Re: [ovs-dev] [PATCH v3] Extend OVS IPFIX exporter to export tunnel headers

I have addressed all the comments and sent out another version of the patch, may you help? Thanks Wenyu -Original Message- From: Pravin Shelar [mailto:pshe...@nicira.com] Sent: Friday, July 18, 2014 12:00 PM To: Wenyu Zhang Cc: Romain Lenglet; Ben Pfaff; Jesse Gross; dev@openvswitch.or

Re: [ovs-dev] [PATCH 1/4] lib/classifier: Unify struct classifier and cls_classifier.

Thanks for the reviews, pushed to master. Jarno On Jul 17, 2014, at 10:40 AM, Ben Pfaff wrote: > On Thu, Jul 17, 2014 at 07:09:15AM -0700, Jarno Rajahalme wrote: >> >> On Jul 15, 2014, at 10:18 PM, YAMAMOTO Takashi >> wrote: >> /* A flow classifier. */ struct classifier { -

Re: [ovs-dev] [PATCH 3/4] lib/classifier: Return all matching prefix lengths from trie lookup.

Thanks for the review. Pushed with also the preceding bigger comment updated. Jarno On Jul 17, 2014, at 7:13 AM, Jarno Rajahalme wrote: > > On Jul 15, 2014, at 2:24 PM, Ben Pfaff wrote: > >> On Fri, Jul 11, 2014 at 04:55:56AM -0700, Jarno Rajahalme wrote: >>> Previously we only returned th

Re: [ovs-dev] [PATCH 2/4] lib/classifier: Change a variable name.

Pushed with this change: - Another variable was already named “match_len”, so I had to rename it as well. The following patch removes this other variable. Jarno On Jul 17, 2014, at 10:40 AM, Ben Pfaff wrote: > On Thu, Jul 17, 2014 at 07:15:17AM -0700, Jarno Rajahalme wrote: >> >> On Jul 15

Re: [ovs-dev] [PATCH 4/4] lib/classifier: Clarify subtable skipping.

Thanks for the reviews! Pushed to master, Jarno On Jul 15, 2014, at 10:20 PM, YAMAMOTO Takashi wrote: >> On Fri, Jul 11, 2014 at 04:55:57AM -0700, Jarno Rajahalme wrote: >>> Clarify comments for trie-based subtable skipping. >>> >>> Perform the cheaper check first. >>> >>> Signed-off-by: J

Re: [ovs-dev] [PATCH v2 8/8] datapath: Allow masks for set actions.

Jesse, Thanks for the review, and sorry for the time it has taken me to get back to this! Responses below, Jarno On May 8, 2014, at 1:27 PM, Jesse Gross wrote: > On Fri, Apr 11, 2014 at 4:29 PM, Jarno Rajahalme > wrote: >> Masked set action allows more megaflow wildcarding. Masked set ac

[ovs-dev] Rackspace Users List

Hi, Would you be interested in *Rackspace *users list for your Sale/Marketing campaigns? Information Fields – Name, Title, Email, Company Name, and Company Details like, Physical Address, Web Address, Revenue Size, Employee Size and industry. We also have other Software/Hardware users like

Re: [ovs-dev] [PATCH] ovsdb: Don't add ovsdb-server.c to libovsdb.

On Thu, Jul 17, 2014 at 06:31:28PM -0700, Gurucharan Shetty wrote: > Without this change, with shared libraries, VLOG > constructor for ovsdb-server would get called twice corrupting > the 'vlog_modules' list causing an infinite loop. > > Signed-off-by: Gurucharan Shetty > Reported-by: Gur Stavi

Re: [ovs-dev] [PATCH 03/14] Ignore epoll specific functions in case of MSVC

On Thu, Jul 17, 2014 at 07:40:49PM +, Alin Serdean wrote: > The substitute for epoll would be I/O Completion Ports > http://msdn.microsoft.com/en-us/library/windows/desktop/aa365198%28v=vs.85%29.aspx > . > > I will be back later on with a patch with this matter. > > I have taken your review

