DPDK patches and discussions
 help / color / mirror / Atom feed
From: Thomas Monjalon <thomas@monjalon.net>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: dev@dpdk.org, arybchenko@solarflare.com,
	Wenzhuo Lu <wenzhuo.lu@intel.com>,
	Beilei Xing <beilei.xing@intel.com>,
	Bernard Iremonger <bernard.iremonger@intel.com>
Subject: Re: [dpdk-dev] [PATCH 19/20] app/testpmd: reset port status on close notification
Date: Thu, 24 Sep 2020 14:17:04 +0200	[thread overview]
Message-ID: <5482678.jQLLJtJmLa@thomas> (raw)
In-Reply-To: <7f70403e-79f7-b3de-6a7b-a83571251810@intel.com>

24/09/2020 14:07, Ferruh Yigit:
> On 9/23/2020 9:32 PM, Thomas Monjalon wrote:
> > 23/09/2020 18:45, Ferruh Yigit:
> >> On 9/13/2020 11:07 PM, Thomas Monjalon wrote:
> >>> Since rte_eth_dev_release_port() is called on all port close operations,
> >>> the event RTE_ETH_EVENT_DESTROY can be reliably used for resetting
> >>> the port status on the application side.
> >>>
> >>> The intermediate state RTE_PORT_HANDLING is removed in close_port()
> >>> because a port can also be closed by a PMD in a device remove operation.
> >>>
> >>> In case multiple ports are closed, calling remove_invalid_ports()
> >>> only once is enough.
> >>>
> >>> Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
> >>
> >> <...>
> >>
> >>> @@ -3118,6 +3093,13 @@ eth_event_callback(portid_t port_id, enum rte_eth_event_type type, void *param,
> >>>    				rmv_port_callback, (void *)(intptr_t)port_id))
> >>>    			fprintf(stderr, "Could not set up deferred device removal\n");
> >>>    		break;
> >>> +	case RTE_ETH_EVENT_DESTROY:
> >>> +		if (rte_atomic16_cmpset(&(ports[port_id].port_status),
> >>> +					RTE_PORT_STOPPED,
> >>> +					RTE_PORT_CLOSED) == 0)
> >>> +			printf("Port %d cannot be set to closed\n", port_id);
> >>> +		printf("Port %u is closed\n", port_id);
> >>> +		break;
> >>
> >> This is failing if a port closed without application port stop command,
> >> PMD may be doing port stop within the close function but since
> >> application didn't give the stop command, the port status is not
> >> 'RTE_PORT_STOPPED', hence 'port_status' is not updated correctly.
> > 
> > Do you think we should give up with the atomic state transition,
> > and just assign the state as closed?
> > 
> 
> It can be better, if the DESTROY event received it should be in closed 
> state.

I don't understand your proposal.
Note that we are not managing ethdev states here.
Despite the misleading RTE prefix, it is a testpmd state.

> Right now device can be hot removed while running, or it may be closed 
> before it has been stopped, both cases state transition assumption from 
> stopped->closed will be wrong, and final state will be wrong.

I think we should force the state, not matter what it was before.



  reply	other threads:[~2020-09-24 12:25 UTC|newest]

