crypto: powerpc/aes - key handling

Key generation for big endian core routines.
Signed-off-by: Markus Stockhausen <stockhausen@collogia.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

arch/powerpc/crypto/aes-spe-keys.S

+/*
+ * Key handling functions for PPC AES implementation
+ *
+ * Copyright (c) 2015 Markus Stockhausen <stockhausen@collogia.de>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <asm/ppc_asm.h>
+
+#ifdef __BIG_ENDIAN__
+#define LOAD_KEY(d, s, off) \
+	lwz		d,off(s);
+#else
+#define LOAD_KEY(d, s, off) \
+	li		r0,off; \
+	lwbrx		d,s,r0;
+#endif
+
+#define INITIALIZE_KEY \
+	stwu		r1,-32(r1);	/* create stack frame		*/ \
+	stw		r14,8(r1);	/* save registers		*/ \
+	stw		r15,12(r1);					   \
+	stw		r16,16(r1);
+
+#define FINALIZE_KEY \
+	lwz		r14,8(r1);	/* restore registers		*/ \
+	lwz		r15,12(r1);					   \
+	lwz		r16,16(r1);					   \
+	xor		r5,r5,r5;	/* clear sensitive data		*/ \
+	xor		r6,r6,r6;					   \
+	xor		r7,r7,r7;					   \
+	xor		r8,r8,r8;					   \
+	xor		r9,r9,r9;					   \
+	xor		r10,r10,r10;					   \
+	xor		r11,r11,r11;					   \
+	xor		r12,r12,r12;					   \
+	addi		r1,r1,32;	/* cleanup stack		*/
+
+#define LS_BOX(r, t1, t2) \
+	lis		t2,PPC_AES_4K_ENCTAB@h;				   \
+	ori		t2,t2,PPC_AES_4K_ENCTAB@l;			   \
+	rlwimi		t2,r,4,20,27;					   \
+	lbz		t1,8(t2);					   \
+	rlwimi		r,t1,0,24,31;					   \
+	rlwimi		t2,r,28,20,27;					   \
+	lbz		t1,8(t2);					   \
+	rlwimi		r,t1,8,16,23;					   \
+	rlwimi		t2,r,20,20,27;					   \
+	lbz		t1,8(t2);					   \
+	rlwimi		r,t1,16,8,15;					   \
+	rlwimi		t2,r,12,20,27;					   \
+	lbz		t1,8(t2);					   \
+	rlwimi		r,t1,24,0,7;
+
+#define GF8_MUL(out, in, t1, t2) \
+	lis t1,0x8080;			/* multiplication in GF8	*/ \
+	ori t1,t1,0x8080; 						   \
+	and t1,t1,in; 							   \
+	srwi t1,t1,7; 							   \
+	mulli t1,t1,0x1b; 						   \
+	lis t2,0x7f7f; 							   \
+	ori t2,t2,0x7f7f; 						   \
+	and t2,t2,in; 							   \
+	slwi t2,t2,1; 							   \
+	xor out,t1,t2;
+
+/*
+ * ppc_expand_key_128(u32 *key_enc, const u8 *key)
+ *
+ * Expand 128 bit key into 176 bytes encryption key. It consists of
+ * key itself plus 10 rounds with 16 bytes each
+ *
+ */
+_GLOBAL(ppc_expand_key_128)
+	INITIALIZE_KEY
+	LOAD_KEY(r5,r4,0)
+	LOAD_KEY(r6,r4,4)
+	LOAD_KEY(r7,r4,8)
+	LOAD_KEY(r8,r4,12)
+	stw		r5,0(r3)	/* key[0..3] = input data	*/
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	li		r16,10		/* 10 expansion rounds		*/
+	lis		r0,0x0100	/* RCO(1)			*/
+ppc_expand_128_loop:
+	addi		r3,r3,16
+	mr		r14,r8		/* apply LS_BOX to 4th temp	*/
+	rotlwi		r14,r14,8
+	LS_BOX(r14, r15, r4)
+	xor		r14,r14,r0
+	xor		r5,r5,r14	/* xor next 4 keys		*/
+	xor		r6,r6,r5
+	xor		r7,r7,r6
+	xor		r8,r8,r7
+	stw		r5,0(r3)	/* store next 4 keys		*/
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	GF8_MUL(r0, r0, r4, r14)	/* multiply RCO by 2 in GF	*/
+	subi		r16,r16,1
+	cmpwi		r16,0
+	bt		eq,ppc_expand_128_end
+	b		ppc_expand_128_loop
+ppc_expand_128_end:
+	FINALIZE_KEY
+	blr
+
+/*
+ * ppc_expand_key_192(u32 *key_enc, const u8 *key)
+ *
+ * Expand 192 bit key into 208 bytes encryption key. It consists of key
+ * itself plus 12 rounds with 16 bytes each
+ *
+ */
+_GLOBAL(ppc_expand_key_192)
+	INITIALIZE_KEY
+	LOAD_KEY(r5,r4,0)
+	LOAD_KEY(r6,r4,4)
+	LOAD_KEY(r7,r4,8)
+	LOAD_KEY(r8,r4,12)
+	LOAD_KEY(r9,r4,16)
+	LOAD_KEY(r10,r4,20)
