No free hugepages reported in hugepages-1048576kB/d"])
+/EAL: No free .* hugepages reported/d"])
This will likely conflict with my patchset on system-dpdk, but this is
easy to fix.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Thu, Dec 2, 2021 at 1:18 PM Ilya Maximets wrote:
>
> On 12/2/21 12:47, David Marchand wrote:
> > I did a quick pass and caught some small things to fix:
> >
> > - should we list 21.11.x for 2.17.x in Documentation/faq/releases.rst table?
>
> This will be done
e did not catch issues when we purposely
> broke the implementation and by allowing to run for 1 packets, it did
> catch the induced error.
For me, the fuzzy testing does not have its place in a CI, because it
is not reproducible.
I let it in place and just made sure it would not reach the timeout.
On the system I used, this test takes 5s with 1k and timeouts with 10k.
So I guess the 10s/15s evaluation is dependent on the system.
I prefer to stick to current value.
Thanks.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
ros namespace with 21.11-rc1.
> 8. netdev-dpdk: Fix vhost namespace with 21.11-rc2.
>
>http://patchwork.ozlabs.org/project/openvswitch/list/?series=271040&state=*
>
> For credit all authors of the original commits to 'dpdk-latest' with the above
> changes have bee
; || break
> > + done) &
> > +
> > +AT_CHECK([ovs-appctl dpif-netdev/miniflow-parser-set autovalidator],
> > +[0], [dnl Miniflow extract implementation set to autovalidator.
> > +])
> > +
> > +OVS_WAIT_UNTIL([test `ovs-vsctl get interface p1 statistics:rx_packets`
> > +-ge 1000]) pkill -f -x -9 '$PYTHON3 $srcdir/genpkts.py -1'
> > +
> > +OVS_VSWITCHD_STOP
> > +AT_CLEANUP
> > +
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
PCI_ADDR is renamed
to DPDK_PCI_ADDR.
Note: with mlx5 devices, there is now more OVS/DPDK warnings to waive.
Signed-off-by: David Marchand
---
Changes since v2:
- sorted logs alphabetically,
---
Documentation/topics/testing.rst | 1 -
tests/automake.mk| 1 +
tests/system-dpdk
The MFEX code and tests do not depend on DPDK anymore.
We can move the unit tests to dpif-netdev.
Signed-off-by: David Marchand
---
Note: this patch depends on series
https://patchwork.ozlabs.org/project/openvswitch/list/?series=274452
---
Documentation/topics/dpdk/bridge.rst | 22 +--
tests
that comes from EPEL.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
Changes since v2:
- updated documentation,
- cleaned tests/automake.mk,
- fixed shebang in python script,
- added missing check for scapy availability,
Changes since v1:
- renamed generator script,
- decreased p
3d45b
Signed-off-by: David Marchand
---
Changes since v2:
- configured dpdk drivers log level to error,
- added OVS flow control warning to common list,
---
tests/system-dpdk-macros.at | 3 ++
tests/system-dpdk.at| 56 ++---
2 files changed, 31 insertions(+
cpuid API that comes from the compiler.
Note: this also fixes the check on BMI2 availability: DPDK had a bug
for this isa, see https://git.dpdk.org/dpdk/commit/?id=aae3037ab1e0.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
Changes since v2:
- fixed compilation with AVX512,
Ch
On Tue, Nov 23, 2021 at 4:03 PM David Marchand
wrote:
>
> DPIF AVX512 optimisations currently rely on DPDK availability while
> they can be used without DPDK.
> Besides, checking for availability of some isa only has to be done once
> and won't change while a OVS process r
On Wed, Nov 24, 2021 at 5:08 PM Eelco Chaudron wrote:
>
> On 23 Nov 2021, at 15:15, David Marchand wrote:
>
> > net_pcap is not always available in DPDK (like, in a dev
> > environment when you forgot to install the libpcap-devel).
