diff options
author | Andy Polyakov <appro@openssl.org> | 2007-09-14 21:06:14 +0000 |
---|---|---|
committer | Andy Polyakov <appro@openssl.org> | 2007-09-14 21:06:14 +0000 |
commit | 563d3e5948a17328819f90b2ab20477256ada1f4 (patch) | |
tree | edf9dfa26ed81b62b25b0ebcd548932a142406e4 /crypto | |
parent | 1187ee7dad8a3809192d24e6f1c69718767caa88 (diff) |
Engage new x86 assembler modules.
Diffstat (limited to 'crypto')
-rw-r--r-- | crypto/sha/Makefile | 18 | ||||
-rw-r--r-- | crypto/sha/asm/sha512-586.pl | 14 |
2 files changed, 22 insertions, 10 deletions
diff --git a/crypto/sha/Makefile b/crypto/sha/Makefile index 1231f1da38..e02d9f1dc3 100644 --- a/crypto/sha/Makefile +++ b/crypto/sha/Makefile @@ -45,18 +45,24 @@ lib: $(LIBOBJ) # ELF sx86-elf.s: asm/sha1-586.pl ../perlasm/x86asm.pl (cd asm; $(PERL) sha1-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@) -s512sse2-elf.s: asm/sha512-sse2.pl ../perlasm/x86asm.pl - (cd asm; $(PERL) sha512-sse2.pl elf $(CFLAGS) $(PROCESSOR) > ../$@) +sha256x86-elf.s: asm/sha256-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha256-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@) +sha512x86-elf.s: asm/sha512-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha512-586.pl elf $(CFLAGS) $(PROCESSOR) > ../$@) # COFF sx86-cof.s: asm/sha1-586.pl ../perlasm/x86asm.pl (cd asm; $(PERL) sha1-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@) -s512sse2-cof.s: asm/sha512-sse2.pl ../perlasm/x86asm.pl - (cd asm; $(PERL) sha512-sse2.pl coff $(CFLAGS) $(PROCESSOR) > ../$@) +sha256x86-cof.s: asm/sha256-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha256-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@) +sha512x86-cof.s: asm/sha512-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha512-586.pl coff $(CFLAGS) $(PROCESSOR) > ../$@) # a.out sx86-out.s: asm/sha1-586.pl ../perlasm/x86asm.pl (cd asm; $(PERL) sha1-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@) -s512sse2-out.s: asm/sha512-sse2.pl ../perlasm/x86asm.pl - (cd asm; $(PERL) sha512-sse2.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@) +sha256x86-out.s: asm/sha256-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha256-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@) +sha512x86-out.s: asm/sha512-586.pl ../perlasm/x86asm.pl + (cd asm; $(PERL) sha512-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@) sha1-ia64.s: asm/sha1-ia64.pl (cd asm; $(PERL) sha1-ia64.pl $(CFLAGS) ) > $@ diff --git a/crypto/sha/asm/sha512-586.pl b/crypto/sha/asm/sha512-586.pl index 8f215ac247..49a9814544 100644 --- a/crypto/sha/asm/sha512-586.pl +++ b/crypto/sha/asm/sha512-586.pl @@ -38,6 +38,11 @@ require "x86asm.pl"; &asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386"); +$sse2=0; +for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); } + +&external_label("OPENSSL_ia32cap_P") if ($sse2); + $Tlo=&DWP(0,"esp"); $Thi=&DWP(4,"esp"); $Alo=&DWP(8,"esp"); $Ahi=&DWP(8+4,"esp"); $Blo=&DWP(16,"esp"); $Bhi=&DWP(16+4,"esp"); @@ -70,7 +75,7 @@ sub BODY_00_15_sse2 { &movq ("mm1",$E); # %mm1 is sliding right &movq ("mm2",$E); # %mm2 is sliding left &psrlq ("mm1",14); - &movq ($Esse2,$E); # module-scheduled save e + &movq ($Esse2,$E); # modulo-scheduled save e &psllq ("mm2",23); &movq ("mm3","mm1"); # %mm3 is T1 &psrlq ("mm1",4); @@ -109,7 +114,7 @@ sub BODY_00_15_sse2 { &pxor ("mm7","mm6"); &psllq ("mm6",6); &pxor ("mm7","mm5"); - &movq (&QWP(0,"esp"),$A); # module-scheduled save a + &movq (&QWP(0,"esp"),$A); # modulo-scheduled save a &pxor ("mm7","mm6"); # T2=Sigma0_512(a) &movq ("mm5",$A); # %mm5=a @@ -274,7 +279,8 @@ sub BODY_00_15_x86 { &mov (&DWP(8,"esp"),"eax"); # inp+num*128 &mov (&DWP(12,"esp"),"ebx"); # saved sp - &picmeup("edx","OPENSSL_ia32cap_P",$K512,&label("pic_point")); +if ($sse2) { + &picmeup("edx","OPENSSL_ia32cap_P",$K512,&label("K512")); &bt (&DWP(0,"edx"),26); &jnc (&label("loop_x86")); @@ -403,7 +409,7 @@ sub BODY_00_15_x86 { &emms (); &mov ("esp",&DWP(8*10+12,"esp")); # restore sp &function_end_A(); - +} &set_label("loop_x86",16); # copy input block to stack reversing byte and qword order for ($i=0;$i<8;$i++) { |