On Wed, Oct 03, 2012 at 12:17:16PM -0700, Jesse Gross wrote:
> From: Leo Alterman
>
> Allow datapath to recognize and extract MPLS labels into flow keys
> and execute actions which push, pop, and set labels on packets.
>
> Signed-off-by: Leo Alterman
Hi,
As a method of reviewing this I've be
Thanks for the time. I found my mistake. Without the use of a key to
differentiate between the two tunnels, so if you give a different key to
both tunnels, you can set up as many as you would want.
But I do still have this query of GRE handler. Can it only be invoked once?
I try to use it for NVGR
On Thu, Oct 4, 2012 at 1:58 AM, Simon Horman wrote:
> On Wed, Oct 03, 2012 at 12:17:16PM -0700, Jesse Gross wrote:
>> From: Leo Alterman
>>
>> Allow datapath to recognize and extract MPLS labels into flow keys
>> and execute actions which push, pop, and set labels on packets.
>>
>> Signed-off-by:
You can't have multiple handlers for the same IP type because there
would be no way to distinguish which one to hand packets to.
However, NV-GRE shouldn't be implemented directly in the kernel
anyways. There is a lot of tunnel work ongoing which is going to both
completely change the tunnel code
On Wed, Oct 03, 2012 at 05:32:40PM -0700, Ethan Jackson wrote:
> In the ESX build system, autom4te is in a funny location specified
> by an environment variable. Before this patch, this environment
> variable was overridden by the tests automake file.
>
> Signed-off-by: Ethan Jackson
We can't (
On Wed, Oct 03, 2012 at 05:32:39PM -0700, Ethan Jackson wrote:
> For some reason, the ESX build tools seem to be confused about the
> location of 'lib'. This patch specifies it more directly.
>
> Signed-off-by: Ethan Jackson
Looks good, thanks.
(There are lots of pitfalls in Makefiles, though,
On Thu, Sep 20, 2012 at 01:10:41PM +0900, Simon Horman wrote:
> This allows for encoding and decoding Open Flow 1.1 and 1.2 Port Stats
> Request and Reply message
>
> Signed-off-by: Simon Horman
I see a few careless errors here. print_port_stat() should have its
return type on a separate line.
On Thu, Oct 04, 2012 at 11:41:15AM +0900, Simon Horman wrote:
> This seires adds a command line options to ovs-vswitchd and various
> utilities to allow the range of Open Flow versions to be set at run-time.
>
> The options are --max-ofp-version and --min-ofp-version.
> The values accepted are: Op
On Thu, Oct 04, 2012 at 11:41:16AM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman
Applied to master, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Thu, Oct 04, 2012 at 11:41:20AM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman
I think the commit message should mention connmgr_run() instead of
connmgr_destroy(). The commit doesn't affect connmgr_destroy()
as far as I can tell.
The layering seems wrong to me. I don't think that
On Thu, Oct 04, 2012 at 11:41:21AM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman
I think this becomes unnecessary if we move the acceptable versions to
pvconn_open().
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/lis
Thanks, Ralf. I will add
Signed-off-by: Ralf Spenneberg
to the commit log.
On Thu, Oct 04, 2012 at 07:40:55AM +0200, Ralf Spenneberg wrote:
> Hi,
>
> sorry for the delay. Yesterday was a national holiday in Germany. Yes I sure
> sign off this work.
>
> Kind regards,
>
> Ralf
> Am Die
Thank you for the reply Jesse. I was thinking of implementing it in the
same way as Pravin implemented GRE64 i.e. without actually defining a
separate protocol handler since NV-GRE is merely an extension to GRE.
I have been trying to keep myself up to date with what you guys have been
doing. I hop
On Tue, Sep 11, 2012 at 4:21 AM, Pravin B Shelar wrote:
> From: Pravin Shelar
>
> Extend GRE to have a 64-bit key. Use GRE sequence number to store
> upper 32-bits of the key.
>
> Bug #13186
> Signed-off-by: Pravin B Shelar
I received a sparse warning when applying this:
CHECK /home/jgross
Yep, ran that before sending the patch to review. Works without problems.
Ethan
> (There are lots of pitfalls in Makefiles, though, so I recommend
> running "make distcheck" before pushing it.)
