From: Gleb Natapov <gleb@cloudius-systems.com>
To: Vladimir Medvedkin <medvedkinv@gmail.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] Add toeplitz hash algorithm
Date: Thu, 9 Apr 2015 09:37:42 +0300 [thread overview]
Message-ID: <20150409063742.GA16818@cloudius-systems.com> (raw)
In-Reply-To: <1428519973-10550-1-git-send-email-medvedkinv@gmail.com>
On Wed, Apr 08, 2015 at 03:06:13PM -0400, Vladimir Medvedkin wrote:
> Software implementation of the Toeplitz hash function used by RSS.
> Can be used either for packet distribution on single queue NIC
> or for simulating of RSS computation on specific NIC (for example
> after GRE header decapsulating).
>
> Signed-off-by: Vladimir Medvedkin <medvedkinv@gmail.com>
> ---
> lib/librte_hash/Makefile | 1 +
> lib/librte_hash/rte_thash.h | 179 ++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 180 insertions(+)
> create mode 100644 lib/librte_hash/rte_thash.h
>
> diff --git a/lib/librte_hash/Makefile b/lib/librte_hash/Makefile
> index 3696cb1..083a9e5 100644
> --- a/lib/librte_hash/Makefile
> +++ b/lib/librte_hash/Makefile
> @@ -50,6 +50,7 @@ SYMLINK-$(CONFIG_RTE_LIBRTE_HASH)-include := rte_hash.h
> SYMLINK-$(CONFIG_RTE_LIBRTE_HASH)-include += rte_hash_crc.h
> SYMLINK-$(CONFIG_RTE_LIBRTE_HASH)-include += rte_jhash.h
> SYMLINK-$(CONFIG_RTE_LIBRTE_HASH)-include += rte_fbk_hash.h
> +SYMLINK-$(CONFIG_RTE_LIBRTE_HASH)-include += rte_thash.h
>
> # this lib needs eal
> DEPDIRS-$(CONFIG_RTE_LIBRTE_HASH) += lib/librte_eal lib/librte_malloc
> diff --git a/lib/librte_hash/rte_thash.h b/lib/librte_hash/rte_thash.h
> new file mode 100644
> index 0000000..1acfa3a
> --- /dev/null
> +++ b/lib/librte_hash/rte_thash.h
> @@ -0,0 +1,179 @@
> +/*-
> + * BSD LICENSE
> + *
> + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
> + * All rights reserved.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions
> + * are met:
> + *
> + * * Redistributions of source code must retain the above copyright
> + * notice, this list of conditions and the following disclaimer.
> + * * Redistributions in binary form must reproduce the above copyright
> + * notice, this list of conditions and the following disclaimer in
> + * the documentation and/or other materials provided with the
> + * distribution.
> + * * Neither the name of Intel Corporation nor the names of its
> + * contributors may be used to endorse or promote products derived
> + * from this software without specific prior written permission.
> + *
> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
> + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
> + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
> + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> + */
> +
> +#ifndef _RTE_THASH_H
> +#define _RTE_THASH_H
> +
> +/**
> + * @file
> + *
> + * toeplitz hash functions.
> + */
> +
> +#ifdef __cplusplus
> +extern "C" {
> +#endif
> +
> +/**
> + * Software implementation of the Toeplitz hash function used by RSS.
> + * Can be used either for packet distribution on single queue NIC
> + * or for simulating of RSS computation on specific NIC (for example
> + * after GRE header decapsulating)
> + */
> +
> +#include <stdint.h>
> +#include <rte_byteorder.h>
> +
> +enum rte_thash_flag {
> + RTE_THASH_L3 = 0, //calculate hash tacking into account only l3 header
> + RTE_THASH_L4 //calculate hash tacking into account l4 + l4 headers
> +};
> +
> +/**
> + * Prepare special converted key to use with rte_softrss_be()
> + * @param orig
> + * pointer to original RSS key
> + * @param targ
> + * pointer to target RSS key
> + */
> +
> +static inline void
> +rte_convert_rss_key(uint32_t *orig, uint32_t *targ)
> +{
> + int i;
> + for (i = 0; i < 10; i++) {
> + targ[i] = rte_be_to_cpu_32(orig[i]);
> + }
> +}
> +
> +/**
> + * Generic implementation. Can be used with original rss_key
> + * All ip's and ports have to be CPU byte order.
> + * @param sip
> + * Source ip address.
> + * @param dip
> + * Destination ip address.
ipv4, what about ipv6? Why not define rss function that works on byte
buffer and let caller build it according to whatever fields it want to
hash?
> + * @param sp
> + * Source TCP|UDP port.
> + * @param dp
> + * Destination TCP|UDP port.
> + * @param flag
> + * RTE_THASH_L3: calculate hash tacking into account only sip and dip
> + * RTE_THASH_L4: calculate hash tacking into account sip, dip, sp and dp
> + * @param *rss_key
> + * Pointer to 40-byte RSS hash key.
i40e has 52 byte RSS hash key.
> + * @return
> + * Calculated hash value.
