* [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher
@ 2026-03-21 22:52 Eric Biggers
2026-03-23 7:02 ` Subbaraya Sundeep
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Eric Biggers @ 2026-03-21 22:52 UTC (permalink / raw)
To: netdev
Cc: Andrew Lunn, David S . Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Sunil Goutham, Geetha sowjanya, Subbaraya Sundeep,
hariprasad, Bharat Bhushan, Sabrina Dubroca, Eric Biggers,
linux-crypto, linux-kernel
cn10k_ecb_aes_encrypt() just encrypts a single block with AES. That is
much more easily and efficiently done with the AES library than
crypto_skcipher. Use the AES library instead.
Signed-off-by: Eric Biggers <ebiggers@kernel.org>
---
.../net/ethernet/marvell/octeontx2/Kconfig | 1 +
.../marvell/octeontx2/nic/cn10k_macsec.c | 53 +++++--------------
2 files changed, 13 insertions(+), 41 deletions(-)
diff --git a/drivers/net/ethernet/marvell/octeontx2/Kconfig b/drivers/net/ethernet/marvell/octeontx2/Kconfig
index 35c4f5f64f58..47e549c581f0 100644
--- a/drivers/net/ethernet/marvell/octeontx2/Kconfig
+++ b/drivers/net/ethernet/marvell/octeontx2/Kconfig
@@ -31,10 +31,11 @@ config NDC_DIS_DYNAMIC_CACHING
config OCTEONTX2_PF
tristate "Marvell OcteonTX2 NIC Physical Function driver"
select OCTEONTX2_MBOX
select NET_DEVLINK
select PAGE_POOL
+ select CRYPTO_LIB_AES if MACSEC
depends on (64BIT && COMPILE_TEST) || ARM64
select DIMLIB
depends on PCI
depends on PTP_1588_CLOCK_OPTIONAL
depends on MACSEC || !MACSEC
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
index 4649996dc7da..2cc1bdfd9b2e 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
@@ -2,11 +2,11 @@
/* Marvell MACSEC hardware offload driver
*
* Copyright (C) 2022 Marvell.
*/
-#include <crypto/skcipher.h>
+#include <crypto/aes.h>
#include <linux/rtnetlink.h>
#include <linux/bitfield.h>
#include "otx2_common.h"
#define MCS_TCAM0_MAC_DA_MASK GENMASK_ULL(47, 0)
@@ -44,55 +44,26 @@
#define MCS_TCI_C 0x04 /* changed text */
#define CN10K_MAX_HASH_LEN 16
#define CN10K_MAX_SAK_LEN 32
-static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, u8 *sak,
- u16 sak_len, u8 *hash)
+static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, const u8 *sak,
+ u16 sak_len, u8 hash[CN10K_MAX_HASH_LEN])
{
- u8 data[CN10K_MAX_HASH_LEN] = { 0 };
- struct skcipher_request *req = NULL;
- struct scatterlist sg_src, sg_dst;
- struct crypto_skcipher *tfm;
- DECLARE_CRYPTO_WAIT(wait);
- int err;
-
- tfm = crypto_alloc_skcipher("ecb(aes)", 0, 0);
- if (IS_ERR(tfm)) {
- dev_err(pfvf->dev, "failed to allocate transform for ecb-aes\n");
- return PTR_ERR(tfm);
- }
-
- req = skcipher_request_alloc(tfm, GFP_KERNEL);
- if (!req) {
- dev_err(pfvf->dev, "failed to allocate request for skcipher\n");
- err = -ENOMEM;
- goto free_tfm;
- }
+ static const u8 zeroes[CN10K_MAX_HASH_LEN];
+ struct aes_enckey aes;
- err = crypto_skcipher_setkey(tfm, sak, sak_len);
- if (err) {
- dev_err(pfvf->dev, "failed to set key for skcipher\n");
- goto free_req;
+ if (aes_prepareenckey(&aes, sak, sak_len) != 0) {
+ dev_err(pfvf->dev, "invalid AES key length: %d\n", sak_len);
+ return -EINVAL;
}
- /* build sg list */
- sg_init_one(&sg_src, data, CN10K_MAX_HASH_LEN);
- sg_init_one(&sg_dst, hash, CN10K_MAX_HASH_LEN);
-
- skcipher_request_set_callback(req, 0, crypto_req_done, &wait);
- skcipher_request_set_crypt(req, &sg_src, &sg_dst,
- CN10K_MAX_HASH_LEN, NULL);
+ static_assert(CN10K_MAX_HASH_LEN == AES_BLOCK_SIZE);
