From: Ilya Maximets <i.maximets@samsung.com>
To: dev@dpdk.org, David Marchand <david.marchand@6wind.com>,
Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>,
Thomas Monjalon <thomas@monjalon.net>
Cc: Heetae Ahn <heetae82.ahn@samsung.com>,
Yuanhan Liu <yliu@fridaylinux.org>,
Jianfeng Tan <jianfeng.tan@intel.com>,
Neil Horman <nhorman@tuxdriver.com>,
Yulong Pei <yulong.pei@intel.com>,
Ilya Maximets <i.maximets@samsung.com>
Subject: [dpdk-dev] [PATCH v3 1/2] mem: balanced allocation of hugepages
Date: Tue, 06 Jun 2017 09:22:17 +0300 [thread overview]
Message-ID: <1496730138-32056-2-git-send-email-i.maximets@samsung.com> (raw)
In-Reply-To: <1496730138-32056-1-git-send-email-i.maximets@samsung.com>
Currently EAL allocates hugepages one by one not paying attention
from which NUMA node allocation was done.
Such behaviour leads to allocation failure if number of available
hugepages for application limited by cgroups or hugetlbfs and
memory requested not only from the first socket.
Example:
# 90 x 1GB hugepages availavle in a system
cgcreate -g hugetlb:/test
# Limit to 32GB of hugepages
cgset -r hugetlb.1GB.limit_in_bytes=34359738368 test
# Request 4GB from each of 2 sockets
cgexec -g hugetlb:test testpmd --socket-mem=4096,4096 ...
EAL: SIGBUS: Cannot mmap more hugepages of size 1024 MB
EAL: 32 not 90 hugepages of size 1024 MB allocated
EAL: Not enough memory available on socket 1!
Requested: 4096MB, available: 0MB
PANIC in rte_eal_init():
Cannot init memory
This happens beacause all allocated pages are
on socket 0.
Fix this issue by setting mempolicy MPOL_PREFERRED for each hugepage
to one of requested nodes using following schema:
1) Allocate essential hugepages:
1.1) Allocate as many hugepages from numa N to
only fit requested memory for this numa.
1.2) repeat 1.1 for all numa nodes.
2) Try to map all remaining free hugepages in a round-robin
fashion.
3) Sort pages and choose the most suitable.
In this case all essential memory will be allocated and all remaining
pages will be fairly distributed between all requested nodes.
libnuma added as a general dependency for EAL.
Fixes: 77988fc08dc5 ("mem: fix allocating all free hugepages")
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
---
lib/librte_eal/Makefile | 2 +
lib/librte_eal/linuxapp/eal/eal_memory.c | 87 ++++++++++++++++++++++++++++++--
mk/rte.app.mk | 1 +
3 files changed, 86 insertions(+), 4 deletions(-)
diff --git a/lib/librte_eal/Makefile b/lib/librte_eal/Makefile
index 5690bb4..0a1af3a 100644
--- a/lib/librte_eal/Makefile
+++ b/lib/librte_eal/Makefile
@@ -37,4 +37,6 @@ DEPDIRS-linuxapp := common
DIRS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += bsdapp
DEPDIRS-bsdapp := common
+LDLIBS += -lnuma
+
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
index 9c9baf6..35e5bce 100644
--- a/lib/librte_eal/linuxapp/eal/eal_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
@@ -54,6 +54,7 @@
#include <sys/time.h>
#include <signal.h>
#include <setjmp.h>
+#include <numaif.h>
#include <rte_log.h>
#include <rte_memory.h>
@@ -358,6 +359,19 @@ static int huge_wrap_sigsetjmp(void)
return sigsetjmp(huge_jmpenv, 1);
}
+#ifndef ULONG_SIZE
+#define ULONG_SIZE sizeof(unsigned long)
+#endif
+#ifndef ULONG_BITS
+#define ULONG_BITS (ULONG_SIZE * CHAR_BIT)
+#endif
+#ifndef DIV_ROUND_UP
+#define DIV_ROUND_UP(n, d) (((n) + (d) - 1) / (d))
+#endif
+#ifndef BITS_TO_LONGS
+#define BITS_TO_LONGS(nr) DIV_ROUND_UP(nr, ULONG_SIZE)
+#endif
+
/*
* Mmap all hugepages of hugepage table: it first open a file in
* hugetlbfs, then mmap() hugepage_sz data in it. If orig is set, the
@@ -366,18 +380,71 @@ static int huge_wrap_sigsetjmp(void)
* map continguous physical blocks in contiguous virtual blocks.
