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 51D0AA04B3; Mon, 16 Dec 2019 09:50:19 +0100 (CET) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id B9A5C1BF6D; Mon, 16 Dec 2019 09:50:18 +0100 (CET) Received: from us-smtp-delivery-1.mimecast.com (us-smtp-2.mimecast.com [205.139.110.61]) by dpdk.org (Postfix) with ESMTP id 4ACF61BF6C for ; Mon, 16 Dec 2019 09:50:17 +0100 (CET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1576486216; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:autocrypt:autocrypt; bh=mwyVuTFrMBZmubrWS2tFcBjhi4hZv71tpcll3ZwRADQ=; b=APxt+QOH890njS36ACfTHrMJh/H1Jj5yNY4se8JkxP58XOKNKRp+G3sCEE1qCCkl3yQDAg OecXUuQTVuKXA/bfLDJ2i0/sYiqIiZUGQ/P2FpAqu3FqoyetzdeLvzCqJQBSFwUwRpqbb5 pJTYgHjFpzBQ3Arl0KkbCEGaF6QODA8= Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-412-Thty7juhNeCmOnVNj1Na9g-1; Mon, 16 Dec 2019 03:50:12 -0500 Received: from smtp.corp.redhat.com (int-mx08.intmail.prod.int.phx2.redhat.com [10.5.11.23]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id 98B3E800D4E; Mon, 16 Dec 2019 08:50:10 +0000 (UTC) Received: from [10.36.112.19] (ovpn-112-19.ams2.redhat.com [10.36.112.19]) by smtp.corp.redhat.com (Postfix) with ESMTPS id 8BF7A261AF; Mon, 16 Dec 2019 08:50:06 +0000 (UTC) To: Andrew Rybchenko , Matan Azrad , Tiwei Bie , Thomas Monjalon Cc: Ori Kam , "Liang, Cunming" , "Wang, Xiao W" , "Wang, Zhihong" , "Yigit, Ferruh" , Shahaf Shuler , "dev@dpdk.org" , Slava Ovsiienko , Asaf Penso , Olga Shern References: <17d5139b-3c7c-5ce2-1d2a-a86533a8bc2f@solarflare.com> From: Maxime Coquelin Autocrypt: addr=maxime.coquelin@redhat.com; keydata= mQINBFOEQQIBEADjNLYZZqghYuWv1nlLisptPJp+TSxE/KuP7x47e1Gr5/oMDJ1OKNG8rlNg kLgBQUki3voWhUbMb69ybqdMUHOl21DGCj0BTU3lXwapYXOAnsh8q6RRM+deUpasyT+Jvf3a gU35dgZcomRh5HPmKMU4KfeA38cVUebsFec1HuJAWzOb/UdtQkYyZR4rbzw8SbsOemtMtwOx YdXodneQD7KuRU9IhJKiEfipwqk2pufm2VSGl570l5ANyWMA/XADNhcEXhpkZ1Iwj3TWO7XR uH4xfvPl8nBsLo/EbEI7fbuUULcAnHfowQslPUm6/yaGv6cT5160SPXT1t8U9QDO6aTSo59N jH519JS8oeKZB1n1eLDslCfBpIpWkW8ZElGkOGWAN0vmpLfdyiqBNNyS3eGAfMkJ6b1A24un /TKc6j2QxM0QK4yZGfAxDxtvDv9LFXec8ENJYsbiR6WHRHq7wXl/n8guyh5AuBNQ3LIK44x0 KjGXP1FJkUhUuruGyZsMrDLBRHYi+hhDAgRjqHgoXi5XGETA1PAiNBNnQwMf5aubt+mE2Q5r qLNTgwSo2dpTU3+mJ3y3KlsIfoaxYI7XNsPRXGnZi4hbxmeb2NSXgdCXhX3nELUNYm4ArKBP LugOIT/zRwk0H0+RVwL2zHdMO1Tht1UOFGfOZpvuBF60jhMzbQARAQABtCxNYXhpbWUgQ29x dWVsaW4gPG1heGltZS5jb3F1ZWxpbkByZWRoYXQuY29tPokCOAQTAQIAIgUCV3u/5QIbAwYL CQgHAwIGFQgCCQoLBBYCAwECHgECF4AACgkQyjiNKEaHD4ma2g/+P+Hg9WkONPaY1J4AR7Uf kBneosS4NO3CRy0x4WYmUSLYMLx1I3VH6SVjqZ6uBoYy6Fs6TbF6SHNc7QbB6Qjo3neqnQR1 71Ua1MFvIob8vUEl3jAR/+oaE1UJKrxjWztpppQTukIk4oJOmXbL0nj3d8dA2QgHdTyttZ1H xzZJWWz6vqxCrUqHU7RSH9iWg9R2iuTzii4/vk1oi4Qz7y/q8ONOq6ffOy/t5xSZOMtZCspu Mll2Szzpc/trFO0pLH4LZZfz/nXh2uuUbk8qRIJBIjZH3ZQfACffgfNefLe2PxMqJZ8mFJXc RQO0ONZvwoOoHL6CcnFZp2i0P5ddduzwPdGsPq1bnIXnZqJSl3dUfh3xG5ArkliZ/++zGF1O wvpGvpIuOgLqjyCNNRoR7cP7y8F24gWE/HqJBXs1qzdj/5Hr68NVPV1Tu/l2D1KMOcL5sOrz 2jLXauqDWn1Okk9hkXAP7+0Cmi6QwAPuBT3i6t2e8UdtMtCE4sLesWS/XohnSFFscZR6Vaf3 gKdWiJ/fW64L6b9gjkWtHd4jAJBAIAx1JM6xcA1xMbAFsD8gA2oDBWogHGYcScY/4riDNKXi lw92d6IEHnSf6y7KJCKq8F+Jrj2BwRJiFKTJ6ChbOpyyR6nGTckzsLgday2KxBIyuh4w+hMq