1 files changed, 327 insertions, 0 deletions
diff --git a/src/lib/libcrypto/bn/asm/mips3-mont.pl b/src/lib/libcrypto/bn/asm/mips3-mont.pl
new file mode 100644
index 0000000000..8f9156e02a
--- /dev/null
+++ b/src/lib/libcrypto/bn/asm/mips3-mont.pl
@@ -0,0 +1,327 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+# This module doesn't present direct interest for OpenSSL, because it
+# doesn't provide better performance for longer keys. While 512-bit
+# RSA private key operations are 40% faster, 1024-bit ones are hardly
+# faster at all, while longer key operations are slower by up to 20%.
+# It might be of interest to embedded system developers though, as
+# it's smaller than 1KB, yet offers ~3x improvement over compiler
+# generated code.
+#
+# The module targets N32 and N64 MIPS ABIs and currently is a bit
+# IRIX-centric, i.e. is likely to require adaptation for other OSes.
+# int bn_mul_mont(
+$rp="a0";       # BN_ULONG *rp,
+$ap="a1";       # const BN_ULONG *ap,
+$bp="a2";       # const BN_ULONG *bp,
+$np="a3";       # const BN_ULONG *np,
+$n0="a4";       # const BN_ULONG *n0,
+$num="a5";      # int num);
+$lo0="a6";
+$hi0="a7";
+$lo1="v0";
+$hi1="v1";
+$aj="t0";
+$bi="t1";
+$nj="t2";
+$tp="t3";
+$alo="s0";
+$ahi="s1";
+$nlo="s2";
+$nhi="s3";
+$tj="s4";
+$i="s5";
+$j="s6";
+$fp="t8";
+$m1="t9";
+$FRAME=8*(2+8);
+$code=<<___;
+#include <asm.h>
+#include <regdef.h>
+.text
+.set    noat
+.set    reorder
+.align  5
+.globl  bn_mul_mont
+.ent    bn_mul_mont
+bn_mul_mont:
+        .set    noreorder
+        PTR_SUB sp,64
+        move    $fp,sp
+        .frame  $fp,64,ra
+        slt     AT,$num,4
+        li      v0,0
+        beqzl   AT,.Lproceed
+        nop
+        jr      ra
+        PTR_ADD sp,$fp,64
+        .set    reorder
+.align  5
+.Lproceed:
+        ld      $n0,0($n0)
+        ld      $bi,0($bp)      # bp[0]
+        ld      $aj,0($ap)      # ap[0]
+        ld      $nj,0($np)      # np[0]
+        PTR_SUB sp,16           # place for two extra words
+        sll     $num,3
+        li      AT,-4096
+        PTR_SUB sp,$num
+        and     sp,AT
+        sd      s0,0($fp)
+        sd      s1,8($fp)
+        sd      s2,16($fp)
+        sd      s3,24($fp)
+        sd      s4,32($fp)
+        sd      s5,40($fp)
+        sd      s6,48($fp)
+        sd      s7,56($fp)
+        dmultu  $aj,$bi
+        ld      $alo,8($ap)
+        ld      $nlo,8($np)
+        mflo    $lo0
+        mfhi    $hi0
+        dmultu  $lo0,$n0
+        mflo    $m1
+        dmultu  $alo,$bi
+        mflo    $alo
+        mfhi    $ahi
+        dmultu  $nj,$m1
+        mflo    $lo1
+        mfhi    $hi1
+        dmultu  $nlo,$m1
+        daddu   $lo1,$lo0
+        sltu    AT,$lo1,$lo0
+        daddu   $hi1,AT
+        mflo    $nlo
+        mfhi    $nhi
+        move    $tp,sp
+        li      $j,16
+.align  4
+.L1st:
+        .set    noreorder
+        PTR_ADD $aj,$ap,$j
+        ld      $aj,($aj)