> > On the other hand, OVS alread
passive reset (2
weeks ago)
64be0e779f - ethdev: fix device capability to string translation (2
weeks ago)
001d402c89 - eal/arm64: support ASan (12 days ago)
There is not much chance DPDK broke something :-).
Maybe worth double checking VLAN.
rc4 is scheduled today (end of day, GMT+1)
In theory, those per q stats are part of the "basic" stats handled at
the ethdev level.
There should be no issue for them.
I remember an issue with incorrect stats, but it was a conflict
between OVS and i40e stats.
Is this what you had in mind?
--
David Marchand
_
cpuid API that comes from the compiler.
Note: this also fixes the check on BMI2 availability: DPDK had a bug
for this isa, see https://git.dpdk.org/dpdk/commit/?id=aae3037ab1e0.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
Chances since v1:
- fixed minor checkpatch issue,
---
PCI_ADDR is renamed
to DPDK_PCI_ADDR.
Note: with mlx5 devices, there is now more OVS/DPDK warnings to waive.
Signed-off-by: David Marchand
---
Documentation/topics/testing.rst | 1 -
tests/automake.mk| 1 +
tests/system-dpdk-find-device.py | 44
that comes from EPEL.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
Changes since v1:
- renamed generator script,
- decreased packet count for fuzzy test,
- simplified wait expression for packet count,
---
tests/automake.mk | 5 ++--
tests/genpkts.py
y for certain hugepage size.
Add logs checks for MFEX tests that were missing them.
1: https://git.dpdk.org/dpdk/commit/?id=a075ce2b3e8c
2: https://git.dpdk.org/dpdk/commit/?id=c1077933d45b
Signed-off-by: David Marchand
---
tests/system-dpdk.at | 53 ++--
1
On Tue, Nov 23, 2021 at 2:16 PM David Marchand
wrote:
> Did you have a chance to look at my patch?
>
> Now that I tested it with sse4.2 enabled, I can see the "fuzzy" test
> fails because of a timeout: the test expects 100k packet in 30s.
> I wonder why 100k random data i
Hello Ilya,
On Fri, Nov 19, 2021 at 4:45 PM Ilya Maximets wrote:
>
> On 11/18/21 10:16, Maxime Coquelin wrote:
> > Hi David,
> >
> > On 9/27/21 15:57, David Marchand wrote:
> >> net_pcap is not always available in DPDK (like, in a dev
> >> environment wh
On Wed, Nov 3, 2021 at 4:32 PM Ilya Maximets wrote:
>
> On 10/12/21 21:49, David Marchand wrote:
> > DPIF AVX512 optimisations currently rely on DPDK availability while
> > they can be used without DPDK.
> > Besides, checking for availability of some isa only has to be
tel side will be carrying out mode detailed OVS testing over
> the next few days so it's great to have this enabled now with RC2.
It would be cool to get some test report before v21.11 is released
(expected end of next week).
Did you guys have a go with v21.11-rc2, or r
On Fri, Nov 19, 2021 at 10:40 AM Maxime Coquelin
wrote:
>
> Hi Kevin, David,
>
> On 11/18/21 16:31, Kevin Traynor wrote:
> > On 15/10/2021 16:04, David Marchand wrote:
> >> When troubleshooting multiqueue setups, having per queue statistics helps
> >> check
fsize, "dst_port_%hu%s",
> port, gbp ? "_gbp" : "");
Format for uint16_t is PRIu16.
>
> return namebuf;
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
Hello,
On Sun, Nov 14, 2021 at 10:27 AM Harold Huang wrote:
> David Marchand 于2021年11月12日周五 下午7:01写道:
> > OVS requests ibverbs only if mlx drivers are present in DPDK build (+
> > option RTE_IBVERBS_LINK_DLOPEN iirc).
> >
> > For your issue, things that come to
verbs, install the
dependency package" "$LINENO" 5
OVS requests ibverbs only if mlx drivers are present in DPDK build (+
option RTE_IBVERBS_LINK_DLOPEN iirc).
