Date: Wed, 14 Jan 2015 21:39:32 +0000 (UTC) From: Xin LI <delphij@FreeBSD.org> To: src-committers@freebsd.org, svn-src-user@freebsd.org Subject: svn commit: r277196 - in user/cperciva/freebsd-update-build/patches: 10.0-RELEASE 10.1-RELEASE 8.4-RELEASE 9.3-RELEASE Message-ID: <201501142139.t0ELdWL6026812@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: delphij Date: Wed Jan 14 21:39:31 2015 New Revision: 277196 URL: https://svnweb.freebsd.org/changeset/base/277196 Log: Add patches for today's batch. Added: user/cperciva/freebsd-update-build/patches/10.0-RELEASE/16-SA-15:01.openssl user/cperciva/freebsd-update-build/patches/10.1-RELEASE/4-SA-15:01.openssl user/cperciva/freebsd-update-build/patches/8.4-RELEASE/22-SA-15:01.openssl user/cperciva/freebsd-update-build/patches/9.3-RELEASE/8-SA-15:01.openssl Added: user/cperciva/freebsd-update-build/patches/10.0-RELEASE/16-SA-15:01.openssl ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ user/cperciva/freebsd-update-build/patches/10.0-RELEASE/16-SA-15:01.openssl Wed Jan 14 21:39:31 2015 (r277196) @@ -0,0 +1,2083 @@ +Index: crypto/openssl/crypto/asn1/a_bitstr.c +=================================================================== +--- crypto/openssl/crypto/asn1/a_bitstr.c (revision 276867) ++++ crypto/openssl/crypto/asn1/a_bitstr.c (working copy) +@@ -136,11 +136,16 @@ ASN1_BIT_STRING *c2i_ASN1_BIT_STRING(ASN1_BIT_STRI + + p= *pp; + i= *(p++); ++ if (i > 7) ++ { ++ i=ASN1_R_INVALID_BIT_STRING_BITS_LEFT; ++ goto err; ++ } + /* We do this to preserve the settings. If we modify + * the settings, via the _set_bit function, we will recalculate + * on output */ + ret->flags&= ~(ASN1_STRING_FLAG_BITS_LEFT|0x07); /* clear */ +- ret->flags|=(ASN1_STRING_FLAG_BITS_LEFT|(i&0x07)); /* set */ ++ ret->flags|=(ASN1_STRING_FLAG_BITS_LEFT|i); /* set */ + + if (len-- > 1) /* using one because of the bits left byte */ + { +Index: crypto/openssl/crypto/asn1/a_type.c +=================================================================== +--- crypto/openssl/crypto/asn1/a_type.c (revision 276867) ++++ crypto/openssl/crypto/asn1/a_type.c (working copy) +@@ -113,7 +113,7 @@ IMPLEMENT_STACK_OF(ASN1_TYPE) + IMPLEMENT_ASN1_SET_OF(ASN1_TYPE) + + /* Returns 0 if they are equal, != 0 otherwise. */ +-int ASN1_TYPE_cmp(ASN1_TYPE *a, ASN1_TYPE *b) ++int ASN1_TYPE_cmp(const ASN1_TYPE *a, const ASN1_TYPE *b) + { + int result = -1; + +Index: crypto/openssl/crypto/asn1/a_verify.c +=================================================================== +--- crypto/openssl/crypto/asn1/a_verify.c (revision 276867) ++++ crypto/openssl/crypto/asn1/a_verify.c (working copy) +@@ -90,6 +90,12 @@ int ASN1_verify(i2d_of_void *i2d, X509_ALGOR *a, A + ASN1err(ASN1_F_ASN1_VERIFY,ASN1_R_UNKNOWN_MESSAGE_DIGEST_ALGORITHM); + goto err; + } ++ ++ if (signature->type == V_ASN1_BIT_STRING && signature->flags & 0x7) ++ { ++ ASN1err(ASN1_F_ASN1_VERIFY, ASN1_R_INVALID_BIT_STRING_BITS_LEFT); ++ goto err; ++ } + + inl=i2d(data,NULL); + buf_in=OPENSSL_malloc((unsigned int)inl); +@@ -146,6 +152,12 @@ int ASN1_item_verify(const ASN1_ITEM *it, X509_ALG + return -1; + } + ++ if (signature->type == V_ASN1_BIT_STRING && signature->flags & 0x7) ++ { ++ ASN1err(ASN1_F_ASN1_ITEM_VERIFY, ASN1_R_INVALID_BIT_STRING_BITS_LEFT); ++ return -1; ++ } ++ + EVP_MD_CTX_init(&ctx); + + /* Convert signature OID into digest and public key OIDs */ +Index: crypto/openssl/crypto/asn1/asn1.h +=================================================================== +--- crypto/openssl/crypto/asn1/asn1.h (revision 276867) ++++ crypto/openssl/crypto/asn1/asn1.h (working copy) +@@ -776,7 +776,7 @@ DECLARE_ASN1_FUNCTIONS_fname(ASN1_TYPE, ASN1_ANY, + int ASN1_TYPE_get(ASN1_TYPE *a); + void ASN1_TYPE_set(ASN1_TYPE *a, int type, void *value); + int ASN1_TYPE_set1(ASN1_TYPE *a, int type, const void *value); +-int ASN1_TYPE_cmp(ASN1_TYPE *a, ASN1_TYPE *b); ++int ASN1_TYPE_cmp(const ASN1_TYPE *a, const ASN1_TYPE *b); + + ASN1_OBJECT * ASN1_OBJECT_new(void ); + void ASN1_OBJECT_free(ASN1_OBJECT *a); +@@ -1329,6 +1329,7 @@ void ERR_load_ASN1_strings(void); + #define ASN1_R_ILLEGAL_TIME_VALUE 184 + #define ASN1_R_INTEGER_NOT_ASCII_FORMAT 185 + #define ASN1_R_INTEGER_TOO_LARGE_FOR_LONG 128 ++#define ASN1_R_INVALID_BIT_STRING_BITS_LEFT 220 + #define ASN1_R_INVALID_BMPSTRING_LENGTH 129 + #define ASN1_R_INVALID_DIGIT 130 + #define ASN1_R_INVALID_MIME_TYPE 205 +Index: crypto/openssl/crypto/asn1/asn1_err.c +=================================================================== +--- crypto/openssl/crypto/asn1/asn1_err.c (revision 276867) ++++ crypto/openssl/crypto/asn1/asn1_err.c (working copy) +@@ -246,6 +246,7 @@ static ERR_STRING_DATA ASN1_str_reasons[]= + {ERR_REASON(ASN1_R_ILLEGAL_TIME_VALUE) ,"illegal time value"}, + {ERR_REASON(ASN1_R_INTEGER_NOT_ASCII_FORMAT),"integer not ascii format"}, + {ERR_REASON(ASN1_R_INTEGER_TOO_LARGE_FOR_LONG),"integer too large for long"}, ++{ERR_REASON(ASN1_R_INVALID_BIT_STRING_BITS_LEFT),"invalid bit string bits left"}, + {ERR_REASON(ASN1_R_INVALID_BMPSTRING_LENGTH),"invalid bmpstring length"}, + {ERR_REASON(ASN1_R_INVALID_DIGIT) ,"invalid digit"}, + {ERR_REASON(ASN1_R_INVALID_MIME_TYPE) ,"invalid mime type"}, +Index: crypto/openssl/crypto/asn1/x_algor.c +=================================================================== +--- crypto/openssl/crypto/asn1/x_algor.c (revision 276867) ++++ crypto/openssl/crypto/asn1/x_algor.c (working copy) +@@ -142,3 +142,14 @@ void X509_ALGOR_set_md(X509_ALGOR *alg, const EVP_ + X509_ALGOR_set0(alg, OBJ_nid2obj(EVP_MD_type(md)), param_type, NULL); + + } ++ ++int X509_ALGOR_cmp(const X509_ALGOR *a, const X509_ALGOR *b) ++ { ++ int rv; ++ rv = OBJ_cmp(a->algorithm, b->algorithm); ++ if (rv) ++ return rv; ++ if (!a->parameter && !b->parameter) ++ return 0; ++ return ASN1_TYPE_cmp(a->parameter, b->parameter); ++ } +Index: crypto/openssl/crypto/bio/bss_dgram.c +=================================================================== +--- crypto/openssl/crypto/bio/bss_dgram.c (revision 276867) ++++ crypto/openssl/crypto/bio/bss_dgram.c (working copy) +@@ -982,7 +982,12 @@ static int dgram_sctp_free(BIO *a) + return 0; + + data = (bio_dgram_sctp_data *)a->ptr; +- if(data != NULL) OPENSSL_free(data); ++ if(data != NULL) ++ { ++ if(data->saved_message.data != NULL) ++ OPENSSL_free(data->saved_message.data); ++ OPENSSL_free(data); ++ } + + return(1); + } +@@ -1099,6 +1104,7 @@ static int dgram_sctp_read(BIO *b, char *out, int + dgram_sctp_write(data->saved_message.bio, data->saved_message.data, + data->saved_message.length); + OPENSSL_free(data->saved_message.data); ++ data->saved_message.data = NULL; + data->saved_message.length = 0; + } + +@@ -1258,9 +1264,11 @@ static int dgram_sctp_write(BIO *b, const char *in + if (data->save_shutdown && !BIO_dgram_sctp_wait_for_dry(b)) + { + data->saved_message.bio = b; +- data->saved_message.length = inl; ++ if (data->saved_message.data) ++ OPENSSL_free(data->saved_message.data); + data->saved_message.data = OPENSSL_malloc(inl); + memcpy(data->saved_message.data, in, inl); ++ data->saved_message.length = inl; + return inl; + } + +Index: crypto/openssl/crypto/bn/asm/mips.pl +=================================================================== +--- crypto/openssl/crypto/bn/asm/mips.pl (revision 276867) ++++ crypto/openssl/crypto/bn/asm/mips.pl (working copy) +@@ -1874,8 +1874,43 @@ ___ + + ($a_4,$a_5,$a_6,$a_7)=($b_0,$b_1,$b_2,$b_3); + ++sub add_c2 () { ++my ($hi,$lo,$c0,$c1,$c2, ++ $warm, # !$warm denotes first call with specific sequence of ++ # $c_[XYZ] when there is no Z-carry to accumulate yet; ++ $an,$bn # these two are arguments for multiplication which ++ # result is used in *next* step [which is why it's ++ # commented as "forward multiplication" below]; ++ )=@_; + $code.=<<___; ++ mflo $lo ++ mfhi $hi ++ $ADDU $c0,$lo ++ sltu $at,$c0,$lo ++ $MULTU $an,$bn # forward multiplication ++ $ADDU $c0,$lo ++ $ADDU $at,$hi ++ sltu $lo,$c0,$lo ++ $ADDU $c1,$at ++ $ADDU $hi,$lo ++___ ++$code.=<<___ if (!$warm); ++ sltu $c2,$c1,$at ++ $ADDU $c1,$hi ++ sltu $hi,$c1,$hi ++ $ADDU $c2,$hi ++___ ++$code.=<<___ if ($warm); ++ sltu $at,$c1,$at ++ $ADDU $c1,$hi ++ $ADDU $c2,$at ++ sltu $hi,$c1,$hi ++ $ADDU $c2,$hi ++___ ++} + ++$code.=<<___; ++ + .align 5 + .globl bn_sqr_comba8 + .ent bn_sqr_comba8 +@@ -1922,25 +1957,14 @@ $code.=<<___; + sltu $at,$c_2,$t_1 + $ADDU $c_3,$t_2,$at + $ST $c_2,$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_1,$a_1); # mul_add_c(a[1],b[1],c3,c1,c2); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_1 # mul_add_c(a[1],b[1],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_3,$t_1 + sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 + $MULTU $a_0,$a_3 # mul_add_c2(a[0],b[3],c1,c2,c3); + $ADDU $t_2,$at + $ADDU $c_1,$t_2 +@@ -1947,71 +1971,23 @@ $code.=<<___; + sltu $at,$c_1,$t_2 + $ADDU $c_2,$at + $ST $c_3,2*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_3,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_2 # mul_add_c2(a[1],b[2],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_4,$a_0 # mul_add_c2(a[4],b[0],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at ++___ ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0, ++ $a_1,$a_2); # mul_add_c2(a[1],b[2],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_4,$a_0); # mul_add_c2(a[4],b[0],c2,c3,c1); ++$code.=<<___; + $ST $c_1,3*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0, ++ $a_3,$a_1); # mul_add_c2(a[3],b[1],c2,c3,c1); ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1, ++ $a_2,$a_2); # mul_add_c(a[2],b[2],c2,c3,c1); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_1,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_3,$a_1 # mul_add_c2(a[3],b[1],c2,c3,c1); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_2,$t_1 + sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_1,$at +- $MULTU $a_2,$a_2 # mul_add_c(a[2],b[2],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 + $MULTU $a_0,$a_5 # mul_add_c2(a[0],b[5],c3,c1,c2); + $ADDU $t_2,$at + $ADDU $c_3,$t_2 +@@ -2018,101 +1994,27 @@ $code.