From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id 76946A04DB; Thu, 15 Oct 2020 12:11:43 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 405AA1DE04; Thu, 15 Oct 2020 12:11:42 +0200 (CEST) Received: from shelob.oktetlabs.ru (shelob.oktetlabs.ru [91.220.146.113]) by dpdk.org (Postfix) with ESMTP id 49ED51DDE5 for ; Thu, 15 Oct 2020 12:11:40 +0200 (CEST) Received: from [192.168.38.17] (aros.oktetlabs.ru [192.168.38.17]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by shelob.oktetlabs.ru (Postfix) with ESMTPSA id EC7CE7F4ED; Thu, 15 Oct 2020 13:11:38 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 shelob.oktetlabs.ru EC7CE7F4ED DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=oktetlabs.ru; s=default; t=1602756699; bh=A8tBWIZVSoBM6r6mHsOfTvoVfSuWyIVgc4LXjWpZY0Y=; h=Subject:To:Cc:References:From:Date:In-Reply-To; b=KfxYlutfb5/O5xWJTBBdUZ32ojKzYWWiNjJVhgz7WTwTaPY3CEm7xQwjn4rKAwLL8 /78MYM1giVRQtOrkpCmKsthxOs3xKxZyH+v2lL50H2q3RX8r/S3Kzk+eZSMy65Izro tWNjk9oblq4Q9ijNj5GwGXTBjzB4d/KMGKC+J+lg= To: Viacheslav Ovsiienko , dev@dpdk.org Cc: thomas@monjalon.net, stephen@networkplumber.org, ferruh.yigit@intel.com, olivier.matz@6wind.com, jerinjacobk@gmail.com, maxime.coquelin@redhat.com, david.marchand@redhat.com, arybchenko@solarflare.com References: <1602699122-15737-1-git-send-email-viacheslavo@nvidia.com> <1602699122-15737-2-git-send-email-viacheslavo@nvidia.com> From: Andrew Rybchenko Organization: OKTET Labs Message-ID: Date: Thu, 15 Oct 2020 13:11:38 +0300 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.3.1 MIME-Version: 1.0 In-Reply-To: <1602699122-15737-2-git-send-email-viacheslavo@nvidia.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Subject: Re: [dpdk-dev] [PATCH v6 1/6] ethdev: introduce Rx buffer split X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On 10/14/20 9:11 PM, Viacheslav Ovsiienko wrote: > The DPDK datapath in the transmit direction is very flexible. > An application can build the multi-segment packet and manages > almost all data aspects - the memory pools where segments > are allocated from, the segment lengths, the memory attributes > like external buffers, registered for DMA, etc. > > In the receiving direction, the datapath is much less flexible, > an application can only specify the memory pool to configure the > receiving queue and nothing more. In order to extend receiving > datapath capabilities it is proposed to add the way to provide > extended information how to split the packets being received. > > The following structure is introduced to specify the Rx packet > segment: > > struct rte_eth_rxseg { > struct rte_mempool *mp; /* memory pools to allocate segment from */ > uint16_t length; /* segment maximal data length, > configures "split point" */ > uint16_t offset; /* data offset from beginning > of mbuf data buffer */ > uint32_t reserved; /* reserved field */ > }; > > The segment descriptions are added to the rte_eth_rxconf structure: > rx_seg - pointer the array of segment descriptions, each element > describes the memory pool, maximal data length, initial > data offset from the beginning of data buffer in mbuf. > This array allows to specify the different settings for > each segment in individual fashion. > rx_nseg - number of elements in the array > > If the extended segment descriptions is provided with these new > fields the mp parameter of the rte_eth_rx_queue_setup must be > specified as NULL to avoid ambiguity. > > There are two options to specifiy Rx buffer configuration: > - mp is not NULL, rx_conf.rx_seg is NULL, rx_conf.rx_nseg is zero, > it is compatible configuraion, follows existing implementation, > provides single pool and no description for segment sizes > and offsets. > - mp is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not > zero, it provides the extended configuration, individually for > each segment. > > The new offload flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT in device > capabilities is introduced to present the way for PMD to report to > application about supporting Rx packet split to configurable > segments. Prior invoking the rte_eth_rx_queue_setup() routine > application should check RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag. > > If