1 files changed, 418 insertions, 202 deletions

diff --git a/arch/arm/crypto/ghash-ce-glue.c b/arch/arm/crypto/ghash-ce-glue.c
index f13401f3e669..a52dcc8c1e33 100644
--- a/arch/arm/crypto/ghash-ce-glue.c
+++ b/arch/arm/crypto/ghash-ce-glue.c
@@ -2,44 +2,52 @@
 /*
  * Accelerated GHASH implementation with ARMv8 vmull.p64 instructions.
  *
- * Copyright (C) 2015 - 2018 Linaro Ltd. <ard.biesheuvel@linaro.org>
+ * Copyright (C) 2015 - 2018 Linaro Ltd.
+ * Copyright (C) 2023 Google LLC.
  */
 
 #include <asm/hwcap.h>
 #include <asm/neon.h>
-#include <asm/simd.h>
-#include <asm/unaligned.h>
+#include <crypto/aes.h>
 #include <crypto/b128ops.h>
-#include <crypto/cryptd.h>
-#include <crypto/internal/hash.h>
-#include <crypto/internal/simd.h>
+#include <crypto/gcm.h>
 #include <crypto/gf128mul.h>
+#include <crypto/ghash.h>
+#include <crypto/internal/aead.h>
+#include <crypto/internal/hash.h>
+#include <crypto/internal/skcipher.h>
+#include <crypto/scatterwalk.h>
 #include <linux/cpufeature.h>
-#include <linux/crypto.h>
+#include <linux/errno.h>
 #include <linux/jump_label.h>
+#include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/string.h>
+#include <linux/unaligned.h>
 
 MODULE_DESCRIPTION("GHASH hash function using ARMv8 Crypto Extensions");
-MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
-MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Ard Biesheuvel <ardb@kernel.org>");
+MODULE_LICENSE("GPL");
 MODULE_ALIAS_CRYPTO("ghash");
+MODULE_ALIAS_CRYPTO("gcm(aes)");
+MODULE_ALIAS_CRYPTO("rfc4106(gcm(aes))");
 
-#define GHASH_BLOCK_SIZE	16
-#define GHASH_DIGEST_SIZE	16
+#define RFC4106_NONCE_SIZE	4
 
 struct ghash_key {
 	be128	k;
 	u64	h[][2];
 };
 
-struct ghash_desc_ctx {
-	u64 digest[GHASH_DIGEST_SIZE/sizeof(u64)];
-	u8 buf[GHASH_BLOCK_SIZE];
-	u32 count;
+struct gcm_key {
+	u64	h[4][2];
+	u32	rk[AES_MAX_KEYLENGTH_U32];
+	int	rounds;
+	u8	nonce[];	// for RFC4106 nonce
 };
 
-struct ghash_async_ctx {
-	struct cryptd_ahash *cryptd_tfm;
+struct arm_ghash_desc_ctx {
+	u64 digest[GHASH_DIGEST_SIZE/sizeof(u64)];
 };
 
 asmlinkage void pmull_ghash_update_p64(int blocks, u64 dg[], const char *src,
@@ -52,94 +60,69 @@ static __ro_after_init DEFINE_STATIC_KEY_FALSE(use_p64);
 
 static int ghash_init(struct shash_desc *desc)
 {
-	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	struct arm_ghash_desc_ctx *ctx = shash_desc_ctx(desc);
 
-	*ctx = (struct ghash_desc_ctx){};
+	*ctx = (struct arm_ghash_desc_ctx){};
 	return 0;
 }
 
