From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <aburakov@ecsmtp.ir.intel.com>
Received: from mga01.intel.com (mga01.intel.com [192.55.52.88])
by dpdk.org (Postfix) with ESMTP id 6B9674CBB
for <dev@dpdk.org>; Thu, 22 Mar 2018 13:36:08 +0100 (CET)
X-Amp-Result: SKIPPED(no attachment in message)
X-Amp-File-Uploaded: False
Received: from orsmga001.jf.intel.com ([10.7.209.18])
by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384;
22 Mar 2018 05:36:07 -0700
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.48,344,1517904000"; d="scan'208";a="41238534"
Received: from irvmail001.ir.intel.com ([163.33.26.43])
by orsmga001.jf.intel.com with ESMTP; 22 Mar 2018 05:36:05 -0700
Received: from sivswdev01.ir.intel.com (sivswdev01.ir.intel.com
[10.237.217.45])
by irvmail001.ir.intel.com (8.14.3/8.13.6/MailSET/Hub) with ESMTP id
w2MCa4ai000540; Thu, 22 Mar 2018 12:36:04 GMT
Received: from sivswdev01.ir.intel.com (localhost [127.0.0.1])
by sivswdev01.ir.intel.com with ESMTP id w2MCa4bD029483;
Thu, 22 Mar 2018 12:36:04 GMT
Received: (from aburakov@localhost)
by sivswdev01.ir.intel.com with LOCAL id w2MCa4q4029462;
Thu, 22 Mar 2018 12:36:04 GMT
From: Anatoly Burakov <anatoly.burakov@intel.com>
To: dev@dpdk.org
Cc: Bruce Richardson <bruce.richardson@intel.com>, thomas@monjalon.net,
chaozhu@linux.vnet.ibm.com, gowrishankar.m@linux.vnet.ibm.com
Date: Thu, 22 Mar 2018 12:36:04 +0000
Message-Id: <9ff09ae09b3e9e1b149f73a693b552f8c94a01a2.1521722141.git.anatoly.burakov@intel.com>
X-Mailer: git-send-email 1.7.0.7
In-Reply-To: <9755302ee8a1439a152408d4b1c69cc1012153e4.1521715978.git.anatoly.burakov@intel.com>
References: <9755302ee8a1439a152408d4b1c69cc1012153e4.1521715978.git.anatoly.burakov@intel.com>
Subject: [dpdk-dev] [PATCH v6] eal: provide API for querying valid socket
id's
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://dpdk.org/ml/options/dev>,
<mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://dpdk.org/ml/listinfo/dev>,
<mailto:dev-request@dpdk.org?subject=subscribe>
X-List-Received-Date: Thu, 22 Mar 2018 12:36:09 -0000
During lcore scan, find all socket ID's and store them, and
provide public API to query valid socket id's. This will break
the ABI, so bump ABI version.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
---
Notes:
v6:
- Fixed meson ABI version header
v5:
- Move API to experimental
- Store list of valid socket id's instead of simply
recording the biggest one
v4:
- Remove backwards ABI compatibility, bump ABI instead
v3:
- Added ABI compatibility
v2:
- checkpatch changes
- check socket before deciding if the core is not to be used
lib/librte_eal/bsdapp/eal/Makefile | 2 +-
lib/librte_eal/common/eal_common_lcore.c | 75 ++++++++++++++++++++++++++-----
lib/librte_eal/common/include/rte_eal.h | 3 ++
lib/librte_eal/common/include/rte_lcore.h | 30 +++++++++++++
lib/librte_eal/linuxapp/eal/Makefile | 2 +-
lib/librte_eal/meson.build | 2 +-
lib/librte_eal/rte_eal_version.map | 2 +
7 files changed, 101 insertions(+), 15 deletions(-)
diff --git a/lib/librte_eal/bsdapp/eal/Makefile b/lib/librte_eal/bsdapp/eal/Makefile
index dd455e6..ed1d17b 100644
--- a/lib/librte_eal/bsdapp/eal/Makefile
+++ b/lib/librte_eal/bsdapp/eal/Makefile
@@ -21,7 +21,7 @@ LDLIBS += -lgcc_s
EXPORT_MAP := ../../rte_eal_version.map
-LIBABIVER := 6
+LIBABIVER := 7
# specific to bsdapp exec-env
SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) := eal.c
