On 7/31/15 6:30 PM, Flavio Leitner wrote:
This RFC is based on the vhost multiple queues work on
dpdk-dev: http://dpdk.org/ml/archives/dev/2015-June/019345.html
Signed-off-by: Flavio Leitner f...@redhat.com
---
lib/netdev-dpdk.c | 61 ---
1
This is cleaner than I was expecting. Do you think we want to backport it to
2.3?
--Justin
On Aug 6, 2015, at 12:30 PM, Jarno Rajahalme jrajaha...@nicira.com wrote:
Ethan,
This ended up being a very clean patch, thank you!
IMO this is good to go, but I’d like to see a test case that
Acked-by: Jarno Rajahalme jrajaha...@nicira.com
On Aug 5, 2015, at 9:59 PM, Ben Pfaff b...@nicira.com wrote:
OpenFlow 1.5 says:
If the group-mod request specifies more than one bucket for a group of
type Indirect, the switch must refuse to add the group entry and must
send an
Packets were always resubmmited to the remote output table, which
handles output to a remote chassis. Broadcast packets were only sent
remotely and not to local ports on the same logical switch. Unicast
packets between 2 local ports also didn't work.
This patch makes it so packets are sent to
On Fri, Aug 07, 2015 at 12:54:12AM +0530, laxmikanta.beh...@gmail.com wrote:
Hi,
we are sending two different stream from same port using TG.
Do you want to say that only one host can connect in each port. My use case
is I connect multiple host in a single host.
Yes.
Please let us know
Looks good to me, but some comments below.
Acked-by: Jarno Rajahalme jrajaha...@nicira.com
On Aug 5, 2015, at 9:59 PM, Ben Pfaff b...@nicira.com wrote:
OpenFlow 1.0.1 says:
The dl_vlan_pcp field must be ignored when the OFPFW_DL_VLAN wildcard
bit is set or when the dl_vlan value is
Packets were always resubmmited to the remote output table, which
handles output to a remote chassis. Broadcast packets were only sent
remotely and not to local ports on the same logical switch. Unicast
packets between 2 local ports also didn't work.
This patch makes it so packets are sent to
On 6 August 2015 at 00:54, Ilya Maximets i.maxim...@samsung.com wrote:
ping.
On 30.07.2015 18:29, Ilya Maximets wrote:
In multiqueue mode several pmd threads may process one
port, but different queues. Flow doesn't depend on queue.
So, while miss upcall processing, all threads (except first
On 05.08.2015 20:04, Daniele Di Proietto wrote:
On 05/08/2015 17:47, Ilya Maximets i.maxim...@samsung.com wrote:
On 05.08.2015 19:26, Daniele Di Proietto wrote:
On 05/08/2015 16:42, Ilya Maximets i.maxim...@samsung.com wrote:
Sorry, I agree that example is incorrect. It is really
Packets were always resubmmited to the remote output table, which
handles output to a remote chassis. Broadcast packets were only sent
remotely and not to local ports on the same logical switch. Unicast
packets between 2 local ports also didn't work.
This patch makes it so packets are sent to
On Thu, Aug 06, 2015 at 03:24:29PM -0400, Thomas F Herbert wrote:
On 8/6/15 1:40 PM, Flavio Leitner wrote:
On Thu, Aug 06, 2015 at 12:39:29PM -0400, Thomas F Herbert wrote:
On 7/31/15 6:30 PM, Flavio Leitner wrote:
This RFC is based on the vhost multiple queues work on
dpdk-dev:
On Thu, Aug 06, 2015 at 04:00:36PM -0700, Jesse Gross wrote:
On Wed, Aug 5, 2015 at 10:20 PM, Simon Horman
simon.hor...@netronome.com wrote:
There appears to be a miss-match between the handling of miss-sized
neighbour discovery options in implementation of
parse_icmpv6() in user-space and
On Thu, Aug 6, 2015 at 7:43 PM, Simon Horman simon.hor...@netronome.com wrote:
On Thu, Aug 06, 2015 at 04:00:36PM -0700, Jesse Gross wrote:
On Wed, Aug 5, 2015 at 10:20 PM, Simon Horman
simon.hor...@netronome.com wrote:
There appears to be a miss-match between the handling of miss-sized
On Thu, Aug 06, 2015 at 08:27:02PM -0700, Jesse Gross wrote:
On Thu, Aug 6, 2015 at 7:43 PM, Simon Horman simon.hor...@netronome.com
wrote:
On Thu, Aug 06, 2015 at 04:00:36PM -0700, Jesse Gross wrote:
On Wed, Aug 5, 2015 at 10:20 PM, Simon Horman
simon.hor...@netronome.com wrote:
There
Struct miniflow is now sometimes used just as a map. Define a new
struct flowmap for that purpose. The flowmap is defined as an array of
maps, and it is automatically sized according to the size of struct
flow, so it will be easier to maintain in the future.
