OK, can you add this to the commit message (assuming Jesse doesn't
disagree)? Thanks.
On Tue, Nov 06, 2012 at 05:29:24PM -0800, Pravin Shelar wrote:
> Most of protocol in kernel are registered with RTNL lock, so GRE protocol
> registration shld also use same lock rather than introducing new lock.
On Tue, Nov 6, 2012 at 6:32 PM, Jesse Gross wrote:
> If a key used to find a port then we mark it as being not present
> (previously this was accomplished by setting it to zero). However,
> we also still pass the key to userspace which could lead to an
> inconsistent state. This also zeros out
On Nov 6, 2012, at 8:45 PM, Jesse Gross wrote:
> On Tue, Nov 6, 2012 at 6:39 PM, Kyle Mestery (kmestery)
> wrote:
>
> On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
>
> > Tunnel ports now always include full outer IP information, even if
> > userspace can't understand it. Since our flows o
On Tue, Nov 6, 2012 at 6:39 PM, Kyle Mestery (kmestery)
wrote:
>
> On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
>
> > Tunnel ports now always include full outer IP information, even if
> > userspace can't understand it. Since our flows our exact match this
> > information must also be provided
On Nov 6, 2012, at 8:32 PM, Jesse Gross wrote:
> If a key used to find a port then we mark it as being not present
> (previously this was accomplished by setting it to zero). However,
> we also still pass the key to userspace which could lead to an
> inconsistent state. This also zeros out the k
On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
> Tunnel ports now always include full outer IP information, even if
> userspace can't understand it. Since our flows our exact match this
> information must also be provided when setting up flows. Since flows
> with only OVS_KEY_ATTR_TUN_ID keys
On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
> When the IPV4_TUNNEL action is executed, a pointer in the skb is
> directly assigned the address of the action, which is protected by
> RCU. If a TUN_ID action is later executed it will write into the
> action, which is not allowed by RCU. This av
On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
> There's no reason why GRE64 can't be configured using flow based
> tunneling. Since that's the direction we are moving, everything
> should support that. Once this special casing is removed, it's
> possible to unify the tunnel parameter parsing co
On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
> The names for the flags used by flow based tunneling are pretty long.
> This shortens them a little by removing the word FLOW, which is a
> distinction that won't be meaningful in the near future.
>
> Signed-off-by: Jesse Gross
Acked-by: Kyle Me
On Nov 6, 2012, at 6:50 PM, Jesse Gross wrote:
> During development it was preferable to keep OVS_KEY_ATTR_IPV4_TUNNEL
> in the non-upstream range of identifiers to avoid conflicts or
> compatibility issues as it evolved. However, since the intention is
> to get it upstream, it makes sense to mov
If a key used to find a port then we mark it as being not present
(previously this was accomplished by setting it to zero). However,
we also still pass the key to userspace which could lead to an
inconsistent state. This also zeros out the key as before in order
to make our reported results consi
Most of protocol in kernel are registered with RTNL lock, so GRE protocol
registration shld also use same lock rather than introducing new lock.
On Tue, Nov 6, 2012 at 5:09 PM, Ben Pfaff wrote:
> On Tue, Nov 06, 2012 at 04:59:01PM -0800, Pravin B Shelar wrote:
> > Currently gre demultiplexer us
On Tue, Nov 06, 2012 at 04:59:01PM -0800, Pravin B Shelar wrote:
> Currently gre demultiplexer uses its own spin-lock. We should rather
> RTNL lock.
The commit message doesn't make it clear why this change is desirable.
___
dev mailing list
dev@openvswit
Use UDP Tunnel segmention.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 27 +++
1 file changed, 23 insertions(+), 4 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 983a239..07e43bb 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net
Signed-off-by: Pravin B Shelar
---
include/linux/skbuff.h |3 ++
net/ipv4/af_inet.c |6 ++-
net/ipv4/udp.c | 106
net/ipv6/af_inet6.c|1 +
net/ipv6/udp.c |7 +++-
5 files changed, 96 insertions(+), 27 deleti
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 22 ++
1 file changed, 22 insertions(+)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 8faaeb8..983a239 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -638,6 +638,21 @@ static netdev_tx_t
IP Tunnel modules defines generic rtnl-netlink ops for tunneling
devices VXLAN can directly use them.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 105 ++-
1 file changed, 37 insertions(+), 68 deletions(-)
diff --git a/drivers/net/vxl
Export vxlan_create_socket() so that other modules, like Open vSwitch,
can have configurable destination port for VXLAN.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 46 +++---
1 file changed, 23 insertions(+), 23 deletions(-)
diff --git a/d
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 26 ++
1 file changed, 2 insertions(+), 24 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index be2064b..dacd20f 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -604,28 +604,6 @@
IP Tunnel protocol handler allows multiple modules register for
same protocol, e.g. linux vxlan devices and Open vSwitch both
can register for VXLAN device.
