When tx queue is shared among CPUS,the pkts always be flush in
'netdev_dpdk_eth_send'
So it is unnecessarily for flushing in netdev_dpdk_rxq_recv
Otherwise tx will be accessed without locking
Signed-off-by: Wei li http://openvswitch.org/mailman/listinfo/dev> >
---
lib/netdev-dpdk.c | 7 +--
On Fri, Jun 5, 2015 at 4:23 PM, Alin Serdean
wrote:
> Rebased and added the sign off:
>
> https://github.com/openvswitch/ovs/pull/50
Thank you Alin! I applied this. There are a couple of things that I noticed.
1. While installing, I got a dialog box which said: "The file
OVSExt.sys on Open vSwitc
On Sun, Jun 7, 2015 at 11:14 PM, Ben Pfaff wrote:
> On Sun, Jun 07, 2015 at 10:15:29PM +0300, Andrey Korolyov wrote:
>> On Sun, Jun 7, 2015 at 7:48 PM, Ben Pfaff wrote:
>> > Setting CFLAGS by hand before invoking dpkg-buildflags is ineffective,
>> > because dpkg-buildflags overrides it.
>> >
>> >
On Sun, Jun 7, 2015 at 1:50 PM, Ben Pfaff wrote:
> Acked-by: Ben Pfaff
>
Thx, applied to master, no need to backport,~
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Sun, Jun 07, 2015 at 01:33:51PM -0700, Alex Wang wrote:
> This commit fixes unit test failure caused by commit f171fa8 (odp-util:
> Make sure vlan tci mask has exact match for VLAN_CFI.).
>
> Signed-off-by: Alex Wang
I tested that this fixes the test failure for me.
Acked-by: Ben Pfaff
This commit fixes unit test failure caused by commit f171fa8 (odp-util:
Make sure vlan tci mask has exact match for VLAN_CFI.).
Signed-off-by: Alex Wang
---
tests/ofproto-dpif.at |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif
On Sun, Jun 07, 2015 at 10:15:29PM +0300, Andrey Korolyov wrote:
> On Sun, Jun 7, 2015 at 7:48 PM, Ben Pfaff wrote:
> > Setting CFLAGS by hand before invoking dpkg-buildflags is ineffective,
> > because dpkg-buildflags overrides it.
> >
> > Reported-by: Andrey Korolyov
> > Signed-off-by: Ben Pfaf
On Sun, Jun 7, 2015 at 7:48 PM, Ben Pfaff wrote:
> Setting CFLAGS by hand before invoking dpkg-buildflags is ineffective,
> because dpkg-buildflags overrides it.
>
> Reported-by: Andrey Korolyov
> Signed-off-by: Ben Pfaff
> ---
> debian/rules | 9 +
> 1 file changed, 1 insertion(+), 8 d
Applied to master, branch-2.3, branch-2.1,
On Sun, Jun 7, 2015 at 11:33 AM, Alex Wang wrote:
> >
>> > Before this commit, this is not enforced, so OpenFlow flow like
>> > "vlan_tci=0x000a/0x0fff,action=output:local" can generate datapath
>> > flow like "vlan(vid=5/0xfff,pcp=0/0x0,cfi=1/0)".
>> >
On Tue, Jun 02, 2015 at 01:55:28PM -0400, Thomas F Herbert wrote:
> From: "Thomas F. Herbert"
>
> Changes to allow the tpid to be specified and all vlan tpid checking to be
> generalized.
>
> Signed-off-by: Thomas F Herbert
Applied to master, thanks!
___
>
> >
> > Before this commit, this is not enforced, so OpenFlow flow like
> > "vlan_tci=0x000a/0x0fff,action=output:local" can generate datapath
> > flow like "vlan(vid=5/0xfff,pcp=0/0x0,cfi=1/0)".
> >
>
> I'd think exact match on hex 'a' would yield vid=10 instead of 5.
This really bites me!, I
Sorry for missing this discussion,
Yeah, this is the correct fix. I think I must have put it there since I
wanted my
own display to be sorted... But I should really do it in my own .bashrc.