It would have been tempting to use
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/lib/classifier.c b/lib/classifier.c
index fc6a1f5..0a6855e 100644
--- a/lib/classifier.c
+++ b/lib/classifier.c
@@ -283,11 +283,10 @@
GCC (4.7) sees too wide shifts when there are none, refactor to
circumvent the false error.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/flow.c | 35 ++-
1 file changed, 22 insertions(+), 13 deletions(-)
diff --git a/lib/flow.c b/lib/flow.c
index
This makes stage mask computation happen only when a subtable is
inserted and allows simplification of the main lookup function.
Classifier benchmark shows that this speeds up the classification
(with wildcards) about 5%.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
Add a benchmark command for classifier lookup performance testing.
Usage:
ovstest test-classifier benchmark n_rules n_priorities n_subtables
n_threads n_lookups
Where:
n_rules - The number of rules to install for lookups. More rules
makes misses less likely.
Remove unused cls_rule_hash() and minimatch_hash() functions.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier-private.h | 7 ---
lib/classifier.c | 7 ---
lib/classifier.h | 1 -
3 files changed, 15 deletions(-)
diff --git
minimask_hash() can be simplified as each value is known to be non-zero.
Move miniflow_hash() into test-classifier.c as miniflow_hash__() as it
is no longer needed elsewhere.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier-private.h | 38
This email newsletter was sent to you in graphical HTML format.
If you're seeing this version, your email program prefers plain text emails.
You can read the original version online:
http://ymlp103.com/zBTq0h
This is a false positive sparse warning. I changed the flowmap_aux initializer
to use designated initializer instead of an inline function and that got rid of
the sparse warning.
I also separated some changes into separate patches.
I’ll send a v2 soon.
Jarno
On Aug 5, 2015, at 1:17 PM,
Remove unused cls_rule_hash() and minimatch_hash() functions.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier-private.h | 7 ---
lib/classifier.c | 7 ---
lib/classifier.h | 1 -
3 files changed, 15 deletions(-)
diff --git
Add a benchmark command for classifier lookup performance testing.
Usage:
ovstest test-classifier benchmark n_rules n_priorities n_subtables
n_threads n_lookups
Where:
n_rules - The number of rules to install for lookups. More rules
makes misses less likely.
This makes stage mask computation happen only when a subtable is
inserted and allows simplification of the main lookup function.
Classifier benchmark shows that this speeds up the classification
(with wildcards) about 5%.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
GCC (4.7) sees too wide shifts when there are none, refactor to
circumvent the false error.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/flow.c | 35 ++-
1 file changed, 22 insertions(+), 13 deletions(-)
diff --git a/lib/flow.c b/lib/flow.c
index
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/lib/classifier.c b/lib/classifier.c
index fc6a1f5..0a6855e 100644
--- a/lib/classifier.c
+++ b/lib/classifier.c
@@ -283,11 +283,10 @@
On Aug 6, 2015, at 16:29, Jesse Gross je...@nicira.com wrote:
On Thu, Aug 6, 2015 at 4:17 PM, Jarno Rajahalme jrajaha...@nicira.com
wrote:
Struct miniflow is now sometimes used just as a map. Define a new
struct flowmap for that purpose. The flowmap is defined as an array of
maps, and
On Thu, Aug 6, 2015 at 3:58 PM, Joe Stringer joestrin...@nicira.com wrote:
On 30 July 2015 at 12:51, Joe Stringer joestrin...@nicira.com wrote:
On 30 July 2015 at 11:57, Andy Zhou az...@nicira.com wrote:
On Wed, Jul 29, 2015 at 4:52 PM, Joe Stringer joestrin...@nicira.com
wrote:
We already
On Wed, Aug 5, 2015 at 10:20 PM, Simon Horman
simon.hor...@netronome.com wrote:
There appears to be a miss-match between the handling of miss-sized
neighbour discovery options in implementation of
parse_icmpv6() in user-space and in the kernel datapath.
