MIPS assembly pack: adapt it for MIPS[32|64]R6.

MIPS[32|64]R6 is binary and source incompatible with previous MIPS ISA
specifications. Fortunately it's still possible to resolve differences
in source code with standard pre-processor and switching to trap-free
version of addition and subtraction instructions.

Reviewed-by: Richard Levitte <levitte@openssl.org>

1 files changed, 65 insertions, 22 deletions

diff --git a/crypto/aes/asm/aes-mips.pl b/crypto/aes/asm/aes-mips.pl
index 439578d9c2..0eb1474a3a 100644
--- a/crypto/aes/asm/aes-mips.pl
+++ b/crypto/aes/asm/aes-mips.pl
@@ -65,8 +65,8 @@ $flavour = shift || "o32"; # supported flavours are o32,n32,64,nubi32,nubi64
 
 if ($flavour =~ /64|n32/i) {
 	$PTR_LA="dla";
-	$PTR_ADD="dadd";	# incidentally works even on n32
-	$PTR_SUB="dsub";	# incidentally works even on n32
+	$PTR_ADD="daddu";	# incidentally works even on n32
+	$PTR_SUB="dsubu";	# incidentally works even on n32
 	$PTR_INS="dins";
 	$REG_S="sd";
 	$REG_L="ld";
@@ -74,8 +74,8 @@ if ($flavour =~ /64|n32/i) {
 	$SZREG=8;
 } else {
 	$PTR_LA="la";
-	$PTR_ADD="add";
-	$PTR_SUB="sub";
+	$PTR_ADD="addu";
+	$PTR_SUB="subu";
 	$PTR_INS="ins";
 	$REG_S="sw";
 	$REG_L="lw";
@@ -102,15 +102,13 @@ open STDOUT,">$output";
 my ($MSB,$LSB)=(0,3);	# automatically converted to little-endian
 
 $code.=<<___;
+#include "mips_arch.h"
+
 .text
 #ifdef OPENSSL_FIPSCANISTER
 # include <openssl/fipssyms.h>
 #endif
 
-#if defined(__mips_smartmips) && !defined(_MIPS_ARCH_MIPS32R2)
-#define _MIPS_ARCH_MIPS32R2
-#endif
-
 #if !defined(__mips_eabi) && (!defined(__vxworks) || defined(__pic__))
 .option	pic2
 #endif
@@ -146,7 +144,7 @@ _mips_AES_encrypt:
 	xor	$s2,$t2
 	xor	$s3,$t3
 
-	sub	$cnt,1
+	subu	$cnt,1
 #if defined(__mips_smartmips)
 	ext	$i0,$s1,16,8
 .Loop_enc:
@@ -218,7 +216,7 @@ _mips_AES_encrypt:
 	xor	$t2,$t6
 	xor	$t3,$t7
 
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key0,16
 	xor	$s0,$t0
 	xor	$s1,$t1
@@ -409,7 +407,7 @@ _mips_AES_encrypt:
 	xor	$t2,$t6
 	xor	$t3,$t7
 
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key0,16
 	xor	$s0,$t0
 	xor	$s1,$t1
@@ -657,6 +655,12 @@ $code.=<<___;
 	.set	reorder
 	$PTR_LA	$Tbl,AES_Te		# PIC-ified 'load address'
 
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	lw	$s0,0($inp)
+	lw	$s1,4($inp)
+	lw	$s2,8($inp)
+	lw	$s3,12($inp)
+#else
 	lwl	$s0,0+$MSB($inp)
 	lwl	$s1,4+$MSB($inp)
 	lwl	$s2,8+$MSB($inp)
@@ -665,9 +669,16 @@ $code.=<<___;
 	lwr	$s1,4+$LSB($inp)
 	lwr	$s2,8+$LSB($inp)
 	lwr	$s3,12+$LSB($inp)
+#endif
 
 	bal	_mips_AES_encrypt
 
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	sw	$s0,0($out)
+	sw	$s1,4($out)
+	sw	$s2,8($out)
+	sw	$s3,12($out)
+#else
 	swr	$s0,0+$LSB($out)
 	swr	$s1,4+$LSB($out)
 	swr	$s2,8+$LSB($out)
@@ -676,6 +687,7 @@ $code.=<<___;
 	swl	$s1,4+$MSB($out)
 	swl	$s2,8+$MSB($out)
 	swl	$s3,12+$MSB($out)
+#endif
 
 	.set	noreorder
 	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
@@ -720,7 +732,7 @@ _mips_AES_decrypt:
 	xor	$s2,$t2
 	xor	$s3,$t3
 
-	sub	$cnt,1
+	subu	$cnt,1
 #if defined(__mips_smartmips)
 	ext	$i0,$s3,16,8
 .Loop_dec:
@@ -792,7 +804,7 @@ _mips_AES_decrypt:
 	xor	$t2,$t6
 	xor	$t3,$t7
 
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key0,16
 	xor	$s0,$t0
 	xor	$s1,$t1
@@ -985,7 +997,7 @@ _mips_AES_decrypt:
 	xor	$t2,$t6
 	xor	$t3,$t7
 
