On Fri, Oct 27, 2017 at 11:19 PM, Ben Pfaff wrote:
> On Fri, Oct 27, 2017 at 10:03:07AM -0700, Guru Shetty wrote:
> > On 27 October 2017 at 08:35, Ben Pfaff wrote:
> >
> > > This patch is one where I need help. I asked Guru because it's an area
> > > he knows, but
add dpdk
2017-10-31 10:43 GMT+08:00 Sam :
> Hi all,
>
> I found in ovs-dpdk it is failing to release memory when QEMU exits. Is
> this a clear bug in ovs-dpdk? If it is, when did this bug fixed? Detail is
> bellow:
>
>
> For qemu-2.6.0 and ovs-dpdk, in huge page (1G)
Hi all,
I found in ovs-dpdk it is failing to release memory when QEMU exits. Is
this a clear bug in ovs-dpdk? If it is, when did this bug fixed? Detail is
bellow:
For qemu-2.6.0 and ovs-dpdk, in huge page (1G) environment, after kill the
qemu process, memory which is alloc for the vm could not
Hi Ben,
Thanks for the comments!
The multipath feature can separete traffics to multi-gateways. If you
get two gateway routers and both of them pin on chassis. (gateway1 pin on
chassis-A, gateway2 pin on chassis-B) The multipath can seperate traffic to
those chassis automaticly.
From: Kaige Fu
IFF_NO_QUEUE is checked twice. This cause a situation
that HAVE_IFF_NO_QUEUE is defined no matter whether
the flag IFF_NO_QUEUE is defined in kernel or not.
Signed-off-by: Kaige Fu
---
datapath/linux/compat/include/linux/netdevice.h | 3
On Thu, Sep 21, 2017 at 07:45:23PM +0800, wang.qia...@zte.com.cn wrote:
> From 80260a2950f10544e307d6f20cb1cfe8c9bb885f Mon Sep 17 00:00:00 2001
> From: wang qianyu
> Date: Thu, 21 Sep 2017 18:05:16 +0800
> Subject: [PATCH] ovn: reduce logical flow applied to
On Tue, Sep 19, 2017 at 04:38:04PM +0800, wang.qia...@zte.com.cn wrote:
> To support taas function, this patch add two type of logica_switch_port,
> "mirror" and "taas". port with type "mirror" is used as inport for monitor
> flow in logica_switch, and port with type "taas" is used as outport
On Tue, Aug 29, 2017 at 10:06:54AM -0700, Greg Rose wrote:
> On 08/28/2017 08:23 AM, Lance Richardson wrote:
> >Sparse version 0.5.1 will be released in the near future. Prepare
> >for it by eliminating the only new warning (as of 0.5.1-rc5):
> >
> >ofproto/fail-open.c:134:22: error: Initializer
On Tue, Aug 29, 2017 at 08:25:28AM -0700, Greg Rose wrote:
> On 08/25/2017 07:27 PM, fukaige wrote:
> >From: Kaige Fu
> >
> >Signed-off-by: Kaige Fu
> >---
> > lib/dpif.c | 10 +-
> > 1 file changed, 5 insertions(+), 5 deletions(-)
> >
> >diff
On Thu, Aug 17, 2017 at 03:10:34PM -0700, Greg Rose wrote:
> On 08/14/2017 12:11 AM, fukaige wrote:
> >From: Kaige Fu
> >
> >Start a virtual machine with its backend tap device attached to a brought up
> >linux bridge.
> >If we delete the linux bridge when vm is still
On Mon, Aug 07, 2017 at 08:48:07PM +0530, Jai Singh Rana wrote:
> Currently ovn-controller uses default unix domain socket in Open
> vSwitch's "run" directory to connect to OpenFlow switch. If
> ovn-controller/ovsdb-server and vswitchd are running on different
> systems, remote connection method
On Thu, Jul 27, 2017 at 11:38:00PM -0700, Tonghao Zhang wrote:
> When inserting or updating (e.g. emc_insert) a flow to EMC,
> we compare (e.g the hash and miniflow ) the netdev_flow_key.
