On Tue, Apr 29, 2014 at 11:41:57AM -0700, Jesse Gross wrote:
> On Mon, Apr 28, 2014 at 5:13 PM, Simon Horman wrote:
> > On Mon, Apr 28, 2014 at 02:37:47PM -0700, Jesse Gross wrote:
> >> On Mon, Apr 28, 2014 at 12:00 AM, Simon Horman wrote:
> >> > On Fri, Apr 25, 2014 at 12:57:20PM -0700, Jesse Gr
Thanks for the review. Pushed.
On Tue, Apr 29, 2014 at 2:13 PM, Pritesh Kothari (pritkoth)
wrote:
> works for me here.
>
> Acked-by: Pritesh Kothari
>
> On Apr 29, 2014, at 1:22 PM, Andy Zhou wrote:
>
>> Caught by clang-3.5.
>>
>> Reported-by: Simon Horman
>> Signed-off-by: Andy Zhou
>> ---
>
On Tue, Apr 29, 2014 at 4:34 PM, Pravin B Shelar wrote:
> Patch fixes following warning:
> datapath/linux/flow_table.c:580:40: warning: symbol 'flow' shadows an earlier
> one
> datapath/linux/flow_table.c:558:24: originally declared here
>
> Reported-by: Ben Pfaff
> Signed-off-by: Pravin B Shela
On Wed, Apr 30, 2014 at 10:24:46AM +0900, YAMAMOTO Takashi wrote:
> The problem mentioned by Simon Horman in the following mail.
> http://openvswitch.org/pipermail/dev/2014-April/039492.html
>
> Cc: Simon Horman
> Signed-off-by: YAMAMOTO Takashi
One of the most pointless differences between OF1
The problem mentioned by Simon Horman in the following mail.
http://openvswitch.org/pipermail/dev/2014-April/039492.html
Cc: Simon Horman
Signed-off-by: YAMAMOTO Takashi
---
ofproto/ofproto-dpif-xlate.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/ofproto/ofproto-d
> The Open vSwitch software switch now supports all the required features of
> OpenFlow 1.0 through 1.3, with one known trivial exception[*]. Enable them
> by default in ovs-vswitchd.
>
> For now, ovs-ofctl only enables OpenFlow 1.0 by default. This is
> because ovs-ofctl implements command such
> On Tue, Apr 29, 2014 at 06:34:22PM +0900, Simon Horman wrote:
>> Add per-table counters. This resolves some short-comings
>> in the data provided in a table stats reply message.
>>
>> * Lookups and matches are calculated based on table
>> accesses rather than datapath flow hits and misses.
>>
This commit adds a new configuration "stats-update-interval" in
"other_config" of Open_Vswitch table. So users can control the
statistics update frequency. A possible use case is that, users
can lower the update frequency to reduce the cpu consumption of
the ovs-vswitchd thread.
The configured v
Thanks for the review Joe,
On Tue, Apr 29, 2014 at 4:45 PM, Joe Stringer wrote:
> On 30 April 2014 05:29, Alex Wang wrote:
>
>> +dnl disable the stats update, the following 'recv' should not be updated.
>> +AT_CHECK([ovs-vsctl set O . other_config:stats-update-interval=1000])
>> +for i in `
On 30 April 2014 05:29, Alex Wang wrote:
>
> +dnl disable the stats update, the following 'recv' should not be updated.
> +AT_CHECK([ovs-vsctl set O . other_config:stats-update-interval=1000])
> +for i in `seq 0 10`; do ovs-appctl time/warp 1000; done
> +for i in `seq 1 5`; do
> +AT_CHECK(
Thanks. Sent a fix.