*/
static unsigned
-map_all_hugepages(struct hugepage_file *hugepg_tbl,
- struct hugepage_info *hpi, int orig)
+map_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi,
+ uint64_t *essential_memory, int orig)
{
int fd;
unsigned i;
void *virtaddr;
void *vma_addr = NULL;
size_t vma_len = 0;
+ unsigned long nodemask[BITS_TO_LONGS(RTE_MAX_NUMA_NODES)] = {0UL};
+ unsigned long maxnode = 0;
+ int node_id = -1;
+
+ if (orig) {
+ for (i = 0; i < RTE_MAX_NUMA_NODES; i++)
+ if (internal_config.socket_mem[i])
+ maxnode = i + 1;
+ }
for (i = 0; i < hpi->num_pages[0]; i++) {
uint64_t hugepage_sz = hpi->hugepage_sz;
+ if (maxnode) {
+ unsigned int j;
+
+ for (j = 0; j < RTE_MAX_NUMA_NODES; j++)
+ if (essential_memory[j])
+ break;
+
+ if (j == RTE_MAX_NUMA_NODES) {
+ node_id = (node_id + 1) % RTE_MAX_NUMA_NODES;
+ while (!internal_config.socket_mem[node_id]) {
+ node_id++;
+ node_id %= RTE_MAX_NUMA_NODES;
+ }
+ } else {
+ node_id = j;
+ if (essential_memory[j] < hugepage_sz)
+ essential_memory[j] = 0;
+ else
+ essential_memory[j] -= hugepage_sz;
+ }
+
+ nodemask[node_id / ULONG_BITS] =
+ 1UL << (node_id % ULONG_BITS);
+
+ RTE_LOG(DEBUG, EAL,
+ "Setting policy MPOL_PREFERRED for socket %d\n",
+ node_id);
+ /*
+ * Due to old linux kernel bug (feature?) we have to
+ * increase maxnode by 1. It will be unconditionally
+ * decreased back to normal value inside the syscall
+ * handler.
+ */
+ if (set_mempolicy(MPOL_PREFERRED,
+ nodemask, maxnode + 1) < 0) {
+ RTE_LOG(ERR, EAL,
+ "Failed to set policy MPOL_PREFERRED: "
+ "%s\n", strerror(errno));
+ return i;
+ }
+
+ nodemask[node_id / ULONG_BITS] = 0UL;
+ }
+
if (orig) {
hugepg_tbl[i].file_id = i;
hugepg_tbl[i].size = hugepage_sz;
@@ -488,6 +555,9 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl,
vma_len -= hugepage_sz;
}
+ if (maxnode && set_mempolicy(MPOL_DEFAULT, NULL, 0) < 0)
+ RTE_LOG(ERR, EAL, "Failed to set mempolicy MPOL_DEFAULT\n");
+
return i;
}
@@ -572,6 +642,9 @@ find_numasocket(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi)
if (hugepg_tbl[i].orig_va == va) {
hugepg_tbl[i].socket_id = socket_id;
hp_count++;
+ RTE_LOG(DEBUG, EAL,
+ "Hugepage %s is on socket %d\n",
+ hugepg_tbl[i].filepath, socket_id);
}
}
}
@@ -1010,6 +1083,11 @@ rte_eal_hugepage_init(void)
huge_register_sigbus();
+ /* make a copy of socket_mem, needed for balanced allocation. */
+ for (i = 0; i < RTE_MAX_NUMA_NODES; i++)
+ memory[i] = internal_config.socket_mem[i];
+
+
/* map all hugepages and sort them */
for (i = 0; i < (int)internal_config.num_hugepage_sizes; i ++){
unsigned pages_old, pages_new;
@@ -1027,7 +1105,8 @@ rte_eal_hugepage_init(void)
/* map all hugepages available */
pages_old = hpi->num_pages[0];
- pages_new = map_all_hugepages(&tmp_hp[hp_offset], hpi, 1);
+ pages_new = map_all_hugepages(&tmp_hp[hp_offset], hpi,
+ memory, 1);
if (pages_new < pages_old) {
RTE_LOG(DEBUG, EAL,
"%d not %d hugepages of size %u MB allocated\n",
@@ -1070,7 +1149,7 @@ rte_eal_hugepage_init(void)
sizeof(struct hugepage_file), cmp_physaddr);
/* remap all hugepages */
- if (map_all_hugepages(&tmp_hp[hp_offset], hpi, 0) !=
+ if (map_all_hugepages(&tmp_hp[hp_offset], hpi, NULL, 0) !=
hpi->num_pages[0]) {
RTE_LOG(ERR, EAL, "Failed to remap %u MB pages\n",
(unsigned)(hpi->hugepage_sz / 0x100000));
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index bcaf1b3..b208e88 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -186,6 +186,7 @@ ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),n)
