From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by dpdk.org (Postfix) with ESMTP id 0A361C47E for ; Mon, 29 Jun 2015 11:21:28 +0200 (CEST) Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by orsmga101.jf.intel.com with ESMTP; 29 Jun 2015 02:21:20 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.13,698,1427785200"; d="scan'208";a="736740037" Received: from pgsmsx102.gar.corp.intel.com ([10.221.44.80]) by fmsmga001.fm.intel.com with ESMTP; 29 Jun 2015 02:21:18 -0700 Received: from kmsmsx154.gar.corp.intel.com (172.21.73.14) by PGSMSX102.gar.corp.intel.com (10.221.44.80) with Microsoft SMTP Server (TLS) id 14.3.224.2; Mon, 29 Jun 2015 17:20:34 +0800 Received: from shsmsx102.ccr.corp.intel.com (10.239.4.154) by KMSMSX154.gar.corp.intel.com (172.21.73.14) with Microsoft SMTP Server (TLS) id 14.3.224.2; Mon, 29 Jun 2015 17:20:34 +0800 Received: from shsmsx101.ccr.corp.intel.com ([169.254.1.246]) by shsmsx102.ccr.corp.intel.com ([169.254.2.165]) with mapi id 14.03.0224.002; Mon, 29 Jun 2015 17:20:33 +0800 From: "Qiu, Michael" To: "Iremonger, Bernard" , "dev@dpdk.org" Thread-Topic: [PATCH 1/2 v4] fm10k: Free queues when close port Thread-Index: AQHQr+pPvWjWc3reIUWJMy4rOvx5Tg== Date: Mon, 29 Jun 2015 09:20:32 +0000 Message-ID: <533710CFB86FA344BFBF2D6802E60286046A6343@SHSMSX101.ccr.corp.intel.com> References: <1434702717-11877-1-git-send-email-michael.qiu@intel.com> <1435307390-20140-1-git-send-email-michael.qiu@intel.com> <1435307390-20140-2-git-send-email-michael.qiu@intel.com> <8CEF83825BEC744B83065625E567D7C204A422AD@IRSMSX108.ger.corp.intel.com> <533710CFB86FA344BFBF2D6802E60286046A5FD7@SHSMSX101.ccr.corp.intel.com> <8CEF83825BEC744B83065625E567D7C204A42BB8@IRSMSX108.ger.corp.intel.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Cc: "He, Shaopeng" Subject: Re: [dpdk-dev] [PATCH 1/2 v4] fm10k: Free queues when close port X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 29 Jun 2015 09:21:29 -0000 On 6/29/2015 4:57 PM, Iremonger, Bernard wrote:=0A= >> -----Original Message-----=0A= >> From: Qiu, Michael=0A= >> Sent: Monday, June 29, 2015 9:17 AM=0A= >> To: Iremonger, Bernard; dev@dpdk.org=0A= >> Cc: Chen, Jing D; He, Shaopeng=0A= >> Subject: Re: [PATCH 1/2 v4] fm10k: Free queues when close port=0A= >>=0A= >> On 6/26/2015 7:02 PM, Iremonger, Bernard wrote:=0A= >>>> -----Original Message-----=0A= >>>> From: Qiu, Michael=0A= >>>> Sent: Friday, June 26, 2015 9:30 AM=0A= >>>> To: dev@dpdk.org=0A= >>>> Cc: Chen, Jing D; He, Shaopeng; Iremonger, Bernard; Qiu, Michael=0A= >>>> Subject: [PATCH 1/2 v4] fm10k: Free queues when close port=0A= >>>>=0A= >>>> When close a port, lots of memory should be released, such as=0A= >>>> software rings, queues, etc.=0A= >>>>=0A= >>>> Signed-off-by: Michael Qiu =0A= >>>> ---=0A= >>> Hi Michael,=0A= >>>=0A= >>> There are 2 comments inline=0A= >>>=0A= >>>> drivers/net/fm10k/fm10k_ethdev.c | 37=0A= >>>> +++++++++++++++++++++++++++++++++----=0A= >>>> 1 file changed, 33 insertions(+), 4 deletions(-)=0A= >>>>=0A= >>>> diff --git a/drivers/net/fm10k/fm10k_ethdev.c=0A= >>>> b/drivers/net/fm10k/fm10k_ethdev.c=0A= >>>> index 406c350..eba7228 100644=0A= >>>> --- a/drivers/net/fm10k/fm10k_ethdev.c=0A= >>>> +++ b/drivers/net/fm10k/fm10k_ethdev.c=0A= >>>> @@ -65,6 +65,8 @@ static void=0A= >>>> fm10k_MAC_filter_set(struct rte_eth_dev *dev, const u8 *mac, bool=0A= >>>> add); static void fm10k_MACVLAN_remove_all(struct