Thx, applied to master
On Tue, Sep 2, 2014 at 7:27 PM, Pravin Shelar wrote:
> On Tue, Sep 2, 2014 at 6:44 PM, Alex Wang wrote:
> > This commit fixes a bug which prevents the display of interface
> > status for dpdk0.
> >
> > Found by inspection.
> >
> > Signed-off-by: Alex Wang
> Looks good.
On Tue, Sep 2, 2014 at 8:51 PM, Ben Pfaff wrote:
> On Tue, Sep 02, 2014 at 06:43:31PM -0700, Jesse Gross wrote:
>> On Tue, Sep 2, 2014 at 6:23 PM, Ben Pfaff wrote:
>> > On Tue, Sep 02, 2014 at 06:00:08PM -0700, Jesse Gross wrote:
>> >> The system defined ICMPv6 header doesn't have sparse annotati
>
> > @@ -1864,7 +1874,10 @@ pmd_load_queues(struct pmd_thread *f,
> > index = 0;
> >
> > CMAP_FOR_EACH (port, node, &f->dp->ports) {
> > -if (netdev_is_pmd(port->netdev)) {
> > +/* Calls port_try_ref() to prevent the main thread
> > + * from deleting the port. */
Please don't drop the mailing list.
The ovs-ofctl manpage is in the source tree and installed as part of
Open vSwitch.
On Wed, Sep 03, 2014 at 08:56:30AM +0800, loy wolfe wrote:
> Hi Ben,
>
> By your means, where is the official manpage?
>
>
> On Wed, Sep 3, 2014 at 12:11 AM, Ben Pfaff wrote:
On Tue, Sep 02, 2014 at 06:43:31PM -0700, Jesse Gross wrote:
> On Tue, Sep 2, 2014 at 6:23 PM, Ben Pfaff wrote:
> > On Tue, Sep 02, 2014 at 06:00:08PM -0700, Jesse Gross wrote:
> >> The system defined ICMPv6 header doesn't have sparse annotation,
> >> so this adds a definition so that endianness c
Hi Team,
I am working on addition of the "IMPORTANCE" field in the add-flow command
as per the Openflow-specs 1.4.
As I am new to openvswitch and have just started exploring its code, can
anyone tell me from where to start. I mean the files where changes are to
be done or from where to
start
On 2 September 2014 22:48, Joe Stringer wrote:
> If a datapath is created with the flag OVS_DP_F_INDEX_BY_UID, then an
> additional table_instance is added to the flow_table, which is indexed
> by unique identifiers ("UID"). This can be manipulated using the flow
> key as before, or by using the
On Tue, Sep 2, 2014 at 6:44 PM, Alex Wang wrote:
> When pmd thread interates through all ports for queue loading,
> the main thread may unreference and 'rcu-free' a port before
> pmd thread take new reference of it. This could cause pmd
> thread fail the reference and access freed memory later.
>
On Tue, Sep 2, 2014 at 6:44 PM, Alex Wang wrote:
> This commit fixes a bug which prevents the display of interface
> status for dpdk0.
>
> Found by inspection.
>
> Signed-off-by: Alex Wang
Looks good.
