Re: [ovs-dev] OVS Micro Summit Notes

2014-10-21 Thread Thomas Graf
On 10/16/14 at 05:57pm, Jesse Gross wrote: > Proposals on a kernel offload API: Additional notes from the switching offload BoF can be found here: https://etherpad.fr/p/LPC2014_NetOffload Slides: http://goo.gl/ChGFnx (still being uploaded) ___ dev maili

Re: [ovs-dev] [PATCH v3] lib/dpif: Fix crash in format_odp_actions, actions = 0x0

2014-10-21 Thread Ben Pfaff
On Sat, Oct 18, 2014 at 01:18:01PM -0700, Madhu Challa wrote: > When flow_get fails (in this case flow does not exist) simply log > the key part of the get and erase the rest of the flow because it > is invalid. > > verified the fix by doing ovs-ofctl del-flows when traffic is running > > 2014-10

Re: [ovs-dev] [PATCH v3] lib/dpif: Fix crash in format_odp_actions, actions = 0x0

2014-10-21 Thread Madhu Challa
sorry forgot to put that. On Tue, Oct 21, 2014 at 7:43 AM, Ben Pfaff wrote: > On Sat, Oct 18, 2014 at 01:18:01PM -0700, Madhu Challa wrote: > > When flow_get fails (in this case flow does not exist) simply log > > the key part of the get and erase the rest of the flow because it > > is invalid.

Re: [ovs-dev] [PATCH v3] lib/dpif: Fix crash in format_odp_actions, actions = 0x0

2014-10-21 Thread Ben Pfaff
On Tue, Oct 21, 2014 at 08:08:02AM -0700, Madhu Challa wrote: > sorry forgot to put that. > > On Tue, Oct 21, 2014 at 7:43 AM, Ben Pfaff wrote: > > > On Sat, Oct 18, 2014 at 01:18:01PM -0700, Madhu Challa wrote: > > > When flow_get fails (in this case flow does not exist) simply log > > > the ke

Re: [ovs-dev] [PATCH v1 3/6] datapath-windows: Calling OvsAddPidInstance and OvsDelPidInstance

2014-10-21 Thread Eitan Eliahu
Please find comments inline -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Ankur Sharma Sent: Monday, October 20, 2014 5:35 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH v1 3/6] datapath-windows: Calling OvsAddPidInstance and OvsDelPidInstance Sign

Re: [ovs-dev] [PATCH v1 5/6] datapath-windows: Refactor CreateQueue function to handle vport pid.

2014-10-21 Thread Eitan Eliahu
Inline. Thanks, Eitan -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Ankur Sharma Sent: Monday, October 20, 2014 5:35 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH v1 5/6] datapath-windows: Refactor CreateQueue function to handle vport pid. Refactor

Re: [ovs-dev] [PATCH v1 6/6] datapath-windows: Fixes during integration testing.

2014-10-21 Thread Eitan Eliahu
LG Acked-by: Eitan Eliahu (I notice you fixed an issue in a previous patch) -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Ankur Sharma Sent: Monday, October 20, 2014 5:35 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH v1 6/6] datapath-windows: Fixes

Re: [ovs-dev] [PATCH v1 4/6] datapath-windows: Add code for OvsGetQueue.

2014-10-21 Thread Eitan Eliahu
Please consider removing the lock from OvsGetQueue() and holding it by the caller so getting the queue and inserting elements to it would be atopmic? Thanks, Eirtan -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Ankur Sharma Sent: Monday, October 20, 2014 5

[ovs-dev] [PATCH 1/3] datapath: Replace __force type cast with rcu_dereference_raw().

2014-10-21 Thread Pravin B Shelar
rcu_dereference_raw() api is cleaner way of accessing RCU pointer when no locking is required. Signed-off-by: Pravin B Shelar --- datapath/datapath.c |2 +- datapath/flow_table.c |8 datapath/vport.c |2 +- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git

[ovs-dev] [PATCH 2/3] datapath: Fix comment style.

2014-10-21 Thread Pravin B Shelar
Use netdev comment style. Signed-off-by: Pravin B Shelar --- datapath/actions.c|3 ++- datapath/compat.h |3 ++- datapath/datapath.c | 15 ++- datapath/flow.c |6 -- datapath/flow_netlink.c | 12 d

[ovs-dev] [PATCH 3/3] datapath: Use upstream ipv6_find_hdr().