> + */
> +
> +static inline uint32_t
> +rte_softrss(uint32_t sip, uint32_t dip, uint16_t sp, uint16_t dp, enum rte_thash_flag flag, uint32_t *rss_key)
> +{
> + uint32_t ret = 0;
> + int i;
> + for (i = 0; i < 32; i++) {
> + if (sip & (1 << (31 - i))) {
> + ret ^= (rte_cpu_to_be_32(*rss_key) << i)|(uint32_t)((uint64_t)(rte_cpu_to_be_32(*(rss_key + 1))) >> (32 - i));
> + }
> + }
> + rss_key++;
> + for (i = 0; i < 32; i++) {
> + if (dip & (1 << (31 - i))) {
> + ret ^= (rte_cpu_to_be_32(*rss_key) << i)|(uint32_t)((uint64_t)(rte_cpu_to_be_32(*(rss_key + 1))) >> (32 - i));
> + }
> + }
> + if (flag == RTE_THASH_L4) {
> + rss_key++;
> + for (i = 0; i < 32; i++) {
> + if (((sp<<16)|dp) & (1 << (31 - i))) {
> + ret ^= (rte_cpu_to_be_32(*rss_key) << i)|(uint32_t)((uint64_t)(rte_cpu_to_be_32(*(rss_key + 1))) >> (32 - i));
> + }
> + }
> + }
> + return ret;
> +}
> +
> +/**
> + * Optimized implementation.
> + * If you want the calculated hash value matches NIC RSS value
> + * you have to use special converted key.
> + * All ip's and ports have to be CPU byte order.
> + * @param sip
> + * Source ip address.
> + * @param dip
> + * Destination ip address.
> + * @param sp
> + * Source TCP|UDP port.
> + * @param dp
> + * Destination TCP|UDP port.
> + * @param flag
> + * RTE_THASH_L3: calculate hash tacking into account only sip and dip
> + * RTE_THASH_L4: calculate hash tacking into account sip, dip, sp and dp
> + * @param *rss_key
> + * Pointer to 40-byte RSS hash key.
> + * @return
> + * Calculated hash value.
> + */
> +
> +static inline uint32_t
> +rte_softrss_be(uint32_t sip, uint32_t dip, uint16_t sp, uint16_t dp, enum rte_thash_flag flag, uint32_t *rss_key)
> +{
> + uint32_t ret = 0;
> + int i;
> + for (i = 0; i < 32; i++) {
> + if (sip & (1 << (31 - i))) {
> + ret ^= (*rss_key << i)|(uint32_t)((uint64_t)*(rss_key + 1) >> (32 - i));
> + }
> + }
> + rss_key++;
> + for (i = 0; i < 32; i++) {
> + if (dip & (1 << (31 - i))) {
> + ret ^= (*rss_key << i)|(uint32_t)((uint64_t)*(rss_key + 1) >> (32 - i));
> + }
> + }
> + if (flag == RTE_THASH_L4) {
> + rss_key++;
> + for (i = 0; i < 32; i++) {
> + if (((sp<<16)|dp) & (1 << (31 - i))) {
> + ret ^= (*rss_key << i)|(uint32_t)((uint64_t)*(rss_key + 1) >> (32 - i));
> + }
> + }
> + }
> + return ret;
> +}
> +
> +#ifdef __cplusplus
> +}
> +#endif
> +
> +#endif /* _RTE_THASH_H */
> --
> 1.8.3.2
>
--
Gleb.
next prev parent reply other threads:[~2015-04-09 6:37 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-04-08 19:06 Vladimir Medvedkin
2015-04-08 22:24 ` Stephen Hemminger
2015-04-09 12:50 ` Vladimir Medvedkin
2015-04-09 6:37 ` Gleb Natapov [this message]
2015-04-09 12:57 ` Vladimir Medvedkin
2015-05-05 13:20 ` [dpdk-dev] [PATCH v2] Add toeplitz hash algorithm used by RSS Vladimir Medvedkin
2015-05-05 16:03 ` Chilikin, Andrey
2015-05-07 10:28 ` Vladimir Medvedkin
2015-05-07 11:38 ` Chilikin, Andrey
2015-05-08 14:58 ` Vladimir Medvedkin
2015-05-08 14:58 ` Vladimir Medvedkin
2015-06-03 14:07 ` Thomas Monjalon
2015-06-16 9:07 ` Thomas Monjalon
2015-06-16 10:36 ` Bruce Richardson
2015-06-16 12:29 ` Bruce Richardson
2015-06-16 19:26 ` Vladimir Medvedkin
2015-06-19 14:55 ` [dpdk-dev] [PATCH v4] " Vladimir Medvedkin
2015-06-19 15:59 ` Richardson, Bruce
2015-06-19 16:14 ` Vladimir Medvedkin
2015-06-29 12:18 ` Bruce Richardson
2015-06-19 16:33 ` Mcnamara, John
2015-06-19 17:31 ` [dpdk-dev] [PATCH v5] " Vladimir Medvedkin
2015-06-29 12:40 ` Bruce Richardson
2015-06-30 12:14 ` Vladimir Medvedkin
2015-06-30 12:48 ` Bruce Richardson
2015-06-30 23:40 ` [dpdk-dev] [PATCH v6] " Vladimir Medvedkin
2015-07-01 15:29 ` Bruce Richardson
2015-07-01 21:23 ` Thomas Monjalon
2015-07-22 7:55 ` Tony Lu
2015-07-27 11:57 ` Vladimir Medvedkin
2015-07-29 5:01 ` Qiu, Michael
2015-07-29 14:00 ` Vladimir Medvedkin
2015-06-19 17:32 ` [dpdk-dev] [PATCH v3] Add unit test for thash library Vladimir Medvedkin
2015-06-29 12:42 ` Bruce Richardson
2015-06-30 23:41 ` [dpdk-dev] [PATCH v4] " Vladimir Medvedkin
2015-07-01 15:30 ` Bruce Richardson
2015-07-01 21:28 ` Thomas Monjalon
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=20150409063742.GA16818@cloudius-systems.com \
--to=gleb@cloudius-systems.com \
--cc=dev@dpdk.org \
--cc=medvedkinv@gmail.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).