the Rx queue is configured with new settings the packets being > received will be split into multiple segments pushed to the mbufs > with specified attributes. The PMD will split the received packets > into multiple segments according to the specification in the > description array: > > - the first network buffer will be allocated from the memory pool, > specified in the first segment description element, the second > network buffer - from the pool in the second segment description > element and so on. If there is no enough elements to describe > the buffer for entire packet of maximal length the pool from the > last valid element will be used to allocate the buffers from for the > rest of segments > > - the offsets from the segment description elements will provide > the data offset from the buffer beginning except the first mbuf - > for this one the offset is added to the RTE_PKTMBUF_HEADROOM to get > actual offset from the buffer beginning. If there is no enough > elements to describe the buffer for entire packet of maximal length > the offsets for the rest of segment will be supposed to be zero. > > - the data length being received to each segment is limited by the > length specified in the segment description element. The data > receiving starts with filling up the first mbuf data buffer, if the > specified maximal segment length is reached and there are data > remaining (packet is longer than buffer in the first mbuf) the > following data will be pushed to the next segment up to its own > maximal length. If the first two segments is not enough to store > all the packet remaining data the next (third) segment will > be engaged and so on. If the length in the segment description > element is zero the actual buffer size will be deduced from > the appropriate memory pool properties. If there is no enough > elements to describe the buffer for entire packet of maximal > length the buffer size will be deduced from the pool of the last > valid element for the remaining segments. > > For example, let's suppose we configured the Rx queue with the > following segments: > seg0 - pool0, len0=14B, off0=2 > seg1 - pool1, len1=20B, off1=128B > seg2 - pool2, len2=20B, off2=0B > seg3 - pool3, len3=512B, off3=0B > > The packet 46 bytes long will look like the following: > seg0 - 14B long @ RTE_PKTMBUF_HEADROOM + 2 in mbuf from pool0 > seg1 - 20B long @ 128 in mbuf from pool1 > seg2 - 12B long @ 0 in mbuf from pool2 > > The packet 1500 bytes long will look like the following: > seg0 - 14B @ RTE_PKTMBUF_HEADROOM + 2 in mbuf from pool0 > seg1 - 20B @ 128 in mbuf from pool1 > seg2 - 20B @ 0 in mbuf from pool2 > seg3 - 512B @ 0 in mbuf from pool3 > seg4 - 512B @ 0 in mbuf from pool3 > seg5 - 422B @ 0 in mbuf from pool3 > > The offload RTE_ETH_RX_OFFLOAD_SCATTER must be present and > configured to support new buffer split feature (if rx_nseg > is greater than one). > > The new approach would allow splitting the ingress packets into > multiple parts pushed to the memory with different attributes. > For example, the packet headers can be pushed to the embedded > data buffers within mbufs and the application data into > the external buffers attached to mbufs allocated from the > different memory pools. The memory attributes for the split > parts may differ either - for example the application data > may be pushed into the external memory located on the dedicated > physical device, say GPU or NVMe. This would improve the DPDK > receiving datapath flexibility with preserving compatibility > with existing API. > > Signed-off-by: Viacheslav Ovsiienko > --- > doc/guides/nics/features.rst | 15 +++++ > doc/guides/rel_notes/deprecation.rst | 5 -- > doc/guides/rel_notes/release_20_11.rst | 9 +++ > lib/librte_ethdev/rte_ethdev.c | 111 +++++++++++++++++++++++++-------- > lib/librte_ethdev/rte_ethdev.h | 62 +++++++++++++++++- > 5 files changed, 171 insertions(+), 31 deletions(-) > > diff --git a/doc/guides/nics/features.rst b/doc/guides/nics/features.rst > index dd8c955..832ea3b 100644 > --- a/doc/guides/nics/features.rst > +++ b/doc/guides/nics/features.rst > @@ -185,6 +185,21 @@ Supports receiving segmented mbufs. > * **[related] eth_dev_ops**: ``rx_pkt_burst``. > > > +.. _nic_features_buffer_split: > + > +Buffer Split on Rx > +------------------ > + > +Scatters the packets being received on specified boundaries to segmented mbufs. > + > +* **[uses] rte_eth_rxconf,rte_eth_rxmode**: ``offloads:RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT``. > +* **[uses] rte_eth_rxconf**: ``rx_conf.rx_seg, rx_conf.rx_nseg``. > +* **[implements] datapath**: ``Buffer Split functionality``. > +* **[provides] rte_eth_dev_info**: ``rx_offload_capa:RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT``. > +* **[provides] eth_dev_ops**: ``rxq_info_get:buffer_split``. > +* **[related] API**: ``rte_eth_rx_queue_setup()``. > + > + > .. _nic_features_lro: > > LRO > diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst > index 584e720..232cd54 100644 > --- a/doc/guides/rel_notes/deprecation.rst > +++ b/doc/guides/rel_notes/deprecation.rst > @@ -138,11 +138,6 @@ Deprecation Notices > In 19.11 PMDs will still update the field even when the offload is not > enabled. > > -* ethdev: Add new fields to ``rte_eth_rxconf`` to configure the receiving > - queues to split ingress packets into multiple segments according to the > - specified lengths into the buffers allocated from the specified > - memory pools. The backward compatibility to existing API is preserved. > - > * ethdev: ``rx_descriptor_done`` dev_ops and ``rte_eth_rx_descriptor_done`` > will be removed in 21.11. > Existing ``rte_eth_rx_descriptor_status`` and ``rte_eth_tx_descriptor_status`` > diff --git a/doc/guides/rel_notes/release_20_11.rst b/doc/guides/rel_notes/release_20_11.rst > index bcc0fc2..bcc2479 100644 > --- a/doc/guides/rel_notes/release_20_11.rst > +++ b/doc/guides/rel_notes/release_20_11.rst > @@ -60,6 +60,12 @@ New Features > Added the FEC API which provides functions for query FEC capabilities and > current FEC mode from device. Also, API for configuring FEC mode is also provided. > > +* **Introduced extended buffer description for receiving.** > + > + Added the extended Rx buffer description for Rx queue setup routine > + providing the individual settings for each Rx segment with maximal size, > + buffer offset and memory pool to allocate data buffers from. > + > * **Updated Broadcom bnxt driver.** > > Updated the Broadcom bnxt driver with new features and improvements, including: > @@ -253,6 +259,9 @@ API Changes > As the data of ``uint8_t`` will be truncated when queue number under > a TC is greater than 256. > > +* ethdev: Added fields rx_seg and rx_nseg to rte_eth_rxconf structure > + to provide extended description of the receiving buffer. > + > * vhost: Moved vDPA APIs from experimental to stable. > > * rawdev: Added a structure size parameter to the functions > diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c > index 892c246..96ecb91 100644 > --- a/lib/librte_ethdev/rte_ethdev.c > +++ b/lib/librte_ethdev/rte_ethdev.c > @@ -105,6 +105,9 @@ struct rte_eth_xstats_name_off { > #define RTE_RX_OFFLOAD_BIT2STR(_name) \ > { DEV_RX_OFFLOAD_##_name, #_name } > > +#define RTE_ETH_RX_OFFLOAD_BIT2STR(_name) \ > + { RTE_ETH_RX_OFFLOAD_##_name, #_name } > + > static const struct { > uint64_t offload; > const char *name; > @@ -128,9 +131,11 @@ struct rte_eth_xstats_name_off { > RTE_RX_OFFLOAD_BIT2STR(SCTP_CKSUM), > RTE_RX_OFFLOAD_BIT2STR(OUTER_UDP_CKSUM), > RTE_RX_OFFLOAD_BIT2STR(RSS_HASH), > + RTE_ETH_RX_OFFLOAD_BIT2STR(BUFFER_SPLIT), > }; > > #undef RTE_RX_OFFLOAD_BIT2STR > +#undef RTE_ETH_RX_OFFLOAD_BIT2STR > > #define RTE_TX_OFFLOAD_BIT2STR(_name) \ > { DEV_TX_OFFLOAD_##_name, #_name } > @@ -1784,38 +1789,94 @@ struct rte_eth_dev * > return -EINVAL; > } > > - if (mp == NULL) { > - RTE_ETHDEV_LOG(ERR, "Invalid null mempool pointer\n"); > - return -EINVAL; > - } > - > RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_setup, -ENOTSUP); > > - /* > - * Check the size of the mbuf data buffer. > - * This value must be provided in the private data of the memory pool. > - * First check that the memory pool has a valid private data. > - */ > ret = rte_eth_dev_info_get(port_id, &dev_info); > if (ret != 0) > return ret; > > - if (mp->private_data_size < sizeof(struct rte_pktmbuf_pool_private)) { > - RTE_ETHDEV_LOG(ERR, "%s