Re: [ovs-dev] [PATCH 0/6] Bond port megaflow using recirculation

2014-02-13 Thread Andy Zhou
Right for bonding rules. I think MPLS handling may benefit from it -- so that we don't have maintain special hmap for recirc ids. Then we may need to resubmit from table 254. At any rate, in case it is needed, I don't see a show stopper. On Thu, Feb 13, 2014 at 8:47 PM, Ben Pfaff wrote: > On

Re: [ovs-dev] [PATCH 0/6] Bond port megaflow using recirculation

2014-02-13 Thread Ben Pfaff
On Fri, Feb 14, 2014 at 01:07:30PM +0900, Simon Horman wrote: > On Thu, Feb 13, 2014 at 08:32:11AM -0800, Ben Pfaff wrote: > > On Wed, Feb 12, 2014 at 05:12:05PM +0900, Simon Horman wrote: > > > One thing that concerns me with using special tables and the like is that > > > it may not work well in

Re: [ovs-dev] [PATCH 0/6] Bond port megaflow using recirculation

2014-02-13 Thread Simon Horman
On Thu, Feb 13, 2014 at 08:32:11AM -0800, Ben Pfaff wrote: > On Wed, Feb 12, 2014 at 05:12:05PM +0900, Simon Horman wrote: > > One thing that concerns me with using special tables and the like is that > > it may not work well in conjunction with goto table and resubmit. Although > > my implementati

Re: [ovs-dev] Group Tables Support

2014-02-13 Thread Ben Pfaff
On Thu, Feb 13, 2014 at 06:48:22PM -0700, Mohammad Ali Salahuddin wrote: > I am unable to find Group Table support in Open vSwitch. So, are Group > Tables not supported in the latest version of OVS? Do you know of > other Kernel Switches supporting Group Tables, which I can use in > Mininet? OVS 2

[ovs-dev] Group Tables Support

2014-02-13 Thread Mohammad Ali Salahuddin
Hello, I am unable to find Group Table support in Open vSwitch. So, are Group Tables not supported in the latest version of OVS? Do you know of other Kernel Switches supporting Group Tables, which I can use in Mininet? I have used a User-Space switch, however, I noticed performance issues. Any he

Re: [ovs-dev] [PATCH] upcall: Allow max_idle be configured through ovs-vsctl.

2014-02-13 Thread Ethan Jackson
I think the use of the atomic variable in this patch is a bit odd. ofproto_max_idle, is a regular old variable, even though it can be written and read from different threads: the flow dumper and the main thread. But when shared among the revalidators, then we create a special atomic in the struct

Re: [ovs-dev] [PATCH 4/4] lockfile: Implementation for Windows platform.

2014-02-13 Thread Gurucharan Shetty
On Thu, Feb 13, 2014 at 11:51 AM, Ben Pfaff wrote: > On Thu, Feb 13, 2014 at 11:22:23AM -0800, Gurucharan Shetty wrote: >> Start with a simple implementation that does not support >> symbolic links. >> >> Signed-off-by: Gurucharan Shetty > > This seems OK to me. > > ovs_lasterror_to_string() woul

[ovs-dev] modify the openflow code in openvswitch

2014-02-13 Thread ghaith alwerfally
hi guy i want test the performane of openflow controllers in both modes proactive and reactive, so how can i set up the openflow switch to works in proactive mode or reactive mode as i wish ? openflow by default works in reactive mode best regards -- غيث ابراهيم احمد ابومخلب Mr.Gheit Abomakh

Re: [ovs-dev] [PATCH 1/2] ovs-thread: Add support for pthread adaptive mutex

2014-02-13 Thread Jarno Rajahalme
On Feb 13, 2014, at 8:30 AM, Ben Pfaff wrote: > On Wed, Feb 12, 2014 at 02:01:41PM -0800, Jarno Rajahalme wrote: >> Signed-off-by: Jarno Rajahalme > > Acked-by: Ben Pfaff Pushed to master, Thanks, Jarno ___ dev mailing list dev@openvswitch.org

Re: [ovs-dev] [PATCH 2/2] ofproto-dpif: Use adaptive mutex for stats.

