From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <dev-bounces@dpdk.org>
Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124])
by inbox.dpdk.org (Postfix) with ESMTP id A336BA00C5;
Wed, 14 Sep 2022 21:35:39 +0200 (CEST)
Received: from [217.70.189.124] (localhost [127.0.0.1])
by mails.dpdk.org (Postfix) with ESMTP id 592F44021D;
Wed, 14 Sep 2022 21:35:39 +0200 (CEST)
Received: from us-smtp-delivery-124.mimecast.com
(us-smtp-delivery-124.mimecast.com [170.10.133.124])
by mails.dpdk.org (Postfix) with ESMTP id 060E340156
for <dev@dpdk.org>; Wed, 14 Sep 2022 21:35:37 +0200 (CEST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com;
s=mimecast20190719; t=1663184137;
h=from:from:reply-to:subject:subject:date:date:message-id:message-id:
to:to:cc:cc:mime-version:mime-version:content-type:content-type:
content-transfer-encoding:content-transfer-encoding:
in-reply-to:in-reply-to:references:references;
bh=ffhs2Rd19PkAkHkjaXTf1hDc9xaM1tqwczvqkn00WoA=;
b=YObHbSgKkIJdY2gHWZt2N+RmZiUpHdPW+nQ8sURluqP90Q9aEONduyvKNjloDQ/QmG5gwI
DiFx0xwiQIug0YeXRWDmrL+NrjuG4934QRTZ1FtUXLV6YqgmFpGWepI1+lH37WWrdroRWS
HXXsMikB790GVOE1/++BIQ3HxdRfnts=
Received: from mimecast-mx02.redhat.com (mx3-rdu2.redhat.com
[66.187.233.73]) by relay.mimecast.com with ESMTP with STARTTLS
(version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id
us-mta-612-oEWyGwFkOjW0Km3v5KxO2Q-1; Wed, 14 Sep 2022 15:35:36 -0400
X-MC-Unique: oEWyGwFkOjW0Km3v5KxO2Q-1
Received: from smtp.corp.redhat.com (int-mx03.intmail.prod.int.rdu2.redhat.com
[10.11.54.3])
(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))
(No client certificate requested)
by mimecast-mx02.redhat.com (Postfix) with ESMTPS id ED40F3C02590;
Wed, 14 Sep 2022 19:35:35 +0000 (UTC)
Received: from [10.39.208.12] (unknown [10.39.208.12])
by smtp.corp.redhat.com (Postfix) with ESMTPS id 80EAC1121314;
Wed, 14 Sep 2022 19:35:34 +0000 (UTC)
Message-ID: <7b356859-afb9-74ce-121e-b661a6b6f4d9@redhat.com>
Date: Wed, 14 Sep 2022 21:35:33 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
Thunderbird/91.12.0
To: Hernan Vargas <hernan.vargas@intel.com>, dev@dpdk.org,
gakhil@marvell.com, trix@redhat.com
Cc: nicolas.chautru@intel.com, qi.z.zhang@intel.com
References: <20220820023157.189047-1-hernan.vargas@intel.com>
<20220820023157.189047-8-hernan.vargas@intel.com>
From: Maxime Coquelin <maxime.coquelin@redhat.com>
Subject: Re: [PATCH v2 07/37] baseband/acc100: add LDPC encoder padding
function
In-Reply-To: <20220820023157.189047-8-hernan.vargas@intel.com>
X-Scanned-By: MIMEDefang 3.1 on 10.11.54.3
X-Mimecast-Spam-Score: 0
X-Mimecast-Originator: redhat.com
Content-Language: en-US
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
X-BeenThere: dev@dpdk.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://mails.dpdk.org/options/dev>,
<mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://mails.dpdk.org/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://mails.dpdk.org/listinfo/dev>,
<mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
On 8/20/22 04:31, Hernan Vargas wrote:
> LDPC Encoder input may need to be padded to avoid small beat for ACC100.
> Padding 5GDL input buffer length (BLEN) to avoid case (BLEN % 64) <= 8.
> Adding protection for corner case to avoid for 5GDL occurrence of last
> beat within the ACC100 fabric with <= 8B which might trigger a fabric
> corner case hang issue.
Again, it looks like a bug fix.
