lib/crypto: arm64/sha512: Migrate optimized SHA-512 code to library

Instead of exposing the arm64-optimized SHA-512 code via arm64-specific
crypto_shash algorithms, instead just implement the sha512_blocks()
library function.  This is much simpler, it makes the SHA-512 (and
SHA-384) library functions be arm64-optimized, and it fixes the
longstanding issue where the arm64-optimized SHA-512 code was disabled
by default.  SHA-512 still remains available through crypto_shash, but
individual architectures no longer need to handle it.

To match sha512_blocks(), change the type of the nblocks parameter of
the assembly functions from int or 'unsigned int' to size_t.  Update the
ARMv8 CE assembly function accordingly.  The scalar assembly function
actually already treated it as size_t.

Acked-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20250630160320.2888-9-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@kernel.org>

arch/arm64/configs/defconfig

@@ -1744,7 +1744,6 @@ CONFIG_CRYPTO_ANSI_CPRNG=y
 CONFIG_CRYPTO_USER_API_RNG=m
 CONFIG_CRYPTO_GHASH_ARM64_CE=y
 CONFIG_CRYPTO_SHA1_ARM64_CE=y
-CONFIG_CRYPTO_SHA512_ARM64_CE=m
 CONFIG_CRYPTO_SHA3_ARM64=m
 CONFIG_CRYPTO_SM3_ARM64_CE=m
 CONFIG_CRYPTO_AES_ARM64_CE_BLK=y

arch/arm64/crypto/Kconfig

@@ -36,25 +36,6 @@ config CRYPTO_SHA1_ARM64_CE
 	  Architecture: arm64 using:
 	  - ARMv8 Crypto Extensions
 
-config CRYPTO_SHA512_ARM64
-	tristate "Hash functions: SHA-384 and SHA-512"
-	select CRYPTO_HASH
-	help
-	  SHA-384 and SHA-512 secure hash algorithms (FIPS 180)
-
-	  Architecture: arm64
-
-config CRYPTO_SHA512_ARM64_CE
-	tristate "Hash functions: SHA-384 and SHA-512 (ARMv8 Crypto Extensions)"
-	depends on KERNEL_MODE_NEON
-	select CRYPTO_HASH
-	select CRYPTO_SHA512_ARM64
-	help
-	  SHA-384 and SHA-512 secure hash algorithms (FIPS 180)
-
-	  Architecture: arm64 using:
-	  - ARMv8 Crypto Extensions
-
 config CRYPTO_SHA3_ARM64
 	tristate "Hash functions: SHA-3 (ARMv8.2 Crypto Extensions)"
 	depends on KERNEL_MODE_NEON

arch/arm64/crypto/Makefile

@@ -8,9 +8,6 @@
 obj-$(CONFIG_CRYPTO_SHA1_ARM64_CE) += sha1-ce.o
 sha1-ce-y := sha1-ce-glue.o sha1-ce-core.o
 
-obj-$(CONFIG_CRYPTO_SHA512_ARM64_CE) += sha512-ce.o
-sha512-ce-y := sha512-ce-glue.o sha512-ce-core.o
-
 obj-$(CONFIG_CRYPTO_SHA3_ARM64) += sha3-ce.o
 sha3-ce-y := sha3-ce-glue.o sha3-ce-core.o
 
@@ -53,9 +50,6 @@ aes-ce-blk-y := aes-glue-ce.o aes-ce.o
 obj-$(CONFIG_CRYPTO_AES_ARM64_NEON_BLK) += aes-neon-blk.o
 aes-neon-blk-y := aes-glue-neon.o aes-neon.o
 
-obj-$(CONFIG_CRYPTO_SHA512_ARM64) += sha512-arm64.o
-sha512-arm64-y := sha512-glue.o sha512-core.o
-
 obj-$(CONFIG_CRYPTO_NHPOLY1305_NEON) += nhpoly1305-neon.o
 nhpoly1305-neon-y := nh-neon-core.o nhpoly1305-neon-glue.o
 
