From: Ferruh Yigit <ferruh.yigit@intel.com>
To: Shahaf Shuler <shahafs@mellanox.com>,
Rafal Kozik <rk@semihalf.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: "mw@semihalf.com" <mw@semihalf.com>,
"mk@semihalf.com" <mk@semihalf.com>,
"gtzalik@amazon.com" <gtzalik@amazon.com>,
"evgenys@amazon.com" <evgenys@amazon.com>,
"matua@amazon.com" <matua@amazon.com>,
"igorch@amazon.com" <igorch@amazon.com>
Subject: Re: [dpdk-dev] [PATCH 1/2] net/ena: convert to new Tx offloads API
Date: Wed, 17 Jan 2018 18:58:21 +0000 [thread overview]
Message-ID: <40a3b871-094f-0873-789d-d24a5ad4dc19@intel.com> (raw)
In-Reply-To: <VI1PR05MB3149A725C6AE527E388A9A36C3E90@VI1PR05MB3149.eurprd05.prod.outlook.com>
On 1/17/2018 6:56 AM, Shahaf Shuler wrote:
> Tuesday, January 16, 2018 1:53 PM, Rafal Kozik:
>> Subject: [dpdk-dev] [PATCH 1/2] net/ena: convert to new Tx offloads API
>>
>> Ethdev Tx offloads API has changed since:
>>
>> commit cba7f53b717d ("ethdev: introduce Tx queue offloads API")
>>
>> This commit support the new Tx offloads API. Queue configuration is stored
>> in ena_ring.offloads. During preparing mbufs for tx, offloads are allowed only
>> if appropriate flags in this field are set.
>>
>> Signed-off-by: Rafal Kozik <rk@semihalf.com>
>> ---
>> drivers/net/ena/ena_ethdev.c | 73
>> +++++++++++++++++++++++++++++++++++---------
>> drivers/net/ena/ena_ethdev.h | 3 ++
>> 2 files changed, 61 insertions(+), 15 deletions(-)
>>
>> diff --git a/drivers/net/ena/ena_ethdev.c b/drivers/net/ena/ena_ethdev.c
>> index 22db895..6473776 100644
>> --- a/drivers/net/ena/ena_ethdev.c
>> +++ b/drivers/net/ena/ena_ethdev.c
>> @@ -164,6 +164,14 @@ static const struct ena_stats
>> ena_stats_ena_com_strings[] = {
>> #define ENA_STATS_ARRAY_RX ARRAY_SIZE(ena_stats_rx_strings)
>> #define ENA_STATS_ARRAY_ENA_COM
>> ARRAY_SIZE(ena_stats_ena_com_strings)
>>
>> +#define QUEUE_OFFLOADS (DEV_TX_OFFLOAD_TCP_CKSUM |\
>> + DEV_TX_OFFLOAD_UDP_CKSUM |\
>> + DEV_TX_OFFLOAD_IPV4_CKSUM |\
>> + DEV_TX_OFFLOAD_TCP_TSO)
>> +#define MBUF_OFFLOADS (PKT_TX_L4_MASK |\
>> + PKT_TX_IP_CKSUM |\
>> + PKT_TX_TCP_SEG)
>> +
>> /** Vendor ID used by Amazon devices */ #define
>> PCI_VENDOR_ID_AMAZON 0x1D0F
>> /** Amazon devices */
>> @@ -227,6 +235,8 @@ static int ena_rss_reta_query(struct rte_eth_dev
>> *dev,
>> struct rte_eth_rss_reta_entry64 *reta_conf,
>> uint16_t reta_size);
>> static int ena_get_sset_count(struct rte_eth_dev *dev, int sset);
>> +static bool ena_are_tx_queue_offloads_allowed(struct ena_adapter
>> *adapter,
>> + uint64_t offloads);
>>
>> static const struct eth_dev_ops ena_dev_ops = {
>> .dev_configure = ena_dev_configure,
>> @@ -280,21 +290,24 @@ static inline void ena_rx_mbuf_prepare(struct
>> rte_mbuf *mbuf, }
>>
>> static inline void ena_tx_mbuf_prepare(struct rte_mbuf *mbuf,
>> - struct ena_com_tx_ctx *ena_tx_ctx)
>> + struct ena_com_tx_ctx *ena_tx_ctx,
>> + uint64_t queue_offloads)
>> {
>> struct ena_com_tx_meta *ena_meta = &ena_tx_ctx->ena_meta;
>>
>> - if (mbuf->ol_flags &
>> - (PKT_TX_L4_MASK | PKT_TX_IP_CKSUM | PKT_TX_TCP_SEG)) {
>> + if ((mbuf->ol_flags & MBUF_OFFLOADS) &&
>> + (queue_offloads & QUEUE_OFFLOADS)) {
>> /* check if TSO is required */
>> - if (mbuf->ol_flags & PKT_TX_TCP_SEG) {
>> + if ((mbuf->ol_flags & PKT_TX_TCP_SEG) &&
>> + (queue_offloads & DEV_TX_OFFLOAD_TCP_TSO)) {
>> ena_tx_ctx->tso_enable = true;
>>
>> ena_meta->l4_hdr_len = GET_L4_HDR_LEN(mbuf);
