All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric Biggers <ebiggers@kernel.org>
To: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, Ard Biesheuvel <ardb@kernel.org>,
	"Jason A . Donenfeld" <Jason@zx2c4.com>,
	Herbert Xu <herbert@gondor.apana.org.au>,
	linux-arm-kernel@lists.infradead.org,
	linuxppc-dev@lists.ozlabs.org, linux-riscv@lists.infradead.org,
	linux-s390@vger.kernel.org, x86@kernel.org
Subject: Re: [PATCH 00/19] GHASH library
Date: Mon, 23 Mar 2026 17:50:08 -0700	[thread overview]
Message-ID: <20260324005008.GA230053@quark> (raw)
In-Reply-To: <20260319061723.1140720-1-ebiggers@kernel.org>

On Wed, Mar 18, 2026 at 11:17:01PM -0700, Eric Biggers wrote:
> This series is targeting libcrypto-next.  It can also be retrieved from:
> 
>     git fetch https://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux.git ghash-lib-v1
> 
> This series migrates the standalone GHASH code to lib/crypto/, then
> converts the "gcm" template and AES-GCM library code to use it.  (GHASH
> is the universal hash function used by GCM mode.)  As was the case with
> POLYVAL and Poly1305 as well, the library is a much better fit for it.
> 
> Since GHASH and POLYVAL are closely related and it often makes sense to
> implement one in terms of the other, the existing "polyval" library
> module is renamed to "gf128hash" and the GHASH support is added to it.
> 
> The generic implementation of GHASH is also replaced with a better one
> utilizing the existing polyval_mul_generic().
> 
> Note that some GHASH implementations, often faster ones using more
> recent CPU features, still exist in arch/*/crypto/ as internal
> components of AES-GCM implementations.  Those are left as-is for now.
> The goal with this GHASH library is just to provide parity with the
> existing standalone GHASH support, which is used when a full
> implementation of AES-GCM (or ${someothercipher}-GCM, if another block
> cipher is being used) is unavailable.  Migrating the
> architecture-optimized AES-GCM code to lib/crypto/ will be a next step.
> 

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux.git/log/?h=libcrypto-next

As usual, the s390 code will need to be tested by someone who has the
privilege of access to a z/Architecture mainframe.  That is the only way
to test that code, given that the s390 community has not yet updated
QEMU to support the CPACF_KIMD_GHASH instruction.

From another review pass I also folded in some trivial cleanups that
don't seem worth sending a v2 for unless something else comes up.
Removed a definition I forgot to remove, dropped unnecessary rename of
'h' to 'k', improved consistency in a couple places, etc.

diff --git a/arch/arm64/crypto/ghash-ce-glue.c b/arch/arm64/crypto/ghash-ce-glue.c
index c74066d430fa..eaf2932ceaf5 100644
--- a/arch/arm64/crypto/ghash-ce-glue.c
+++ b/arch/arm64/crypto/ghash-ce-glue.c
@@ -35,10 +35,6 @@ struct arm_ghash_key {
 	u64			h[4][2];
 };
 
-struct arm_ghash_desc_ctx {
-	u64 digest[GHASH_DIGEST_SIZE/sizeof(u64)];
-};
-
 struct gcm_aes_ctx {
 	struct aes_enckey	aes_key;
 	u8			nonce[RFC4106_NONCE_SIZE];
diff --git a/lib/crypto/arm/gf128hash.h b/lib/crypto/arm/gf128hash.h
index cb929bed29d5..c33c8cbe51fe 100644
--- a/lib/crypto/arm/gf128hash.h
+++ b/lib/crypto/arm/gf128hash.h
@@ -12,7 +12,7 @@
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_neon);
 
 void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
-			   const u8 *src, const struct polyval_elem *k);
+			   const u8 *src, const struct polyval_elem *h);
 
 #define ghash_blocks_arch ghash_blocks_arch
 static void ghash_blocks_arch(struct polyval_elem *acc,
diff --git a/lib/crypto/arm/ghash-neon-core.S b/lib/crypto/arm/ghash-neon-core.S
index bf423fb06a75..eeffd12504a9 100644
--- a/lib/crypto/arm/ghash-neon-core.S
+++ b/lib/crypto/arm/ghash-neon-core.S
@@ -181,7 +181,7 @@
 	/*
 	 * void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
 	 *			      const u8 *src,
-	 *			      const struct polyval_elem *k)
+	 *			      const struct polyval_elem *h)
 	 */
 ENTRY(pmull_ghash_update_p8)
 	vld1.64		{SHASH}, [r3]
diff --git a/lib/crypto/arm64/gf128hash.h b/lib/crypto/arm64/gf128hash.h
index d5ef1b1b77e1..b2c85585b758 100644
--- a/lib/crypto/arm64/gf128hash.h
+++ b/lib/crypto/arm64/gf128hash.h
@@ -12,14 +12,14 @@
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_asimd);
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_pmull);
 