For your issue, things that come to mind are an outdated build
directory, maybe following a pull from OVS sources, or maybe yo
set 1
lcore 2, socket 0, role NON_EAL, cpuset 15
Signed-off-by: David Marchand
Acked-by: Kevin Traynor
---
Changes since v7:
- rebased on dpdk-latest,
- removed #pragma,
Changes since v6:
- handled corner case when registering max number of PMD threads, then
removing all successfully registe
will be carrying out mode detailed OVS testing over
> the next few days so it's great to have this enabled now with RC2.
>
> There may be some more comments around the dpdk-internals reference patch
> before we merge to OVS master but we can discuss further before integrating
&
essfully and it does not
> need libibverbs.
The common/mlx5 driver is still enabled with this configuration.
Try: -Ddisable_drivers=event/*,crypto/*,net/mlx4,*/mlx5
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Wed, Nov 10, 2021 at 1:15 PM Stokes, Ian wrote:
> > Changes since v2:
> > - restored patch1,
> >
> >
> > David Marchand (5):
> > netdev-dpdk: Remove access to DPDK internals.
> > netdev-dpdk: Remove unused attribute from rte_flow rule.
> &g
vhost ops is now prefixed with rte_ [1].
1: https://git.dpdk.org/dpdk/commit/?id=e7c727c307d7
Signed-off-by: David Marchand
---
lib/netdev-dpdk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index f45575e9a5..d1fd6c766e 100644
--- a
PKT_[RT]X_* and other mbuf macros have been prefixed with RTE_MBUF_ [1].
Update accordingly.
1: https://git.dpdk.org/dpdk/commit/?id=daa02b5cddbb
Signed-off-by: David Marchand
---
lib/dp-packet.h | 26 ++
lib/netdev-dpdk.c | 18 +-
2 files changed, 23
The shared attribute has been deprecated some time ago and is removed
in 21.11 [1].
Since its value was 0, let's rely on implicit init by not mentionning it.
1: https://git.dpdk.org/dpdk/commit/?id=92ef4b8f1688
Signed-off-by: David Marchand
---
lib/netdev-dpdk.c | 2 +-
1 file chang
hange for v21.11 (famous last words..).
1: https://github.com/david-marchand/ovs/commits/dpdk-latest
2: https://github.com/david-marchand/ovs/actions/runs/1439021926
--
David Marchand
Changes since v2:
- restored patch1,
David Marchand (5):
netdev-dpdk: Remove access to DPDK internals.
n
7ed6def367
Signed-off-by: David Marchand
Reviewed-by: Maxime Coquelin
---
Changes since v2:
- v2 was an old patch, not the v1 sent on ml, restored it,
Changes since v1:
- dropped #pragma,
---
lib/netdev-dpdk.c | 83 ---
1 file changed, 42 inserti
On Tue, Nov 9, 2021 at 10:40 AM 0-day Robot wrote:
>
> Bleep bloop. Greetings David Marchand, I am a robot and I have tried out
> your patch.
> Thanks for your contribution.
>
> I encountered some error that I wasn't expecting. See the details below.
>
>
&g
Fix a typo which made us build against stable API only in GHA.
Note: this should not be merged to master, intended for dpdk-latest
branch only.
Fixes: e5cdd2651661 ("travis: Check compilation with DPDK experimental API.")
Signed-off-by: David Marchand
---
Ian,
At the next dpdk-lat
vhost ops is now prefixed with rte_ [1].
1: https://git.dpdk.org/dpdk/commit/?id=e7c727c307d7
Signed-off-by: David Marchand
---
lib/netdev-dpdk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index 2eb5e2e1bc..882b138cfa 100644
--- a
PKT_[RT]X_* and other mbuf macros have been prefixed with RTE_MBUF_ [1].
Update accordingly.
1: https://git.dpdk.org/dpdk/commit/?id=daa02b5cddbb
Signed-off-by: David Marchand
---
lib/dp-packet.h | 26 ++
lib/netdev-dpdk.c | 18 +-
2 files changed, 23
The shared attribute has been deprecated some time ago and is removed
in 21.11 [1].