> If the key is matched, we will update it. If we didn’t find
> the miniflow in the cache, the new flow will
On Mon, Oct 02, 2017 at 03:49:28PM +0100, Michal Weglicki wrote:
> This commit extends netdev_dpdk_get_status API to include additional
> driver-related information: if_type and if_descr.
>
> v2->v3: Code rebase.
> v3->v4: Minor comments applied.
>
> Co-authored-by: Michal Weglicki
On Mon, Oct 30, 2017 at 04:00:20PM -0700, Ben Pfaff wrote:
> On Mon, Oct 30, 2017 at 03:52:38PM +0300, Ilya Maximets wrote:
> > CC: Robert Wojciechowicz
> > CC: Antonio Fischetti
> > Fixes: d555d9bded5f ("netdev-dpdk: Create separate
On Mon, Oct 30, 2017 at 03:52:38PM +0300, Ilya Maximets wrote:
> CC: Robert Wojciechowicz
> CC: Antonio Fischetti
> Fixes: d555d9bded5f ("netdev-dpdk: Create separate memory pool for each
> port.")
> Fixes: 65056fd79694
Great, thanks for the info.
On Mon, Oct 30, 2017 at 09:57:49PM +, Darrell Ball wrote:
> I believe this series is superceded by the HWOL series from Yuanhan Liu
> y...@fridaylinux.org
>
>
> On 10/30/17, 2:22 PM, "ovs-dev-boun...@openvswitch.org on behalf of Ben
> Pfaff"
I believe this series is superceded by the HWOL series from Yuanhan Liu
y...@fridaylinux.org
On 10/30/17, 2:22 PM, "ovs-dev-boun...@openvswitch.org on behalf of Ben Pfaff"
wrote:
On Wed, Jul 05, 2017 at 12:27:08PM +, Shachar
OK, great, thanks.
On Mon, Oct 30, 2017 at 09:53:01PM +, Darrell Ball wrote:
> This patch was merged to dpdk_merge on Sept 6 and later merged to ovs.
>
> https://mail.openvswitch.org/pipermail/ovs-dev/2017-September/338413.html
>
>
> On 10/30/17, 2:38 PM, "ovs-dev-boun...@openvswitch.org
Thank you for sending this series. I feel bad that this series did not
receive a timely review. Let me help a little bit.
I don't entirely understand the series. This is mostly because of a
lack of context, which the series could be revised to provide.
One issue I have is that the term
This patch was merged to dpdk_merge on Sept 6 and later merged to ovs.
https://mail.openvswitch.org/pipermail/ovs-dev/2017-September/338413.html
On 10/30/17, 2:38 PM, "ovs-dev-boun...@openvswitch.org on behalf of Ben Pfaff"
wrote:
Hi Ian and Gao, it looks like this patch never got applied. Gao, is it
still relevant and, if so, Ian, will you add it to your next batch of
patches?
Thanks,
Ben.
On Sun, Sep 03, 2017 at 10:46:24AM +, Stokes, Ian wrote:
> OK,
>
> In my own testing I’m seeing the same behavior with a tap
On Wed, Jul 12, 2017 at 07:22:58AM +, Zoltán Balogh wrote:
>
> It turned out, checking datapath flow statistics during system-userspace
> test is not reliable. Unwanted packets can be injected depending on
> system configuration. As a workaround, this commit removes checking
> statistics of
On Wed, Jul 05, 2017 at 12:27:08PM +, Shachar Beiser wrote:
> The HW set the mark id that represents matching rule.
> The hw-pipeline reads the mark id from fdir.hi
>
> Signed-off-by: Shachar Beiser
I'm currently going through old patches. It looks like this series
Las mejores técnicas para la recuperación de su cartera vencida
Habilidades de cobranza altamente eficaces
10 de noviembre - MCE. Abdón Guzmán Santana Zárate9am-6pm
El arte de la cobranza requiere de técnicas de negociación y habilidades para
la interacción y el trato humano. El éxito de esta
On Fri, Oct 27, 2017 at 02:18:43PM +0200, Timothy Redaelli wrote:
> python-* package names are deprecated on Fedora
> (https://fedoraproject.org/wiki/Packaging:Naming#Python2_binary_package_naming)
> so use python2-sphinx, when available, instead.