Acked-by: Pravin B Shelar
> ---
> lib/netdev-dpdk.c |2 +-
> 1 file changed, 1 insertio
On Tue, Sep 2, 2014 at 6:55 PM, Jesse Gross wrote:
> On Mon, Sep 1, 2014 at 1:10 AM, Simon Horman
> wrote:
>> On Thu, Aug 28, 2014 at 10:12:49AM +0900, Simon Horman wrote:
>>> On Wed, Aug 27, 2014 at 03:03:53PM -0500, Jesse Gross wrote:
>>> > On Wed, Aug 27, 2014 at 11:51 AM, Ben Pfaff wrote:
>
On Mon, Sep 1, 2014 at 1:10 AM, Simon Horman wrote:
> On Thu, Aug 28, 2014 at 10:12:49AM +0900, Simon Horman wrote:
>> On Wed, Aug 27, 2014 at 03:03:53PM -0500, Jesse Gross wrote:
>> > On Wed, Aug 27, 2014 at 11:51 AM, Ben Pfaff wrote:
>> > > On Wed, Aug 27, 2014 at 10:26:14AM +0900, Simon Horman
oops, I missed it. I'll drop this then
On Tue, Sep 2, 2014 at 6:44 PM, Jesse Gross wrote:
> On Tue, Sep 2, 2014 at 6:36 PM, Ethan Jackson wrote:
>> warning: incorrect type in argument 1 (different base types)
>> expected restricted ovs_be16 [usertype] old_csum
>> got unsigned short [uns
On Tue, Sep 2, 2014 at 6:36 PM, Ethan Jackson wrote:
> warning: incorrect type in argument 1 (different base types)
> expected restricted ovs_be16 [usertype] old_csum
> got unsigned short [unsigned] [usertype] icmp6_cksum
>
> Signed-off-by: Ethan Jackson
I actually just sent out a patch
On Tue, Sep 2, 2014 at 6:23 PM, Ben Pfaff wrote:
> On Tue, Sep 02, 2014 at 06:00:08PM -0700, Jesse Gross wrote:
>> The system defined ICMPv6 header doesn't have sparse annotation,
>> so this adds a definition so that endianness can be checked.
>>
>> Reported-by: Alex Wang
>> Signed-off-by: Jesse
This commit fixes a bug which prevents the display of interface
status for dpdk0.
Found by inspection.
Signed-off-by: Alex Wang
---
lib/netdev-dpdk.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index 049ea5e..a4ed355 100644
--- a/
When pmd thread interates through all ports for queue loading,
the main thread may unreference and 'rcu-free' a port before
pmd thread take new reference of it. This could cause pmd
thread fail the reference and access freed memory later.
This commit fixes this race by introducing port_try_ref()
warning: incorrect type in argument 1 (different base types)
expected restricted ovs_be16 [usertype] old_csum
got unsigned short [unsigned] [usertype] icmp6_cksum
Signed-off-by: Ethan Jackson
---
lib/packets.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/lib/packets.c b/lib/pack
Your message could not be delivered
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Tue, Sep 02, 2014 at 06:00:08PM -0700, Jesse Gross wrote:
> The system defined ICMPv6 header doesn't have sparse annotation,
> so this adds a definition so that endianness can be checked.
>
> Reported-by: Alex Wang
> Signed-off-by: Jesse Gross
In some cases we've avoided warnings on system t
---
datapath-windows/ovsext/Types.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/datapath-windows/ovsext/Types.h b/datapath-windows/ovsext/Types.h
index e48df7a..b2ef48c 100644
--- a/datapath-windows/ovsext/Types.h
+++ b/datapath-windows/ovsext/Types.h
@@ -31,6 +31,7 @@ typedef uint8 __u8;
In this change we have added the APIs for putting
netlink headers, attributes in a buffer.
The buffer is managed through NetlinkBuf.[c|h].
---
datapath-windows/ovsext/Netlink/Netlink.c | 403 +-
datapath-windows/ovsext/Netlink/Netlink.h | 28 +++
2 files changed, 427
This series contains changes for providing Netlink PUT Apis in
windows kernel.
Ankur Sharma (5):
datapath-windows/Netlink: Move netlink files to a new directory.
NetlinkBuf.c: Netlink buffer mgmt apis.
NetlinkProto.h: Minor fix for typos and new macro for padding.
OvsTypes.h : Added s
In this change we have introduced buffer mgmt apis which will be
used while creating netlink messages. The basic functionality provided
by apis is on similar lines to ofpbuf in userspace with an exception
that it will not do run time buffer reallocation.
Signed-off-by: Ankur Sharma
Tested-by: Ank
Added a new macro for calculating the number of bytes required
for padding. Fixed a minor typo.
---
datapath-windows/ovsext/Netlink/NetlinkProto.h | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/datapath-windows/ovsext/Netlink/NetlinkProto.h
b/datapath-windows/ovsext/Netli
In this change we have created a new directory named Netlink
inside datapath-windows/ovsext/. This directory will be used to
keep all the netlink related files.