# The static libraries do not know their dependencies.
# So linking with static library requires explicit dependencies.
_LDLIBS-$(CONFIG_RTE_LIBRTE_EAL) += -lrt
+_LDLIBS-$(CONFIG_RTE_LIBRTE_EAL) += -lnuma
_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED) += -lm
_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED) += -lrt
_LDLIBS-$(CONFIG_RTE_LIBRTE_METER) += -lm
--
2.7.4
next prev parent reply other threads:[~2017-06-06 6:22 UTC|newest]
Thread overview: 99+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CGME20170216130139eucas1p2512567d6f5db9eaac5ee840b56bf920a@eucas1p2.samsung.com>
2017-02-16 13:01 ` [dpdk-dev] [PATCH] " Ilya Maximets
2017-02-16 13:26 ` Tan, Jianfeng
2017-02-16 13:55 ` Ilya Maximets
2017-02-16 13:57 ` Ilya Maximets
2017-02-16 13:31 ` Bruce Richardson
2017-03-06 9:34 ` Ilya Maximets
2017-03-08 13:46 ` Sergio Gonzalez Monroy
2017-03-09 12:57 ` Ilya Maximets
2017-03-27 13:01 ` Sergio Gonzalez Monroy
2017-03-27 14:43 ` Ilya Maximets
2017-04-07 15:14 ` Ilya Maximets
2017-04-07 15:44 ` Thomas Monjalon
2017-04-10 7:11 ` Ilya Maximets
2017-04-10 7:51 ` Sergio Gonzalez Monroy
2017-04-10 8:05 ` Ilya Maximets
[not found] ` <CGME20170410080425eucas1p27fd424ae58151f13b1a7a3723aa4ad1e@eucas1p2.samsung.com>
2017-04-10 8:04 ` [dpdk-dev] [PATCH v2] " Ilya Maximets
2017-04-10 10:03 ` Thomas Monjalon
[not found] ` <CGME20170606062227eucas1p2c49a95fb0fe11a4cadd5b4ceeb9712b1@eucas1p2.samsung.com>
2017-06-06 6:22 ` [dpdk-dev] [PATCH v3 0/2] Balanced " Ilya Maximets
[not found] ` <CGME20170606062232eucas1p11d2c304a28353d32b93ddfbd134d4da9@eucas1p1.samsung.com>
2017-06-06 6:22 ` Ilya Maximets [this message]
[not found] ` <CGME20170606062237eucas1p1de58fdde1bff816e480e50308804ba7a@eucas1p1.samsung.com>
2017-06-06 6:22 ` [dpdk-dev] [PATCH v3 2/2] config: enable vhost numa awareness by default Ilya Maximets
[not found] ` <CGME20170606081359eucas1p2f7eafa1abc346c5bb910c783df1d1520@eucas1p2.samsung.com>
2017-06-06 8:13 ` [dpdk-dev] [PATCH v4 0/2] Balanced allocation of hugepages Ilya Maximets
[not found] ` <CGME20170606081403eucas1p20c561b9177a51cfe58dd53b76cbfaaf7@eucas1p2.samsung.com>
2017-06-06 8:13 ` [dpdk-dev] [PATCH v4 1/2] mem: balanced " Ilya Maximets
[not found] ` <CGME20170606081409eucas1p2eed4a7dc49f1028c723f8c0a7a61fadf@eucas1p2.samsung.com>
2017-06-06 8:13 ` [dpdk-dev] [PATCH v4 2/2] config: enable vhost numa awareness by default Ilya Maximets
[not found] ` <CGME20170606133348eucas1p1cc5c3c05f88b2101c2ea47b26e0cac24@eucas1p1.samsung.com>
2017-06-06 13:33 ` [dpdk-dev] [PATCH v5 0/2] Balanced allocation of hugepages Ilya Maximets
[not found] ` <CGME20170606133352eucas1p13d1e860e996057a50a084f9365189e4d@eucas1p1.samsung.com>
2017-06-06 13:33 ` [dpdk-dev] [PATCH v5 1/2] mem: balanced " Ilya Maximets
[not found] ` <CGME20170606133354eucas1p284ae347e9ff07d6e8ab2bc09344ad1e5@eucas1p2.samsung.com>
