From: Pravin B Shelar
Date: Tue, 23 Apr 2013 10:48:30 -0700
> All genl callbacks are serialized by genl-mutex. This can become
> bottleneck in multi threaded case.
> Following patch adds an parameter to genl_family so that a
> particular family can get concurrent netlink callback without
> genl_l
From: Pravin B Shelar
Date: Tue, 23 Apr 2013 10:48:48 -0700
> OVS locking was recently changed to have private OVS lock which
> simplified overall locking. Therefore there is no need to have
> another global genl lock to protect OVS data structures. Following
> patch uses of parallel_ops genl f
On Wed, Apr 24, 2013 at 10:51:37AM -0700, Jesse Gross wrote:
> On Wed, Apr 24, 2013 at 2:12 AM, Simon Horman wrote:
> > On Tue, Apr 23, 2013 at 06:47:36PM -0700, Jesse Gross wrote:
> >> On Tue, Apr 23, 2013 at 12:58 AM, Simon Horman wrote:
> >> > On Tue, Apr 23, 2013 at 07:41:37AM +, Rajahalm
hi,
>> +static int
>> +netdev_bsd_get_next_hop(const struct in_addr *host, struct in_addr
>> *next_hop,
>> +char **netdev_name)
>> +{
>> +#if defined(__NetBSD__)
>
> Acked-by: Ed Maste
>
> This is close to being usable on FreeBSD as well, with the exception
> that we do
hi,
thanks for reviewing.
> Yamamoto-san,
>
> Thanks for this work. I'm happy that it will result in BSD systems as
> a group getting more OVS test coverage.
>
> The patch set builds and passes unit tests for me on FreeBSD, with the
> exception of the missing #if defined() guard described in a
Actually it wasn't just tunnel ports. Anything that didn't have an ifIndex was
mapping to the same sFlow datasource. Even if it was on another bridge.
Besides, sFlow wouldn't want to model every tunnel as a separate datasource
either -- even if they did have ifIndex numbers. There might be th
Hi,
Thanks, I've made my comments on that thread. Another related question
came to my mind: I've seen from time to time some
"ofproto_dpif|WARN|unexpected flow from datapath" messages, not just
with these patches but with the original XS 6.1 ovs. I couldn't see a
pattern in the type of flows.
I've reviewed and tested this. In the patch description there are one
small mistakes:
-execute_controller_action(), by passing true instead of false as its
+execute_controller_action(), by passing false instead of true as its
But more importantly, todo in handle_miss_upcalls are freed twice, an
On Wed, Apr 24, 2013 at 11:12:51AM -0700, Justin Pettit wrote:
> On Apr 24, 2013, at 10:26 AM, Ben Pfaff wrote:
>
> > On Thu, Mar 21, 2013 at 03:26:44PM -0700, Ben Pfaff wrote:
> >> Signed-off-by: Ben Pfaff
> >
> > Still needs a review.
>
> I thought you got a review from Kyle on 3/22. You fo
Send an optional OVS_PACKET_ATTR_TIMESTAMP attribute in upcalls with
the packet's timestamp. Add a corresponding tstamp attribute into
struct dpif_upcall.
Enable Linux packet timestamping, only if at least one flow requires
it since Linux packet timestamping is set on or off for all packets.
Defi
On Apr 24, 2013, at 10:26 AM, Ben Pfaff wrote:
> On Thu, Mar 21, 2013 at 03:26:44PM -0700, Ben Pfaff wrote:
>> Signed-off-by: Ben Pfaff
>
> Still needs a review.
I thought you got a review from Kyle on 3/22. You followed up with a question
to him:
http://openvswitch.org/pipermail/de
On Wed, Apr 24, 2013 at 10:46 AM, Neil Mckee wrote:
> Adjust sFlow packet samplers to better reflect the underlying per-bridge
> sampling that is really happening, rather than modeling it as per-interface
> sampling. This fixes an aliasing issue: tunnel traffic sampled on different
> bridges was
On Wed, Apr 24, 2013 at 2:12 AM, Simon Horman wrote:
> On Tue, Apr 23, 2013 at 06:47:36PM -0700, Jesse Gross wrote:
>> On Tue, Apr 23, 2013 at 12:58 AM, Simon Horman wrote:
>> > On Tue, Apr 23, 2013 at 07:41:37AM +, Rajahalme, Jarno (NSN -
>> > FI/Espoo) wrote:
>> >>
>> >> On Apr 23, 2013, a
Adjust sFlow packet samplers to better reflect the underlying per-bridge
sampling that is really happening, rather than modeling it as per-interface
sampling. This fixes an aliasing issue: tunnel traffic sampled on different
bridges was all being reported against the same sFlow datasource.
The sF
On Thu, Mar 21, 2013 at 03:26:44PM -0700, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Still needs a review.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Wed, Apr 24, 2013 at 06:02:35PM +0900, Simon Horman wrote:
> 1. mpls or mplsm should be included in the formatted match instead of
>dl_type=0x if the dl_type of the match is ETH_TYPE_MPLS or
>ETH_TYPE_MPLS_MCAST respectively. This is consistent with the treatment
>of other dl_ty
On Wed, Apr 24, 2013 at 12:12:20PM -0400, Ed Maste wrote:
> On 24 April 2013 11:56, Ben Pfaff wrote:
> > On Wed, Apr 24, 2013 at 11:28:45AM -0400, Ed Maste wrote:
> >> On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
> >> > +/*
> >> > + * note: UINT64_MAX means unsupported
> >> > + */
On Thu, Apr 18, 2013 at 06:07:42PM +0300, Jarno Rajahalme wrote:
> Do not clear tunnel metadata on tunnel input. This is used by a later patch.