diff --git a/lib/librte_eal/common/eal_common_lcore.c b/lib/librte_eal/common/eal_common_lcore.c
index 7724fa4..50d9f82 100644
--- a/lib/librte_eal/common/eal_common_lcore.c
+++ b/lib/librte_eal/common/eal_common_lcore.c
@@ -7,6 +7,7 @@
#include <string.h>
#include <dirent.h>
+#include <rte_errno.h>
#include <rte_log.h>
#include <rte_eal.h>
#include <rte_lcore.h>
@@ -16,6 +17,19 @@
#include "eal_private.h"
#include "eal_thread.h"
+static int
+socket_id_cmp(const void *a, const void *b)
+{
+ const int *lcore_id_a = a;
+ const int *lcore_id_b = b;
+
+ if (*lcore_id_a < *lcore_id_b)
+ return -1;
+ if (*lcore_id_a > *lcore_id_b)
+ return 1;
+ return 0;
+}
+
/*
* Parse /sys/devices/system/cpu to get the number of physical and logical
* processors on the machine. The function will fill the cpu_info
@@ -28,6 +42,8 @@ rte_eal_cpu_init(void)
struct rte_config *config = rte_eal_get_configuration();
unsigned lcore_id;
unsigned count = 0;
+ unsigned int socket_id, prev_socket_id;
+ int lcore_to_socket_id[RTE_MAX_LCORE];
/*
* Parse the maximum set of logical cores, detect the subset of running
@@ -39,6 +55,19 @@ rte_eal_cpu_init(void)
/* init cpuset for per lcore config */
CPU_ZERO(&lcore_config[lcore_id].cpuset);
+ /* find socket first */
+ socket_id = eal_cpu_socket_id(lcore_id);
+ if (socket_id >= RTE_MAX_NUMA_NODES) {
+#ifdef RTE_EAL_ALLOW_INV_SOCKET_ID
+ socket_id = 0;
+#else
+ RTE_LOG(ERR, EAL, "Socket ID (%u) is greater than RTE_MAX_NUMA_NODES (%d)\n",
+ socket_id, RTE_MAX_NUMA_NODES);
+ return -1;
+#endif
+ }
+ lcore_to_socket_id[lcore_id] = socket_id;
+
/* in 1:1 mapping, record related cpu detected state */
lcore_config[lcore_id].detected = eal_cpu_detected(lcore_id);
if (lcore_config[lcore_id].detected == 0) {
@@ -54,18 +83,7 @@ rte_eal_cpu_init(void)
config->lcore_role[lcore_id] = ROLE_RTE;
lcore_config[lcore_id].core_role = ROLE_RTE;
lcore_config[lcore_id].core_id = eal_cpu_core_id(lcore_id);
- lcore_config[lcore_id].socket_id = eal_cpu_socket_id(lcore_id);
- if (lcore_config[lcore_id].socket_id >= RTE_MAX_NUMA_NODES) {
-#ifdef RTE_EAL_ALLOW_INV_SOCKET_ID
- lcore_config[lcore_id].socket_id = 0;
-#else
- RTE_LOG(ERR, EAL, "Socket ID (%u) is greater than "
- "RTE_MAX_NUMA_NODES (%d)\n",
- lcore_config[lcore_id].socket_id,
- RTE_MAX_NUMA_NODES);
- return -1;
-#endif
- }
+ lcore_config[lcore_id].socket_id = socket_id;
RTE_LOG(DEBUG, EAL, "Detected lcore %u as "
"core %u on socket %u\n",
lcore_id, lcore_config[lcore_id].core_id,
@@ -79,5 +97,38 @@ rte_eal_cpu_init(void)
RTE_MAX_LCORE);
RTE_LOG(INFO, EAL, "Detected %u lcore(s)\n", config->lcore_count);
+ /* sort all socket id's in ascending order */
+ qsort(lcore_to_socket_id, RTE_DIM(lcore_to_socket_id),
+ sizeof(lcore_to_socket_id[0]), socket_id_cmp);
+
+ prev_socket_id = -1;
+ config->numa_node_count = 0;
+ for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
+ socket_id = lcore_to_socket_id[lcore_id];
+ if (socket_id != prev_socket_id)
+ config->numa_nodes[config->numa_node_count++] =
+ socket_id;
+ prev_socket_id = socket_id;
+ }
+ RTE_LOG(INFO, EAL, "Detected %u NUMA nodes\n", config->numa_node_count);
+
return 0;
}
+
+unsigned int __rte_experimental
+rte_num_socket_ids(void)
+{
+ const struct rte_config *config = rte_eal_get_configuration();
+ return config->numa_node_count;
+}
+
+int __rte_experimental
+rte_socket_id_by_idx(unsigned int idx)
+{