The reason we have created new directory is that for 'put' related
APIs we will be adding netlink buffer mgmt files as well. These files
The system defined ICMPv6 header doesn't have sparse annotation,
so this adds a definition so that endianness can be checked.
Reported-by: Alex Wang
Signed-off-by: Jesse Gross
---
lib/packets.c | 5 +++--
lib/packets.h | 9 +
2 files changed, 12 insertions(+), 2 deletions(-)
diff --git
On Tue, Sep 02, 2014 at 09:03:24AM -0700, Ben Pfaff wrote:
> On Mon, Sep 01, 2014 at 04:59:22PM +0900, Simon Horman wrote:
> > On Thu, Aug 28, 2014 at 10:11:57AM +0900, Simon Horman wrote:
> > > On Wed, Aug 27, 2014 at 09:51:59AM -0700, Ben Pfaff wrote:
> > > > On Wed, Aug 27, 2014 at 10:26:14AM +0
It would be clearer to use tagged initializers (as suggested by Samuel). I
think Nithin probably just forgot to add that change. Seems like a minor
cleanup that can be made when we add the next command.
>In this patch, we add support for the GET_DP netlink command to dump
>the datpaaths. The user
>I didn't realize earlier that version in a netlink message was a
>UINT8. So, fixing that here.
>
>Also, some of the commands don't pass a valid DP value. Hence adding
>a field to identify such commands.
>
>Signed-off-by: Nithin Raju
>Signed-off-by: Ankur Sharma
Acked-by: Saurabh Shah
_
>In this patch we add a context structure for collecting all the parameters
>passed from usersapce in one place. The idea is to reduce the number of
>parameters being passed to the netlink command handler functions.
>
>It can be argued that not all functions require all the arguments, but
>this
>
I am fine with either approach - index array or separate variables. So
this works.
>Signed-off-by: Nithin Raju
Acked-by: Saurabh Shah
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 9/2/14, 3:28 PM, "Pravin Shelar" wrote:
>On Tue, Sep 2, 2014 at 2:44 PM, Daniele Di Proietto
> wrote:
>> On 9/2/14, 1:39 PM, "Pravin Shelar" wrote:
>>
>>>On Fri, Aug 29, 2014 at 4:52 PM, Daniele Di Proietto
>>> wrote:
dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are
dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are disabled
(if the datapath is being freed upcalls should be disabled) 'dp->upcall_rwlock'
is taken and freeing it causes an assertion to fail.
This commit takes makes sure that the upcalls are disabled and releases
'dp->upcall_rwl
Thx, applied this to master.
On Sat, Aug 30, 2014 at 12:01 PM, Pravin Shelar wrote:
> On Mon, Aug 11, 2014 at 9:56 PM, Alex Wang wrote:
> > This commit makes the memory pool name contain the socket id.
> > Since dpdk library do not allow creation of memory pool with
> > same name, this commit
On Tue, Sep 2, 2014 at 2:44 PM, Daniele Di Proietto
wrote:
> On 9/2/14, 1:39 PM, "Pravin Shelar" wrote:
>
>>On Fri, Aug 29, 2014 at 4:52 PM, Daniele Di Proietto
>> wrote:
>>> dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are
>>>disabled
>>> (if the datapath is being freed upcal
On Fri, Aug 29, 2014 at 10:34 PM, Andy Zhou wrote:
> Since kernel stack is limited in size, it is not wise to using
> recursive function with large stack frames.
>
> This patch provides an alternative implementation of recirc action
> without using recursion.
>
> A per CPU fixed sized, 'deferred a
On Fri, Aug 29, 2014 at 10:34 PM, Andy Zhou wrote:
> Future patches will change the recirc action implementation to not
> using recursion. The stack depth detection is no longer necessary.
>
> Signed-off-by: Andy Zhou
looks good.