=<<___; + sltu $at,$c_3,$t_2 + $ADDU $c_1,$at + $ST $c_2,4*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_4 # mul_add_c2(a[1],b[4],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_2,$at +- $MULTU $a_2,$a_3 # mul_add_c2(a[2],b[3],c3,c1,c2); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $MULTU $a_6,$a_0 # mul_add_c2(a[6],b[0],c1,c2,c3); +- $ADDU $c_2,$at +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_1,$a_4); # mul_add_c2(a[1],b[4],c3,c1,c2); ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1, ++ $a_2,$a_3); # mul_add_c2(a[2],b[3],c3,c1,c2); ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1, ++ $a_6,$a_0); # mul_add_c2(a[6],b[0],c1,c2,c3); ++$code.=<<___; + $ST $c_3,5*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0, ++ $a_5,$a_1); # mul_add_c2(a[5],b[1],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_4,$a_2); # mul_add_c2(a[4],b[2],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_3,$a_3); # mul_add_c(a[3],b[3],c1,c2,c3); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_3,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_5,$a_1 # mul_add_c2(a[5],b[1],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_1,$t_1 + sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_4,$a_2 # mul_add_c2(a[4],b[2],c1,c2,c3); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_3,$a_3 # mul_add_c(a[3],b[3],c1,c2,c3); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 + $MULTU $a_0,$a_7 # mul_add_c2(a[0],b[7],c2,c3,c1); + $ADDU $t_2,$at + $ADDU $c_2,$t_2 +@@ -2119,116 +2021,29 @@ $code.=<<___; + sltu $at,$c_2,$t_2 + $ADDU $c_3,$at + $ST $c_1,6*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_1,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_6 # mul_add_c2(a[1],b[6],c2,c3,c1); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_1,$at +- $MULTU $a_2,$a_5 # mul_add_c2(a[2],b[5],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_1,$at +- $MULTU $a_3,$a_4 # mul_add_c2(a[3],b[4],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_1,$at +- $MULTU $a_7,$a_1 # mul_add_c2(a[7],b[1],c3,c1,c2); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at ++___ ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0, ++ $a_1,$a_6); # mul_add_c2(a[1],b[6],c2,c3,c1); ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1, ++ $a_2,$a_5); # mul_add_c2(a[2],b[5],c2,c3,c1); ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1, ++ $a_3,$a_4); # mul_add_c2(a[3],b[4],c2,c3,c1); ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1, ++ $a_7,$a_1); # mul_add_c2(a[7],b[1],c3,c1,c2); ++$code.