Your message dated Wed, 14 Aug 2013 06:18:13 +
with message-id
and subject line Bug#710125: fixed in openvswitch 1.9.2+git20130813-1
has caused the Debian Bug report #710125,
regarding openvswitch: unable to compile kernel module with linux-3.8
to be marked as done.
This means that you claim
Your message dated Wed, 14 Aug 2013 06:18:13 +
with message-id
and subject line Bug#701758: fixed in openvswitch 1.9.2+git20130813-1
has caused the Debian Bug report #701758,
regarding openvswitch: new version upstream
to be marked as done.
This means that you claim that the problem has been
Accepted:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
Format: 1.8
Date: Fri, 07 Jun 2013 14:15:01 -0700
Source: openvswitch
Binary: openvswitch-datapath-source openvswitch-datapath-dkms
openvswitch-common openvswitch-switch openvswitch-ipsec openvswitch-pki
openvswitch-controller openvswi
openvswitch_1.9.2+git20130813-1_i386.changes uploaded successfully to localhost
along with the files:
openvswitch_1.9.2+git20130813-1.dsc
openvswitch_1.9.2+git20130813.orig.tar.xz
openvswitch_1.9.2+git20130813-1.debian.tar.gz
openvswitch-common_1.9.2+git20130813-1_i386.deb
openvswitch-swi
From: gyang
With megaflow support, there is API to convert mask to nlattr key based
format. This change introduces API to do the reverse conversion. We
leverage the existing odp_flow_key_to_flow() API to resue the code.
Signed-off-by: Guolin Yang
---
lib/odp-util.c | 301 +
Thanks.
I'm going to spend a little more time tomorrow trying to reproduce this
and figure out the root cause. If I can't do that, I'll go with what's
here.
I was debugging this on branch-1.11 (that's where my bug report is) but
Justin pointed out that it's not just there, it also affects other
Glad you found it. We use SW_FLOW_KEY_PUT/MEMCPY to also keep track of
mask range as we populate the fields in any order.
On Tue, Aug 13, 2013 at 5:15 PM, roampune wrote:
> actually found it, hidden way beneath.. ok so it is at:
> datapath/flow.c:53: update_range__ it update the ranges for tun
On Wed, Aug 14, 2013 at 09:29:48AM +0800, Ethan Jackson wrote:
> > - Make sparse work with clang. (I don't think it's too hard,
> > at least on systems where GCC is also installed.)
> >
> > - Add bitwise (or similar) support to Clang (and upstream it).
>
> Either of thes
Thanks for the review, will work on this after I send the bfd decay patch
later this evening.
On Tue, Aug 13, 2013 at 6:24 PM, Ethan Jackson wrote:
> Now that multithreading is in. Could you rework this patch to
> actually reconfigure the ofproto-dpif-upcall module with the
> appropriate setti
> - Make sparse work with clang. (I don't think it's too hard,
> at least on systems where GCC is also installed.)
>
> - Add bitwise (or similar) support to Clang (and upstream it).
Either of these seems fine to me. Suspect that just getting sparse to
work with clang is
Now that multithreading is in. Could you rework this patch to
actually reconfigure the ofproto-dpif-upcall module with the
appropriate setting? Basically the udpif_recv_set() function needs to
be called with the appropriate value on boot, and whenever the setting
changes.
I don't see any particu
I'm having a look at it.
On Thu, Aug 8, 2013 at 6:42 AM, Alex Wang wrote:
> Hey Ben,
>
> Could you help review this patch?
>
> Kind Regards,
> Alex Wang,
>
>
> On Wed, Aug 7, 2013 at 3:51 PM, Alex Wang wrote:
>>
>> This commit adds a new column "n-worker-threads" to ovsdb Open_vSwitch
>> TABLE.
Acked-by: Ethan Jackson
On Wed, Aug 14, 2013 at 8:44 AM, Ben Pfaff wrote:
> When the ofp_port argument to bundle_add_port() refers to an ofport_dpif
> that already belongs to some other bundle, bundle_add_port() removed
> the port from the other bundle, correctly, with bundle_del_port().
> If t
On Mon, Jul 29, 2013 at 3:47 PM, Pravin B Shelar wrote:
> CC: Lori Jakab
> Signed-off-by: Pravin B Shelar
> ---
> datapath/vport-lisp.c | 48 ++--
> 1 files changed, 2 insertions(+), 46 deletions(-)
I think idea to use zero as the protocol is a goo
On Mon, Jul 29, 2013 at 3:47 PM, Pravin B Shelar wrote:
> diff --git a/datapath/vport-lisp.c b/datapath/vport-lisp.c
> index 6027a16..6e37b2f 100644
> --- a/datapath/vport-lisp.c
> +++ b/datapath/vport-lisp.c
> +static struct sk_buff *handle_offloads(struct sk_buff *skb)
> +{
> + int err;
>
When the ofp_port argument to bundle_add_port() refers to an ofport_dpif
that already belongs to some other bundle, bundle_add_port() removed
the port from the other bundle, correctly, with bundle_del_port().