@@ -64,11 +58,3 @@ aes-arm64-y := aes-cipher-core.o aes-cipher-glue.o
 
 obj-$(CONFIG_CRYPTO_AES_ARM64_BS) += aes-neon-bs.o
 aes-neon-bs-y := aes-neonbs-core.o aes-neonbs-glue.o
-
-quiet_cmd_perlasm = PERLASM $@
-      cmd_perlasm = $(PERL) $(<) void $(@)
-
-$(obj)/sha512-core.S: $(src)/../lib/crypto/sha2-armv8.pl
-	$(call cmd,perlasm)
-
-clean-files += sha512-core.S

arch/arm64/crypto/sha512-ce-glue.c

@@ -1,96 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0
-/*
- * sha512-ce-glue.c - SHA-384/SHA-512 using ARMv8 Crypto Extensions
- *
- * Copyright (C) 2018 Linaro Ltd <ard.biesheuvel@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include <asm/neon.h>
-#include <crypto/internal/hash.h>
-#include <crypto/sha2.h>
-#include <crypto/sha512_base.h>
-#include <linux/cpufeature.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-
-MODULE_DESCRIPTION("SHA-384/SHA-512 secure hash using ARMv8 Crypto Extensions");
-MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
-MODULE_LICENSE("GPL v2");
-MODULE_ALIAS_CRYPTO("sha384");
-MODULE_ALIAS_CRYPTO("sha512");
-
-asmlinkage int __sha512_ce_transform(struct sha512_state *sst, u8 const *src,
-				     int blocks);
-
-static void sha512_ce_transform(struct sha512_state *sst, u8 const *src,
-				int blocks)
-{
-	do {
-		int rem;
-
-		kernel_neon_begin();
-		rem = __sha512_ce_transform(sst, src, blocks);
-		kernel_neon_end();
-		src += (blocks - rem) * SHA512_BLOCK_SIZE;
-		blocks = rem;
-	} while (blocks);
-}
-
-static int sha512_ce_update(struct shash_desc *desc, const u8 *data,
-			    unsigned int len)
-{
-	return sha512_base_do_update_blocks(desc, data, len,
-					    sha512_ce_transform);
-}
-
-static int sha512_ce_finup(struct shash_desc *desc, const u8 *data,
-			   unsigned int len, u8 *out)
-{
-	sha512_base_do_finup(desc, data, len, sha512_ce_transform);
-	return sha512_base_finish(desc, out);
-}
-
-static struct shash_alg algs[] = { {
-	.init			= sha384_base_init,
-	.update			= sha512_ce_update,
-	.finup			= sha512_ce_finup,
-	.descsize		= SHA512_STATE_SIZE,
-	.digestsize		= SHA384_DIGEST_SIZE,
-	.base.cra_name		= "sha384",
-	.base.cra_driver_name	= "sha384-ce",
-	.base.cra_priority	= 200,
-	.base.cra_flags		= CRYPTO_AHASH_ALG_BLOCK_ONLY |
-				  CRYPTO_AHASH_ALG_FINUP_MAX,
-	.base.cra_blocksize	= SHA512_BLOCK_SIZE,
-	.base.cra_module	= THIS_MODULE,
-}, {
-	.init			= sha512_base_init,
-	.update			= sha512_ce_update,
-	.finup			= sha512_ce_finup,
-	.descsize		= SHA512_STATE_SIZE,
-	.digestsize		= SHA512_DIGEST_SIZE,
-	.base.cra_name		= "sha512",
-	.base.cra_driver_name	= "sha512-ce",
-	.base.cra_priority	= 200,
-	.base.cra_flags		= CRYPTO_AHASH_ALG_BLOCK_ONLY |
-				  CRYPTO_AHASH_ALG_FINUP_MAX,
-	.base.cra_blocksize	= SHA512_BLOCK_SIZE,
-	.base.cra_module	= THIS_MODULE,
-} };
-
-static int __init sha512_ce_mod_init(void)
-{
-	return crypto_register_shashes(algs, ARRAY_SIZE(algs));
-}
-
-static void __exit sha512_ce_mod_fini(void)
-{
-	crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
-}
-
-module_cpu_feature_match(SHA512, sha512_ce_mod_init);
-module_exit(sha512_ce_mod_fini);