Acked-by: Pravin B Shelar
On 9/2/14, 1:39 PM, "Pravin Shelar" wrote:
>On Fri, Aug 29, 2014 at 4:52 PM, Daniele Di Proietto
> wrote:
>> dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are
>>disabled
>> (if the datapath is being freed upcalls should be disabled)
>>'dp->upcall_rwlock'
>> is taken and freeing
From: Ariel Tubaltsev
This commit adds default values for some BFD configuration keys
(bfd_config_local:bfd_dst_mac and bfd_params:enable). It also adds new
BFD status keys (bfd_enabled and bfd_info).
Signed-off-by: Ariel Tubaltsev
Signed-off-by: Bruce Davie
---
vtep/vtep.xml | 16 +
On Fri, Aug 29, 2014 at 4:52 PM, Daniele Di Proietto
wrote:
> dp_netdev_free() must free 'dp->upcall_rwlock', but when upcalls are disabled
> (if the datapath is being freed upcalls should be disabled)
> 'dp->upcall_rwlock'
> is taken and freeing it causes an assertion to fail.
>
> This commit ta
On Tue, Sep 2, 2014 at 6:14 AM, Thomas Graf wrote:
> On 09/02/14 at 08:52pm, roy.qing...@gmail.com wrote:
>> @@ -304,9 +303,11 @@ int ovs_dp_upcall(struct datapath *dp, struct sk_buff
>> *skb,
>> if (err)
>> goto err;
>>
>> + consume_skb(skb);
>> return 0;
>
> What a
On Tue, Sep 2, 2014 at 5:52 AM, wrote:
> From: Li RongQing
>
> The user_skb maybe be leaked if the operation on it failed and codes
> skipped into the label "out:" without calling genlmsg_unicast.
>
> Cc: Pravin Shelar
> Signed-off-by: Li RongQing
Looks good.
Acked-by: Pravin B Shelar
> ---
I see.
I applied v2 to master. Thanks!
On Tue, Sep 02, 2014 at 07:52:44PM +, Eitan Eliahu wrote:
>
> I was not sure if I submitted the latest. Basically, v2 and v3 are just the
> spaces to tab fix.
> Thanks,
> Eitan
>
> -Original Message-
> From: Ben Pfaff [mailto:b...@nicira.com
I was not sure if I submitted the latest. Basically, v2 and v3 are just the
spaces to tab fix.
Thanks,
Eitan
-Original Message-
From: Ben Pfaff [mailto:b...@nicira.com]
Sent: Tuesday, September 02, 2014 12:47 PM
To: Eitan Eliahu
Cc: dev@openvswitch.org
Subject: Re: [ovs-dev] [PATCH v3]
On Tue, Sep 02, 2014 at 08:15:27PM -0700, Eitan Eliahu wrote:
> Unlike the compilation mode used for OVS x64 Linux Windows long word is
> 4 bytes for both 32 and 64 bit builds.
> Replaced _UNCONST macro with CONST_CAST to avoid the intermediate casting
> to an integer.
>
> Testing: 32 and 64 Windo
Unlike the compilation mode used for OVS x64 Linux Windows long word is
4 bytes for both 32 and 64 bit builds.
Replaced _UNCONST macro with CONST_CAST to avoid the intermediate casting
to an integer.
Testing: 32 and 64 Windows builds.
Signed-off-by: Eitan Eliahu elia...@vmware.com
---
lib/getopt_
It included it originally, but removed since updates for diagnostic field are
already inside:
In case of a problem, set to an error message that reports what the
local BFD session thinks is wrong. The error messages are defined
in section 4.1 of [RFC 5880].
On Tue, Sep 2, 2014 at 11:27 AM, wrote:
> It also brings the usage
> of the BFD diagnostic keys into line with the recent clarifications
> made for OVS.
>
Does this commit miss this part? did not see any change/line move related
to diagnostic,
___
dev
Unlike the compilation mode used for OVS x64 Linux Windows long word is
4 bytes for both 32 and 64 bit builds.
Replaced _UNCONST macro with CONST_CAST to avoid the intermediate casting
to an integer.
Testing: 32 and 64 Windows builds.
Signed-off-by: Eitan Eliahu elia...@vmware.com
---
lib/getopt_
On Tue, Sep 02, 2014 at 11:28:45AM -0700, Alex Wang wrote:
> >
> > > > Generally we should describe subject in less than 70 characters.