2014-10-21 Thread Pravin B Shelar
ipv6_find_hdr() already fixed in newer upstram kernel, we can start using this API safely. This patch also backports fix (ipv6: ipv6_find_hdr restore prev functionality) to compat ipv6_find_hdr(). CC: Ansis Atteka Signed-off-by: Pravin B Shelar --- acinclude.m4 |

Re: [ovs-dev] [PATCH v1 0/4] openvswitch: DPDK Tunneling.

2014-10-21 Thread David Christensen
> Following series adds support for tunneling entirely in userspace. > Even though this is targeted for DPDK based device, this should > work on any platform that support netdev datapath. > > To make tunneling work it need special configuration, There is > README-native-tunneling file which has de

Re: [ovs-dev] [PATCH v1 0/4] openvswitch: DPDK Tunneling.

2014-10-21 Thread Pravin Shelar
On Tue, Oct 21, 2014 at 10:15 AM, David Christensen wrote: >> Following series adds support for tunneling entirely in userspace. >> Even though this is targeted for DPDK based device, this should >> work on any platform that support netdev datapath. >> >> To make tunneling work it need special con

Re: [ovs-dev] [PATCH 1/4 v2] datapath-windows: event read should not fail when no events

2014-10-21 Thread Ben Pfaff
On Sat, Oct 18, 2014 at 11:39:35AM -0700, Nithin Raju wrote: > The semantics are read operation are generally to return 0 bytes and > STATUS_SUCCESS when there are no events. > > Also, added a fix to assign the PID to the synthetic OVS_MESSAGE formed > for the command validation. > > Signed-off-b

Re: [ovs-dev] [PATCH 2/4 v2] datapath-windows: fixes in Flow.c in key parsing

2014-10-21 Thread Ben Pfaff
On Sat, Oct 18, 2014 at 11:39:36AM -0700, Nithin Raju wrote: > Signed-off-by: Nithin Raju Applied, thanks! ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev

Re: [ovs-dev] [PATCH 3/4 v2] datapath-windows: Fixes in packet created for userspace

2014-10-21 Thread Ben Pfaff
On Sat, Oct 18, 2014 at 11:39:37AM -0700, Nithin Raju wrote: > A couple of miscellaneous fixes in code that creates a packet for > userspace as well as when we copy the packet to memory specified by > userspace. > > Signed-off-by: Nithin Raju > Acked-by: Ankur Sharma Applied, thanks! __

Re: [ovs-dev] [PATCH 4/4 v2] lib/netlink-socket.c: fixes in nl_sock_recv__() on Windows

2014-10-21 Thread Ben Pfaff
On Sat, Oct 18, 2014 at 11:39:38AM -0700, Nithin Raju wrote: > In nl_sock_recv__() on Windows, we realloc a new ofpbuf to copy received > data if the caller specified buffer is small. While we do so, we need > reset some of the other stack variables to point to the new ofpbuf. > > Other fixes are

Re: [ovs-dev] [PATCH] datapath-windows: check for 'gOvsSwitchContext' in ValidateNetlinkCmd()

2014-10-21 Thread Alin Serdean
Acked-by: Alin Gabriel Serdean -Mesaj original- De la: dev [mailto:dev-boun...@openvswitch.org] În numele Nithin Raju Trimis: Wednesday, October 15, 2014 9:31 PM Către: dev@openvswitch.org Subiect: [ovs-dev] [PATCH] datapath-windows: check for 'gOvsSwitchContext' in ValidateNetlinkCmd

Re: [ovs-dev] [PATCH v1 1/4] route-table: get rid of name-table

2014-10-21 Thread Thomas Graf
On 10/16/14 at 11:38am, Pravin B Shelar wrote: > name table maintains device ifindex to name mapping. On any name > table changes it invalidate name table and routing table. So > rather than building two tables this patch moves dev name to routing > entry and build routing table on any name table

[ovs-dev] [PATCH] Build: shared libraries and versioning

2014-10-21 Thread Scott Mann
These changes cause shared libraries to be built by default. In particular, lib/libopenvwitch.so, lib/libsflow.so, ofproto/libofproto.so, and ovsdb/libovsdb.so will be built. Original behavior of building static objects excusively may be accomplished by providing the --disable-shared argument to co

Re: [ovs-dev] [PATCH v1 2/4] route-table: Use classifier to store routing table.

2014-10-21 Thread Thomas Graf
On 10/16/14 at 11:38am, Pravin B Shelar wrote: > Rather than using hmap for storing routing entries we can directly use > classifier which has support for priority and wildcard entries. > This makes route lookup lockless. This help when we use route lookup > for native tunneling. > > Signed-off-by

Re: [ovs-dev] [PATCH v1 3/4] route-table: extract gw information.