This patch addresses that by modifying
On 30 July 2015 at 12:51, Joe Stringer joestrin...@nicira.com wrote:
On 30 July 2015 at 11:57, Andy Zhou az...@nicira.com wrote:
On Wed, Jul 29, 2015 at 4:52 PM, Joe Stringer joestrin...@nicira.com wrote:
We already queue the removal of the kernel module in OVS_VSWITCHD_START,
via an ON_EXIT()
Struct miniflow is now sometimes used just as a map. Define a new
struct flowmap for that purpose. The flowmap is defined as an array of
maps, and it is automatically sized according to the size of struct
flow, so it will be easier to maintain in the future.
It would have been tempting to use
minimask_hash() can be simplified as each value is known to be non-zero.
Move miniflow_hash() into test-classifier.c as miniflow_hash__() as it
is no longer needed elsewhere.
Signed-off-by: Jarno Rajahalme jrajaha...@nicira.com
---
lib/classifier-private.h | 38
On Thu, Aug 6, 2015 at 4:17 PM, Jarno Rajahalme jrajaha...@nicira.com wrote:
Struct miniflow is now sometimes used just as a map. Define a new
struct flowmap for that purpose. The flowmap is defined as an array of
maps, and it is automatically sized according to the size of struct
flow, so
I am Mrs. Gu Kailai and i intend to make a DONATION. please send your reply to
my personal email for more details via : w...@langsonenergy.com
===
The information in this email is confidential, and is intended solely for the
--verbose makes the ovs-ofctl process more verbose by allowing the
VLOG_DBG() functions to print to the terminal.
For kernel debugging, you want printk() or OVS_NLERR(), and these will
print to the kernel log (which shows up in dmesg). If you're not able
to see it there, then it may mean that
Great, I'll merge both today.
Thanks a lot,
Ethan
On Thu, Aug 6, 2015 at 9:00 AM, Stokes, Ian ian.sto...@intel.com wrote:
No problem,
I've sent a patch to the mailing list with these changes for INSTALL.DPDK.md.
Thanks
Ian
-Original Message-
From: Ethan Jackson
Is there a way of adding messages myself and see them thorugh --verbose? I
was told that printk() did that but I'm not able to see it anywhere :(
Thank you for your answer,
Kind regards,
Raúl
2015-08-05 22:40 GMT+02:00 Joe Stringer joestrin...@nicira.com:
On 5 August 2015 at 13:02, Raul
On Fri, Aug 07, 2015 at 12:00:14AM +0530, laxmikanta.beh...@gmail.com wrote:
Hi ,
IGMP Snooping table erasing multicast group IP after receiving the 1st
leave message from host. vswitch is not waiting till receive the last
leave message for the same multicast group IP from different
Ethan,
This ended up being a very clean patch, thank you!
IMO this is good to go, but I’d like to see a test case that shows the modify
operation in the logs to prevent regression in the future. Maybe work off some
of the bond rebalance test cases to prove that modify is used in the case we
Just to top-post briefly:
qemu reverted the vhost-user multiqueue support sometime back.
So this patch isn't really usable.
Regards,
-Karthick
commit f73ca7363440240b7ee5ee7f7ddb1c64751efb54
Merge: 7135847 f9d6dbf
Author: Peter Maydell peter.mayd...@linaro.org
Date: Mon Jul 20 13:25:28 2015
On Thu, Aug 06, 2015 at 10:54:16AM -0700, Karthick, A.R. wrote:
Just to top-post briefly:
qemu reverted the vhost-user multiqueue support sometime back.