private_data_size %d < %d\n", > - mp->name, (int)mp->private_data_size, > - (int)sizeof(struct rte_pktmbuf_pool_private)); > - return -ENOSPC; > - } > - mbp_buf_size = rte_pktmbuf_data_room_size(mp); > + if (mp) { Please, compare vs NULL [1] [1] https://doc.dpdk.org/guides/contributing/coding_style.html#null-pointers > + /* Single pool configuration check. */ > + if (rx_conf->rx_seg || rx_conf->rx_nseg) { Please, compare vs NULL and 0. IMHO, rx_nsegs check is sufficient. If it is 0, nobody cares what is in rx_seg. > + RTE_ETHDEV_LOG(ERR, > + "Ambiguous segment configuration\n"); > + return -EINVAL; > + } > + /* > + * Check the size of the mbuf data buffer, this value > + * must be provided in the private data of the memory pool. > + * First check that the memory pool(s) has a valid private data. > + */ > + if (mp->private_data_size < > + sizeof(struct rte_pktmbuf_pool_private)) { > + RTE_ETHDEV_LOG(ERR, "%s private_data_size %u < %u\n", > + mp->name, mp->private_data_size, > + (unsigned int) > + sizeof(struct rte_pktmbuf_pool_private)); > + return -ENOSPC; > + } > + mbp_buf_size = rte_pktmbuf_data_room_size(mp); > + if (mbp_buf_size < dev_info.min_rx_bufsize + > + RTE_PKTMBUF_HEADROOM) { > + RTE_ETHDEV_LOG(ERR, > + "%s mbuf_data_room_size %u < %u" > + " (RTE_PKTMBUF_HEADROOM=%u +" > + " min_rx_bufsize(dev)=%u)\n", Do not split format string. It is not a problem that it is long. > + mp->name, mbp_buf_size, > + RTE_PKTMBUF_HEADROOM + > + dev_info.min_rx_bufsize, > + RTE_PKTMBUF_HEADROOM, > + dev_info.min_rx_bufsize); > + return -EINVAL; > + } > + } else { > + const struct rte_eth_rxseg *rx_seg = rx_conf->rx_seg; > + uint16_t n_seg = rx_conf->rx_nseg; > + uint16_t seg_idx; > > - if (mbp_buf_size < dev_info.min_rx_bufsize + RTE_PKTMBUF_HEADROOM) { > - RTE_ETHDEV_LOG(ERR, > - "%s mbuf_data_room_size %d < %d (RTE_PKTMBUF_HEADROOM=%d + min_rx_bufsize(dev)=%d)\n", > - mp->name, (int)mbp_buf_size, > - (int)(RTE_PKTMBUF_HEADROOM + dev_info.min_rx_bufsize), > - (int)RTE_PKTMBUF_HEADROOM, > - (int)dev_info.min_rx_bufsize); > - return -EINVAL; > + /* Extended multi-segment configuration check. */ > + if (!rx_conf->rx_seg || !rx_conf->rx_nseg) { Please, compare vs NULL and 0 > + RTE_ETHDEV_LOG(ERR, > + "Memory pool is null and no" > + " extended configuration provided\n"); Do not split format string. It is not a problem that it is long. > + return -EINVAL; > + } > + /* > + * Check the sizes and offsets against buffer sizes > + * for each segment specified in extended configuration. > + */ > + for (seg_idx = 0; seg_idx < n_seg; seg_idx++) { > + struct rte_mempool *mpl = rx_seg[seg_idx].mp; > + uint32_t length = rx_seg[seg_idx].length; > + uint32_t offset = rx_seg[seg_idx].offset; > + uint32_t head_room = seg_idx ? 0 : RTE_PKTMBUF_HEADROOM; Why? Shouldn't it be in offset? IMHO too many offsets this way. > + > + if (mpl == NULL) { > + RTE_ETHDEV_LOG(ERR, "null mempool pointer\n"); > + return -EINVAL; > + } > + if (mpl->private_data_size < > + sizeof(struct rte_pktmbuf_pool_private)) { > + RTE_ETHDEV_LOG(ERR, > + "%s private_data_size %u < %u\n", > + mpl->name, > + mpl->private_data_size, > + (unsigned int)sizeof(struct > + rte_pktmbuf_pool_private)); > + return -ENOSPC; > + } > + mbp_buf_size = rte_pktmbuf_data_room_size(mpl); > + length = length ? length : (mbp_buf_size - head_room); Compare length with 0. What does ensure that mbp_buf_size is greater or equal to head_room size? > + if (mbp_buf_size < length + offset + head_room) { > + RTE_ETHDEV_LOG(ERR, > + "%s mbuf_data_room_size %u < %u" > + " (segment length=%u +" > + " segment offset=%u)\n", Do not split format string. It is not a problem that it is long. > + mpl->name, mbp_buf_size, > + length + offset, length, offset); > + return -EINVAL; > + } > + } > } > > /* Use default specified by driver, if nb_rx_desc is zero */ > diff --git a/lib/librte_ethdev/rte_ethdev.h b/lib/librte_ethdev/rte_ethdev.h > index 5bcfbb8..e019f4a 100644 > --- a/lib/librte_ethdev/rte_ethdev.h > +++ b/lib/librte_ethdev/rte_ethdev.h > @@ -970,6 +970,16 @@ struct rte_eth_txmode { > }; > > /** > + * A structure used to configure an RX packet segment to split. RX -> Rx > + */ > +struct rte_eth_rxseg { > + struct rte_mempool *mp; /**< Memory