>
> CC: Lance Richardson
>
On Fri, Oct 27, 2017 at 02:03:15PM +0200, Timothy Redaelli wrote:
> python-* package names are deprecated
> (https://fedoraproject.org/wiki/Packaging:Naming#Python2_binary_package_naming)
> so use python2-sphinx instead.
>
> CC: Gurucharan Shetty
> Fixes: fc9669525f3f ("rhel,
On Tue, Jul 11, 2017 at 11:22:26AM -0700, Ben Pfaff wrote:
> On Tue, Jun 20, 2017 at 05:43:56PM +0200, Timothy Redaelli wrote:
> > Add --enable-static-binaries configuration option for enabling or disabling
> > static linking with libopenvswitch, libsflow, libovsdb, libvtep, libovn and
> >
Justin, I don't think there's been a review of this yet. It's been
almost 8 months. Are you planning to take a look?
On Wed, Mar 08, 2017 at 03:58:28PM -0800, Ben Pfaff wrote:
> Justin, you've reviewed previous versions of this patch. Will you look
> at this one?
>
> On Sun, Feb 19, 2017 at
On Tue, Aug 29, 2017 at 03:27:53PM -0700, Han Zhou wrote:
> Add parameters in local Open_vSwitch DB external-ids for rate-
> limiting the log file writing:
> ovn-acl-log-rl-rate
> ovn-acl-log-rl-burst
>
> Note: this has nothing to do with packet-in rate-limiting.
> Signed-off-by: Han Zhou
On Mon, Oct 30, 2017 at 12:49 PM, Ben Pfaff wrote:
> On Mon, Oct 30, 2017 at 10:27:55AM -0700, William Tu wrote:
>> Clang reports potiential leak of memory pointed to by 'pname'.
>> We already free the previous subporgram name when setting the
>> new subprogram name, and once the
On Mon, Oct 30, 2017 at 10:27:55AM -0700, William Tu wrote:
> Clang reports potiential leak of memory pointed to by 'pname'.
> We already free the previous subporgram name when setting the
> new subprogram name, and once the running thread exits, the
> name of the process will be free by the
On Mon, Oct 30, 2017 at 10:27:54AM -0700, William Tu wrote:
> Clang reports value stored to 'sb' is never read.
> Fix it by removing it.
>
> Signed-off-by: William Tu
Thanks, applied to master.
___
dev mailing list
On Mon, Oct 30, 2017 at 10:27:53AM -0700, William Tu wrote:
> Clang reports value stored to 'eah' in never read.
> Fix it by removing it.
>
> Signed-off-by: William Tu
Thanks a lot! Applied to master.
___
dev mailing list
On Mon, Oct 30, 2017 at 10:27:52AM -0700, William Tu wrote:
> Clang reports possibly bad free of 'ofm' when it comes from the stack
> instead of malloc because Clang is not able to verify whether the previous
> if condition 'ctx->xin->xcache' still hold the same. Fix it by
> adding additional
On Sat, Oct 28, 2017 at 10:31:56AM -0700, William Tu wrote:
> Clang reports possible null pointer '>masks.ipv6_src' to memset.
> Workaround it by adding extra pointer check.
>
> Signed-off-by: William Tu
> ---
> lib/tnl-neigh-cache.c | 3 +++
> 1 file changed, 3
I'd recommend just something like "if (n_flows) { qsort(...); }".
On Mon, Oct 30, 2017 at 12:05:55PM -0700, William Tu wrote:
> Hi Mark,
> Thanks for the review. I will fix it and resubmit v2.
> William
>
> On Mon, Oct 30, 2017 at 8:12 AM, Mark Michelson wrote:
> > This is
On Sat, Oct 28, 2017 at 10:31:54AM -0700, William Tu wrote:
> Clang reports possible null pointer passes to dst of memcmp, >ip6_src.
