crypto: algapi - make crypto_xor() take separate dst and src arguments

There are quite a number of occurrences in the kernel of the pattern

  if (dst != src)
          memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
  crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);

or

  crypto_xor(keystream, src, nbytes);
  memcpy(dst, keystream, nbytes);

where crypto_xor() is preceded or followed by a memcpy() invocation
that is only there because crypto_xor() uses its output parameter as
one of the inputs. To avoid having to add new instances of this pattern
in the arm64 code, which will be refactored to implement non-SIMD
fallbacks, add an alternative implementation called crypto_xor_cpy(),
taking separate input and output arguments. This removes the need for
the separate memcpy().
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

arch/arm/crypto/aes-ce-glue.c

@@ -285,9 +285,7 @@ static int ctr_encrypt(struct skcipher_request *req)
 
 		ce_aes_ctr_encrypt(tail, NULL, (u8 *)ctx->key_enc,
 				   num_rounds(ctx), blocks, walk.iv);
-		if (tdst != tsrc)
-			memcpy(tdst, tsrc, nbytes);
-		crypto_xor(tdst, tail, nbytes);
+		crypto_xor_cpy(tdst, tsrc, tail, nbytes);
 		err = skcipher_walk_done(&walk, 0);
 	}
 	kernel_neon_end();

arch/arm/crypto/aes-neonbs-glue.c

@@ -221,9 +221,8 @@ static int ctr_encrypt(struct skcipher_request *req)
 			u8 *dst = walk.dst.virt.addr + blocks * AES_BLOCK_SIZE;
 			u8 *src = walk.src.virt.addr + blocks * AES_BLOCK_SIZE;
 
-			if (dst != src)
-				memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
-			crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
+			crypto_xor_cpy(dst, src, final,
+				       walk.total % AES_BLOCK_SIZE);
 
 			err = skcipher_walk_done(&walk, 0);
 			break;

arch/arm64/crypto/aes-glue.c

@@ -241,9 +241,7 @@ static int ctr_encrypt(struct skcipher_request *req)
 
 		aes_ctr_encrypt(tail, NULL, (u8 *)ctx->key_enc, rounds,
 				blocks, walk.iv, first);
-		if (tdst != tsrc)
-			memcpy(tdst, tsrc, nbytes);
-		crypto_xor(tdst, tail, nbytes);
+		crypto_xor_cpy(tdst, tsrc, tail, nbytes);
 		err = skcipher_walk_done(&walk, 0);
 	}
 	kernel_neon_end();

arch/arm64/crypto/aes-neonbs-glue.c

@@ -224,9 +224,8 @@ static int ctr_encrypt(struct skcipher_request *req)
 			u8 *dst = walk.dst.virt.addr + blocks * AES_BLOCK_SIZE;
 			u8 *src = walk.src.virt.addr + blocks * AES_BLOCK_SIZE;
 
-			if (dst != src)
-				memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
-			crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
+			crypto_xor_cpy(dst, src, final,
+				       walk.total % AES_BLOCK_SIZE);
 
 			err = skcipher_walk_done(&walk, 0);
 			break;

arch/sparc/crypto/aes_glue.c

@@ -344,8 +344,7 @@ static void ctr_crypt_final(struct crypto_sparc64_aes_ctx *ctx,
 
 	ctx->ops->ecb_encrypt(&ctx->key[0], (const u64 *)ctrblk,
 			      keystream, AES_BLOCK_SIZE);
-	crypto_xor((u8 *) keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, (u8 *) keystream, src, nbytes);
 	crypto_inc(ctrblk, AES_BLOCK_SIZE);
 }
 

arch/x86/crypto/aesni-intel_glue.c

@@ -475,8 +475,8 @@ static void ctr_crypt_final(struct crypto_aes_ctx *ctx,
 	unsigned int nbytes = walk->nbytes;
 
 	aesni_enc(ctx, keystream, ctrblk);
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
+
 	crypto_inc(ctrblk, AES_BLOCK_SIZE);
 }
 

arch/x86/crypto/blowfish_glue.c

@@ -271,8 +271,7 @@ static void ctr_crypt_final(struct bf_ctx *ctx, struct blkcipher_walk *walk)
 	unsigned int nbytes = walk->nbytes;
 
 	blowfish_enc_blk(ctx, keystream, ctrblk);
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
 
 	crypto_inc(ctrblk, BF_BLOCK_SIZE);
 }

arch/x86/crypto/cast5_avx_glue.c

@@ -256,8 +256,7 @@ static void ctr_crypt_final(struct blkcipher_desc *desc,
 	unsigned int nbytes = walk->nbytes;
 