 static void ghash_do_update(int blocks, u64 dg[], const char *src,
 			    struct ghash_key *key, const char *head)
 {
-	if (likely(crypto_simd_usable())) {
-		kernel_neon_begin();
-		if (static_branch_likely(&use_p64))
-			pmull_ghash_update_p64(blocks, dg, src, key->h, head);
-		else
-			pmull_ghash_update_p8(blocks, dg, src, key->h, head);
-		kernel_neon_end();
-	} else {
-		be128 dst = { cpu_to_be64(dg[1]), cpu_to_be64(dg[0]) };
-
-		do {
-			const u8 *in = src;
-
-			if (head) {
-				in = head;
-				blocks++;
-				head = NULL;
-			} else {
-				src += GHASH_BLOCK_SIZE;
-			}
-
-			crypto_xor((u8 *)&dst, in, GHASH_BLOCK_SIZE);
-			gf128mul_lle(&dst, &key->k);
-		} while (--blocks);
-
-		dg[0] = be64_to_cpu(dst.b);
-		dg[1] = be64_to_cpu(dst.a);
-	}
+	kernel_neon_begin();
+	if (static_branch_likely(&use_p64))
+		pmull_ghash_update_p64(blocks, dg, src, key->h, head);
+	else
+		pmull_ghash_update_p8(blocks, dg, src, key->h, head);
+	kernel_neon_end();
 }
 
 static int ghash_update(struct shash_desc *desc, const u8 *src,
 			unsigned int len)
 {
-	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
-	unsigned int partial = ctx->count % GHASH_BLOCK_SIZE;
-
-	ctx->count += len;
+	struct ghash_key *key = crypto_shash_ctx(desc->tfm);
+	struct arm_ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	int blocks;
 
-	if ((partial + len) >= GHASH_BLOCK_SIZE) {
-		struct ghash_key *key = crypto_shash_ctx(desc->tfm);
-		int blocks;
+	blocks = len / GHASH_BLOCK_SIZE;
+	ghash_do_update(blocks, ctx->digest, src, key, NULL);
+	return len - blocks * GHASH_BLOCK_SIZE;
+}
 
-		if (partial) {
-			int p = GHASH_BLOCK_SIZE - partial;
+static int ghash_export(struct shash_desc *desc, void *out)
+{
+	struct arm_ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	u8 *dst = out;
 
-			memcpy(ctx->buf + partial, src, p);
-			src += p;
-			len -= p;
-		}
+	put_unaligned_be64(ctx->digest[1], dst);
+	put_unaligned_be64(ctx->digest[0], dst + 8);
+	return 0;
+}
 
-		blocks = len / GHASH_BLOCK_SIZE;
-		len %= GHASH_BLOCK_SIZE;
+static int ghash_import(struct shash_desc *desc, const void *in)
+{
+	struct arm_ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	const u8 *src = in;
 
-		ghash_do_update(blocks, ctx->digest, src, key,
-				partial ? ctx->buf : NULL);
-		src += blocks * GHASH_BLOCK_SIZE;
-		partial = 0;
-	}
-	if (len)
-		memcpy(ctx->buf + partial, src, len);
+	ctx->digest[1] = get_unaligned_be64(src);
+	ctx->digest[0] = get_unaligned_be64(src + 8);
 	return 0;
 }
 
-static int ghash_final(struct shash_desc *desc, u8 *dst)
+static int ghash_finup(struct shash_desc *desc, const u8 *src,
+		       unsigned int len, u8 *dst)
 {
-	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
-	unsigned int partial = ctx->count % GHASH_BLOCK_SIZE;
+	struct ghash_key *key = crypto_shash_ctx(desc->tfm);
+	struct arm_ghash_desc_ctx *ctx = shash_desc_ctx(desc);
 
-	if (partial) {
-		struct ghash_key *key = crypto_shash_ctx(desc->tfm);
+	if (len) {
+		u8 buf[GHASH_BLOCK_SIZE] = {};
 
-		memset(ctx->buf + partial, 0, GHASH_BLOCK_SIZE - partial);
-		ghash_do_update(1, ctx->digest, ctx->buf, key, NULL);
+		memcpy(buf, src, len);
+		ghash_do_update(1, ctx->digest, buf, key, NULL);
+		memzero_explicit(buf, sizeof(buf));
 	}
-	put_unaligned_be64(ctx->digest[1], dst);
-	put_unaligned_be64(ctx->digest[0], dst + 8);
-
-	*ctx = (struct ghash_desc_ctx){};
-	return 0;
+	return ghash_export(desc, dst);
 }
 
 static void ghash_reflect(u64 h[], const be128 *k)
@@ -184,165 +167,393 @@ static struct shash_alg ghash_alg = {
 	.digestsize		= GHASH_DIGEST_SIZE,
 	.init			= ghash_init,
 	.update			= ghash_update,
-	.final			= ghash_final,
+	.finup			= ghash_finup,
 	.setkey			= ghash_setkey,
-	.descsize		= sizeof(struct ghash_desc_ctx),
+	.export			= ghash_export,
+	.import			= ghash_import,
+	.descsize		= sizeof(struct arm_ghash_desc_ctx),
+	.statesize		= sizeof(struct ghash_desc_ctx),
 