TGDSp2rmWGJjASq5Ag0EVPSbkwEQAMkaNc084Qvql+XW+wcUIY+Dn9A2D1gMr2BVwdSfVDN7 0ZYxo9PvSkzh6eQmnZNQtl8WSHl3VG3IEDQzsMQ2ftZn2sxjcCadexrQQv3Lu60Tgj7YVYRM H+fLYt9W5YuWduJ+FPLbjIKynBf6JCRMWr75QAOhhhaI0tsie3eDsKQBA0w7WCuPiZiheJaL 4MDe9hcH4rM3ybnRW7K2dLszWNhHVoYSFlZGYh+MGpuODeQKDS035+4H2rEWgg+iaOwqD7bg CQXwTZ1kSrm8NxIRVD3MBtzp9SZdUHLfmBl/tLVwDSZvHZhhvJHC6Lj6VL4jPXF5K2+Nn/Su CQmEBisOmwnXZhhu8ulAZ7S2tcl94DCo60ReheDoPBU8PR2TLg8rS5f9w6mLYarvQWL7cDtT d2eX3Z6TggfNINr/RTFrrAd7NHl5h3OnlXj7PQ1f0kfufduOeCQddJN4gsQfxo/qvWVB7PaE 1WTIggPmWS+Xxijk7xG6x9McTdmGhYaPZBpAxewK8ypl5+yubVsE9yOOhKMVo9DoVCjh5To5 aph7CQWfQsV7cd9PfSJjI2lXI0dhEXhQ7lRCFpf3V3mD6CyrhpcJpV6XVGjxJvGUale7+IOp sQIbPKUHpB2F+ZUPWds9yyVxGwDxD8WLqKKy0WLIjkkSsOb9UBNzgRyzrEC9lgQ/ABEBAAGJ Ah8EGAECAAkFAlT0m5MCGwwACgkQyjiNKEaHD4nU8hAAtt0xFJAy0sOWqSmyxTc7FUcX+pbD KVyPlpl6urKKMk1XtVMUPuae/+UwvIt0urk1mXi6DnrAN50TmQqvdjcPTQ6uoZ8zjgGeASZg jj0/bJGhgUr9U7oG7Hh2F8vzpOqZrdd65MRkxmc7bWj1k81tOU2woR/Gy8xLzi0k0KUa8ueB iYOcZcIGTcs9CssVwQjYaXRoeT65LJnTxYZif2pfNxfINFzCGw42s3EtZFteczClKcVSJ1+L +QUY/J24x0/ocQX/M1PwtZbB4c/2Pg/t5FS+s6UB1Ce08xsJDcwyOPIH6O3tccZuriHgvqKP yKz/Ble76+NFlTK1mpUlfM7PVhD5XzrDUEHWRTeTJSvJ8TIPL4uyfzhjHhlkCU0mw7Pscyxn DE8G0UYMEaNgaZap8dcGMYH/96EfE5s/nTX0M6MXV0yots7U2BDb4soLCxLOJz4tAFDtNFtA wLBhXRSvWhdBJZiig/9CG3dXmKfi2H+wdUCSvEFHRpgo7GK8/Kh3vGhgKmnnxhl8ACBaGy9n fxjSxjSO6rj4/MeenmlJw1yebzkX8ZmaSi8BHe+n6jTGEFNrbiOdWpJgc5yHIZZnwXaW54QT UhhSjDL1rV2B4F28w30jYmlRmm2RdN7iCZfbyP3dvFQTzQ4ySquuPkIGcOOHrvZzxbRjzMx1 Mwqu3GQ= Message-ID: <186093dc-5330-965e-5283-4432b7e996f3@redhat.com> Date: Mon, 16 Dec 2019 09:50:04 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.1.0 MIME-Version: 1.0 In-Reply-To: <17d5139b-3c7c-5ce2-1d2a-a86533a8bc2f@solarflare.com> Content-Language: en-US X-Scanned-By: MIMEDefang 2.84 on 10.5.11.23 X-MC-Unique: Thty7juhNeCmOnVNj1Na9g-1 X-Mimecast-Spam-Score: 0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [dpdk-dev] discussion: creating a new class for vdpa drivers 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 Andrew, On 12/16/19 9:46 AM, Andrew Rybchenko wrote: > On 12/16/19 11:29 AM, Matan Azrad wrote: >> >> Hi all >> >> I understand all of you agree \ not object with the new class for vdpa d= rivers. >=20 > I have two control questions: >=20 > 1. If so, is it allowed to have vDPA driver in > drivers/net/ if it is better from code sharing point > of view? If it has something to share, I think we should move the common bits to the common directory. > 2. If drivers/common is used, is exported functions which are > used by drivers/net/ and drivers/vdpa/ and > data structures are a part of public API/ABI? Hopefully not, > but I'd like to double-check and ensure that it is solved in > the case of shared libraries build. Common functions and data should not be part of the API/ABI I agree. I guess we should use relative paths for including the common headers. >> Based on that, I'm going to start it. >> >> From: Tiwei Bie >>> On Tue, Dec 10, 2019 at 09:00:33AM +0100, Thomas Monjalon wrote: >>>> 10/12/2019 03:41, Tiwei Bie: >>>>> On Mon, Dec 09, 2019 at 02:22:27PM +0300, Andrew Rybchenko wrote: >>>>>> On 12/9/19 1:41 PM, Ori Kam wrote: >>>>>>> From: Andrew Rybchenko >>>>>>>> On 12/8/19 10:06 AM, Matan Azrad wrote: >>>>>>>>> From: Andrew Rybchenko >>>>>>>>>> On 12/6/19 8:32 AM, Liang, Cunming wrote: >>>>>>>>>>> From: Bie, Tiwei >>>>>>>>>>>> On Thu, Dec 05, 2019 at 01:26:36PM +0000, Matan Azrad wrote: >>>>>>>>>>>>> Hi all >>>>>>>>>>>>> >>>>>>>>>>>>> As described in RFC =E2=80=9C[RFC] net: new vdpa PMD for Mell= anox >>>>>>>>>>>>> devices=E2=80=9D, a new vdpa drivers is going to be added for >>>>>>>>>>>>> Mellanox devices =E2=80=93 mlx5_vdpa >>>>>>>>>>>>> >>>>>>>>>>>>> The only vdpa driver now is the IFC driver that is located >>>>>>>>>>>>> in net >>>>>>>> directory. >>>>>>>>>>>>> >>>>>>>>>>>>> The IFC driver and the new mlx5_vdpa driver provide the >>>>>>>>>>>>> vdpa ops >>>>>>>> and >>>>>>>>>>>>> not the eth_dev ops. >>>>>>>>>>>>> >>>>>>>>>>>>> All the others drivers in net provide the eth-dev ops. >>>>>>>>>>>>> >>>>>>>>>>>>> I suggest to create a new class for vdpa drivers, to move >>>>>>>>>>>>> IFC to this class and to add the mlx5_vdpa to this class too. >>>>>>>>>>>>> >>>>>>>>>>>>> Later, all the new drivers that implements the vdpa ops >>>>>>>>>>>>> will be added to the vdpa class. >>>>>>>>>>>> >>>>>>>>>>>> +1. Sounds like a good idea to me. >>>>>>>>>>> +1 >>>>>>>>>> >>>>>>>>>> vDPA drivers are vendor-specific and expected to talk to vendor >>> NIC. I.e. >>>>>>>>>> there are significant chances to share code with network drivers >>> (e.g. >>>>>>>> base >>>>>>>>>> driver). Should base driver be moved to drivers/common in >>>>>>>>>> this case or is >>>>>>>> it >>>>>>>>>> still allows to have vdpa driver in drivers/net together with et= hdev >>> driver? >>>>>>>>> >>>>>>>>> Yes, I think this should be the method, shared code should be >>>>>>>>> moved to >>>>>>>> the drivers/common directory. >>>>>>>>> I think there is a precedence with shared code in common which >>>>>>>>> shares a >>>>>>>> vendor specific code between crypto and net. >>>>>>>> >>>>>>>> I see motivation behind driver/vdpa. However, vdpa and net >>>>>>>> drivers tightly related and I would prefer to avoid extra >>>>>>>> complexity here. Right now simplicity over-weights for me. >>>>>>>> No strong opinion on the topic, but it definitely requires >>>>>>>> better and more clear motivation why a new class should be >>>>>>>> introduced and existing drivers restructured. >>>>>>>> >>>>>>> >>>>>>> Why do you think there is extra complexity? >>>>>> >>>>>> Even grep becomes a bit more complicated J >>>>>> >>>>>>> I think from design correctness it is more correct to create a dedi= cated >>> class for the following reasons: >>>>>>> 1. All of the classes implements a different set of ops. For >>>>>>> example the cryptodev has a defined set of ops, same goes for the >>> compress driver and the ethdev driver. Each ones of them has different = ops. >>> Going by this definition since VDPA has a different set of ops, it make= s sense >>> that it will be in a different class. >>>>>>> >>>>>>> 2. even that both drivers (eth/vdpa) handle traffic from the nic >>>>>>> most of the code is different (the difference is also dependent on = the >>> manufacture) So even the shared code base is not large