DPDK patches and discussions
 help / color / mirror / Atom feed
From: Yuanhan Liu <yliu@fridaylinux.org>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: Wenzhuo Lu <wenzhuo.lu@intel.com>,
	Jingjing Wu <jingjing.wu@intel.com>,
	dev@dpdk.org, Zhiyong Yang <zhiyong.yang@intel.com>,
	Bernard Iremonger <bernard.iremonger@intel.com>,
	Thomas Monjalon <thomas@monjalon.net>,
	Maxime Coquelin <maxime.coquelin@redhat.com>
Subject: Re: [dpdk-dev] [PATCH] app/testpmd: fix failsafe PMD failure on exit
Date: Sun, 27 May 2018 12:06:30 +0800	[thread overview]
Message-ID: <20180527040630.x3wrjsblyqbxk3bz@yuanhanliu-NB0.tencent.com> (raw)
In-Reply-To: <20180522183509.66644-1-ferruh.yigit@intel.com>

On Tue, May 22, 2018 at 07:35:08PM +0100, Ferruh Yigit wrote:
> vdevs detach on testpmd exit implemented as workaround to fix
> a virtio-user issue. The issue was virtio-user cleanup is not
> called and existing socket file not cleaned up which will fail
> next run.
> 
> The vdev cleanup causing problems in failsafe PMD.
> 
> Reduce the cleanup to only virtio-user and add a comment that this
> workaround should be converted to a proper cleanup, not something
> specific to virtio-user, and not something specific to vdev and
> testpmd.
> 
> Fixes: fe890955114d ("app/testpmd: fix exit for vdevs")
> 
...
>  pmd_test_exit(void)
>  {
> -	const struct rte_bus *bus;
>  	struct rte_device *device;
>  	portid_t pt_id;
>  	int ret;
> @@ -2025,13 +2024,21 @@ pmd_test_exit(void)
>  	if (ports != NULL) {
>  		no_link_check = 1;
>  		RTE_ETH_FOREACH_DEV(pt_id) {
> -			device = rte_eth_devices[pt_id].device;
> -			bus = rte_bus_find_by_device(device);
>  			printf("\nShutting down port %d...\n", pt_id);
>  			fflush(stdout);
>  			stop_port(pt_id);
>  			close_port(pt_id);
> -			if (bus && !strcmp(bus->name, "vdev"))
> +
> +			/*
> +			 * This is a workaround to fix a virtio-user issue that
> +			 * requires to call clean-up routine to remove existing
> +			 * socket.

I came across this patch while I was cherry-picking patches to 17.11.4
release. And this patch seems wrong to me.

Any particular reason why the socket removal can not be done in virtio-user
pmd, say at its close method?

	--yliu
> +			 * This workaround valid only for testpmd, needs a fix
> +			 * valid for all applications.
> +			 * TODO: Implement proper resource cleanup
> +			 */
> +			device = rte_eth_devices[pt_id].device;
> +			if (device && !strcmp(device->driver->name, "net_virtio_user"))
>  				detach_port(pt_id);
>  		}
>  	}
> -- 
> 2.14.3

  parent reply	other threads:[~2018-05-27  4:06 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-22 18:35 Ferruh Yigit
2018-05-22 19:47 ` Thomas Monjalon
2018-05-22 20:49   ` Ferruh Yigit
2018-05-22 20:50     ` Ferruh Yigit
2018-05-27  4:06 ` Yuanhan Liu [this message]
2018-05-27 12:37   ` Thomas Monjalon
2018-06-02  5:01     ` Yuanhan Liu

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=20180527040630.x3wrjsblyqbxk3bz@yuanhanliu-NB0.tencent.com \
    --to=yliu@fridaylinux.org \
    --cc=bernard.iremonger@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=jingjing.wu@intel.com \
    --cc=maxime.coquelin@redhat.com \
    --cc=thomas@monjalon.net \
    --cc=wenzhuo.lu@intel.com \
    --cc=zhiyong.yang@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).