+        PTR_ADD $nj,$np,$j
+        ld      $nj,($nj)
+        dmultu  $aj,$bi
+        daddu   $lo0,$alo,$hi0
+        daddu   $lo1,$nlo,$hi1
+        sltu    AT,$lo0,$hi0
+        sltu    s7,$lo1,$hi1
+        daddu   $hi0,$ahi,AT
+        daddu   $hi1,$nhi,s7
+        mflo    $alo
+        mfhi    $ahi
+        daddu   $lo1,$lo0
+        sltu    AT,$lo1,$lo0
+        dmultu  $nj,$m1
+        daddu   $hi1,AT
+        addu    $j,8
+        sd      $lo1,($tp)
+        sltu    s7,$j,$num
+        mflo    $nlo
+        mfhi    $nhi
+        bnez    s7,.L1st
+        PTR_ADD $tp,8
+        .set    reorder
+        daddu   $lo0,$alo,$hi0
+        sltu    AT,$lo0,$hi0
+        daddu   $hi0,$ahi,AT
+        daddu   $lo1,$nlo,$hi1
+        sltu    s7,$lo1,$hi1
+        daddu   $hi1,$nhi,s7
+        daddu   $lo1,$lo0
+        sltu    AT,$lo1,$lo0
+        daddu   $hi1,AT
+        sd      $lo1,($tp)
+        daddu   $hi1,$hi0
+        sltu    AT,$hi1,$hi0
+        sd      $hi1,8($tp)
+        sd      AT,16($tp)
+        li      $i,8
+.align  4
+.Louter:
+        PTR_ADD $bi,$bp,$i
+        ld      $bi,($bi)
+        ld      $aj,($ap)
+        ld      $alo,8($ap)
+        ld      $tj,(sp)
+        dmultu  $aj,$bi
+        ld      $nj,($np)
+        ld      $nlo,8($np)
+        mflo    $lo0
+        mfhi    $hi0
+        daddu   $lo0,$tj
+        dmultu  $lo0,$n0
+        sltu    AT,$lo0,$tj
+        daddu   $hi0,AT
+        mflo    $m1
+        dmultu  $alo,$bi
+        mflo    $alo
+        mfhi    $ahi
+        dmultu  $nj,$m1
+        mflo    $lo1
+        mfhi    $hi1
+        dmultu  $nlo,$m1
+        daddu   $lo1,$lo0
+        sltu    AT,$lo1,$lo0
+        daddu   $hi1,AT
+        mflo    $nlo
+        mfhi    $nhi
+        move    $tp,sp
+        li      $j,16
+        ld      $tj,8($tp)
+.align  4
+.Linner:
+        .set    noreorder
+        PTR_ADD $aj,$ap,$j
+        ld      $aj,($aj)
+        PTR_ADD $nj,$np,$j
+        ld      $nj,($nj)
+        dmultu  $aj,$bi
+        daddu   $lo0,$alo,$hi0
+        daddu   $lo1,$nlo,$hi1
+        sltu    AT,$lo0,$hi0
+        sltu    s7,$lo1,$hi1
+        daddu   $hi0,$ahi,AT
+        daddu   $hi1,$nhi,s7
+        mflo    $alo
+        mfhi    $ahi
+        daddu   $lo0,$tj
+        addu    $j,8
+        dmultu  $nj,$m1
+        sltu    AT,$lo0,$tj
+        daddu   $lo1,$lo0
+        daddu   $hi0,AT
+        sltu    s7,$lo1,$lo0
+        ld      $tj,16($tp)
+        daddu   $hi1,s7
+        sltu    AT,$j,$num
+        mflo    $nlo
+        mfhi    $nhi
+        sd      $lo1,($tp)
+        bnez    AT,.Linner
+        PTR_ADD $tp,8
+        .set    reorder
+        daddu   $lo0,$alo,$hi0
+        sltu    AT,$lo0,$hi0
+        daddu   $hi0,$ahi,AT
+        daddu   $lo0,$tj
+        sltu    s7,$lo0,$tj
+        daddu   $hi0,s7
+        ld      $tj,16($tp)
+        daddu   $lo1,$nlo,$hi1
+        sltu    AT,$lo1,$hi1
