linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Kai Ye <yekai13@huawei.com>
To: <herbert@gondor.apana.org.au>
Cc: <linux-crypto@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<wangzhou1@hisilicon.com>, <yekai13@huawei.com>,
	<liulongfang@huawei.com>
Subject: [PATCH v3 3/6] crypto: hisilicon/sec - support the larger packets for digest mode
Date: Fri, 22 Oct 2021 17:10:52 +0800	[thread overview]
Message-ID: <20211022091055.15369-4-yekai13@huawei.com> (raw)
In-Reply-To: <20211022091055.15369-1-yekai13@huawei.com>

Support the larger packets inputing for digest mode. By split the larger
packet, the driver input one block size packet to the hardware every time.
Setting one block size is PAGE_SIZE(4K). Based on the long hash feature,
the soft waits until the previous packet is complete and then sends the
next packet.

Signed-off-by: Kai Ye <yekai13@huawei.com>
---
 drivers/crypto/hisilicon/sec2/sec_crypto.c | 76 ++++++++++++++++++++++
 1 file changed, 76 insertions(+)

diff --git a/drivers/crypto/hisilicon/sec2/sec_crypto.c b/drivers/crypto/hisilicon/sec2/sec_crypto.c
index c2010995d4b8..3f95d7b598ea 100644
--- a/drivers/crypto/hisilicon/sec2/sec_crypto.c
+++ b/drivers/crypto/hisilicon/sec2/sec_crypto.c
@@ -2732,6 +2732,80 @@ static int sec_ahash_finup(struct ahash_request *req)
 	return sec_ahash_process(req);
 }
 
+static int digest_hardware_update(struct sec_req *sreq, struct scatterlist *src,
+				  u32 start, u32 nbytes)
+{
+	struct sec_ahash_req *sareq = &sreq->hash_req;
+	struct scatterlist *pingpong_sg;
+	struct sec_ctx *ctx = sreq->ctx;
+	int nents = sg_nents(src);
+	int sid = sareq->sid;
+	u8 idx;
+
+	idx = ctx->pingpong_idx[sid];
+	pingpong_sg = ctx->pingpong_sg[sid][idx].sgl;
+
+	sareq->op = SEC_SHA_UPDATE;
+	sareq->pp_sg = pingpong_sg;
+
+	/* The pingpong buffer can hold the one block input of this request */
+	sg_pcopy_to_buffer(src, nents, sg_virt(pingpong_sg), nbytes, start);
+	ctx->pingpong_sg[sid][idx].len = nbytes;
+
+	sareq->req_data_len = 0;
+	sareq->pp_data_len = nbytes;
+	sareq->block_data_len = nbytes;
+	sareq->total_data_len += nbytes;
+
+	return ctx->req_op->process(ctx, sreq);
+}
+
+static int sec_ahash_larger_digest(struct ahash_request *req)
+{
+	struct sec_req *sreq = ahash_request_ctx(req);
+	struct sec_ahash_req *sareq = &sreq->hash_req;
+	u32 input_len = req->nbytes;
+	struct scatterlist *pingpong_sg;
+	struct sec_ctx *ctx = sreq->ctx;
+	u32 sid = sareq->sid;
+	u8 idx = ctx->pingpong_idx[sid];
+	u32 start = 0;
+	int ret;
+
+	while (input_len > SEC_SID_BUF_LEN) {
+		/* Setting one block size is PAGE_SIZE */
+		req->nbytes = SEC_SID_BUF_LEN;
+		input_len -= SEC_SID_BUF_LEN;
+
+		ret = digest_hardware_update(sreq, req->src, start,
+					     req->nbytes);
+		if (unlikely(ret == -EINVAL)) {
+			pr_err("ahash digest: hardware update process is error!\n");
+			return ret;
+		}
+
+		stream_hash_wait(sareq);
+
+		start += SEC_SID_BUF_LEN;
+	}
+
+	/* Last packet send to the hardware */
+	req->nbytes = input_len;
+	sareq->req_data_len = 0;
+
+	pingpong_sg = ctx->pingpong_sg[sid][idx].sgl;
+	sg_pcopy_to_buffer(req->src, sg_nents(req->src), sg_virt(pingpong_sg),
+			   input_len, start);
+	ctx->pingpong_sg[sid][idx].len = input_len;
+	sareq->pp_data_len = input_len;
+	sareq->block_data_len = input_len;
+	sareq->total_data_len += input_len;
+
+	sareq->op = SEC_SHA_FINAL;
+
+	return ctx->req_op->process(ctx, sreq);
+}
+
 static int sec_ahash_digest(struct ahash_request *req)
 {
 	struct sec_req *sreq = ahash_request_ctx(req);
@@ -2748,6 +2822,8 @@ static int sec_ahash_digest(struct ahash_request *req)
 
 	ctx = sreq->ctx;
 	a_ctx = &ctx->a_ctx;
+	if (req->nbytes > SEC_HW_MAX_LEN)
+		return sec_ahash_larger_digest(req);
 
 	sareq->op = SEC_SHA_DIGEST;
 
-- 
2.33.0


  parent reply	other threads:[~2021-10-22  9:13 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-22  9:10 [PATCH v3 0/6] crypto: hisilicon - supports hash algorithm for SEC engine Kai Ye
2021-10-22  9:10 ` [PATCH v3 1/6] crypto: hisilicon/sec - add ping-pong buffer for ahash Kai Ye
2021-10-22  9:10 ` [PATCH v3 2/6] crypto: hisilicon/sec - add ahash alg features for Kunpeng920 Kai Ye
2021-11-13  6:01   ` Herbert Xu
2021-10-22  9:10 ` Kai Ye [this message]
2021-10-22  9:10 ` [PATCH v3 4/6] crypto: hisilicon/sec - ahash adapt to Kunpeng930 SQE Kai Ye
2021-10-22  9:10 ` [PATCH v3 5/6] crypto: hisilicon/sec - add fallback tfm supporting for ahash Kai Ye
2021-10-22  9:10 ` [PATCH v3 6/6] crypto: hisilicon/sec - add sm3/md5 generic selection " Kai Ye
2021-10-30  1:06 ` [PATCH v3 0/6] crypto: hisilicon - supports hash algorithm for SEC engine yekai(A)
2021-11-02  3:52   ` Herbert Xu

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=20211022091055.15369-4-yekai13@huawei.com \
    --to=yekai13@huawei.com \
    --cc=herbert@gondor.apana.org.au \
    --cc=linux-crypto@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=liulongfang@huawei.com \
    --cc=wangzhou1@hisilicon.com \
    /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).