2017-06-06 13:33 ` [dpdk-dev] [PATCH v5 2/2] config: enable vhost numa awareness by default Ilya Maximets
2017-06-08 11:21 ` [dpdk-dev] [PATCH v5 0/2] Balanced allocation of hugepages Ilya Maximets
2017-06-08 12:14 ` Bruce Richardson
2017-06-08 15:44 ` Sergio Gonzalez Monroy
2017-06-14 6:11 ` Ilya Maximets
2017-06-19 11:10 ` Hemant Agrawal
2017-06-20 13:07 ` Thomas Monjalon
2017-06-20 13:58 ` Ilya Maximets
2017-06-20 14:35 ` Thomas Monjalon
2017-06-20 14:58 ` Sergio Gonzalez Monroy
2017-06-20 15:41 ` Jerin Jacob
2017-06-20 15:51 ` Sergio Gonzalez Monroy
2017-06-21 8:14 ` Hemant Agrawal
2017-06-21 8:25 ` Sergio Gonzalez Monroy
2017-06-21 8:36 ` Ilya Maximets
2017-06-21 8:41 ` Jerin Jacob
2017-06-21 8:49 ` Thomas Monjalon
2017-06-21 9:27 ` Jerin Jacob
2017-06-21 9:58 ` Thomas Monjalon
2017-06-21 10:29 ` Jerin Jacob
2017-06-21 10:36 ` Ilya Maximets
2017-06-21 11:22 ` Jerin Jacob
2017-06-21 11:29 ` Thomas Monjalon
2017-06-27 9:13 ` Hemant Agrawal
2017-06-27 9:26 ` Thomas Monjalon
2017-06-27 9:48 ` Hemant Agrawal
[not found] ` <CGME20170621080434eucas1p18d3d4e4133c1cf885c849d022806408d@eucas1p1.samsung.com>
2017-06-21 8:04 ` [dpdk-dev] [PATCH v6 " Ilya Maximets
[not found] ` <CGME20170621080441eucas1p2dc01b29e7c8e4c1546ace6cd76ae51ff@eucas1p2.samsung.com>
2017-06-21 8:04 ` [dpdk-dev] [PATCH v6 1/2] mem: balanced " Ilya Maximets
2017-06-21 8:51 ` Thomas Monjalon
2017-06-21 8:58 ` Bruce Richardson
2017-06-21 9:25 ` Ilya Maximets
2017-06-21 9:34 ` Bruce Richardson
2017-06-21 9:28 ` Thomas Monjalon
[not found] ` <CGME20170621080448eucas1p28951fac6e4910cc599fe88d7edac9734@eucas1p2.samsung.com>
2017-06-21 8:04 ` [dpdk-dev] [PATCH v6 2/2] config: enable vhost numa awareness by default Ilya Maximets
[not found] ` <CGME20170621100837eucas1p1c570092cac733a66d939ca7ff04ac9e6@eucas1p1.samsung.com>
2017-06-21 10:08 ` [dpdk-dev] [PATCH v7 0/2] Balanced allocation of hugepages Ilya Maximets
[not found] ` <CGME20170621100841eucas1p1114078b1d8a38920c3633e9bddbabc02@eucas1p1.samsung.com>
2017-06-21 10:08 ` [dpdk-dev] [PATCH v7 1/2] mem: balanced " Ilya Maximets
[not found] ` <CGME20170621100845eucas1p2a457b1694d20de8e2d8126df679c43ae@eucas1p2.samsung.com>
2017-06-21 10:08 ` [dpdk-dev] [PATCH v7 2/2] config: enable vhost numa awareness by default Ilya Maximets
2017-06-27 9:20 ` Hemant Agrawal
2017-06-26 10:44 ` [dpdk-dev] [PATCH v7 0/2] Balanced allocation of hugepages Ilya Maximets
2017-06-26 14:07 ` Jerin Jacob
2017-06-26 15:33 ` Sergio Gonzalez Monroy
2017-06-27 8:42 ` Ilya Maximets
[not found] ` <CGME20170627084632eucas1p28133ee4b425b3938e2564fca03e1140b@eucas1p2.samsung.com>
2017-06-27 8:46 ` [dpdk-dev] [PATCH v8 " Ilya Maximets
[not found] ` <CGME20170627084637eucas1p2c591db905fa9f143fa5dbb3c08fae82f@eucas1p2.samsung.com>
2017-06-27 8:46 ` [dpdk-dev] [PATCH v8 1/2] mem: balanced " Ilya Maximets