So this patch isn't really usable.
That's correct. There is a new proposal enhancing the protocol
which will require work on the DPDK
On 08/03/2015 07:53 PM, Russell Bryant wrote:
On 08/03/2015 07:47 PM, Ben Pfaff wrote:
On Mon, Aug 03, 2015 at 07:44:03PM -0400, Russell Bryant wrote:
On 08/03/2015 07:42 PM, Ben Pfaff wrote:
Thank you for such a careful review!
I incorporated all of these comments and applied this to
Hi ,
IGMP Snooping table erasing multicast group IP after receiving the 1st
leave message from host. vswitch is not waiting till receive the last
leave message for the same multicast group IP from different host.
*Description*
1. Host 1 and 2 both connected with port number 2.
2.
Hi,
I know this is a low hanging fruit.
But still wanted to know if this can be merged.
Regards,
-Karthi
On Tue, Jul 28, 2015 at 11:33 AM, A.R. Karthick ar.karth...@cyaninc.com
wrote:
Hi,
Currently, the ovs netdev-dpdk ring type dpdkr, restricts dpdk ring name
size to 10 bytes (includes
On Thu, Aug 06, 2015 at 12:39:29PM -0400, Thomas F Herbert wrote:
On 7/31/15 6:30 PM, Flavio Leitner wrote:
This RFC is based on the vhost multiple queues work on
dpdk-dev: http://dpdk.org/ml/archives/dev/2015-June/019345.html
Signed-off-by: Flavio Leitner f...@redhat.com
---
Hi,
we are sending two different stream from same port using TG.
Do you want to say that only one host can connect in each port. My use case
is I connect multiple host in a single host.
Please let us know if OVS support multiple host can connect in a single
port.
Thanks Regards,
Laxmikanta
On 8/6/15 1:40 PM, Flavio Leitner wrote:
On Thu, Aug 06, 2015 at 12:39:29PM -0400, Thomas F Herbert wrote:
On 7/31/15 6:30 PM, Flavio Leitner wrote:
This RFC is based on the vhost multiple queues work on
dpdk-dev: http://dpdk.org/ml/archives/dev/2015-June/019345.html
Signed-off-by: Flavio
On Tue, Jul 28, 2015 at 11:33:09AM -0700, A.R. Karthick wrote:
Hi,
Currently, the ovs netdev-dpdk ring type dpdkr, restricts dpdk ring name
size to 10 bytes (includes the _rx/tx suffix).
This only gives us space to accommodate 10 rings as truncation happens
with dpdkr10_rx/tx.
The
ping.
On 30.07.2015 18:29, Ilya Maximets wrote:
In multiqueue mode several pmd threads may process one
port, but different queues. Flow doesn't depend on queue.
So, while miss upcall processing, all threads (except first
for that port) will receive error = ENOSPC due to
ukey_install
Currently tx_qid is equal to pmd-core_id. This leads to unexpected
behavior if pmd-cpu-mask different from '/(0*)(1|3|7)?(f*)/',
e.g. if core_ids are not sequential, or doesn't start from 0, or both.
Example:
starting 2 pmd threads with 1 port, 2 rxqs per port,
pmd-cpu-mask =
No problem,
I've sent a patch to the mailing list with these changes for INSTALL.DPDK.md.
Thanks
Ian
-Original Message-
From: Ethan Jackson [mailto:et...@nicira.com]
Sent: Tuesday, August 04, 2015 10:46 PM
To: Stokes, Ian
Cc: Ben Pfaff; Justin Pettit; Pravin Shelar; Traynor,
Currently there are restrictions regarding the use of the XL710 network
interface with OVS and DPDK. This patch details those restrictions in
INSTALL.DPDK.md.
Signed-off-by: Ian Stokes ian.sto...@intel.com
---
INSTALL.DPDK.md | 13 +
1 files changed, 13 insertions(+), 0
54 matches
Mail list logo