diff options
author | Andy Polyakov <appro@openssl.org> | 2017-02-13 22:34:51 +0100 |
---|---|---|
committer | Andy Polyakov <appro@openssl.org> | 2017-02-15 15:43:05 +0100 |
commit | 399976c7ba8dcc9eba3641e03d0bb41e4d137ed1 (patch) | |
tree | 73162c8a6b4f070c83754d1b366f13b149ff0100 /crypto | |
parent | ed874fac6399d5064d6eb8fe2022b918aeaf75af (diff) |
sha/asm/*-x86_64.pl: add CFI annotations.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Diffstat (limited to 'crypto')
-rw-r--r-- | crypto/sha/asm/sha1-mb-x86_64.pl | 46 | ||||
-rw-r--r-- | crypto/sha/asm/sha256-mb-x86_64.pl | 46 | ||||
-rwxr-xr-x | crypto/sha/asm/sha512-x86_64.pl | 92 |
3 files changed, 183 insertions, 1 deletions
diff --git a/crypto/sha/asm/sha1-mb-x86_64.pl b/crypto/sha/asm/sha1-mb-x86_64.pl index 2f6b35f355..56e15292a7 100644 --- a/crypto/sha/asm/sha1-mb-x86_64.pl +++ b/crypto/sha/asm/sha1-mb-x86_64.pl @@ -363,6 +363,7 @@ $code.=<<___; .type sha1_multi_block,\@function,3 .align 32 sha1_multi_block: +.cfi_startproc mov OPENSSL_ia32cap_P+4(%rip),%rcx bt \$61,%rcx # check SHA bit jc _shaext_shortcut @@ -373,8 +374,11 @@ $code.=<<___ if ($avx); ___ $code.=<<___; mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbx ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -393,6 +397,7 @@ $code.=<<___; sub \$`$REG_SZ*18`,%rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody: lea K_XX_XX(%rip),$Tbl lea `$REG_SZ*16`(%rsp),%rbx @@ -487,6 +492,7 @@ $code.=<<___; .Ldone: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 ___ $code.=<<___ if ($win64); movaps -0xb8(%rax),%xmm6 @@ -502,10 +508,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue: ret +.cfi_endproc .size sha1_multi_block,.-sha1_multi_block ___ {{{ @@ -517,10 +527,14 @@ $code.=<<___; .type sha1_multi_block_shaext,\@function,3 .align 32 sha1_multi_block_shaext: +.cfi_startproc _shaext_shortcut: mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -756,10 +770,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_shaext: ret +.cfi_endproc .size sha1_multi_block_shaext,.-sha1_multi_block_shaext ___ }}} @@ -1002,6 +1020,7 @@ $code.=<<___; .type sha1_multi_block_avx,\@function,3 .align 32 sha1_multi_block_avx: +.cfi_startproc _avx_shortcut: ___ $code.=<<___ if ($avx>1); @@ -1016,8 +1035,11 @@ $code.=<<___ if ($avx>1); ___ $code.=<<___; mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -1036,6 +1058,7 @@ $code.=<<___; sub \$`$REG_SZ*18`, %rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody_avx: lea K_XX_XX(%rip),$Tbl lea `$REG_SZ*16`(%rsp),%rbx @@ -1125,6 +1148,7 @@ $code.=<<___; .Ldone_avx: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1141,10 +1165,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx: ret +.cfi_endproc .size sha1_multi_block_avx,.-sha1_multi_block_avx ___ @@ -1164,14 +1192,22 @@ $code.=<<___; .type sha1_multi_block_avx2,\@function,3 .align 32 sha1_multi_block_avx2: +.cfi_startproc _avx2_shortcut: mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -1190,6 +1226,7 @@ $code.=<<___; sub \$`$REG_SZ*18`, %rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody_avx2: lea K_XX_XX(%rip),$Tbl shr \$1,$num @@ -1280,6 +1317,7 @@ $code.