and can be share= d >>> using the common directory. For example ifc doesn't have any shared cod= e. >>>>>>> >>>>>>> What do you think? >>>>>> >>>>>> The true reason is: if the difference in ops implemented is a key >>>>>> difference which should enforce location in different directories. >>>>>> Or underlying device class is a key. >>>>>> Roughly: >>>>>> - net driver is a control+data path >>>>>> - vdpa driver is a control path only My fear is that control path >>>>>> will grow more and more (Rx mode, RSS, filters and so on) >>>>> >>>>> I think this is a reasonable concern. >>>>> >>>>> One thing needs to be clarified is that, the control path (ops) in >>>>> vdpa driver is something very different with the control path in net >>>>> driver. vdpa is very generic (or more precisely vhost-related), >>>>> instead of network related: >>>>> >>>>> https://eur03.safelinks.protection.outlook.com/?url=3Dhttps%3A%2F%2Fg= i >>>>> >>> thub.com%2FDPDK%2Fdpdk%2Fblob%2Faef1d0733179%2Flib%2Flibrte_vhos >>> t%2F >>>>> rte_vdpa.h%23L40- >>> L78&data=3D02%7C01%7Cmatan%40mellanox.com%7Cfac15 >>>>> >>> 729a67c4c81ee7608d77d7434a2%7Ca652971c7d2e4d9ba6a4d149256f461b%7C >>> 0%7 >>>>> >>> C0%7C637115810358231304&sdata=3D%2BZa39vxadtKx5Ov7vmqcU3RuZhz >>> kOP9o >>>>> 8roEB0d5j6M%3D&reserved=3D0 >>>>> >>>>> It's built on top of vhost-user protocol, manipulates virtqueues, >>>>> virtio/vhost features, memory table, ... >>>>> >>>>> Technically, it's possible to have blk, crypto, ... >>>>> vdpa devices as well (we already have vhost-user-blk, >>>>> vhost-user-crypto, ...). >>>>> >>>>> But network specific features will come eventually, e.g. RSS. One >>>>> possible way to solve it is to define a generic event callback in >>>>> vdpa ops, and vdpa driver can request the corresponding info from >>>>> vhost based on the event received. >>>>> >>>>> Another thing needs to be clarified is that, the control path >>>>> supposed to be used by DPDK apps directly in vdpa should always be >>>>> generic, it should just be something like: >>>>> >>>>> int rte_vdpa_find_device_id(struct rte_vdpa_dev_addr *addr); int >>>>> rte_vhost_driver_attach_vdpa_device(const char *path, int did); int >>>>> rte_vhost_driver_detach_vdpa_device(const char *path); ... >>>>> >>>>> That is to say, users just need to bind the vdpa device to the vhost >>>>> connection. The control path ops in vdpa is supposed to be called by >>>>> vhost-library transparently based on the events on the vhost-user >>>>> connection, i.e. >>>>> the vdpa device will be configured (including RSS) by the guest >>>>> driver in QEMU "directly" via the vhost-user protocol instead of the >>>>> DPDK app in the host. >>>> >>>> Tiwei, in order to be clear, >>>> You think vDPA drivers should be in drivers/vdpa directory? >>> >>> I was just trying to clarify two facts in vDPA to address Andrew's conc= ern. >>> And back to the question, to make sure that we don't miss anything >>> important, (although maybe not very related) it might be helpful to als= o >>> clarify how to support vDPA in OvS at the same time which isn't quite c= lear to >>> me yet.. >>> >>> Regards, >>> Tiwei >>> >>>> >>>> >=20