This patch fixes the compilation errors of OVS on 4.19+ kernels.
Signed-off-by: Yifeng Sun
---
datapath/linux/compat/include/net/netfilter/nf_conntrack_count.h | 4 ++--
datapath/linux/compat/nf_conncount.c | 6 +-
2 files changed, 7 insertions(+), 3 deletions(-)
From: Florian Westphal
Upstream Commit:
commit 93e66024b0249cec81e91328c55a754efd3192e0
Author: Florian Westphal
Date: Wed Sep 12 15:19:07 2018 +0200
netfilter: conntrack: pass nf_hook_state to packet and error handlers
nf_hook_state contains all the hook meta-informa
Upstream kernel commit a0ae2562 ("netfilter: conntrack: remove l3proto
abstraction") removed header file net/netfilter/nf_conntrack_l3proto.h.
This patch detects it and fixes compilation errors of OVS on 4.19+ kernels.
Signed-off-by: Yifeng Sun
---
acinclude.m4 | 3
This patch updated acinclude.m4 so that OVS can be compiled on 4.19.x
and 4.20.x kernels.
This patch also updated travis files so that latest kernel versions
are used during travis test builds.
Signed-off-by: Yifeng Sun
---
.travis.yml | 18 ++
NEWS | 2 ++
acinclude.m
This patch updated acinclude.m4 so that OVS can be compiled on 4.19.x
and 4.20.x kernels.
This patch also updated travis files so that latest kernel versions
are used during travis test builds.
Signed-off-by: Yifeng Sun
---
.travis.yml | 18 ++
NEWS | 2 ++
acinclude.m
From: Gao Feng
Upstream commit:
commit 440534d3c56be04abfb26850ee882d19d223557a
Author: Gao Feng
Date: Mon Jul 9 18:06:33 2018 +0800
netfilter: Remove useless param helper of nf_ct_helper_ext_add
The param helper of nf_ct_helper_ext_add is useless now, then remove
From: Florian Westphal
Upstream commit:
commit 60e3be94e6a1c5162a0763c9aafb5190b2b1fdce
Author: Florian Westphal
Date: Mon Jun 25 17:55:32 2018 +0200
openvswitch: use nf_ct_get_tuplepr, invert_tuplepr
These versions deal with the l3proto/l4proto details internally.
From: Florian Westphal
Upstream commit:
commit 70b095c84326640eeacfd69a411db8fc36e8ab1a
Author: Florian Westphal
Date: Sat Jul 14 01:14:01 2018 +0200
ipv6: remove dependency of nf_defrag_ipv6 on ipv6 module
IPV6=m
DEFRAG_IPV6=m
CONNTRACK=y yields:
net
This series of patches enabled OVS to support kernel 4.19.x and 4.20.x.
Tests shows that this series of patches passed check-kmod tests or
didn't introduce new failed check-kmod tests for the following
kernel versions:
3.10.0-957.12.1.el7.x86_64 (rhel)
4.4.0-31-generic (ubuntu 14.04)
4.4.0-146-gen
On Mon, May 6, 2019 at 9:48 AM Darrell Ball wrote:
>
> Userspace conntrack cares about IPv4 checksums, so this is a
> prerequisite for adding zone limit support to userspace conntrack.
>
> Fixes: 3f1087c70cf9 ("system-traffic: Add conntrack per zone limit test
> case.")
> CC: Yi-Hung Wei
> Signe
On Mon, May 6, 2019 at 1:02 PM Aaron Conole wrote:
>
> > lib/.deps/dpif-netdev.Tpo -c lib/dpif-netdev.c -o lib/dpif-netdev.o
> > lib/dpif-netdev.c: In function 'dp_execute_cb':
> > lib/dpif-netdev.c:7112:13: error: enumeration value
> > 'OVS_CT_ATTR_TIMEOUT' not handled in switch [-Werror=switch]
0-day Robot writes:
> Bleep bloop. Greetings Yi-Hung Wei, 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.
>
>
> checkpatch:
> ERROR: Author Dan Carpenter needs to sign off.
> WARNING:
On Thu, May 02, 2019 at 05:29:16PM -0700, Darrell Ball wrote:
> There are a few cases where structure copy can be replaced by
> memcpy(), for possible portability benefit. This is because
> the structures involved have padding and elements of the
> structure are used to generate hashes.
>
> Signe
Hi. Thanks for testing!
> Hi,
> Initial tests show that offload is broken as netdev_rte_offloads_flow_put()
> is not called.
> It is not called because it is not registered.
> When starting debugging it I noticed in function
> netdev_register_flow_api_provider()
> that we may return an uninitial
H�rm%9tli sat1F meneceri!XahiF edirik biz%9 %9n yaxF1 qiym%9t
g�nd%9rin. Flav%9 olaraq t%9tbiq olunmaq ���n qiym%9t v%9 m�vcudluq F =
[MLT-2] iki s%9f%9r. (Apr - Sep) %9lav%9 olaraq F = [MLT-2] [MLT ^ -2] F =
[MLT-2] kg = K�tl%9 (M) G�c vahidi kq.m / s ^ 2 H�rm%9tl%9, WUNG zuhan
Userspace conntrack cares about IPv4 checksums, so this is a
prerequisite for adding zone limit support to userspace conntrack.