=<<___; .Ldone_avx2: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1296,14 +1334,22 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -48(%rax),%r15 +.cfi_restore %r15 mov -40(%rax),%r14 +.cfi_restore %r14 mov -32(%rax),%r13 +.cfi_restore %r13 mov -24(%rax),%r12 +.cfi_restore %r12 mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx2: ret +.cfi_endproc .size sha1_multi_block_avx2,.-sha1_multi_block_avx2 ___ } }}} diff --git a/crypto/sha/asm/sha256-mb-x86_64.pl b/crypto/sha/asm/sha256-mb-x86_64.pl index b8a77c7fce..24276f9e67 100644 --- a/crypto/sha/asm/sha256-mb-x86_64.pl +++ b/crypto/sha/asm/sha256-mb-x86_64.pl @@ -244,6 +244,7 @@ $code.=<<___; .type sha256_multi_block,\@function,3 .align 32 sha256_multi_block: +.cfi_startproc mov OPENSSL_ia32cap_P+4(%rip),%rcx bt \$61,%rcx # check SHA bit jc _shaext_shortcut @@ -254,8 +255,11 @@ $code.=<<___ if ($avx); ___ $code.=<<___; mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -274,6 +278,7 @@ $code.=<<___; sub \$`$REG_SZ*18`, %rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody: lea K256+128(%rip),$Tbl lea `$REG_SZ*16`(%rsp),%rbx @@ -391,6 +396,7 @@ $code.=<<___; .Ldone: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 ___ $code.=<<___ if ($win64); movaps -0xb8(%rax),%xmm6 @@ -406,10 +412,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue: ret +.cfi_endproc .size sha256_multi_block,.-sha256_multi_block ___ {{{ @@ -421,10 +431,14 @@ $code.=<<___; .type sha256_multi_block_shaext,\@function,3 .align 32 sha256_multi_block_shaext: +.cfi_startproc _shaext_shortcut: mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -758,10 +772,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_shaext: ret +.cfi_endproc .size sha256_multi_block_shaext,.-sha256_multi_block_shaext ___ }}} @@ -921,6 +939,7 @@ $code.=<<___; .type sha256_multi_block_avx,\@function,3 .align 32 sha256_multi_block_avx: +.cfi_startproc _avx_shortcut: ___ $code.=<<___ if ($avx>1); @@ -935,8 +954,11 @@ $code.=<<___ if ($avx>1); ___ $code.=<<___; mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -955,6 +977,7 @@ $code.=<<___; sub \$`$REG_SZ*18`, %rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody_avx: lea K256+128(%rip),$Tbl lea `$REG_SZ*16`(%rsp),%rbx @@ -1070,6 +1093,7 @@ $code.=<<___; .Ldone_avx: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1086,10 +1110,14 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx: ret +.cfi_endproc .size sha256_multi_block_avx,.-sha256_multi_block_avx ___ if ($avx>1) { @@ -1105,14 +1133,22 @@ $code.=<<___; .type sha256_multi_block_avx2,\@function,3 .align 32 sha256_multi_block_avx2: +.cfi_startproc _avx2_shortcut: mov %rsp,%rax +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 ___ $code.=<<___ if ($win64); lea -0xa8(%rsp),%rsp @@ -1131,6 +1167,7 @@ $code.=<<___; sub \$`$REG_SZ*18`, %rsp and \$-256,%rsp mov %rax,`$REG_SZ*17`(%rsp) # original %rsp +.cfi_cfa_expression %rsp+`$REG_SZ*17`,deref,+8 .Lbody_avx2: lea K256+128(%rip),$Tbl lea 0x80($ctx),$ctx # size optimization @@ -1246,6 +1283,7 @@ $code.