 	.base.cra_name		= "ghash",
-	.base.cra_driver_name	= "ghash-ce-sync",
-	.base.cra_priority	= 300 - 1,
+	.base.cra_driver_name	= "ghash-ce",
+	.base.cra_priority	= 300,
+	.base.cra_flags		= CRYPTO_AHASH_ALG_BLOCK_ONLY,
 	.base.cra_blocksize	= GHASH_BLOCK_SIZE,
 	.base.cra_ctxsize	= sizeof(struct ghash_key) + sizeof(u64[2]),
 	.base.cra_module	= THIS_MODULE,
 };
 
-static int ghash_async_init(struct ahash_request *req)
+void pmull_gcm_encrypt(int blocks, u64 dg[], const char *src,
+		       struct gcm_key const *k, char *dst,
+		       const char *iv, int rounds, u32 counter);
+
+void pmull_gcm_enc_final(int blocks, u64 dg[], char *tag,
+			 struct gcm_key const *k, char *head,
+			 const char *iv, int rounds, u32 counter);
+
+void pmull_gcm_decrypt(int bytes, u64 dg[], const char *src,
+		       struct gcm_key const *k, char *dst,
+		       const char *iv, int rounds, u32 counter);
+
+int pmull_gcm_dec_final(int bytes, u64 dg[], char *tag,
+			struct gcm_key const *k, char *head,
+			const char *iv, int rounds, u32 counter,
+			const char *otag, int authsize);
+
+static int gcm_aes_setkey(struct crypto_aead *tfm, const u8 *inkey,
+			  unsigned int keylen)
 {
-	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
-	struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
-	struct crypto_shash *child = cryptd_ahash_child(cryptd_tfm);
-
-	desc->tfm = child;
-	return crypto_shash_init(desc);
+	struct gcm_key *ctx = crypto_aead_ctx(tfm);
+	struct crypto_aes_ctx aes_ctx;
+	be128 h, k;
+	int ret;
+
+	ret = aes_expandkey(&aes_ctx, inkey, keylen);
+	if (ret)
+		return -EINVAL;
+
+	aes_encrypt(&aes_ctx, (u8 *)&k, (u8[AES_BLOCK_SIZE]){});
+
+	memcpy(ctx->rk, aes_ctx.key_enc, sizeof(ctx->rk));
+	ctx->rounds = 6 + keylen / 4;
+
+	memzero_explicit(&aes_ctx, sizeof(aes_ctx));
+
+	ghash_reflect(ctx->h[0], &k);
+
+	h = k;
+	gf128mul_lle(&h, &k);
+	ghash_reflect(ctx->h[1], &h);
+
+	gf128mul_lle(&h, &k);
+	ghash_reflect(ctx->h[2], &h);
+
+	gf128mul_lle(&h, &k);
+	ghash_reflect(ctx->h[3], &h);
+
+	return 0;
 }
 
-static int ghash_async_update(struct ahash_request *req)
+static int gcm_aes_setauthsize(struct crypto_aead *tfm, unsigned int authsize)
 {
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
-
-	if (!crypto_simd_usable() ||
-	    (in_atomic() && cryptd_ahash_queued(cryptd_tfm))) {
-		memcpy(cryptd_req, req, sizeof(*req));
-		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
-		return crypto_ahash_update(cryptd_req);
-	} else {
-		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
-		return shash_ahash_update(req, desc);
-	}
+	return crypto_gcm_check_authsize(authsize);
 }
 
