sha1-x86_64.pl
@@ -222,7 +222,6 @@ $code.=<<___;
.type sha1_block_data_order,\@function,3
.align 16
sha1_block_data_order:
- endbr64
mov OPENSSL_ia32cap_P+0(%rip),%r9d
mov OPENSSL_ia32cap_P+4(%rip),%r8d
test \$IA32CAP_MASK1_SSSE3,%r8d # check SSSE3 bit
@@ -310,7 +309,6 @@ $code.=<<___;
.align 16
sha1_block_data_order_ssse3:
_ssse3_shortcut:
- endbr64
push %rbx
push %rbp
push %r12
@@ -731,7 +729,6 @@ $code.=<<___;
.align 16
sha1_block_data_order_avx:
_avx_shortcut:
- endbr64
push %rbx
push %rbp
push %r12
@@ -1102,7 +1099,6 @@ $code.=<<___;
.type se_handler,\@abi-omnipotent
.align 16
se_handler:
- endbr64
push %rsi
push %rdi
push %rbx
diff --git a/src/lib/libcrypto/sha/asm/sha512-x86_64.pl b/src/lib/libcrypto/sha/asm/sha512-x86_64.pl
index 120693fee..0517eab66 100755
--- a/src/lib/libcrypto/sha/asm/sha512-x86_64.pl
+++ b/src/lib/libcrypto/sha/asm/sha512-x86_64.pl
@@ -175,7 +175,6 @@ $code=<<___;
.type $func,\@function,4
.align 16
$func:
- endbr64
push %rbx
push %rbp
push %r12
diff --git a/src/lib/libcrypto/whrlpool/asm/wp-x86_64.pl b/src/lib/libcrypto/whrlpool/asm/wp-x86_64.pl
index 7958f6d28..de5d3acfb 100644
--- a/src/lib/libcrypto/whrlpool/asm/wp-x86_64.pl
+++ b/src/lib/libcrypto/whrlpool/asm/wp-x86_64.pl
@@ -57,7 +57,6 @@ $code=<<___;
.type $func,\@function,3
.align 16
$func:
- endbr64
push %rbx
push %rbp
push %r12
diff --git a/src/lib/libcrypto/x86_64cpuid.pl b/src/lib/libcrypto/x86_64cpuid.pl
index dc56732a2..1b67d1110 100644
--- a/src/lib/libcrypto/x86_64cpuid.pl
+++ b/src/lib/libcrypto/x86_64cpuid.pl
@@ -18,7 +18,6 @@ print<<___;
.extern OPENSSL_cpuid_setup
.hidden OPENSSL_cpuid_setup
.section .init
- endbr64
call OPENSSL_cpuid_setup
.extern OPENSSL_ia32cap_P
@@ -30,7 +29,6 @@ print<<___;
.type OPENSSL_ia32_cpuid,\@abi-omnipotent
.align 16
OPENSSL_ia32_cpuid:
- endbr64
mov %rbx,%r8 # save %rbx
xor %eax,%eax
--
2.42.0
Reference in New Issue View Git Blame Copy Permalink