With complex flow tables, facet_check_consistency() can be
expensive enough to show up in flow setup performance benchmarks.
In my testing this patch gives us a roughly 10% improvement in
TCP_CRR and ovs-benchmark.
Signed-off-by: Ethan Jackson
---
ofproto/ofproto-dpif.c |7 ++-
1 file ch
In the TCP_CRR benchmark, ovs-vswitchd spends so much time in
update_stats() that it has a significant impact on flow setup
performance. Further work is needed in this area, but for now,
simply rate limiting facet_learn() has a roughly 10% improvement
with complex flow tables.
Signed-off-by: Etha
We have a controller that puts many rules with different metadata values
into the flow table, where metadata is used (by "resubmit"s) to distinguish
stages in a pipeline. Thus, any given flow only needs to be hashed into
classifier "cls_table"s that contain a match for the flow's metadata value.
T
These seem slightly nicer than e.g. htons(UINT16_MAX).
Signed-off-by: Ben Pfaff
---
include/openvswitch/types.h |6 +-
lib/match.c | 30 +++---
lib/meta-flow.c | 11 +--
lib/nx-match.c | 10 +-
lib/ofp
Signed-off-by: Ben Pfaff
---
lib/classifier.c |3 +-
lib/flow.c | 122 ++
lib/flow.h | 10 +++--
lib/match.c | 31 +-
lib/match.h | 18
5 files changed, 143 insertions(+), 41 deletions(-)
di
On Mon, Mar 25, 2013 at 12:03 PM, Jarno Rajahalme
wrote:
> Eliminate the extra call to ip_select_ident(), and place the
> __ip_select_ident() call where the ip_select_ident() call was.
> This fixes two problems: Before, the call to ip_select_ident() did
> always zero out the value set earlier by _
On Mon, Mar 25, 2013 at 12:03 PM, Jarno Rajahalme
wrote:
>
> Signed-off-by: Jarno Rajahalme
Applied, thanks.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Je veux que tu m'aides à réaliser dans ton pays un projet de donation d'une
somme de 2.025.000€. Je suis actuellement hospitalisée à l'hôpital A
LONDRES pour le traitement d’une tumeur cérébrale que traine depuis 8
ans. En effet je vous contacte car mon défunt mari Ingénieur consultant
au Bénin
On Mon, Mar 25, 2013 at 12:03 PM, Jarno Rajahalme
wrote:
> Changes the default tunnel dont_fragment from "true" (don't
> fragment) to "false" (allow fragmentation). Tunnel outer headers
> will not have the DF bit set by default, and if "df=true" option is
> given for a tunnel, also local fragment
On Mon, Mar 25, 2013 at 12:20 PM, Pravin Shelar wrote:
> On Mon, Mar 25, 2013 at 9:39 AM, Jesse Gross wrote:
>> On Thu, Mar 21, 2013 at 8:08 PM, Pravin Shelar wrote:
>>> On Wed, Mar 20, 2013 at 4:44 PM, Jesse Gross wrote:
Allocation of the Netlink notification skb can potentially fail
These descriptions were vague to the point of being incorrect.
Reported-by: Kenneth Duda
Signed-off-by: Ben Pfaff
---
vswitchd/vswitch.xml | 10 ++
1 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index 2af04bd..00e5d19 100644
On Tue, Mar 26, 2013 at 6:09 AM, Kyle Mestery (kmestery)
wrote:
> On Mar 25, 2013, at 2:01 PM, Pravin B Shelar wrote:
>> While sending packet over tunnel it can get segmented or reallocated.
>> In that case we loose OVS_CB(). But some tunneling protocol
>> implementation needs OVS_CB(), e.g. vxla
On Mon, Mar 25, 2013 at 11:27 PM, 王国栋 wrote:
> could you explain it more specifically, what's the difference between
> internal device and other device?
> where dose it get the packet or when would the internal_dev_xmit procedure
> be called?
I would recommend reading the user documentation first
On Mar 25, 2013, at 2:01 PM, Pravin B Shelar wrote:
> While sending packet over tunnel it can get segmented or reallocated.
> In that case we loose OVS_CB(). But some tunneling protocol
> implementation needs OVS_CB(), e.g. vxlan needs OVS_CB()->flow which
> leads to kernel panic.
> Following patc
14 matches
Mail list logo