If the other bundle now contained no ports, however, this violated the
invariant that a bu
On Mon, Jul 29, 2013 at 3:47 PM, Pravin B Shelar wrote:
> Signed-off-by: Pravin B Shelar
> ---
> datapath/compat.h | 42 ++
> datapath/tunnel.c | 38 --
> datapath/tunnel.h |4
> 3 files changed, 42 insertio
On Mon, Jul 29, 2013 at 3:47 PM, Pravin B Shelar wrote:
> This makes ovs-module more in-sync with upstream ovs-module.
>
> Signed-off-by: Pravin B Shelar
Acked-by: Jesse Gross
X-CudaMail-Whitelist-To: dev@openvswitch.org
___
dev mailing list
dev@openv
actually found it, hidden way beneath.. ok so it is at:
datapath/flow.c:53: update_range__ it update the ranges for tun_id and
other parameters, basically called during flow-add as Andy mentioned below.
Thanks guys for the help.
Regards.
On Tue, Aug 13, 2013 at 5:10 PM, roampune wrote:
>
> On
On Tue, Aug 13, 2013 at 4:34 PM, Andy Zhou wrote:
> Mask list gets build as the kernel receives the flow-add netlink message
> from the user space. If the downloaded flow does not mention tun_id, we
> will treat is as wildcard, this is in general the same with other netlink
> attributes.
>
very
This can be a bug, when we configure bfd:min_rx and bfd:decay_min_rx
together. (at the time we are in decay)
On Tue, Aug 13, 2013 at 4:51 PM, Alex Wang wrote:
> Currently, when there are multiple bfd configuration changes,
> the bfd_poll() will only update one change at a time with the
> other
Mask list gets build as the kernel receives the flow-add netlink message
from the user space. If the downloaded flow does not mention tun_id, we
will treat is as wildcard, this is in general the same with other netlink
attributes.
On Tue, Aug 13, 2013 at 4:05 PM, roampune wrote:
> Hi Guys,
>
>
Currently, when there are multiple bfd configuration changes,
the bfd_poll() will only update one change at a time with the
other side. This commit moves the call to bfd_poll() at the
end of configuration processing function, so that bfd_poll()
will update all configuration changes together.
Signe
Hi Guys,
I am trying to trace the code for vxlan in datapath directory.
I know how the skb traverse from udp socket as shown below.
udp socket -> vxlan_udp_encap_recv -> vxlan_rcv ->
ovs_vport_receive -> ovs_dp_process_received_packet -> ...
but i don't quite understand where does the tun_id com
On Tue, Aug 13, 2013 at 10:50 AM, Alex Wang wrote:
> I think the intention is right here, but it's a bit odd. For example,
>> if in_decay is true, we don't drop out of it with this logic in all
>> cases. How about we add a boolean to this function
>> "cfm_min_rx_changed" and set it to true here
From: gyang
With megaflow support, there is API to convert mask to nlattr key based
format. This change introduces API to do the reverse conversion. We
leverage the existing odp_flow_key_to_flow() API to resue the code.
Signed-off-by: Guolin Yang
---
lib/odp-util.c | 257 +
Thanks for the review, I pushed the series to master.
On Tue, Aug 13, 2013 at 1:48 PM, Andy Zhou wrote:
> All 4 patches look good to me.
>
> Acked-by: Andy Zhou
>
>
>
> On Tue, Aug 6, 2013 at 12:57 PM, Jesse Gross wrote:
>>
>> The Linux kernel datapath enables matching and setting the skb mark
Looks good.
--Justin
On Aug 13, 2013, at 12:54 PM, Ben Pfaff wrote:
> dpkg-buildflags has not always supported --export=configure, but commit
> 6c2d4c8780 (debian: Apply hardening options to build.) used it
> unconditionally, causing the build to fail on old Debian distributions.
> This fixes
From: YAMAMOTO Takashi
Signed-off-by: YAMAMOTO Takashi
---
lib/netdev-bsd.c | 63 ++--
1 file changed, 61 insertions(+), 2 deletions(-)
diff --git a/lib/netdev-bsd.c b/lib/netdev-bsd.c
index 50fb520..180ce7f 100644
--- a/lib/netdev-bsd.c
+++
All 4 patches look good to me.
Acked-by: Andy Zhou
On Tue, Aug 6, 2013 at 12:57 PM, Jesse Gross wrote:
> The Linux kernel datapath enables matching and setting the skb mark
> but this functionality is currently used only internally by
> ovs-vswitchd. This exposes it through NXM to enable ext
dpkg-buildflags has not always supported --export=configure, but commit
6c2d4c8780 (debian: Apply hardening options to build.) used it
unconditionally, causing the build to fail on old Debian distributions.
This fixes the problem.
Signed-off-by: Ben Pfaff
---
debian/rules |9 +++--
1 fil
On Tue, Aug 13, 2013 at 12:32:55PM -0700, Ben Pfaff wrote:
> dpkg-buildflags has not always supported --export=configure, but commit
> 6c2d4c8780 (debian: Apply hardening options to build.) used it
> unconditionally, causing the build to fail on old Debian distributions.
