All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/2] crypto: shash optimizations
@ 2023-10-09  7:32 Eric Biggers
  2023-10-09  7:32 ` [PATCH 1/2] crypto: shash - optimize the default digest and finup Eric Biggers
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: Eric Biggers @ 2023-10-09  7:32 UTC (permalink / raw)
  To: linux-crypto

This series fixes some inefficiencies in crypto_shash_digest() and
crypto_shash_finup(), particularly in cases where the algorithm doesn't
implement ->digest or ->finup respectively.

Eric Biggers (2):
  crypto: shash - optimize the default digest and finup
  crypto: shash - fold shash_digest_unaligned() into
    crypto_shash_digest()

 crypto/shash.c | 27 +++++++++++++++++++--------
 1 file changed, 19 insertions(+), 8 deletions(-)


base-commit: 8468516f9f93a41dc65158b6428a1a1039c68f20
-- 
2.42.0


^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH 1/2] crypto: shash - optimize the default digest and finup
  2023-10-09  7:32 [PATCH 0/2] crypto: shash optimizations Eric Biggers
@ 2023-10-09  7:32 ` Eric Biggers
  2023-10-09  7:32 ` [PATCH 2/2] crypto: shash - fold shash_digest_unaligned() into crypto_shash_digest() Eric Biggers
  2023-10-20  5:51 ` [PATCH 0/2] crypto: shash optimizations Herbert Xu
  2 siblings, 0 replies; 4+ messages in thread
From: Eric Biggers @ 2023-10-09  7:32 UTC (permalink / raw)
  To: linux-crypto

From: Eric Biggers <ebiggers@google.com>

For an shash algorithm that doesn't implement ->digest, currently
crypto_shash_digest() with aligned input makes 5 indirect calls: 1 to
shash_digest_unaligned(), 1 to ->init, 2 to ->update ('alignmask + 1'
bytes, then the rest), then 1 to ->final.  This is true even if the
algorithm implements ->finup.  This is caused by an unnecessary fallback
to code meant to handle unaligned inputs.  In fact,
crypto_shash_digest() already does the needed alignment check earlier.
Therefore, optimize the number of indirect calls for aligned inputs to 3
when the algorithm implements ->finup.  It remains at 5 when the
algorithm implements neither ->finup nor ->digest.

Similarly, for an shash algorithm that doesn't implement ->finup,
currently crypto_shash_finup() with aligned input makes 4 indirect
calls: 1 to shash_finup_unaligned(), 2 to ->update, and
1 to ->final.  Optimize this to 3 calls.

Signed-off-by: Eric Biggers <ebiggers@google.com>
---
 crypto/shash.c | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/crypto/shash.c b/crypto/shash.c
index 1fadb6b59bdcc..d99dc2f94c65f 100644
--- a/crypto/shash.c
+++ b/crypto/shash.c
@@ -184,20 +184,29 @@ int crypto_shash_final(struct shash_desc *desc, u8 *out)
 }
 EXPORT_SYMBOL_GPL(crypto_shash_final);
 
 static int shash_finup_unaligned(struct shash_desc *desc, const u8 *data,
 				 unsigned int len, u8 *out)
 {
 	return shash_update_unaligned(desc, data, len) ?:
 	       shash_final_unaligned(desc, out);
 }
 
+static int shash_default_finup(struct shash_desc *desc, const u8 *data,
+			       unsigned int len, u8 *out)
+{
+	struct shash_alg *shash = crypto_shash_alg(desc->tfm);
+
+	return shash->update(desc, data, len) ?:
+	       shash->final(desc, out);
+}
+
 int crypto_shash_finup(struct shash_desc *desc, const u8 *data,
 		       unsigned int len, u8 *out)
 {
 	struct crypto_shash *tfm = desc->tfm;
 	struct shash_alg *shash = crypto_shash_alg(tfm);
 	unsigned long alignmask = crypto_shash_alignmask(tfm);
 	int err;
 
