Cover Detail
Show a cover letter.
GET /api/covers/68762/?format=api
{ "id": 68762, "url": "http://patches.dpdk.org/api/covers/68762/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/cover/20200417133639.14019-1-konstantin.ananyev@intel.com/", "project": { "id": 1, "url": "http://patches.dpdk.org/api/projects/1/?format=api", "name": "DPDK", "link_name": "dpdk", "list_id": "dev.dpdk.org", "list_email": "dev@dpdk.org", "web_url": "http://core.dpdk.org", "scm_url": "git://dpdk.org/dpdk", "webscm_url": "http://git.dpdk.org/dpdk", "list_archive_url": "https://inbox.dpdk.org/dev", "list_archive_url_format": "https://inbox.dpdk.org/dev/{}", "commit_url_format": "" }, "msgid": "<20200417133639.14019-1-konstantin.ananyev@intel.com>", "list_archive_url": "https://inbox.dpdk.org/dev/20200417133639.14019-1-konstantin.ananyev@intel.com", "date": "2020-04-17T13:36:30", "name": "[v4,0/9] New sync modes for ring", "submitter": { "id": 33, "url": "http://patches.dpdk.org/api/people/33/?format=api", "name": "Ananyev, Konstantin", "email": "konstantin.ananyev@intel.com" }, "mbox": "http://patches.dpdk.org/project/dpdk/cover/20200417133639.14019-1-konstantin.ananyev@intel.com/mbox/", "series": [ { "id": 9464, "url": "http://patches.dpdk.org/api/series/9464/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/list/?series=9464", "date": "2020-04-17T13:36:30", "name": "New sync modes for ring", "version": 4, "mbox": "http://patches.dpdk.org/series/9464/mbox/" } ], "comments": "http://patches.dpdk.org/api/covers/68762/comments/", "headers": { "Return-Path": "<dev-bounces@dpdk.org>", "X-Original-To": "patchwork@inbox.dpdk.org", "Delivered-To": "patchwork@inbox.dpdk.org", "Received": [ "from dpdk.org (dpdk.org [92.243.14.124])\n\tby inbox.dpdk.org (Postfix) with ESMTP id 7608CA058A;\n\tFri, 17 Apr 2020 15:36:53 +0200 (CEST)", "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id C89211E895;\n\tFri, 17 Apr 2020 15:36:51 +0200 (CEST)", "from mga17.intel.com (mga17.intel.com [192.55.52.151])\n by dpdk.org (Postfix) with ESMTP id 03C881E886\n for <dev@dpdk.org>; Fri, 17 Apr 2020 15:36:49 +0200 (CEST)", "from orsmga003.jf.intel.com ([10.7.209.27])\n by fmsmga107.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384;\n 17 Apr 2020 06:36:49 -0700", "from sivswdev08.ir.intel.com ([10.237.217.47])\n by orsmga003.jf.intel.com with ESMTP; 17 Apr 2020 06:36:47 -0700" ], "IronPort-SDR": [ "\n 4Ex/a7CJwTZnJh2dkcZkyHMTJXWJ2ZH7JbB5JHo6PcwjyaTQ+cBzW9F49F0NYy7xHwxiRLATzH\n TuMFvRgvW6uA==", "\n +yH/9MW5umymHj+xzHyiIN7GSY2WLpLZrJ9nPICpyWbwcXcUbgSqofaDAXHs13IuFLmHYtC6mi\n sqoddYvreOGA==" ], "X-Amp-Result": "SKIPPED(no attachment in message)", "X-Amp-File-Uploaded": "False", "X-ExtLoop1": "1", "X-IronPort-AV": "E=Sophos;i=\"5.72,395,1580803200\"; d=\"scan'208\";a=\"254209863\"", "From": "Konstantin Ananyev <konstantin.ananyev@intel.com>", "To": "dev@dpdk.org", "Cc": "honnappa.nagarahalli@arm.com, david.marchand@redhat.com,\n jielong.zjl@antfin.com, Konstantin Ananyev <konstantin.ananyev@intel.com>", "Date": "Fri, 17 Apr 2020 14:36:30 +0100", "Message-Id": "<20200417133639.14019-1-konstantin.ananyev@intel.com>", "X-Mailer": "git-send-email 2.18.0", "In-Reply-To": "<20200403174235.23308-1-konstantin.ananyev@intel.com>", "References": "<20200403174235.23308-1-konstantin.ananyev@intel.com>", "Subject": "[dpdk-dev] [PATCH v4 0/9] New sync modes for ring", "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 <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 <mailto:dev-request@dpdk.org?subject=subscribe>", "Errors-To": "dev-bounces@dpdk.org", "Sender": "\"dev\" <dev-bounces@dpdk.org>" }, "content": "V3 - V4 changes:\nAddress comments from Honnappa:\n1. for new sync modes make legacy API wrappers around _elem_ calls\n2. remove rte_ring_(hts|rts)_generic.h\n3. few changes in C11 version\n4. peek API - add missing functions for _elem_\n5. remove _IS_SP/_IS_MP, etc. internal macros\n6. fix param types (obj_table) for _elem_functions\n7. fix