On Tue, Jun 12, 2012 at 12:13 PM, Ben Pfaff wrote:
> On Tue, Jun 12, 2012 at 11:40:22AM -0700, ravi kerur wrote:
>> On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
>> > Ravi, I'm more or less happy with the userspace code here. I mostly
>> > have style kinds of concerns. (I haven't fully scr
I am currently debugging this and had planned to send an email asking
for suggestions when code review was nearing completion. This bug has
been elusive for a long time. The packet format themselves looked fine
as I have found a way to test them and have successfully tested them.
The problem is th
On Tue, Jun 12, 2012 at 11:40:22AM -0700, ravi kerur wrote:
> On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
> > Ravi, I'm more or less happy with the userspace code here. I mostly
> > have style kinds of concerns. (I haven't fully scrutinized every line
> > though.)
> >
>
> is it coding s
Oh, I saw one testsuite failure also:
448: ofproto-dpif - controller FAILED
(ofproto-dpif.at:323)
Does this pass for you.
On Tue, Jun 12, 2012 at 11:11 AM, Ben Pfaff wrote:
> Ravi, I'm more or less happy with the userspace code here. I mostly
> have style kinds of concerns.
This is based off
commit 73c0ce349ba8d13a63a249a56aad0bec6e6caf26
Author: Joe Stringer
Date: Tue May 29 00:38:21 2012 +1200
it can wait until your return. I will probably do some performance
testing in the mean time for both mpls + qinq tcp offload.
Thanks,
Ravi
On Fri, Jun 1, 2012 at 2:29 P
On Thu, May 31, 2012 at 10:16:51AM -0700, ravi kerur wrote:
> I have generated single patch for OF 1.1 and 1.3 changes assuming it
> would be easier for code-review but as I suggested earlier, I will
> probably have 2 separate commits. Please let me know if you want me to
> separate them for code r
On Thu, May 31, 2012 at 6:54 PM, Jesse Gross wrote:
> On Thu, May 31, 2012 at 6:04 PM, ravi kerur wrote:
>> Thanks Jesse, I don't think I was clear. For mpls and qinq, offloading
>> goes through ovs emulation code in vport-netdev.c. In addiition to
>> TSO, there is checksum-offload and I believe
On Thu, May 31, 2012 at 6:04 PM, ravi kerur wrote:
> Thanks Jesse, I don't think I was clear. For mpls and qinq, offloading
> goes through ovs emulation code in vport-netdev.c. In addiition to
> TSO, there is checksum-offload and I believe this will be handled in
> the NIC and the driver sets the
Thanks Jesse, I don't think I was clear. For mpls and qinq, offloading
goes through ovs emulation code in vport-netdev.c. In addiition to
TSO, there is checksum-offload and I believe this will be handled in
the NIC and the driver sets the flags in DMA descriptor before sending
to the NIC. This, I d
On Thu, May 31, 2012 at 3:44 PM, ravi kerur wrote:
> Hi Jesse, Ben,
>
> I think I might have figured out performance issue with mpls/qinq and
> tcp offload mentioned earlier. The driver(intel e1000e driver) checks
> skb->protocol to be either ETH_P_IP/ETH_P_IPV6 and sets the flag in
> DMA descript
Hi Jesse, Ben,
I think I might have figured out performance issue with mpls/qinq and
tcp offload mentioned earlier. The driver(intel e1000e driver) checks
skb->protocol to be either ETH_P_IP/ETH_P_IPV6 and sets the flag in
DMA descriptors based on layer-4. This logic is basically used in
checksum
Thanks Jesse, I will check the code in vport-netdev.c file and get
back with the diffs. I did modify the code to get earlier version
working. Anyways will get back to you.
On Tue, May 29, 2012 at 3:32 PM, Jesse Gross wrote:
> On Tue, May 29, 2012 at 1:32 PM, ravi kerur wrote:
>> Thanks Jesse. Wh
On Tue, May 29, 2012 at 1:32 PM, ravi kerur wrote:
> Thanks Jesse. What I have tested is 2 Ubuntu 11.04 and 12.04 physical
> systems connected via a ethernet switch. I have tested scp downloads,
> iperf + netperf. Flows are setup such that MPLS tag is pushed for
> every outgoing TCP segments. I do
Jesse, Does this make sense?
Thanks,
Ravi
On Tue, May 29, 2012 at 1:32 PM, ravi kerur wrote:
> Thanks Jesse. What I have tested is 2 Ubuntu 11.04 and 12.04 physical
> systems connected via a ethernet switch. I have tested scp downloads,
> iperf + netperf. Flows are setup such that MPLS tag is pu
Thanks Jesse. What I have tested is 2 Ubuntu 11.04 and 12.04 physical
systems connected via a ethernet switch. I have tested scp downloads,
iperf + netperf. Flows are setup such that MPLS tag is pushed for
every outgoing TCP segments. I do see via tcpdump that a jumbo packet
(~15K) is sent out by s
On Tue, May 29, 2012 at 12:12 PM, ravi kerur wrote:
> Additional changes are needed here esp. in NIC driver(looked into
> e1000e driver code from Intel) as it requires handling of new
> eth_types. For these changes I probably need to work with you and
> Jesse on how best we can incorporate them.
On Tue, May 29, 2012 at 11:14 AM, Ben Pfaff wrote:
> Thanks so much for applying my comments. Will you please post the
> newest version of the MPLS patch? At this time it would be good to
> get your
> Signed-off-by: Ravi Kerur
> in the commit log, since it is getting closer. (You can re
Thanks so much for applying my comments. Will you please post the
newest version of the MPLS patch? At this time it would be good to
get your
Signed-off-by: Ravi Kerur
in the commit log, since it is getting closer. (You can read about
the purpose and meaning of this line in SubmittingPa
Hi Ben,
All your code review comments have been taken care for MPLS. I did
retest the code after incorporating code review comments. I also
modified the code for QinQ (common comments which apply to QinQ code
as well), added additional automation tests in odp.at for QinQ. In
addition, patch to sup
19 matches
Mail list logo
