On Tue, Sep 11, 2012 at 10:37:48PM -0700, Ben Pfaff wrote:
> On Tue, Sep 11, 2012 at 01:30:04PM +0900, Simon Horman wrote:
> > From: Isaku Yamahata
> >
> > Use a uninion mf_subvalue instead of a uint64_t for
> > the value member of struct ofpact_reg_load.
> >
> > set_field action needs to hold v
Hi,
This series adds preliminary support for set-field actions.
I have picked up this series from Yamahata-san on his request.
It is based on the current master branch.
Changes for each patch are noted inline.
I believe this series addresses raised by Ben Pfaff in respect to
v4 of this series.
A
From: Isaku Yamahata
Signed-off-by: Isaku Yamahata
Signed-off-by: Simon Horman
---
v5 [Simon Horman]
* Do not add mf_check_dst_oxm(), just use mf_check_dst().
v4 [Simon Horman]
* Remove utilities/ovs-ofctl.c hunk, introduced in v3,
which belongs in a different patch
* Make use of new oxm_w
From: Isaku Yamahata
Use a uninion mf_subvalue instead of a uint64_t for
the value member of struct ofpact_reg_load.
set_field action needs to hold values wider than 64 bits.
This is preparation for set_field action.
Signed-off-by: Isaku Yamahata
Signed-off-by: Simon Horman
---
v5 [Simon Ho
By passing a flow to the action parser the pre-requisites
of set-feild actions will be checked. If the flow is NULL,
for instance in test code such as ofctl_parse_ofp11_instructions(),
then the check is skiped, as it always was before this change.
Unfortunately I don't think that this check is cor
From: Isaku Yamahata
Signed-off-by: Isaku Yamahata
Signed-off-by: Simon Horman
---
v5 [Simon Horman]
* No change
v4 [Simon Horman]
* Update call to ofpacts_pull_openflow11_instructions()
in ofctl_parse_ofp11_instructions() to include new ofp_version argument
* Add a special case for OFPAT1
From: Isaku Yamahata
Signed-off-by: Isaku Yamahata
Signed-off-by: Simon Horman
---
v5 [Simon Horman]
* No change
v4 [Simon Horman]
* Manual Rebase
* Add a limited number of tests
v3 [Simon Horman]
* Manual Rebase
- Add ofpacts_put_openflow12_instructions() and
ofpacts_put_openflow11_a
One note inline
On Sep 11, 2555 BE, at 23:17, Isaku Yamahata wrote:
> Since Transaction._substitute doesn't substitute elements of list/tuple,
> setting list references results in transaction error. Teach it such case.
>
> Example:
> {"op": "update",
> "row":{"bridges":["set",[["uuid",
>
Contextless phone review here also.
Perhaps raise a value error if location and schema are both provided?
Isaku/Ben, are there any side effects of keeping this data in memory before it
is used, and removing entirely it after the 'schema =' line? Before it loaded
it on demand. Isn't necessarily
On Tue, Sep 11, 2012 at 9:11 PM, Saleha Asad wrote:
>
> Yes the standard is not completely defined yet but Microsoft server 2012
> has this protocol implemented and they are using the following commands to
> allot VSIDs to the packet:
>
> New-NetVirtualizationLookupRecord -VirtualSubnetID “5001″
Patches 1/5, 2/5, 5/5 look fine to me.
On Sep 11, 2555 BE, at 23:17, Isaku Yamahata wrote:
> In some cases getattr(Row instance, attrname) doesn't raise AttributeError,
> but TypeError
>
>> File "python/ovs/db/idl.py", line 554, in __getattr__
>>datum = self._data[column_name]
>> TypeError:
On Wed, Sep 12, 2012 at 05:44:28PM +0900, Simon Horman wrote:
> From: Isaku Yamahata
>
> Use a uninion mf_subvalue instead of a uint64_t for
> the value member of struct ofpact_reg_load.