2014-10-21 Thread Thomas Graf
On 10/16/14 at 11:38am, Pravin B Shelar wrote: > Routing table will be used by ovs userspace tunneling, it need to > know gw address, following commit extract gw information from > netlink message so that ovs can populate it in ovs route table. > > Signed-off-by: Pravin B Shelar > --- > lib/rout

Re: [ovs-dev] [PATCHv7 10/11] dpif: Index flows using unique identifiers.

2014-10-21 Thread Joe Stringer
On 13 October 2014 17:09, Ben Pfaff wrote: > On Tue, Oct 07, 2014 at 12:23:37AM +1300, Joe Stringer wrote: > > This patch modifies the dpif interface to allow flows to be manipulated > > using a 128-bit identifier. This allows revalidator threads to perform > > datapath operations faster, as they

[ovs-dev] [PATCH 0/5] OVS-on-HV: userspace support for packet receive

2014-10-21 Thread Nithin Raju
In this series of patches we add support for packet receive in Windows userspace. With this patch, we can ping between 2 VMs on the same host on the same OVS bridge. Nithin Raju (5): lib/util.c: add ovs_windows_only() cpp macro datapath-windows: return success when duplicate flow is added li

[ovs-dev] [PATCH 2/5] datapath-windows: return success when duplicate flow is added

2014-10-21 Thread Nithin Raju
If we are trying to insert a flow while there's already a key with the same flow, return success instead of failure. It can be argued that we should probably return a transactional error EEXIST, but we'll handle this in a subsequent commit. I've added a comment to address this later. Signed-off-by

[ovs-dev] [PATCH 1/5] lib/util.c: add ovs_windows_only() cpp macro

2014-10-21 Thread Nithin Raju
It is convenient to have a macro which we can use to wrap Windows specific code without using the #ifdef _WIN32/#endif combo each time. Signed-off-by: Nithin Raju --- lib/util.h |8 1 files changed, 8 insertions(+), 0 deletions(-) diff --git a/lib/util.h b/lib/util.h index f171dbf.

[ovs-dev] [PATCH 3/5] lib/netlink-socket.c: packet subscribe functionality on Windows

2014-10-21 Thread Nithin Raju
In this patch, we add support in userspace for packet subscribe API similar to the join/leave MC group API that is used for port events. The kernel code has already been commited. Acked-by: Nithin Raju --- lib/netlink-socket.c | 59 ++ 1 files ch

[ovs-dev] [PATCH 4/5] lib/netlink-socket.c: use poll_immediate_wake() on Windows

2014-10-21 Thread Nithin Raju
We have not yet tested the wakup via pending IRP functionality on Windows yet. Hence we use poll_immediate_wake(). Signed-off-by: Nithin Raju --- lib/netlink-socket.c |3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/lib/netlink-socket.c b/lib/netlink-socket.c index 56f4

[ovs-dev] [PATCH 5/5] lib/dpif-netlink.c: add support for packet receive on Windows

2014-10-21 Thread Nithin Raju
In this patch, we add support in dpif-netlink.c to receive packets on Windows. Windows does not natively support epoll(). Even though there are mechanisms/interfaces that provide functionality similar to epoll(), we take a simple approach of using a pool of sockets. Here are some details of the im

Re: [ovs-dev] [PATCH 1/5] lib/util.c: add ovs_windows_only() cpp macro

2014-10-21 Thread Eitan Eliahu
Acked-by: Eitan Eliahu Thanks, Eitan -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Nithin Raju Sent: Tuesday, October 21, 2014 4:11 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH 1/5] lib/util.c: add ovs_windows_only() cpp macro It is convenient to

Re: [ovs-dev] [PATCH 2/5] datapath-windows: return success when duplicate flow is added

2014-10-21 Thread Eitan Eliahu
Acked-by: Eitan Eliahu -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Nithin Raju Sent: Tuesday, October 21, 2014 4:11 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH 2/5] datapath-windows: return success when duplicate flow is added If we are tryin

Re: [ovs-dev] [PATCH 3/5] lib/netlink-socket.c: packet subscribe functionality on Windows

2014-10-21 Thread Eitan Eliahu
Hi Nithin, Please remove the following line: sock->read_ioctl = OVS_IOCTL_READ_PACKET; // XXX: The socket type changes only if the IOCTL is successful. (otherwise user mode could call unsubscribe) Thanks, Eitan -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of

