[ovs-dev] [PATCH v2 1/3] system-dpdk: cleanup stale hugepage files after tests

Gowrishankar Muthukrishnan gmuthukr at redhat.com
Wed May 27 02:10:06 UTC 2020


I tried --in-memory but finding issues with existing tests unable to pass
mainly, being unable to ping between kernel veth and dpdk tap. Problem
could be because of implicit --no-shconf ? .Can we handle this separately
as it is blocking other patches ? Your opinions please.

Regards.

On Thu, May 21, 2020 at 7:54 PM Ilya Maximets <i.maximets at ovn.org> wrote:

> On 5/16/20 7:53 AM, Gowrishankar Muthukrishnan wrote:
> > After dpdk tests completes, cleaning up hugepage map files
> > created by tests is helpful to release used memory into
> > hugepage memory allocator.
> >
> > Signed-off-by: Gowrishankar Muthukrishnan <gmuthukr at redhat.com>
> > ---
> >  tests/system-dpdk-macros.at | 13 +++++++++++++
> >  tests/system-dpdk.at        |  7 +++++++
> >  2 files changed, 20 insertions(+)
> >
> > diff --git a/tests/system-dpdk-macros.at b/tests/system-dpdk-macros.at
> > index c6708ca..d3a3aea 100644
> > --- a/tests/system-dpdk-macros.at
> > +++ b/tests/system-dpdk-macros.at
> > @@ -63,3 +63,16 @@ m4_define([OVS_DPDK_START],
> >     AT_CAPTURE_FILE([ovs-vswitchd.log])
> >     on_exit "kill_ovs_vswitchd `cat ovs-vswitchd.pid`"
> >  ])
> > +
> > +
> > +# OVS_DPDK_HUGEPAGE_CLEANUP([file])
> > +#
> > +# Cleanup system for stale hugepages.
> > +#
> > +m4_define([OVS_DPDK_HUGEPAGE_CLEANUP],
> > +  [dnl Cleanup mapping files in hugetlbfs mount point
> > +   AT_CHECK([cat /proc/mounts | grep hugetlbfs], [], [stdout])
> > +   AT_CHECK([cut -d ' ' -f 2  stdout], [], [stdout])
> > +   AT_CHECK([rm -f $(cat stdout)/$1], [], [])
> > +
> > +])
>
> Can we just use --in-memory option and avoid having files at all?
>
> General note about sending patches: Please, don't send new versions in
> reply
> to the previous one.  This messes up mailboxes.
>
> Best regards, Ilya Maximets.
>
>

-- 
Gowrishankar M


More information about the dev mailing list