>
> set_field action needs to hold values wider than 64 bits.
> This is preparation for set_field action.
>
On Wed, Sep 12, 2012 at 05:44:29PM +0900, Simon Horman wrote:
> +enum ofperr
> +nxm_reg_load_from_openflow12_set_field(
> +const struct ofp12_action_set_field * oasf, struct ofpbuf *ofpacts)
> +{
> +uint16_t oasf_len = ntohs(oasf->len);
> +ovs_be32 *p = (ovs_be32*)oasf->field;
> +ui
This is what I intended when I suggested using mf_subvalue, but I didn't
notice the difference until after applying the patch.
Signed-off-by: Ben Pfaff
---
lib/learn.c |4 +++-
lib/meta-flow.c |2 +-
lib/nx-match.c| 25 +
lib/ofp-actions.h |2 +-
On Wed, Sep 12, 2012 at 03:17:22PM +0900, Isaku Yamahata wrote:
> add register_table method to SchemaHelper as counter part of
> ovsdb_idl_add_table().
>
> Signed-off-by: Isaku Yamahata
This seems like a good idea, but would you mind adding a doc comment
and updating comments on other functions
On Wed, Sep 12, 2012 at 03:17:21PM +0900, Isaku Yamahata wrote:
> The changeset of ed815d9bd28422a490fe370d7804d24bcd676806 eliminated
> the use of select.poll for eventlet/gevent.
> It forgot to select.poll in socket_util.
>
> Signed-off-by: Isaku Yamahata
Applied to master, thanks!
___
I agree about raising an error if both are provided.
There doesn't seem to be an issue with reading the file earlier. It
won't bother any in-tree users anyhow.
On Wed, Sep 12, 2012 at 08:41:55AM -0700, Reid Price wrote:
> Contextless phone review here also.
>
> Perhaps raise a value error if lo
Thanks as always for the reviews, Reid!
On Wed, Sep 12, 2012 at 08:49:49AM -0700, Reid Price wrote:
> Patches 1/5, 2/5, 5/5 look fine to me.
>
> On Sep 11, 2555 BE, at 23:17, Isaku Yamahata wrote:
>
> > In some cases getattr(Row instance, attrname) doesn't raise AttributeError,
> > but TypeErro
On Wed, Sep 12, 2012 at 05:44:30PM +0900, Simon Horman wrote:
> From: Isaku Yamahata
>
> Signed-off-by: Isaku Yamahata
> Signed-off-by: Simon Horman
This seems reasonable once the issues in patch 2 are cleaned up.
Thanks,
Ben.
___
dev mailing list
On Wed, Sep 12, 2012 at 05:44:31PM +0900, Simon Horman wrote:
> From: Isaku Yamahata
>
> Signed-off-by: Isaku Yamahata
> Signed-off-by: Simon Horman
I think that, if I'm reading this right, then attempting to convert
many actions to OF1.2 will cause an assertion failure. We need to do
somethi
On Wed, Sep 12, 2012 at 05:44:32PM +0900, Simon Horman wrote:
> By passing a flow to the action parser the pre-requisites
> of set-feild actions will be checked. If the flow is NULL,
> for instance in test code such as ofctl_parse_ofp11_instructions(),
> then the check is skiped, as it always was b
Cabestan. Layout Simple
Asegúrate de no perderte ninguna oferta, añade ofer...@globalbono.com a tu
lista de contactos.
Si no ves correctamente las imágenes, pulsa [
http://email.globalbono.com/E12092012134620.cfm?WL=1293&WS=208833_8707685&WA=610
] aquí
[ http://email.globalbono.com/Go/index.c
On Wed, Sep 12, 2012 at 11:37:53AM -0700, Ben Pfaff wrote:
> This is what I intended when I suggested using mf_subvalue, but I didn't
> notice the difference until after applying the patch.