> This is due to *nh comes from dp_packet_l3, and dp_packet_l3 might
> return NULL. Fix it by adding ovs_assert check.
>
> Signed-off-by: William Tu
On Sat, Oct 28, 2017 at 10:31:53AM -0700, William Tu wrote:
> Clang reports possible null pointer to memmove due to
> dp_packet_data might retun null. Fix it by adding
> ovs_assert check.
>
> Signed-off-by: William Tu
> ---
> lib/packets.c | 1 +
> 1 file changed, 1
On Sat, Oct 28, 2017 at 10:31:50AM -0700, William Tu wrote:
> Clang reports possible optarg as null pointer passed to atoi.
> Fix it by adding ovs_assert check before.
>
> Signed-off-by: William Tu
> ---
> utilities/ovs-appctl.c | 1 +
> 1 file changed, 1 insertion(+)
>
>
On Sat, Oct 28, 2017 at 10:31:48AM -0700, William Tu wrote:
> Clang reports possible null pointer argument to the memcpy src.
> This is due to at dp_packet_clone_data_with_headroom, the
> dp_packet *b might have a NULL base due to allocating a dp_packet
> with size = 0. Fix it by adding
On Mon, Oct 30, 2017 at 11:27 AM, Ben Pfaff wrote:
>
> On Fri, Oct 27, 2017 at 05:40:56PM -0700, Han Zhou wrote:
> > IDL index should be able to be used without having to be in a
> > transaction. However, current implementation leads to crash if
> > a reference type column is being
IDL index should be able to be used without having to be in a
transaction. However, current implementation leads to crash if
a reference type column is being set in an index row for querying
purpose when it is not in a transaction. It is because of the
uninitialized arcs and unnecessary updates of
On Sat, Oct 28, 2017 at 10:31:47AM -0700, William Tu wrote:
> Before the patch, the scan-build reports 46 bugs found.
> The patch series fix the clang static checker bug group:
> "Argument with nonnull attribute passed null"
> Most of the fixes are adding "ovs_assert" to tell the checker
> that
On Sat, Oct 28, 2017 at 10:31:49AM -0700, William Tu wrote:
> Clang reports possible null pointer in_dhcp_opt passing to memcmp.
> This might due to dp_packet_get_udp_payload retuning null. Fix it
> by adding ovs_assert.
>
> Signed-off-by: William Tu
> ---
>
Hi Mark,
Thanks for the review. I will fix it and resubmit v2.
William
On Mon, Oct 30, 2017 at 8:12 AM, Mark Michelson wrote:
> This is not the proper solution. If you start ovn-northd and do not add any
> logical switches or routers, then there will be no logical flows
On Fri, Oct 27, 2017 at 05:40:56PM -0700, Han Zhou wrote:
> IDL index should be able to be used without having to be in a
> transaction. However, current implementation leads to crash if
> a reference type column is being set in an index row for querying
> purpose when it is not in a transaction.
On Mon, Oct 30, 2017 at 10:46:42AM -0700, William Tu wrote:
> Replace path 'lib/ofp-msgs.h' with 'include/openvswitch/ofp-msgs.h'
>
> Signed-off-by: William Tu
Thanks, applied to master.
___
dev mailing list
d...@openvswitch.org
Replace path 'lib/ofp-msgs.h' with 'include/openvswitch/ofp-msgs.h'
Signed-off-by: William Tu
---
Documentation/faq/contributing.rst | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/Documentation/faq/contributing.rst
Thanks for the review.
Guru and I had some offline discussion. We have concern about possible
packet lost when unconditionally remove the interface and bring it
again. Currently if the userspace ovs-vswitchd dies, since the kernel
datapath module still remains, packets can still go through.
If
On Thu, Oct 26, 2017 at 2:24 PM, Yi-Hung Wei wrote:
> Three conntrack system-traffic tests are broken because of a recent
> change 7827edcaebd8 ("Add dl_type to flow metadata for correct
> interpretation of conntrack metadata"). It can be reproduced by
> $ make
Clang reports potiential leak of memory pointed to by 'pname'.