> This fixes the problem.
>
Good idea, I added:
# We send each packet twice because the first packet in each flow causes the
# flow table to change and thus revalidations, which (depending on timing)
# can keep a megaflow from being installed. The revalidations are done by
# the second iteration, allowing the flows to be ins
dpkg-buildflags has not always supported --export=configure, but commit
6c2d4c8780 (debian: Apply hardening options to build.) used it
unconditionally, causing the build to fail on old Debian distributions.
This fixes the problem.
Signed-off-by: Ben Pfaff
---
debian/rules |4 ++--
1 file cha
Looks good. Do you think it's worth adding a comment explaining this?
--Justin
On Aug 13, 2013, at 11:28 AM, Ben Pfaff wrote:
> Threaded ofproto-dpif uses a queue to pass packets from the forwarding
> threads to the main thread for (mega)flow setup and for learning. When
> learning occurs, c
Looks good.
--Justin
On Aug 13, 2013, at 11:28 AM, Ben Pfaff wrote:
> This test had two problems. First, it had a bizarre dependency on stats
> that were not up-to-date: the "ovs-ofctl dump-flows" assumed that only
> the first one of ten packets sent through the switch had been accounted
> to
Threaded ofproto-dpif uses a queue to pass packets from the forwarding
threads to the main thread for (mega)flow setup and for learning. When
learning occurs, causing revalidations, this races against flow setup, so
that sometimes a datapath (mega)flow does get set up for a packet that
causes lear
This test had two problems. First, it had a bizarre dependency on stats
that were not up-to-date: the "ovs-ofctl dump-flows" assumed that only
the first one of ten packets sent through the switch had been accounted
to OpenFlow flow statistics. Adding a 1-second time warp fixed this
problem by ens
I've got a lot more code I can upstream, but I haven't seen any reviews for
the first patch. Any comments? (Or did I submit it incorrectly?)
On Fri, Aug 9, 2013 at 4:29 PM, Casey Barker wrote:
> OK, I just sent a patch, and I'll try to post more as time allows.
> Unfortunately, I'm working from
Thanks very much for the review,
On Mon, Aug 12, 2013 at 11:47 PM, Ethan Jackson wrote:
>
> >>> +/* Always resets decay_min_rx when cfg_min_rx is updated. */
> >>> +bfd->decay_min_rx = 0;
> >>> +bfd_poll(bfd);
>
> I think the intention is right here, but it's a bit odd.
On Tue, Aug 13, 2013 at 10:02:41AM -0700, Alex Wang wrote:
> This commit chops off the trailing whitespace in the 'flag' field
> of 'bfd/show' output. This is for the string matching in bfd
> unit test.
>
> Signed-off-by: Alex Wang
Applied, thanks.
__
On Tue, Aug 13, 2013 at 02:47:43PM +0800, Ethan Jackson wrote:
> Thanks this version is much closer. Just some minor points which
> should be pretty easy to clean up. Ben, so we don't have a 12 hour
> review cycle on this, would you mind looking at the next version of
> this patch (which should b
On Tue, Aug 13, 2013 at 08:49:19AM +0800, Ethan Jackson wrote:
> Acked-by: Ethan Jackson
I applied all of these. Thank you for the reviews.
> We've really got to get sparse to work with clang, this sort of thing
> is going to get annoying. C'est la vie.
I think we have a few options assuming
This commit chops off the trailing whitespace in the 'flag' field
of 'bfd/show' output. This is for the string matching in bfd
unit test.
Signed-off-by: Alex Wang
---
lib/bfd.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/lib/bfd.c b/lib/bfd.c
index 81fd178..7c7c0b7 100644
--- a/lib
Thanks for the answer,
On Tue, Aug 13, 2013 at 9:29 AM, Ben Pfaff wrote:
> "dpkg-buildpackage -b" then lintian on the .changes file that produces.
>
> On Mon, Aug 12, 2013 at 05:47:41PM -0700, Alex Wang wrote:
> > This change looks good to me,
> >
> > During my run of "fakeroot debian/rules bin
Thanks for the reviews, I applied these to master and branch-1.9,
branch-1.10, branch-1.11.
On Tue, Aug 13, 2013 at 09:29:19AM -0700, Ben Pfaff wrote:
> "dpkg-buildpackage -b" then lintian on the .changes file that produces.
>
> On Mon, Aug 12, 2013 at 05:47:41PM -0700, Alex Wang wrote:
> > This
"dpkg-buildpackage -b" then lintian on the .changes file that produces.
On Mon, Aug 12, 2013 at 05:47:41PM -0700, Alex Wang wrote:
> This change looks good to me,
>
> During my run of "fakeroot debian/rules binary", there is no warning.
>
> Could you hint me how to reproduce it?
>
>
> On Mon,
I think this would be a bit cleaner if we ditched the has_rx flag.
Instead, each time we go through the run loop, we can check if we've
received any packets, if we have, we reset to rx_detect_time to now +
bfd_min_rx * mult. Then in bfd_forwarding() when forwarding_if_rx is
enabled, we only have t
48 matches
Mail list logo