+ aes_encrypt(&aes, hash, zeroes);
- err = crypto_skcipher_encrypt(req);
- err = crypto_wait_req(err, &wait);
-
-free_req:
- skcipher_request_free(req);
-free_tfm:
- crypto_free_skcipher(tfm);
- return err;
+ memzero_explicit(&aes, sizeof(aes));
+ return 0;
}
static struct cn10k_mcs_txsc *cn10k_mcs_get_txsc(struct cn10k_mcs_cfg *cfg,
struct macsec_secy *secy)
{
base-commit: fb78a629b4f0eb399b413f6c093a3da177b3a4eb
--
2.53.0
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher
2026-03-21 22:52 [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher Eric Biggers
@ 2026-03-23 7:02 ` Subbaraya Sundeep
2026-03-25 18:08 ` Subbaraya Sundeep
2026-03-26 10:50 ` patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: Subbaraya Sundeep @ 2026-03-23 7:02 UTC (permalink / raw)
To: Eric Biggers
Cc: netdev, Andrew Lunn, David S . Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Sunil Goutham, Geetha sowjanya,
hariprasad, Bharat Bhushan, Sabrina Dubroca, linux-crypto,
linux-kernel
Hi,
On 2026-03-22 at 04:22:08, Eric Biggers (ebiggers@kernel.org) wrote:
> cn10k_ecb_aes_encrypt() just encrypts a single block with AES. That is
> much more easily and efficiently done with the AES library than
> crypto_skcipher. Use the AES library instead.
>
Thanks for the patch. Give me couple of days, I will test and
ack.
Sundeep
> Signed-off-by: Eric Biggers <ebiggers@kernel.org>
> ---
> .../net/ethernet/marvell/octeontx2/Kconfig | 1 +
> .../marvell/octeontx2/nic/cn10k_macsec.c | 53 +++++--------------
> 2 files changed, 13 insertions(+), 41 deletions(-)
>
> diff --git a/drivers/net/ethernet/marvell/octeontx2/Kconfig b/drivers/net/ethernet/marvell/octeontx2/Kconfig
> index 35c4f5f64f58..47e549c581f0 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/Kconfig
> +++ b/drivers/net/ethernet/marvell/octeontx2/Kconfig
> @@ -31,10 +31,11 @@ config NDC_DIS_DYNAMIC_CACHING
> config OCTEONTX2_PF
> tristate "Marvell OcteonTX2 NIC Physical Function driver"
> select OCTEONTX2_MBOX
> select NET_DEVLINK
> select PAGE_POOL
> + select CRYPTO_LIB_AES if MACSEC
> depends on (64BIT && COMPILE_TEST) || ARM64
> select DIMLIB
> depends on PCI
> depends on PTP_1588_CLOCK_OPTIONAL
> depends on MACSEC || !MACSEC
> diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> index 4649996dc7da..2cc1bdfd9b2e 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> @@ -2,11 +2,11 @@
> /* Marvell MACSEC hardware offload driver
> *
> * Copyright (C) 2022 Marvell.
> */
>
> -#include <crypto/skcipher.h>
> +#include <crypto/aes.h>
> #include <linux/rtnetlink.h>
> #include <linux/bitfield.h>
> #include "otx2_common.h"
>
> #define MCS_TCAM0_MAC_DA_MASK GENMASK_ULL(47, 0)
> @@ -44,55 +44,26 @@
> #define MCS_TCI_C 0x04 /* changed text */
>
> #define CN10K_MAX_HASH_LEN 16
> #define CN10K_MAX_SAK_LEN 32
>
> -static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, u8 *sak,
> - u16 sak_len, u8 *hash)
> +static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, const u8 *sak,
> + u16 sak_len, u8 hash[CN10K_MAX_HASH_LEN])
> {
> - u8 data[CN10K_MAX_HASH_LEN] = { 0 };
> - struct skcipher_request *req = NULL;
> - struct scatterlist sg_src, sg_dst;
> - struct crypto_skcipher *tfm;
> - DECLARE_CRYPTO_WAIT(wait);
> - int err;
> -
> - tfm = crypto_alloc_skcipher("ecb(aes)", 0, 0);
> - if (IS_ERR(tfm)) {
> - dev_err(pfvf->dev, "failed to allocate transform for ecb-aes\n");
> - return PTR_ERR(tfm);