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 68 +++
1 file changed, 53 inser
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 119 ---
1 file changed, 56 insertions(+), 63 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 030559d..c8939ab2 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/v
Signed-off-by: Pravin B Shelar
---
drivers/net/vxlan.c | 71 +--
1 file changed, 7 insertions(+), 64 deletions(-)
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index c8939ab2..d2449f3 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vx
Signed-off-by: Pravin B Shelar
---
include/linux/skbuff.h | 12 ++
net/ipv4/af_inet.c |1 +
net/ipv4/gre.c | 95
net/ipv4/ip_gre.c |7
net/ipv4/tcp.c |1 +
net/ipv4/udp.c |3 +-
net/ipv6
Reuse common ip-tunneling code which is re-factored from GRE module.
Signed-off-by: Pravin B Shelar
---
net/ipv4/ipip.c | 743 ---
1 file changed, 101 insertions(+), 642 deletions(-)
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index d3b004
Currently gre demultiplexer uses its own spin-lock. We should rather
RTNL lock.
Suggested-by: Stephen Hemminger
Signed-off-by: Pravin B Shelar
---
drivers/net/ppp/pptp.c |5 +
net/ipv4/gre.c | 18 +-
net/ipv4/ip_gre.c |5 +
3 files changed, 15 inse
On Tue, Nov 06, 2012 at 04:49:07PM -0800, Ben Pfaff wrote:
> On Wed, Nov 07, 2012 at 09:45:59AM +0900, Simon Horman wrote:
> > Do you have any thoughts on if it is worth implementing something
> > similar for bitmap_scan() using ffs()?
>
> I've never seen bitmap_scan() show up in profiling data so
Tunnel ports now always include full outer IP information, even if
userspace can't understand it. Since our flows our exact match this
information must also be provided when setting up flows. Since flows
with only OVS_KEY_ATTR_TUN_ID keys don't contain all of this information
they can never be hi
There's no reason why GRE64 can't be configured using flow based
tunneling. Since that's the direction we are moving, everything
should support that. Once this special casing is removed, it's
possible to unify the tunnel parameter parsing code since both
GRE and CAPWAP are the same.
Signed-off-b
The names for the flags used by flow based tunneling are pretty long.
This shortens them a little by removing the word FLOW, which is a
distinction that won't be meaningful in the near future.
Signed-off-by: Jesse Gross
---
datapath/flow.c | 12 ++--
datapath/tunnel.c
When the IPV4_TUNNEL action is executed, a pointer in the skb is
directly assigned the address of the action, which is protected by
RCU. If a TUN_ID action is later executed it will write into the
action, which is not allowed by RCU. This avoids the problem by
making a copy of the data and writin
During development it was preferable to keep OVS_KEY_ATTR_IPV4_TUNNEL
in the non-upstream range of identifiers to avoid conflicts or
compatibility issues as it evolved. However, since the intention is
to get it upstream, it makes sense to move it down now to avoid issues
with compatibility when up
On Wed, Nov 07, 2012 at 09:45:59AM +0900, Simon Horman wrote:
> Do you have any thoughts on if it is worth implementing something
> similar for bitmap_scan() using ffs()?
I've never seen bitmap_scan() show up in profiling data so it's a low
priority.
The current implementation has the advantage o
On Sun, Nov 04, 2012 at 10:27:48PM -0800, Ben Pfaff wrote:
> On Fri, Oct 26, 2012 at 11:36:06AM +0900, Simon Horman wrote:
> > Make use of __builtin_clz if available which should optimize
> > ofputil_version_bitmap_scanr() by replacing a loop with
> > a single CLZ instruction when available.
> >
>
On Sun, Nov 04, 2012 at 10:20:09PM -0800, Ben Pfaff wrote:
> On Thu, Nov 01, 2012 at 08:33:04AM +0800, Simon Horman wrote:
> > On Wed, Oct 31, 2012 at 11:16:06AM -0700, Ben Pfaff wrote:
> > > On Fri, Oct 26, 2012 at 11:35:51AM +0900, Simon Horman wrote:
> > > > Allow encoding and decoding of versio
Hey Jesse,
This sounds like a plan. It is pretty current (I think I last updated my github
last week). I'll go ahead and update things, make sure things work, and apply
your comments below.
Thanks!