=<<___; .Ldone_avx2: mov `$REG_SZ*17`(%rsp),%rax # original %rsp +.cfi_def_cfa %rax,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1262,14 +1300,22 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -48(%rax),%r15 +.cfi_restore %r15 mov -40(%rax),%r14 +.cfi_restore %r14 mov -32(%rax),%r13 +.cfi_restore %r13 mov -24(%rax),%r12 +.cfi_restore %r12 mov -16(%rax),%rbp +.cfi_restore %rbp mov -8(%rax),%rbx +.cfi_restore %rbx lea (%rax),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx2: ret +.cfi_endproc .size sha256_multi_block_avx2,.-sha256_multi_block_avx2 ___ } }}} diff --git a/crypto/sha/asm/sha512-x86_64.pl b/crypto/sha/asm/sha512-x86_64.pl index 7a8ed7c7f2..bdf1d28fda 100755 --- a/crypto/sha/asm/sha512-x86_64.pl +++ b/crypto/sha/asm/sha512-x86_64.pl @@ -176,7 +176,7 @@ $Tbl="%rbp"; $_ctx="16*$SZ+0*8(%rsp)"; $_inp="16*$SZ+1*8(%rsp)"; $_end="16*$SZ+2*8(%rsp)"; -$_rsp="16*$SZ+3*8(%rsp)"; +$_rsp="`16*$SZ+3*8`(%rsp)"; $framesz="16*$SZ+4*8"; @@ -269,6 +269,7 @@ $code=<<___; .type $func,\@function,3 .align 16 $func: +.cfi_startproc ___ $code.=<<___ if ($SZ==4 || $avx); lea OPENSSL_ia32cap_P(%rip),%r11 @@ -302,12 +303,19 @@ $code.=<<___ if ($SZ==4); ___ $code.=<<___; mov %rsp,%rax # copy %rsp +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 shl \$4,%rdx # num*16 sub \$$framesz,%rsp lea ($inp,%rdx,$SZ),%rdx # inp+num*16*$SZ @@ -316,6 +324,7 @@ $code.=<<___; mov $inp,$_inp # save inp, 2nd arh mov %rdx,$_end # save end pointer, "3rd" arg mov %rax,$_rsp # save copy of %rsp +.cfi_cfa_expression $_rsp,deref,+8 .Lprologue: mov $SZ*0($ctx),$A @@ -382,15 +391,24 @@ $code.=<<___; jb .Lloop mov $_rsp,%rsi +.cfi_def_cfa %rsi,8 mov -48(%rsi),%r15 +.cfi_restore %r15 mov -40(%rsi),%r14 +.cfi_restore %r14 mov -32(%rsi),%r13 +.cfi_restore %r13 mov -24(%rsi),%r12 +.cfi_restore %r12 mov -16(%rsi),%rbp +.cfi_restore %rbp mov -8(%rsi),%rbx +.cfi_restore %rbx lea (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lepilogue: ret +.cfi_endproc .size $func,.-$func ___ @@ -760,14 +778,22 @@ $code.=<<___; .type ${func}_ssse3,\@function,3 .align 64 ${func}_ssse3: +.cfi_startproc .Lssse3_shortcut: mov %rsp,%rax # copy %rsp +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 shl \$4,%rdx # num*16 sub \$`$framesz+$win64*16*4`,%rsp lea ($inp,%rdx,$SZ),%rdx # inp+num*16*$SZ @@ -776,6 +802,7 @@ ${func}_ssse3: mov $inp,$_inp # save inp, 2nd arh mov %rdx,$_end # save end pointer, "3rd" arg mov %rax,$_rsp # save copy of %rsp +.cfi_cfa_expression $_rsp,deref,+8 ___ $code.=<<___ if ($win64); movaps %xmm6,16*$SZ+32(%rsp) @@ -1074,6 +1101,7 @@ $code.=<<___; jb .Lloop_ssse3 mov $_rsp,%rsi +.cfi_def_cfa %rsi,8 ___ $code.=<<___ if ($win64); movaps 16*$SZ+32(%rsp),%xmm6 @@ -1083,14 +1111,22 @@ $code.=<<___ if ($win64); ___ $code.=<<___; mov -48(%rsi),%r15 +.cfi_restore %r15 mov -40(%rsi),%r14 +.cfi_restore %r14 mov -32(%rsi),%r13 +.cfi_restore %r13 mov -24(%rsi),%r12 +.cfi_restore %r12 mov -16(%rsi),%rbp +.cfi_restore %rbp mov -8(%rsi),%rbx +.cfi_restore %rbx lea (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_ssse3: ret +.cfi_endproc .size ${func}_ssse3,.-${func}_ssse3 ___ } @@ -1104,14 +1140,22 @@ $code.=<<___; .type ${func}_xop,\@function,3 .align 64 ${func}_xop: +.cfi_startproc .Lxop_shortcut: mov %rsp,%rax # copy %rsp +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 shl \$4,%rdx # num*16 sub \$`$framesz+$win64*16*($SZ==4?4:6)`,%rsp lea ($inp,%rdx,$SZ),%rdx # inp+num*16*$SZ @@ -1120,6 +1164,7 @@ ${func}_xop: mov $inp,$_inp # save inp, 2nd arh mov %rdx,$_end # save end pointer, "3rd" arg mov %rax,$_rsp # save copy of %rsp +.cfi_cfa_expression $_rsp,deref,+8 ___ $code.