Thread overview: 201+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-13 22:06 [dpdk-dev] [PATCH 00/20] cleanup ethdev close operation Thomas Monjalon
2020-09-13 22:06 ` [dpdk-dev] [PATCH 01/20] ethdev: reset device and interrupt pointers on release Thomas Monjalon
2020-09-23 16:41   ` Ferruh Yigit
2020-09-13 22:06 ` [dpdk-dev] [PATCH 02/20] ethdev: allow drivers to return error on close Thomas Monjalon
2020-09-23 16:41   ` Ferruh Yigit
2020-09-23 20:53     ` Thomas Monjalon
2020-09-23 21:02       ` Stephen Hemminger
2020-09-23 21:06         ` Thomas Monjalon
2020-09-23 21:47           ` Stephen Hemminger
2020-09-23 21:52             ` Thomas Monjalon
2020-09-23 22:02               ` Stephen Hemminger
2020-09-23 22:35                 ` Thomas Monjalon
2020-09-24  9:12                   ` Bruce Richardson
2020-09-24 10:07                     ` Thomas Monjalon
2020-09-24 12:09                       ` Ferruh Yigit
2020-09-24 14:48                     ` Stephen Hemminger
2020-09-13 22:06 ` [dpdk-dev] [PATCH 03/20] net/af_packet: release port upon close Thomas Monjalon
2020-09-23 16:41   ` Ferruh Yigit
2020-09-13 22:06 ` [dpdk-dev] [PATCH 04/20] net/atlantic: " Thomas Monjalon
2020-09-16 15:14   ` Igor Russkikh
2020-09-23 16:42   ` Ferruh Yigit
2020-09-23 20:50     ` Thomas Monjalon
2020-09-13 22:06 ` [dpdk-dev] [PATCH 05/20] net/axgbe: " Thomas Monjalon
2020-09-13 22:06 ` [dpdk-dev] [PATCH 06/20] net/bonding: " Thomas Monjalon
2020-09-13 22:06 ` [dpdk-dev] [PATCH 07/20] net/failsafe: " Thomas Monjalon
2020-09-23 21:24   ` Thomas Monjalon
2020-09-13 22:06 ` [dpdk-dev] [PATCH 08/20] net/iavf: " Thomas Monjalon
2020-09-13 22:07 ` [dpdk-dev] [PATCH 09/20] net/mlx4: " Thomas Monjalon
2020-09-13 22:07 ` [dpdk-dev] [PATCH 10/20] net/null: " Thomas Monjalon
2020-09-23 16:44   ` Ferruh Yigit
2020-09-23 20:47     ` Thomas Monjalon
2020-09-24 21:58       ` Thomas Monjalon
2020-09-25  8:52         ` Ferruh Yigit
2020-09-25 13:13           ` Thomas Monjalon
2020-09-13 22:07 ` [dpdk-dev] [PATCH 11/20] net/octeontx: " Thomas Monjalon
2020-09-13 22:07 ` [dpdk-dev] [PATCH 12/20] net/pcap: " Thomas Monjalon
2020-09-23 16:44   ` Ferruh Yigit
2020-09-23 20:44     ` Thomas Monjalon
2020-09-24 11:56       ` Ferruh Yigit
2020-09-13 22:07 ` [dpdk-dev] [PATCH 13/20] net/ring: " Thomas Monjalon
2020-09-14  8:51   ` Bruce Richardson
2020-09-13 22:07 ` [dpdk-dev] [PATCH 14/20] net/softnic: " Thomas Monjalon
2020-09-14 15:21   ` Dumitrescu, Cristian
2020-09-13 22:07 ` [dpdk-dev] [PATCH 15/20] net/tap: " Thomas Monjalon
2020-09-13 22:07 ` [dpdk-dev] [PATCH 16/20] ethdev: remove old close behaviour Thomas Monjalon
2020-09-23 16:44   ` Ferruh Yigit
2020-09-23 20:41     ` Thomas Monjalon
2020-09-24 12:00       ` Ferruh Yigit
2020-09-25  4:31     ` Rasesh Mody
2020-09-13 22:07 ` [dpdk-dev] [PATCH 17/20] drivers/net: accept removing device without any port Thomas Monjalon
2020-09-23 16:45   ` Ferruh Yigit
2020-09-27  6:25   ` Xu, Rosen
2020-09-13 22:07 ` [dpdk-dev] [PATCH 18/20] drivers/net: remove redundant MAC addresses freeing Thomas Monjalon
2020-09-23 16:45   ` Ferruh Yigit
2020-09-13 22:07 ` [dpdk-dev] [PATCH 19/20] app/testpmd: reset port status on close notification Thomas Monjalon
2020-09-23 16:45   ` Ferruh Yigit
2020-09-23 20:32     ` Thomas Monjalon
2020-09-24 12:07       ` Ferruh Yigit
2020-09-24 12:17         ` Thomas Monjalon [this message]
2020-09-24 13:06           ` Ferruh Yigit
2020-09-13 22:07 ` [dpdk-dev] [PATCH 20/20] app/testpmd: align behaviour of multi-port detach Thomas Monjalon
2020-09-23 16:44 ` [dpdk-dev] [PATCH 00/20] cleanup ethdev close operation Ferruh Yigit
2020-09-27 23:42 ` [dpdk-dev] [PATCH v2 00/25] " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 01/25] ethdev: reset device and interrupt pointers on release Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 02/25] ethdev: allow drivers to return error on close Thomas Monjalon
2020-09-28  0:46     ` Xu, Rosen
2020-09-28  9:51     ` Sachin Saxena (OSS)
2020-09-28 18:26     ` [dpdk-dev] [EXT] " Liron Himi