Kyle
On Nov 6, 2012, at 1:27 PM, Jesse Gross
wrote:
> Hi Kyle,
>
> Things are continuing to
Hi Kyle,
Things are continuing to move along and I think that most of the major
infrastructure pieces are either complete or coming along nicely. One area
that might make sense for you to work on is the VXLAN dataplane support. I
know that you've been maintaining Ben's patch out of tree so hopef
Populate the interface table's 'ofport_request' column
with the value in the column 'ofport' while saving flows.
When vswitchd restarts, it will use the ofport_request
value to set the value of ofport.
Signed-off-by: Gurucharan Shetty
---
utilities/ovs-save | 19 ++-
1 file ch
Now that we have a single datapath and the datapath
port number has no relation to openflow port numbers,
it does not make sense to save and restore kernel
datapath configuration for the purpose of restoring
openflow flows.
Signed-off-by: Gurucharan Shetty
---
utilities/ovs-ctl.8 |9 +++---
On Tue, Nov 06, 2012 at 08:41:50PM +0200, Ansis Atteka wrote:
> Implementation of the IPv6 'set' action depends on ipv6_find_exthdr()
> function to find routing header. By looking at the routing headers,
> it is possible to tell, whether L4 checksums will need to be
> recalculated, whenever the des
On Tue, Nov 6, 2012 at 10:41 AM, Ansis Atteka wrote:
> Implementation of the IPv6 'set' action depends on ipv6_find_exthdr()
> function to find routing header. By looking at the routing headers,
> it is possible to tell, whether L4 checksums will need to be
> recalculated, whenever the destinatio
This patch prepares ipv6_find_hdr() function so that it could be
able to skip routing headers, where segements_left is 0. This is
required for us to handle multiple routing header case correctly.
Signed-off-by: Ansis Atteka
---
datapath/linux/compat/exthdrs_core.c | 31
Implementation of the IPv6 'set' action depends on ipv6_find_exthdr()
function to find routing header. By looking at the routing headers,
it is possible to tell, whether L4 checksums will need to be
recalculated, whenever the destination address is changed in
the main IPv6 header.
This function wi
On Tue, Nov 6, 2012 at 8:26 PM, Ben Pfaff wrote:
> On Tue, Nov 06, 2012 at 08:25:41PM +0200, Ansis Atteka wrote:
>> On Tue, Nov 6, 2012 at 8:19 PM, Ben Pfaff wrote:
>> > On Tue, Nov 06, 2012 at 10:16:47AM -0800, Jesse Gross wrote:
>> >> On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
>> >>
>
On Tue, Nov 06, 2012 at 08:25:41PM +0200, Ansis Atteka wrote:
> On Tue, Nov 6, 2012 at 8:19 PM, Ben Pfaff wrote:
> > On Tue, Nov 06, 2012 at 10:16:47AM -0800, Jesse Gross wrote:
> >> On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
> >>
> >> > This patch prepares ipv6_find_hdr() function so th
On Tue, Nov 6, 2012 at 8:19 PM, Ben Pfaff wrote:
> On Tue, Nov 06, 2012 at 10:16:47AM -0800, Jesse Gross wrote:
>> On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
>>
>> > This patch prepares ipv6_find_hdr() function so that it could be
>> > able to skip routing headers, where segements_left i
On Tue, Nov 6, 2012 at 10:21 AM, Ansis Atteka wrote:
> On Tue, Nov 6, 2012 at 8:16 PM, Jesse Gross wrote:
> > On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
> >>
> >> This patch prepares ipv6_find_hdr() function so that it could be
> >> able to skip routing headers, where segements_left is
On Tue, Nov 6, 2012 at 10:03 AM, Ahsan Rafique <
ahsan.rafi...@xflowresearch.com> wrote:
> **
> Hi,
> I have added the nvgre support in OVS.Now i want to do interop testing of
> this tunnel with the microsoft windows server 2012.The ping is successful
> when i ping from one host to another both
On Tue, Nov 06, 2012 at 10:16:47AM -0800, Jesse Gross wrote:
> On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
>
> > This patch prepares ipv6_find_hdr() function so that it could be
> > able to skip routing headers, where segements_left is 0. This is
> > required for us to handle multiple rou
On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
> This patch prepares ipv6_find_hdr() function so that it could be
> able to skip routing headers, where segements_left is 0. This is
> required for us to handle multiple routing header case correctly.