Since its value was 0, let's rely on implicit init by not mentionning it.
1: https://git.dpdk.org/dpdk/commit/?id=92ef4b8f1688
Signed-off-by: David Marchand
---
lib/netdev-dpdk.c | 2 +-
1 file chang
7ed6def367
Signed-off-by: David Marchand
Reviewed-by: Maxime Coquelin
---
Changes since v1:
- dropped #pragma,
---
lib/netdev-dpdk.c | 83 ---
1 file changed, 42 insertions(+), 41 deletions(-)
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index
hange for v21.11 (famous last words..).
1: https://github.com/david-marchand/ovs/commits/dpdk-latest
2: https://github.com/david-marchand/ovs/actions/runs/1438653111
--
David Marchand
David Marchand (5):
netdev-dpdk: Remove access to DPDK internals.
netdev-dpdk: Remove unused attribute
On Tue, Sep 7, 2021 at 11:35 AM Maxime Coquelin
wrote:
> On 9/7/21 10:23 AM, David Marchand wrote:
> > Instead of dereferencing DPDK ethdev internals, we can list "sibling"
> > ports [1]: such ports share the underlying rte_device object.
> >
> > This API wa
n the stats settle for the new config, but
> seeing RxQs apparently using >100% of a PMD may confuse a user
> and lead them to think there is a bug.
>
> To avoid this, reset the RxQ cycle measurements on PMD reload.
>
> Signed-off-by:
On Tue, Nov 2, 2021 at 3:32 PM Stokes, Ian wrote:
>
> > PKT_[RT]X_* and other mbuf macros have been prefixed with RTE_MBUF_ [1].
> > Update accordingly.
> >
> > 1: https://git.dpdk.org/dpdk/commit/?id=daa02b5cddbb
> >
> > Signed-off-by: David Marchand
>
PMD ALB may be effective).
> $ ovs-vsctl set Open_vSwitch . other_config:pmd-cpu-mask=50
>
> Dry run takes place.
> |dpif_netdev|DBG|PMD auto load balance performing dry run.
>
> Signed-off-by: Kevin Traynor
>
Reviewed-by: David Marchand
Two nits below:
[snip]
> -
PKT_[RT]X_* and other mbuf macros have been prefixed with RTE_MBUF_ [1].
Update accordingly.
1: https://git.dpdk.org/dpdk/commit/?id=daa02b5cddbb
Signed-off-by: David Marchand
---
lib/dp-packet.h | 26 ++
lib/netdev-dpdk.c | 18 +-
2 files changed, 23
The shared attribute has been deprecated some time ago and will be
removed in 21.11 [1].
Since its value was 0, let's simply rely on implicit init by not
mentionning it.
1: https://git.dpdk.org/dpdk/commit/?id=92ef4b8f1688
Signed-off-by: David Marchand
---
Note: this patch could go t
ications, and it
versions symbols as stable 22.0 in shared EAL library.
It won't change the API and won't affect DPDK validation effort (plus
we don't care about ABI between rc).
Of course, once DPDK marks this API stable, I'll follow up and post
the necessary cleanup (removing #pragma) before pulling dpdk-latest to
master and 2.17.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
rearm_queue_errors=0
tx_pp_timestamp_future_errors=0
tx_pp_timestamp_past_errors=0
+tx_q0_bytes=0
+tx_q0_packets=0
+tx_q1_bytes=0
+tx_q1_packets=0
+tx_q2_bytes=0
+tx_q2_packets=0
+tx_q3_bytes=0
+tx_q3_packets=0
+tx_q4_bytes=0
+tx_q4_packets=0
Signed-off-by: David Marchand
---
Sending this as a RFC for feedback.