+asmlinkage void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
+				      const u8 *src,
+				      const struct polyval_elem *h);
 asmlinkage void polyval_mul_pmull(struct polyval_elem *a,
 				  const struct polyval_elem *b);
 asmlinkage void polyval_blocks_pmull(struct polyval_elem *acc,
 				     const struct polyval_key *key,
 				     const u8 *data, size_t nblocks);
-asmlinkage void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
-				      const u8 *src,
-				      const struct polyval_elem *k);
 
 #define polyval_preparekey_arch polyval_preparekey_arch
 static void polyval_preparekey_arch(struct polyval_key *key,
@@ -91,8 +91,8 @@ static void ghash_blocks_arch(struct polyval_elem *acc,
 	if (static_branch_likely(&have_asimd) && may_use_simd()) {
 		do {
 			/* Allow rescheduling every 4 KiB. */
-			size_t n =
-				min_t(size_t, nblocks, 4096 / GHASH_BLOCK_SIZE);
+			size_t n = min_t(size_t, nblocks,
+					 4096 / GHASH_BLOCK_SIZE);
 
 			scoped_ksimd()
 				pmull_ghash_update_p8(n, acc, data, &key->h);
diff --git a/lib/crypto/arm64/ghash-neon-core.S b/lib/crypto/arm64/ghash-neon-core.S
index eadd6da47247..85b20fcd98fe 100644
--- a/lib/crypto/arm64/ghash-neon-core.S
+++ b/lib/crypto/arm64/ghash-neon-core.S
@@ -180,7 +180,7 @@
 	/*
 	 * void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
 	 *			      const u8 *src,
-	 *			      const struct polyval_elem *k)
+	 *			      const struct polyval_elem *h)
 	 */
 SYM_FUNC_START(pmull_ghash_update_p8)
 	ld1		{SHASH.2d}, [x3]
diff --git a/lib/crypto/riscv/ghash-riscv64-zvkg.S b/lib/crypto/riscv/ghash-riscv64-zvkg.S
index 2839ff1a990c..6a2a2f2bc7c8 100644
--- a/lib/crypto/riscv/ghash-riscv64-zvkg.S
+++ b/lib/crypto/riscv/ghash-riscv64-zvkg.S
@@ -55,6 +55,8 @@
 // void ghash_zvkg(u8 accumulator[GHASH_BLOCK_SIZE],
 //		   const u8 key[GHASH_BLOCK_SIZE],
 //		   const u8 *data, size_t nblocks);
+//
+// |nblocks| must be nonzero.
 SYM_FUNC_START(ghash_zvkg)
 	vsetivli	zero, 4, e32, m1, ta, ma
 	vle32.v		v1, (ACCUMULATOR)
diff --git a/lib/crypto/tests/Kconfig b/lib/crypto/tests/Kconfig
index 279ff1a339be..5b60d5c3644b 100644
--- a/lib/crypto/tests/Kconfig
+++ b/lib/crypto/tests/Kconfig
@@ -41,7 +41,7 @@ config CRYPTO_LIB_GHASH_KUNIT_TEST
 	default KUNIT_ALL_TESTS
 	select CRYPTO_LIB_BENCHMARK_VISIBLE
 	help
-	  KUnit tests for GHASH library functions.
+	  KUnit tests for the GHASH library functions.
 
 config CRYPTO_LIB_MD5_KUNIT_TEST
 	tristate "KUnit tests for MD5" if !KUNIT_ALL_TESTS


WARNING: multiple messages have this Message-ID (diff)
From: Eric Biggers <ebiggers@kernel.org>
To: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, Ard Biesheuvel <ardb@kernel.org>,
	"Jason A . Donenfeld" <Jason@zx2c4.com>,
	Herbert Xu <herbert@gondor.apana.org.au>,
	linux-arm-kernel@lists.infradead.org,
	linuxppc-dev@lists.ozlabs.org, linux-riscv@lists.infradead.org,
	linux-s390@vger.kernel.org, x86@kernel.org
Subject: Re: [PATCH 00/19] GHASH library
Date: Mon, 23 Mar 2026 17:50:08 -0700	[thread overview]
Message-ID: <20260324005008.GA230053@quark> (raw)
In-Reply-To: <20260319061723.1140720-1-ebiggers@kernel.org>