=<<___; + $ST $c_2,7*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_6,$a_2); # mul_add_c2(a[6],b[2],c3,c1,c2); ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1, ++ $a_5,$a_3); # mul_add_c2(a[5],b[3],c3,c1,c2); ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1, ++ $a_4,$a_4); # mul_add_c(a[4],b[4],c3,c1,c2); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_6,$a_2 # mul_add_c2(a[6],b[2],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_3,$t_1 + sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_2,$at +- $MULTU $a_5,$a_3 # mul_add_c2(a[5],b[3],c3,c1,c2); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_2,$at +- $MULTU $a_4,$a_4 # mul_add_c(a[4],b[4],c3,c1,c2); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 + $MULTU $a_2,$a_7 # mul_add_c2(a[2],b[7],c1,c2,c3); + $ADDU $t_2,$at + $ADDU $c_1,$t_2 +@@ -2235,86 +2050,25 @@ $code.=<<___; + sltu $at,$c_1,$t_2 + $ADDU $c_2,$at + $ST $c_3,8*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_3,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_3,$a_6 # mul_add_c2(a[3],b[6],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_4,$a_5 # mul_add_c2(a[4],b[5],c1,c2,c3); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_7,$a_3 # mul_add_c2(a[7],b[3],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at ++___ ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0, ++ $a_3,$a_6); # mul_add_c2(a[3],b[6],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_4,$a_5); # mul_add_c2(a[4],b[5],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_7,$a_3); # mul_add_c2(a[7],b[3],c2,c3,c1); ++$code.=<<___; + $ST $c_1,9*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0, ++ $a_6,$a_4); # mul_add_c2(a[6],b[4],c2,c3,c1); ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,1, ++ $a_5,$a_5); # mul_add_c(a[5],b[5],c2,c3,c1); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_1,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_6,$a_4 # mul_add_c2(a[6],b[4],c2,c3,c1); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_2,$t_1 + sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_1,$at +- $MULTU $a_5,$a_5 # mul_add_c(a[5],b[5],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 + $MULTU $a_4,$a_7 # mul_add_c2(a[4],b[7],c3,c1,c2); + $ADDU $t_2,$at + $ADDU $c_3,$t_2 +@@ -2321,56 +2075,21 @@ $code.=<<___; + sltu $at,$c_3,$t_2 + $ADDU $c_1,$at + $ST $c_2,10*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_5,$a_6 # mul_add_c2(a[5],b[6],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_2,$at +- $MULTU $a_7,$a_5 # mul_add_c2(a[7],b[5],c1,c2,c3); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_5,$a_6); # mul_add_c2(a[5],b[6],c3,c1,c2); ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,1, ++ $a_7,$a_5); # mul_add_c2(a[7],b[5],c1,c2,c3); ++$code.=<<___; + $ST $c_3,11*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0, ++ $a_6,$a_6); # mul_add_c(a[6],b[6],c1,c2,c3); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_3,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_6,$a_6 # mul_add_c(a[6],b[6],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_1,$t_1 + sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 + $MULTU $a_6,$a_7 # mul_add_c2(a[6],b[7],c2,c3,c1); + $ADDU $t_2,$at + $ADDU $c_2,$t_2 +@@ -2377,21 +2096,10 @@ $code.=<<___; + sltu $at,$c_2,$t_2 + $ADDU $c_3,$at + $ST $c_1,12*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_1,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_7,$a_7 # mul_add_c(a[7],b[7],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at ++___ ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0, ++ $a_7,$a_7); # mul_add_c(a[7],b[7],c3,c1,c2); ++$code.