+ const struct rte_config *config = rte_eal_get_configuration();
+ if (idx >= config->numa_node_count) {
+ rte_errno = EINVAL;
+ return -1;
+ }
+ return config->numa_nodes[idx];
+}
diff --git a/lib/librte_eal/common/include/rte_eal.h b/lib/librte_eal/common/include/rte_eal.h
index 93ca4cc..6109472 100644
--- a/lib/librte_eal/common/include/rte_eal.h
+++ b/lib/librte_eal/common/include/rte_eal.h
@@ -57,6 +57,9 @@ enum rte_proc_type_t {
struct rte_config {
uint32_t master_lcore; /**< Id of the master lcore */
uint32_t lcore_count; /**< Number of available logical cores. */
+ uint32_t numa_node_count; /**< Number of detected NUMA nodes. */
+ uint32_t numa_nodes[RTE_MAX_NUMA_NODES];
+ /**< List of detected numa nodes. */
uint32_t service_lcore_count;/**< Number of available service cores. */
enum rte_lcore_role_t lcore_role[RTE_MAX_LCORE]; /**< State of cores. */
diff --git a/lib/librte_eal/common/include/rte_lcore.h b/lib/librte_eal/common/include/rte_lcore.h
index 0472220..c6511a9 100644
--- a/lib/librte_eal/common/include/rte_lcore.h
+++ b/lib/librte_eal/common/include/rte_lcore.h
@@ -132,6 +132,36 @@ rte_lcore_index(int lcore_id)
unsigned rte_socket_id(void);
/**
+ * Return number of physical sockets detected on the system.
+ *
+ * Note that number of nodes may not be correspondent to their physical id's:
+ * for example, a system may report two socket id's, but the actual socket id's
+ * may be 0 and 8.
+ *
+ * @return
+ * the number of physical sockets as recognized by EAL
+ */
+unsigned int __rte_experimental
+rte_num_socket_ids(void);
+
+/**
+ * Return socket id with a particular index.
+ *
+ * This will return socket id at a particular position in list of all detected
+ * physical socket id's. For example, on a machine with sockets [0, 8], passing
+ * 1 as a parameter will return 8.
+ *
+ * @param idx
+ * index of physical socket id to return
+ *
+ * @return
+ * - physical socket id as recognized by EAL
+ * - -1 on error, with errno set to EINVAL
+ */
+int __rte_experimental
+rte_socket_id_by_idx(unsigned int idx);
+
+/**
* Get the ID of the physical socket of the specified lcore
*
* @param lcore_id
diff --git a/lib/librte_eal/linuxapp/eal/Makefile b/lib/librte_eal/linuxapp/eal/Makefile
index 7e5bbe8..b9c7727 100644
--- a/lib/librte_eal/linuxapp/eal/Makefile
+++ b/lib/librte_eal/linuxapp/eal/Makefile
@@ -10,7 +10,7 @@ ARCH_DIR ?= $(RTE_ARCH)
EXPORT_MAP := ../../rte_eal_version.map
VPATH += $(RTE_SDK)/lib/librte_eal/common/arch/$(ARCH_DIR)
-LIBABIVER := 6
+LIBABIVER := 7
VPATH += $(RTE_SDK)/lib/librte_eal/common
diff --git a/lib/librte_eal/meson.build b/lib/librte_eal/meson.build
index d9ba385..8f0ce1b 100644
--- a/lib/librte_eal/meson.build
+++ b/lib/librte_eal/meson.build
@@ -43,7 +43,7 @@ else
error('unsupported system type @0@'.format(hostmachine.system()))
endif
-version = 6 # the version of the EAL API
+version = 7 # the version of the EAL API
allow_experimental_apis = true
deps += 'compat'
cflags += '-D_GNU_SOURCE'
diff --git a/lib/librte_eal/rte_eal_version.map b/lib/librte_eal/rte_eal_version.map
index 1d88437..6a4c355 100644
--- a/lib/librte_eal/rte_eal_version.map
+++ b/lib/librte_eal/rte_eal_version.map
@@ -229,6 +229,8 @@ EXPERIMENTAL {
rte_mp_request;
rte_mp_request_async;
rte_mp_reply;
+ rte_num_socket_ids;
+ rte_socket_id_by_idx;
rte_service_attr_get;
rte_service_attr_reset_all;
rte_service_component_register;
--
2.7.4