DPDK patches and discussions
 help / color / mirror / Atom feed
From: Declan Doherty <declan.doherty@intel.com>
To: Eric Kinzie <ehkinzie@gmail.com>,
	Andriy Berestovskyy <aber@semihalf.com>
Cc: dev@dpdk.org, Eric Kinzie <ekinzie@brocade.com>
Subject: Re: [dpdk-dev] [PATCH 6/8] bond: handle slaves with fewer queues than bonding device
Date: Tue, 5 Jan 2016 13:46:05 +0000	[thread overview]
Message-ID: <568BC91D.8090806@intel.com> (raw)
In-Reply-To: <20151204191831.GA20647@roosta.home>

On 04/12/15 19:18, Eric Kinzie wrote:
> On Fri Dec 04 19:36:09 +0100 2015, Andriy Berestovskyy wrote:
>> Hi guys,
>> I'm not quite sure if we can support less TX queues on a slave that easy:
>>
>>> queue_id = bond_slave_txqid(internals, i, bd_tx_q->queue_id);
>>> num_tx_slave = rte_eth_tx_burst(slaves[i], queue_id,
>>>       slave_bufs[i], slave_nb_pkts[i]);
>>
>> It seems that two different lcores might end up writing to the same
>> slave queue at the same time, isn't it?
>>
>> Regards,
>> Andriy
>
> Andriy, I think you're probably right about this.  Perhaps it should
> instead refuse to add or refuse to activate a slave with too few
> tx queues.  Could probably fix this with another layer of buffering
> so that an lcore with a valid tx queue could pick up the mbufs later,
> but this doesn't seem very appealing.
>
> Eric
>
>
>> On Fri, Dec 4, 2015 at 6:14 PM, Stephen Hemminger
>> <stephen@networkplumber.org> wrote:
>>> From: Eric Kinzie <ekinzie@brocade.com>
>>>
>>> In the event that the bonding device has a greater number of tx and/or rx
>>> queues than the slave being added, track the queue limits of the slave.
>>> On receive, ignore queue identifiers beyond what the slave interface
>>> can support.  During transmit, pick a different queue id to use if the
>>> intended queue is not available on the slave.
>>>
>>> Signed-off-by: Eric Kinzie <ekinzie@brocade.com>
>>> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
>>> ---
...


I don't there is any straight forward way of supporting slaves with 
different numbers of queues, the initial library was written with the 
assumption that the number of tx/rx queues would always be the same on 
each slave. This is why,when a slave is added to a bonded device we 
reconfigure the queues. For features like RSS we have to have the same 
number of rx queues otherwise the flow distribution to an application 
could change in the case of a fail over event. Also by supporting 
different numbers of queues between slaves we would be no longer be 
supporting the standard behavior of ethdevs in DPDK were we expect that 
by using different queues we don't require locking to be thread safe.

  reply	other threads:[~2016-01-05 13:47 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-04 17:14 [dpdk-dev] [PATCH 0/8] bonding: fixes and enhancements Stephen Hemminger
2015-12-04 17:14 ` [dpdk-dev] [PATCH 1/8] bond: use existing enslaved device queues Stephen Hemminger
2016-01-05 13:32   ` Declan Doherty
2015-12-04 17:14 ` [dpdk-dev] [PATCH 2/8] bond mode 4: copy entire config structure Stephen Hemminger
2016-01-05 13:32   ` Declan Doherty
2015-12-04 17:14 ` [dpdk-dev] [PATCH 3/8] bond mode 4: do not ignore multicast Stephen Hemminger
2016-01-05 13:32   ` Declan Doherty
2015-12-04 17:14 ` [dpdk-dev] [PATCH 4/8] bond mode 4: allow external state machine Stephen Hemminger
2016-01-05 13:33   ` Declan Doherty
2015-12-04 17:14 ` [dpdk-dev] [PATCH 5/8] bond: active slaves with no primary Stephen Hemminger
2016-01-05 13:34   ` Declan Doherty
2015-12-04 17:14 ` [dpdk-dev] [PATCH 6/8] bond: handle slaves with fewer queues than bonding device Stephen Hemminger
2015-12-04 18:36   ` Andriy Berestovskyy
2015-12-04 19:18     ` Eric Kinzie
2016-01-05 13:46       ` Declan Doherty [this message]
2016-01-05 15:31         ` Stephen Hemminger
2016-02-03 11:28       ` Bruce Richardson
2016-02-03 15:17         ` Declan Doherty
2016-02-03 15:21           ` Thomas Monjalon
2016-02-18 10:26             ` Iremonger, Bernard
2016-02-19 19:17               ` [dpdk-dev] [PATCH v2 0/6] bonding: fixes and enhancements Eric Kinzie
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 1/6] bond: use existing enslaved device queues Eric Kinzie
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 2/6] bond mode 4: copy entire config structure Eric Kinzie
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 3/6] bond mode 4: do not ignore multicast Eric Kinzie
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 4/6] bond mode 4: allow external state machine Eric Kinzie
2016-02-22 13:03                   ` Panu Matilainen
2016-02-25 15:22                     ` Iremonger, Bernard
2016-03-01 17:31                       ` [dpdk-dev] [PATCH V3 0/4] bonding: fixes and enhancements Eric Kinzie
2016-03-01 17:31                         ` [dpdk-dev] [PATCH V3 1/4] bond mode 4: copy entire config structure Eric Kinzie
2016-03-01 17:32                         ` [dpdk-dev] [PATCH V3 2/4] bond mode 4: do not ignore multicast Eric Kinzie
2016-03-01 17:32                         ` [dpdk-dev] [PATCH V3 3/4] bond: active slaves with no primary Eric Kinzie
2016-03-01 17:32                         ` [dpdk-dev] [PATCH V3 4/4] bond: do not activate slave twice Eric Kinzie
2016-03-10 15:41                         ` [dpdk-dev] [PATCH V3 0/4] bonding: fixes and enhancements Bruce Richardson
2016-03-01 17:40                       ` [dpdk-dev] [PATCH v2 4/6] bond mode 4: allow external state machine Eric Kinzie
2016-03-02  9:49                         ` Iremonger, Bernard
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 5/6] bond: active slaves with no primary Eric Kinzie
2016-02-19 19:17                 ` [dpdk-dev] [PATCH v2 6/6] bond: do not activate slave twice Eric Kinzie
2015-12-04 17:14 ` [dpdk-dev] [PATCH 7/8] bond: per-slave intermediate rx ring Stephen Hemminger
2015-12-04 17:14 ` [dpdk-dev] [PATCH 8/8] bond: do not activate slave twice Stephen Hemminger
2016-01-05 13:47   ` Declan Doherty
2015-12-23 10:51 ` [dpdk-dev] [PATCH 0/8] bonding: fixes and enhancements Iremonger, Bernard

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=568BC91D.8090806@intel.com \
    --to=declan.doherty@intel.com \
    --cc=aber@semihalf.com \
    --cc=dev@dpdk.org \
    --cc=ehkinzie@gmail.com \
    --cc=ekinzie@brocade.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).