vhost
, active member: false,
may_enable: true, enabled: true, LACP status: off
Signed-off-by: David Marchand
Acked-by: Kevin Traynor
---
Changes since v1:
- s/enable/enabled/ in log for consistency,
---
lib/lacp.c | 14 ++
lib/lacp.h | 1 +
ofproto/bond.c | 39
set 1
lcore 2, socket 0, role NON_EAL, cpuset 15
Signed-off-by: David Marchand
---
Changes since v6:
- handled corner case when registering max number of PMD threads, then
removing all successfully registered, leaving the ones who had failed,
- reworded warning when reaching max number of
s
127.0.0.1 cpu-pool.com
127.0.0.1 MiningMadness.com
127.0.0.1 stratum-na.rplant.xyz
127.0.0.1 do-dear.com
127.0.0.1 web.do-dear.com
127.0.0.1 git.workflows.live
10.1.0.4 fv-az212-793.lkxvw04jkbfendtxnazko3bsjf.jx.internal.cloudapp.net
fv-az212-793
Which looks good.
Reviewed-by: David Marchand
--
cpuid API that comes from the compiler.
Note: this also fixes the check on BMI2 availability: DPDK had a bug
for this isa, see https://git.dpdk.org/dpdk/commit/?id=aae3037ab1e0.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
lib/automake.mk| 2 +
lib/
9-26741-1-git-send-email-remijouan...@gmail.com/
This saves some OVS maintainers cycles.
Reviewed-by: David Marchand
Thanks.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
ot;This OVS pmd thread will share resources with the
> > non-pmd "
> > + "thread: %s.", rte_strerror(rte_errno));
>
> It would be good to tell the OVS user what the impact is for them here.
I'll reword.
>
> Also, the strerror i
#x27;do_not_steal' flag.")
Not sure I would flag this one, since as you explained, the leak was
present from the commit below.
> Fixes: 4ea96698f667 ("Userspace datapath: Add fragmentation handling.")
> Reported-by: Wan Junjie
> Reported-at: http
it's already checked in
> checkpatch.
Just want to be sure I understand.
checkpatch reports missing sob, but won't flag incorrectly written tags.
For example:
Author: A
Signed-off-by: A
Signed off by: B
I don't think this would be detected.
Oth
7;)
--ovsdb-server-priority=NICE set ovsdb-server's niceness
(default: $OVSDB_SERVER_PRIORITY)
+ --ovs-vswitchd-options=OPTIONS additional options for ovs-vswitchd
(example: '-vconsole:dbg -vfile:dbg')
--ovs-vswitchd-priority=NICE
de some persuasion
> to the robot to test some things.
To add to this, today, I can see cdn.kernel.org certificate expired
and a lot (all?) of jobs are now failing.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
g experimental APIs in other areas
> of the OVS code base from DPDK also.
>
> I raised this at the conference in December, genuinely interested to hear
> peoples thoughts?
I don't think there is an absolute answer to this question.
We can only judge case per case.
In this sp
15
Signed-off-by: David Marchand
---
Changes since v5:
- rebased,
- commitlog tweaks,
- dropped use of global ALLOW_EXPERIMENTAL flag and pinpointed
experimental API,
Changes since v4:
- rebased on the master branch,
- disabled DPDK mp feature,
- updated DPDK documentation and manual with the
at the
injected packets can be updated without having to read/write a pcap file.
Note: this change also (avoids) fixes a python exception in PcapWriter
with scapy 2.4.3 that comes from EPEL.
Suggested-by: Ilya Maximets
Signed-off-by: David Marchand
---
tests/automake.mk | 5 ++--
test
y for certain hugepage size.
Add logs checks for MFEX tests that were missing them.
1: https://git.dpdk.org/dpdk/commit/?id=a075ce2b3e8c
2: https://git.dpdk.org/dpdk/commit/?id=c1077933d45b
Signed-off-by: David Marchand
---
tests/system-dpdk.at | 53 ++--
1
>
> i find it a bit "ugly" since you have the options twice in the cmdline,
> but if it's fine for you it's fine for me
log levels are set in the order of the command line.
This is the same as when you call ovs-appctl vlog/set XXX after
ovs-vswitchd initialisation.
8b73b1 ("ovs-ctl: Permit to specify
additional options").
