From: Matan Azrad <matan@mellanox.com>
To: Neil Horman <nhorman@tuxdriver.com>,
Thomas Monjalon <thomas@monjalon.net>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
"Bruce Richardson" <bruce.richardson@intel.com>,
"Ananyev, Konstantin" <konstantin.ananyev@intel.com>,
"Gaëtan Rivet" <gaetan.rivet@6wind.com>,
"Wu, Jingjing" <jingjing.wu@intel.com>
Subject: Re: [dpdk-dev] [PATCH 2/5] ethdev: add port ownership
Date: Thu, 21 Dec 2017 19:37:06 +0000 [thread overview]
Message-ID: <HE1PR0502MB3659B8A71663742533E45148D20D0@HE1PR0502MB3659.eurprd05.prod.outlook.com> (raw)
In-Reply-To: <20171221174304.GB23958@hmswarspite.think-freely.org>
Hi
> -----Original Message-----
> From: Neil Horman [mailto:nhorman@tuxdriver.com]
> Sent: Thursday, December 21, 2017 7:43 PM
> To: Thomas Monjalon <thomas@monjalon.net>
> Cc: dev@dpdk.org; Bruce Richardson <bruce.richardson@intel.com>; Matan
> Azrad <matan@mellanox.com>; Ananyev, Konstantin
> <konstantin.ananyev@intel.com>; Gaëtan Rivet <gaetan.rivet@6wind.com>;
> Wu, Jingjing <jingjing.wu@intel.com>
> Subject: Re: [dpdk-dev] [PATCH 2/5] ethdev: add port ownership
>
> On Thu, Dec 21, 2017 at 06:06:48PM +0100, Thomas Monjalon wrote:
> > 08/12/2017 13:31, Neil Horman:
> > > On Fri, Dec 08, 2017 at 12:35:18PM +0100, Thomas Monjalon wrote:
> > > > 05/12/2017 11:05, Bruce Richardson:
> > > > > > I think you suggest to make all the ethdev configuration race
> > > > > > safe, it is behind to this thread. Current ethdev
> > > > > > implementation leave the race management to applications, so
> > > > > > port ownership as any other port configurations should be designed
> in the same method.
> > > > >
> > > > > One key difference, though, being that port ownership itself
> > > > > could be used to manage the thread-safety of the ethdev
> > > > > configuration. It's also a little different from other APIs in
> > > > > that I find it hard to come up with a scenario where it would be
> > > > > very useful to an application without also having some form of
> > > > > locking present in it. For other config/control APIs we can have
> > > > > the control plane APIs work without locks e.g. by having a
> > > > > single designated thread/process manage all configuration
> > > > > updates. However, as Neil points out, in such a scenario, the
> > > > > ownership concept doesn't provide any additional benefit so can
> > > > > be skipped completely. I'd view it a bit like the reference
> > > > > counting of mbufs - we can provide a lockless/non-atomic version,
> but for just about every real use-case, you want the atomic version.
> > > >
> > > > I think we need to clearly describe what is the tread-safety
> > > > policy in DPDK (especially in ethdev as a first example).
> > > > Let's start with obvious things:
> > > >
> > > > 1/ A queue is not protected for races with multiple Rx or Tx
> > > > - no planned change because of performance
> purpose
> > > > 2/ The list of devices is racy
> > > > - to be fixed with atomics
> > > > 3/ The configuration of different devices is thread-safe
> > > > - the configurations are different per-device
> > > > 4/ The configuration of a given device is racy
> > > > - can be managed by the owner of the device
> > > > 5/ The device ownership is racy
> > > > - to be fixed with atomics
> > > >
> > > > What am I missing?
> > > >
Thank you Thomas for this order.
Actually the port ownership is a good opportunity to redefine the synchronization rules in ethdev :)
> > > There is fan out to consider here:
> > >
> > > 1) Is device configuration racy with ownership? That is to say, can
> > > I change ownership of a device safely while another thread that
> > > currently owns it modifies its configuration?