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 03/25] net/af_packet: release port upon close Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 04/25] net/atlantic: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 05/25] net/axgbe: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 06/25] net/bnx2x: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 07/25] net/bonding: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 08/25] net/failsafe: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 09/25] net/mlx4: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 10/25] net/null: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 11/25] net/octeontx: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 12/25] net/pcap: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 13/25] net/qede: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 14/25] net/ring: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 15/25] net/softnic: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 16/25] net/tap: " Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 17/25] ethdev: remove old close behaviour Thomas Monjalon
2020-09-28 18:25     ` [dpdk-dev] [EXT] " Liron Himi
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 18/25] drivers/net: accept removing device without any port Thomas Monjalon
2020-09-28  0:47     ` Xu, Rosen
2020-09-28  9:54     ` Sachin Saxena (OSS)
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 19/25] drivers/net: check process type in close operation Thomas Monjalon
2020-09-27 23:52     ` Thomas Monjalon
2020-09-28  0:50     ` Xu, Rosen
2020-09-28  9:55     ` Sachin Saxena (OSS)
2020-09-28 14:57       ` Ajit Khaparde
2020-09-28 18:25     ` [dpdk-dev] [EXT] " Liron Himi
2020-09-28 18:51     ` [dpdk-dev] " Stephen Hemminger
2020-09-28 20:57       ` Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 20/25] drivers/net: remove redundant MAC addresses freeing Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 21/25] app/testpmd: reset port status on close notification Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 22/25] app/testpmd: align behaviour of multi-port detach Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 23/25] ethdev: remove forcing stopped state upon close Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 24/25] ethdev: reset all when releasing a port Thomas Monjalon
2020-09-27 23:42   ` [dpdk-dev] [PATCH v2 25/25] ethdev: allow close function to return an error Thomas Monjalon
2020-09-28 18:24     ` [dpdk-dev] [EXT] " Liron Himi
2020-09-28 23:14 ` [dpdk-dev] [PATCH v3 00/29] cleanup ethdev close operation Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 01/29] ethdev: reset device and interrupt pointers on release Thomas Monjalon
2020-09-29 10:52     ` Andrew Rybchenko
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 02/29] ethdev: allow drivers to return error on close Thomas Monjalon
2020-09-29  2:16     ` Wang, Haiyue
2020-09-29  5:56     ` Guo, Jia
2020-09-29 10:53     ` Andrew Rybchenko
2020-09-30 12:12       ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 03/29] net/af_packet: release port upon close Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 04/29] net/atlantic: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 05/29] net/axgbe: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 06/29] net/bnx2x: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 07/29] net/bonding: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 08/29] net/dpaa: " Thomas Monjalon
2020-09-29  4:53     ` Hemant Agrawal
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 09/29] net/dpaa2: " Thomas Monjalon
2020-09-29  4:53     ` Hemant Agrawal
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 10/29] net/enetc: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 11/29] net/failsafe: " Thomas Monjalon
2020-10-05 10:19     ` Gaëtan Rivet
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 12/29] net/mlx4: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 13/29] net/null: " Thomas Monjalon
2020-09-29 16:47     ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 14/29] net/octeontx: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 15/29] net/pcap: " Thomas Monjalon
2020-09-29 16:49     ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 16/29] net/pfe: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 17/29] net/qede: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 18/29] net/ring: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 19/29] net/softnic: " Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 20/29] net/tap: " Thomas Monjalon
2020-09-30  8:34     ` wangyunjian
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 21/29] ethdev: remove old close behaviour Thomas Monjalon