+        daddu   $hi1,$nhi,AT
+        daddu   $lo1,$lo0
+        sltu    s7,$lo1,$lo0
+        daddu   $hi1,s7
+        sd      $lo1,($tp)
+        daddu   $lo1,$hi1,$hi0
+        sltu    $hi1,$lo1,$hi0
+        daddu   $lo1,$tj
+        sltu    AT,$lo1,$tj
+        daddu   $hi1,AT
+        sd      $lo1,8($tp)
+        sd      $hi1,16($tp)
+        addu    $i,8
+        sltu    s7,$i,$num
+        bnez    s7,.Louter
+        .set    noreorder
+        PTR_ADD $tj,sp,$num     # &tp[num]
+        move    $tp,sp
+        move    $ap,sp
+        li      $hi0,0          # clear borrow bit
+.align  4
+.Lsub:  ld      $lo0,($tp)
+        ld      $lo1,($np)
+        PTR_ADD $tp,8
+        PTR_ADD $np,8
+        dsubu   $lo1,$lo0,$lo1  # tp[i]-np[i]
+        sgtu    AT,$lo1,$lo0
+        dsubu   $lo0,$lo1,$hi0
+        sgtu    $hi0,$lo0,$lo1
+        sd      $lo0,($rp)
+        or      $hi0,AT
+        sltu    AT,$tp,$tj
+        bnez    AT,.Lsub
+        PTR_ADD $rp,8
+        dsubu   $hi0,$hi1,$hi0  # handle upmost overflow bit
+        move    $tp,sp
+        PTR_SUB $rp,$num        # restore rp
+        not     $hi1,$hi0
+        and     $ap,$hi0,sp
+        and     $bp,$hi1,$rp
+        or      $ap,$ap,$bp     # ap=borrow?tp:rp
+.align  4
+.Lcopy: ld      $aj,($ap)
+        PTR_ADD $ap,8
+        PTR_ADD $tp,8
+        sd      zero,-8($tp)
+        sltu    AT,$tp,$tj
+        sd      $aj,($rp)
+        bnez    AT,.Lcopy
+        PTR_ADD $rp,8
+        ld      s0,0($fp)
+        ld      s1,8($fp)
+        ld      s2,16($fp)
+        ld      s3,24($fp)
+        ld      s4,32($fp)
+        ld      s5,40($fp)
+        ld      s6,48($fp)
+        ld      s7,56($fp)
+        li      v0,1
+        jr      ra
+        PTR_ADD sp,$fp,64
+        .set    reorder
+END(bn_mul_mont)
+.rdata
+.asciiz "Montgomery Multiplication for MIPS III/IV, CRYPTOGAMS by <appro\@openssl.org>"
+___
+print $code;
+close STDOUT;

diff --git a/src/lib/libcrypto/bn/asm/mips3-mont.pl b/src/lib/libcrypto/bn/asm/mips3-mont.pl new file mode 100644 index 0000000000..8f9156e02a --- /dev/null +++ b/src/lib/libcrypto/bn/asm/mips3-mont.pl
@@ -0,0 +1,327 @@
	1	#!/usr/bin/env perl
	2	#
	3	# ====================================================================
	4	# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
	5	# project. The module is, however, dual licensed under OpenSSL and
	6	# CRYPTOGAMS licenses depending on where you obtain it. For further
	7	# details see http://www.openssl.org/~appro/cryptogams/.
	8	# ====================================================================
	9
	10	# This module doesn't present direct interest for OpenSSL, because it
	11	# doesn't provide better performance for longer keys. While 512-bit
	12	# RSA private key operations are 40% faster, 1024-bit ones are hardly
	13	# faster at all, while longer key operations are slower by up to 20%.
	14	# It might be of interest to embedded system developers though, as