>> }
>>
>> /* check if L3 checksum is needed */
>> - if (mbuf->ol_flags & PKT_TX_IP_CKSUM)
>> + if ((mbuf->ol_flags & PKT_TX_IP_CKSUM) &&
>> + (queue_offloads & DEV_TX_OFFLOAD_IPV4_CKSUM))
>> ena_tx_ctx->l3_csum_enable = true;
>>
>> if (mbuf->ol_flags & PKT_TX_IPV6) {
>> @@ -310,19 +323,17 @@ static inline void ena_tx_mbuf_prepare(struct
>> rte_mbuf *mbuf,
>> }
>>
>> /* check if L4 checksum is needed */
>> - switch (mbuf->ol_flags & PKT_TX_L4_MASK) {
>> - case PKT_TX_TCP_CKSUM:
>> + if ((mbuf->ol_flags & PKT_TX_TCP_CKSUM) &&
>> + (queue_offloads & DEV_TX_OFFLOAD_TCP_CKSUM)) {
>> ena_tx_ctx->l4_proto =
>> ENA_ETH_IO_L4_PROTO_TCP;
>> ena_tx_ctx->l4_csum_enable = true;
>> - break;
>> - case PKT_TX_UDP_CKSUM:
>> + } else if ((mbuf->ol_flags & PKT_TX_UDP_CKSUM) &&
>> + (queue_offloads &
>> DEV_TX_OFFLOAD_UDP_CKSUM)) {
>> ena_tx_ctx->l4_proto =
>> ENA_ETH_IO_L4_PROTO_UDP;
>> ena_tx_ctx->l4_csum_enable = true;
>> - break;
>> - default:
>> + } else {
>> ena_tx_ctx->l4_proto =
>> ENA_ETH_IO_L4_PROTO_UNKNOWN;
>> ena_tx_ctx->l4_csum_enable = false;
>> - break;
>> }
>>
>> ena_meta->mss = mbuf->tso_segsz;
>> @@ -945,7 +956,7 @@ static int ena_tx_queue_setup(struct rte_eth_dev
>> *dev,
>> uint16_t queue_idx,
>> uint16_t nb_desc,
>> __rte_unused unsigned int socket_id,
>> - __rte_unused const struct rte_eth_txconf
>> *tx_conf)
>> + const struct rte_eth_txconf *tx_conf)
>> {
>> struct ena_com_create_io_ctx ctx =
>> /* policy set to _HOST just to satisfy icc compiler */ @@ -
>> 982,6 +993,11 @@ static int ena_tx_queue_setup(struct rte_eth_dev *dev,
>> return -EINVAL;
>> }
>>
>> + if (!ena_are_tx_queue_offloads_allowed(adapter, tx_conf-
>>> offloads)) {
>> + RTE_LOG(ERR, PMD, "Unsupported queue offloads\n");
>> + return -EINVAL;
>> + }
>
> Here is it better to check also the ETH_TXQ_FLAGS_IGNORE.
> If application has not yet moved to the new API, then it won't set any port Tx offloads. So for old applications, the ena_are_tx_queue_offloads_allowed is not necessary.
But ethdev layer will set the offloads if ETH_TXQ_FLAGS_IGNORE is missing, can't
PMD always only rely on tx_conf->offloads ?
next prev parent reply other threads:[~2018-01-17 18:58 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-01-16 11:52 [dpdk-dev] [PATCH 0/2] net/ena: convert to new " Rafal Kozik
2018-01-16 11:52 ` [dpdk-dev] [PATCH 1/2] net/ena: convert to new Tx " Rafal Kozik
2018-01-17 6:56 ` Shahaf Shuler
2018-01-17 18:58 ` Ferruh Yigit [this message]
2018-01-18 6:26 ` Shahaf Shuler
2018-01-18 14:49 ` Ferruh Yigit
2018-01-17 8:23 ` [dpdk-dev] [PATCH v2 " Rafal Kozik
2018-01-17 18:58 ` Ferruh Yigit
2018-01-18 9:18 ` Rafał Kozik
2018-01-18 14:49 ` Ferruh Yigit
2018-01-18 14:53 ` Ferruh Yigit
2018-01-18 14:59 ` Michał Krawczyk
2018-01-18 15:12 ` Ferruh Yigit
2018-01-16 11:52 ` [dpdk-dev] [PATCH 2/2] net/ena: convert to new Rx " Rafal Kozik
2018-01-17 6:57 ` Shahaf Shuler
2018-01-17 8:26 ` [dpdk-dev] [PATCH v2 " Rafal Kozik
2018-01-18 15:06 ` Michał Krawczyk
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=40a3b871-094f-0873-789d-d24a5ad4dc19@intel.com \
--to=ferruh.yigit@intel.com \
--cc=dev@dpdk.org \
--cc=evgenys@amazon.com \
--cc=gtzalik@amazon.com \
--cc=igorch@amazon.com \
--cc=matua@amazon.com \
--cc=mk@semihalf.com \
--cc=mw@semihalf.com \
--cc=rk@semihalf.com \
--cc=shahafs@mellanox.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).