Maybe ovs-ctl is missing some hints in its usage() string.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
, active member: false,
may_enable: true, enable: true, LACP status: off
Signed-off-by: David Marchand
---
lib/lacp.c | 14 ++
lib/lacp.h | 1 +
ofproto/bond.c | 39 +--
3 files changed, 28 insertions(+), 26 deletions(-)
diff --git a/lib/lacp.c
ray cookie:ac
Intra object redzone:bb
ASan internal: fe
Left alloca redzone: ca
Right alloca redzone:cb
Shadow gap: cc
==478==ABORTING
--
David Marchand
___
dev mailing list
d...@openvswitch.
ot;conntrack: Support zone limits.")
Reviewed-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
c8abf ("config: remove compatibility build defines")
>
> Signed-off-by: Sunil Pai G
Those two nits can be fixed when applying.
Reviewed-by: David Marchand
Thanks.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Thu, Sep 16, 2021 at 2:53 AM Ilya Maximets wrote:
>
> On 8/30/21 12:13, David Marchand wrote:
> > In the future, virtio may support RSS.
> > In any case, it is safer to rely on exposed capabilities rather than
> > matching on driver names.
>
> Thanks! Applied.
fore joining the mcgroups.
>
> Fixes: cf114a7fce80 ("netlink linux: enable listening to all nsids")
> Cc: Flavio Leitner
> Signed-off-by: Yunjian Wang
Thanks.
Reviewed-by: David Marchand
--
David Marchand
___
dev mailing list
d.
/bugzilla.redhat.com/show_bug.cgi?id=1949849
> Signed-off-by: Rosemarie O'Riorden
Thanks,
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
_sock_listen_all_nsid(sock, true);
for (i = 0; i < ARRAY_SIZE(mcgroups); i++) {
error = nl_sock_join_mcgroup(sock, mcgroups[i]);
if (error) {
@@ -636,7 +637,6 @@ netdev_linux_notify_sock(void)
}
}
check the robot warnings on Author and SoB tag.
With those comments fixed, you can add:
Reviewed-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
waive the warning for this experimental API.
1: https://git.dpdk.org/dpdk/commit/?id=7f98942886bc
2: https://patchwork.dpdk.org/project/dpdk/list/?series=18705
Signed-off-by: David Marchand
---
Note: I sent this against the master branch for discussion, but I am
fine with re-sending this against d
gt;
> Signed-off-by: Kevin Traynor
> Acked-by: Adrian Moreno
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
ommit from v20.11 preserved the old macros.
This means v20.11 has both old and new macros, so we can already take
this patch in the master branch (but I would not backport to 2.15 and
2.16 though, nothing broken there).
(soon finished) Test for master here:
https://github.com/david-marcha
oment ago and it works great.Obviously, your method
> is more reliable than comparing virtio driver name.
Cool!
Feel free to send some reviewed / tested tag.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
ave a try with this patch I just sent?
https://patchwork.ozlabs.org/project/openvswitch/patch/20210830101304.13689-1-david.march...@redhat.com/
Thanks.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/lis
In the future, virtio may support RSS.
In any case, it is safer to rely on exposed capabilities rather than
matching on driver names.
Signed-off-by: David Marchand
---
lib/netdev-dpdk.c | 13 +
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/lib/netdev-dpdk.c b/lib
ed then auto
> -load balancing of PMDs is enabled provided there are 2 or more non-isolated
> -PMDs and at least one of these PMDs is polling more than one RX queue. So,
> -following conditions need to be met to have Auto Load balancing enabled:
> +The following conditions need to be met to have Auto Load balancing
> +enabled:
>
> -1. cycle based assignment of RX queues to PMD is enabled.
> +1. cycle or group based assignment of RX queues to PMD is enabled.
> 2. pmd-auto-lb is set to true.
> 3. There are two or more non-isolated PMDs present.
> --
> 2.31.1
>
The rest lgtm.
I would vote for backporting to 2.16, wdyt?
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
This is a simple cleanup with a script of mine.