[ovs-dev] [PATCH 1/1] Always save ofports on kmod reload

Not saving the ofport on kmod reloads can lead to datapath outages with some scenarios. Further, there doesn't seem to be a reason not to save ofports. Signed-off-by: Andy Hill ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/list

Re: [ovs-dev] [PATCH 1/1] Always save ofports on kmod reload

On Fri, Jul 18, 2014 at 02:10:04PM -0400, Andy Hill wrote: > Not saving the ofport on kmod reloads can lead to datapath outages with some > scenarios. Further, there doesn't seem to be a reason not to save ofports. > > Signed-off-by: Andy Hill ofports are only saved when upgrading from 1.10 beca

Re: [ovs-dev] [PATCHv3] datapath: Cache netlink flow key, mask on flow_put.

On Thu, Jul 17, 2014 at 3:53 PM, Joe Stringer wrote: > On 18 July 2014 05:19, Ben Pfaff wrote: >> - If the userspace flow key includes more fields than the >> > kernel's, for example if userspace decoded an IPv6 header but >> > the kernel stopped at the Ethernet type, then userspa

Re: [ovs-dev] issue on building openvswitch 2.1.2 on RHEL 7.0

You will likely have better luck posting this on a mailing list and also asking Red Hat people. On Sun, Jul 13, 2014 at 9:27 PM, JIANGANG Zhang wrote: > Hey Jesse, > > I am trying to build openvswitch 2.1.2 on RHEL/CentOS 7.0, but getting > similar problem as in > http://openvswitch.org/pipermail

Re: [ovs-dev] [PATCH v4] Extend OVS IPFIX exporter to export tunnel headers

sparse warning CHECK /home/pravin/ovs/w8/datapath/linux/vport.c /home/pravin/ovs/w8/datapath/linux/../flow.h:126:15: warning: memset with byte count of 0 CC [M] /home/pravin/ovs/w8/datapath/linux/vport.o On Fri, Jul 18, 2014 at 1:59 AM, Wenyu Zhang wrote: > Extend IPFIX exporter to expo

[ovs-dev] [PATCH] stp: Add more logging points for debug.

This commit adds more logging points in stp module for debugging. Also, it makes the log print out the port name. Signed-off-by: Alex Wang --- lib/stp.c | 30 ++ lib/stp.h |1 + ofproto/ofproto-dpif.c |1 + 3 files changed, 32 inser

[ovs-dev] [PATCH] netlink-socket: Do not make flow_dump block on netlink socket.

Commit 93295354 (netlink-socket: Simplify multithreaded dumping to match Linux reality.) makes the call to recvmsg() block if no messages are available. This can cause revalidator threads hanging for long time or even deadlock when main thread tries to stop the revalidator threads. This commit fi

Re: [ovs-dev] [PATCH] netlink-socket: Do not make flow_dump block on netlink socket.

On Fri, Jul 18, 2014 at 02:45:47PM -0700, Alex Wang wrote: > Commit 93295354 (netlink-socket: Simplify multithreaded dumping > to match Linux reality.) makes the call to recvmsg() block if no > messages are available. This can cause revalidator threads hanging > for long time or even deadlock when

Re: [ovs-dev] [PATCH] netlink-socket: Do not make flow_dump block on netlink socket.

*Sure, when I tried to delete my br-int, ovs hangs* *Basically, main thread joins the revalidator thread, revalidator threads are either blocking at recvmsg() or the mutex.* *Following is the trace:* Id Target Id Frame 47 Thread 0x7f18bed6a700 (LWP 338) "revalidator57" 0x7f18

Re: [ovs-dev] [PATCH] netlink-socket: Do not make flow_dump block on netlink socket.