>
> Signed-off-by: Ansis Atteka
This loo
Hi,
I have added the nvgre support in OVS.Now i want to do interop testing of this
tunnel with the microsoft windows server 2012.The ping is successful when i ping
from one host to another both having OVS.But when i ping from host 1 having OVS
with nvgre support to host 2 having windows server 2012
On 11/06/2012 11:10 AM, Ben Pfaff wrote:
> On Fri, Nov 02, 2012 at 02:26:20PM -0700, Ben Pfaff wrote:
>> On Fri, Nov 02, 2012 at 04:21:56PM -0500, Adam Heath wrote:
>>> On 11/02/2012 04:19 PM, Ben Pfaff wrote:
I still don't understand. Why would an existing script call ovs-vsctl
with --m
On Tue, Nov 6, 2012 at 7:25 AM, Ansis Atteka wrote:
> Implementation of the IPv6 'set' action depends on ipv6_find_exthdr()
> function to find routing header. By looking at the routing headers,
> it is possible to tell, whether L4 checksums will need to be
> recalculated, whenever the destination
On Mon, Nov 5, 2012 at 8:01 PM, Ben Pfaff wrote:
> Jesse has been back for a week or so.
>
> He's focusing on the tunneling changes. I hope that he will get to MPLS
> soon after that is ready to go.
Yes, that sounds right.
___
dev mailing list
dev@op
On Mon, Nov 05, 2012 at 06:29:02PM -0800, Ethan Jackson wrote:
> This patch fixes the following warning caused by a switch case
> which was not handled.
>
> lib/netdev-vport.c:144:5: error: enumeration value
> ‘OVS_VPORT_TYPE_FT_GRE’ not handled in switch
>
> Signed-off-by: Ethan Jackson
Looks
On Mon, Nov 5, 2012 at 6:29 PM, Ethan Jackson wrote:
> This patch fixes the following warning caused by a switch case
> which was not handled.
>
> lib/netdev-vport.c:144:5: error: enumeration value
> ‘OVS_VPORT_TYPE_FT_GRE’ not handled in switch
>
> Signed-off-by: Ethan Jackson
Looks good, tha
On Mon, Nov 5, 2012 at 5:51 PM, Pravin Shelar wrote:
> On Mon, Nov 5, 2012 at 3:50 PM, Jesse Gross wrote:
>>
>> diff --git a/datapath/datapath.c b/datapath/datapath.c
>> index 9c253e1..976a5f2 100644
>> --- a/datapath/datapath.c
>> +++ b/datapath/datapath.c
>> @@ -818,13 +818,13 @@ static int ov
On Fri, Nov 02, 2012 at 02:26:20PM -0700, Ben Pfaff wrote:
> On Fri, Nov 02, 2012 at 04:21:56PM -0500, Adam Heath wrote:
> > On 11/02/2012 04:19 PM, Ben Pfaff wrote:
> > > I still don't understand. Why would an existing script call ovs-vsctl
> > > with --may-exist as a global option? The call wou
This patch prepares ipv6_find_hdr() function so that it could be
able to skip routing headers, where segements_left is 0. This is
required for us to handle multiple routing header case correctly.
Signed-off-by: Ansis Atteka
---
datapath/linux/compat/exthdrs_core.c | 41
Implementation of the IPv6 'set' action depends on ipv6_find_exthdr()
function to find routing header. By looking at the routing headers,
it is possible to tell, whether L4 checksums will need to be
recalculated, whenever the destination address is changed in
the main IPv6 header.
This function wi
Jesse and Pravin:
Curious what the current state of the tunnel work is. About 4 weeks back, Jesse
had sent out an email with some work items. I have some cycles now to continue
working on this, but wanted to touch base with you guys to see where things are
at and what areas you may need help wi
Bonjour Bien Aimé
Ce sont nos œuvres qui resteront quand tombera sur nous, le rideau de la
mort.
En mourant nous n'emportons aucun bien matériel avec nous. Les prières et la
foi en son seigneur valent beaucoup plus que l'or ; le diamant ou tous autres
pier
On Tue, Nov 06, 2012 at 05:03:22PM +0900, Isaku Yamahata wrote:
> Except minor typo,
Thanks, I will fix the typo.
> Reviewed-by: Isaku Yamahata
>
> thanks,
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Except minor typo,
Reviewed-by: Isaku Yamahata
thanks,
On Mon, Nov 05, 2012 at 10:22:34AM +0800, Simon Horman wrote:
> Hi,
>
> This series implements basic MPLS actions and matches based on work by
> Ravi K, Leo Alterman and Yamahata-san.
>
> The main limitation of this implementation is that
Minor nitpick
On Mon, Nov 05, 2012 at 10:22:40AM +0800, Simon Horman wrote:
> This adds support for the OpenFlow 1.1+ copy_ttl_out action.
typo. copy-n-paste
> And also adds an NX copy_ttl_out action.
>
> The implementation does not support copy
64 matches
Mail list logo