-static int ghash_async_final(struct ahash_request *req)
+static void gcm_update_mac(u64 dg[], const u8 *src, int count, u8 buf[],
+			   int *buf_count, struct gcm_key *ctx)
 {
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
-
-	if (!crypto_simd_usable() ||
-	    (in_atomic() && cryptd_ahash_queued(cryptd_tfm))) {
-		memcpy(cryptd_req, req, sizeof(*req));
-		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
-		return crypto_ahash_final(cryptd_req);
-	} else {
-		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
-		return crypto_shash_final(desc, req->result);
+	if (*buf_count > 0) {
+		int buf_added = min(count, GHASH_BLOCK_SIZE - *buf_count);
+
+		memcpy(&buf[*buf_count], src, buf_added);
+
+		*buf_count += buf_added;
+		src += buf_added;
+		count -= buf_added;
+	}
+
+	if (count >= GHASH_BLOCK_SIZE || *buf_count == GHASH_BLOCK_SIZE) {
+		int blocks = count / GHASH_BLOCK_SIZE;
+
+		pmull_ghash_update_p64(blocks, dg, src, ctx->h,
+				       *buf_count ? buf : NULL);
+
+		src += blocks * GHASH_BLOCK_SIZE;
+		count %= GHASH_BLOCK_SIZE;
+		*buf_count = 0;
+	}
+
+	if (count > 0) {
+		memcpy(buf, src, count);
+		*buf_count = count;
 	}
 }
 
-static int ghash_async_digest(struct ahash_request *req)
+static void gcm_calculate_auth_mac(struct aead_request *req, u64 dg[], u32 len)
 {
-	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
-
-	if (!crypto_simd_usable() ||
-	    (in_atomic() && cryptd_ahash_queued(cryptd_tfm))) {
-		memcpy(cryptd_req, req, sizeof(*req));
-		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
-		return crypto_ahash_digest(cryptd_req);
-	} else {
-		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
-		struct crypto_shash *child = cryptd_ahash_child(cryptd_tfm);
-
-		desc->tfm = child;
-		return shash_ahash_digest(req, desc);
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct gcm_key *ctx = crypto_aead_ctx(aead);
+	u8 buf[GHASH_BLOCK_SIZE];
+	struct scatter_walk walk;
+	int buf_count = 0;
+
+	scatterwalk_start(&walk, req->src);
+
+	do {
+		unsigned int n;
+
+		n = scatterwalk_next(&walk, len);
+		gcm_update_mac(dg, walk.addr, n, buf, &buf_count, ctx);
+		scatterwalk_done_src(&walk,  n);
+
+		if (unlikely(len / SZ_4K > (len - n) / SZ_4K)) {
+			kernel_neon_end();
+			kernel_neon_begin();
+		}
+
+		len -= n;
+	} while (len);
+
+	if (buf_count) {
+		memset(&buf[buf_count], 0, GHASH_BLOCK_SIZE - buf_count);
+		pmull_ghash_update_p64(1, dg, buf, ctx->h, NULL);
 	}
 }
 
-static int ghash_async_import(struct ahash_request *req, const void *in)
+static int gcm_encrypt(struct aead_request *req, const u8 *iv, u32 assoclen)
 {
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct gcm_key *ctx = crypto_aead_ctx(aead);
+	struct skcipher_walk walk;
+	u8 buf[AES_BLOCK_SIZE];
+	u32 counter = 2;
+	u64 dg[2] = {};
+	be128 lengths;
+	const u8 *src;
+	u8 *tag, *dst;
+	int tail, err;
+
+	err = skcipher_walk_aead_encrypt(&walk, req, false);
+
+	kernel_neon_begin();
+
+	if (assoclen)
+		gcm_calculate_auth_mac(req, dg, assoclen);
+
+	src = walk.src.virt.addr;
+	dst = walk.dst.virt.addr;
+
+	while (walk.nbytes >= AES_BLOCK_SIZE) {
+		int nblocks = walk.nbytes / AES_BLOCK_SIZE;
+
+		pmull_gcm_encrypt(nblocks, dg, src, ctx, dst, iv,
+				  ctx->rounds, counter);
+		counter += nblocks;
+
+		if (walk.nbytes == walk.total) {
+			src += nblocks * AES_BLOCK_SIZE;
+			dst += nblocks * AES_BLOCK_SIZE;
+			break;
+		}
+
+		kernel_neon_end();
+
+		err = skcipher_walk_done(&walk,
+					 walk.nbytes % AES_BLOCK_SIZE);
+		if (err)
+			return err;
+
+		src = walk.src.virt.addr;
+		dst = walk.dst.virt.addr;
+
+		kernel_neon_begin();
+	}
+
+
+	lengths.a = cpu_to_be64(assoclen * 8);
+	lengths.b = cpu_to_be64(req->cryptlen * 8);
+
+	tag = (u8 *)&lengths;
+	tail = walk.nbytes % AES_BLOCK_SIZE;
+
+	/*
+	 * Bounce via a buffer unless we are encrypting in place and src/dst
+	 * are not pointing to the start of the walk buffer. In that case, we
+	 * can do a NEON load/xor/store sequence in place as long as we move
+	 * the plain/ciphertext and keystream to the start of the register. If
+	 * not, do a memcpy() to the end of the buffer so we can reuse the same
+	 * logic.
+	 */
+	if (unlikely(tail && (tail == walk.nbytes || src != dst)))
+		src = memcpy(buf + sizeof(buf) - tail, src, tail);
 