On Wed, Mar 18, 2026 at 11:17:01PM -0700, Eric Biggers wrote:
> This series is targeting libcrypto-next.  It can also be retrieved from:
> 
>     git fetch https://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux.git ghash-lib-v1
> 
> This series migrates the standalone GHASH code to lib/crypto/, then
> converts the "gcm" template and AES-GCM library code to use it.  (GHASH
> is the universal hash function used by GCM mode.)  As was the case with
> POLYVAL and Poly1305 as well, the library is a much better fit for it.
> 
> Since GHASH and POLYVAL are closely related and it often makes sense to
> implement one in terms of the other, the existing "polyval" library
> module is renamed to "gf128hash" and the GHASH support is added to it.
> 
> The generic implementation of GHASH is also replaced with a better one
> utilizing the existing polyval_mul_generic().
> 
> Note that some GHASH implementations, often faster ones using more
> recent CPU features, still exist in arch/*/crypto/ as internal
> components of AES-GCM implementations.  Those are left as-is for now.
> The goal with this GHASH library is just to provide parity with the
> existing standalone GHASH support, which is used when a full
> implementation of AES-GCM (or ${someothercipher}-GCM, if another block
> cipher is being used) is unavailable.  Migrating the
> architecture-optimized AES-GCM code to lib/crypto/ will be a next step.
> 

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux.git/log/?h=libcrypto-next

As usual, the s390 code will need to be tested by someone who has the
privilege of access to a z/Architecture mainframe.  That is the only way
to test that code, given that the s390 community has not yet updated
QEMU to support the CPACF_KIMD_GHASH instruction.

From another review pass I also folded in some trivial cleanups that
don't seem worth sending a v2 for unless something else comes up.
Removed a definition I forgot to remove, dropped unnecessary rename of
'h' to 'k', improved consistency in a couple places, etc.

diff --git a/arch/arm64/crypto/ghash-ce-glue.c b/arch/arm64/crypto/ghash-ce-glue.c
index c74066d430fa..eaf2932ceaf5 100644
--- a/arch/arm64/crypto/ghash-ce-glue.c
+++ b/arch/arm64/crypto/ghash-ce-glue.c
@@ -35,10 +35,6 @@ struct arm_ghash_key {
 	u64			h[4][2];
 };
 
-struct arm_ghash_desc_ctx {
-	u64 digest[GHASH_DIGEST_SIZE/sizeof(u64)];
-};
-
 struct gcm_aes_ctx {
 	struct aes_enckey	aes_key;
 	u8			nonce[RFC4106_NONCE_SIZE];
diff --git a/lib/crypto/arm/gf128hash.h b/lib/crypto/arm/gf128hash.h
index cb929bed29d5..c33c8cbe51fe 100644
--- a/lib/crypto/arm/gf128hash.h
+++ b/lib/crypto/arm/gf128hash.h
@@ -12,7 +12,7 @@
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_neon);
 
 void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
-			   const u8 *src, const struct polyval_elem *k);
+			   const u8 *src, const struct polyval_elem *h);
 
 #define ghash_blocks_arch ghash_blocks_arch
 static void ghash_blocks_arch(struct polyval_elem *acc,
diff --git a/lib/crypto/arm/ghash-neon-core.S b/lib/crypto/arm/ghash-neon-core.S
index bf423fb06a75..eeffd12504a9 100644
--- a/lib/crypto/arm/ghash-neon-core.S
+++ b/lib/crypto/arm/ghash-neon-core.S
@@ -181,7 +181,7 @@
 	/*
 	 * void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
 	 *			      const u8 *src,
-	 *			      const struct polyval_elem *k)
+	 *			      const struct polyval_elem *h)
 	 */
 ENTRY(pmull_ghash_update_p8)
 	vld1.64		{SHASH}, [r3]
diff --git a/lib/crypto/arm64/gf128hash.h b/lib/crypto/arm64/gf128hash.h
index d5ef1b1b77e1..b2c85585b758 100644
--- a/lib/crypto/arm64/gf128hash.h
+++ b/lib/crypto/arm64/gf128hash.h
@@ -12,14 +12,14 @@
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_asimd);
 static __ro_after_init DEFINE_STATIC_KEY_FALSE(have_pmull);
 
+asmlinkage void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
+				      const u8 *src,
+				      const struct polyval_elem *h);
 asmlinkage void polyval_mul_pmull(struct polyval_elem *a,
 				  const struct polyval_elem *b);
 asmlinkage void polyval_blocks_pmull(struct polyval_elem *acc,
 				     const struct polyval_key *key,
 				     const u8 *data, size_t nblocks);
-asmlinkage void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
-				      const u8 *src,
-				      const struct polyval_elem *k);
 