2014-02-13 Thread Jarno Rajahalme
On Feb 13, 2014, at 8:31 AM, Ben Pfaff wrote: > On Wed, Feb 12, 2014 at 02:01:42PM -0800, Jarno Rajahalme wrote: >> Stats critical sections are held for very short time, so it makes >> sense to spin on them for a while rather than sleeping immediately. >> >> Signed-off-by: Jarno Rajahalme > >

Re: [ovs-dev] [PATCH v2 03/13] datapath: Fix race.

2014-02-13 Thread Jarno Rajahalme
Thanks for the review. Pushed to master, 2.1, 2.0, 1.11, 1.10, and 1.9, Jarno On Feb 12, 2014, at 3:35 PM, Pravin Shelar wrote: > On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme > wrote: >> ovs_vport_cmd_dump() did rcu_read_lock() only after getting the >> datapath, which could have been

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread David Miller
From: Jiri Pirko Date: Thu, 13 Feb 2014 21:07:53 +0100 > Thu, Feb 13, 2014 at 08:26:54PM CET, da...@davemloft.net wrote: >>From: Jiri Pirko >>Date: Thu, 13 Feb 2014 19:12:57 +0100 >> > [ 3014.524118] [] ? __dev_queue_xmit+0x5/0x6a0 >> ... >>> I fail to see where bh is disabled. ovs_dp_proc

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread Jiri Pirko
Thu, Feb 13, 2014 at 08:26:54PM CET, da...@davemloft.net wrote: >From: Jiri Pirko >Date: Thu, 13 Feb 2014 19:12:57 +0100 > [ 3014.524118] [] ? __dev_queue_xmit+0x5/0x6a0 > ... >> I fail to see where bh is disabled. ovs_dp_process_received_packet() is >> called with rcu_read_lock... > >It ha

Re: [ovs-dev] [PATCH 4/4] lockfile: Implementation for Windows platform.

2014-02-13 Thread Ben Pfaff
On Thu, Feb 13, 2014 at 11:22:23AM -0800, Gurucharan Shetty wrote: > Start with a simple implementation that does not support > symbolic links. > > Signed-off-by: Gurucharan Shetty This seems OK to me. ovs_lasterror_to_string() would be more convenient if it just returned sock_strerror(GetLastE

Re: [ovs-dev] [PATCH 1/4] util: Make xreadlink a static function.

2014-02-13 Thread Ben Pfaff
On Thu, Feb 13, 2014 at 11:22:20AM -0800, Gurucharan Shetty wrote: > Signed-off-by: Gurucharan Shetty I guess that's because it's only used in util.c. Acked-by: Ben Pfaff ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo

Re: [ovs-dev] [PATCH 3/4] util: follow_symlinks() for windows.

2014-02-13 Thread Ben Pfaff
On Thu, Feb 13, 2014 at 11:22:22AM -0800, Gurucharan Shetty wrote: > Start with not supporting symbolic links for windows. > This is useful for an upcoming commit. > > Signed-off-by: Gurucharan Shetty Acked-by: Ben Pfaff ___ dev mailing list dev@openv

Re: [ovs-dev] [PATCH 2/4] string: strerror_r() for windows.

2014-02-13 Thread Ben Pfaff
On Thu, Feb 13, 2014 at 11:22:21AM -0800, Gurucharan Shetty wrote: > Signed-off-by: Gurucharan Shetty Acked-by: Ben Pfaff ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev

[ovs-dev] [PATCH 2/4] string: strerror_r() for windows.

2014-02-13 Thread Gurucharan Shetty
Signed-off-by: Gurucharan Shetty --- lib/string.h.in |1 + 1 file changed, 1 insertion(+) diff --git a/lib/string.h.in b/lib/string.h.in index 9e8fba9..bbdaeb4 100644 --- a/lib/string.h.in +++ b/lib/string.h.in @@ -35,6 +35,7 @@ #define strtok_r strtok_s #define strcasecmp _stricmp #defin

[ovs-dev] [PATCH 3/4] util: follow_symlinks() for windows.

2014-02-13 Thread Gurucharan Shetty
Start with not supporting symbolic links for windows. This is useful for an upcoming commit. Signed-off-by: Gurucharan Shetty --- lib/util.c |5 + 1 file changed, 5 insertions(+) diff --git a/lib/util.c b/lib/util.c index 6490b47..45efdb7 100644 --- a/lib/util.c +++ b/lib/util.c @@ -779

[ovs-dev] [PATCH 1/4] util: Make xreadlink a static function.

