From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by dpdk.space (Postfix) with ESMTP id B85C2A0679 for ; Fri, 5 Apr 2019 16:36:42 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 69D351B524; Fri, 5 Apr 2019 16:36:41 +0200 (CEST) Received: from mga05.intel.com (mga05.intel.com [192.55.52.43]) by dpdk.org (Postfix) with ESMTP id BE9721B51A; Fri, 5 Apr 2019 16:36:35 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by fmsmga105.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 05 Apr 2019 07:36:24 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.60,312,1549958400"; d="scan'208";a="161746982" Received: from mhajkowx-mobl.ger.corp.intel.com ([10.104.116.181]) by fmsmga001.fm.intel.com with ESMTP; 05 Apr 2019 07:36:23 -0700 From: Hajkowski To: david.hunt@intel.com Cc: dev@dpdk.org, Marcin Hajkowski , stable@dpdk.org, Anatoly Burakov Date: Fri, 5 Apr 2019 16:35:35 +0200 Message-Id: <20190405143535.2028-1-marcinx.hajkowski@intel.com> X-Mailer: git-send-email 2.20.1.windows.1 In-Reply-To: <20190318115647.14784-1-marcinx.hajkowski@intel.com> References: <20190318115647.14784-1-marcinx.hajkowski@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Subject: [dpdk-stable] [PATCH] power: fix non thread-safe power env modification 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" From: Marcin Hajkowski Due to lack of thread safety in exisiting solution use spinlock mechanism for atomic modification of power environment related data. Fixes: 445c6528b5 ("power: common interface for guest and host") Cc: stable@dpdk.org Signed-off-by: Marcin Hajkowski Acked-by: Anatoly Burakov --- doc/guides/rel_notes/release_19_05.rst | 2 ++ lib/librte_power/rte_power.c | 30 ++++++++++++++++++-------- lib/librte_power/rte_power.h | 2 +- 3 files changed, 24 insertions(+), 10 deletions(-) diff --git a/doc/guides/rel_notes/release_19_05.rst b/doc/guides/rel_notes/release_19_05.rst index c9d443e83..79f8ba76d 100644 --- a/doc/guides/rel_notes/release_19_05.rst +++ b/doc/guides/rel_notes/release_19_05.rst @@ -176,6 +176,8 @@ API Changes ``rte_vfio_container_dma_unmap`` have been extended with an option to request mapping or un-mapping to the default vfio container fd. +* power: ``rte_power_set_env`` and ``rte_power_unset_env`` functions + have been modified to be thread safe. ABI Changes ----------- diff --git a/lib/librte_power/rte_power.c b/lib/librte_power/rte_power.c index a05fbef94..540d69be7 100644 --- a/lib/librte_power/rte_power.c +++ b/lib/librte_power/rte_power.c @@ -2,7 +2,7 @@ * Copyright(c) 2010-2014 Intel Corporation */ -#include +#include #include "rte_power.h" #include "power_acpi_cpufreq.h" @@ -12,7 +12,7 @@ enum power_management_env global_default_env = PM_ENV_NOT_SET; -volatile uint32_t global_env_cfg_status = 0; +static rte_spinlock_t global_env_cfg_lock = RTE_SPINLOCK_INITIALIZER; /* function pointers */ rte_power_freqs_t rte_power_freqs = NULL; @@ -30,9 +30,15 @@ rte_power_get_capabilities_t rte_power_get_capabilities; int rte_power_set_env(enum power_management_env env) { - if (rte_atomic32_cmpset(&global_env_cfg_status, 0, 1) == 0) { + rte_spinlock_lock(&global_env_cfg_lock); + + if (global_default_env != PM_ENV_NOT_SET) { + rte_spinlock_unlock(&global_env_cfg_lock); return 0; } + + int ret = 0; + if (env == PM_ENV_ACPI_CPUFREQ) { rte_power_freqs = power_acpi_cpufreq_freqs; rte_power_get_freq = power_acpi_cpufreq_get_freq; @@ -73,18 +79,24 @@ rte_power_set_env(enum power_management_env env) } else { RTE_LOG(ERR, POWER, "Invalid Power Management Environment(%d) set\n", env); - rte_power_unset_env(); - return -1; + ret = -1; } - global_default_env = env; - return 0; + + if (ret == 0) + global_default_env = env; + else + global_default_env = PM_ENV_NOT_SET; + + rte_spinlock_unlock(&global_env_cfg_lock); + return ret; } void rte_power_unset_env(void) { - if (rte_atomic32_cmpset(&global_env_cfg_status, 1, 0) != 0) - global_default_env = PM_ENV_NOT_SET; + rte_spinlock_lock(&global_env_cfg_lock); + global_default_env = PM_ENV_NOT_SET; + rte_spinlock_unlock(&global_env_cfg_lock); } enum power_management_env diff --git a/lib/librte_power/rte_power.h b/lib/librte_power/rte_power.h index dee7af345..47db69f33 100644 --- a/lib/librte_power/rte_power.h +++ b/lib/librte_power/rte_power.h @@ -26,7 +26,7 @@ enum power_management_env {PM_ENV_NOT_SET, PM_ENV_ACPI_CPUFREQ, PM_ENV_KVM_VM, /** * Set the default power management implementation. If this is not called prior * to rte_power_init(), then auto-detect of the environment will take place. - * It is not thread safe. + * It is thread safe. * * @param env * env. The environment in which to initialise Power Management for. -- 2.17.2