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key0,16
 	xor	$s0,$t0
 	xor	$s1,$t1
@@ -1228,6 +1240,12 @@ $code.=<<___;
 	.set	reorder
 	$PTR_LA	$Tbl,AES_Td		# PIC-ified 'load address'
 
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	lw	$s0,0($inp)
+	lw	$s1,4($inp)
+	lw	$s2,8($inp)
+	lw	$s3,12($inp)
+#else
 	lwl	$s0,0+$MSB($inp)
 	lwl	$s1,4+$MSB($inp)
 	lwl	$s2,8+$MSB($inp)
@@ -1236,9 +1254,16 @@ $code.=<<___;
 	lwr	$s1,4+$LSB($inp)
 	lwr	$s2,8+$LSB($inp)
 	lwr	$s3,12+$LSB($inp)
+#endif
 
 	bal	_mips_AES_decrypt
 
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	sw	$s0,0($out)
+	sw	$s1,4($out)
+	sw	$s2,8($out)
+	sw	$s3,12($out)
+#else
 	swr	$s0,0+$LSB($out)
 	swr	$s1,4+$LSB($out)
 	swr	$s2,8+$LSB($out)
@@ -1247,6 +1272,7 @@ $code.=<<___;
 	swl	$s1,4+$MSB($out)
 	swl	$s2,8+$MSB($out)
 	swl	$s3,12+$MSB($out)
+#endif
 
 	.set	noreorder
 	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
@@ -1295,35 +1321,52 @@ _mips_AES_set_encrypt_key:
 	$PTR_ADD $rcon,$Tbl,256
 
 	.set	reorder
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	lw	$rk0,0($inp)		# load 128 bits
+	lw	$rk1,4($inp)
+	lw	$rk2,8($inp)
+	lw	$rk3,12($inp)
+#else
 	lwl	$rk0,0+$MSB($inp)	# load 128 bits
 	lwl	$rk1,4+$MSB($inp)
 	lwl	$rk2,8+$MSB($inp)
 	lwl	$rk3,12+$MSB($inp)
-	li	$at,128
 	lwr	$rk0,0+$LSB($inp)
 	lwr	$rk1,4+$LSB($inp)
 	lwr	$rk2,8+$LSB($inp)
 	lwr	$rk3,12+$LSB($inp)
+#endif
+	li	$at,128
 	.set	noreorder
 	beq	$bits,$at,.L128bits
 	li	$cnt,10
 
 	.set	reorder
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	lw	$rk4,16($inp)		# load 192 bits
+	lw	$rk5,20($inp)
+#else
 	lwl	$rk4,16+$MSB($inp)	# load 192 bits
 	lwl	$rk5,20+$MSB($inp)
-	li	$at,192
 	lwr	$rk4,16+$LSB($inp)
 	lwr	$rk5,20+$LSB($inp)
+#endif
+	li	$at,192
 	.set	noreorder
 	beq	$bits,$at,.L192bits
 	li	$cnt,8
 
 	.set	reorder
+#if defined(_MIPS_ARCH_MIPS32R6) || defined(_MIPS_ARCH_MIPS64R6)
+	lw	$rk6,24($inp)		# load 256 bits
+	lw	$rk7,28($inp)
+#else
 	lwl	$rk6,24+$MSB($inp)	# load 256 bits
 	lwl	$rk7,28+$MSB($inp)
-	li	$at,256
 	lwr	$rk6,24+$LSB($inp)
 	lwr	$rk7,28+$LSB($inp)
+#endif
+	li	$at,256
 	.set	noreorder
 	beq	$bits,$at,.L256bits
 	li	$cnt,7
@@ -1353,7 +1396,7 @@ _mips_AES_set_encrypt_key:
 	sw	$rk1,4($key)
 	sw	$rk2,8($key)
 	sw	$rk3,12($key)
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key,16
 
 	_bias	$i0,24
@@ -1410,7 +1453,7 @@ _mips_AES_set_encrypt_key:
 	sw	$rk3,12($key)
 	sw	$rk4,16($key)
 	sw	$rk5,20($key)
-	sub	$cnt,1
+	subu	$cnt,1
 	$PTR_ADD $key,24
 
 	_bias	$i0,24
@@ -1471,7 +1514,7 @@ _mips_AES_set_encrypt_key:
 	sw	$rk5,20($key)
 	sw	$rk6,24($key)
 	sw	$rk7,28($key)
-	sub	$cnt,1
+	subu	$cnt,1
 
 	_bias	$i0,24
 	_bias	$i1,16
@@ -1653,7 +1696,7 @@ $code.=<<___;
 
 	lw	$tp1,16($key)		# modulo-scheduled
 	lui	$x80808080,0x8080
-	sub	$cnt,1
+	subu	$cnt,1
 	or	$x80808080,0x8080
 	sll	$cnt,2
 	$PTR_ADD $key,16
@@ -1716,7 +1759,7 @@ $code.=<<___;
 	lw	$tp1,4($key)		# modulo-scheduled
 	xor	$tpe,$tp2
 #endif
-	sub	$cnt,1
+	subu	$cnt,1
 	sw	$tpe,0($key)
 	$PTR_ADD $key,4
 	bnez	$cnt,.Lmix