 	if (IS_ENABLED(CONFIG_CRYPTO_STATS)) {
 		struct crypto_istat_hash *istat = shash_get_stat(shash);
@@ -217,20 +226,29 @@ int crypto_shash_finup(struct shash_desc *desc, const u8 *data,
 EXPORT_SYMBOL_GPL(crypto_shash_finup);
 
 static int shash_digest_unaligned(struct shash_desc *desc, const u8 *data,
 				  unsigned int len, u8 *out)
 {
 	return crypto_shash_init(desc) ?:
 	       shash_update_unaligned(desc, data, len) ?:
 	       shash_final_unaligned(desc, out);
 }
 
+static int shash_default_digest(struct shash_desc *desc, const u8 *data,
+				unsigned int len, u8 *out)
+{
+	struct shash_alg *shash = crypto_shash_alg(desc->tfm);
+
+	return shash->init(desc) ?:
+	       shash->finup(desc, data, len, out);
+}
+
 int crypto_shash_digest(struct shash_desc *desc, const u8 *data,
 			unsigned int len, u8 *out)
 {
 	struct crypto_shash *tfm = desc->tfm;
 	struct shash_alg *shash = crypto_shash_alg(tfm);
 	unsigned long alignmask = crypto_shash_alignmask(tfm);
 	int err;
 
 	if (IS_ENABLED(CONFIG_CRYPTO_STATS)) {
 		struct crypto_istat_hash *istat = shash_get_stat(shash);
@@ -649,23 +667,23 @@ static int shash_prepare_alg(struct shash_alg *alg)
 		return -EINVAL;
 
 	err = hash_prepare_alg(&alg->halg);
 	if (err)
 		return err;
 
 	base->cra_type = &crypto_shash_type;
 	base->cra_flags |= CRYPTO_ALG_TYPE_SHASH;
 
 	if (!alg->finup)
-		alg->finup = shash_finup_unaligned;
+		alg->finup = shash_default_finup;
 	if (!alg->digest)
-		alg->digest = shash_digest_unaligned;
+		alg->digest = shash_default_digest;
 	if (!alg->export) {
 		alg->export = shash_default_export;
 		alg->import = shash_default_import;
 		alg->halg.statesize = alg->descsize;
 	}
 	if (!alg->setkey)
 		alg->setkey = shash_no_setkey;
 
 	return 0;
 }
-- 
2.42.0


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH 2/2] crypto: shash - fold shash_digest_unaligned() into crypto_shash_digest()
  2023-10-09  7:32 [PATCH 0/2] crypto: shash optimizations Eric Biggers
  2023-10-09  7:32 ` [PATCH 1/2] crypto: shash - optimize the default digest and finup Eric Biggers
@ 2023-10-09  7:32 ` Eric Biggers
  2023-10-20  5:51 ` [PATCH 0/2] crypto: shash optimizations Herbert Xu
  2 siblings, 0 replies; 4+ messages in thread
From: Eric Biggers @ 2023-10-09  7:32 UTC (permalink / raw)
  To: linux-crypto

From: Eric Biggers <ebiggers@google.com>

Fold shash_digest_unaligned() into its only remaining caller.  Also,
avoid a redundant check of CRYPTO_TFM_NEED_KEY by replacing the call to
crypto_shash_init() with shash->init(desc).  Finally, replace
shash_update_unaligned() + shash_final_unaligned() with
shash_finup_unaligned() which does exactly that.