Will apply the patch soon,
Acked-by: Alex Wang
Thanks,
Alex Wang,
On Sun, Jun 7, 2015 at 11:16 AM, Be
[adding Alex, the author of the script]
On Thu, May 28, 2015 at 10:57:42AM -0300, Flavio Leitner wrote:
> On Thu, May 28, 2015 at 12:07:14PM +0800, Zang MingJie wrote:
> > Signed-off-by: Zang MingJie
> > ---
> > utilities/ovs-appctl-bashcomp.bash | 5 +
> > 1 file changed, 1 insertion(+), 4
On Mon, Jun 01, 2015 at 03:34:52PM -0700, Andy Zhou wrote:
> Measure user space only instruction counters for commit each
> transaction. This measurement is mainly useful for
> tuning OVSDB internal implementation, such as how OVSDB scales over
> number of client connections.
>
> A simple usage ex
On Mon, Jun 01, 2015 at 03:34:51PM -0700, Andy Zhou wrote:
> Replace the original PERF_FUNCTION_BEGIN and PERF_FUNCTION_END pair
> with a single PERF macro. This design is also more flexible, removing
> the restriction of have only one measurement per function.
>
> The next patch will make use of
On Mon, Jun 01, 2015 at 06:22:20PM -0700, Joe Stringer wrote:
> This function doesn't need to be exported in the public OVS headers, and
> it had an inconsistent name compared to uuid_equals(). Rename and move.
>
> Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
On Mon, Jun 01, 2015 at 06:22:19PM -0700, Joe Stringer wrote:
> Unique flow identifiers are really a UUID of sorts, so it makes sense to
> reuse the UUID string representations for UFID.
>
> Suggested-by: Ben Pfaff
> Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
_
On Mon, Jun 01, 2015 at 01:49:30PM -0700, Jesse Gross wrote:
> We have a special flow_metadata structure to represent the parts
> of a packet that aren't carried in the payload itself. This is
> used in the case where we need to send the packet as a Packet In
> to an OpenFlow controller. This is a
Two small nits, otherwise:
Acked-by: Jarno Rajahalme
Jarno
> On Jun 6, 2015, at 23:37, Alex Wang wrote:
>
> OVS datapath has check which prevents the installation of flow
> that matches VLAN TCI but does not have exact match for VLAN_CFI
> bit. To follow this rule, ovs userspace must mak
On Mon, Jun 01, 2015 at 11:24:19AM -0700, Alex Wang wrote:
> This commit implements a prototype of Header Space Analysis of the
> OVS OpenFlow table.
>
> What It Does
>
>
> 1. Dump all OpenFlow rules from a specified bridge.
> 2. Generate all-unmasked header space with specified
On Mon, Jun 01, 2015 at 11:24:18AM -0700, Alex Wang wrote:
> This commit adds functions that check if the 'struct flow_wildcards'
> member is fully masked (exact-match) or fully unmasked (don't care
> the entire field at all). Also, this commit adds a function to
> change a more wildcarded flow to
On Mon, Jun 01, 2015 at 11:24:17AM -0700, Alex Wang wrote:
> For conducting Header Space Analysis (HSA), we convert the wildcarded
> OpenFlow flow represented by 'struct match' into an encoded byte array.
> To further save memory, we use a sparse array to represent such byte
> array in the same way
Setting CFLAGS by hand before invoking dpkg-buildflags is ineffective,
because dpkg-buildflags overrides it.
Reported-by: Andrey Korolyov
Signed-off-by: Ben Pfaff
---
debian/rules | 9 +
1 file changed, 1 insertion(+), 8 deletions(-)
diff --git a/debian/rules b/debian/rules
index 38ecd
Debian likes to enable -Wformat-zero-length, even over our code trying to
disable it. It isn't too hard to make our code warning-free against this
option, so this commit both stops disabling it and fixes the warnings.
The first fix is to change set_subprogram_name() to take a plain string
instead
On Sat, Jun 06, 2015 at 11:37:08PM -0700, Alex Wang wrote:
> OVS datapath has check which prevents the installation of flow
> that matches VLAN TCI but does not have exact match for VLAN_CFI
> bit. To follow this rule, ovs userspace must make sure the
> flow key for datapath flow matching VLAN TCI
Send me a F#ckFriends request so we can hook up
My usename is BadChristina93
http://cc4.co/SOPXC";>my profile is here CANT WAIT
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
26 matches
Mail list logo