arch/arm64/crypto/sha512-glue.c

@@ -1,83 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * Linux/arm64 port of the OpenSSL SHA512 implementation for AArch64
- *
- * Copyright (c) 2016 Linaro Ltd. <ard.biesheuvel@linaro.org>
- */
-
-#include <crypto/internal/hash.h>
-#include <crypto/sha2.h>
-#include <crypto/sha512_base.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-
-MODULE_DESCRIPTION("SHA-384/SHA-512 secure hash for arm64");
-MODULE_AUTHOR("Andy Polyakov <appro@openssl.org>");
-MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
-MODULE_LICENSE("GPL v2");
-MODULE_ALIAS_CRYPTO("sha384");
-MODULE_ALIAS_CRYPTO("sha512");
-
-asmlinkage void sha512_blocks_arch(u64 *digest, const void *data,
-				   unsigned int num_blks);
-
-static void sha512_arm64_transform(struct sha512_state *sst, u8 const *src,
-				   int blocks)
-{
-	sha512_blocks_arch(sst->state, src, blocks);
-}
-
-static int sha512_update_arm64(struct shash_desc *desc, const u8 *data,
-			       unsigned int len)
-{
-	return sha512_base_do_update_blocks(desc, data, len,
-					    sha512_arm64_transform);
-}
-
-static int sha512_finup(struct shash_desc *desc, const u8 *data,
-			unsigned int len, u8 *out)
-{
-	sha512_base_do_finup(desc, data, len, sha512_arm64_transform);
-	return sha512_base_finish(desc, out);
-}
-
-static struct shash_alg algs[] = { {
-	.digestsize		= SHA512_DIGEST_SIZE,
-	.init			= sha512_base_init,
-	.update			= sha512_update_arm64,
-	.finup			= sha512_finup,
-	.descsize		= SHA512_STATE_SIZE,
-	.base.cra_name		= "sha512",
-	.base.cra_driver_name	= "sha512-arm64",
-	.base.cra_priority	= 150,
-	.base.cra_flags		= CRYPTO_AHASH_ALG_BLOCK_ONLY |
-				  CRYPTO_AHASH_ALG_FINUP_MAX,
-	.base.cra_blocksize	= SHA512_BLOCK_SIZE,
-	.base.cra_module	= THIS_MODULE,
-}, {
-	.digestsize		= SHA384_DIGEST_SIZE,
-	.init			= sha384_base_init,
-	.update			= sha512_update_arm64,
-	.finup			= sha512_finup,
-	.descsize		= SHA512_STATE_SIZE,
-	.base.cra_name		= "sha384",
-	.base.cra_driver_name	= "sha384-arm64",
-	.base.cra_priority	= 150,
-	.base.cra_flags		= CRYPTO_AHASH_ALG_BLOCK_ONLY |
-				  CRYPTO_AHASH_ALG_FINUP_MAX,
-	.base.cra_blocksize	= SHA384_BLOCK_SIZE,
-	.base.cra_module	= THIS_MODULE,
-} };
-
-static int __init sha512_mod_init(void)
-{
-	return crypto_register_shashes(algs, ARRAY_SIZE(algs));
-}
-
-static void __exit sha512_mod_fini(void)
-{
-	crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
-}
-
-module_init(sha512_mod_init);
-module_exit(sha512_mod_fini);

lib/crypto/Kconfig

@@ -178,6 +178,7 @@ config CRYPTO_LIB_SHA512_ARCH
 	bool
 	depends on CRYPTO_LIB_SHA512 && !UML
 	default y if ARM && !CPU_V7M
+	default y if ARM64
 
 config CRYPTO_LIB_SM3
 	tristate

lib/crypto/Makefile

@@ -5,6 +5,9 @@ aflags-thumb2-$(CONFIG_THUMB2_KERNEL)  := -U__thumb2__ -D__thumb2__=1
 quiet_cmd_perlasm = PERLASM $@
       cmd_perlasm = $(PERL) $(<) > $(@)
 
+quiet_cmd_perlasm_with_args = PERLASM $@
+      cmd_perlasm_with_args = $(PERL) $(<) void $(@)
+
 obj-$(CONFIG_CRYPTO_LIB_UTILS)			+= libcryptoutils.o
 libcryptoutils-y				:= memneq.o utils.o
 
@@ -82,6 +85,13 @@ clean-files += arm/sha512-core.S
 AFLAGS_arm/sha512-core.o += $(aflags-thumb2-y)
 endif
 
+ifeq ($(CONFIG_ARM64),y)
+libsha512-y += arm64/sha512-core.o
+$(obj)/arm64/sha512-core.S: $(src)/../../arch/arm64/lib/crypto/sha2-armv8.pl
+	$(call cmd,perlasm_with_args)
+clean-files += arm64/sha512-core.S
+libsha512-$(CONFIG_KERNEL_MODE_NEON) += arm64/sha512-ce-core.o
+endif
 endif # CONFIG_CRYPTO_LIB_SHA512_ARCH
 
 obj-$(CONFIG_MPILIB) += mpi/

lib/crypto/arm64/.gitignore

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+sha512-core.S

lib/crypto/arm64/sha512-ce-core.S

@@ -102,8 +102,8 @@
 	.endm
 
 	/*
-	 * int __sha512_ce_transform(struct sha512_state *sst, u8 const *src,
+	 * size_t __sha512_ce_transform(struct sha512_block_state *state,
-	 *			     int blocks)
+	 *				const u8 *data, size_t nblocks);
 	 */
 	.text
 SYM_FUNC_START(__sha512_ce_transform)
@@ -117,7 +117,7 @@ SYM_FUNC_START(__sha512_ce_transform)
 	/* load input */
 0:	ld1		{v12.2d-v15.2d}, [x1], #64
 	ld1		{v16.2d-v19.2d}, [x1], #64
-	sub		w2, w2, #1
+	sub		x2, x2, #1
 
 CPU_LE(	rev64		v12.16b, v12.16b	)
 CPU_LE(	rev64		v13.16b, v13.16b	)
@@ -197,10 +197,10 @@ CPU_LE(	rev64		v19.16b, v19.16b	)
 
 	cond_yield	3f, x4, x5
 	/* handled all input blocks? */
-	cbnz		w2, 0b
+	cbnz		x2, 0b
 
 	/* store new state */
 3:	st1		{v8.2d-v11.2d}, [x0]
-	mov		w0, w2
+	mov		x0, x2
 	ret
 SYM_FUNC_END(__sha512_ce_transform)

lib/crypto/arm64/sha512.h

@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * arm64-optimized SHA-512 block function
+ *
+ * Copyright 2025 Google LLC
+ */
+
+#include <asm/neon.h>
+#include <crypto/internal/simd.h>
+#include <linux/cpufeature.h>
+
+static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_sha512_insns);
+
+asmlinkage void sha512_blocks_arch(struct sha512_block_state *state,
+				   const u8 *data, size_t nblocks);
+asmlinkage size_t __sha512_ce_transform(struct sha512_block_state *state,
+					const u8 *data, size_t nblocks);
+
+static void sha512_blocks(struct sha512_block_state *state,
+			  const u8 *data, size_t nblocks)
+{
+	if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) &&
+	    static_branch_likely(&have_sha512_insns) &&
+	    likely(crypto_simd_usable())) {
+		do {
+			size_t rem;
+
+			kernel_neon_begin();
+			rem = __sha512_ce_transform(state, data, nblocks);
+			kernel_neon_end();
+			data += (nblocks - rem) * SHA512_BLOCK_SIZE;
+			nblocks = rem;
+		} while (nblocks);
+	} else {
+		sha512_blocks_arch(state, data, nblocks);
+	}
+}
+
+#ifdef CONFIG_KERNEL_MODE_NEON
+#define sha512_mod_init_arch sha512_mod_init_arch
+static inline void sha512_mod_init_arch(void)
+{
+	if (cpu_have_named_feature(SHA512))
+		static_branch_enable(&have_sha512_insns);
+}
+#endif /* CONFIG_KERNEL_MODE_NEON */