We already free the previous subporgram name when setting the
new subprogram name, and once the running thread exits, the
name of the process will be free by the system. Thus, we don't
need to explicitly free it, so exclude the clang
Clang reports value stored to 'sb' is never read.
Fix it by removing it.
Signed-off-by: William Tu
---
ovn/utilities/ovn-sbctl.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/ovn/utilities/ovn-sbctl.c b/ovn/utilities/ovn-sbctl.c
index
Clang reports value stored to 'eah' in never read.
Fix it by removing it.
Signed-off-by: William Tu
---
lib/ofp-actions.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
index 71eb70c3c239..79041c281ed0 100644
---
The patch series fix three bug types reported by clang,
dead assignment, bad free, and memory leak.
William Tu (4):
ofproto-dpif-xlate: Fix bad memory free.
ofp-actions: Fix dead assignment.
ovn-sbctl: Fix dead assignment.
util: Fix potential leak of memory.
lib/ofp-actions.c
Clang reports possibly bad free of 'ofm' when it comes from the stack
instead of malloc because Clang is not able to verify whether the previous
if condition 'ctx->xin->xcache' still hold the same. Fix it by
adding additional condition.
Signed-off-by: William Tu
---
On Thu, Oct 26, 2017 at 03:08:00PM -0700, Yi-Hung Wei wrote:
> On Mon, Oct 9, 2017 at 10:33 AM, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
> > ---
> > Documentation/faq/contributing.rst | 44
> > +++---
> >
On Sat, Oct 28, 2017 at 04:38:30PM +0100, Richard Oliver wrote:
> [Problem]
> Compilation error on newer versions of macOS (Sierra onwards) due to
> multiple declarations of clock_gettime.
>
> [Solution]
> Have configure check for clock_gettime and check this result in
> timeval to avoid
On Mon, Oct 30, 2017 at 10:51:46AM -0200, Flavio Leitner wrote:
> On Fri, 27 Oct 2017 09:59:33 -0700
> Ben Pfaff wrote:
>
> > "echo -n" is not POSIX and has spotty support in shells.
> >
> > CC: Timothy Redaelli
> > CC: Flavio Leitner
> >
One possible use for this DNS feature is to identify cluster members for
OVSDB clusters. That might be an appropriate use for a low TTL.
On Mon, Oct 30, 2017 at 10:52:14AM +0100, Miguel Angel Ajo Pelayo wrote:
> I don't believe rounding up TTLs would be a good practice. The
> administrators
>
I've reviewed everything except for patches 1, 6, 7, and 9. I'm not very
familiar with those areas of code and do not feel as comfortable giving an
ACK to those.
On Sat, Oct 28, 2017 at 12:32 PM William Tu wrote:
> Before the patch, the scan-build reports 46 bugs found.
>
Looks good to me
On Sat, Oct 28, 2017 at 12:38 PM William Tu wrote:
> Clang reports possible optarg as null passed to atoi.
> Fix it by adding ovs_assert check.
>
> Signed-off-by: William Tu
>
Acked-by: Mark Michelson
> ---
>
This looks fine to me.
If you end up needing to re-work this patch series, I think it would be
appropriate to combine this with patch #4 since this is essentially fixing
the same possible NULL pointer.
On Sat, Oct 28, 2017 at 12:38 PM William Tu wrote:
> Clang reports
Looks good to me.
On Sat, Oct 28, 2017 at 12:35 PM William Tu wrote:
> Clang reports possible null pointer of optarg as argument to strtoul.
> Fix it by adding ovs_assert check.
>
> Signed-off-by: William Tu
>
Acked-by: Mark Michelson
Looks good to me.
On Sat, Oct 28, 2017 at 12:35 PM William Tu wrote:
> Clang reports possible null pointer due to process_register could
> take the name from argv[0]. Fix it by adding ovs_assert check.
>
> Signed-off-by: William Tu
>
Acked-by: Mark
Looks good to me.
