From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id 60AC1A0352; Mon, 23 Dec 2019 10:35:51 +0100 (CET) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id BA9631C01; Mon, 23 Dec 2019 10:35:49 +0100 (CET) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id 48242F72; Mon, 23 Dec 2019 10:35:46 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga005.fm.intel.com ([10.253.24.32]) by fmsmga103.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 23 Dec 2019 01:35:45 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.69,347,1571727600"; d="scan'208";a="417206322" Received: from fmsmsx108.amr.corp.intel.com ([10.18.124.206]) by fmsmga005.fm.intel.com with ESMTP; 23 Dec 2019 01:35:45 -0800 Received: from FMSMSX110.amr.corp.intel.com (10.18.116.10) by FMSMSX108.amr.corp.intel.com (10.18.124.206) with Microsoft SMTP Server (TLS) id 14.3.439.0; Mon, 23 Dec 2019 01:35:45 -0800 Received: from shsmsx107.ccr.corp.intel.com (10.239.4.96) by fmsmsx110.amr.corp.intel.com (10.18.116.10) with Microsoft SMTP Server (TLS) id 14.3.439.0; Mon, 23 Dec 2019 01:35:45 -0800 Received: from shsmsx101.ccr.corp.intel.com ([169.254.1.19]) by SHSMSX107.ccr.corp.intel.com ([169.254.9.164]) with mapi id 14.03.0439.000; Mon, 23 Dec 2019 17:35:43 +0800 From: "Li, Xiaoyun" To: Gavin Hu , "Wu, Jingjing" CC: "dev@dpdk.org" , "Maslekar, Omkar" , "stable@dpdk.org" , nd , "jerinj@marvell.com" , Honnappa Nagarahalli , "Richardson, Bruce" , nd Thread-Topic: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier issue Thread-Index: AQHVs7Tq7x2YYQcLw0mJCFIUj/Hkaqe8D3eAgAtTD0D//4iOAIAAlKPQ Date: Mon, 23 Dec 2019 09:35:42 +0000 Message-ID: References: <20191204151916.12607-1-xiaoyun.li@intel.com> <20191216015854.28725-1-xiaoyun.li@intel.com> In-Reply-To: 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 Subject: Re: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier issue X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Hi Still, stability and correctness are much more important than performance. As I said, with WC can benefit more than 20X perf. Comparing to this benefi= t, the difference between rte_wmb and rte_io_wmb is not that important. And in my test, the performance is not that bad with rte_wmb especially wit= h large packets which are the normal use cases. BTW, I've searched linux kernel codes and don't see any NTB device on arm p= latform. So I don't think you need to consider the perf hurt to arm. Best Regards Xiaoyun Li > -----Original Message----- > From: Gavin Hu [mailto:Gavin.Hu@arm.com] > Sent: Monday, December 23, 2019 16:38 > To: Li, Xiaoyun ; Wu, Jingjing > Cc: dev@dpdk.org; Maslekar, Omkar ; > stable@dpdk.org; nd ; jerinj@marvell.com; Honnappa > Nagarahalli ; Richardson, Bruce > ; nd > Subject: RE: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier issu= e >=20 > Hi Xiaoyun, >=20 > > -----Original Message----- > > From: Li, Xiaoyun > > Sent: Monday, December 23, 2019 3:52 PM > > To: Gavin Hu ; Wu, Jingjing > > Cc: dev@dpdk.org; Maslekar, Omkar ; > > stable@dpdk.org; nd > > Subject: RE: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier > > issue > > > > Hi > > I reconsidered and retested about this issue. > > I still need to use rte_wmb instead of using rte_io_wmb. > > > > Because to achieve high performance, ntb needs to turn on WC(write > > combining) feature. The perf difference with and without WC enabled is > > more than 20X. > > And when WC enabled, rte_io_wmb cannot make sure