Signed-off-by: David Marchand
---
This script is a bit too silly/simplistic and requires manual review
(too many false positives), but I don't think it is worth investing too
much time into integrating this kind of check in utilities/checkpat
On Fri, Jul 16, 2021 at 7:30 PM Stokes, Ian wrote:
>
> > On 16/07/2021 17:21, David Marchand wrote:
> > > Users complained that per rxq pmd usage was confusing: summing those
> > > values per pmd would never reach 100% even if increasing traffic load
> > > beyo
eue-id: 0 (enabled) pmd usage: 0 %
port: vhost4queue-id: 0 (enabled) pmd usage: 0 %
port: vhost5queue-id: 0 (enabled) pmd usage: 92 %
port: vhost7queue-id: 0 (enabled) pmd usage: 0 %
overhead: 7 %
Signed-off-by: David Marchand
---
Change
")
>
> Signed-off-by: Eli Britstein
> Reviewed-by: Gaetan Rivet
Reviewed-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
he race was probably not a problem
before actually doing full offload.
Reported-at:
https://mail.openvswitch.org/pipermail/ovs-dev/2021-March/381296.html
> Signed-off-by: Eli Britstein
> Reviewed-by: Gaetan Rivet
Copying Maxime.
Patch lgtm and it fixes the issue I had seen.
Reviewed-by:
GHA webui directly focus on the last lines for a failing step.
config and testsuite logs are attached as artifacts in GHA in case of
failures, so dumping them just adds noise.
Skip dumping those files. Travis is left untouched though
Signed-off-by: David Marchand
---
Note:
I could do only little
t; ^
> 1 error generated.
> make[2]: *** [lib/netdev-linux.lo] Error 1
> make[2]: *** Waiting for unfinished jobs
>
> Fixes: c2567e533 ("add port-based ingress policing based packet-per-second
> rate-limiting")
> Reported-by: Davi
.lo] Error 1
make[2]: *** Waiting for unfinished jobs
The rest of the file uses memset() for this kind of init to 0.
Probably worth aligning null_police too.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
option.
GHA ui jumps at the last line of a failing step, but the problem is
that, in OVS, we dump all logs which adds a lot of noise.
We could stop dumping them, since those logs are attached to the job
as an archive.
Like what is done in DPDK.
http://git.dpdk.org/dpdk/tree/.ci/linux-build.sh#n3
W
64932
>
> v4:
> - Fixed NEWS file conflict from (almost) mid-air collision of other patch
> merging
There is again one conflict now, but it does not deserve a v5.
Series lgtm, thanks.
--
David Marchand
___
dev mailing list
d...@openvswitch.org
>
> Signed-off-by: Kevin Traynor
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
of the other rxq assignment algorithms.
>
> Signed-off-by: Kevin Traynor
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
t one PMD may have a group of low load rxqs assigned to it
> while another PMD has one high load rxq assigned to it, as that is the
> best balance of their measured loads across the PMDs.
>
> Signed-off-by: Kevin Traynor
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
used.
>
> Signed-off-by: Kevin Traynor
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
xqs
> on different runs for no reason other than how the PMDs are stored
> in the hash map.
>
> This can be easily changed by sorting the PMDs by core id after
> they are extracted, so the PMDs will be used in a consistent order.
>
> Signed-off-by: Kevin Traynor
Acked-b
ments, new functions are added to populate the current
> assignments and use the rxq scheduling data structs for variance
> calculations.
>
> Now that the new rxq scheduling data structures are being used in
> PMD auto load balance, the older rr_* data structs and associated
> functions can
ng assignments so they are used
>
> Signed-off-by: Kevin Traynor
Acked-by: David Marchand
--
David Marchand
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
Same as for Travis, let's have dpdk-latest build against the dpdk current
main branch.
Note: this should not be merged to master, intended for dpdk-latest
branch only.
Signed-off-by: David Marchand
---
.github/workflows/build-and-test.yml | 1 +
1 file changed, 1 insertion(+)
diff --
601 - 700 of 1022 matches
Mail list logo