On Tue, Apr 29, 2014 at 3:28 PM, Ben Pfaff wrote:
> GCC is telling me that ovs_flow_tbl_lookup_stats() has two variables
> both named 'flow':
>
> datapath/linux/flow_table.c:580:40: warning: symbol 'flow' shadows an
> earlier one
> datapath/linux/flow_table.c:558:24:
Patch fixes following warning:
datapath/linux/flow_table.c:580:40: warning: symbol 'flow' shadows an earlier
one
datapath/linux/flow_table.c:558:24: originally declared here
Reported-by: Ben Pfaff
Signed-off-by: Pravin B Shelar
---
datapath/flow_table.c |1 -
1 file changed, 1 deletion(-)
This commit adds a requirement that bfd session must receive at least
one bfd control packet every 100 * bfd->cfg_min_rx amount of time in
forwarding_if_rx mode. Otherwise, even if the data packets are received
on the monitored interface, the bfd->forwarding is still false.
Since the datapath flo
This commit adds a new requirement that cfm session must receive
at least one ccm every 100 * cfm_interval amount of time in demand
mode. Otherwise, even if the data packets are received on the
monitored interface, the cfm session still reports "[recv]" fault.
Since the datapath flow is not purge
On Apr 28, 2014, at 7:17 PM, Kmindg G wrote:
> On Tue, Apr 29, 2014 at 9:43 AM, Ethan Jackson wrote:
>> Acked-by: Ethan Jackson
>>
>> I'm not sure I totally follow Kmindg's comment, perhaps he could
>> explain further? At any rate, once addressed I"m happy with this.
>>
>
> After a second
On Apr 19, 2014, at 9:51 PM, Kmindg G wrote:
> On Sat, Apr 19, 2014 at 3:42 AM, Jarno Rajahalme
> wrote:
>>
>> static struct cls_match *
>> cls_match_alloc(struct cls_rule *rule)
>> {
>> -struct cls_match *cls_match = xmalloc(sizeof *cls_match);
>> +int count = count_1bits(rule->match
Series pushed!
Thanks for the review, Ethan!
Jarno
On Apr 29, 2014, at 2:35 PM, Ethan Jackson wrote:
> Meh I changed my mind again. Just merge the series instead of holding
> back, I'll make my tweaks on top of master.
>
> Ethan
>
> On Tue, Apr 29, 2014 at 2:34 PM, Ethan Jackson wrote:
>
On Apr 29, 2014, at 2:51 PM, Ben Pfaff wrote:
> On Fri, Apr 25, 2014 at 08:49:10AM -0700, Jarno Rajahalme wrote:
>> Allocate struct rule_actions and the space for the actions at once.
>> This reduces one memory indirection and helps reduce cache misses
>> visible in perf annotations.
>>
>> Fix
Pushed, thanks for the review!
Jarno
On Apr 29, 2014, at 3:11 PM, Ben Pfaff wrote:
> On Tue, Apr 29, 2014 at 03:01:47PM -0700, Jarno Rajahalme wrote:
>>
>> On Apr 29, 2014, at 2:55 PM, Jarno Rajahalme wrote:
>>
>>>
>>> On Apr 24, 2014, at 9:18 AM, Ben Pfaff wrote:
>>>
On Thu, Apr 2
GCC is telling me that ovs_flow_tbl_lookup_stats() has two variables
both named 'flow':
datapath/linux/flow_table.c:580:40: warning: symbol 'flow' shadows an
earlier one
datapath/linux/flow_table.c:558:24: originally declared here
If it's not a bug, it still seems worth fixing.
Thanks,
Signed-off-by: Pritesh Kothari
---
v6: rebase for changes in datapath/datapath.c
v5: move skb_clear_rxhash() from compat.h to skbuff.h,
simplify skb_get_hash() integration,
fix random.h order in acinclude.m4,
move pcpu_sw_netstats to netdevice.h.
v4: rebase for changes in datapath/acti
I pushed both.
Thanks for the reminder about backporting. I backported both of them.
