DPDK patches and discussions
 help / color / mirror / Atom feed
From: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
To: Ilya Maximets <i.maximets@samsung.com>,
	dev@dpdk.org, David Marchand <david.marchand@6wind.com>
Cc: Heetae Ahn <heetae82.ahn@samsung.com>,
	Yuanhan Liu <yuanhan.liu@linux.intel.com>,
	Jianfeng Tan <jianfeng.tan@intel.com>,
	Neil Horman <nhorman@tuxdriver.com>,
	Yulong Pei <yulong.pei@intel.com>,
	stable@dpdk.org, Thomas Monjalon <thomas.monjalon@6wind.com>,
	Bruce Richardson <bruce.richardson@intel.com>
Subject: Re: [dpdk-dev] [PATCH] mem: balanced allocation of hugepages
Date: Wed, 8 Mar 2017 13:46:26 +0000	[thread overview]
Message-ID: <f50d0fa1-9530-436c-d532-0e6123f4e06d@intel.com> (raw)
In-Reply-To: <50517d4c-5174-f4b2-e77e-143f7aac2c00@samsung.com>

Hi Ilya,

I have done similar tests and as you already pointed out, 'numactl 
--interleave' does not seem to work as expected.
I have also checked that the issue can be reproduced with quota limit on 
hugetlbfs mount point.

I would be inclined towards *adding libnuma as dependency* to DPDK to 
make memory allocation a bit more reliable.

Currently at a high level regarding hugepages per numa node:
1) Try to map all free hugepages. The total number of mapped hugepages 
depends if there were any limits, such as cgroups or quota in mount point.
2) Find out numa node of each hugepage.
3) Check if we have enough hugepages for requested memory in each numa 
socket/node.

Using libnuma we could try to allocate hugepages per numa:
1) Try to map as many hugepages from numa 0.
2) Check if we have enough hugepages for requested memory in numa 0.
3) Try to map as many hugepages from numa 1.
4) Check if we have enough hugepages for requested memory in numa 1.

This approach would improve failing scenarios caused by limits but It 
would still not fix issues regarding non-contiguous hugepages (worst 
case each hugepage is a memseg).
The non-contiguous hugepages issues are not as critical now that 
mempools can span over multiple memsegs/hugepages, but it is still a 
problem for any other library requiring big chunks of memory.

Potentially if we were to add an option such as 'iommu-only' when all 
devices are bound to vfio-pci, we could have a reliable way to allocate 
hugepages by just requesting the number of pages from each numa.

Thoughts?

Sergio

On 06/03/2017 09:34, Ilya Maximets wrote:
> Hi all.
>
> So, what about this change?
>
> Best regards, Ilya Maximets.
>
> On 16.02.2017 16:01, Ilya Maximets wrote:
>> 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 in a round-robin fashion.
>> In this case all allocated pages will be fairly distributed
>> between all requested nodes.
>>
>> New config option RTE_LIBRTE_EAL_NUMA_AWARE_HUGEPAGES
>> introduced and disabled by default because of external
>> dependency from libnuma.
>>
>> Cc: <stable@dpdk.org>
>> Fixes: 77988fc08dc5 ("mem: fix allocating all free hugepages")
>>
>> Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
>> ---
>>   config/common_base                       |  1 +
>>   lib/librte_eal/Makefile                  |  4 ++
>>   lib/librte_eal/linuxapp/eal/eal_memory.c | 66 ++++++++++++++++++++++++++++++++
>>   mk/rte.app.mk                            |  3 ++
>>   4 files changed, 74 insertions(+)
>>

  reply	other threads:[~2017-03-08 13:46 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 ` 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 [this message]
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             ` [dpdk-dev] [PATCH v3 1/2] mem: balanced " Ilya Maximets
     [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=f50d0fa1-9530-436c-d532-0e6123f4e06d@intel.com \
    --to=sergio.gonzalez.monroy@intel.com \
    --cc=bruce.richardson@intel.com \
    --cc=david.marchand@6wind.com \
    --cc=dev@dpdk.org \
    --cc=heetae82.ahn@samsung.com \
    --cc=i.maximets@samsung.com \
    --cc=jianfeng.tan@intel.com \
    --cc=nhorman@tuxdriver.com \
    --cc=stable@dpdk.org \
    --cc=thomas.monjalon@6wind.com \
    --cc=yuanhan.liu@linux.intel.com \
    --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).