> > > > Commit msg should explain why we are introducing this change. It is
> > > > not clear from the patch the relation between number of core on socket
> > > >
>
> > > Generally we should describe subject in less than 70 characters.
> > > Commit msg should explain why we are introducing this change. It is
> > > not clear from the patch the relation between number of core on socket
> > > and rx queues.
> > >
> > >
> >
> > Thx, I'll use the 70 characters ru
From: Ariel Tubaltsev
This commit adds default values for some BFD configuration keys
(bfd_config_local:bfd_dst_mac and bfd_params:enable). It also adds new
BFD status keys (bfd_enabled and bfd_info). It also brings the usage
of the BFD diagnostic keys into line with the recent clarifications
mad
On Tue, Sep 02, 2014 at 06:21:09PM +, Ariel Tubaltsev wrote:
> Sending patch as attachment, to avoid mangling by mail client
Your mail client mangled the attachment so badly that it left it out
entirely.
___
dev mailing list
dev@openvswitch.org
http:
Sending patch as attachment, to avoid mangling by mail client
Ariel
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Sep 01, 2014 at 11:41:24AM -0700, Alex Wang wrote:
> On Sat, Aug 30, 2014 at 12:02 PM, Pravin Shelar wrote:
>
> > On Mon, Aug 11, 2014 at 9:56 PM, Alex Wang wrote:
> > > Before this commit, ovs only creates one tx/rx queue for each
> > > dpdk interface and uses only one poll thread for h
On Tue, Sep 02, 2014 at 06:11:41PM -0700, Eitan Eliahu wrote:
> Unlike the compilation mode used for OVS x64 Linux Windows long word is
> 4 bytes for both 32 and 64 bit builds.
> Replaced _UNCONST macro with CONST_CAST to avoid the intermediate casting
> to an integer.
>
> Testing: 32 and 64 Windo
Unlike the compilation mode used for OVS x64 Linux Windows long word is
4 bytes for both 32 and 64 bit builds.
Replaced _UNCONST macro with CONST_CAST to avoid the intermediate casting
to an integer.
Testing: 32 and 64 Windows builds.
Signed-off-by: Eitan Eliahu elia...@vmware.com
---
lib/getopt_
On Mon, Sep 1, 2014 at 11:56 AM, Alex Wang wrote:
>> > static struct dp_netdev_port *dp_netdev_lookup_port(const struct
>> > dp_netdev *dp,
>> > @@ -281,6 +281,15 @@ struct dp_netdev_actions
>> > *dp_netdev_flow_get_actions(
>> > const struct dp_netdev_flow *);
>> > static void dp_netdev_ac
On Mon, Sep 1, 2014 at 11:41 AM, Alex Wang wrote:
>
>
>
> On Sat, Aug 30, 2014 at 12:02 PM, Pravin Shelar wrote:
>>
>> On Mon, Aug 11, 2014 at 9:56 PM, Alex Wang wrote:
>> > Before this commit, ovs only creates one tx/rx queue for each
>> > dpdk interface and uses only one poll thread for handli
That's not a manual, that's an early and incomplete draft.
Read the ovs-ofctl manpage.
On Mon, Sep 01, 2014 at 11:39:01AM +0800, loy wolfe wrote:
> sorry, forgot the link
>
> http://benpfaff.org/~blp/ovs-fields.pdf
>
>
> On Mon, Sep 1, 2014 at 11:38 AM, loy wolfe wrote:
>
> > Hi,
> >
> > I s
On Tue, Sep 02, 2014 at 09:06:23AM -0700, Ben Pfaff wrote:
> On Tue, Sep 02, 2014 at 06:04:33PM +0200, Thomas Graf wrote:
> > On 09/02/14 at 08:35am, Ben Pfaff wrote:
> > > Reported-by: Thomas Graf
> > > Signed-off-by: Ben Pfaff
> >
> > Good catch
> >
> > Acked-by: Thomas Graf
>
> Thanks, app
On Tue, Sep 02, 2014 at 06:04:33PM +0200, Thomas Graf wrote:
> On 09/02/14 at 08:35am, Ben Pfaff wrote:
> > Reported-by: Thomas Graf
> > Signed-off-by: Ben Pfaff
>
> Good catch
>
> Acked-by: Thomas Graf
Thanks, applied to master.
___
dev mailing lis
On Sun, Aug 31, 2014 at 12:24:46AM -0700, Srini Seetharaman wrote:
> Fixing issue where "resubmit" action in a group action set was not
> considered sufficient to retain the full action set. This patch allows
> a group action set (considered terminal with OF1.4 and earlier spec)
> to have the "outp
On 09/02/14 at 08:35am, Ben Pfaff wrote:
> Reported-by: Thomas Graf
> Signed-off-by: Ben Pfaff
Good catch
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Sep 01, 2014 at 04:59:22PM +0900, Simon Horman wrote:
> On Thu, Aug 28, 2014 at 10:11:57AM +0900, Simon Horman wrote:
> > On Wed, Aug 27, 2014 at 09:51:59AM -0700, Ben Pfaff wrote:
> > > On Wed, Aug 27, 2014 at 10:26:14AM +0900, Simon Horman wrote:
>
> [snip]
>
> > > The proposal seems re
On Mon, Sep 01, 2014 at 08:50:55AM +, Niels van Adrichem wrote:
> 2) This was mainly a point of discussion that I am curious towards. I
> am concerned that checking all BFD, OAM, CFM, etc., sequentially may
> harm the throughput of the function odp_port_is_alive(). [...]
I doubt that's a probl
On Mon, Sep 01, 2014 at 06:10:26PM +0200, Thomas Graf wrote:
> Signed-off-by: Thomas Graf
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Sep 01, 2014 at 06:09:57PM +0200, Thomas Graf wrote:
> Signed-off-by: Thomas Graf
I don't think it's really much of a leak to fail to close a file in the
main function for a short-lived test program, but it doesn't hurt so I
applied this. Thanks!
_
On Mon, Sep 01, 2014 at 06:09:21PM +0200, Thomas Graf wrote:
> Signed-off-by: Thomas Graf
I think this is a correctness bug, not a leak. I sent a patch:
http://openvswitch.org/pipermail/dev/2014-September/045014.html
___
dev mailing list
dev@op
Reported-by: Thomas Graf
Signed-off-by: Ben Pfaff
---
tests/ovs-vsctl.at| 3 +++
utilities/ovs-vsctl.c | 14 +-
2 files changed, 12 insertions(+), 5 deletions(-)
diff --git a/tests/ovs-vsctl.at b/tests/ovs-vsctl.at
index 135f768..4ea47ef 100644
--- a/tests/ovs-vsctl.at
+++ b/te
On Mon, Sep 01, 2014 at 02:52:13PM +0200, Thomas Graf wrote:
> Signed-off-by: Thomas Graf
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Sep 01, 2014 at 05:16:49PM +0530, Hiteshi Madan wrote:
> Hi Ben and Team,
>
> I have checked the features list supported in open Vswitch.I found that
> following feature is not yet supported:
>
> ?-??? 802.1ad (provider bridging)
>
> We are thinking to contribute in above feature.
>
>
On Mon, Sep 01, 2014 at 08:11:54PM +, Alin Serdean wrote:
> MSVC does not like variable length array either.
>
> This patch treats the following error:
>
> lib/dpif-netdev.c(2272) :
> error C2057: expected constant expression
> lib/dpif-netdev.c(2272) :
> error C2466: cannot allocate an array
From: Li RongQing
distinguish between the dropped and consumed skb, not assume the skb
is consumed always
Cc: Pravin Shelar
Signed-off-by: Li RongQing
---
net/openvswitch/datapath.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/openvswitch/datapath.c b/net/openv
From: Li RongQing
The user_skb maybe be leaked if the operation on it failed and codes
skipped into the label "out:" without calling genlmsg_unicast.
