From: "Zhoujian (jay)" <jianjay.zhou@huawei.com>
To: Fan Zhang <roy.fan.zhang@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: "pablo.de.lara.guarch@intel.com" <pablo.de.lara.guarch@intel.com>,
"stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [dpdk-stable] [PATCH] crypto/virtio: fix iv physical address
Date: Fri, 15 Jun 2018 04:41:30 +0000 [thread overview]
Message-ID: <B2D15215269B544CADD246097EACE7473B6B3EA8@dggemm508-mbx.china.huawei.com> (raw)
In-Reply-To: <20180614110257.10967-1-roy.fan.zhang@intel.com>
Hi Fan,
> -----Original Message-----
> From: Fan Zhang [mailto:roy.fan.zhang@intel.com]
> Sent: Thursday, June 14, 2018 7:03 PM
> To: dev@dpdk.org
> Cc: pablo.de.lara.guarch@intel.com; Zhoujian (jay) <jianjay.zhou@huawei.com>;
> stable@dpdk.org
> Subject: [PATCH] crypto/virtio: fix iv physical address
>
> The physical address of IV required by Virtio was computed using crypto
> operations' physical address plus the offset. However not all crypto ops will
> have physical address field initialized and compute it runtimely is costly.
> This patch fixes this problem by adding iv field in virtio_crypto_op_cookie
> and does a memcpy of iv instead.
>
> Fixes: 82adb12a1fce ("crypto/virtio: support burst enqueue/dequeue")
> Cc: stable@dpdk.org
>
> Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
> ---
> drivers/crypto/virtio/virtio_cryptodev.c | 6 ++++++
> drivers/crypto/virtio/virtio_cryptodev.h | 3 +++
> drivers/crypto/virtio/virtio_rxtx.c | 8 +++++++-
> 3 files changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/crypto/virtio/virtio_cryptodev.c
> b/drivers/crypto/virtio/virtio_cryptodev.c
> index df88953f6..6ffa7619c 100644
> --- a/drivers/crypto/virtio/virtio_cryptodev.c
> +++ b/drivers/crypto/virtio/virtio_cryptodev.c
> @@ -1223,6 +1223,12 @@ virtio_crypto_sym_pad_op_ctrl_req(
> /* Get cipher xform from crypto xform chain */
> cipher_xform = virtio_crypto_get_cipher_xform(xform);
> if (cipher_xform) {
> + if (cipher_xform->iv.length > VIRTIO_CRYPTO_MAX_IV_SIZE) {
Considering the consistency with iv.length, could we use VIRTIO_CRYPTO_MAX_IV_LENGTH
instead of VIRTIO_CRYPTO_MAX_IV_SIZE?
> + VIRTIO_CRYPTO_SESSION_LOG_ERR(
> + "cipher IV cannot longer than %u",
s/cipher IV/cipher IV length/ would be better, I think.
> + VIRTIO_CRYPTO_MAX_IV_SIZE);
> + return -1;
> + }
> if (is_chainned)
> ret = virtio_crypto_sym_pad_cipher_param(
> &ctrl->u.sym_create_session.u.chain.para
> diff --git a/drivers/crypto/virtio/virtio_cryptodev.h
> b/drivers/crypto/virtio/virtio_cryptodev.h
> index e402c0309..676e008d9 100644
> --- a/drivers/crypto/virtio/virtio_cryptodev.h
> +++ b/drivers/crypto/virtio/virtio_cryptodev.h
> @@ -16,6 +16,8 @@
>
> #define NUM_ENTRY_VIRTIO_CRYPTO_OP 7
>
> +#define VIRTIO_CRYPTO_MAX_IV_SIZE 32
> +
> extern uint8_t cryptodev_virtio_driver_id;
>
> enum virtio_crypto_cmd_id {
> @@ -29,6 +31,7 @@ struct virtio_crypto_op_cookie {
> struct virtio_crypto_op_data_req data_req;
> struct virtio_crypto_inhdr inhdr;
> struct vring_desc desc[NUM_ENTRY_VIRTIO_CRYPTO_OP];
> + uint8_t iv[VIRTIO_CRYPTO_MAX_IV_SIZE];
> };
>
> /*
> diff --git a/drivers/crypto/virtio/virtio_rxtx.c
> b/drivers/crypto/virtio/virtio_rxtx.c
> index 450392843..1b16221d0 100644
> --- a/drivers/crypto/virtio/virtio_rxtx.c
> +++ b/drivers/crypto/virtio/virtio_rxtx.c
> @@ -203,6 +203,8 @@ virtqueue_crypto_sym_enqueue_xmit(
> uint16_t req_data_len = sizeof(struct virtio_crypto_op_data_req);
> uint32_t indirect_vring_addr_offset = req_data_len +
> sizeof(struct virtio_crypto_inhdr);
> + uint32_t indirect_iv_addr_offset = indirect_vring_addr_offset +
> + sizeof(struct vring_desc) * NUM_ENTRY_VIRTIO_CRYPTO_OP;
> struct rte_crypto_sym_op *sym_op = cop->sym;
> struct virtio_crypto_session *session =
> (struct virtio_crypto_session *)get_session_private_data( @@ -
> 259,7 +261,11 @@ virtqueue_crypto_sym_enqueue_xmit(
>
> /* indirect vring: iv of cipher */
> if (session->iv.length) {
> - desc[idx].addr = cop->phys_addr + session->iv.offset;
> + rte_memcpy(crypto_op_cookie->iv, rte_crypto_op_ctod_offset(cop,
> + uint8_t *, session->iv.offset),
> + session->iv.length);
> + desc[idx].addr = indirect_op_data_req_phys_addr +
> + indirect_iv_addr_offset;
This ops of rte_memcpy could be avoided if cop->phys_addr is initialized,
so how about:
if (cop->phys_addr) {
desc[idx].addr = cop->phys_addr + session->iv.offset;
} else {
...
}
Regards,
Jay
> desc[idx].len = session->iv.length;
> desc[idx++].flags = VRING_DESC_F_NEXT;
> }
> --
> 2.13.6
next prev parent reply other threads:[~2018-06-15 4:41 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-06-14 11:02 Fan Zhang
2018-06-14 12:38 ` [dpdk-stable] [dpdk-dev] " Akhil Goyal
2018-06-15 4:45 ` Zhoujian (jay)
2018-06-15 4:41 ` Zhoujian (jay) [this message]
2018-06-26 2:10 ` [dpdk-stable] [PATCH v2] " Fan Zhang
2018-07-03 12:29 ` Zhoujian (jay)
2018-07-05 10:53 ` De Lara Guarch, Pablo
2018-07-06 1:12 ` Zhoujian (jay)
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=B2D15215269B544CADD246097EACE7473B6B3EA8@dggemm508-mbx.china.huawei.com \
--to=jianjay.zhou@huawei.com \
--cc=dev@dpdk.org \
--cc=pablo.de.lara.guarch@intel.com \
--cc=roy.fan.zhang@intel.com \
--cc=stable@dpdk.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 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).