* [dpdk-stable] [PATCH v2 07/27] crypto/qat: fix KASUMI authentication [not found] ` <20170626102126.56563-1-pablo.de.lara.guarch@intel.com> @ 2017-06-26 10:21 ` Pablo de Lara 0 siblings, 0 replies; 2+ messages in thread From: Pablo de Lara @ 2017-06-26 10:21 UTC (permalink / raw) To: declan.doherty, zbigniew.bodek, jerin.jacob, akhil.goyal, hemant.agrawal Cc: Pablo de Lara, stable QAT PMD was assuming that cipher IV was always prepended, before the input buffer, but it is not necessary to have it there, only the auth IV (COUNT and FRESH) and the input buffer needs to be contiguous, with the direction bit after. If AAD (containing the IV for authentication) is not just before the message, the authentication will fail. Therefore, the headroom of the input buffer is used to copy the AAD, and then it is removed after the operation is finished. It was also assuming that the IV was starting at offset 0, which is not always the case. Fixes: d4f2745300e0 ("crypto/qat: add KASUMI") CC: stable@dpdk.org Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com> --- drivers/crypto/qat/qat_crypto.c | 37 ++++++++++++++++++++++++++----------- 1 file changed, 26 insertions(+), 11 deletions(-) diff --git a/drivers/crypto/qat/qat_crypto.c b/drivers/crypto/qat/qat_crypto.c index 9b294e4..bdd5bab 100644 --- a/drivers/crypto/qat/qat_crypto.c +++ b/drivers/crypto/qat/qat_crypto.c @@ -812,6 +812,11 @@ qat_pmd_dequeue_op_burst(void *qp, struct rte_crypto_op **ops, (rx_op->sym->session->_private); if (sess->bpi_ctx) qat_bpicipher_postprocess(sess, rx_op); + if (sess->qat_hash_alg == + ICP_QAT_HW_AUTH_ALGO_KASUMI_F9) + /* Trim area used for authentication IV. */ + rte_pktmbuf_adj(rx_op->sym->m_src, + ICP_QAT_HW_KASUMI_BLK_SZ); rx_op->status = RTE_CRYPTO_OP_STATUS_SUCCESS; } @@ -1012,15 +1017,25 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, if (ctx->qat_hash_alg == ICP_QAT_HW_AUTH_ALGO_KASUMI_F9) { - if (do_cipher) { - auth_len = auth_len + auth_ofs + 1 - - ICP_QAT_HW_KASUMI_BLK_SZ; - auth_ofs = ICP_QAT_HW_KASUMI_BLK_SZ; - } else { - auth_len = auth_len + auth_ofs + 1; - auth_ofs = 0; - } - } + /* + * Prepend IV in mbuf, as IV and the plaintext + * needs to be contiguous + */ + uint8_t *auth_iv_ptr_dst = + (uint8_t *) rte_pktmbuf_prepend( + op->sym->m_src, + ICP_QAT_HW_KASUMI_BLK_SZ); + const uint8_t *auth_iv_ptr_src = + op->sym->auth.aad.data; + rte_memcpy(auth_iv_ptr_dst, auth_iv_ptr_src, + ICP_QAT_HW_KASUMI_BLK_SZ); + /* Auth IV and message is contiguous + direction bit */ + auth_len = auth_len + ICP_QAT_HW_KASUMI_BLK_SZ + 1; + /* Buffer to cipher starts after auth IV */ + if (do_cipher) + cipher_ofs += ICP_QAT_HW_KASUMI_BLK_SZ; + } else + auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; } else if (ctx->qat_hash_alg == ICP_QAT_HW_AUTH_ALGO_GALOIS_128 || @@ -1028,6 +1043,8 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, ICP_QAT_HW_AUTH_ALGO_GALOIS_64) { auth_ofs = op->sym->cipher.data.offset; auth_len = op->sym->cipher.data.length; + + auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; } else { auth_ofs = op->sym->auth.data.offset; auth_len = op->sym->auth.data.length; @@ -1036,8 +1053,6 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, auth_param->auth_res_addr = op->sym->auth.digest.phys_addr; - auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; - } if (op->sym->m_src->next || (op->sym->m_dst && op->sym->m_dst->next)) -- 2.9.4 ^ permalink raw reply [flat|nested] 2+ messages in thread
[parent not found: <20170626102300.56637-1-pablo.de.lara.guarch@intel.com>]