Cc: Pravin Shelar
Signed-off-by: Li RongQing
---
net/openvswitch/datapath.c |4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff
If a datapath is created with the flag OVS_DP_F_INDEX_BY_UID, then an
additional table_instance is added to the flow_table, which is indexed
by unique identifiers ("UID"). This can be manipulated using the flow
key as before, or by using the new UID field. If both are specified,
then UID takes prec
Signed-off-by: Joe Stringer
---
v3: Rebase.
v2: No change.
---
lib/dpif-netdev.c | 211 ---
lib/flow.h |6 ++
tests/dpif-netdev.at|3 +
tests/ofproto-dpif.at | 20 +++--
tests/ofproto-macros.at |1 +
5 files chang
Signed-off-by: Joe Stringer
---
v3: Rebase.
v2: No change.
---
lib/dpif-linux.c | 91 +-
1 file changed, 42 insertions(+), 49 deletions(-)
diff --git a/lib/dpif-linux.c b/lib/dpif-linux.c
index a32a550..9e4f623 100644
--- a/lib/dpif-linux.c
+
When handler threads are installing ukeys, they hold the ukey mutex for
the duration of flow setup. If a revalidator thread dumps this flow
while the handler holds the lock, it will attempt to lock it using
ovs_mutex_trylock(), then if it cannot grab the lock, skip the flow.
Attempting to lock on
Future patches will make use of the 'struct dump_op' in a broader sense,
so this patch renames it to make things a bit clearer.
Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
---
v3: Rebase.
v2: No change.
RFC: First post.
---
ofproto/ofproto-dpif-upcall.c | 95 +-
This patch modifies the way that ukey lookup is done, to use a 128-bit
hash of the flow key as Netlink attributes. This hash is generated at
flow installation time, and passed down to the dpif so that datapaths
which support UID can index based on this ID rather than the flow.
For backward compati
RFCv3:
- Add datapath implementation
- Minor fixes
- Rebased
RFCv2:
- Revised early patches from v1 feedback
- Add Acks from Ben
- Rebased
RFCv1:
http://openvswitch.org/pipermail/dev/2014-August/044383.html
Joe Stringer (13):
revalidator: Use 'cmap' for storing ukeys.
revalidator: Protect uk
Currently, when a revalidator thread first dumps a flow, it creates a
'udpif_key' object and caches a copy of a kernel flow key. This allows
us to perform lookups in the classifier to attribute stats and validate
the correctness of the datapath flow.
This patch sets up this cache from the handler
Signed-off-by: Joe Stringer
Acked-by: Ben Pfaff
---
v3: Rebase.
v2: Call ovsrcu_quiesce() unconditionally.
RFC: Initial Post.
---
ofproto/ofproto-dpif-upcall.c | 61 ++---
1 file changed, 33 insertions(+), 28 deletions(-)
diff --git a/ofproto/ofproto-dpif-u
Currently, udpif_keys are protected during revalidator_sweep__() as only
one thread accesses the ukey at a time. This is ensured using barriers:
all revalidators will be in the GC phase, so they will only access their
own ukey collection.
A future patch will change the access patterns to allow the
This allows us to ignore most fields of a flow_dump, requiring only the
flow key for looking up the ukey. Fetching flows can also be avoided in
the corner case where a flow is missed from a dump but revalidation is
required.
A future patch will modify the datapath interface to make these cached
fi
An upcoming patch will change the access patterns for ukey maps to
increase the number of writers, and shift write-access from revalidator
threads to upcall handler threads. As such, it no longer makes sense to
tie these maps to revalidators in a 1:1 relationship.
This patch separates the ukey map
One of the limiting factors on the number of flows that can be supported
in the datapath is the overhead of assembling flow dump messages in the
datapath. This patch adds a new alternative to dumping the key, mask and
actions from the datapath, which is a 128-bit unique identifier for the
flow. For
Add the 128-bit murmurhash by Austin Appleby, for 32-bit systems from:
http://code.google.com/p/smhasher/source/browse/trunk/MurmurHash3.cpp
Signed-off-by: Joe Stringer
---
v3: Rebase.
There is also a version for 64-bit systems which I haven't tried yet,
mostly because this version provides the
89 matches
Mail list logo