-	desc->tfm = cryptd_ahash_child(ctx->cryptd_tfm);
+	pmull_gcm_enc_final(tail, dg, tag, ctx, (u8 *)src, iv,
+			    ctx->rounds, counter);
+	kernel_neon_end();
 
-	return crypto_shash_import(desc, in);
+	if (unlikely(tail && src != dst))
+		memcpy(dst, src, tail);
+
+	if (walk.nbytes) {
+		err = skcipher_walk_done(&walk, 0);
+		if (err)
+			return err;
+	}
+
+	/* copy authtag to end of dst */
+	scatterwalk_map_and_copy(tag, req->dst, req->assoclen + req->cryptlen,
+				 crypto_aead_authsize(aead), 1);
+
+	return 0;
 }
 
-static int ghash_async_export(struct ahash_request *req, void *out)
+static int gcm_decrypt(struct aead_request *req, const u8 *iv, u32 assoclen)
 {
-	struct ahash_request *cryptd_req = ahash_request_ctx(req);
-	struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct gcm_key *ctx = crypto_aead_ctx(aead);
+	int authsize = crypto_aead_authsize(aead);
+	struct skcipher_walk walk;
+	u8 otag[AES_BLOCK_SIZE];
+	u8 buf[AES_BLOCK_SIZE];
+	u32 counter = 2;
+	u64 dg[2] = {};
+	be128 lengths;
+	const u8 *src;
+	u8 *tag, *dst;
+	int tail, err, ret;
+
+	scatterwalk_map_and_copy(otag, req->src,
+				 req->assoclen + req->cryptlen - authsize,
+				 authsize, 0);
+
+	err = skcipher_walk_aead_decrypt(&walk, req, false);
+
+	kernel_neon_begin();
+
+	if (assoclen)
+		gcm_calculate_auth_mac(req, dg, assoclen);
+
+	src = walk.src.virt.addr;
+	dst = walk.dst.virt.addr;
+
+	while (walk.nbytes >= AES_BLOCK_SIZE) {
+		int nblocks = walk.nbytes / AES_BLOCK_SIZE;
+
+		pmull_gcm_decrypt(nblocks, dg, src, ctx, dst, iv,
+				  ctx->rounds, counter);
+		counter += nblocks;
+
+		if (walk.nbytes == walk.total) {
+			src += nblocks * AES_BLOCK_SIZE;
+			dst += nblocks * AES_BLOCK_SIZE;
+			break;
+		}
+
+		kernel_neon_end();
+
+		err = skcipher_walk_done(&walk,
+					 walk.nbytes % AES_BLOCK_SIZE);
+		if (err)
+			return err;
+
+		src = walk.src.virt.addr;
+		dst = walk.dst.virt.addr;
+
+		kernel_neon_begin();
+	}
 
-	return crypto_shash_export(desc, out);
+	lengths.a = cpu_to_be64(assoclen * 8);
+	lengths.b = cpu_to_be64((req->cryptlen - authsize) * 8);
+
+	tag = (u8 *)&lengths;
+	tail = walk.nbytes % AES_BLOCK_SIZE;
+
+	if (unlikely(tail && (tail == walk.nbytes || src != dst)))
+		src = memcpy(buf + sizeof(buf) - tail, src, tail);
+
+	ret = pmull_gcm_dec_final(tail, dg, tag, ctx, (u8 *)src, iv,
+				  ctx->rounds, counter, otag, authsize);
+	kernel_neon_end();
+
+	if (unlikely(tail && src != dst))
+		memcpy(dst, src, tail);
+
+	if (walk.nbytes) {
+		err = skcipher_walk_done(&walk, 0);
+		if (err)
+			return err;
+	}
+
+	return ret ? -EBADMSG : 0;
 }
 