>
> Signed-off-by: Ben Pfaff
Acked-by: Simon Horman
___
dev
On Wed, Sep 12, 2012 at 11:13:49AM -0700, Ben Pfaff wrote:
> On Wed, Sep 12, 2012 at 05:44:29PM +0900, Simon Horman wrote:
> > +enum ofperr
> > +nxm_reg_load_from_openflow12_set_field(
> > +const struct ofp12_action_set_field * oasf, struct ofpbuf *ofpacts)
> > +{
> > +uint16_t oasf_len = n
On Wed, Sep 12, 2012 at 12:06:50PM -0700, Ben Pfaff wrote:
> On Wed, Sep 12, 2012 at 05:44:31PM +0900, Simon Horman wrote:
> > From: Isaku Yamahata
> >
> > Signed-off-by: Isaku Yamahata
> > Signed-off-by: Simon Horman
>
> I think that, if I'm reading this right, then attempting to convert
> ma
On Wed, Sep 12, 2012 at 11:45:12AM -0700, Ben Pfaff wrote:
> On Wed, Sep 12, 2012 at 03:17:22PM +0900, Isaku Yamahata wrote:
> > add register_table method to SchemaHelper as counter part of
> > ovsdb_idl_add_table().
> >
> > Signed-off-by: Isaku Yamahata
>
> This seems like a good idea, but woul
On Wed, Sep 12, 2012 at 11:52:09AM -0700, Ben Pfaff wrote:
> I agree about raising an error if both are provided.
>
> There doesn't seem to be an issue with reading the file earlier. It
> won't bother any in-tree users anyhow.
Schema file isn't such large in practice. for example,
The size of vs
On Wed, Sep 12, 2012 at 08:32:15AM -0700, Reid Price wrote:
> One note inline
>
> On Sep 11, 2555 BE, at 23:17, Isaku Yamahata wrote:
>
> > Since Transaction._substitute doesn't substitute elements of list/tuple,
> > setting list references results in transaction error. Teach it such case.
> >
On Wed, Sep 12, 2012 at 12:09:41PM -0700, Ben Pfaff wrote:
> On Wed, Sep 12, 2012 at 05:44:32PM +0900, Simon Horman wrote:
> > By passing a flow to the action parser the pre-requisites
> > of set-feild actions will be checked. If the flow is NULL,
> > for instance in test code such as ofctl_parse_o
On Thu, Sep 13, 2012 at 11:22:53AM +0900, Isaku Yamahata wrote:
> On Wed, Sep 12, 2012 at 11:45:12AM -0700, Ben Pfaff wrote:
> > On Wed, Sep 12, 2012 at 03:17:22PM +0900, Isaku Yamahata wrote:
> > > add register_table method to SchemaHelper as counter part of
> > > ovsdb_idl_add_table().
> > >
> >
On Thu, Sep 13, 2012 at 11:27:30AM +0900, Isaku Yamahata wrote:
> On Wed, Sep 12, 2012 at 11:52:09AM -0700, Ben Pfaff wrote:
> > I agree about raising an error if both are provided.
> >
> > There doesn't seem to be an issue with reading the file earlier. It
> > won't bother any in-tree users anyh
On Thu, Sep 13, 2012 at 11:30:23AM +0900, Isaku Yamahata wrote:
> From: Isaku Yamahata
> Date: Mon, 10 Sep 2012 18:21:20 +0900
> Subject: [PATCH] python/ovs/db/idl.py: Transaction._substitute doesn't handle
> list/tuple
>
> Since Transaction._substitute doesn't substitute elements of list/tuple,
If by chance you haven't pushed yet, s/preresentation//
On Wed, Sep 12, 2012 at 8:48 PM, Ben Pfaff wrote:
> On Thu, Sep 13, 2012 at 11:27:30AM +0900, Isaku Yamahata wrote:
> > On Wed, Sep 12, 2012 at 11:52:09AM -0700, Ben Pfaff wrote:
> > > I agree about raising an error if both are provided.