=<<___ if ($win64); movaps %xmm6,16*$SZ+32(%rsp) @@ -1446,6 +1491,7 @@ $code.=<<___; jb .Lloop_xop mov $_rsp,%rsi +.cfi_def_cfa %rsi,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1460,14 +1506,22 @@ $code.=<<___ if ($win64 && $SZ>4); ___ $code.=<<___; mov -48(%rsi),%r15 +.cfi_restore %r15 mov -40(%rsi),%r14 +.cfi_restore %r14 mov -32(%rsi),%r13 +.cfi_restore %r13 mov -24(%rsi),%r12 +.cfi_restore %r12 mov -16(%rsi),%rbp +.cfi_restore %rbp mov -8(%rsi),%rbx +.cfi_restore %rbx lea (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_xop: ret +.cfi_endproc .size ${func}_xop,.-${func}_xop ___ } @@ -1480,14 +1534,22 @@ $code.=<<___; .type ${func}_avx,\@function,3 .align 64 ${func}_avx: +.cfi_startproc .Lavx_shortcut: mov %rsp,%rax # copy %rsp +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 shl \$4,%rdx # num*16 sub \$`$framesz+$win64*16*($SZ==4?4:6)`,%rsp lea ($inp,%rdx,$SZ),%rdx # inp+num*16*$SZ @@ -1496,6 +1558,7 @@ ${func}_avx: mov $inp,$_inp # save inp, 2nd arh mov %rdx,$_end # save end pointer, "3rd" arg mov %rax,$_rsp # save copy of %rsp +.cfi_cfa_expression $_rsp,deref,+8 ___ $code.=<<___ if ($win64); movaps %xmm6,16*$SZ+32(%rsp) @@ -1754,6 +1817,7 @@ $code.=<<___; jb .Lloop_avx mov $_rsp,%rsi +.cfi_def_cfa %rsi,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -1768,14 +1832,22 @@ $code.=<<___ if ($win64 && $SZ>4); ___ $code.=<<___; mov -48(%rsi),%r15 +.cfi_restore %r15 mov -40(%rsi),%r14 +.cfi_restore %r14 mov -32(%rsi),%r13 +.cfi_restore %r13 mov -24(%rsi),%r12 +.cfi_restore %r12 mov -16(%rsi),%rbp +.cfi_restore %rbp mov -8(%rsi),%rbx +.cfi_restore %rbx lea (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx: ret +.cfi_endproc .size ${func}_avx,.-${func}_avx ___ @@ -1831,14 +1903,22 @@ $code.=<<___; .type ${func}_avx2,\@function,3 .align 64 ${func}_avx2: +.cfi_startproc .Lavx2_shortcut: mov %rsp,%rax # copy %rsp +.cfi_def_cfa_register %rax push %rbx +.cfi_push %rbx push %rbp +.cfi_push %rbp push %r12 +.cfi_push %r12 push %r13 +.cfi_push %r13 push %r14 +.cfi_push %r14 push %r15 +.cfi_push %r15 sub \$`2*$SZ*$rounds+4*8+$win64*16*($SZ==4?4:6)`,%rsp shl \$4,%rdx # num*16 and \$-256*$SZ,%rsp # align stack frame @@ -1848,6 +1928,7 @@ ${func}_avx2: mov $inp,$_inp # save inp, 2nd arh mov %rdx,$_end # save end pointer, "3rd" arg mov %rax,$_rsp # save copy of %rsp +.cfi_cfa_expression $_rsp,deref,+8 ___ $code.=<<___ if ($win64); movaps %xmm6,16*$SZ+32(%rsp) @@ -2128,6 +2209,7 @@ $code.=<<___; .Ldone_avx2: lea ($Tbl),%rsp mov $_rsp,%rsi +.cfi_def_cfa %rsi,8 vzeroupper ___ $code.=<<___ if ($win64); @@ -2142,14 +2224,22 @@ $code.=<<___ if ($win64 && $SZ>4); ___ $code.=<<___; mov -48(%rsi),%r15 +.cfi_restore %r15 mov -40(%rsi),%r14 +.cfi_restore %r14 mov -32(%rsi),%r13 +.cfi_restore %r13 mov -24(%rsi),%r12 +.cfi_restore %r12 mov -16(%rsi),%rbp +.cfi_restore %rbp mov -8(%rsi),%rbx +.cfi_restore %rbx lea (%rsi),%rsp +.cfi_def_cfa_register %rsp .Lepilogue_avx2: ret +.cfi_endproc .size ${func}_avx2,.-${func}_avx2 ___ }} |