-static int ghash_async_setkey(struct crypto_ahash *tfm, const u8 *key,
-			      unsigned int keylen)
+static int gcm_aes_encrypt(struct aead_request *req)
 {
-	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
-	struct crypto_ahash *child = &ctx->cryptd_tfm->base;
+	return gcm_encrypt(req, req->iv, req->assoclen);
+}
 
-	crypto_ahash_clear_flags(child, CRYPTO_TFM_REQ_MASK);
-	crypto_ahash_set_flags(child, crypto_ahash_get_flags(tfm)
-			       & CRYPTO_TFM_REQ_MASK);
-	return crypto_ahash_setkey(child, key, keylen);
+static int gcm_aes_decrypt(struct aead_request *req)
+{
+	return gcm_decrypt(req, req->iv, req->assoclen);
 }
 
-static int ghash_async_init_tfm(struct crypto_tfm *tfm)
+static int rfc4106_setkey(struct crypto_aead *tfm, const u8 *inkey,
+			  unsigned int keylen)
 {
-	struct cryptd_ahash *cryptd_tfm;
-	struct ghash_async_ctx *ctx = crypto_tfm_ctx(tfm);
+	struct gcm_key *ctx = crypto_aead_ctx(tfm);
+	int err;
 
-	cryptd_tfm = cryptd_alloc_ahash("ghash-ce-sync", 0, 0);
-	if (IS_ERR(cryptd_tfm))
-		return PTR_ERR(cryptd_tfm);
-	ctx->cryptd_tfm = cryptd_tfm;
-	crypto_ahash_set_reqsize(__crypto_ahash_cast(tfm),
-				 sizeof(struct ahash_request) +
-				 crypto_ahash_reqsize(&cryptd_tfm->base));
+	keylen -= RFC4106_NONCE_SIZE;
+	err = gcm_aes_setkey(tfm, inkey, keylen);
+	if (err)
+		return err;
 
+	memcpy(ctx->nonce, inkey + keylen, RFC4106_NONCE_SIZE);
 	return 0;
 }
 
-static void ghash_async_exit_tfm(struct crypto_tfm *tfm)
+static int rfc4106_setauthsize(struct crypto_aead *tfm, unsigned int authsize)
 {
-	struct ghash_async_ctx *ctx = crypto_tfm_ctx(tfm);
+	return crypto_rfc4106_check_authsize(authsize);
+}
 
-	cryptd_free_ahash(ctx->cryptd_tfm);
+static int rfc4106_encrypt(struct aead_request *req)
+{
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct gcm_key *ctx = crypto_aead_ctx(aead);
+	u8 iv[GCM_AES_IV_SIZE];
+
+	memcpy(iv, ctx->nonce, RFC4106_NONCE_SIZE);
+	memcpy(iv + RFC4106_NONCE_SIZE, req->iv, GCM_RFC4106_IV_SIZE);
+
+	return crypto_ipsec_check_assoclen(req->assoclen) ?:
+	       gcm_encrypt(req, iv, req->assoclen - GCM_RFC4106_IV_SIZE);
 }
 
