From: gowrishankar muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
To: "Burakov, Anatoly" <anatoly.burakov@intel.com>
Cc: dev@dpdk.org, Bruce Richardson <bruce.richardson@intel.com>,
Chao Zhu <chaozhu@linux.vnet.ibm.com>
Subject: Re: [dpdk-dev] [PATCH 18.05 v4] eal: add function to return number of detected sockets
Date: Thu, 22 Mar 2018 10:46:09 +0530 [thread overview]
Message-ID: <da9b1d2f-0384-55f0-cf31-91f226a775f5@linux.vnet.ibm.com> (raw)
In-Reply-To: <db9677b5-63f5-0d92-6c2c-4cefe4b7b801@intel.com>
On Wednesday 21 March 2018 03:54 PM, Burakov, Anatoly wrote:
>
>>> + config->numa_node_count = max_socket_id + 1;
>>
>> In some IBM servers, socket ID number does not seem to be in
>> sequence. For an instance, 0 and 8 for a 2 node server.
>>
>> In this case, numa_node_count would mislead users if wrongly
>> understood by its variable name IMO (see below)
>>> + RTE_LOG(INFO, EAL, "Detected %u NUMA nodes\n",
>>> config->numa_node_count);
>>
>> For an instance, reading above message would tell 'EAL detected 8
>> nodes' in my server, but actually there are only two nodes.
>>
>> Could its name better be 'numa_node_id_max' ?. Also, we store in
>> actual count of numa nodes in _count variable.
>>
>> Also, there could be a case when there is no local memory available
>> to a numa node too.
>>
>> Thanks,
>> Gowrishankar
>
> The point of this patchset is to (pre)allocate memory only on existing
> sockets.
>
> If we don't know how many sockets there are, we are forced to
> preallocate VA space per each *possible* NUMA node - that is, reserve
> e.g. 8x128G of memory, 6 of which will go unused on a 2-socket system.
> We can't know if there is no memory on socket in advance, but we can
> at least avoid preallocating VA space for sockets that don't exist in
> the first place.
>
Sounds good Anatoly.
May be, sysfs/ might help to confirm if a numa node has local memory ?.
Anyway, for the context of this particular patch (return numa nodes),
below approach you mentioned is good.
> How about we store all possible socket id's instead? e.g. something like:
>
> static int numa_node_ids[MAX_NUMA_NODES];
> <...>
> int rte_eal_cpu_init() {
> int sockets[RTE_MAX_LCORE];
> <...>
> for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> core_to_socket[lcore_id] = socket;
sockets[lcore_id] = eal_cpu_socket_id(lcore_id);
> }
> <...>
> qsort(sockets);
> <...>
> // store all unique sockets in numa_node_ids in ascending order
Just thinking that, is there a purpose of retaining a numa ID which does
not have local memory attached ?
but sockets[] is suppose to reflect all available nodes though (and
assuming, its calling place to ensure
for the existence of numa local memory).
> }
> <...>
>
> on a 2 socket system we then get:
>
> rte_num_sockets() => return 2
> rte_get_socket_id(int idx) => return numa_node_ids[idx]
rte_get_socket_mem(idx) might help to validate for local memory existence ?
>
> Would that be suitable?
>
Thanks,
Gowrishankar
next prev parent reply other threads:[~2018-03-22 5:16 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-12-22 11:58 [dpdk-dev] [PATCH] " Anatoly Burakov
2017-12-22 12:41 ` [dpdk-dev] [PATCH v2] " Anatoly Burakov
2018-01-11 22:20 ` Thomas Monjalon
2018-01-12 11:44 ` Burakov, Anatoly
2018-01-12 11:50 ` Thomas Monjalon
2018-01-16 11:56 ` Burakov, Anatoly
2018-01-16 12:20 ` Thomas Monjalon
2018-01-16 15:05 ` Burakov, Anatoly
2018-01-16 17:34 ` Thomas Monjalon
2018-01-16 17:38 ` Burakov, Anatoly
2018-01-16 18:26 ` Thomas Monjalon
2018-01-16 17:53 ` [dpdk-dev] [PATCH] doc: add ABI change notice for numa_node_count in eal Anatoly Burakov
2018-01-23 10:39 ` Mcnamara, John
2018-02-07 10:10 ` Jerin Jacob
2018-02-09 14:42 ` Bruce Richardson
2018-02-14 0:04 ` Thomas Monjalon
2018-02-14 14:25 ` Thomas Monjalon
2018-02-12 16:00 ` Jonas Pfefferle
[not found] ` <cover.1517848624.git.anatoly.burakov@intel.com>
2018-02-05 16:37 ` [dpdk-dev] [PATCH v3] eal: add function to return number of detected sockets Anatoly Burakov
2018-02-05 17:39 ` Burakov, Anatoly
2018-02-05 22:45 ` Thomas Monjalon
2018-02-06 9:28 ` Burakov, Anatoly
2018-02-06 9:47 ` Thomas Monjalon
2018-02-07 9:58 ` [dpdk-dev] [PATCH 18.05 v4] Add " Anatoly Burakov
2018-02-07 9:58 ` [dpdk-dev] [PATCH 18.05 v4] eal: add " Anatoly Burakov
2018-03-08 12:12 ` Bruce Richardson
2018-03-08 14:38 ` Burakov, Anatoly
2018-03-09 16:32 ` Bruce Richardson
2018-03-20 22:43 ` Thomas Monjalon
2018-03-21 4:59 ` gowrishankar muthukrishnan
2018-03-21 10:24 ` Burakov, Anatoly
2018-03-22 5:16 ` gowrishankar muthukrishnan [this message]
2018-03-22 9:04 ` Burakov, Anatoly
2018-03-22 10:58 ` [dpdk-dev] [PATCH v5] eal: provide API for querying valid socket id's Anatoly Burakov
2018-03-22 11:45 ` Burakov, Anatoly
2018-03-22 12:36 ` [dpdk-dev] [PATCH v6] " Anatoly Burakov
2018-03-22 17:07 ` gowrishankar muthukrishnan
2018-03-27 16:24 ` Thomas Monjalon
2018-03-31 13:35 ` Burakov, Anatoly
2018-04-02 15:27 ` Thomas Monjalon
2018-03-31 17:08 ` [dpdk-dev] [PATCH v7] " Anatoly Burakov
2018-04-04 22:31 ` 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=da9b1d2f-0384-55f0-cf31-91f226a775f5@linux.vnet.ibm.com \
--to=gowrishankar.m@linux.vnet.ibm.com \
--cc=anatoly.burakov@intel.com \
--cc=bruce.richardson@intel.com \
--cc=chaozhu@linux.vnet.ibm.com \
--cc=dev@dpdk.org \
/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).