2020-09-29  2:27     ` Wang, Haiyue
2020-09-29  5:55     ` Guo, Jia
2020-09-29 10:38     ` Andrew Rybchenko
2020-09-29 17:08     ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 22/29] drivers/net: accept removing device without any port Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 23/29] drivers/net: check process type in close operation Thomas Monjalon
2020-09-29  2:39     ` Wang, Haiyue
2020-09-29  5:53     ` Guo, Jia
2020-09-29 10:42     ` Andrew Rybchenko
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 24/29] drivers/net: remove redundant MAC addresses freeing Thomas Monjalon
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 25/29] app/testpmd: reset port status on close notification Thomas Monjalon
2020-09-30 12:15     ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 26/29] app/testpmd: align behaviour of multi-port detach Thomas Monjalon
2020-09-30 12:17     ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 27/29] ethdev: remove forcing stopped state upon close Thomas Monjalon
2020-09-29 10:44     ` Andrew Rybchenko
2020-09-29 16:01     ` Ferruh Yigit
2020-09-29 16:06       ` Thomas Monjalon
2020-09-29 16:39         ` Ferruh Yigit
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 28/29] ethdev: reset all when releasing a port Thomas Monjalon
2020-09-29  2:34     ` Wang, Haiyue
2020-09-29  5:51     ` Guo, Jia
2020-09-29 10:26     ` Maxime Coquelin
2020-09-29 10:36       ` Thomas Monjalon
2020-09-29 11:58         ` Wang, Haiyue
2020-09-29 15:50           ` Ferruh Yigit
2020-09-29 16:02             ` Thomas Monjalon
2020-09-29 16:35               ` Ferruh Yigit
2020-09-30 12:17                 ` Ferruh Yigit
2020-09-29 10:50     ` Andrew Rybchenko
2020-09-28 23:14   ` [dpdk-dev] [PATCH v3 29/29] ethdev: allow close function to return an error Thomas Monjalon
2020-09-29 11:05     ` Andrew Rybchenko
2020-09-29 11:47       ` Thomas Monjalon
2020-09-29 11:54         ` Andrew Rybchenko
2020-09-28 23:33   ` [dpdk-dev] [PATCH v3 00/29] cleanup ethdev close operation Stephen Hemminger
2020-09-30 12:22   ` Ferruh Yigit
2020-10-05 17:08 ` [dpdk-dev] [PATCH v4 0/3] " Thomas Monjalon
2020-10-05 17:08   ` [dpdk-dev] [PATCH v4 1/3] ethdev: remove forcing stopped state upon close Thomas Monjalon
2020-10-05 17:08   ` [dpdk-dev] [PATCH v4 2/3] ethdev: reset all when releasing a port Thomas Monjalon
2020-10-05 17:08   ` [dpdk-dev] [PATCH v4 3/3] ethdev: allow close function to return an error Thomas Monjalon
2020-10-06  9:43     ` Ferruh Yigit
2020-10-06 10:57       ` Thomas Monjalon
2020-10-13  8:40         ` Andrew Rybchenko
2020-10-13  8:55           ` Thomas Monjalon
2020-10-13  9:33             ` Ferruh Yigit
2020-10-13 10:06 ` [dpdk-dev] [PATCH v5 0/3] cleanup ethdev close operation Thomas Monjalon
2020-10-13 10:06   ` [dpdk-dev] [PATCH v5 1/3] ethdev: remove forcing stopped state upon close Thomas Monjalon
2020-10-13 12:36     ` Ferruh Yigit
2020-10-13 12:49       ` Thomas Monjalon
2020-10-13 12:45     ` Ferruh Yigit
2020-10-13 12:51       ` Thomas Monjalon
2020-10-13 17:54         ` Ferruh Yigit
2020-10-13 17:59           ` Thomas Monjalon
2020-10-13 10:06   ` [dpdk-dev] [PATCH v5 2/3] ethdev: reset all when releasing a port Thomas Monjalon
2020-10-13 13:10     ` Ferruh Yigit
2020-10-13 10:06   ` [dpdk-dev] [PATCH v5 3/3] ethdev: allow close function to return an error Thomas Monjalon
2020-10-13 10:41     ` Andrew Rybchenko
2020-10-13 10:43       ` Thomas Monjalon
2020-10-13 13:10     ` Ferruh Yigit
2020-10-16 13:32 ` [dpdk-dev] [PATCH v6 0/3] cleanup ethdev close operation Thomas Monjalon
2020-10-16 13:32   ` [dpdk-dev] [PATCH v6 1/3] ethdev: remove forcing stopped state upon close Thomas Monjalon
2020-10-16 13:32   ` [dpdk-dev] [PATCH v6 2/3] ethdev: reset all when releasing a port Thomas Monjalon
2020-10-16 15:21     ` Ajit Khaparde
2020-10-16 13:32   ` [dpdk-dev] [PATCH v6 3/3] ethdev: allow close function to return an error Thomas Monjalon
2020-10-16 17:55   ` [dpdk-dev] [PATCH v6 0/3] cleanup ethdev close operation Ferruh Yigit
2020-10-20 12:24     ` Bruce Richardson

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=5482678.jQLLJtJmLa@thomas \
    --to=thomas@monjalon.net \
    --cc=arybchenko@solarflare.com \
    --cc=beilei.xing@intel.com \
    --cc=bernard.iremonger@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=wenzhuo.lu@intel.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).