___
dev mailing list
dev@openvswitch.org
http://openvswitc
Thanks, that's reassuring.
On Thu, Oct 04, 2012 at 12:53:12PM -0700, Ethan Jackson wrote:
> Yep, ran that before sending the patch to review. Works without problems.
>
> Ethan
>
> > (There are lots of pitfalls in Makefiles, though, so I recommend
> > running "make distcheck" before pushing it.)
On Oct 3, 2012, at 11:38 AM, Jesse Gross wrote:
> On Mon, Oct 1, 2012 at 8:52 AM, Kyle Mestery wrote:
>> This is a first pass at providing a tun_key which can be
>> used as the basis for flow-based tunnelling. The
>> tun_key includes and replaces the tun_id in both struct
>> ovs_skb_cb and struct
On Thu, Oct 4, 2012 at 1:12 PM, Kyle Mestery (kmestery)
wrote:
> On Oct 3, 2012, at 11:38 AM, Jesse Gross wrote:
>> Also, when I was working on the userspace portions I defined three
>> flags to be used instead of the current tunnel port configuration:
>>
>> Flow lib/flow.h:
>> #define FLOW_TNL_F_
In the ESX build system, autom4te is in a funny location specified
by an environment variable. Before this patch, this environment
variable was overridden by the tests automake file.
Signed-off-by: Ethan Jackson
---
configure.ac | 5 +
tests/automake.mk | 1 -
2 files changed, 5 insert
If an attempt is made to transmit a packet that is over the device's
MTU then we log it using the datapath's name. However, it is much
more helpful to use the device name instead.
Signed-off-by: Jesse Gross
---
datapath/vport-netdev.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
di
On Thu, Oct 04, 2012 at 01:34:53PM -0700, Ethan Jackson wrote:
> In the ESX build system, autom4te is in a funny location specified
> by an environment variable. Before this patch, this environment
> variable was overridden by the tests automake file.
>
> Signed-off-by: Ethan Jackson
s/auto4mte
On Oct 4, 2012, at 3:37 PM, Jesse Gross wrote:
> If an attempt is made to transmit a packet that is over the device's
> MTU then we log it using the datapath's name. However, it is much
> more helpful to use the device name instead.
>
> Signed-off-by: Jesse Gross
Much more helpful.
Acked-by: K
Thanks, fixed and merged.
Ethan
On Thu, Oct 4, 2012 at 1:38 PM, Ben Pfaff wrote:
> On Thu, Oct 04, 2012 at 01:34:53PM -0700, Ethan Jackson wrote:
>> In the ESX build system, autom4te is in a funny location specified
>> by an environment variable. Before this patch, this environment
>> variable
On Thu, Oct 4, 2012 at 1:41 PM, Kyle Mestery (kmestery)
wrote:
> On Oct 4, 2012, at 3:37 PM, Jesse Gross wrote:
>> If an attempt is made to transmit a packet that is over the device's
>> MTU then we log it using the datapath's name. However, it is much
>> more helpful to use the device name inste
Commit 296e07ace0f (flow: Extend struct flow to contain tunnel outer
header.) changed the tunnel ID parameter of flow_extract() from an integer
passed by value to a structure passed by pointer. Before flow_extract()
reads the tunnel ID, it zeros the entire flow parameter. This means that,
if a ca
Before commit e3a880272 (lib: Specify idl location more precisely.), the
files named in VSWITCH_IDL_FILES were relative to the source directory.
That commit made them inconsistent: one remained relative to the source
directory, the other became relative to the build directory. This meant
that if t
Commit 7fd3e43db07 (tests: Allow environment specified autom4te.) added a
requirement for Autom4te without anyone noticing it in advance. This
caused a build failure in the XenServer DDK, which doesn't have Autom4te.
Since Autom4te is only needed if the testsuite is modified, this commit
makes it
Hmm strange, this worked for me without the patch when I tried
building from an alternative directory. At any rate, thanks for
fixing this. s/change/chance in the commit message and it's good.