=<<___; + $ST $c_2,13*$BNSZ($a0) + + mflo $t_1 +@@ -2459,25 +2167,14 @@ $code.=<<___; + sltu $at,$c_2,$t_1 + $ADDU $c_3,$t_2,$at + $ST $c_2,$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_1,$a_1); # mul_add_c(a[1],b[1],c3,c1,c2); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_1 # mul_add_c(a[1],b[1],c3,c1,c2); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_3,$t_1 + sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 + $MULTU $a_0,$a_3 # mul_add_c2(a[0],b[3],c1,c2,c3); + $ADDU $t_2,$at + $ADDU $c_1,$t_2 +@@ -2484,56 +2181,21 @@ $code.=<<___; + sltu $at,$c_1,$t_2 + $ADDU $c_2,$at + $ST $c_3,2*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_3,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_1,$a_2 # mul_add_c(a2[1],b[2],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at +- mflo $t_1 +- mfhi $t_2 +- slt $at,$t_2,$zero +- $ADDU $c_3,$at +- $MULTU $a_3,$a_1 # mul_add_c2(a[3],b[1],c2,c3,c1); +- $SLL $t_2,1 +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_1,$t_1 +- sltu $at,$c_1,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_2,$t_2 +- sltu $at,$c_2,$t_2 +- $ADDU $c_3,$at ++___ ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,0, ++ $a_1,$a_2); # mul_add_c2(a2[1],b[2],c1,c2,c3); ++ &add_c2($t_2,$t_1,$c_1,$c_2,$c_3,1, ++ $a_3,$a_1); # mul_add_c2(a[3],b[1],c2,c3,c1); ++$code.=<<___; + $ST $c_1,3*$BNSZ($a0) +- ++___ ++ &add_c2($t_2,$t_1,$c_2,$c_3,$c_1,0, ++ $a_2,$a_2); # mul_add_c(a[2],b[2],c2,c3,c1); ++$code.=<<___; + mflo $t_1 + mfhi $t_2 +- slt $c_1,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_2,$a_2 # mul_add_c(a[2],b[2],c2,c3,c1); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 + $ADDU $c_2,$t_1 + sltu $at,$c_2,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_3,$t_2 +- sltu $at,$c_3,$t_2 +- $ADDU $c_1,$at +- mflo $t_1 +- mfhi $t_2 +- $ADDU $c_2,$t_1 +- sltu $at,$c_2,$t_1 + $MULTU $a_2,$a_3 # mul_add_c2(a[2],b[3],c3,c1,c2); + $ADDU $t_2,$at + $ADDU $c_3,$t_2 +@@ -2540,21 +2202,10 @@ $code.=<<___; + sltu $at,$c_3,$t_2 + $ADDU $c_1,$at + $ST $c_2,4*$BNSZ($a0) +- +- mflo $t_1 +- mfhi $t_2 +- slt $c_2,$t_2,$zero +- $SLL $t_2,1 +- $MULTU $a_3,$a_3 # mul_add_c(a[3],b[3],c1,c2,c3); +- slt $a2,$t_1,$zero +- $ADDU $t_2,$a2 +- $SLL $t_1,1 +- $ADDU $c_3,$t_1 +- sltu $at,$c_3,$t_1 +- $ADDU $t_2,$at +- $ADDU $c_1,$t_2 +- sltu $at,$c_1,$t_2 +- $ADDU $c_2,$at ++___ ++ &add_c2($t_2,$t_1,$c_3,$c_1,$c_2,0, ++ $a_3,$a_3); # mul_add_c(a[3],b[3],c1,c2,c3); ++$code.=<<___; + $ST $c_3,5*$BNSZ($a0) + + mflo $t_1 +Index: crypto/openssl/crypto/bn/asm/x86_64-gcc.c +=================================================================== +--- crypto/openssl/crypto/bn/asm/x86_64-gcc.c (revision 276867) ++++ crypto/openssl/crypto/bn/asm/x86_64-gcc.c (working copy) +@@ -273,6 +273,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN + /* sqr_add_c(a,i,c0,c1,c2) -- c+=a[i]^2 for three word number c=(c2,c1,c0) */ + /* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */ + ++/* ++ * Keep in mind that carrying into high part of multiplication result ++ * can not overflow, because it cannot be all-ones. ++ */ + #if 