rte_eth_dev=0A= >> *dev);=0A= >>>> +static void fm10k_tx_queue_release(void *queue); static void=0A= >>>> +fm10k_rx_queue_release(void *queue);=0A= >>>>=0A= >>>> static void=0A= >>>> fm10k_mbx_initlock(struct fm10k_hw *hw) @@ -809,11 +811,37 @@=0A= >>>> fm10k_dev_stop(struct rte_eth_dev *dev)=0A= >>>>=0A= >>>> PMD_INIT_FUNC_TRACE();=0A= >>>>=0A= >>>> - for (i =3D 0; i < dev->data->nb_tx_queues; i++)=0A= >>>> - fm10k_dev_tx_queue_stop(dev, i);=0A= >>>> + if (dev->data->tx_queues)=0A= >>>> + for (i =3D 0; i < dev->data->nb_tx_queues; i++)=0A= >>>> + fm10k_dev_tx_queue_stop(dev, i);=0A= >>>>=0A= >>>> - for (i =3D 0; i < dev->data->nb_rx_queues; i++)=0A= >>>> - fm10k_dev_rx_queue_stop(dev, i);=0A= >>>> + if (dev->data->rx_queues)=0A= >>>> + for (i =3D 0; i < dev->data->nb_rx_queues; i++)=0A= >>>> + fm10k_dev_rx_queue_stop(dev, i);=0A= >>>> +}=0A= >>>> +=0A= >>>> +static void=0A= >>>> +fm10k_dev_queue_release(struct rte_eth_dev *dev) {=0A= >>>> + int i;=0A= >>>> +=0A= >>>> + PMD_INIT_FUNC_TRACE();=0A= >>>> +=0A= >>>> + if (dev->data->tx_queues) {=0A= >>>> + for (i =3D 0; i < dev->data->nb_tx_queues; i++)=0A= >>>> + fm10k_tx_queue_release(dev->data-=0A= >>>>> tx_queues[i]);=0A= >>>> + rte_free(dev->data->tx_queues);=0A= >>>> + dev->data->tx_queues =3D NULL;=0A= >>> The memory for dev->data->tx_queues is not allocated in the fm10k=0A= >>> PMD, so it should probably not be released here.=0A= >>> I have submitted a patch today for rte_eth_dev.c to do this.=0A= >>> /dev/patchwork/patch/5829/=0A= >>>=0A= >>>> + dev->data->nb_tx_queues =3D 0;=0A= >>>> + }=0A= >>>> +=0A= >>>> + if (dev->data->rx_queues) {=0A= >>>> + for (i =3D 0; i < dev->data->nb_rx_queues; i++)=0A= >>>> + fm10k_rx_queue_release(dev->data-=0A= >>>>> rx_queues[i]);=0A= >>>> + rte_free(dev->data->rx_queues);=0A= >>>> + dev->data->rx_queues =3D NULL;=0A= >>> The memory for dev->data->rx_queues is not allocated in the fm10k=0A= >>> PMD, so it should probably not be released here.=0A= >>> I have submitted a patch today for rte_eth_dev.c to do this.=0A= >>> /dev/patchwork/patch/5829/=0A= >> Is it a good idea? What about to close the port for twice at a time?=0A= >> I think it is better to do it in rte_eth_dev_close(), I will give the co= mments to=0A= >> you.=0A= >>=0A= >> Thanks,=0A= >> Michael=0A= > Hi Michael,=0A= > Could you take a look at the comments on http://dpdk.org/dev/patchwork/pa= tch/5829/=0A= =0A= Hi, Bernard=0A= =0A= I have give comments on it.=0A= =0A= > The consensus is that memory should be freed in the component that alloca= ted it.=0A= > In my pmd hotplug patches I have used a flag to ensure that dev_close is = not called twice.=0A= > In the e1000 patch I have added a stopped flag to struct e1000_adapter. = =0A= > http://dpdk.org/dev/patchwork/patch/5655/=0A= =0A= =0A= =0A= I reviewed your patch about ixgbe and fvl before. But forget e1000.=0A= =0A= In my logic, when dev->data->rx_queues is NULL, that means this device=0A= has been closed before. What else, we even do not care about whether it=0A= has been closed or not, when close() function be called, all memory=0A= should be freed if exist am I right?=0A= =0A= So, check dev->data->rx_queues whether it is NULL will be recommend in=0A= close function, only this could avoid unsafe situations for pointer.=0A= =0A= Thanks,=0A= Michael=0A= > Regards,=0A= >=0A= > Bernard.=0A= >=0A= > =0A= >=0A= >=0A= >=0A= =0A=