> - }
> -
> - req = skcipher_request_alloc(tfm, GFP_KERNEL);
> - if (!req) {
> - dev_err(pfvf->dev, "failed to allocate request for skcipher\n");
> - err = -ENOMEM;
> - goto free_tfm;
> - }
> + static const u8 zeroes[CN10K_MAX_HASH_LEN];
> + struct aes_enckey aes;
>
> - err = crypto_skcipher_setkey(tfm, sak, sak_len);
> - if (err) {
> - dev_err(pfvf->dev, "failed to set key for skcipher\n");
> - goto free_req;
> + if (aes_prepareenckey(&aes, sak, sak_len) != 0) {
> + dev_err(pfvf->dev, "invalid AES key length: %d\n", sak_len);
> + return -EINVAL;
> }
>
> - /* build sg list */
> - sg_init_one(&sg_src, data, CN10K_MAX_HASH_LEN);
> - sg_init_one(&sg_dst, hash, CN10K_MAX_HASH_LEN);
> -
> - skcipher_request_set_callback(req, 0, crypto_req_done, &wait);
> - skcipher_request_set_crypt(req, &sg_src, &sg_dst,
> - CN10K_MAX_HASH_LEN, NULL);
> + static_assert(CN10K_MAX_HASH_LEN == AES_BLOCK_SIZE);
> + aes_encrypt(&aes, hash, zeroes);
>
> - err = crypto_skcipher_encrypt(req);
> - err = crypto_wait_req(err, &wait);
> -
> -free_req:
> - skcipher_request_free(req);
> -free_tfm:
> - crypto_free_skcipher(tfm);
> - return err;
> + memzero_explicit(&aes, sizeof(aes));
> + return 0;
> }
>
> static struct cn10k_mcs_txsc *cn10k_mcs_get_txsc(struct cn10k_mcs_cfg *cfg,
> struct macsec_secy *secy)
> {
>
> base-commit: fb78a629b4f0eb399b413f6c093a3da177b3a4eb
> --
> 2.53.0
>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher
2026-03-21 22:52 [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher Eric Biggers
2026-03-23 7:02 ` Subbaraya Sundeep
@ 2026-03-25 18:08 ` Subbaraya Sundeep
2026-03-26 10:50 ` patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: Subbaraya Sundeep @ 2026-03-25 18:08 UTC (permalink / raw)
To: Eric Biggers
Cc: netdev, Andrew Lunn, David S . Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Sunil Goutham, Geetha sowjanya,
hariprasad, Bharat Bhushan, Sabrina Dubroca, linux-crypto,
linux-kernel
On 2026-03-22 at 04:22:08, Eric Biggers (ebiggers@kernel.org) wrote:
> cn10k_ecb_aes_encrypt() just encrypts a single block with AES. That is
> much more easily and efficiently done with the AES library than
> crypto_skcipher. Use the AES library instead.
>
> Signed-off-by: Eric Biggers <ebiggers@kernel.org>
Reviewed-by: Subbaraya Sundeep <sbhatta@marvell.com>
Thanks,
Sundeep
> ---
> .../net/ethernet/marvell/octeontx2/Kconfig | 1 +
> .../marvell/octeontx2/nic/cn10k_macsec.c | 53 +++++--------------
> 2 files changed, 13 insertions(+), 41 deletions(-)
>
> diff --git a/drivers/net/ethernet/marvell/octeontx2/Kconfig b/drivers/net/ethernet/marvell/octeontx2/Kconfig
> index 35c4f5f64f58..47e549c581f0 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/Kconfig
> +++ b/drivers/net/ethernet/marvell/octeontx2/Kconfig
> @@ -31,10 +31,11 @@ config NDC_DIS_DYNAMIC_CACHING
> config OCTEONTX2_PF
> tristate "Marvell OcteonTX2 NIC Physical Function driver"
> select OCTEONTX2_MBOX
> select NET_DEVLINK
> select PAGE_POOL
> + select CRYPTO_LIB_AES if MACSEC
> depends on (64BIT && COMPILE_TEST) || ARM64
> select DIMLIB
> depends on PCI
> depends on PTP_1588_CLOCK_OPTIONAL
> depends on MACSEC || !MACSEC
> diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> index 4649996dc7da..2cc1bdfd9b2e 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k_macsec.c
> @@ -2,11 +2,11 @@
> /* Marvell MACSEC hardware offload driver
> *
> * Copyright (C) 2022 Marvell.