 #define polyval_preparekey_arch polyval_preparekey_arch
 static void polyval_preparekey_arch(struct polyval_key *key,
@@ -91,8 +91,8 @@ static void ghash_blocks_arch(struct polyval_elem *acc,
 	if (static_branch_likely(&have_asimd) && may_use_simd()) {
 		do {
 			/* Allow rescheduling every 4 KiB. */
-			size_t n =
-				min_t(size_t, nblocks, 4096 / GHASH_BLOCK_SIZE);
+			size_t n = min_t(size_t, nblocks,
+					 4096 / GHASH_BLOCK_SIZE);
 
 			scoped_ksimd()
 				pmull_ghash_update_p8(n, acc, data, &key->h);
diff --git a/lib/crypto/arm64/ghash-neon-core.S b/lib/crypto/arm64/ghash-neon-core.S
index eadd6da47247..85b20fcd98fe 100644
--- a/lib/crypto/arm64/ghash-neon-core.S
+++ b/lib/crypto/arm64/ghash-neon-core.S
@@ -180,7 +180,7 @@
 	/*
 	 * void pmull_ghash_update_p8(size_t blocks, struct polyval_elem *dg,
 	 *			      const u8 *src,
-	 *			      const struct polyval_elem *k)
+	 *			      const struct polyval_elem *h)
 	 */
 SYM_FUNC_START(pmull_ghash_update_p8)
 	ld1		{SHASH.2d}, [x3]
diff --git a/lib/crypto/riscv/ghash-riscv64-zvkg.S b/lib/crypto/riscv/ghash-riscv64-zvkg.S
index 2839ff1a990c..6a2a2f2bc7c8 100644
--- a/lib/crypto/riscv/ghash-riscv64-zvkg.S
+++ b/lib/crypto/riscv/ghash-riscv64-zvkg.S
@@ -55,6 +55,8 @@
 // void ghash_zvkg(u8 accumulator[GHASH_BLOCK_SIZE],
 //		   const u8 key[GHASH_BLOCK_SIZE],
 //		   const u8 *data, size_t nblocks);
+//
+// |nblocks| must be nonzero.
 SYM_FUNC_START(ghash_zvkg)
 	vsetivli	zero, 4, e32, m1, ta, ma
 	vle32.v		v1, (ACCUMULATOR)
diff --git a/lib/crypto/tests/Kconfig b/lib/crypto/tests/Kconfig
index 279ff1a339be..5b60d5c3644b 100644
--- a/lib/crypto/tests/Kconfig
+++ b/lib/crypto/tests/Kconfig
@@ -41,7 +41,7 @@ config CRYPTO_LIB_GHASH_KUNIT_TEST
 	default KUNIT_ALL_TESTS
 	select CRYPTO_LIB_BENCHMARK_VISIBLE
 	help
-	  KUnit tests for GHASH library functions.
+	  KUnit tests for the GHASH library functions.
 
 config CRYPTO_LIB_MD5_KUNIT_TEST
 	tristate "KUnit tests for MD5" if !KUNIT_ALL_TESTS

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

  parent reply	other threads:[~2026-03-24  0:50 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-19  6:17 [PATCH 00/19] GHASH library Eric Biggers
2026-03-19  6:17 ` Eric Biggers
2026-03-19  6:17 ` [PATCH 01/19] lib/crypto: gf128hash: Rename polyval module to gf128hash Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 02/19] lib/crypto: gf128hash: Support GF128HASH_ARCH without all POLYVAL functions Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 03/19] lib/crypto: gf128hash: Add GHASH support Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 04/19] lib/crypto: tests: Add KUnit tests for GHASH Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 05/19] crypto: arm/ghash - Make the "ghash" crypto_shash NEON-only Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 06/19] crypto: arm/ghash - Move NEON GHASH assembly into its own file Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 07/19] lib/crypto: arm/ghash: Migrate optimized code into library Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 08/19] crypto: arm64/ghash - Move NEON GHASH assembly into its own file Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 09/19] lib/crypto: arm64/ghash: Migrate optimized code into library Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 10/19] crypto: arm64/aes-gcm - Rename struct ghash_key and make fixed-sized Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 11/19] lib/crypto: powerpc/ghash: Migrate optimized code into library Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 12/19] lib/crypto: riscv/ghash: " Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 13/19] lib/crypto: s390/ghash: " Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 14/19] lib/crypto: x86/ghash: " Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 15/19] crypto: gcm - Use GHASH library instead of crypto_ahash Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 16/19] crypto: ghash - Remove ghash from crypto_shash API Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 17/19] lib/crypto: gf128mul: Remove unused 4k_lle functions Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 18/19] lib/crypto: gf128hash: Remove unused content from ghash.h Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-19  6:17 ` [PATCH 19/19] lib/crypto: aesgcm: Use GHASH library API Eric Biggers
2026-03-19  6:17   ` Eric Biggers
2026-03-23 14:14 ` [PATCH 00/19] GHASH library Ard Biesheuvel
2026-03-23 14:14   ` Ard Biesheuvel
2026-03-24  0:50 ` Eric Biggers [this message]
2026-03-24  0:50   ` Eric Biggers

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260324005008.GA230053@quark \
    --to=ebiggers@kernel.org \
    --cc=Jason@zx2c4.com \
    --cc=ardb@kernel.org \
    --cc=herbert@gondor.apana.org.au \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-crypto@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.