Acked-by: Ethan Jackson
On Thu, Oct 4, 2012 at 3:55 PM, Ben Pfaff wrote:
> Before commit e3a880272
Acked-by: Ethan Jackson
On Thu, Oct 4, 2012 at 3:55 PM, Ben Pfaff wrote:
> Commit 7fd3e43db07 (tests: Allow environment specified autom4te.) added a
> requirement for Autom4te without anyone noticing it in advance. This
> caused a build failure in the XenServer DDK, which doesn't have Autom4te
Thanks, I pushed these.
On Thu, Oct 04, 2012 at 04:16:08PM -0700, Ethan Jackson wrote:
> Hmm strange, this worked for me without the patch when I tried
> building from an alternative directory. At any rate, thanks for
> fixing this. s/change/chance in the commit message and it's good.
>
> Acked
On Thu, Oct 4, 2012 at 3:11 PM, Ben Pfaff wrote:
> Commit 296e07ace0f (flow: Extend struct flow to contain tunnel outer
> header.) changed the tunnel ID parameter of flow_extract() from an integer
> passed by value to a structure passed by pointer. Before flow_extract()
> reads the tunnel ID, it
On Thu, Oct 04, 2012 at 04:55:59PM -0700, Jesse Gross wrote:
> On Thu, Oct 4, 2012 at 3:11 PM, Ben Pfaff wrote:
> > Commit 296e07ace0f (flow: Extend struct flow to contain tunnel outer
> > header.) changed the tunnel ID parameter of flow_extract() from an integer
> > passed by value to a structure
On Thu, Oct 04, 2012 at 08:52:52AM -0700, Jesse Gross wrote:
> On Thu, Oct 4, 2012 at 1:58 AM, Simon Horman wrote:
> > On Wed, Oct 03, 2012 at 12:17:16PM -0700, Jesse Gross wrote:
> >> From: Leo Alterman
> >>
> >> Allow datapath to recognize and extract MPLS labels into flow keys
> >> and execute
The timer_create() system call is not supported in ESX and returns
an error when called. Aborting when this system call fails seems a
bit extreme. So instead, this patch simply falls back to disabling
the cached time optimization.
Signed-off-by: Ethan Jackson
---
lib/timeval.c | 33
Before this patch, time_alarm() used the SIGALRM handler to notify
the poll loop that it should exit the program. Instead, this patch
simply implements time_alarm() directly in the pool loop. This
significantly simplifies the code, while removing a call to
timer_create() which is not currently su
v1-v2:
- Fixed according to comments from Ben and Jesse.
--8<--cut here-->8--
Extend GRE to have a 64-bit key. Use GRE sequence number to
store upper 32-bits of the key, this is not standard way of
using GRE sequence number.
Bug #13186
Signe
On Thu, Oct 04, 2012 at 05:33:12AM -0700, Pravin B Shelar wrote:
> +- 64-bit GRE: GRE with 64-bit key.
How about a little more elaborate: "New support for a nonstandard form
of GRE that supports a 64-bit key."
___
dev mailing list
dev@openvswitch.org
Hey Isaku,
Thanks for the patch.
I have applied it to master.
thanx!
mehak
On Wed, Sep 26, 2012 at 12:12 AM, Isaku Yamahata wrote:
> Thus ovsdb-client aborts as follows.
>
> > # ovs-vsctl set-manager ptcp:6634
> > # ovsdb-client get-schema tcp:127.0.0.1:6634
> > 2012-09-14T05:38:26Z|1|jsonr
This patch adds instruction OF11 apply-actions/goto-table supports.
Write-actions support is on my TODO.
Change v5 -> v6:
- revised syntax
- eliminated check of zero padding
- fixed unit test
- dropped half baked write-actions patch
Change v4 -> v5:
- improved ofpact_instruction_name_from_type
-
Signed-off-by: Isaku Yamahata
---
v5
- manual rebase
- remove check of zero padding
v4
- clear-actions part is split outed.
- check zero padding
- man page
---
lib/ofp-actions.c| 28 ++--
lib/ofp-actions.h| 13 -
lib/ofp-parse.c |
Signed-off-by: Isaku Yamahata
---
v6
- manual rebase
- simplify logic
- remove all zero check
- update unit tests
v5
- manual rebase
v4
- squashed goto-table instruction part into single patch
- only introduce goto-table. Other instruction will be addressed
by other patches.
- check zero paddi
41 matches
Mail list logo