On Fri, Jul 18, 2014 at 03:10:27PM -0700, Alex Wang wrote: > *Sure, when I tried to delete my br-int, ovs hangs* > > *Basically, main thread joins the revalidator thread, revalidator threads > are either blocking at recvmsg() or the mutex.* Thanks, after some reading and experimentation I underst

Re: [ovs-dev] [PATCH 8/8] lib/classifier: Use cmap.

I think I found the reason for the slowness of cmap on XenServer. I was debugging on a XenServer tonight and found out that the configure test for GCC atomics had apparently failed, as I noticed actual lock and unlock operations on gdb disassembly, and then by the fact the changes I made to the

Re: [ovs-dev] [PATCH 3/4] dpif-windows: Implement datapath interface for windows.

From: Ben Pfaff mailto:b...@nicira.com>> Date: Wednesday, June 25, 2014 at 4:54 PM To: Saurabh Shah mailto:ssaur...@vmware.com>> Cc: "dev@openvswitch.org" mailto:dev@openvswitch.org>>, Guolin Yang mailto:gy...@vmware.com>> Subject: Re: [ovs-dev] [PATCH 3/4] dpif-wind

Re: [ovs-dev] [PATCH] netlink-socket: Do not make flow_dump block on netlink socket.

Thx Ben, the suggestion makes sense. Applied to master with the comments. On Fri, Jul 18, 2014 at 4:16 PM, Ben Pfaff wrote: > On Fri, Jul 18, 2014 at 03:10:27PM -0700, Alex Wang wrote: > > *Sure, when I tried to delete my br-int, ovs hangs* > > > > *Basically, main thread joins the revalidat

Re: [ovs-dev] [PATCH 8/8] lib/classifier: Use cmap.

>From config.log: configure:13921: checking whether gcc -std=gnu99 supports GCC 4.0+ atomic built-ins configure:13988: gcc -std=gnu99 -o conftest -g -O2 conftest.c -lrt -lm >&5 /tmp/ccZ2M4lr.o: In function `main': /usr/src/redhat/BUILD/openvswitch-2.3.90.37691/conftest.c:57: undefined referen

Re: [ovs-dev] [PATCH 1/4] datapath-windows: Base code for developing the Hyper-V switch entension.

From: Ben Pfaff mailto:b...@nicira.com>> Date: Wednesday, June 25, 2014 at 2:57 PM To: Saurabh Shah mailto:ssaur...@vmware.com>> Cc: "dev@openvswitch.org" mailto:dev@openvswitch.org>>, Guolin Yang mailto:gy...@vmware.com>> Subject: Re: [ovs-dev] [PATCH 1/4] datapath-w

Re: [ovs-dev] [PATCH 1/4] datapath-windows: Base code for developing the Hyper-V switch entension.

HTML formatting probably doesn¹t work. Bubbling up my comment. "The licensing is a bit tricky. We made a call to rewrite the entire base code. The new code will have no dependency on any of the MS-LPL code and will be licensed under APACHE itself. I will submit a V2 patch with the new base code &

Re: [ovs-dev] [PATCH 4/4] datapath-windows: Implement the virtual switch forwarding extension.

On Mon, Jun 23, 2014 at 05:34:59PM -0700, Saurabh Shah wrote: With this implementation we support bridge backed forwarding (with VLAN tagging) & tunneling over VXLAN. Co-Authored-By: Eitan Eliahu mailto:elia...@vmware.com>> Signed-off-by: Eitan Eliahu mailto:elia...@vmware.com>> Co-Authored-By: Gu

Re: [ovs-dev] [PATCH 3/4] dpif-windows: Implement datapath interface for windows.

Converted to plain text. Hopefully this will make my inline comments more apparent. > > >From: Ben Pfaff mailto:b...@nicira.com>> >Date: Wednesday, June 25, 2014 at 4:54 PM >To: Saurabh Shah mailto:ssaur...@vmware.com>> >Cc: "dev@openvswitch.org" >mailto:dev@openvswitc

[ovs-dev] [PATCH 1/5] lib/ovs-rcu: evaluate argument of ovsrcu_get only once.