+	stw		r5,0(r3)
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	stw		r9,16(r3)
+	stw		r10,20(r3)
+	li		r16,8		/* 8 expansion rounds		*/
+	lis		r0,0x0100	/* RCO(1)			*/
+ppc_expand_192_loop:
+	addi		r3,r3,24
+	mr		r14,r10		/* apply LS_BOX to 6th temp	*/
+	rotlwi		r14,r14,8
+	LS_BOX(r14, r15, r4)
+	xor		r14,r14,r0
+	xor		r5,r5,r14	/* xor next 6 keys		*/
+	xor		r6,r6,r5
+	xor		r7,r7,r6
+	xor		r8,r8,r7
+	xor		r9,r9,r8
+	xor		r10,r10,r9
+	stw		r5,0(r3)
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	subi		r16,r16,1
+	cmpwi		r16,0		/* last round early kick out	*/
+	bt		eq,ppc_expand_192_end
+	stw		r9,16(r3)
+	stw		r10,20(r3)
+	GF8_MUL(r0, r0, r4, r14)	/* multiply RCO GF8		*/
+	b		ppc_expand_192_loop
+ppc_expand_192_end:
+	FINALIZE_KEY
+	blr
+
+/*
+ * ppc_expand_key_256(u32 *key_enc, const u8 *key)
+ *
+ * Expand 256 bit key into 240 bytes encryption key. It consists of key
+ * itself plus 14 rounds with 16 bytes each
+ *
+ */
+_GLOBAL(ppc_expand_key_256)
+	INITIALIZE_KEY
+	LOAD_KEY(r5,r4,0)
+	LOAD_KEY(r6,r4,4)
+	LOAD_KEY(r7,r4,8)
+	LOAD_KEY(r8,r4,12)
+	LOAD_KEY(r9,r4,16)
+	LOAD_KEY(r10,r4,20)
+	LOAD_KEY(r11,r4,24)
+	LOAD_KEY(r12,r4,28)
+	stw		r5,0(r3)
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	stw		r9,16(r3)
+	stw		r10,20(r3)
+	stw		r11,24(r3)
+	stw		r12,28(r3)
+	li		r16,7		/* 7 expansion rounds		*/
+	lis		r0,0x0100	/* RCO(1)			*/
+ppc_expand_256_loop:
+	addi		r3,r3,32
+	mr		r14,r12		/* apply LS_BOX to 8th temp	*/
+	rotlwi		r14,r14,8
+	LS_BOX(r14, r15, r4)
+	xor		r14,r14,r0
+	xor		r5,r5,r14	/* xor 4 keys			*/
+	xor		r6,r6,r5
+	xor		r7,r7,r6
+	xor		r8,r8,r7
+	mr		r14,r8
+	LS_BOX(r14, r15, r4)		/* apply LS_BOX to 4th temp	*/
+	xor		r9,r9,r14	/* xor 4 keys			*/
+	xor		r10,r10,r9
+	xor		r11,r11,r10
+	xor		r12,r12,r11
+	stw		r5,0(r3)
+	stw		r6,4(r3)
+	stw		r7,8(r3)
+	stw		r8,12(r3)
+	subi		r16,r16,1
+	cmpwi		r16,0		/* last round early kick out	*/
+	bt		eq,ppc_expand_256_end
+	stw		r9,16(r3)
+	stw		r10,20(r3)
+	stw		r11,24(r3)
+	stw		r12,28(r3)
+	GF8_MUL(r0, r0, r4, r14)
+	b		ppc_expand_256_loop
+ppc_expand_256_end:
+	FINALIZE_KEY
+	blr
+
+/*
+ * ppc_generate_decrypt_key: derive decryption key from encryption key
+ * number of bytes to handle are calculated from length of key (16/24/32)
+ *
+ */
+_GLOBAL(ppc_generate_decrypt_key)
+	addi		r6,r5,24
+	slwi		r6,r6,2
+	lwzx		r7,r4,r6	/* first/last 4 words are same	*/
+	stw		r7,0(r3)
+	lwz		r7,0(r4)
+	stwx		r7,r3,r6
+	addi		r6,r6,4
+	lwzx		r7,r4,r6
+	stw		r7,4(r3)
+	lwz		r7,4(r4)
+	stwx		r7,r3,r6
+	addi		r6,r6,4
+	lwzx		r7,r4,r6
+	stw		r7,8(r3)
+	lwz		r7,8(r4)
+	stwx		r7,r3,r6
+	addi		r6,r6,4
+	lwzx		r7,r4,r6
+	stw		r7,12(r3)
+	lwz		r7,12(r4)
+	stwx		r7,r3,r6
+	addi		r3,r3,16
+	add		r4,r4,r6
+	subi		r4,r4,28
+	addi		r5,r5,20
+	srwi		r5,r5,2
+ppc_generate_decrypt_block:
+	li	r6,4
+	mtctr	r6
+ppc_generate_decrypt_word:
+	lwz		r6,0(r4)
+	GF8_MUL(r7, r6, r0, r7)
+	GF8_MUL(r8, r7, r0, r8)
+	GF8_MUL(r9, r8, r0, r9)
+	xor		r10,r9,r6
+	xor		r11,r7,r8
+	xor		r11,r11,r9
+	xor		r12,r7,r10
+	rotrwi		r12,r12,24
+	xor		r11,r11,r12
+	xor		r12,r8,r10
+	rotrwi		r12,r12,16
+	xor		r11,r11,r12
+	rotrwi		r12,r10,8
+	xor		r11,r11,r12
+	stw		r11,0(r3)
+	addi		r3,r3,4
+	addi		r4,r4,4
+	bdnz		ppc_generate_decrypt_word
+	subi		r4,r4,32
+	subi		r5,r5,1
+	cmpwi		r5,0
+	bt		gt,ppc_generate_decrypt_block
+	blr