2014-02-13 Thread Gurucharan Shetty
Signed-off-by: Gurucharan Shetty --- lib/util.c |4 +++- lib/util.h |1 - 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/lib/util.c b/lib/util.c index fc990ef..6490b47 100644 --- a/lib/util.c +++ b/lib/util.c @@ -55,6 +55,8 @@ DEFINE_STATIC_PER_THREAD_DATA(struct { char s[

[ovs-dev] [PATCH 4/4] lockfile: Implementation for Windows platform.

2014-02-13 Thread Gurucharan Shetty
Start with a simple implementation that does not support symbolic links. Signed-off-by: Gurucharan Shetty --- lib/lockfile.c | 95 +--- 1 file changed, 90 insertions(+), 5 deletions(-) diff --git a/lib/lockfile.c b/lib/lockfile.c index d8f39

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread David Miller
From: Jiri Pirko Date: Thu, 13 Feb 2014 19:12:57 +0100 >>> [ 3014.524118] [] ? __dev_queue_xmit+0x5/0x6a0 ... > I fail to see where bh is disabled. ovs_dp_process_received_packet() is > called with rcu_read_lock... It happens in __dev_queue_xmit(), which does rcu_read_lock_bh(). _

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread Pravin Shelar
On Thu, Feb 13, 2014 at 10:12 AM, Jiri Pirko wrote: > Thu, Feb 13, 2014 at 06:42:03PM CET, pshe...@nicira.com wrote: >>On Thu, Feb 13, 2014 at 9:13 AM, Jiri Pirko wrote: >>> Hi. >>> >>> On current net-next I'm getting following message once I add a dp: >>> >>> [ 3014.523665] =

Re: [ovs-dev] [PATCH 7/7] meta-flow: Add mf_mask_subfield_wc()

2014-02-13 Thread Jarno Rajahalme
I tested again with some different parameters, and this actually made the performance little worse, so I’m holding off on this patch for now. Jarno On Feb 11, 2014, at 5:43 PM, Ben Pfaff wrote: > On Tue, Feb 11, 2014 at 04:30:49PM -0800, Jarno Rajahalme wrote: >> The code setting wc masks sh

Re: [ovs-dev] [PATCH 2/7] ofproto: Lock for vlan splinters only if have them.

2014-02-13 Thread Jarno Rajahalme
Pushed to master. Thanks On Feb 12, 2014, at 9:45 AM, Ben Pfaff wrote: > On Wed, Feb 12, 2014 at 08:59:18AM -0800, Jarno Rajahalme wrote: >> >> On Feb 11, 2014, at 4:40 PM, Ben Pfaff wrote: >> >>> On Tue, Feb 11, 2014 at 04:30:44PM -0800, Jarno Rajahalme wrote: Reading the hmap count

Re: [ovs-dev] [PATCH v2 11/13] datapath: Remove memory allocations from ovs_flow_cmd_execute.

2014-02-13 Thread Jarno Rajahalme
I had an offline discussion on this with Pravin. I wasn’t fully aware of the kernel stack restrictions, and as this specific patch seemed to have little impact, we’ll drop this for now. Jarno On Feb 12, 2014, at 3:48 PM, Pravin Shelar wrote: > On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalm

[ovs-dev] [PATCH v7] datapath: Per NUMA node flow stats.

2014-02-13 Thread Jarno Rajahalme
Keep kernel flow stats for each NUMA node rather than each (logical) CPU. This avoids using the per-CPU allocator and removes most of the kernel-side OVS locking overhead otherwise on the top of perf reports and allows OVS to scale better with higher number of threads. With 9 handlers and 4 reval

Re: [ovs-dev] [PATCH v2 02/13] datapath: Per NUMA node flow stats.

