Yes, that's the branch I meant (I'm starting from it as a place to get
the ddlog code up to speed with OVN master).
That makes sense. Thanks.
On Fri, Feb 07, 2020 at 04:01:04PM -0800, Justin Pettit wrote:
> I think this is in the ddlog-dev-v2 branch, not master, correct? That branch
> was a
I think this is in the ddlog-dev-v2 branch, not master, correct? That branch
was a working directory for me, and I probably just pulled those patches in the
wrong order. Does that make sense?
--Justin
> On Feb 7, 2020, at 2:58 PM, Ben Pfaff wrote:
>
> Hi Justin.
>
> In the OVN repo,
If a TCP connection is in SYN_SENT state, receiving another SYN packet
would just renew the timeout of that conntrack entry rather than create
a new one. Thus, tcp_conn_update() should return CT_UPDATE_VALID_NEW.
This also fixes regressions of a couple of OVN system tests.
Fixes: a867c010ee91
Hi Justin.
In the OVN repo, commit 00523aa8e970 ("tests: Add function to handle
ddlog's id allocation.") adds a stableid utility and uses it in ovn.at.
A few commits later, commit 15e5a8b00cdf ("ovn.at: Break "ovn.at" into
"ovn-components.at" and "ovn-e2e.at".") stops using it and also drops a
Done.
On Fri, Feb 07, 2020 at 10:02:06PM +0530, Martin Varghese wrote:
> Thanks Ben,
> Please keep martin.vargh...@nokia.com
>
> On Fri, Feb 7, 2020 at 9:33 PM Ben Pfaff wrote:
>
> > Of course.
> >
> > I see a couple of email addresses for you. Which one is best for the
> > authors file?
> >
If not supplied by the user --socket-mem and --socket-limit EAL options
are set to a default value based system topology. The assumption that
numa nodes must be numbered consecutive is removed by this change.
These options can be seen in the ovs-vswitchd.log. For example:
a system containing
Ovs-numa currently makes the assumption that numa node ids and cpu core ids
will be numbered consecutively. Current Power systems don't always follow this
model. Furthermore, cpus on Power may be on/off lined based the setting of
Simultaneous multithreading (SMT). The result can be gaps in the
This change removes the assumption that numa nodes and cores are numbered
consecutively in linux. This change is required to support some Power
systems.
An additional check has been added to verify that cores are online,
offline cores result in non-consecutively numbered cores.
I manually
Hello,
Further looking at the code, when upcall_xlate is performed, we create a
hardcoded status block to represent the first packet:
stats.n_packets = 1;
stats.n_bytes = dp_packet_size(upcall->packet);
stats.used = time_msec();
stats.tcp_flags = ntohs(upcall->flow->tcp_flags);
Thanks Ben,
Please keep martin.vargh...@nokia.com
On Fri, Feb 7, 2020 at 9:33 PM Ben Pfaff wrote:
> Of course.
>
> I see a couple of email addresses for you. Which one is best for the
> authors file?
>
> On Fri, Feb 07, 2020 at 04:47:42PM +0530, Martin Varghese wrote:
> > HI
> > I checked the
On 2/7/20 12:50 PM, yang_y...@126.com wrote:
> From: Yi Yang
>
> We can avoid high system call overhead by using TPACKET_V3
> and using DPDK-like poll to receive and send packets (Note: send
> still needs to call sendto to trigger final packet transmission).
>
>>From Linux kernel 3.10 on,
Bleep bloop. Greetings Flavio Leitner, I am a robot and I have tried out your
patch.
Thanks for your contribution.
I encountered some error that I wasn't expecting. See the details below.
checkpatch:
WARNING: Line lacks whitespace around operator
#27 FILE: lib/dp-packet.h:1052:
There is a cache miss when accessing mbuf->data_off while cloning
a batch and using prefetch improved the throughput by ~2.3%.
Before: 13709416.30 pps
After: 14031475.80 pps
Fixes: d48771848560 ("dp-packet: preserve headroom when cloning a pkt batch")
Signed-off-by: Flavio Leitner
---
From: Yi Yang
We can avoid high system call overhead by using TPACKET_V3
and using DPDK-like poll to receive and send packets (Note: send
still needs to call sendto to trigger final packet transmission).
>From Linux kernel 3.10 on, TPACKET_V3 has been supported,
so all the Linux kernels current
HI
I checked the Authors file today and coudn't find my name. I would like to
see my name there :) if i meet the criteria
Thanks,
Martin
On Fri, Nov 22, 2019 at 10:57 PM Ben Pfaff wrote:
> On Fri, Nov 22, 2019 at 08:32:53AM -0800, Gregory Rose wrote:
> > On 11/21/2019 10:07 PM, Martin Varghese
On 2/6/20 2:35 PM, Ilya Maximets wrote:
> DPDK drivers has different implementations of transmit functions.
> Enabled offloading may cause driver to choose slower variant
> significantly affecting performance if userspace TSO wasn't requested.
>
> Fixes: 29cf9c1b3b9c ("userspace: Add TCP
Just repeating last mail with fixed indentations (mail was in wrong format):
Hello Ben, Pravin,
Thank you for your consideration.
>> It is also simpler to fix the stats issue using this approach.
>There's no stats issue. Userspace just counts the number of packets it
>sent and adds them
Hello Ben, Pravin,
Thank you for your consideration.
> It is also simpler to fix the stats issue using this approach.
There's no stats issue. Userspace just counts the number of packets it
sent and adds them in.
Regarding the stats issue. In case of LTE Broadcast, I have tight
synchronization
From: Anton Ivanov
1. Provides signigificant performance improvement in large OVSDB
installs such as southdb in OVN by removing the current "feature" of
"trying to break an EAGAIN wall" in all streams. All streams will
attempt IO IF AND ONLY IF there is IO pending. If there is no IO
signalled by
From: Anton Ivanov
Run JSON RPC every time there is a backlog and not only if
backlog was created on this iteration.
Signed-off-by: Anton Ivanov
---
lib/jsonrpc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/jsonrpc.c b/lib/jsonrpc.c
index ed748dbde..fc354e517
From: Anton Ivanov
In order to allow switching between poll and epoll we need to
switch from the direct use of POLLXXX macros to OVS_POLLXXX which
can be either POLLXXX or EPOLLXXX depending on the actual backend.
Signed-off-by: Anton Ivanov
---
include/openvswitch/poll-loop.h | 6 ++
21 matches
Mail list logo