Fixes: 3f1087c70cf9 ("system-traffic: Add conntrack per zone limit test case.")
CC: Yi-Hung Wei
Signed-off-by: Darrell Ball
---
tests/system-traffic.at | 10 +-
For performance and code simplification reasons, add rcu support for
conntrack. The array of hmaps is replaced by a cmap as part of this
conversion. Using a single map also simplifies the handling of NAT
and allows the removal of the nat_conn map and friends. Per connection
entry locks are introd
Hi,
Initial tests show that offload is broken as netdev_rte_offloads_flow_put() is
not called.
It is not called because it is not registered.
When starting debugging it I noticed in function
netdev_register_flow_api_provider()
that we may return an uninitialized error value.
Because of this in m
On 30.04.2019 15:17, David Marchand wrote:
> No need for a latch here since we don't have to wait.
> A simple boolean flag is enough.
>
> Fixes: e4cfed38b159 ("dpif-netdev: Add poll-mode-device thread.")
> Signed-off-by: David Marchand
> ---
> lib/dpif-netdev.c | 9 -
> 1 file changed, 4
On Mon, May 06, 2019 at 07:37:18AM -0700, Darrell Ball wrote:
> Fixes: 57593fd24378 ( conntrack: Stop exporting internal datastructures.)
> Signed-off-by: Darrell Ball
Thanks, applied.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.o
On 30.04.2019 15:17, David Marchand wrote:
> We have been testing the rebalance code in different situations while
> having traffic going through OVS.
> Those tests have shown that part of the observed packets losses is due to
> some time wasted in signaling/waiting for the pmd threads to reload th
Current issues with Flow API:
* OVS calls offloading functions regardless of successful
flow API initialization. (ex. on init_flow_api failure)
* Static initilaization of Flow API for a netdev_class forbids
having different offloading types for different instances
of netdev with the same net
On Mon, May 6, 2019 at 8:01 AM Ben Pfaff wrote:
> On Thu, May 02, 2019 at 09:34:05PM -0700, Darrell Ball wrote:
> > For performance and code simplification reasons, add rcu support for
> > conntrack. The array of hmaps is replaced by a cmap as part of this
> > conversion. Using a single map also
Flow API providers renamed to be consistent with parent module
'netdev-offload' and look more like each other.
'_rte_' replaced with more convenient '_dpdk_'.
We'll have following structure:
Common code:
lib/netdev-offload-provider.h
lib/netdev-offload.c
lib/netdev-offload.h
Pro
New module 'netdev-offload' created to manage different flow API
implementations. All the generic and provider independent code moved
there from the 'netdev' module.
Flow API providers further encapsulated.
The only function that was changed is 'netdev_any_oor'.
Now it uses offloading related hma
This is the combined patch-set for dynamic flow API along with
the further work about splitting up netdev offloading from the
generic netdev related code and making different netdev flow
offloading implementations more or less consistent.
Version 2:
* No functional changes.
* Patches combined
On Thu, May 02, 2019 at 09:34:05PM -0700, Darrell Ball wrote:
> For performance and code simplification reasons, add rcu support for
> conntrack. The array of hmaps is replaced by a cmap as part of this
> conversion. Using a single map also simplifies the handling of NAT
> and allows the removal o
Fixes: 57593fd24378 ( conntrack: Stop exporting internal datastructures.)
Signed-off-by: Darrell Ball
---
lib/conntrack.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/lib/conntrack.c b/lib/conntrack.c
index 83fc8f2..49f6325 100644
--- a/lib/conntrack.c
+++ b/lib/conntrack.c
@@ -391,6 +391,
On 03.05.2019 22:02, William Tu wrote:
> +static struct xsk_socket_info *
> +xsk_configure(int ifindex, int xdp_queue_id, int xdpmode)
> +{
> +struct xsk_socket_info *xsk;
> +struct xsk_umem_info *umem;
> +void *bufs;
> +int ret;
> +
> +/* umem memory region */
> +ret = posi
Hi. Thanks for a new version.
Quick review inline.
Best regards, Ilya Maximets.
On 03.05.2019 22:02, William Tu wrote:
> The patch introduces experimental AF_XDP support for OVS netdev.
> AF_XDP, Address Family of the eXpress Data Path, is a new Linux socket type
> built upon the eBPF and XDP te
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
Background
==
OVS HW offload solution is consisted of forwarding and control. HW implements
embedded switch that connects SRIOV VF's and forwards packets according to the
dynamically configured HW rules (packets can be altered by HW rules). Packets
that have no forwarding rule, called exce
Background
==
OVS HW offload solution is consisted of forwarding and control. HW implements
embedded switch that connects SRIOV VF's and forwards packets according to the
dynamically configured HW rules (packets can be altered by HW rules). Packets
that have no forwarding rule, called exce
33 matches
Mail list logo