	15	# it's smaller than 1KB, yet offers ~3x improvement over compiler
	16	# generated code.
	17	#
	18	# The module targets N32 and N64 MIPS ABIs and currently is a bit
	19	# IRIX-centric, i.e. is likely to require adaptation for other OSes.
	20
	21	# int bn_mul_mont(
	22	$rp="a0"; # BN_ULONG *rp,
	23	$ap="a1"; # const BN_ULONG *ap,
	24	$bp="a2"; # const BN_ULONG *bp,
	25	$np="a3"; # const BN_ULONG *np,
	26	$n0="a4"; # const BN_ULONG *n0,
	27	$num="a5"; # int num);
	28
	29	$lo0="a6";
	30	$hi0="a7";
	31	$lo1="v0";
	32	$hi1="v1";
	33	$aj="t0";
	34	$bi="t1";
	35	$nj="t2";
	36	$tp="t3";
	37	$alo="s0";
	38	$ahi="s1";
	39	$nlo="s2";
	40	$nhi="s3";
	41	$tj="s4";
	42	$i="s5";
	43	$j="s6";
	44	$fp="t8";
	45	$m1="t9";
	46
	47	$FRAME=8*(2+8);
	48
	49	$code=<<___;
	50	#include <asm.h>
	51	#include <regdef.h>
	52
	53	.text
	54
	55	.set noat
	56	.set reorder
	57
	58	.align 5
	59	.globl bn_mul_mont
	60	.ent bn_mul_mont
	61	bn_mul_mont:
	62	.set noreorder
	63	PTR_SUB sp,64
	64	move $fp,sp
	65	.frame $fp,64,ra
	66	slt AT,$num,4
	67	li v0,0
	68	beqzl AT,.Lproceed
	69	nop
	70	jr ra
	71	PTR_ADD sp,$fp,64
	72	.set reorder
	73	.align 5
	74	.Lproceed:
	75	ld $n0,0($n0)
	76	ld $bi,0($bp) # bp[0]
	77	ld $aj,0($ap) # ap[0]
	78	ld $nj,0($np) # np[0]
	79	PTR_SUB sp,16 # place for two extra words
	80	sll $num,3
	81	li AT,-4096
	82	PTR_SUB sp,$num
	83	and sp,AT
	84
	85	sd s0,0($fp)
	86	sd s1,8($fp)
	87	sd s2,16($fp)
	88	sd s3,24($fp)
	89	sd s4,32($fp)
	90	sd s5,40($fp)
	91	sd s6,48($fp)
	92	sd s7,56($fp)
	93
	94	dmultu $aj,$bi
	95	ld $alo,8($ap)
	96	ld $nlo,8($np)
	97	mflo $lo0
	98	mfhi $hi0
	99	dmultu $lo0,$n0
	100	mflo $m1
	101
	102	dmultu $alo,$bi
	103	mflo $alo
	104	mfhi $ahi
	105
	106	dmultu $nj,$m1
	107	mflo $lo1
	108	mfhi $hi1
	109	dmultu $nlo,$m1
	110	daddu $lo1,$lo0
	111	sltu AT,$lo1,$lo0
	112	daddu $hi1,AT
	113	mflo $nlo
	114	mfhi $nhi
	115
	116	move $tp,sp
	117	li $j,16
	118	.align 4
	119	.L1st:
	120	.set noreorder
	121	PTR_ADD $aj,$ap,$j
	122	ld $aj,($aj)
	123	PTR_ADD $nj,$np,$j
	124	ld $nj,($nj)
	125
	126	dmultu $aj,$bi
	127	daddu $lo0,$alo,$hi0
	128	daddu $lo1,$nlo,$hi1
	129	sltu AT,$lo0,$hi0
	130	sltu s7,$lo1,$hi1
	131	daddu $hi0,$ahi,AT
	132	daddu $hi1,$nhi,s7
	133	mflo $alo
	134	mfhi $ahi
	135
	136	daddu $lo1,$lo0
	137	sltu AT,$lo1,$lo0
	138	dmultu $nj,$m1
	139	daddu $hi1,AT
	140	addu $j,8
	141	sd $lo1,($tp)
	142	sltu s7,$j,$num
	143	mflo $nlo
	144	mfhi $nhi
	145
	146	bnez s7,.L1st
	147	PTR_ADD $tp,8
	148	.set reorder
	149
	150	daddu $lo0,$alo,$hi0
	151	sltu AT,$lo0,$hi0
	152	daddu $hi0,$ahi,AT
	153
	154	daddu $lo1,$nlo,$hi1
	155	sltu s7,$lo1,$hi1
	156	daddu $hi1,$nhi,s7
	157	daddu $lo1,$lo0
	158	sltu AT,$lo1,$lo0
	159	daddu $hi1,AT
	160
	161	sd $lo1,($tp)
	162
	163	daddu $hi1,$hi0
	164	sltu AT,$hi1,$hi0
	165	sd $hi1,8($tp)
	166	sd AT,16($tp)
	167
	168	li $i,8
	169	.align 4
	170	.Louter:
	171	PTR_ADD $bi,$bp,$i
	172	ld $bi,($bi)
	173	ld $aj,($ap)