Signed-off-by: Eric Biggers <ebiggers@google.com>
---
 crypto/shash.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)

diff --git a/crypto/shash.c b/crypto/shash.c
index d99dc2f94c65f..15fee57cca8ef 100644
--- a/crypto/shash.c
+++ b/crypto/shash.c
@@ -218,28 +218,20 @@ int crypto_shash_finup(struct shash_desc *desc, const u8 *data,
 	if (((unsigned long)data | (unsigned long)out) & alignmask)
 		err = shash_finup_unaligned(desc, data, len, out);
 	else
 		err = shash->finup(desc, data, len, out);
 
 
 	return crypto_shash_errstat(shash, err);
 }
 EXPORT_SYMBOL_GPL(crypto_shash_finup);
 
-static int shash_digest_unaligned(struct shash_desc *desc, const u8 *data,
-				  unsigned int len, u8 *out)
-{
-	return crypto_shash_init(desc) ?:
-	       shash_update_unaligned(desc, data, len) ?:
-	       shash_final_unaligned(desc, out);
-}
-
 static int shash_default_digest(struct shash_desc *desc, const u8 *data,
 				unsigned int len, u8 *out)
 {
 	struct shash_alg *shash = crypto_shash_alg(desc->tfm);
 
 	return shash->init(desc) ?:
 	       shash->finup(desc, data, len, out);
 }
 
 int crypto_shash_digest(struct shash_desc *desc, const u8 *data,
@@ -253,21 +245,22 @@ int crypto_shash_digest(struct shash_desc *desc, const u8 *data,
 	if (IS_ENABLED(CONFIG_CRYPTO_STATS)) {
 		struct crypto_istat_hash *istat = shash_get_stat(shash);
 
 		atomic64_inc(&istat->hash_cnt);
 		atomic64_add(len, &istat->hash_tlen);
 	}
 
 	if (crypto_shash_get_flags(tfm) & CRYPTO_TFM_NEED_KEY)
 		err = -ENOKEY;
 	else if (((unsigned long)data | (unsigned long)out) & alignmask)
-		err = shash_digest_unaligned(desc, data, len, out);
+		err = shash->init(desc) ?:
+		      shash_finup_unaligned(desc, data, len, out);
 	else
 		err = shash->digest(desc, data, len, out);
 
 	return crypto_shash_errstat(shash, err);
 }
 EXPORT_SYMBOL_GPL(crypto_shash_digest);
 
 int crypto_shash_tfm_digest(struct crypto_shash *tfm, const u8 *data,
 			    unsigned int len, u8 *out)
 {
-- 
2.42.0


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH 0/2] crypto: shash optimizations
  2023-10-09  7:32 [PATCH 0/2] crypto: shash optimizations Eric Biggers
  2023-10-09  7:32 ` [PATCH 1/2] crypto: shash - optimize the default digest and finup Eric Biggers
  2023-10-09  7:32 ` [PATCH 2/2] crypto: shash - fold shash_digest_unaligned() into crypto_shash_digest() Eric Biggers
@ 2023-10-20  5:51 ` Herbert Xu
  2 siblings, 0 replies; 4+ messages in thread
From: Herbert Xu @ 2023-10-20  5:51 UTC (permalink / raw)
  To: Eric Biggers; +Cc: linux-crypto

Eric Biggers <ebiggers@kernel.org> wrote:
> This series fixes some inefficiencies in crypto_shash_digest() and
> crypto_shash_finup(), particularly in cases where the algorithm doesn't
> implement ->digest or ->finup respectively.
> 
> Eric Biggers (2):
>  crypto: shash - optimize the default digest and finup
>  crypto: shash - fold shash_digest_unaligned() into
>    crypto_shash_digest()
> 
> crypto/shash.c | 27 +++++++++++++++++++--------
> 1 file changed, 19 insertions(+), 8 deletions(-)
> 
> 
> base-commit: 8468516f9f93a41dc65158b6428a1a1039c68f20

All applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2023-10-20  5:51 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-10-09  7:32 [PATCH 0/2] crypto: shash optimizations Eric Biggers
2023-10-09  7:32 ` [PATCH 1/2] crypto: shash - optimize the default digest and finup Eric Biggers
2023-10-09  7:32 ` [PATCH 2/2] crypto: shash - fold shash_digest_unaligned() into crypto_shash_digest() Eric Biggers
2023-10-20  5:51 ` [PATCH 0/2] crypto: shash optimizations Herbert Xu

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.