> >
> > If an entity steals ownership to another one, either it is agreed
> > earlier, or it is done by a central authority.
> > When it is acked that ownership can be moved, there should not be any
> > configuration in progress.
> > So it is more a communication issue than a race.
> >
> But if thats the case (specifically that mutual exclusion between port
> ownership and configuration is an exercize left to an application developer),
> then port ownership itself is largely meaningless within the dpdk, because
> the notion of who owns the port needs to be codified within the application
> anyway.
>
Bruce, As I understand it, only the dpdk entity who took ownership of a port successfully can configure the device by default, if other dpdk entities want to configure it too they must to be synchronized with the port owner while it is not recommended after the port ownership integration.
So, for example, if the dpdk entity is an application, the application should take ownership of the port and manage the synchronization of this port configuration between the application threads and its EAL host thread callbacks, no other dpdk entity should configure the same port because they should fail when they try to take ownership of the same port too.
Each dpdk entity which wants to take ownership must to be able to synchronize the port configuration in its level.
>
> > > 2) Is device configuration racy with device addition/removal? That
> > > is to say, can one thread remove a device while another configures it.
> >
> > I think it is the same as two threads configuring the same device
> > (item 4/ above). It can be managed with port ownership.
> >
> Only if you assert that application is required to have the owning port be
> responsible for the ports deletion, which we can say, but that leads to the
> issue above again.
>
>
As Thomas said in item 2 the port creation must be synchronized by ethdev and we need to add it there.
I think it is obvious that port removal must to be done only by the port owner.
I think we need to add synchronization for port ownership management in this patch V2 and add port creation synchronization in ethdev in separate patch to fill the new rules Thomas suggested.
What do you think?
> > > There are probably many subsystem interactions that need to be
> addressed here.
> > >
> > > Neil
> > >
> > > > I am also wondering whether the device ownership can be a separate
> > > > library used in several device class interfaces?
> >
> >
> >
next prev parent reply other threads:[~2017-12-21 19:37 UTC|newest]
Thread overview: 214+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-28 11:57 [dpdk-dev] [PATCH 0/5] ethdev: Port ownership Matan Azrad
2017-11-28 11:57 ` [dpdk-dev] [PATCH 1/5] ethdev: free a port by a dedicated API Matan Azrad
2017-11-28 11:57 ` [dpdk-dev] [PATCH 2/5] ethdev: add port ownership Matan Azrad
2017-11-30 12:36 ` Neil Horman
2017-11-30 13:24 ` Gaëtan Rivet
2017-11-30 14:30 ` Matan Azrad
2017-11-30 15:09 ` Gaëtan Rivet
2017-11-30 15:43 ` Matan Azrad
2017-12-01 12:09 ` Neil Horman
2017-12-03 8:04 ` Matan Azrad
2017-12-03 11:10 ` Ananyev, Konstantin
2017-12-03 13:46 ` Matan Azrad
2017-12-04 16:01 ` Neil Horman
2017-12-04 18:10 ` Matan Azrad
