X-Git-Url: https://git.openssl.org/?a=blobdiff_plain;f=crypto%2Fx86_64cpuid.pl;h=f228c24ba17c694d081fb24836b996ee0f3aeab4;hb=2ca8bbe55084253eb954caa440e469961b981f79;hp=3a1adeecccde79af8d0fb6e0e5870f34aa7469f7;hpb=c5cd28bd64fa2b02f29e74486539e4b2f6741114;p=openssl.git diff --git a/crypto/x86_64cpuid.pl b/crypto/x86_64cpuid.pl index 3a1adeeccc..f228c24ba1 100644 --- a/crypto/x86_64cpuid.pl +++ b/crypto/x86_64cpuid.pl @@ -1,4 +1,11 @@ -#!/usr/bin/env perl +#! /usr/bin/env perl +# Copyright 2005-2016 The OpenSSL Project Authors. All Rights Reserved. +# +# Licensed under the OpenSSL license (the "License"). You may not use +# this file except in compliance with the License. You can obtain a copy +# in the file LICENSE in the source distribution or at +# https://www.openssl.org/source/license.html + $flavour = shift; $output = shift; @@ -11,7 +18,8 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1; ( $xlate="${dir}perlasm/x86_64-xlate.pl" and -f $xlate) or die "can't locate x86_64-xlate.pl"; -open STDOUT,"| \"$^X\" $xlate $flavour $output"; +open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\""; +*STDOUT=*OUT; ($arg1,$arg2,$arg3,$arg4)=$win64?("%rcx","%rdx","%r8", "%r9") : # Win64 order ("%rdi","%rsi","%rdx","%rcx"); # Unix order @@ -55,10 +63,12 @@ OPENSSL_rdtsc: .type OPENSSL_ia32_cpuid,\@function,1 .align 16 OPENSSL_ia32_cpuid: +.cfi_startproc mov %rbx,%r8 # save %rbx +.cfi_register %rbx,%r8 xor %eax,%eax - mov %eax,8(%rdi) # clear 3rd word + mov %eax,8(%rdi) # clear extended feature flags cpuid mov %eax,%r11d # max value for standard query level @@ -126,14 +136,6 @@ OPENSSL_ia32_cpuid: shr \$14,%r10d and \$0xfff,%r10d # number of cores -1 per L1D - cmp \$7,%r11d - jb .Lnocacheinfo - - mov \$7,%eax - xor %ecx,%ecx - cpuid - mov %ebx,8(%rdi) - .Lnocacheinfo: mov \$1,%eax cpuid @@ -143,8 +145,19 @@ OPENSSL_ia32_cpuid: or \$0x40000000,%edx # set reserved bit#30 on Intel CPUs and \$15,%ah cmp \$15,%ah # examine Family ID - jne .Lnotintel + jne .LnotP4 or \$0x00100000,%edx # set reserved bit#20 to engage RC4_CHAR +.LnotP4: + cmp \$6,%ah + jne .Lnotintel + and \$0x0fff0ff0,%eax + cmp \$0x00050670,%eax # Knights Landing + je .Lknights + cmp \$0x00080650,%eax # Knights Mill (according to sde) + jne .Lnotintel +.Lknights: + and \$0xfbffffff,%ecx # clear XSAVE flag to mimic Silvermont + .Lnotintel: bt \$28,%edx # test hyper-threading bit jnc .Lgeneric @@ -163,23 +176,46 @@ OPENSSL_ia32_cpuid: or %ecx,%r9d # merge AMD XOP flag mov %edx,%r10d # %r9d:%r10d is copy of %ecx:%edx + + cmp \$7,%r11d + jb .Lno_extended_info + mov \$7,%eax + xor %ecx,%ecx + cpuid + bt \$26,%r9d # check XSAVE bit, cleared on Knights + jc .Lnotknights + and \$0xfff7ffff,%ebx # clear ADCX/ADOX flag +.Lnotknights: + mov %ebx,8(%rdi) # save extended feature flags +.Lno_extended_info: + bt \$27,%r9d # check OSXSAVE bit jnc .Lclear_avx xor %ecx,%ecx # XCR0 .byte 0x0f,0x01,0xd0 # xgetbv + and \$0xe6,%eax # isolate XMM, YMM and ZMM state support + cmp \$0xe6,%eax + je .Ldone + andl \$0xfffeffff,8(%rdi) # clear AVX512F, ~(1<<16) + # note that we don't touch other AVX512 + # extensions, because they can be used + # with YMM (without opmasking though) and \$6,%eax # isolate XMM and YMM state support cmp \$6,%eax je .Ldone .Lclear_avx: mov \$0xefffe7ff,%eax # ~(1<<28|1<<12|1<<11) and %eax,%r9d # clear AVX, FMA and AMD XOP bits - andl \$0xffffffdf,8(%rdi) # cleax AVX2, ~(1<<5) + mov \$0x3fdeffdf,%eax # ~(1<<31|1<<30|1<<21|1<<16|1<<5) + and %eax,8(%rdi) # clear AVX2 and AVX512* bits .Ldone: shl \$32,%r9 mov %r10d,%eax mov %r8,%rbx # restore %rbx +.cfi_restore %rbx or %r9,%rax ret +.cfi_endproc .size OPENSSL_ia32_cpuid,.-OPENSSL_ia32_cpuid .globl OPENSSL_cleanse @@ -216,6 +252,28 @@ OPENSSL_cleanse: jne .Little ret .size OPENSSL_cleanse,.-OPENSSL_cleanse + +.globl CRYPTO_memcmp +.type CRYPTO_memcmp,\@abi-omnipotent +.align 16 +CRYPTO_memcmp: + xor %rax,%rax + xor %r10,%r10 + cmp \$0,$arg3 + je .Lno_data +.Loop_cmp: + mov ($arg1),%r10b + lea 1($arg1),$arg1 + xor ($arg2),%r10b + lea 1($arg2),$arg2 + or %r10b,%al + dec $arg3 + jnz .Loop_cmp + neg %rax + shr \$63,%rax +.Lno_data: + ret +.size CRYPTO_memcmp,.-CRYPTO_memcmp ___ print<<___ if (!$win64); @@ -363,21 +421,67 @@ OPENSSL_instrument_bus2: ___ } +sub gen_random { +my $rdop = shift; print<<___; -.globl OPENSSL_ia32_rdrand -.type OPENSSL_ia32_rdrand,\@abi-omnipotent +.globl OPENSSL_ia32_${rdop} +.type OPENSSL_ia32_${rdop},\@abi-omnipotent .align 16 -OPENSSL_ia32_rdrand: +OPENSSL_ia32_${rdop}: mov \$8,%ecx -.Loop_rdrand: - rdrand %rax - jc .Lbreak_rdrand - loop .Loop_rdrand -.Lbreak_rdrand: +.Loop_${rdop}: + ${rdop} %rax + jc .Lbreak_${rdop} + loop .Loop_${rdop} +.Lbreak_${rdop}: cmp \$0,%rax cmove %rcx,%rax ret -.size OPENSSL_ia32_rdrand,.-OPENSSL_ia32_rdrand +.size OPENSSL_ia32_${rdop},.-OPENSSL_ia32_${rdop} + +.globl OPENSSL_ia32_${rdop}_bytes +.type OPENSSL_ia32_${rdop}_bytes,\@abi-omnipotent +.align 16 +OPENSSL_ia32_${rdop}_bytes: + xor %rax, %rax # return value + cmp \$0,$arg2 + je .Ldone_${rdop}_bytes + + mov \$8,%r11 +.Loop_${rdop}_bytes: + ${rdop} %r10 + jc .Lbreak_${rdop}_bytes + dec %r11 + jnz .Loop_${rdop}_bytes + jmp .Ldone_${rdop}_bytes + +.align 16 +.Lbreak_${rdop}_bytes: + cmp \$8,$arg2 + jb .Ltail_${rdop}_bytes + mov %r10,($arg1) + lea 8($arg1),$arg1 + add \$8,%rax + sub \$8,$arg2 + jz .Ldone_${rdop}_bytes + mov \$8,%r11 + jmp .Loop_${rdop}_bytes + +.align 16 +.Ltail_${rdop}_bytes: + mov %r10b,($arg1) + lea 1($arg1),$arg1 + inc %rax + shr \$8,%r8 + dec $arg2 + jnz .Ltail_${rdop}_bytes + +.Ldone_${rdop}_bytes: + ret +.size OPENSSL_ia32_${rdop}_bytes,.-OPENSSL_ia32_${rdop}_bytes ___ +} +gen_random("rdrand"); +gen_random("rdseed"); close STDOUT; # flush