the instructions are > > in order only rte_wmb can make sure that. > > > > And in my retest, when sending 64 bytes packets, using rte_io_wmb will > > cause out-of-order issue and cause memory corruption on rx side. > > And using rte_wmb is fine. > That's true, as it is declared as 'write combine' region, even x86 is kno= wn as > strong ordered, it is the interconnect or PCI RC may do the reordering, '= write > combine', 'write coalescing', which caused this problem. > IMO, rte_io_*mb barriers on x86 should be promoted to stronger is WC is > involved(but that will sap performance for non-WC memories?). > https://code.dpdk.org/dpdk/latest/source/lib/librte_eal/common/include/ar= ch/ > x86/rte_atomic.h#L78 >=20 > Using rte_wmb will hurt performance for aarch64 also, as pci device memor= y > accesses to a single device are strongly ordered therefore the strongest > rte_wmb is not necessary. > > So I can only use v1 patch and suspend v2 patch in patchwork. > > > > Best Regards > > Xiaoyun Li > > > > > -----Original Message----- > > > From: Gavin Hu (Arm Technology China) [mailto:Gavin.Hu@arm.com] > > > Sent: Monday, December 16, 2019 18:50 > > > To: Li, Xiaoyun ; Wu, Jingjing > > > > > Cc: dev@dpdk.org; Maslekar, Omkar ; > > > stable@dpdk.org; nd > > > Subject: RE: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier > > issue > > > > > > > > > > > > > -----Original Message----- > > > > From: dev On Behalf Of Xiaoyun Li > > > > Sent: Monday, December 16, 2019 9:59 AM > > > > To: jingjing.wu@intel.com > > > > Cc: dev@dpdk.org; omkar.maslekar@intel.com; Xiaoyun Li > > > > ; stable@dpdk.org > > > > Subject: [dpdk-dev] [PATCH v2] raw/ntb: fix write memory barrier > > > > issue > > > > > > > > All buffers and ring info should be written before tail register up= date. > > > > This patch relocates the write memory barrier before updating tail > > > > register to avoid potential issues. > > > > > > > > Fixes: 11b5c7daf019 ("raw/ntb: add enqueue and dequeue functions") > > > > Cc: stable@dpdk.org > > > > > > > > Signed-off-by: Xiaoyun Li > > > > --- > > > > v2: > > > > * Replaced rte_wmb with rte_io_wmb since rte_io_wmb is enough. > > > > --- > > > > drivers/raw/ntb/ntb.c | 4 ++-- > > > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > > > > > diff --git a/drivers/raw/ntb/ntb.c b/drivers/raw/ntb/ntb.c index > > > > ad7f6abfd..c7de86f36 100644 > > > > --- a/drivers/raw/ntb/ntb.c > > > > +++ b/drivers/raw/ntb/ntb.c > > > > @@ -683,8 +683,8 @@ ntb_enqueue_bufs(struct rte_rawdev *dev, > > > > sizeof(struct ntb_used) * nb1); > > > > rte_memcpy(txq->tx_used_ring, tx_used + nb1, > > > > sizeof(struct ntb_used) * nb2); > > > > + rte_io_wmb(); > > > As both txq->tx_used_ring and *txq->used_cnt are physically reside > > > in the > > PCI > > > device side, rte_io_wmb is correct to ensure the ordering. > > > > > > > *txq->used_cnt =3D txq->last_used; > > > > - rte_wmb(); > > > > > > > > /* update queue stats */ > > > > hw->ntb_xstats[NTB_TX_BYTES_ID + off] +=3D bytes; @@ - > > 789,8 > > > +789,8 @@ > > > > ntb_dequeue_bufs(struct rte_rawdev *dev, > > > > sizeof(struct ntb_desc) * nb1); > > > > rte_memcpy(rxq->rx_desc_ring, rx_desc + nb1, > > > > sizeof(struct ntb_desc) * nb2); > > > > + rte_io_wmb(); > > > > *rxq->avail_cnt =3D rxq->last_avail; > > > > - rte_wmb(); > > > > > > > > /* update queue stats */ > > > > off =3D NTB_XSTATS_NUM * ((size_t)context + 1); > > > > -- > > > > 2.17.1 > > > > > > Reviewed-by: Gavin Hu