-static struct ahash_alg ghash_async_alg = {
-	.init			= ghash_async_init,
-	.update			= ghash_async_update,
-	.final			= ghash_async_final,
-	.setkey			= ghash_async_setkey,
-	.digest			= ghash_async_digest,
-	.import			= ghash_async_import,
-	.export			= ghash_async_export,
-	.halg.digestsize	= GHASH_DIGEST_SIZE,
-	.halg.statesize		= sizeof(struct ghash_desc_ctx),
-	.halg.base		= {
-		.cra_name	= "ghash",
-		.cra_driver_name = "ghash-ce",
-		.cra_priority	= 300,
-		.cra_flags	= CRYPTO_ALG_ASYNC,
-		.cra_blocksize	= GHASH_BLOCK_SIZE,
-		.cra_ctxsize	= sizeof(struct ghash_async_ctx),
-		.cra_module	= THIS_MODULE,
-		.cra_init	= ghash_async_init_tfm,
-		.cra_exit	= ghash_async_exit_tfm,
-	},
-};
+static int rfc4106_decrypt(struct aead_request *req)
+{
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct gcm_key *ctx = crypto_aead_ctx(aead);
+	u8 iv[GCM_AES_IV_SIZE];
+
+	memcpy(iv, ctx->nonce, RFC4106_NONCE_SIZE);
+	memcpy(iv + RFC4106_NONCE_SIZE, req->iv, GCM_RFC4106_IV_SIZE);
+
+	return crypto_ipsec_check_assoclen(req->assoclen) ?:
+	       gcm_decrypt(req, iv, req->assoclen - GCM_RFC4106_IV_SIZE);
+}
+
+static struct aead_alg gcm_aes_algs[] = {{
+	.ivsize			= GCM_AES_IV_SIZE,
+	.chunksize		= AES_BLOCK_SIZE,
+	.maxauthsize		= AES_BLOCK_SIZE,
+	.setkey			= gcm_aes_setkey,
+	.setauthsize		= gcm_aes_setauthsize,
+	.encrypt		= gcm_aes_encrypt,
+	.decrypt		= gcm_aes_decrypt,
+
+	.base.cra_name		= "gcm(aes)",
+	.base.cra_driver_name	= "gcm-aes-ce",
+	.base.cra_priority	= 400,
+	.base.cra_blocksize	= 1,
+	.base.cra_ctxsize	= sizeof(struct gcm_key),
+	.base.cra_module	= THIS_MODULE,
+}, {
+	.ivsize			= GCM_RFC4106_IV_SIZE,
+	.chunksize		= AES_BLOCK_SIZE,
+	.maxauthsize		= AES_BLOCK_SIZE,
+	.setkey			= rfc4106_setkey,
+	.setauthsize		= rfc4106_setauthsize,
+	.encrypt		= rfc4106_encrypt,
+	.decrypt		= rfc4106_decrypt,
+
+	.base.cra_name		= "rfc4106(gcm(aes))",
+	.base.cra_driver_name	= "rfc4106-gcm-aes-ce",
+	.base.cra_priority	= 400,
+	.base.cra_blocksize	= 1,
+	.base.cra_ctxsize	= sizeof(struct gcm_key) + RFC4106_NONCE_SIZE,
+	.base.cra_module	= THIS_MODULE,
+}};
 
 static int __init ghash_ce_mod_init(void)
 {
@@ -352,28 +563,33 @@ static int __init ghash_ce_mod_init(void)
 		return -ENODEV;
 
 	if (elf_hwcap2 & HWCAP2_PMULL) {
+		err = crypto_register_aeads(gcm_aes_algs,
+					    ARRAY_SIZE(gcm_aes_algs));
+		if (err)
+			return err;
 		ghash_alg.base.cra_ctxsize += 3 * sizeof(u64[2]);
 		static_branch_enable(&use_p64);
 	}
 
 	err = crypto_register_shash(&ghash_alg);
 	if (err)
-		return err;
-	err = crypto_register_ahash(&ghash_async_alg);
-	if (err)
-		goto err_shash;
+		goto err_aead;
 
 	return 0;
 
-err_shash:
-	crypto_unregister_shash(&ghash_alg);
+err_aead:
+	if (elf_hwcap2 & HWCAP2_PMULL)
+		crypto_unregister_aeads(gcm_aes_algs,
+					ARRAY_SIZE(gcm_aes_algs));
 	return err;
 }
 
 static void __exit ghash_ce_mod_exit(void)
 {
-	crypto_unregister_ahash(&ghash_async_alg);
 	crypto_unregister_shash(&ghash_alg);
+	if (elf_hwcap2 & HWCAP2_PMULL)
+		crypto_unregister_aeads(gcm_aes_algs,
+					ARRAY_SIZE(gcm_aes_algs));
 }
 
 module_init(ghash_ce_mod_init);