2017-12-04 22:30 ` Neil Horman
2017-12-05 6:08 ` Matan Azrad
2017-12-05 10:05 ` Bruce Richardson
2017-12-08 11:35 ` Thomas Monjalon
2017-12-08 12:31 ` Neil Horman
2017-12-21 17:06 ` Thomas Monjalon
2017-12-21 17:43 ` Neil Horman
2017-12-21 19:37 ` Matan Azrad [this message]
2017-12-21 20:14 ` Neil Horman
2017-12-21 21:57 ` Matan Azrad
2017-12-22 14:26 ` Neil Horman
2017-12-23 22:36 ` Matan Azrad
2017-12-29 16:56 ` Neil Horman
2017-12-05 19:26 ` Neil Horman
2017-12-08 11:06 ` Thomas Monjalon
2017-12-05 11:12 ` Ananyev, Konstantin
2017-12-05 11:44 ` Ananyev, Konstantin
2017-12-05 11:53 ` Thomas Monjalon
2017-12-05 14:56 ` Bruce Richardson
2017-12-05 14:57 ` Ananyev, Konstantin
2017-12-05 11:47 ` Thomas Monjalon
2017-12-05 15:13 ` Ananyev, Konstantin
2017-12-05 15:49 ` Thomas Monjalon
2017-11-28 11:57 ` [dpdk-dev] [PATCH 3/5] net/failsafe: free an eth port by a dedicated API Matan Azrad
2017-11-28 11:58 ` [dpdk-dev] [PATCH 4/5] net/failsafe: use ownership mechanism to own ports Matan Azrad
2017-11-28 11:58 ` [dpdk-dev] [PATCH 5/5] app/testpmd: adjust ethdev port ownership Matan Azrad
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 0/6] ethdev: " Matan Azrad
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 1/6] ethdev: fix port data reset timing Matan Azrad
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 2/6] ethdev: add port ownership Matan Azrad
2018-01-10 13:36 ` Ananyev, Konstantin
2018-01-10 16:58 ` Matan Azrad
2018-01-11 12:40 ` Ananyev, Konstantin
2018-01-11 14:51 ` Matan Azrad
2018-01-12 0:02 ` Ananyev, Konstantin
2018-01-12 7:24 ` Matan Azrad
2018-01-15 11:45 ` Ananyev, Konstantin
2018-01-15 13:09 ` Matan Azrad
2018-01-15 18:43 ` Ananyev, Konstantin
2018-01-16 8:04 ` Matan Azrad
2018-01-16 19:11 ` Ananyev, Konstantin
2018-01-16 20:32 ` Matan Azrad
2018-01-17 11:24 ` Ananyev, Konstantin
2018-01-17 12:05 ` Matan Azrad
2018-01-17 12:54 ` Ananyev, Konstantin
2018-01-17 13:10 ` Matan Azrad
2018-01-17 16:52 ` Ananyev, Konstantin
2018-01-17 18:02 ` Matan Azrad
2018-01-17 20:34 ` Matan Azrad
2018-01-18 14:17 ` Ananyev, Konstantin
2018-01-18 14:26 ` Matan Azrad
2018-01-18 14:41 ` Ananyev, Konstantin
2018-01-18 14:45 ` Matan Azrad
2018-01-18 14:51 ` Ananyev, Konstantin
2018-01-18 15:00 ` Matan Azrad
2018-01-17 14:00 ` Neil Horman
2018-01-17 17:01 ` Ananyev, Konstantin
2018-01-18 13:10 ` Neil Horman
2018-01-18 14:00 ` Matan Azrad
2018-01-18 16:54 ` Neil Horman
2018-01-18 17:20 ` Matan Azrad
2018-01-18 18:41 ` Neil Horman
2018-01-18 20:21 ` Matan Azrad
2018-01-19 1:41 ` Neil Horman
2018-01-19 7:14 ` Matan Azrad
2018-01-19 9:30 ` Bruce Richardson
2018-01-19 10:44 ` Matan Azrad
2018-01-19 13:30 ` Neil Horman
2018-01-19 13:57 ` Matan Azrad
2018-01-19 14:13 ` Thomas Monjalon
2018-01-19 15:27 ` Neil Horman
2018-01-19 17:17 ` Thomas Monjalon
2018-01-19 17:43 ` Neil Horman
2018-01-19 18:12 ` Thomas Monjalon
2018-01-19 19:47 ` Neil Horman
2018-01-19 20:19 ` Thomas Monjalon
2018-01-19 22:52 ` Neil Horman
2018-01-20 3:38 ` Tuxdriver
2018-01-20 12:54 ` Ananyev, Konstantin
2018-01-20 14:02 ` Thomas Monjalon
2018-01-19 12:55 ` Neil Horman
2018-01-19 13:52 ` Neil Horman
2018-01-18 16:27 ` Neil Horman
2018-01-17 17:58 ` Matan Azrad
2018-01-18 13:20 ` Neil Horman
2018-01-18 14:52 ` Matan Azrad
2018-01-19 13:57 ` Neil Horman
2018-01-19 14:07 ` Thomas Monjalon
2018-01-19 14:32 ` Neil Horman
2018-01-19 17:09 ` Thomas Monjalon
2018-01-19 17:37 ` Neil Horman
2018-01-19 18:10 ` Thomas Monjalon
2018-01-21 22:12 ` Ferruh Yigit
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 3/6] ethdev: synchronize port allocation Matan Azrad
2018-01-07 9:58 ` Matan Azrad
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 4/6] net/failsafe: free an eth port by a dedicated API Matan Azrad
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 5/6] net/failsafe: use ownership mechanism to own ports Matan Azrad
2018-01-08 10:32 ` Gaëtan Rivet
2018-01-08 11:16 ` Matan Azrad
2018-01-08 11:35 ` Gaëtan Rivet
2018-01-07 9:45 ` [dpdk-dev] [PATCH v2 6/6] app/testpmd: adjust ethdev port ownership Matan Azrad
2018-01-08 11:39 ` Gaëtan Rivet
2018-01-08 12:30 ` Matan Azrad
2018-01-08 13:30 ` Gaëtan Rivet
2018-01-08 13:55 ` Matan Azrad
2018-01-08 14:21 ` Gaëtan Rivet
2018-01-08 14:42 ` Matan Azrad
2018-01-16 5:53 ` Lu, Wenzhuo
2018-01-16 8:15 ` Matan Azrad
2018-01-17 0:46 ` Lu, Wenzhuo
2018-01-17 8:51 ` Matan Azrad
2018-01-18 0:53 ` Lu, Wenzhuo
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 0/7] Port ownership and syncronization Matan Azrad
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 1/7] ethdev: fix port data reset timing Matan Azrad
2018-01-18 17:00 ` Thomas Monjalon
2018-01-19 12:38 ` Ananyev, Konstantin
2018-03-05 11:24 ` [dpdk-dev] [dpdk-stable] " Ferruh Yigit
2018-03-05 14:52 ` Matan Azrad
2018-03-05 15:06 ` Ferruh Yigit
2018-03-05 15:12 ` Matan Azrad
2018-03-27 22:37 ` Ferruh Yigit
2018-03-28 12:07 ` Matan Azrad
2018-03-30 10:39 ` Ferruh Yigit
2018-04-19 11:07 ` Ferruh Yigit
2018-04-25 12:16 ` Matan Azrad
2018-04-25 12:30 ` Ori Kam
2018-04-25 12:54 ` Ferruh Yigit
2018-04-25 14:01 ` Matan Azrad
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 2/7] ethdev: fix used portid allocation Matan Azrad
2018-01-18 17:00 ` Thomas Monjalon
2018-01-19 12:40 ` Ananyev, Konstantin
2018-01-20 16:48 ` Matan Azrad
2018-01-20 17:26 ` Ananyev, Konstantin
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 3/7] ethdev: add port ownership Matan Azrad
2018-01-18 21:11 ` Thomas Monjalon
2018-01-19 12:41 ` Ananyev, Konstantin
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 4/7] ethdev: synchronize port allocation Matan Azrad
2018-01-18 20:43 ` Thomas Monjalon
2018-01-18 20:52 ` Matan Azrad
2018-01-18 21:17 ` Thomas Monjalon
2018-01-19 12:47 ` Ananyev, Konstantin
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 5/7] net/failsafe: free an eth port by a dedicated API Matan Azrad
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 6/7] net/failsafe: use ownership mechanism to own ports Matan Azrad
2018-01-18 16:35 ` [dpdk-dev] [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership Matan Azrad
2018-01-19 12:37 ` Ananyev, Konstantin
2018-01-19 12:51 ` Matan Azrad
2018-01-19 13:08 ` Ananyev, Konstantin
2018-01-19 13:35 ` Matan Azrad
2018-01-19 15:00 ` Gaëtan Rivet
2018-01-20 18:14 ` Matan Azrad
2018-01-22 10:17 ` Gaëtan Rivet
2018-01-22 11:22 ` Matan Azrad
2018-01-22 12:28 ` Ananyev, Konstantin
2018-01-22 13:22 ` Matan Azrad
