DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Du, Fan" <fan.du@intel.com>
To: "dev@dpdk.org" <dev@dpdk.org>
Subject: [dpdk-dev] huge page didn't get freed when dpdk-app exit(or killed)
Date: Mon, 15 Jun 2015 08:46:49 +0000	[thread overview]
Message-ID: <5A90DA2E42F8AE43BC4A093BF0678848E6B959@SHSMSX104.ccr.corp.intel.com> (raw)

Hi,

I play OpenvSwitch dpdk, when ovs-vswitchd got killed, it won't free up any allocated huge pages, shown by below:
# tail /proc/meminfo
HardwareCorrupted:     0 kB
AnonHugePages:    106496 kB
HugePages_Total:       8
HugePages_Free:        6
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:    1048576 kB

DPDK setup hugepage by below steps, in function rte_eal_hugepage_init
*  1. map N huge pages in separate files in hugetlbfs
 *  2. find associated physical addr
 *  3. find associated NUMA socket ID
 *  4. sort all huge pages by physical address
 *  5. remap these N huge pages in the correct order
 *  6. unmap the first mapping
 *  7. fill memsegs in configuration with contiguous zones

So my env shows below when allocated 2G huge pages.
# ls /dev/hugepages/
libvirt  rtemap_4  rtemap_7

So in scenario where dpdk app is stop, but hugepage will be used by other non-dpdk app,
Would current dpdk huge page behavior(not freed huge page when app exited) be a problem?

I found out this patch: http://dpdk.org/dev/patchwork/patch/993/ tries to provide an interface,
It didn't get merged with additional implementation modification.
Then is there any follow ups for this issue? A pointer will be much appreciated!
Thanks!

                 reply	other threads:[~2015-06-15  8:46 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=5A90DA2E42F8AE43BC4A093BF0678848E6B959@SHSMSX104.ccr.corp.intel.com \
    --to=fan.du@intel.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).