[not found] ` <CGME20170627084641eucas1p182cac065efef74445ffa234a6dcbb23d@eucas1p1.samsung.com>
2017-06-27 8:46 ` [dpdk-dev] [PATCH v8 2/2] config: enable vhost numa awareness by default Ilya Maximets
2017-06-27 9:18 ` Hemant Agrawal
2017-06-27 9:21 ` Thomas Monjalon
2017-06-27 9:41 ` Hemant Agrawal
2017-06-27 9:59 ` Thomas Monjalon
2017-06-27 9:59 ` Jerin Jacob
2017-06-27 12:17 ` Hemant Agrawal
2017-06-27 12:45 ` Jerin Jacob
2017-06-27 13:00 ` Hemant Agrawal
2017-06-27 9:19 ` Thomas Monjalon
2017-06-27 10:26 ` Ilya Maximets
[not found] ` <CGME20170627102447eucas1p15a57bbaaf46944c0935d4ef71b55cd83@eucas1p1.samsung.com>
2017-06-27 10:24 ` [dpdk-dev] [PATCH v9 0/2] Balanced allocation of hugepages Ilya Maximets
[not found] ` <CGME20170627102451eucas1p2254d8679f70e261b9db9d2123aa80091@eucas1p2.samsung.com>
2017-06-27 10:24 ` [dpdk-dev] [PATCH v9 1/2] mem: balanced " Ilya Maximets
2017-06-28 10:30 ` Sergio Gonzalez Monroy
2017-06-29 5:32 ` Hemant Agrawal
2017-06-29 5:48 ` Ilya Maximets
2017-06-29 6:08 ` Ilya Maximets
[not found] ` <CGME20170627102454eucas1p14b2a1024d77158ad0bf40d62e6ad4365@eucas1p1.samsung.com>
2017-06-27 10:24 ` [dpdk-dev] [PATCH v9 2/2] config: enable vhost numa awareness by default Ilya Maximets
2017-06-29 5:31 ` Hemant Agrawal
[not found] ` <CGME20170629055928eucas1p17e823d821cfe95953bfa59dc9883ca4f@eucas1p1.samsung.com>
2017-06-29 5:59 ` [dpdk-dev] [PATCH v10 0/2] Balanced allocation of hugepages Ilya Maximets
[not found] ` <CGME20170629055933eucas1p1e5eba5f07850f63f9afbd48e6ca64c42@eucas1p1.samsung.com>
2017-06-29 5:59 ` [dpdk-dev] [PATCH v10 1/2] mem: balanced " Ilya Maximets
2017-06-29 7:03 ` Hemant Agrawal
[not found] ` <CGME20170629055940eucas1p1c9adcb26bec3ce5de97fe56753fd941a@eucas1p1.samsung.com>
2017-06-29 5:59 ` [dpdk-dev] [PATCH v10 2/2] config: enable vhost numa awareness by default Ilya Maximets
2017-06-30 15:50 ` Thomas Monjalon
2017-06-29 6:29 ` [dpdk-dev] [PATCH v10 0/2] Balanced allocation of hugepages Jerin Jacob
2017-06-30 8:36 ` Ilya Maximets
2017-06-30 16:12 ` [dpdk-dev] [PATCH v11 " Thomas Monjalon
2017-06-30 16:12 ` [dpdk-dev] [PATCH v11 1/2] mem: balanced " Thomas Monjalon
2017-06-30 16:12 ` [dpdk-dev] [PATCH v11 2/2] config: enable vhost NUMA awareness by default Thomas Monjalon
2017-07-01 10:59 ` [dpdk-dev] [PATCH v11 0/2] Balanced allocation of hugepages Thomas Monjalon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1496730138-32056-2-git-send-email-i.maximets@samsung.com \
--to=i.maximets@samsung.com \
--cc=david.marchand@6wind.com \
--cc=dev@dpdk.org \
--cc=heetae82.ahn@samsung.com \
--cc=jianfeng.tan@intel.com \
--cc=nhorman@tuxdriver.com \
--cc=sergio.gonzalez.monroy@intel.com \
--cc=thomas@monjalon.net \
--cc=yliu@fridaylinux.org \
--cc=yulong.pei@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).