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 91128A046B for ; Tue, 23 Jul 2019 15:19:58 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 741021BFE4; Tue, 23 Jul 2019 15:19:58 +0200 (CEST) Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by dpdk.org (Postfix) with ESMTP id 8F5351BFE4; Tue, 23 Jul 2019 15:19:56 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga005.fm.intel.com ([10.253.24.32]) by orsmga101.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 23 Jul 2019 06:19:55 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.64,299,1559545200"; d="scan'208";a="368370640" Received: from silpixa00399498.ir.intel.com (HELO silpixa00399498.ger.corp.intel.com) ([10.237.223.125]) by fmsmga005.fm.intel.com with ESMTP; 23 Jul 2019 06:19:54 -0700 From: Anatoly Burakov To: dev@dpdk.org Cc: stable@dpdk.org Date: Tue, 23 Jul 2019 14:19:53 +0100 Message-Id: <6613e2fb817959e0a1230810b2b0c5b93738f1b8.1563887988.git.anatoly.burakov@intel.com> X-Mailer: git-send-email 2.17.1 Subject: [dpdk-stable] [PATCH] eal: fix proc type auto detection X-BeenThere: stable@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches for DPDK stable branches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: stable-bounces@dpdk.org Sender: "stable" Currently, primary process holds an exclusive lock on the config file, thereby preventing other primaries from spinning up. However, when the primary dies, the lock is no longer being held, even though there might be other secondary processes still running. The fix is two-fold. First of all, downgrade the primary process's exclusive lock to a shared lock once we have it. Second of all, also take out shared locks on the config from the secondaries. We are using fcntl() locks, which get dropped when the file handle is closed, so also remove the closure of config file handle. Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Anatoly Burakov --- lib/librte_eal/linux/eal/eal.c | 37 +++++++++++++++++++++++++++++----- 1 file changed, 32 insertions(+), 5 deletions(-) diff --git a/lib/librte_eal/linux/eal/eal.c b/lib/librte_eal/linux/eal/eal.c index 34db78753..54feb24a3 100644 --- a/lib/librte_eal/linux/eal/eal.c +++ b/lib/librte_eal/linux/eal/eal.c @@ -83,6 +83,13 @@ static struct flock wr_lock = { .l_len = sizeof(early_mem_config.memsegs), }; +static struct flock rd_lock = { + .l_type = F_RDLCK, + .l_whence = SEEK_SET, + .l_start = offsetof(struct rte_mem_config, memsegs), + .l_len = sizeof(early_mem_config.memsegs), +}; + /* Address of global and public configuration */ static struct rte_config rte_config = { .mem_config = &early_mem_config, @@ -343,8 +350,21 @@ rte_eal_config_create(void) if (retval < 0){ close(mem_cfg_fd); mem_cfg_fd = -1; - RTE_LOG(ERR, EAL, "Cannot create lock on '%s'. Is another primary " - "process running?\n", pathname); + RTE_LOG(ERR, EAL, "Cannot create exclusive lock on '%s'. " + "Is another process running?\n", pathname); + return -1; + } + + /* we hold an exclusive lock - now downgrade it to a read lock to allow + * other processes to also hold onto this file while preventing other + * primaries from spinning up. + */ + retval = fcntl(mem_cfg_fd, F_SETLK, &rd_lock); + if (retval < 0) { + close(mem_cfg_fd); + mem_cfg_fd = -1; + RTE_LOG(ERR, EAL, "Cannot downgrade to shared lock on '%s': %s\n", + pathname, strerror(errno)); return -1; } @@ -389,6 +409,16 @@ rte_eal_config_attach(void) return -1; } } + /* lock the file to prevent primary from initializing while this + * process is still running. + */ + if (fcntl(mem_cfg_fd, F_SETLK, &rd_lock) < 0) { + close(mem_cfg_fd); + mem_cfg_fd = -1; + RTE_LOG(ERR, EAL, "Cannot create shared lock on '%s': %s\n", + pathname, strerror(errno)); + return -1; + } /* map it as read-only first */ mem_config = (struct rte_mem_config *) mmap(NULL, sizeof(*mem_config), @@ -427,9 +457,6 @@ rte_eal_config_reattach(void) sizeof(*mem_config), PROT_READ | PROT_WRITE, MAP_SHARED, mem_cfg_fd, 0); - close(mem_cfg_fd); - mem_cfg_fd = -1; - if (mem_config == MAP_FAILED || mem_config != rte_mem_cfg_addr) { if (mem_config != MAP_FAILED) { /* errno is stale, don't use */ -- 2.17.1