[PATCH 0/2] crypto: replace cpumask_next with better API

[PATCH 0/2] crypto: replace cpumask_next with better API

[Yury Norov]

From: Yury Norov [NVIDIA]  <yury.norov@gmail.com>

Recently added cpumask_next_wrap() and cpumask_nth() work better than
cpumask_next(). Use them where appropriate.

Yury Norov [NVIDIA] (2):
  crypto: pcrypt: Optimize pcrypt_aead_init_tfm()
  crypto: caam - Fix opencoded cpumask_next_wrap() in
    caam_drv_ctx_init()

 crypto/pcrypt.c          | 7 ++-----
 drivers/crypto/caam/qi.c | 5 +----
 2 files changed, 3 insertions(+), 9 deletions(-)

-- 
2.43.0

[PATCH 1/2] crypto: pcrypt: Optimize pcrypt_aead_init_tfm()

[Yury Norov]

From: Yury Norov [NVIDIA] <yury.norov@gmail.com>

The function opencodes cpumask_nth(). The dedicated helper is faster
than an open for-loop.

Signed-off-by: Yury Norov [NVIDIA] <yury.norov@gmail.com>
---
 crypto/pcrypt.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/crypto/pcrypt.c b/crypto/pcrypt.c
index c33d29a523e0..c3a9d4f2995c 100644
--- a/crypto/pcrypt.c
+++ b/crypto/pcrypt.c
@@ -178,7 +178,7 @@ static int pcrypt_aead_decrypt(struct aead_request *req)
 
 static int pcrypt_aead_init_tfm(struct crypto_aead *tfm)
 {
-	int cpu, cpu_index;
+	int cpu_index;
 	struct aead_instance *inst = aead_alg_instance(tfm);
 	struct pcrypt_instance_ctx *ictx = aead_instance_ctx(inst);
 	struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(tfm);
@@ -187,10 +187,7 @@ static int pcrypt_aead_init_tfm(struct crypto_aead *tfm)
 	cpu_index = (unsigned int)atomic_inc_return(&ictx->tfm_count) %
 		    cpumask_weight(cpu_online_mask);
 
-	ctx->cb_cpu = cpumask_first(cpu_online_mask);
-	for (cpu = 0; cpu < cpu_index; cpu++)
-		ctx->cb_cpu = cpumask_next(ctx->cb_cpu, cpu_online_mask);
-
+	ctx->cb_cpu = cpumask_nth(cpu_index, cpu_online_mask);
 	cipher = crypto_spawn_aead(&ictx->spawn);
 
 	if (IS_ERR(cipher))
-- 
2.43.0

[PATCH 2/2] crypto: caam - Fix opencoded cpumask_next_wrap() in caam_drv_ctx_init()

[Yury Norov]

From: Yury Norov [NVIDIA] <yury.norov@gmail.com>

The dedicated cpumask_next_wrap() is more verbose and better optimized
comparing to cpumask_next() followed by cpumask_first().

Signed-off-by: Yury Norov [NVIDIA] <yury.norov@gmail.com>
---
 drivers/crypto/caam/qi.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/crypto/caam/qi.c b/drivers/crypto/caam/qi.c
index b6e7c0b29d4e..1e731ed8702b 100644
--- a/drivers/crypto/caam/qi.c
+++ b/drivers/crypto/caam/qi.c
@@ -442,11 +442,8 @@ struct caam_drv_ctx *caam_drv_ctx_init(struct device *qidev,
 	if (!cpumask_test_cpu(*cpu, cpus)) {
 		int *pcpu = &get_cpu_var(last_cpu);
 
-		*pcpu = cpumask_next(*pcpu, cpus);
-		if (*pcpu >= nr_cpu_ids)
-			*pcpu = cpumask_first(cpus);
+		*pcpu = cpumask_next_wrap(*pcpu, cpus);
 		*cpu = *pcpu;
-
 		put_cpu_var(last_cpu);
 	}
 	drv_ctx->cpu = *cpu;
-- 
2.43.0

Re: [PATCH 0/2] crypto: replace cpumask_next with better API

[Herbert Xu]

On Wed, Jun 04, 2025 at 04:47:39PM -0400, Yury Norov wrote:
> From: Yury Norov [NVIDIA]  <yury.norov@gmail.com>
> 
> Recently added cpumask_next_wrap() and cpumask_nth() work better than
> cpumask_next(). Use them where appropriate.
> 
> Yury Norov [NVIDIA] (2):
>   crypto: pcrypt: Optimize pcrypt_aead_init_tfm()
>   crypto: caam - Fix opencoded cpumask_next_wrap() in
>     caam_drv_ctx_init()
> 
>  crypto/pcrypt.c          | 7 ++-----
>  drivers/crypto/caam/qi.c | 5 +----
>  2 files changed, 3 insertions(+), 9 deletions(-)
> 
> -- 
> 2.43.0

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