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 56DDFA3160 for ; Fri, 11 Oct 2019 16:12:25 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 264171EB01; Fri, 11 Oct 2019 16:12:25 +0200 (CEST) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id D68B51EAF1 for ; Fri, 11 Oct 2019 16:12:23 +0200 (CEST) Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 40D67307D868; Fri, 11 Oct 2019 14:12:23 +0000 (UTC) Received: from [10.36.112.48] (ovpn-112-48.ams2.redhat.com [10.36.112.48]) by smtp.corp.redhat.com (Postfix) with ESMTPS id 5B8555C1B2; Fri, 11 Oct 2019 14:12:21 +0000 (UTC) To: Marvin Liu , tiwei.bie@intel.com, zhihong.wang@intel.com, stephen@networkplumber.org, gavin.hu@arm.com Cc: dev@dpdk.org References: <20190925171329.63734-1-yong.liu@intel.com> <20191009133849.69002-1-yong.liu@intel.com> <20191009133849.69002-14-yong.liu@intel.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: <5be3cd3a-4246-65a8-1d56-8ae8ac4751d8@redhat.com> Date: Fri, 11 Oct 2019 16:12:19 +0200 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: <20191009133849.69002-14-yong.liu@intel.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit X-Scanned-By: MIMEDefang 2.79 on 10.5.11.16 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.48]); Fri, 11 Oct 2019 14:12:23 +0000 (UTC) Subject: Re: [dpdk-dev] [PATCH v4 13/14] vhost: check whether disable software pre-fetch 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" On 10/9/19 3:38 PM, Marvin Liu wrote: > Disable software pre-fetch actions on Skylake and later platforms. > Hardware can fetch needed data for vhost, additional software pre-fetch > will impact performance. > > Signed-off-by: Marvin Liu > > diff --git a/lib/librte_vhost/Makefile b/lib/librte_vhost/Makefile > index 30839a001..5f3b42e56 100644 > --- a/lib/librte_vhost/Makefile > +++ b/lib/librte_vhost/Makefile > @@ -16,6 +16,12 @@ CFLAGS += -I vhost_user > CFLAGS += -fno-strict-aliasing > LDLIBS += -lpthread > > +AVX512_SUPPORT=$(shell $(CC) -march=native -dM -E - + > +ifneq ($(AVX512_SUPPORT),) > +CFLAGS += -DDISABLE_SWPREFETCH > +endif That's problematic I think, because the machine running the lib may be different from the machine building it, for example distros. In this case, a Skylake or later may be used to build the package, but with passing "-march=haswell". It would end-up prefetching being disabled whereas we would expect it to be enabled. I see several solutions: - Check for CONFIG_RTE_ENABLE_AVX512 flag. - Keep prefetch instructions (what would be the impact on Skylake and later?) - Remove prefetch instructions (what would be the impact on pre- Skylake?) But really, I think we need some figures before applying such a patch. What performance gain do you measure with this patch? > ifeq ($(RTE_TOOLCHAIN), gcc) > ifeq ($(shell test $(GCC_VERSION) -ge 83 && echo 1), 1) > CFLAGS += -DSUPPORT_GCC_UNROLL_PRAGMA > diff --git a/lib/librte_vhost/meson.build b/lib/librte_vhost/meson.build > index ddf0ee579..5c6f0c0b4 100644 > --- a/lib/librte_vhost/meson.build > +++ b/lib/librte_vhost/meson.build > @@ -15,6 +15,10 @@ elif (toolchain == 'clang' and cc.version().version_compare('>=3.7.0')) > elif (toolchain == 'icc' and cc.version().version_compare('>=16.0.0')) > cflags += '-DSUPPORT_ICC_UNROLL_PRAGMA' > endif > +r = run_command(toolchain, '-march=native', '-dM', '-E', '-', ' +if (r.stdout().strip() != '') > + cflags += '-DDISABLE_SWPREFETCH' > +endif > dpdk_conf.set('RTE_LIBRTE_VHOST_POSTCOPY', > cc.has_header('linux/userfaultfd.h')) > version = 4 > diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c > index 56c2080fb..046e497c2 100644 > --- a/lib/librte_vhost/virtio_net.c > +++ b/lib/librte_vhost/virtio_net.c > @@ -1075,7 +1075,9 @@ virtio_dev_rx_batch_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, > > UNROLL_PRAGMA(UNROLL_PRAGMA_PARAM) > for (i = 0; i < PACKED_BATCH_SIZE; i++) { > +#ifndef DISABLE_SWPREFETCH > rte_prefetch0((void *)(uintptr_t)desc_addrs[i]); > +#endif > hdrs[i] = (struct virtio_net_hdr_mrg_rxbuf *) > (uintptr_t)desc_addrs[i]; > lens[i] = pkts[i]->pkt_len + dev->vhost_hlen; > @@ -1144,7 +1146,9 @@ virtio_dev_rx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, > uint32_t remained = count; > > do { > +#ifndef DISABLE_SWPREFETCH > rte_prefetch0(&vq->desc_packed[vq->last_avail_idx]); > +#endif > > if (remained >= PACKED_BATCH_SIZE) { > if (!virtio_dev_rx_batch_packed(dev, vq, pkts)) { > @@ -1790,7 +1794,9 @@ virtio_dev_tx_batch_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, > > UNROLL_PRAGMA(UNROLL_PRAGMA_PARAM) > for (i = 0; i < PACKED_BATCH_SIZE; i++) { > +#ifndef DISABLE_SWPREFETCH > rte_prefetch0((void *)(uintptr_t)desc_addrs[i]); > +#endif > rte_memcpy(rte_pktmbuf_mtod_offset(pkts[i], void *, 0), > (void *)(uintptr_t)(desc_addrs[i] + buf_offset), > pkts[i]->pkt_len); > @@ -2046,7 +2052,9 @@ virtio_dev_tx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, > uint32_t remained = count; > > do { > +#ifndef DISABLE_SWPREFETCH > rte_prefetch0(&vq->desc_packed[vq->last_avail_idx]); > +#endif > > if (remained >= PACKED_BATCH_SIZE) { > if (!virtio_dev_tx_batch_packed(dev, vq, mbuf_pool, >