On Sat, Oct 28, 2017 at 12:34 PM William Tu wrote:
> Clang reports possible optarg as null pointer passed to atoi.
> Fix it by adding ovs_assert check before.
>
> Signed-off-by: William Tu
>
Acked-by: Mark Michelson
I am Mrs Fofana Maladho Boukhary, I was born on June 8, 1969. I am a widow, I
am so happy to be writing this message because it rejoices my heart if you can
share my passion to get this project I want to get started.
I want you to understand that my current medical status will not allow me to
This looks fine to me. Usually I don't like assertions based on contents of
incoming packets. However, in this case, ovn-northd has set up the logical
flow so that only UDP packets can reach here.
On Sat, Oct 28, 2017 at 12:34 PM William Tu wrote:
> Clang reports possible
On Wed, Oct 25, 2017 at 02:24:15PM +0300, Roi Dayan wrote:
>
>
> On 27/09/2017 12:08, Simon Horman wrote:
> > On Mon, Sep 25, 2017 at 04:31:42PM +0300, Paul Blakey wrote:
> > >
> > >
> > > On 18/09/2017 18:01, Simon Horman wrote:
> > > > On Mon, Sep 18, 2017 at 07:16:03AM +0300, Roi Dayan
Did you receive my previous email regarding your family inheritance?
Y.T. Chan
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
CC: Robert Wojciechowicz
Fixes: d555d9bded5f ("netdev-dpdk: Create separate memory pool for each port.")
Signed-off-by: Ilya Maximets
---
lib/netdev-dpdk.c | 18 --
1 file changed, 4 insertions(+), 14 deletions(-)
diff
Since commit d555d9bded5f ("netdev-dpdk: Create separate memory pool
for each port."), struct dpdk_mp is redundant because each mempool
can be used by single port only and this port already contains all
the information we store in dpdk_mp.
There is no need to duplicate the information.
Fields of
CC: Robert Wojciechowicz
CC: Antonio Fischetti
Fixes: d555d9bded5f ("netdev-dpdk: Create separate memory pool for each port.")
Fixes: b6b26021d2e2 ("netdev-dpdk: fix management of pre-existing mempools.")
Signed-off-by: Ilya Maximets
CC: Robert Wojciechowicz
CC: Antonio Fischetti
Fixes: d555d9bded5f ("netdev-dpdk: Create separate memory pool for each port.")
Fixes: 65056fd79694 ("netdev-dpdk: manage failure in mempool name creation.")
Signed-off-by: Ilya Maximets
On Fri, 27 Oct 2017 09:59:33 -0700
Ben Pfaff wrote:
> "echo -n" is not POSIX and has spotty support in shells.
>
> CC: Timothy Redaelli
> CC: Flavio Leitner
> Signed-off-by: Ben Pfaff
> ---
> Makefile.am | 11
If several bridges are connected via patch ports or the packet is sent
via tunnel and an error occurs during translation in a subsequent bridge,
e.g. Ethernet header cannot be pushed to the packet due to lack of
prerequisites in the second bridge and there were flow actions already
During translation of OF actions on a bridge, we can store the last
valid state of translated actions while iterating over the OF actions
and revert to it in case of error. This can be performed in the
do_xlate_actions() funtion.
Signed-off-by: Zoltan Balogh
When all OF actions have been translated, there could be actions at
the end of list of odp actions which are not needed to be executed.
So, the list can be normalized at the end of xlate_actions().
Signed-off-by: Zoltan Balogh
Signed-off-by: Sugesh Chandran
I don't believe rounding up TTLs would be a good practice. The
administrators
are aware of the risks of having a very low TTL, so if they decide to pick
a low TTL they may have good reasons (like the possibility of a very close
event of the IP being moved, or some sort of failover mechanism, or a
Hello,
Can we proceed further with this patch if there are no more comments?
Thank you in advance.
Br,
Michal.
> -Original Message-
> From: Weglicki, MichalX
> Sent: Thursday, October 12, 2017 10:03 AM
> To: d...@openvswitch.org
> Cc: Greg Rose ; Weglicki,
79 matches
Mail list logo