> Signed-off-by: Hernan Vargas <hernan.vargas@intel.com>
> ---
> drivers/baseband/acc100/rte_acc100_pmd.c | 28 ++++++++++++++----------
> 1 file changed, 16 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/baseband/acc100/rte_acc100_pmd.c b/drivers/baseband/acc100/rte_acc100_pmd.c
> index cc7d146e74..4849d822d1 100644
> --- a/drivers/baseband/acc100/rte_acc100_pmd.c
> +++ b/drivers/baseband/acc100/rte_acc100_pmd.c
> @@ -1304,7 +1304,6 @@ acc100_fcw_td_fill(const struct rte_bbdev_dec_op *op, struct acc100_fcw_td *fcw)
> RTE_BBDEV_TURBO_HALF_ITERATION_EVEN);
> }
>
> -#ifdef RTE_LIBRTE_BBDEV_DEBUG
>
> static inline bool
> is_acc100(struct acc100_queue *q)
> @@ -1317,7 +1316,6 @@ validate_op_required(struct acc100_queue *q)
> {
> return is_acc100(q);
> }
> -#endif
>
> /* Fill in a frame control word for LDPC decoding. */
> static inline void
> @@ -1773,12 +1771,24 @@ acc100_dma_desc_te_fill(struct rte_bbdev_enc_op *op,
> return 0;
> }
>
> +/* May need to pad LDPC Encoder input to avoid small beat for ACC100 */
> +static inline uint16_t
> +pad_le_in(uint16_t blen, struct acc100_queue *q)
> +{
> + if (!is_acc100(q))
> + return blen;
> + uint16_t last_beat = blen % 64;
Do not mix variables declaration & code.
> + if ((last_beat > 0) && (last_beat <= 8))
> + blen += 8;
> + return blen;
> +}
> +
> static inline int
> acc100_dma_desc_le_fill(struct rte_bbdev_enc_op *op,
> struct acc100_dma_req_desc *desc, struct rte_mbuf **input,
> struct rte_mbuf *output, uint32_t *in_offset,
> uint32_t *out_offset, uint32_t *out_length,
> - uint32_t *mbuf_total_left, uint32_t *seg_total_left)
> + uint32_t *mbuf_total_left, uint32_t *seg_total_left, struct acc100_queue *q)
> {
> int next_triplet = 1; /* FCW already done */
> uint16_t K, in_length_in_bits, in_length_in_bytes;
> @@ -1802,8 +1812,7 @@ acc100_dma_desc_le_fill(struct rte_bbdev_enc_op *op,
> }
>
> next_triplet = acc100_dma_fill_blk_type_in(desc, input, in_offset,
> - in_length_in_bytes,
> - seg_total_left, next_triplet);
> + pad_le_in(in_length_in_bytes, q), seg_total_left, next_triplet);
> if (unlikely(next_triplet < 0)) {
> rte_bbdev_log(ERR,
> "Mismatch between data to process and mbuf data length in bbdev_op: %p",
> @@ -2531,7 +2540,7 @@ enqueue_ldpc_enc_n_op_cb(struct acc100_queue *q, struct rte_bbdev_enc_op **ops,
> acc100_header_init(&desc->req);
> desc->req.numCBs = num;
>
> - in_length_in_bytes = ops[0]->ldpc_enc.input.data->data_len;
> + in_length_in_bytes = pad_le_in(ops[0]->ldpc_enc.input.data->data_len, q);
> out_length = (enc->cb_params.e + 7) >> 3;
> desc->req.m2dlen = 1 + num;
> desc->req.d2mlen = num;
> @@ -2600,7 +2609,7 @@ enqueue_ldpc_enc_one_op_cb(struct acc100_queue *q, struct rte_bbdev_enc_op *op,
>
> ret = acc100_dma_desc_le_fill(op, &desc->req, &input, output,
> &in_offset, &out_offset, &out_length, &mbuf_total_left,
> - &seg_total_left);
> + &seg_total_left, q);
>
> if (unlikely(ret < 0))
> return ret;
> @@ -3921,9 +3930,6 @@ dequeue_enc_one_op_cb(struct acc100_queue *q, struct rte_bbdev_enc_op **ref_op,
>
> /* Clearing status, it will be set based on response */
> op->status = 0;
> -
> - op->status |= ((rsp.input_err)
> - ? (1 << RTE_BBDEV_DATA_ERROR) : 0);
Is this really related to the purpose of the patch?
> op->status |= ((rsp.dma_err) ? (1 << RTE_BBDEV_DRV_ERROR) : 0);
> op->status |= ((rsp.fcw_err) ? (1 << RTE_BBDEV_DRV_ERROR) : 0);
>
> @@ -3994,8 +4000,6 @@ dequeue_enc_one_op_tb(struct acc100_queue *q, struct rte_bbdev_enc_op **ref_op,
> rte_bbdev_log_debug("Resp. desc %p: %x", desc,
> rsp.val);
>
> - op->status |= ((rsp.input_err)
> - ? (1 << RTE_BBDEV_DATA_ERROR) : 0);
Ditto.
> op->status |= ((rsp.dma_err) ? (1 << RTE_BBDEV_DRV_ERROR) : 0);
> op->status |= ((rsp.fcw_err) ? (1 << RTE_BBDEV_DRV_ERROR) : 0);
>