* [dpdk-stable] [PATCH v2 07/27] crypto/qat: fix KASUMI authentication [not found] ` <20170626102300.56637-1-pablo.de.lara.guarch@intel.com> @ 2017-06-26 10:22 ` Pablo de Lara 0 siblings, 0 replies; 2+ messages in thread From: Pablo de Lara @ 2017-06-26 10:22 UTC (permalink / raw) To: declan.doherty, zbigniew.bodek, jerin.jacob, akhil.goyal, hemant.agrawal Cc: dev, Pablo de Lara, stable QAT PMD was assuming that cipher IV was always prepended, before the input buffer, but it is not necessary to have it there, only the auth IV (COUNT and FRESH) and the input buffer needs to be contiguous, with the direction bit after. If AAD (containing the IV for authentication) is not just before the message, the authentication will fail. Therefore, the headroom of the input buffer is used to copy the AAD, and then it is removed after the operation is finished. It was also assuming that the IV was starting at offset 0, which is not always the case. Fixes: d4f2745300e0 ("crypto/qat: add KASUMI") CC: stable@dpdk.org Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com> --- drivers/crypto/qat/qat_crypto.c | 37 ++++++++++++++++++++++++++----------- 1 file changed, 26 insertions(+), 11 deletions(-) diff --git a/drivers/crypto/qat/qat_crypto.c b/drivers/crypto/qat/qat_crypto.c index 9b294e4..bdd5bab 100644 --- a/drivers/crypto/qat/qat_crypto.c +++ b/drivers/crypto/qat/qat_crypto.c @@ -812,6 +812,11 @@ qat_pmd_dequeue_op_burst(void *qp, struct rte_crypto_op **ops, (rx_op->sym->session->_private); if (sess->bpi_ctx) qat_bpicipher_postprocess(sess, rx_op); + if (sess->qat_hash_alg == + ICP_QAT_HW_AUTH_ALGO_KASUMI_F9) + /* Trim area used for authentication IV. */ + rte_pktmbuf_adj(rx_op->sym->m_src, + ICP_QAT_HW_KASUMI_BLK_SZ); rx_op->status = RTE_CRYPTO_OP_STATUS_SUCCESS; } @@ -1012,15 +1017,25 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, if (ctx->qat_hash_alg == ICP_QAT_HW_AUTH_ALGO_KASUMI_F9) { - if (do_cipher) { - auth_len = auth_len + auth_ofs + 1 - - ICP_QAT_HW_KASUMI_BLK_SZ; - auth_ofs = ICP_QAT_HW_KASUMI_BLK_SZ; - } else { - auth_len = auth_len + auth_ofs + 1; - auth_ofs = 0; - } - } + /* + * Prepend IV in mbuf, as IV and the plaintext + * needs to be contiguous + */ + uint8_t *auth_iv_ptr_dst = + (uint8_t *) rte_pktmbuf_prepend( + op->sym->m_src, + ICP_QAT_HW_KASUMI_BLK_SZ); + const uint8_t *auth_iv_ptr_src = + op->sym->auth.aad.data; + rte_memcpy(auth_iv_ptr_dst, auth_iv_ptr_src, + ICP_QAT_HW_KASUMI_BLK_SZ); + /* Auth IV and message is contiguous + direction bit */ + auth_len = auth_len + ICP_QAT_HW_KASUMI_BLK_SZ + 1; + /* Buffer to cipher starts after auth IV */ + if (do_cipher) + cipher_ofs += ICP_QAT_HW_KASUMI_BLK_SZ; + } else + auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; } else if (ctx->qat_hash_alg == ICP_QAT_HW_AUTH_ALGO_GALOIS_128 || @@ -1028,6 +1043,8 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, ICP_QAT_HW_AUTH_ALGO_GALOIS_64) { auth_ofs = op->sym->cipher.data.offset; auth_len = op->sym->cipher.data.length; + + auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; } else { auth_ofs = op->sym->auth.data.offset; auth_len = op->sym->auth.data.length; @@ -1036,8 +1053,6 @@ qat_write_hw_desc_entry(struct rte_crypto_op *op, uint8_t *out_msg, auth_param->auth_res_addr = op->sym->auth.digest.phys_addr; - auth_param->u1.aad_adr = op->sym->auth.aad.phys_addr; - } if (op->sym->m_src->next || (op->sym->m_dst && op->sym->m_dst->next)) -- 2.9.4 ^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2017-06-26 18:22 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <1496005522-134934-1-git-send-email-pablo.de.lara.guarch@intel.com>
[not found] ` <20170626102126.56563-1-pablo.de.lara.guarch@intel.com>
2017-06-26 10:21 ` [dpdk-stable] [PATCH v2 07/27] crypto/qat: fix KASUMI authentication Pablo de Lara
[not found] ` <20170626102300.56637-1-pablo.de.lara.guarch@intel.com>
2017-06-26 10:22 ` Pablo de Lara
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).