As ovsrcu_get() looks like a function call, it is reasonable for the callers to expect that the arguments are evaluated only once. CONST_CAST expands its 'POINTER' argument multiple times, and the exact effect of this turned out to be compiler dependent. Fix this by expanding the macro argument be

[ovs-dev] [PATCH 4/5] vswitchd/bridge: Fix setting default prefix fields.

The check for the need of default values was in the wrong place, causing no prefix tracking to be used when database had no configuration for a flow table. Missing configuration means that defaults should be used. To limit clutter on the log, we now log the prefix tracking configuration when it i

[ovs-dev] [PATCH 2/5] lib/cmap, lib/classifier: Avoid aliasing pointers.

Older GCC (e.g. 4.1.2) did not like the pointer casts used for initializing the iteration cursors. This patch changes the code to avoid the void casts for GCC, and makes the classifier interface more similar to that of the cmap. These changes make the code work with GCC 4.1.2 strict aliasing rule

[ovs-dev] [PATCH 3/5] ovs-atomic: Avoid evaluating arguments multiple times.

ovs-atomic-gcc4+ did expand arguments again, if locked set/get was called. Also fix atomic_is_lock_free(). Signed-off-by: Jarno Rajahalme --- lib/ovs-atomic-gcc4+.h |8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/lib/ovs-atomic-gcc4+.h b/lib/ovs-atomic-gcc4+.h ind

[ovs-dev] [PATCH 5/5] lib/dpif-netdev.c: Silence a compiler warning.

'error' would be used without initialization if 'cnt' was 0. Signed-off-by: Jarno Rajahalme --- lib/dpif-netdev.c |2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c index 9270873..90445d1 100644 --- a/lib/dpif-netdev.c +++ b/lib/dpif-netd

Re: [ovs-dev] [PATCH 1/5] lib/ovs-rcu: evaluate argument of ovsrcu_get only once.

On Fri, Jul 18, 2014 at 09:05:48PM -0700, Jarno Rajahalme wrote: > As ovsrcu_get() looks like a function call, it is reasonable for the > callers to expect that the arguments are evaluated only once. > CONST_CAST expands its 'POINTER' argument multiple times, and the > exact effect of this turned o

Re: [ovs-dev] [PATCH 2/5] lib/cmap, lib/classifier: Avoid aliasing pointers.

On Fri, Jul 18, 2014 at 09:05:49PM -0700, Jarno Rajahalme wrote: > Older GCC (e.g. 4.1.2) did not like the pointer casts used for > initializing the iteration cursors. This patch changes the code to > avoid the void casts for GCC, and makes the classifier interface more > similar to that of the cm

Re: [ovs-dev] [PATCH 3/5] ovs-atomic: Avoid evaluating arguments multiple times.

On Fri, Jul 18, 2014 at 09:05:50PM -0700, Jarno Rajahalme wrote: > ovs-atomic-gcc4+ did expand arguments again, if locked set/get was > called. > > Also fix atomic_is_lock_free(). > > Signed-off-by: Jarno Rajahalme Acked-by: Ben Pfaff ___ dev mailing

Re: [ovs-dev] [PATCH 4/5] vswitchd/bridge: Fix setting default prefix fields.

On Fri, Jul 18, 2014 at 09:05:51PM -0700, Jarno Rajahalme wrote: > The check for the need of default values was in the wrong place, > causing no prefix tracking to be used when database had no > configuration for a flow table. Missing configuration means that > defaults should be used. > > To lim

Re: [ovs-dev] [PATCH 5/5] lib/dpif-netdev.c: Silence a compiler warning.

On Fri, Jul 18, 2014 at 09:05:52PM -0700, Jarno Rajahalme wrote: > 'error' would be used without initialization if 'cnt' was 0. > > Signed-off-by: Jarno Rajahalme Acked-by: Ben Pfaff ___ dev mailing list dev@openvswitch.org http://openvswitch.org/mail