	174	ld $alo,8($ap)
	175	ld $tj,(sp)
	176
	177	dmultu $aj,$bi
	178	ld $nj,($np)
	179	ld $nlo,8($np)
	180	mflo $lo0
	181	mfhi $hi0
	182	daddu $lo0,$tj
	183	dmultu $lo0,$n0
	184	sltu AT,$lo0,$tj
	185	daddu $hi0,AT
	186	mflo $m1
	187
	188	dmultu $alo,$bi
	189	mflo $alo
	190	mfhi $ahi
	191
	192	dmultu $nj,$m1
	193	mflo $lo1
	194	mfhi $hi1
	195
	196	dmultu $nlo,$m1
	197	daddu $lo1,$lo0
	198	sltu AT,$lo1,$lo0
	199	daddu $hi1,AT
	200	mflo $nlo
	201	mfhi $nhi
	202
	203	move $tp,sp
	204	li $j,16
	205	ld $tj,8($tp)
	206	.align 4
	207	.Linner:
	208	.set noreorder
	209	PTR_ADD $aj,$ap,$j
	210	ld $aj,($aj)
	211	PTR_ADD $nj,$np,$j
	212	ld $nj,($nj)
	213
	214	dmultu $aj,$bi
	215	daddu $lo0,$alo,$hi0
	216	daddu $lo1,$nlo,$hi1
	217	sltu AT,$lo0,$hi0
	218	sltu s7,$lo1,$hi1
	219	daddu $hi0,$ahi,AT
	220	daddu $hi1,$nhi,s7
	221	mflo $alo
	222	mfhi $ahi
	223
	224	daddu $lo0,$tj
	225	addu $j,8
	226	dmultu $nj,$m1
	227	sltu AT,$lo0,$tj
	228	daddu $lo1,$lo0
	229	daddu $hi0,AT
	230	sltu s7,$lo1,$lo0
	231	ld $tj,16($tp)
	232	daddu $hi1,s7
	233	sltu AT,$j,$num
	234	mflo $nlo
	235	mfhi $nhi
	236	sd $lo1,($tp)
	237	bnez AT,.Linner
	238	PTR_ADD $tp,8
	239	.set reorder
	240
	241	daddu $lo0,$alo,$hi0
	242	sltu AT,$lo0,$hi0
	243	daddu $hi0,$ahi,AT
	244	daddu $lo0,$tj
	245	sltu s7,$lo0,$tj
	246	daddu $hi0,s7
	247
	248	ld $tj,16($tp)
	249	daddu $lo1,$nlo,$hi1
	250	sltu AT,$lo1,$hi1
	251	daddu $hi1,$nhi,AT
	252	daddu $lo1,$lo0
	253	sltu s7,$lo1,$lo0
	254	daddu $hi1,s7
	255	sd $lo1,($tp)
	256
	257	daddu $lo1,$hi1,$hi0
	258	sltu $hi1,$lo1,$hi0
	259	daddu $lo1,$tj
	260	sltu AT,$lo1,$tj
	261	daddu $hi1,AT
	262	sd $lo1,8($tp)
	263	sd $hi1,16($tp)
	264
	265	addu $i,8
	266	sltu s7,$i,$num
	267	bnez s7,.Louter
	268
	269	.set noreorder
	270	PTR_ADD $tj,sp,$num # &tp[num]
	271	move $tp,sp
	272	move $ap,sp
	273	li $hi0,0 # clear borrow bit
	274
	275	.align 4
	276	.Lsub: ld $lo0,($tp)
	277	ld $lo1,($np)
	278	PTR_ADD $tp,8
	279	PTR_ADD $np,8
	280	dsubu $lo1,$lo0,$lo1 # tp[i]-np[i]
	281	sgtu AT,$lo1,$lo0
	282	dsubu $lo0,$lo1,$hi0
	283	sgtu $hi0,$lo0,$lo1
	284	sd $lo0,($rp)
	285	or $hi0,AT
	286	sltu AT,$tp,$tj
	287	bnez AT,.Lsub
	288	PTR_ADD $rp,8
	289
	290	dsubu $hi0,$hi1,$hi0 # handle upmost overflow bit
	291	move $tp,sp
	292	PTR_SUB $rp,$num # restore rp
	293	not $hi1,$hi0
	294
	295	and $ap,$hi0,sp
	296	and $bp,$hi1,$rp
	297	or $ap,$ap,$bp # ap=borrow?tp:rp
	298
	299	.align 4
	300	.Lcopy: ld $aj,($ap)
	301	PTR_ADD $ap,8
	302	PTR_ADD $tp,8
	303	sd zero,-8($tp)
	304	sltu AT,$tp,$tj
	305	sd $aj,($rp)
	306	bnez AT,.Lcopy
	307	PTR_ADD $rp,8
	308
	309	ld s0,0($fp)
	310	ld s1,8($fp)
	311	ld s2,16($fp)
	312	ld s3,24($fp)
	313	ld s4,32($fp)
	314	ld s5,40($fp)
	315	ld s6,48($fp)
	316	ld s7,56($fp)
	317	li v0,1
	318	jr ra
	319	PTR_ADD sp,$fp,64
	320	.set reorder
	321	END(bn_mul_mont)
	322	.rdata
	323	.asciiz "Montgomery Multiplication for MIPS III/IV, CRYPTOGAMS by <appro\@openssl.org>"
	324	___
	325
	326	print $code;
	327	close STDOUT;