>
> Signed-off-by: Jarno Rajahalme
Jesse, I *think* this patch is OK, but we have a history of subtle bugs
here, see commit ef506a7c003232 (ofproto-dpif
I'll repeat myself. We haven't observed that behavior before, when
we've created and destroyed many interfaces, or at least this is the
first time it's been reported to me. Can you figure out why
ovs-vswitchd is trying to commit so many transactions? It would be
better for it to do avoid doing t
> +static int
> +netdev_bsd_get_next_hop(const struct in_addr *host, struct in_addr *next_hop,
> +char **netdev_name)
> +{
> +#if defined(__NetBSD__)
Acked-by: Ed Maste
This is close to being usable on FreeBSD as well, with the exception
that we don't have the RT_ADVANCE
On 24 April 2013 11:56, Ben Pfaff wrote:
> On Wed, Apr 24, 2013 at 11:28:45AM -0400, Ed Maste wrote:
>> On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
>> > +/*
>> > + * note: UINT64_MAX means unsupported
>> > + */
>>
>> Is this checked anywhere (e.g., to avoid displaying unsupported
On Tue, Apr 23, 2013 at 06:30:04PM +, Kyle Mestery (kmestery) wrote:
> On Apr 23, 2013, at 11:30 AM, Ben Pfaff wrote:
> > I keep seeing this question.
> >
> > Signed-off-by: Ben Pfaff
>
>
> Looks good to me Ben, very succinct explanation.
Thanks, applied with Justin's fix.
___
Applied, thanks.
On Tue, Apr 23, 2013 at 03:22:34PM -0700, Justin Pettit wrote:
> Looks good.
>
> --Justin
>
>
> On Apr 23, 2013, at 11:07 AM, Ben Pfaff wrote:
>
> > Other code in the tree uses HAVE_BACKTRACE and then blindly includes
> > if it is present, so this doesn't make anything worse
On Wed, Apr 24, 2013 at 11:28:45AM -0400, Ed Maste wrote:
> On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
> > +/*
> > + * note: UINT64_MAX means unsupported
> > + */
>
> Is this checked anywhere (e.g., to avoid displaying unsupported
> stats)? Or will it result in the display of bo
On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
>
> Signed-off-by: YAMAMOTO Takashi
> ---
> lib/netdev-bsd.c | 60
> +---
> 1 file changed, 53 insertions(+), 7 deletions(-)
Acked-by: Ed Maste
___
d
On Wed, Apr 24, 2013 at 10:39:23AM -0400, Ed Maste wrote:
> I have only one overall comment: I personally don't like the style of
> platform-specific #if #elif cases within the function definition,
> where none of the body is common to the implementations. I'll comment
> on the two cases (netdev_b
On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
>
> Signed-off-by: YAMAMOTO Takashi
> ---
> lib/netdev-bsd.c | 47 +++
> 1 file changed, 43 insertions(+), 4 deletions(-)
>
...
> +#elif defined(__NetBSD__)
> +struct netdev_dev_bsd *netdev_dev =
> +
On 22 April 2013 09:20, YAMAMOTO Takashi wrote:
> where interface renaming is not supported (NetBSD), remember both of
> our netdev name and the correspoinding kernel name separately.
> the latter is necessary to talk with kernel using interface names.
> eg. ifioctls, bpf
Acked-by: Ed Maste
Alt
Yamamoto-san,
Thanks for this work. I'm happy that it will result in BSD systems as
a group getting more OVS test coverage.
The patch set builds and passes unit tests for me on FreeBSD, with the
exception of the missing #if defined() guard described in a previous
email.
I have only one overall
Yes, after backlog comes back to zero, memory is freed, and the memory of
ovs-vswitchd stops growing.
However, if stream_send always returns -EAGAIN, the messages are held: "The
send function will not block. If no bytes can be immediately accepted for
transmission, it returns -EAGAIN immediate
On Tue, 2013-04-23 at 14:57 +0100, Roger Pau Monne wrote:
> On 23/04/13 12:00, Ian Campbell wrote:
> > Based on Waldi's RFC at
> > http://lists.xen.org/archives/html/xen-devel/2012-09/msg00943.html
> >
> > To use it set vif.default.script="vif-openvswitch" in /etc/xen/xl.conf or
> > use
> > scrip
On Tue, Apr 23, 2013 at 06:47:36PM -0700, Jesse Gross wrote:
> On Tue, Apr 23, 2013 at 12:58 AM, Simon Horman wrote:
> > On Tue, Apr 23, 2013 at 07:41:37AM +, Rajahalme, Jarno (NSN - FI/Espoo)
> > wrote:
> >>
> >> On Apr 23, 2013, at 4:51 , ext Simon Horman wrote:
> >>
> >> > On Mon, Apr 22,
1. mpls or mplsm should be included in the formatted match instead of
dl_type=0x if the dl_type of the match is ETH_TYPE_MPLS or
ETH_TYPE_MPLS_MCAST respectively. This is consistent with the treatment
of other dl_types in match_format().
2. Add formatting of MPLS TTL to match_format(
33 matches
Mail list logo