formal API comments\n8. deduplicate code for test_ring_stress\n9. added functional tests for new sync modes\n\nV2 - V3 changes:\n1. Few more compilation fixes (for gcc 4.8.X)\n2. Extra update devtools/libabigail.abignore (workaround) \n\nV1 - V2 changes:\n1. Fix compilation issues\n2. Add C11 atomics support\n3. Updates devtools/libabigail.abignore (workaround)\n\nRFC - V1 changes:\n1. remove ABI brekage (at least I hope I did)\n2. Add support for ring_elem\n3. Rework peek related API a bit\n4. Rework test to make it less verbose and unite all test-cases\n in one command\n5. Add new test-case for MT peek API\n\nTODO list:\n1. Update docs\n\nThese days more and more customers use(/try to use) DPDK based apps within\novercommitted systems (multiple acttive threads over same pysical cores):\nVM, container deployments, etc.\nOne quite common problem they hit:\nLock-Holder-Preemption/Lock-Waiter-Preemption with rte_ring.\nLHP is quite a common problem for spin-based sync primitives\n(spin-locks, etc.) on overcommitted systems.\nThe situation gets much worse when some sort of\nfair-locking technique is used (ticket-lock, etc.).\nAs now not only lock-owner but also lock-waiters scheduling\norder matters a lot (LWP).\nThese two problems are well-known for kernel within VMs:\nhttp://www-archive.xenproject.org/files/xensummitboston08/LHP.pdf\nhttps://www.cs.hs-rm.de/~kaiser/events/wamos2017/Slides/selcuk.pdf\nThe problem with rte_ring is that while head accusion is sort of\nun-fair locking, waiting on tail is very similar to ticket lock schema -\ntail has to be updated in particular order.\nThat makes current rte_ring implementation to perform\nreally pure on some overcommited scenarios.\nIt is probably not possible to completely resolve LHP problem in\nuserspace only (without some kernel communication/intervention).\nBut removing fairness at tail update helps to avoid LWP and\ncan mitigate the situation significantly.\nThis patch proposes two new optional ring synchronization modes:\n1) Head/Tail Sync (HTS) mode\nIn that mode enqueue/dequeue operation is fully serialized:\n only one thread at a time is allowed to perform given op.\n As another enhancement provide ability to split enqueue/dequeue\n operation into two phases:\n - enqueue/dequeue start\n - enqueue/dequeue finish\n That allows user to inspect objects in the ring without removing\n them from it (aka MT safe peek).\n2) Relaxed Tail Sync (RTS)\nThe main difference from original MP/MC algorithm is that\ntail value is increased not by every thread that finished enqueue/dequeue,\nbut only by the last one.\nThat allows threads to avoid spinning on ring tail value,\nleaving actual tail value change to the last thread in the update queue.\n\nNote that these new sync modes are optional.\nFor current rte_ring users nothing should change\n(both in terms of API/ABI and performance).\nExisting sync modes MP/MC,SP/SC kept untouched, set up in the same\nway (via flags and _init_), and MP/MC remains as default one.\nThe only thing that changed:\nFormat of prod/cons now could differ depending on mode selected at _init_.\nSo user has to stick with one sync model through whole ring lifetime.\nIn other words, user can't create a ring for let say SP mode and then\nin the middle of data-path change his mind and start using MP_RTS mode.\nFor existing modes (SP/MP, SC/MC) format remains the same and\nuser can still use them interchangeably, though of course it is an\nerror prone practice.\n\nTest results on IA (see below) show significant improvements\nfor average enqueue/dequeue op times on overcommitted systems.\nFor 'classic' DPDK deployments (one thread per core) original MP/MC\nalgorithm still shows best numbers, though for 64-bit target\nRTS numbers are not that far away.\nNumbers were produced by new UT test-case: ring_stress_autotest, i.e.:\necho ring_stress_autotest | ./dpdk-test -n 4 --lcores='...'\n\nX86_64 @ Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz\nDEQ+ENQ average cycles/obj\n MP/MC HTS RTS\n1thread@1core(--lcores=6-7) 8.00 8.15 8.99\n2thread@2core(--lcores=6-8) 19.14 19.61 20.35\n4thread@4core(--lcores=6-10) 29.43 29.79 31.82\n8thread@8core(--lcores=6-14) 110.59 192.81 119.50\n16thread@16core(--lcores=6-22) 461.03 813.12 495.59\n32thread/@32core(--lcores='6-22,55-70') 982.90 1972.38 1160.51\n\n2thread@1core(--lcores='6,(10-11)@7' 20140.50 23.58 25.14\n4thread@2core(--lcores='6,(10-11)@7,(20-21)@8' 153680.60 76.88 80.05\n8thread@2core(--lcores='6,(10-13)@7,(20-23)@8' 280314.32 294.72 318.79\n16thread@2core(--lcores='6,(10-17)@7,(20-27)@8' 643176.59 1144.02 1175.14\n32thread@2core(--lcores='6,(10-25)@7,(30-45)@8' 4264238.80 4627.48 4892.68\n\n8thread@2core(--lcores='6,(10-17)@(7,8))' 321085.98 298.59 307.47\n16thread@4core(--lcores='6,(20-35)@(7-10))' 1900705.61 575.35 678.29\n32thread@4core(--lcores='6,(20-51)@(7-10))' 5510445.85 2164.36 2714.12\n\ni686 @ Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz\nDEQ+ENQ average cycles/obj\n MP/MC HTS RTS\n1thread@1core(--lcores=6-7) 7.85 12.13 11.31\n2thread@2core(--lcores=6-8) 17.89 24.52 21.86\n8thread@8core(--lcores=6-14) 32.58 354.20 54.58\n32thread/@32core(--lcores='6-22,55-70') 813.77 6072.41 2169.91\n\n2thread@1core(--lcores='6,(10-11)@7' 16095.00 36.06 34.74\n8thread@2core(--lcores='6,(10-13)@7,(20-23)@8' 1140354.54 346.61 361.57\n16thread@2core(--lcores='6,(10-17)@7,(20-27)@8' 1920417.86 1314.90 1416.65\n\n8thread@2core(--lcores='6,(10-17)@(7,8))' 594358.61 332.70 357.74\n32thread@4core(--lcores='6,(20-51)@(7-10))' 5319896.86 2836.44 3028.87\n\nKonstantin Ananyev (9):\n test/ring: add contention stress test\n ring: prepare ring to allow new sync schemes\n ring: introduce RTS ring mode\n test/ring: add contention stress test for RTS ring\n ring: introduce HTS ring mode\n test/ring: add contention stress test for HTS ring\n ring: introduce peek style API\n test/ring: add stress test for MT peek API\n test/ring: add functional tests for new sync modes\n\n app/test/Makefile | 5 +\n app/test/meson.build | 5 +\n app/test/test_pdump.c | 6 +-\n app/test/test_ring.c | 93 ++++--\n app/test/test_ring_hts_stress.c | 32 ++\n app/test/test_ring_mpmc_stress.c | 31 ++\n app/test/test_ring_peek_stress.c | 43 +++\n app/test/test_ring_rts_stress.c | 32 ++\n app/test/test_ring_stress.c | 57 ++++\n app/test/test_ring_stress.h | 38 +++\n app/test/test_ring_stress_impl.h | 396 ++++++++++++++++++++++\n devtools/libabigail.abignore | 7 +\n lib/librte_pdump/rte_pdump.c | 2 +-\n lib/librte_port/rte_port_ring.c | 12 +-\n lib/librte_ring/Makefile | 8 +-\n lib/librte_ring/meson.build | 11 +-\n lib/librte_ring/rte_ring.c | 114 ++++++-\n lib/librte_ring/rte_ring.h | 243 ++++++++------\n lib/librte_ring/rte_ring_c11_mem.h | 44 +++\n lib/librte_ring/rte_ring_core.h | 181 ++++++++++\n lib/librte_ring/rte_ring_elem.h | 141 ++++++--\n lib/librte_ring/rte_ring_generic.h | 48 +++\n lib/librte_ring/rte_ring_hts.h | 332 ++++++++++++++++++\n lib/librte_ring/rte_ring_hts_c11_mem.h | 207 ++++++++++++\n lib/librte_ring/rte_ring_peek.h | 446 +++++++++++++++++++++++++\n lib/librte_ring/rte_ring_rts.h | 439 ++++++++++++++++++++++++\n lib/librte_ring/rte_ring_rts_c11_mem.h | 179 ++++++++++\n 27 files changed, 2978 insertions(+), 174 deletions(-)\n create mode 100644 app/test/test_ring_hts_stress.c\n create mode 100644 app/test/test_ring_mpmc_stress.c\n create mode 100644 app/test/test_ring_peek_stress.c\n create mode 100644 app/test/test_ring_rts_stress.c\n create mode 100644 app/test/test_ring_stress.c\n create mode 100644 app/test/test_ring_stress.h\n create mode 100644 app/test/test_ring_stress_impl.h\n create mode 100644 lib/librte_ring/rte_ring_core.h\n create mode 100644 lib/librte_ring/rte_ring_hts.h\n create mode 100644 lib/librte_ring/rte_ring_hts_c11_mem.h\n create mode 100644 lib/librte_ring/rte_ring_peek.h\n create mode 100644 lib/librte_ring/rte_ring_rts.h\n create mode 100644 lib/librte_ring/rte_ring_rts_c11_mem.h" }