Cover Detail
Show a cover letter.
GET /api/covers/42565/?format=api
http://patches.dpdk.org/api/covers/42565/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/cover/1530906406-289697-1-git-send-email-yipeng1.wang@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": "<1530906406-289697-1-git-send-email-yipeng1.wang@intel.com>", "list_archive_url": "https://inbox.dpdk.org/dev/1530906406-289697-1-git-send-email-yipeng1.wang@intel.com", "date": "2018-07-06T19:46:38", "name": "[v3,0/8] Add read-write concurrency to rte_hash library", "submitter": { "id": 754, "url": "http://patches.dpdk.org/api/people/754/?format=api", "name": "Wang, Yipeng1", "email": "yipeng1.wang@intel.com" }, "mbox": "http://patches.dpdk.org/project/dpdk/cover/1530906406-289697-1-git-send-email-yipeng1.wang@intel.com/mbox/", "series": [ { "id": 463, "url": "http://patches.dpdk.org/api/series/463/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/list/?series=463", "date": "2018-07-06T19:46:38", "name": "Add read-write concurrency to rte_hash library", "version": 3, "mbox": "http://patches.dpdk.org/series/463/mbox/" } ], "comments": "http://patches.dpdk.org/api/covers/42565/comments/", "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 D85BE1BE1E;\n\tSat, 7 Jul 2018 04:53:59 +0200 (CEST)", "from mga04.intel.com (mga04.intel.com [192.55.52.120])\n\tby dpdk.org (Postfix) with ESMTP id B9AED1BE0C\n\tfor <dev@dpdk.org>; Sat, 7 Jul 2018 04:53:56 +0200 (CEST)", "from fmsmga003.fm.intel.com ([10.253.24.29])\n\tby fmsmga104.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384;\n\t06 Jul 2018 19:53:55 -0700", "from skx-yipeng.jf.intel.com ([10.54.81.175])\n\tby FMSMGA003.fm.intel.com with ESMTP; 06 Jul 2018 19:53:44 -0700" ], "X-Amp-Result": "SKIPPED(no attachment in message)", "X-Amp-File-Uploaded": "False", "X-ExtLoop1": "1", "X-IronPort-AV": "E=Sophos;i=\"5.51,319,1526367600\"; d=\"scan'208\";a=\"62888529\"", "From": "Yipeng Wang <yipeng1.wang@intel.com>", "To": "pablo.de.lara.guarch@intel.com", "Cc": "dev@dpdk.org, yipeng1.wang@intel.com, bruce.richardson@intel.com,\n\thonnappa.nagarahalli@arm.com, vguvva@caviumnetworks.com,\n\tbrijesh.s.singh@gmail.com", "Date": "Fri, 6 Jul 2018 12:46:38 -0700", "Message-Id": "<1530906406-289697-1-git-send-email-yipeng1.wang@intel.com>", "X-Mailer": "git-send-email 2.7.4", "In-Reply-To": "<1528455078-328182-1-git-send-email-yipeng1.wang@intel.com>", "References": "<1528455078-328182-1-git-send-email-yipeng1.wang@intel.com>", "Subject": "[dpdk-dev] [PATCH v3 0/8] Add read-write concurrency to rte_hash\n\tlibrary", "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>" }, "content": "This patch set adds the read-write concurrency support in rte_hash.\nA new flag value is added to indicate if read-write concurrency is needed\nduring creation time. Test cases are implemented to do functional and\nperformance tests.\n\nThe new concurrency model is based on rte_rwlock. When Intel TSX is\navailable and the users indicate to use it, the TM version of the\nrte_rwlock will be called. Both multi-writer and read-write concurrency\nare protected by the rte_rwlock instead of the x86 specific RTM\ninstructions, so the x86 specific header rte_cuckoo_hash_x86.h is removed\nand the code is infused into the main .c file.\n\nA new rte_hash_count API is proposed to count how many keys are inserted\ninto the hash table.\n\n\nv2->v3:\n1. hash: Concurrency bug fix: after beginning cuckoo path moving,\nthe last empty slot needs to be verified again in case other writers\nraced into this slot and occupy it. A new commit is added to do this\nbug fix since it applies to master head as well.\n2. hash: Concurrency bug fix: if cuckoo path is detected to be invalid,\nthe current slot needs to be emptied since it is duplicated to its\ntarget bucket.\n3. hash: \"const\" is used for types in multiple locations. (Pablo)\n4. hash: rte_malloc used for readwriter lock used wrong align\nargument. Similar fix applies to master head so a new commit is\ncreated. (Pablo)\n5. hash: ring size calculation fix is moved to front. (Pablo)\n6. hash: search-and-remove function is refactored to be more aligned\nwith other search function. (Pablo)\n7. test: using jhash in functional test for read-write concurrency.\nIt is because jhash with sequential keys incur more cuckoo path.\n8. Multiple coding style, typo, commit message fixes. (Pablo)\n\n\nv1->v2:\n1. Split each commit into two commits for easier review (Pablo).\n2. Add more comments in various places (Pablo).\n3. hash: In key insertion function, move duplicated key checking to\nearlier location and protect it using locks. Checking duplicated key\nshould happen first and data updates should be protected.\n4. hash: In lookup bulk function, put signature comparison in lock,\nsince writers could happen between signature match on two buckets.\n5. hash: Add write locks to reset function as well to protect resets.\n5. test: Fix 32-bit compilation error in read-write test (Pablo).\n6. test: Check total physical core count in read-write test. Don't\ntest with thread count that larger than physical core count.\n7. Other minor fixes such as typos (Pablo).\n\nYipeng Wang (8):\n hash: fix multiwriter lock memory allocation\n hash: fix a multi-writer bug\n hash: fix to have more accurate key slot size\n hash: make duplicated code into functions\n hash: add read and write concurrency support\n test: add tests in hash table perf test\n test: add test case for read write concurrency\n hash: add new API function to query the key count\n\n lib/librte_hash/meson.build | 1 -\n lib/librte_hash/rte_cuckoo_hash.c | 705 +++++++++++++++++++++-------------\n lib/librte_hash/rte_cuckoo_hash.h | 18 +-\n lib/librte_hash/rte_cuckoo_hash_x86.h | 164 --------\n lib/librte_hash/rte_hash.h | 14 +\n lib/librte_hash/rte_hash_version.map | 8 +\n test/test/Makefile | 1 +\n test/test/test_hash.c | 12 +\n test/test/test_hash_multiwriter.c | 9 +\n test/test/test_hash_perf.c | 36 +-\n test/test/test_hash_readwrite.c | 646 +++++++++++++++++++++++++++++++\n 11 files changed, 1167 insertions(+), 447 deletions(-)\n delete mode 100644 lib/librte_hash/rte_cuckoo_hash_x86.h\n create mode 100644 test/test/test_hash_readwrite.c" }{ "id": 42565, "url": "