DPDK patches and discussions
 help / color / mirror / Atom feed
From: Vladimir Medvedkin <medvedkinv@gmail.com>
To: Stephen Hemminger <stephen@networkplumber.org>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH] Add toeplitz hash algorithm
Date: Thu, 9 Apr 2015 15:50:23 +0300	[thread overview]
Message-ID: <CANDrEHk+8YmRnMewBE7i23_yuAEKb9vT+GSPpBpLEqO1rqcZZw@mail.gmail.com> (raw)
In-Reply-To: <20150408152430.0bdedfd6@urahara>

Hi Stephen,



2015-04-09 1:24 GMT+03:00 Stephen Hemminger <stephen@networkplumber.org>:

> On Wed,  8 Apr 2015 15:06:13 -0400
> Vladimir Medvedkin <medvedkinv@gmail.com> 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>
>
> > +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)
> orig should be const
>
> > +{
> > +     int i;
> > +     for (i = 0; i < 10; i++) {
> > +             targ[i] = rte_be_to_cpu_32(orig[i]);
> > +     }
> > +}
>
> > +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)
>
> rss_key should be const
>
> > +{
> > +     uint32_t ret = 0;
> > +     int i;
> > +     for (i = 0; i < 32; i++) {
> blank line after declaration please
>
> > +             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));
>
> Long expression > 80 characters.
> Repeated multiple times (should be inline)
> Extra parens ()
>
Thanks for remarks, I'll fix it.

> Extension to 64 bits is only to avoid compiler warning?
>
No, in case when i = 0 we shift uint32_t left by 32 bits, which leads to
undefined behaviour. In fact, shift counter just masked to 5 bits, so count
range is limited to 0 to 31.

>
>
> > +             }
> > +     }
> > +     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)
> > +{
>
> Same problems as previous code.
> Also lots of copy paste (see Do Not Repeat Yourself principle).
>

  reply	other threads:[~2015-04-09 12:50 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 [this message]
2015-04-09  6:37 ` Gleb Natapov
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=CANDrEHk+8YmRnMewBE7i23_yuAEKb9vT+GSPpBpLEqO1rqcZZw@mail.gmail.com \
    --to=medvedkinv@gmail.com \
    --cc=dev@dpdk.org \
    --cc=stephen@networkplumber.org \
    /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).