> */
>
> -#include <crypto/skcipher.h>
> +#include <crypto/aes.h>
> #include <linux/rtnetlink.h>
> #include <linux/bitfield.h>
> #include "otx2_common.h"
>
> #define MCS_TCAM0_MAC_DA_MASK GENMASK_ULL(47, 0)
> @@ -44,55 +44,26 @@
> #define MCS_TCI_C 0x04 /* changed text */
>
> #define CN10K_MAX_HASH_LEN 16
> #define CN10K_MAX_SAK_LEN 32
>
> -static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, u8 *sak,
> - u16 sak_len, u8 *hash)
> +static int cn10k_ecb_aes_encrypt(struct otx2_nic *pfvf, const u8 *sak,
> + u16 sak_len, u8 hash[CN10K_MAX_HASH_LEN])
> {
> - u8 data[CN10K_MAX_HASH_LEN] = { 0 };
> - struct skcipher_request *req = NULL;
> - struct scatterlist sg_src, sg_dst;
> - struct crypto_skcipher *tfm;
> - DECLARE_CRYPTO_WAIT(wait);
> - int err;
> -
> - tfm = crypto_alloc_skcipher("ecb(aes)", 0, 0);
> - if (IS_ERR(tfm)) {
> - dev_err(pfvf->dev, "failed to allocate transform for ecb-aes\n");
> - return PTR_ERR(tfm);
> - }
> -
> - req = skcipher_request_alloc(tfm, GFP_KERNEL);
> - if (!req) {
> - dev_err(pfvf->dev, "failed to allocate request for skcipher\n");
> - err = -ENOMEM;
> - goto free_tfm;
> - }
> + static const u8 zeroes[CN10K_MAX_HASH_LEN];
> + struct aes_enckey aes;
>
> - err = crypto_skcipher_setkey(tfm, sak, sak_len);
> - if (err) {
> - dev_err(pfvf->dev, "failed to set key for skcipher\n");
> - goto free_req;
> + if (aes_prepareenckey(&aes, sak, sak_len) != 0) {
> + dev_err(pfvf->dev, "invalid AES key length: %d\n", sak_len);
> + return -EINVAL;
> }
>
> - /* build sg list */
> - sg_init_one(&sg_src, data, CN10K_MAX_HASH_LEN);
> - sg_init_one(&sg_dst, hash, CN10K_MAX_HASH_LEN);
> -
> - skcipher_request_set_callback(req, 0, crypto_req_done, &wait);
> - skcipher_request_set_crypt(req, &sg_src, &sg_dst,
> - CN10K_MAX_HASH_LEN, NULL);
> + static_assert(CN10K_MAX_HASH_LEN == AES_BLOCK_SIZE);
> + aes_encrypt(&aes, hash, zeroes);
>
> - err = crypto_skcipher_encrypt(req);
> - err = crypto_wait_req(err, &wait);
> -
> -free_req:
> - skcipher_request_free(req);
> -free_tfm:
> - crypto_free_skcipher(tfm);
> - return err;
> + memzero_explicit(&aes, sizeof(aes));
> + return 0;
> }
>
> static struct cn10k_mcs_txsc *cn10k_mcs_get_txsc(struct cn10k_mcs_cfg *cfg,
> struct macsec_secy *secy)
> {
>
> base-commit: fb78a629b4f0eb399b413f6c093a3da177b3a4eb
> --
> 2.53.0
>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher
2026-03-21 22:52 [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher Eric Biggers
2026-03-23 7:02 ` Subbaraya Sundeep
2026-03-25 18:08 ` Subbaraya Sundeep
@ 2026-03-26 10:50 ` patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: patchwork-bot+netdevbpf @ 2026-03-26 10:50 UTC (permalink / raw)
To: Eric Biggers
Cc: netdev, andrew+netdev, davem, edumazet, kuba, pabeni, sgoutham,
gakula, sbhatta, hkelam, bbhushan2, sd, linux-crypto,
linux-kernel
Hello:
This patch was applied to netdev/net-next.git (main)
by Paolo Abeni <pabeni@redhat.com>:
On Sat, 21 Mar 2026 15:52:08 -0700 you wrote:
> cn10k_ecb_aes_encrypt() just encrypts a single block with AES. That is
> much more easily and efficiently done with the AES library than
> crypto_skcipher. Use the AES library instead.
>
> Signed-off-by: Eric Biggers <ebiggers@kernel.org>
> ---
> .../net/ethernet/marvell/octeontx2/Kconfig | 1 +
> .../marvell/octeontx2/nic/cn10k_macsec.c | 53 +++++--------------
> 2 files changed, 13 insertions(+), 41 deletions(-)
>
> [...]
Here is the summary with links:
- [net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher
https://git.kernel.org/netdev/net-next/c/8f303194b241
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-03-26 10:50 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-21 22:52 [PATCH net-next] octeontx2-pf: macsec: Use AES library instead of ecb(aes) skcipher Eric Biggers
2026-03-23 7:02 ` Subbaraya Sundeep
2026-03-25 18:08 ` Subbaraya Sundeep
2026-03-26 10:50 ` patchwork-bot+netdevbpf
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox