From: Jerin Jacob <jerinjacobk@gmail.com>
To: Jerin Jacob <jerinj@marvell.com>
Cc: dpdk-dev <dev@dpdk.org>,
Nithin Dabilpuram <ndabilpuram@marvell.com>,
Kiran Kumar K <kirankumark@marvell.com>,
John McNamara <john.mcnamara@intel.com>,
Marko Kovacevic <marko.kovacevic@intel.com>,
Vamsi Attunuru <vattunuru@marvell.com>
Subject: Re: [dpdk-dev] [PATCH] net/octeontx2: configure RSS adder as tag lsb bits
Date: Mon, 3 Feb 2020 13:01:39 +0530 [thread overview]
Message-ID: <CALBAE1Mq1O_sGj=R4rGSGcqm+fNMPkV_vMG6GksW3jSxYnz4dA@mail.gmail.com> (raw)
In-Reply-To: <20200129172412.674155-1-jerinj@marvell.com>
On Wed, Jan 29, 2020 at 10:54 PM <jerinj@marvell.com> wrote:
>
> From: Jerin Jacob <jerinj@marvell.com>
>
> Before C0 HW revision, The RSS adder was computed based the following
> static formula.
>
> rss_adder<7:0> = flow_tag<7:0> ^ flow_tag<15:8> ^
> flow_tag<23:16> ^ flow_tag<31:24>
>
> The above scheme has the following drawbacks:
> 1) It is not in line with other standard NIC behavior.
> 2) There can be an SW use case where SW can compute the hash
> upfront using Toeplitz function and predict the queue selection
> to optimize some packet lookup function. The nonstandard
> way of doing XOR makes the consumer to not predict the queue selection.
>
> C0 HW revision onward, The HW can configure the
> rss_adder<7:0> as flow_tag<7:0> to align with standard NICs.
>
> This patch adds an option to select legacy RSS adder mode
> using tag_as_xor=1 devargs option while keeping the standard NIC
> behavior as default.
>
> Signed-off-by: Jerin Jacob <jerinj@marvell.com>
Applied to dpdk-next-net-mrvl/master. Thanks
> ---
> doc/guides/nics/octeontx2.rst | 19 ++++++++++++++++++-
> drivers/common/octeontx2/otx2_mbox.h | 4 +++-
> drivers/net/octeontx2/otx2_ethdev.c | 2 ++
> drivers/net/octeontx2/otx2_ethdev.h | 1 +
> drivers/net/octeontx2/otx2_ethdev_devargs.c | 8 +++++++-
> 5 files changed, 31 insertions(+), 3 deletions(-)
>
> diff --git a/doc/guides/nics/octeontx2.rst b/doc/guides/nics/octeontx2.rst
> index db62a4523f..2ff1dbda0e 100644
> --- a/doc/guides/nics/octeontx2.rst
> +++ b/doc/guides/nics/octeontx2.rst
> @@ -165,7 +165,7 @@ Runtime Config Options
> With the above configuration, each send queue's decscriptor buffer count is
> limited to a maximum of 64 buffers.
>
> -- ``switch header enable`` (default ``none``)
> +- ``Switch header enable`` (default ``none``)
>
> A port can be configured to a specific switch header type by using
> ``switch_header`` ``devargs`` parameter.
> @@ -178,6 +178,23 @@ Runtime Config Options
> traffic on this port should be higig2 traffic only. Supported switch header
> types are "higig2" and "dsa".
>
> +- ``RSS tag as XOR`` (default ``0``)
> +
> + C0 HW revision onward, The HW gives an option to configure the RSS adder as
> +
> + * ``rss_adder<7:0> = flow_tag<7:0> ^ flow_tag<15:8> ^ flow_tag<23:16> ^ flow_tag<31:24>``
> +
> + * ``rss_adder<7:0> = flow_tag<7:0>``
> +
> + Latter one aligns with standard NIC behavior vs former one is a legacy
> + RSS adder scheme used in OCTEON TX2 products.
> +
> + By default, the driver runs in the latter mode from C0 HW revision onward.
> + Setting this flag to 1 to select the legacy mode.
> +
> + For example to select the legacy mode(RSS tag adder as XOR)::
> + -w 0002:02:00.0,tag_as_xor=1
> +
> .. note::
>
> Above devarg parameters are configurable per device, user needs to pass the
> diff --git a/drivers/common/octeontx2/otx2_mbox.h b/drivers/common/octeontx2/otx2_mbox.h
> index 4972b8a6eb..e87a53c7d5 100644
> --- a/drivers/common/octeontx2/otx2_mbox.h
> +++ b/drivers/common/octeontx2/otx2_mbox.h
> @@ -89,7 +89,7 @@ struct mbox_msghdr {
> #define OTX2_MBOX_RSP_SIG (0xbeef)
> /* Signature, for validating corrupted msgs */
> uint16_t __otx2_io sig;
> -#define OTX2_MBOX_VERSION (0x0003)
> +#define OTX2_MBOX_VERSION (0x0004)
> /* Version of msg's structure for this ID */
> uint16_t __otx2_io ver;
> /* Offset of next msg within mailbox region */
> @@ -707,6 +707,8 @@ struct nix_lf_alloc_req {
> uint16_t __otx2_io sso_func;
> uint64_t __otx2_io rx_cfg; /* See NIX_AF_LF(0..127)_RX_CFG */
> uint64_t __otx2_io way_mask;
> +#define NIX_LF_RSS_TAG_LSB_AS_ADDER BIT_ULL(0)
> + uint64_t flags;
> };
>
> struct nix_lf_alloc_rsp {
> diff --git a/drivers/net/octeontx2/otx2_ethdev.c b/drivers/net/octeontx2/otx2_ethdev.c
> index 11f8c786bd..1ec234b3cc 100644
> --- a/drivers/net/octeontx2/otx2_ethdev.c
> +++ b/drivers/net/octeontx2/otx2_ethdev.c
> @@ -70,6 +70,8 @@ nix_lf_alloc(struct otx2_eth_dev *dev, uint32_t nb_rxq, uint32_t nb_txq)
> req->rx_cfg |= BIT_ULL(36 /* CSUM_IL4 */);
> }
> req->rx_cfg |= BIT_ULL(32 /* DROP_RE */);
> + if (dev->rss_tag_as_xor == 0)
> + req->flags = NIX_LF_RSS_TAG_LSB_AS_ADDER;
>
> rc = otx2_mbox_process_msg(mbox, (void *)&rsp);
> if (rc)
> diff --git a/drivers/net/octeontx2/otx2_ethdev.h b/drivers/net/octeontx2/otx2_ethdev.h
> index c075b8d1a4..49fed95e7f 100644
> --- a/drivers/net/octeontx2/otx2_ethdev.h
> +++ b/drivers/net/octeontx2/otx2_ethdev.h
> @@ -285,6 +285,7 @@ struct otx2_eth_dev {
> uintptr_t base;
> uintptr_t lmt_addr;
> uint16_t scalar_ena;
> + uint16_t rss_tag_as_xor;
> uint16_t max_sqb_count;
> uint16_t rx_offload_flags; /* Selected Rx offload flags(NIX_RX_*_F) */
> uint64_t rx_offloads;
> diff --git a/drivers/net/octeontx2/otx2_ethdev_devargs.c b/drivers/net/octeontx2/otx2_ethdev_devargs.c
> index 04da1abbdb..ab1e14ea65 100644
> --- a/drivers/net/octeontx2/otx2_ethdev_devargs.c
> +++ b/drivers/net/octeontx2/otx2_ethdev_devargs.c
> @@ -109,6 +109,7 @@ parse_switch_header_type(const char *key, const char *value, void *extra_args)
> #define OTX2_FLOW_PREALLOC_SIZE "flow_prealloc_size"
> #define OTX2_FLOW_MAX_PRIORITY "flow_max_priority"
> #define OTX2_SWITCH_HEADER_TYPE "switch_header"
> +#define OTX2_RSS_TAG_AS_XOR "tag_as_xor"
>
> int
> otx2_ethdev_parse_devargs(struct rte_devargs *devargs, struct otx2_eth_dev *dev)
> @@ -119,6 +120,7 @@ otx2_ethdev_parse_devargs(struct rte_devargs *devargs, struct otx2_eth_dev *dev)
> uint16_t switch_header_type = 0;
> uint16_t flow_max_priority = 3;
> uint16_t scalar_enable = 0;
> + uint16_t rss_tag_as_xor = 0;
> struct rte_kvargs *kvlist;
>
> if (devargs == NULL)
> @@ -140,10 +142,13 @@ otx2_ethdev_parse_devargs(struct rte_devargs *devargs, struct otx2_eth_dev *dev)
> &parse_flow_max_priority, &flow_max_priority);
> rte_kvargs_process(kvlist, OTX2_SWITCH_HEADER_TYPE,
> &parse_switch_header_type, &switch_header_type);
> + rte_kvargs_process(kvlist, OTX2_RSS_TAG_AS_XOR,
> + &parse_flag, &rss_tag_as_xor);
> rte_kvargs_free(kvlist);
>
> null_devargs:
> dev->scalar_ena = scalar_enable;
> + dev->rss_tag_as_xor = rss_tag_as_xor;
> dev->max_sqb_count = sqb_count;
> dev->rss_info.rss_size = rss_size;
> dev->npc_flow.flow_prealloc_size = flow_prealloc_size;
> @@ -161,4 +166,5 @@ RTE_PMD_REGISTER_PARAM_STRING(net_octeontx2,
> OTX2_MAX_SQB_COUNT "=<8-512>"
> OTX2_FLOW_PREALLOC_SIZE "=<1-32>"
> OTX2_FLOW_MAX_PRIORITY "=<1-32>"
> - OTX2_SWITCH_HEADER_TYPE "=<higig2|dsa>");
> + OTX2_SWITCH_HEADER_TYPE "=<higig2|dsa>"
> + OTX2_RSS_TAG_AS_XOR "=1");
> --
> 2.24.1
>
prev parent reply other threads:[~2020-02-03 7:31 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-29 17:24 jerinj
2020-02-03 7:31 ` Jerin Jacob [this message]
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='CALBAE1Mq1O_sGj=R4rGSGcqm+fNMPkV_vMG6GksW3jSxYnz4dA@mail.gmail.com' \
--to=jerinjacobk@gmail.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=john.mcnamara@intel.com \
--cc=kirankumark@marvell.com \
--cc=marko.kovacevic@intel.com \
--cc=ndabilpuram@marvell.com \
--cc=vattunuru@marvell.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).