>
On Thu, Sep 13, 2012 at 10:08:48AM +0900, Simon Horman wrote:
> On Wed, Sep 12, 2012 at 11:37:53AM -0700, Ben Pfaff wrote:
> > This is what I intended when I suggested using mf_subvalue, but I didn't
> > notice the difference until after applying the patch.
> >
> > Signed-off-by: Ben Pfaff
>
> A
On Thu, Sep 13, 2012 at 10:33:35AM +0900, Simon Horman wrote:
> On Wed, Sep 12, 2012 at 11:13:49AM -0700, Ben Pfaff wrote:
> > Hmm. This looks like it blindly puts an OFPAT12_SET_FIELD action into
> > other versions of OpenFlow. I think we have a few options:
>
> True, but I'm not sure that I un
On Thu, Sep 13, 2012 at 10:44:00AM +0900, Simon Horman wrote:
> On Wed, Sep 12, 2012 at 12:06:50PM -0700, Ben Pfaff wrote:
> > On Wed, Sep 12, 2012 at 05:44:31PM +0900, Simon Horman wrote:
> > > From: Isaku Yamahata
> > >
> > > Signed-off-by: Isaku Yamahata
> > > Signed-off-by: Simon Horman
> >
+else:
+if type(json) == list:
+json = [self._substitute_uuids(elem) for elem in json]
+else:
+json = tuple(self._substitute_uuids(elem) for elem in
json)
tuple/list are treated identically, based upon the
retur
On Wed, Sep 12, 2012 at 09:09:49PM -0700, Ben Pfaff wrote:
> On Thu, Sep 13, 2012 at 10:44:00AM +0900, Simon Horman wrote:
> > On Wed, Sep 12, 2012 at 12:06:50PM -0700, Ben Pfaff wrote:
> > > On Wed, Sep 12, 2012 at 05:44:31PM +0900, Simon Horman wrote:
> > > > From: Isaku Yamahata
> > > >
> > >
On Wed, Sep 12, 2012 at 08:56:41PM -0700, Ben Pfaff wrote:
> On Thu, Sep 13, 2012 at 11:30:23AM +0900, Isaku Yamahata wrote:
> > From: Isaku Yamahata
> > Date: Mon, 10 Sep 2012 18:21:20 +0900
> > Subject: [PATCH] python/ovs/db/idl.py: Transaction._substitute doesn't
> > handle list/tuple
> >
> >
On Thu, Sep 13, 2012 at 01:37:08PM +0900, Simon Horman wrote:
> On Wed, Sep 12, 2012 at 09:09:49PM -0700, Ben Pfaff wrote:
> > On Thu, Sep 13, 2012 at 10:44:00AM +0900, Simon Horman wrote:
> > > On Wed, Sep 12, 2012 at 12:06:50PM -0700, Ben Pfaff wrote:
> > > > On Wed, Sep 12, 2012 at 05:44:31PM +0
On Thu, Sep 13, 2012 at 01:41:04PM +0900, Isaku Yamahata wrote:
> On Wed, Sep 12, 2012 at 08:56:41PM -0700, Ben Pfaff wrote:
> > On Thu, Sep 13, 2012 at 11:30:23AM +0900, Isaku Yamahata wrote:
> > > From: Isaku Yamahata
> > > Date: Mon, 10 Sep 2012 18:21:20 +0900
> > > Subject: [PATCH] python/ovs/
On Tue, Sep 11, 2012 at 09:55:17PM -0700, Ben Pfaff wrote:
> On Mon, Sep 10, 2012 at 10:47:52AM -0700, Ben Pfaff wrote:
> > On Mon, Sep 10, 2012 at 09:59:28AM +0900, Simon Horman wrote:
> > > On Sat, Sep 08, 2012 at 12:15:53PM +0900, Simon Horman wrote:
> > > > On Fri, Sep 07, 2012 at 10:33:00AM -0
42 matches
Mail list logo