2014-02-13 Thread Jarno Rajahalme
On Feb 12, 2014, at 4:12 PM, Pravin Shelar wrote: > On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme > wrote: ... >> +#endif > > How much overhead do we have for non-NUMA systems without this compile > time check. > Since we always allocate node zero stat I do not think update stats > has tha

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread Jiri Pirko
Thu, Feb 13, 2014 at 06:42:03PM CET, pshe...@nicira.com wrote: >On Thu, Feb 13, 2014 at 9:13 AM, Jiri Pirko wrote: >> Hi. >> >> On current net-next I'm getting following message once I add a dp: >> >> [ 3014.523665] = >> [ 3014.524118] [ INFO: inconsistent lock stat

[ovs-dev] Processed: Re: Bug#738863: openvswitch-switch: TCP does not work properly with virtio-net when 3.12 kernel and tx/rx offloading is on

2014-02-13 Thread Debian Bug Tracking System
Processing control commands: > reassign -1 src:openvswitch 1.9.3+git20131029-1.1 Bug #738863 [src:openvswitch-switch] openvswitch-switch: TCP does not work properly with virtio-net when 3.12 kernel and tx/rx offloading is on Warning: Unknown package 'src:openvswitch-switch' Bug reassigned from pa

Re: [ovs-dev] Bug#738863: openvswitch-switch: TCP does not work properly with virtio-net when 3.12 kernel and tx/rx offloading is on

2014-02-13 Thread Andrei POPESCU
Control: reassign -1 src:openvswitch 1.9.3+git20131029-1.1 On Jo, 13 feb 14, 16:05:53, Gary Hawkins wrote: > Source: openvswitch-switch > Version: 1.9.3+git20131029-1.1 > Severity: important > > Dear Maintainer, > >* What led up to the situation? > > I ran apt-get upgrade/aptitude safe-upgr

Re: [ovs-dev] ovs inconsistent lock state

2014-02-13 Thread Pravin Shelar
On Thu, Feb 13, 2014 at 9:13 AM, Jiri Pirko wrote: > Hi. > > On current net-next I'm getting following message once I add a dp: > > [ 3014.523665] = > [ 3014.524118] [ INFO: inconsistent lock state ] > [ 3014.524118] 3.14.0-rc2+ #1 Not tainted > [ 3014.524118] -

[ovs-dev] ovs inconsistent lock state

2014-02-13 Thread Jiri Pirko
Hi. On current net-next I'm getting following message once I add a dp: [ 3014.523665] = [ 3014.524118] [ INFO: inconsistent lock state ] [ 3014.524118] 3.14.0-rc2+ #1 Not tainted [ 3014.524118] - [ 3014.524118] inconsistent {SOFTIRQ-

Re: [ovs-dev] [PATCH 0/6] Bond port megaflow using recirculation

2014-02-13 Thread Ben Pfaff
On Wed, Feb 12, 2014 at 05:12:05PM +0900, Simon Horman wrote: > One thing that concerns me with using special tables and the like is that > it may not work well in conjunction with goto table and resubmit. Although > my implementation does not yet work with those actions either. I don't yet unders

Re: [ovs-dev] [PATCH 2/2] ofproto-dpif: Use adaptive mutex for stats.

2014-02-13 Thread Ben Pfaff
On Wed, Feb 12, 2014 at 02:01:42PM -0800, Jarno Rajahalme wrote: > Stats critical sections are held for very short time, so it makes > sense to spin on them for a while rather than sleeping immediately. > > Signed-off-by: Jarno Rajahalme Acked-by: Ben Pfaff _

Re: [ovs-dev] [PATCH 1/2] ovs-thread: Add support for pthread adaptive mutex

2014-02-13 Thread Ben Pfaff
On Wed, Feb 12, 2014 at 02:01:41PM -0800, Jarno Rajahalme wrote: > Signed-off-by: Jarno Rajahalme Acked-by: Ben Pfaff ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev

Re: [ovs-dev] [PATCH v2 11/13] datapath: Remove memory allocations from ovs_flow_cmd_execute.

2014-02-13 Thread Jarno Rajahalme
On Feb 12, 2014, at 3:48 PM, Pravin Shelar wrote: > On Tue, Feb 11, 2014 at 4:07 PM, Jarno Rajahalme > wrote: >> Signed-off-by: Jarno Rajahalme >> --- >> datapath/datapath.c | 60 >> +++ >> 1 file changed, 32 insertions(+), 28 deletions(-) >>