List cover comments

GET /api/covers/53518/comments/?format=api
HTTP 200 OK
Allow: GET, HEAD, OPTIONS
Content-Type: application/json
Link: 
<http://patches.dpdk.org/api/covers/53518/comments/?format=api&page=1>; rel="first",
<http://patches.dpdk.org/api/covers/53518/comments/?format=api&page=1>; rel="last"
Vary: Accept
[ { "id": 96263, "web_url": "http://patches.dpdk.org/comment/96263/", "msgid": "<CAJFAV8wF6p4h8R3+KZ3Ldr2QtV+3MfRLHrpOJ4dw+DSziRqpUQ@mail.gmail.com>", "list_archive_url": "https://inbox.dpdk.org/dev/CAJFAV8wF6p4h8R3+KZ3Ldr2QtV+3MfRLHrpOJ4dw+DSziRqpUQ@mail.gmail.com", "date": "2019-05-17T13:04:26", "subject": "Re: [dpdk-dev] [PATCH 0/5] vhost: I-cache pressure optimizations", "submitter": { "id": 1173, "url": "http://patches.dpdk.org/api/people/1173/?format=api", "name": "David Marchand", "email": "david.marchand@redhat.com" }, "content": "On Fri, May 17, 2019 at 2:23 PM Maxime Coquelin <maxime.coquelin@redhat.com>\nwrote:\n\n> Some OVS-DPDK PVP benchmarks show a performance drop\n> when switching from DPDK v17.11 to v18.11.\n>\n> With the addition of packed ring layout support,\n> rte_vhost_enqueue_burst and rte_vhost_dequeue_burst\n> became very large, and only a part of the instructions\n> are executed (either packed or split ring used).\n>\n> This series aims at improving the I-cache pressure,\n> first by un-inlining split and packed rings, but\n> also by moving parts considered as cold in dedicated\n> functions (dirty page logging, fragmented descriptors\n> buffer management added for CVE-2018-1059).\n>\n> With the series applied, size of the enqueue and\n> dequeue split paths is reduced significantly:\n>\n> +---------+--------------------+---------------------+\n> | Version | Enqueue split path | Dequeue split path |\n> +---------+--------------------+---------------------+\n> | v19.05 | 16461B | 25521B |\n> | +series | 7286B | 11285B |\n> +---------+--------------------+---------------------+\n>\n> Using perf tool to monitor iTLB-load-misses event\n> while doing PVP benchmark with testpmd as vswitch,\n> we can see the number of iTLB misses being reduced:\n>\n> - v19.05:\n> # perf stat --repeat 10 -C 2,3 -e iTLB-load-miss -- sleep 10\n>\n> Performance counter stats for 'CPU(s) 2,3' (10 runs):\n>\n> 2,438 iTLB-load-miss\n> ( +- 13.43% )\n>\n> 10.00058928 +- 0.00000336 seconds time elapsed ( +- 0.00% )\n>\n> - +series:\n> # perf stat --repeat 10 -C 2,3 -e iTLB-load-miss -- sleep 10\n>\n> Performance counter stats for 'CPU(s) 2,3' (10 runs):\n>\n> 55 iTLB-load-miss\n> ( +- 10.08% )\n>\n> 10.00059466 +- 0.00000283 seconds time elapsed ( +- 0.00% )\n>\n> The series also force the inlining of some rte_memcpy\n> helpers, as by adding packed ring support, some of them\n> were not more inlined but embedded as functions in\n> the virtio_net object file, which was not expected.\n>\n> Finally, the series simplifies the descriptors buffers\n> prefetching, by doing it in the recently introduced\n> descriptor buffer mapping function.