From mboxrd@z Thu Jan  1 00:00:00 1970
From: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Subject: Re: [PATCH] net/i40e: fix wrong device flag setup
Date: Wed, 8 Feb 2017 09:49:51 +0800
Message-ID: <20170208014951.GM20916@yliu-dev.sh.intel.com>
References: <1486164969-2566-1-git-send-email-qi.z.zhang@intel.com>
 <736135ef-b9cf-cd6e-1dfc-ecba9d63e114@intel.com>
 <fb52d206-8464-baf3-dfa2-24fb208c2a3b@intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Qi Zhang <qi.z.zhang@intel.com>, jingjing.wu@intel.com,
 helin.zhang@intel.com, dev@dpdk.org,
 Thomas Monjalon <thomas.monjalon@6wind.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mga07.intel.com (mga07.intel.com [134.134.136.100])
 by dpdk.org (Postfix) with ESMTP id B76E1591E
 for <dev@dpdk.org>; Wed,  8 Feb 2017 02:47:26 +0100 (CET)
Content-Disposition: inline
In-Reply-To: <fb52d206-8464-baf3-dfa2-24fb208c2a3b@intel.com>
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

On Tue, Feb 07, 2017 at 10:32:32AM +0000, Ferruh Yigit wrote:
> On 2/7/2017 10:24 AM, Ferruh Yigit wrote:
> > On 2/3/2017 11:36 PM, Qi Zhang wrote:
> >> dev_flags is wrong overwritten by RTE_ETH_DEV_DETACHABLE after
> >> rte_eth_copy_pci_info.
> >>
...
> Hi Yuanhan,
> 
> I saw while checking for this, there can be an issue in virtio for this
> path, can you please check:
> 
> eth_virtio_dev_init()
>   set-some-eth_dev-dev_flags
>   virtio_init_device()
>     rte_eth_copy_pci_info()
>       eth_dev->data->dev_flags = 0;   <---- Reset previous flags

Hi Ferruh,

Nice catch! And yes, it's buggy. It may be trivially fixed by setting
the orders correctly. But my question is why "rte_eth_copy_pci_info"
does a reset on "dev_flags"? It's an *un-documented* side-effect of
this function, which could be very error-prone, like this one.

	--yliu