Re: [ovs-dev] [PATCH 4/5] lib/netlink-socket.c: use poll_immediate_wake() on Windows

2014-10-21 Thread Eitan Eliahu
Acked-by: Eitan Eliahu -Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Nithin Raju Sent: Tuesday, October 21, 2014 4:11 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH 4/5] lib/netlink-socket.c: use poll_immediate_wake() on Windows We have not yet t

Re: [ovs-dev] [PATCH 5/5] lib/dpif-netlink.c: add support for packet receive on Windows

2014-10-21 Thread Eitan Eliahu
-Original Message- From: dev [mailto:dev-boun...@openvswitch.org] On Behalf Of Nithin Raju Sent: Tuesday, October 21, 2014 4:11 PM To: dev@openvswitch.org Subject: [ovs-dev] [PATCH 5/5] lib/dpif-netlink.c: add support for packet receive on Windows In this patch, we add support in dpif-n

[ovs-dev] [PATCH v2 2/6] datapath-windows: pid-instance hash table APIs.

2014-10-21 Thread Ankur Sharma
In this patch we have added APIs for insert, delete and search APIs. Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/User.c | 57 ++ datapath-windows/ovsext/User.h | 10 2 files changed, 67 insertions(+) diff --git a/datapath-windows/ovse

[ovs-dev] [PATCH v2 1/6] datapath-windows: pid-instance hash table data structure.

2014-10-21 Thread Ankur Sharma
This patch introduces data structure for holding instances hashed by pid. Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/Datapath.h | 2 ++ datapath-windows/ovsext/Switch.c | 15 ++- datapath-windows/ovsext/Switch.h | 3 +++ datapath-windows/ovsext/User.c | 1 + 4

[ovs-dev] [PATCH v2 6/6] datapath-windows: Fixes during integration testing.

2014-10-21 Thread Ankur Sharma
Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/User.c | 4 +++- datapath-windows/ovsext/User.h | 2 +- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c index 6797350..3da34da 100644 --- a/datapath-windows/ovsex

[ovs-dev] [PATCH v2 5/6] datapath-windows: Refactor CreateQueue function to handle vport pid.

2014-10-21 Thread Ankur Sharma
Refactored CreateQueue function so that packets are enqueued to correct corresponding queue. Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/Actions.c | 4 +- datapath-windows/ovsext/PacketIO.c | 2 +- datapath-windows/ovsext/Tunnel.c | 2 +- datapath-windows/ovsext/User.c | 7

[ovs-dev] [PATCH v2 3/6] datapath-windows: Calling OvsAddPidInstance and OvsDelPidInstance

2014-10-21 Thread Ankur Sharma
Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/User.c | 20 1 file changed, 20 insertions(+) diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c index 36787cd..6327959 100644 --- a/datapath-windows/ovsext/User.c +++ b/datapath-windows/ovsext

[ovs-dev] [PATCH v2 4/6] datapath-windows: Add code for OvsGetQueue.

2014-10-21 Thread Ankur Sharma
Signed-off-by: Ankur Sharma --- datapath-windows/ovsext/User.c | 21 - 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/datapath-windows/ovsext/User.c b/datapath-windows/ovsext/User.c index 6327959..54f9682 100644 --- a/datapath-windows/ovsext/User.c +++ b/datapa

Re: [ovs-dev] [PATCHv7 10/11] dpif: Index flows using unique identifiers.

2014-10-21 Thread Joe Stringer
On 14 October 2014 09:49, Ben Pfaff wrote: > On Tue, Oct 07, 2014 at 12:23:37AM +1300, Joe Stringer wrote: > > This patch modifies the dpif interface to allow flows to be manipulated > > using a 128-bit identifier. This allows revalidator threads to perform > > datapath operations faster, as they

Re: [ovs-dev] [PATCHv7 11/11] dpif: Minimize memory copy for revalidation.

2014-10-21 Thread Joe Stringer
On 14 October 2014 10:43, Ben Pfaff wrote: > On Tue, Oct 07, 2014 at 12:23:38AM +1300, Joe Stringer wrote: > > 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 modifies the

Re: [ovs-dev] [PATCHv7 11/11] dpif: Minimize memory copy for revalidation.

2014-10-21 Thread Joe Stringer
On 14 October 2014 10:43, Ben Pfaff wrote: > In the ukey_acquire() comment: > * Returns 0 on success, setting *result to the matching ukey and > returning it > * in a locked state. Otherwise, returns an errno and clears *result. > EBUSY > * indicates that another thread is handling