 	__cast5_encrypt(ctx, keystream, ctrblk);
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
 
 	crypto_inc(ctrblk, CAST5_BLOCK_SIZE);
 }

arch/x86/crypto/des3_ede_glue.c

@@ -277,8 +277,7 @@ static void ctr_crypt_final(struct des3_ede_x86_ctx *ctx,
 	unsigned int nbytes = walk->nbytes;
 
 	des3_ede_enc_blk(ctx, keystream, ctrblk);
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
 
 	crypto_inc(ctrblk, DES3_EDE_BLOCK_SIZE);
 }

crypto/ctr.c

@@ -65,8 +65,7 @@ static void crypto_ctr_crypt_final(struct blkcipher_walk *walk,
 	unsigned int nbytes = walk->nbytes;
 
 	crypto_cipher_encrypt_one(tfm, keystream, ctrblk);
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
 
 	crypto_inc(ctrblk, bsize);
 }

crypto/pcbc.c

@@ -55,8 +55,7 @@ static int crypto_pcbc_encrypt_segment(struct skcipher_request *req,
 	do {
 		crypto_xor(iv, src, bsize);
 		crypto_cipher_encrypt_one(tfm, dst, iv);
-		memcpy(iv, dst, bsize);
-		crypto_xor(iv, src, bsize);
+		crypto_xor_cpy(iv, dst, src, bsize);
 
 		src += bsize;
 		dst += bsize;
@@ -79,8 +78,7 @@ static int crypto_pcbc_encrypt_inplace(struct skcipher_request *req,
 		memcpy(tmpbuf, src, bsize);
 		crypto_xor(iv, src, bsize);
 		crypto_cipher_encrypt_one(tfm, src, iv);
-		memcpy(iv, tmpbuf, bsize);
-		crypto_xor(iv, src, bsize);
+		crypto_xor_cpy(iv, tmpbuf, src, bsize);
 
 		src += bsize;
 	} while ((nbytes -= bsize) >= bsize);
@@ -127,8 +125,7 @@ static int crypto_pcbc_decrypt_segment(struct skcipher_request *req,
 	do {
 		crypto_cipher_decrypt_one(tfm, dst, src);
 		crypto_xor(dst, iv, bsize);
-		memcpy(iv, src, bsize);
-		crypto_xor(iv, dst, bsize);
+		crypto_xor_cpy(iv, dst, src, bsize);
 
 		src += bsize;
 		dst += bsize;
@@ -153,8 +150,7 @@ static int crypto_pcbc_decrypt_inplace(struct skcipher_request *req,
 		memcpy(tmpbuf, src, bsize);
 		crypto_cipher_decrypt_one(tfm, src, src);
 		crypto_xor(src, iv, bsize);
-		memcpy(iv, tmpbuf, bsize);
-		crypto_xor(iv, src, bsize);
+		crypto_xor_cpy(iv, src, tmpbuf, bsize);
 
 		src += bsize;
 	} while ((nbytes -= bsize) >= bsize);

drivers/crypto/vmx/aes_ctr.c

@@ -104,8 +104,7 @@ static void p8_aes_ctr_final(struct p8_aes_ctr_ctx *ctx,
 	pagefault_enable();
 	preempt_enable();
 
-	crypto_xor(keystream, src, nbytes);
-	memcpy(dst, keystream, nbytes);
+	crypto_xor_cpy(dst, keystream, src, nbytes);
 	crypto_inc(ctrblk, AES_BLOCK_SIZE);
 }
 

drivers/md/dm-crypt.c

@@ -758,9 +758,8 @@ static int crypt_iv_tcw_whitening(struct crypt_config *cc,
 	int i, r;
 
 	/* xor whitening with sector number */
-	memcpy(buf, tcw->whitening, TCW_WHITENING_SIZE);
-	crypto_xor(buf, (u8 *)&sector, 8);
-	crypto_xor(&buf[8], (u8 *)&sector, 8);
+	crypto_xor_cpy(buf, tcw->whitening, (u8 *)&sector, 8);
+	crypto_xor_cpy(&buf[8], tcw->whitening + 8, (u8 *)&sector, 8);
 
 	/* calculate crc32 for every 32bit part and xor it */
 	desc->tfm = tcw->crc32_tfm;
@@ -805,10 +804,10 @@ static int crypt_iv_tcw_gen(struct crypt_config *cc, u8 *iv,
 	}
 
 	/* Calculate IV */
-	memcpy(iv, tcw->iv_seed, cc->iv_size);
-	crypto_xor(iv, (u8 *)&sector, 8);
+	crypto_xor_cpy(iv, tcw->iv_seed, (u8 *)&sector, 8);
 	if (cc->iv_size > 8)
-		crypto_xor(&iv[8], (u8 *)&sector, cc->iv_size - 8);
+		crypto_xor_cpy(&iv[8], tcw->iv_seed + 8, (u8 *)&sector,
+			       cc->iv_size - 8);
 
 	return r;
 }

include/crypto/algapi.h

@@ -211,6 +211,25 @@ static inline void crypto_xor(u8 *dst, const u8 *src, unsigned int size)
 	}
 }
 
+static inline void crypto_xor_cpy(u8 *dst, const u8 *src1, const u8 *src2,
+				  unsigned int size)
+{
+	if (IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
+	    __builtin_constant_p(size) &&
+	    (size % sizeof(unsigned long)) == 0) {
+		unsigned long *d = (unsigned long *)dst;
+		unsigned long *s1 = (unsigned long *)src1;
+		unsigned long *s2 = (unsigned long *)src2;
+
+		while (size > 0) {
+			*d++ = *s1++ ^ *s2++;
+			size -= sizeof(unsigned long);
+		}
+	} else {
+		__crypto_xor(dst, src1, src2, size);
+	}
+}
+
 int blkcipher_walk_done(struct blkcipher_desc *desc,
 			struct blkcipher_walk *walk, int err);
 int blkcipher_walk_virt(struct blkcipher_desc *desc,