pool to allocate segment from. */ > + uint16_t length; /**< Segment data length, configures split point. */ > + uint16_t offset; /**< Data offset from beginning of mbuf data buffer. */ > + uint32_t reserved; /**< Reserved field. */ > +}; > + > +/** > * A structure used to configure an RX ring of an Ethernet port. > */ > struct rte_eth_rxconf { > @@ -977,6 +987,43 @@ struct rte_eth_rxconf { > uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */ > uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */ > uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */ > + uint16_t rx_nseg; /**< Number of descriptions in rx_seg array. */ > + /** > + * Points to the array of segment descriptions. Each array element > + * describes the properties for each segment in the receiving > + * buffer. > + * > + * If RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag is set in offloads field, > + * the PMD will split the received packets into multiple segments > + * according to the specification in the description array: > + * > + * - the first network buffer will be allocated from the memory pool, > + * specified in the first array element, the second buffer, from the > + * pool in the second element, and so on. > + * > + * - the offsets from the segment description elements specify > + * the data offset from the buffer beginning except the first mbuf. > + * For this one the offset is added with RTE_PKTMBUF_HEADROOM. > + * > + * - the lengthes in the elements define the maximal data amount > + * being received to each segment. The receiving starts with filling > + * up the first mbuf data buffer up to specified length. If the > + * there are data remaining (packet is longer than buffer in the first > + * mbuf) the following data will be pushed to the next segment > + * up to its own length, and so on. > + * > + * - If the length in the segment description element is zero > + * the actual buffer size will be deduced from the appropriate > + * memory pool properties. > + * > + * - if there is not enough elements to describe the buffer for entire > + * packet of maximal length the following parameters will be used > + * for the all remaining segments: > + * - pool from the last valid element > + * - the buffer size from this pool > + * - zero offset > + */ > + struct rte_eth_rxseg *rx_seg; > /** > * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags. > * Only offloads set on rx_queue_offload_capa or rx_offload_capa > @@ -1260,6 +1307,7 @@ struct rte_eth_conf { > #define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000 > #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000 > #define DEV_RX_OFFLOAD_RSS_HASH 0x00080000 > +#define RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT 0x00100000 > > #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \ > DEV_RX_OFFLOAD_UDP_CKSUM | \ > @@ -2027,9 +2075,21 @@ int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue, > * No need to repeat any bit in rx_conf->offloads which has already been > * enabled in rte_eth_dev_configure() at port level. An offloading enabled > * at port level can't be disabled at queue level. > + * The configuration structure also contains the pointer to the array > + * of the receiving buffer segment descriptions, see rx_seg and rx_nseg > + * fields, this extended configuration might be used by split offloads like > + * RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT. If mp_pool is not NULL, > + * the extended configuration fields must be set to NULL and zero. > * @param mb_pool > * The pointer to the memory pool from which to allocate *rte_mbuf* network > - * memory buffers to populate each descriptor of the receive ring. > + * memory buffers to populate each descriptor of the receive ring. There are > + * two options to provide Rx buffer configuration: > + * - single pool: > + * mb_pool is not NULL, rx_conf.rx_seg is NULL, rx_conf.rx_nseg is 0. > + * - multiple segments description: > + * mb_pool is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not 0. > + * Taken only if flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT is set in offloads. > + * > * @return > * - 0: Success, receive queue correctly set up. > * - -EIO: if device is removed. >