\n>\n> Maxime Coquelin (4):\n> vhost: un-inline dirty pages logging functions\n> vhost: do not inline packed and split functions\n> vhost: do not inline unlikely fragmented buffers code\n> vhost: simplify descriptor's buffer prefetching\n>\n> root (1):\n> eal/x86: force inlining of all memcpy and mov helpers\n>\n\nroot ? \"oops\" :-)", "headers": { "Return-Path": "<dev-bounces@dpdk.org>", "X-Original-To": "patchwork@dpdk.org", "Delivered-To": "patchwork@dpdk.org", "Received": [ "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id 58A725A44;\n\tFri, 17 May 2019 15:04:40 +0200 (CEST)", "from mail-vk1-f194.google.com (mail-vk1-f194.google.com\n\t[209.85.221.194]) by dpdk.org (Postfix) with ESMTP id 9F5842BCE\n\tfor <dev@dpdk.org>; Fri, 17 May 2019 15:04:38 +0200 (CEST)", "by mail-vk1-f194.google.com with SMTP id p24so1990421vki.5\n\tfor <dev@dpdk.org>; Fri, 17 May 2019 06:04:38 -0700 (PDT)" ], "X-Google-DKIM-Signature": "v=1; a=rsa-sha256; c=relaxed/relaxed;\n\td=1e100.net; s=20161025;\n\th=x-gm-message-state:mime-version:references:in-reply-to:from:date\n\t:message-id:subject:to:cc;\n\tbh=tOC2iWCfT8Tz8DelVwmrloErmE9YhMinR5z3Pz3wpKc=;\n\tb=jPbOY8x+0RijYvLW/JXHSnDSPzTDgR6sNYrYi+0ct/WDYfgQfDFEz+1uSv9JY+22IA\n\tPwvE8HuG17Tq3yzMmEW3VAxe8A5FRRPn4pnsyA2i5GSfFmBFxb3lk/0hoTTsQpTUwc++\n\tBYCnCS6uIboCjimeHNbWR07pHnUSZD6bvEOFLvTbZlQqRoQHFt9TeyAMVCUvGXbEMxtN\n\thCIps6uoirt3nfS6LJd5sZAnD0T4df4smbpsmHU5YfVCb0YSQ+iXChSnbQbg60O461+Q\n\t5RMZnLIpD987sOo48nJQPpersARuBWSC6BKzyvbbSu/m/v10gpWwCq+t3g/hEZlJ5sfy\n\tIbbA==", "X-Gm-Message-State": "APjAAAUm+wNp6DLX//1xaZltufoRsVzEOMCOUpXx8QEj5vscFw2oPsgp\n\tCIIzq67bShzFR2Yw/cmCiM1j1x14wB/Mc7gmdMrXtQ==", "X-Google-Smtp-Source": "APXvYqxxbcE60FTrdzZEjB4FnKOoRl0PRTaSC/o8yjrsYif5di/WutTG2AbyQVo13o0+dL4pdHl8ACnVDX/1GbOO/+s=", "X-Received": "by 2002:ac5:c219:: with SMTP id m25mr2221516vkk.53.1558098277692;\n\tFri, 17 May 2019 06:04:37 -0700 (PDT)", "MIME-Version": "1.0", "References": "<20190517122220.31283-1-maxime.coquelin@redhat.com>", "In-Reply-To": "<20190517122220.31283-1-maxime.coquelin@redhat.com>", "From": "David Marchand <david.marchand@redhat.com>", "Date": "Fri, 17 May 2019 15:04:26 +0200", "Message-ID": "<CAJFAV8wF6p4h8R3+KZ3Ldr2QtV+3MfRLHrpOJ4dw+DSziRqpUQ@mail.gmail.com>", "To": "Maxime Coquelin <maxime.coquelin@redhat.com>", "Cc": "dev <dev@dpdk.org>, Tiwei Bie <tiwei.bie@intel.com>, \n\tJens Freimann <jfreimann@redhat.com>,\n\tZhihong Wang <zhihong.wang@intel.com>, \n\tBruce Richardson <bruce.richardson@intel.com>, \n\t\"Ananyev, Konstantin\" <konstantin.ananyev@intel.com>", "Content-Type": "text/plain; charset=\"UTF-8\"", "X-Content-Filtered-By": "Mailman/MimeDel 2.1.15", "Subject": "Re: [dpdk-dev] [PATCH 0/5] vhost: I-cache pressure optimizations", "X-BeenThere": "dev@dpdk.org", "X-Mailman-Version": "2.1.15", "Precedence": "list", "List-Id": "DPDK patches and discussions <dev.dpdk.org>", "List-Unsubscribe": "<https://mails.dpdk.org/options/dev>,\n\t<mailto:dev-request@dpdk.org?subject=unsubscribe>", "List-Archive": "<http://mails.dpdk.org/archives/dev/>", "List-Post": "<mailto:dev@dpdk.org>", "List-Help": "<mailto:dev-request@dpdk.org?subject=help>", "List-Subscribe": "<https://mails.dpdk.org/listinfo/dev>,\n\t<mailto:dev-request@dpdk.org?subject=subscribe>", "Errors-To": "dev-bounces@dpdk.org", "Sender": "\"dev\" <dev-bounces@dpdk.org>" }, "addressed": null }, { "id": 96268, "web_url": "http://patches.dpdk.org/comment/96268/", "msgid": "<e4a7c39c-1fda-a10f-86ab-416967e18a49@redhat.com>", "list_archive_url": "https://inbox.dpdk.org/dev/e4a7c39c-1fda-a10f-86ab-416967e18a49@redhat.com", "date": "2019-05-17T14:42:45", "subject": "Re: [dpdk-dev] [PATCH 0/5] vhost: I-cache pressure optimizations", "submitter": { "id": 512, "url": "http://patches.dpdk.org/api/people/512/?format=api", "name": "Maxime Coquelin", "email": "maxime.coquelin@redhat.com" }, "content": "On 5/17/19 3:04 PM, David Marchand wrote:\n> \n> \n> On Fri, May 17, 2019 at 2:23 PM Maxime Coquelin \n> <maxime.coquelin@redhat.com <mailto:maxime.coquelin@redhat.com>> wrote:\n> \n> Some OVS-DPDK PVP benchmarks show a performance drop\n> when switching from DPDK v17.11 to v18.11.\n> \n> With the addition of packed ring layout support,\n> rte_vhost_enqueue_burst and rte_vhost_dequeue_burst\n> became very large, and only a part of the instructions\n> are executed (either packed or split ring used).\n> \n> This series aims at improving the I-cache pressure,\n> first by un-inlining split and packed rings, but\n> also by moving parts considered as cold in dedicated\n> functions (dirty page logging, fragmented descriptors\n> buffer management added for CVE-2018-1059).\n> \n> With the series applied, size of the enqueue and\n> dequeue split paths is reduced significantly:\n> \n> +---------+--------------------+---------------------+\n> | Version | Enqueue split path |  Dequeue split path |\n> +---------+--------------------+---------------------+\n> | v19.05  | 16461B             | 25521B              |\n> | +series | 7286B              | 11285B              |\n> +---------+--------------------+---------------------+\n> \n> Using perf tool to monitor iTLB-load-misses event\n> while doing PVP benchmark with testpmd as vswitch,\n> we can see the number of iTLB misses being reduced:\n> \n> - v19.05:\n> # perf stat --repeat 10  -C 2,3  -e iTLB-load-miss -- sleep 10\n> \n>  Performance counter stats for 'CPU(s) 2,3' (10 runs):\n> \n>              2,438      iTLB-load-miss \n>                   ( +- 13.43% )\n> \n>        10.00058928 +- 0.00000336 seconds time elapsed  ( +-  0.00% )\n> \n> - +series:\n> # perf stat --repeat 10  -C 2,3  -e iTLB-load-miss -- sleep 10\n> \n>  Performance counter stats for 'CPU(s) 2,3' (10 runs):\n> \n>                 55      iTLB-load-miss \n>                   ( +- 10.08% )\n> \n>        10.00059466 +- 0.00000283 seconds time elapsed  ( +-  0.00% )\n> \n> The series also force the inlining of some rte_memcpy\n> helpers, as by adding packed ring support, some of them\n> were not more inlined but embedded as functions in\n> the virtio_net object file, which was not expected.\n> \n> Finally, the series simplifies the descriptors buffers\n> prefetching, by doing it in the recently introduced\n> descriptor buffer mapping function.