0 + /* original macros are kept for reference purposes */ + #define mul_add_c(a,b,c0,c1,c2) { \ +@@ -287,10 +291,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN + BN_ULONG ta=(a),tb=(b),t0; \ + t1 = BN_UMULT_HIGH(ta,tb); \ + t0 = ta * tb; \ +- t2 = t1+t1; c2 += (t2<t1)?1:0; \ +- t1 = t0+t0; t2 += (t1<t0)?1:0; \ +- c0 += t1; t2 += (c0<t1)?1:0; \ ++ c0 += t0; t2 = t1+((c0<t0)?1:0);\ + c1 += t2; c2 += (c1<t2)?1:0; \ ++ c0 += t0; t1 += (c0<t0)?1:0; \ ++ c1 += t1; c2 += (c1<t1)?1:0; \ + } + #else + #define mul_add_c(a,b,c0,c1,c2) do { \ +@@ -328,22 +332,14 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN + : "=a"(t1),"=d"(t2) \ + : "a"(a),"m"(b) \ + : "cc"); \ +- asm ("addq %0,%0; adcq %2,%1" \ +- : "+d"(t2),"+r"(c2) \ +- : "g"(0) \ +- : "cc"); \ +- asm ("addq %0,%0; adcq %2,%1" \ +- : "+a"(t1),"+d"(t2) \ +- : "g"(0) \ +- : "cc"); \ +- asm ("addq %2,%0; adcq %3,%1" \ +- : "+r"(c0),"+d"(t2) \ +- : "a"(t1),"g"(0) \ +- : "cc"); \ +- asm ("addq %2,%0; adcq %3,%1" \ +- : "+r"(c1),"+r"(c2) \ +- : "d"(t2),"g"(0) \ +- : "cc"); \ ++ asm ("addq %3,%0; adcq %4,%1; adcq %5,%2" \ ++ : "+r"(c0),"+r"(c1),"+r"(c2) \ ++ : "r"(t1),"r"(t2),"g"(0) \ ++ : "cc"); \ ++ asm ("addq %3,%0; adcq %4,%1; adcq %5,%2" \ ++ : "+r"(c0),"+r"(c1),"+r"(c2) \ ++ : "r"(t1),"r"(t2),"g"(0) \ ++ : "cc"); \ + } while (0) + #endif + +Index: crypto/openssl/crypto/bn/bn_asm.c +=================================================================== +--- crypto/openssl/crypto/bn/bn_asm.c (revision 276867) ++++ crypto/openssl/crypto/bn/bn_asm.c (working copy) +@@ -438,6 +438,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG + /* sqr_add_c(a,i,c0,c1,c2) -- c+=a[i]^2 for three word number c=(c2,c1,c0) */ + /* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */ + ++/* ++ * Keep in mind that carrying into high part of multiplication result ++ * can not overflow, because it cannot be all-ones. ++ */ + #ifdef BN_LLONG + #define mul_add_c(a,b,c0,c1,c2) \ + t=(BN_ULLONG)a*b; \ +@@ -478,10 +482,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG + #define mul_add_c2(a,b,c0,c1,c2) { \ + BN_ULONG ta=(a),tb=(b),t0; \ + BN_UMULT_LOHI(t0,t1,ta,tb); \ +- t2 = t1+t1; c2 += (t2<t1)?1:0; \ +- t1 = t0+t0; t2 += (t1<t0)?1:0; \ +- c0 += t1; t2 += (c0<t1)?1:0; \ ++ c0 += t0; t2 = t1+((c0<t0)?1:0);\ + c1 += t2; c2 += (c1<t2)?1:0; \ ++ c0 += t0; t1 += (c0<t0)?1:0; \ ++ c1 += t1; c2 += (c1<t1)?1:0; \ + } + + #define sqr_add_c(a,i,c0,c1,c2) { \ +@@ -508,10 +512,10 @@ BN_ULONG bn_sub_words(BN_ULONG *r, const BN_ULONG + BN_ULONG ta=(a),tb=(b),t0; \ + t1 = BN_UMULT_HIGH(ta,tb); \ + t0 = ta * tb; \ +- t2 = t1+t1; c2 += (t2<t1)?1:0; \ +- t1 = t0+t0; t2 += (t1<t0)?1:0; \ +- c0 += t1; t2 += (c0<t1)?1:0; \ ++ c0 += t0; t2 = t1+((c0<t0)?1:0);\ + c1 += t2; c2 += (c1<t2)?1:0; \ ++ c0 += t0; t1 += (c0<t0)?1:0; \ ++ c1 += t1; c2 += (c1<t1)?1:0; \ + } + + #define sqr_add_c(a,i,c0,c1,c2) { \ +Index: crypto/openssl/crypto/bn/bntest.c +=================================================================== +--- crypto/openssl/crypto/bn/bntest.c (revision 276867) ++++ crypto/openssl/crypto/bn/bntest.c (working copy) +@@ -676,44 +676,98 @@ int test_mul(BIO *bp) + *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201501142139.t0ELdWL6026812>