2018-01-22 20:48 ` Ananyev, Konstantin
2018-01-23 8:54 ` Matan Azrad
2018-01-23 12:56 ` Gaëtan Rivet
2018-01-23 14:30 ` Matan Azrad
2018-01-25 9:36 ` Matan Azrad
2018-01-25 10:05 ` Thomas Monjalon
2018-01-25 11:15 ` Ananyev, Konstantin
2018-01-25 11:33 ` Thomas Monjalon
2018-01-25 11:55 ` Ananyev, Konstantin
2018-01-23 13:34 ` Ananyev, Konstantin
2018-01-23 14:18 ` Thomas Monjalon
2018-01-23 15:12 ` Ananyev, Konstantin
2018-01-23 15:18 ` Ananyev, Konstantin
2018-01-23 17:33 ` Thomas Monjalon
2018-01-23 21:18 ` Ananyev, Konstantin
2018-01-24 8:10 ` Thomas Monjalon
2018-01-24 18:30 ` Ananyev, Konstantin
2018-01-25 10:55 ` Thomas Monjalon
2018-01-25 11:09 ` Ananyev, Konstantin
2018-01-25 11:27 ` Thomas Monjalon
2018-01-23 14:43 ` Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 0/7] Port ownership and syncronization Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 1/7] ethdev: fix port data reset timing Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 2/7] ethdev: fix used portid allocation Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 3/7] ethdev: add port ownership Matan Azrad
2018-01-21 20:43 ` Ferruh Yigit
2018-01-21 20:46 ` Ferruh Yigit
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 4/7] ethdev: synchronize port allocation Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 5/7] net/failsafe: free an eth port by a dedicated API Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 6/7] net/failsafe: use ownership mechanism to own ports Matan Azrad
2018-01-20 21:24 ` [dpdk-dev] [PATCH v4 7/7] app/testpmd: adjust ethdev port ownership Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 0/7] Port ownership and synchronization Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 1/7] ethdev: fix port data reset timing Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 2/7] ethdev: fix used portid allocation Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 3/7] ethdev: add port ownership Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 4/7] ethdev: synchronize port allocation Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 5/7] net/failsafe: free an eth port by a dedicated API Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 6/7] net/failsafe: use ownership mechanism to own ports Matan Azrad
2018-01-22 16:38 ` [dpdk-dev] [PATCH v5 7/7] app/testpmd: adjust ethdev port ownership Matan Azrad
2018-01-25 1:47 ` Lu, Wenzhuo
2018-01-25 8:30 ` Matan Azrad
2018-01-26 0:50 ` Lu, Wenzhuo
2018-01-29 11:21 ` [dpdk-dev] [PATCH v5 0/7] Port ownership and synchronization Matan Azrad
2018-01-31 19:53 ` Thomas Monjalon
2018-01-25 14:35 ` [dpdk-dev] [PATCH v3 0/7] Port ownership and syncronization Ferruh Yigit
2017-12-06 0:40 [dpdk-dev] [PATCH 2/5] ethdev: add port ownership Ananyev, Konstantin
2017-12-06 9:22 ` 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=HE1PR0502MB3659B8A71663742533E45148D20D0@HE1PR0502MB3659.eurprd05.prod.outlook.com \
--to=matan@mellanox.com \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=gaetan.rivet@6wind.com \
--cc=jingjing.wu@intel.com \
--cc=konstantin.ananyev@intel.com \
--cc=nhorman@tuxdriver.com \
--cc=thomas@monjalon.net \
/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).