\n> \n> Maxime Coquelin (4):\n>   vhost: un-inline dirty pages logging functions\n>   vhost: do not inline packed and split functions\n>   vhost: do not inline unlikely fragmented buffers code\n>   vhost: simplify descriptor's buffer prefetching\n> \n> root (1):\n>   eal/x86: force inlining of all memcpy and mov helpers\n> \n> \n> root ? \"oops\" :-)\n\nIndeed... Oops!\n\n> \n> \n> -- \n> David Marchand", "headers": { "Return-Path": "<dev-bounces@dpdk.org>", "X-Original-To": "patchwork@dpdk.org", "Delivered-To": "patchwork@dpdk.org", "Received": [ "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id 809147CB0;\n\tFri, 17 May 2019 16:42:57 +0200 (CEST)", "from mx1.redhat.com (mx1.redhat.com [209.132.183.28])\n\tby dpdk.org (Postfix) with ESMTP id 8932E6CD8\n\tfor <dev@dpdk.org>; Fri, 17 May 2019 16:42:55 +0200 (CEST)", "from smtp.corp.redhat.com\n\t(int-mx03.intmail.prod.int.phx2.redhat.com [10.5.11.13])\n\t(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))\n\t(No client certificate requested)\n\tby mx1.redhat.com (Postfix) with ESMTPS id DB1DF30C1CC1;\n\tFri, 17 May 2019 14:42:54 +0000 (UTC)", "from [10.36.112.59] (ovpn-112-59.ams2.redhat.com [10.36.112.59])\n\tby smtp.corp.redhat.com (Postfix) with ESMTPS id 1935160852;\n\tFri, 17 May 2019 14:42:46 +0000 (UTC)" ], "To": "David Marchand <david.marchand@redhat.com>", "Cc": "dev <dev@dpdk.org>, Tiwei Bie <tiwei.bie@intel.com>,\n\tJens Freimann <jfreimann@redhat.com>,\n\tZhihong Wang <zhihong.wang@intel.com>, \n\tBruce Richardson <bruce.richardson@intel.com>,\n\t\"Ananyev, Konstantin\" <konstantin.ananyev@intel.com>", "References": "<20190517122220.31283-1-maxime.coquelin@redhat.com>\n\t<CAJFAV8wF6p4h8R3+KZ3Ldr2QtV+3MfRLHrpOJ4dw+DSziRqpUQ@mail.gmail.com>", "From": "Maxime Coquelin <maxime.coquelin@redhat.com>", "Message-ID": "<e4a7c39c-1fda-a10f-86ab-416967e18a49@redhat.com>", "Date": "Fri, 17 May 2019 16:42:45 +0200", "User-Agent": "Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101\n\tThunderbird/60.6.1", "MIME-Version": "1.0", "In-Reply-To": "<CAJFAV8wF6p4h8R3+KZ3Ldr2QtV+3MfRLHrpOJ4dw+DSziRqpUQ@mail.gmail.com>", "Content-Type": "text/plain; charset=utf-8; format=flowed", "Content-Language": "en-US", "Content-Transfer-Encoding": "8bit", "X-Scanned-By": "MIMEDefang 2.79 on 10.5.11.13", "X-Greylist": "Sender IP whitelisted, not delayed by milter-greylist-4.5.16\n\t(mx1.redhat.com [10.5.110.46]); Fri, 17 May 2019 14:42:54 +0000 (UTC)", "Subject": "Re: [dpdk-dev] [PATCH 0/5] vhost: I-cache pressure optimizations", "X-BeenThere": "dev@dpdk.org", "X-Mailman-Version": "2.1.15", "Precedence": "list", "List-Id": "DPDK patches and discussions <dev.dpdk.org>", "List-Unsubscribe": "<https://mails.dpdk.org/options/dev>,\n\t<mailto:dev-request@dpdk.org?subject=unsubscribe>", "List-Archive": "<http://mails.dpdk.org/archives/dev/>", "List-Post": "<mailto:dev@dpdk.org>", "List-Help": "<mailto:dev-request@dpdk.org?subject=help>", "List-Subscribe": "<https://mails.dpdk.org/listinfo/dev>,\n\t<mailto:dev-request@dpdk.org?subject=subscribe>", "Errors-To": "dev-bounces@dpdk.org", "Sender": "\"dev\" <dev-bounces@dpdk.org>" }, "addressed": null } ]