On Tue, Apr 29, 2014 at 02:55:24PM -0700, Alex Wang wrote:
> Tested, worked,
>
> Thanks for fixing this,
>
> reminder: need backport, (maybe both patches)
>
> Acked-by: Alex Wang
>
>
> On Tue, Apr 29, 2014
On Tue, Apr 29, 2014 at 03:01:47PM -0700, Jarno Rajahalme wrote:
>
> On Apr 29, 2014, at 2:55 PM, Jarno Rajahalme wrote:
>
> >
> > On Apr 24, 2014, at 9:18 AM, Ben Pfaff wrote:
> >
> >> On Thu, Apr 24, 2014 at 09:09:03AM -0700, Jarno Rajahalme wrote:
> >>> rule_dpif_is_internal is among the t
On Apr 29, 2014, at 2:55 PM, Jarno Rajahalme wrote:
>
> On Apr 24, 2014, at 9:18 AM, Ben Pfaff wrote:
>
>> On Thu, Apr 24, 2014 at 09:09:03AM -0700, Jarno Rajahalme wrote:
>>> rule_dpif_is_internal is among the top ten OVS internal functions in
>>> recent perf reports. Inline it and some oth
Tested, worked,
Thanks for fixing this,
reminder: need backport, (maybe both patches)
Acked-by: Alex Wang
On Tue, Apr 29, 2014 at 2:40 PM, Ben Pfaff wrote:
> In production these may indicate a bug, but in the testsuite they probably
> just indicate a time-warp.
>
> Reported-by: Alex Wang
>
On Apr 24, 2014, at 9:18 AM, Ben Pfaff wrote:
> On Thu, Apr 24, 2014 at 09:09:03AM -0700, Jarno Rajahalme wrote:
>> rule_dpif_is_internal is among the top ten OVS internal functions in
>> recent perf reports. Inline it and some other equally trivial
>> functions.
>>
>> This change removes rule
On Fri, Apr 25, 2014 at 08:49:10AM -0700, Jarno Rajahalme wrote:
> Allocate struct rule_actions and the space for the actions at once.
> This reduces one memory indirection and helps reduce cache misses
> visible in perf annotations.
>
> Fix some old comments referring to ref count, since we now u
Acked-by: Alex Wang
On Tue, Apr 29, 2014 at 2:44 PM, Ben Pfaff wrote:
> The main thread has the empty string as its name, but that's not a good
> log string.
>
> Without this patch we can get log message like
> blocked 1000 ms waiting for to quiesce
> from ovsrcu_synchronize().
>
> Signed
The main thread has the empty string as its name, but that's not a good
log string.
Without this patch we can get log message like
blocked 1000 ms waiting for to quiesce
from ovsrcu_synchronize().
Signed-off-by: Ben Pfaff
---
lib/ovs-rcu.c |4 +++-
1 file changed, 3 insertions(+), 1 de
In production these may indicate a bug, but in the testsuite they probably
just indicate a time-warp.
Reported-by: Alex Wang
Signed-off-by: Ben Pfaff
---
tests/ofproto-macros.at |1 +
1 file changed, 1 insertion(+)
diff --git a/tests/ofproto-macros.at b/tests/ofproto-macros.at
index a82a9b
Meh I changed my mind again. Just merge the series instead of holding
back, I'll make my tweaks on top of master.
Ethan
On Tue, Apr 29, 2014 at 2:34 PM, Ethan Jackson wrote:
>> The almost equivalent function miniflow_equal_flow_in_minimask() does not
>> require the miniflow and minimask to hav
> The almost equivalent function miniflow_equal_flow_in_minimask() does not
> require the miniflow and minimask to have the same map, and it is slower for
> it. However, it is only used by the test-classifier.c, so one option to
> reduce the confusion would be to move it there. What do you think
Sounds good.
Ethan
On Tue, Apr 29, 2014 at 12:58 PM, Jarno Rajahalme wrote:
>
> On Apr 24, 2014, at 6:15 PM, Ethan Jackson wrote:
>
>> In the cache_push_back function, you might consider using the
>> x2nrealloc() function.
>>
>
> I did not know this existed, thanks!
>
>> Does this actually help
works for me here.
Acked-by: Pritesh Kothari
On Apr 29, 2014, at 1:22 PM, Andy Zhou wrote:
> Caught by clang-3.5.
>
> Reported-by: Simon Horman
> Signed-off-by: Andy Zhou
> ---
> ofproto/bond.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/ofproto/bond.c b/ofpro
On Tue, Apr 29, 2014 at 07:53:09AM -0700, Ben Pfaff wrote:
> Looking through OPENFLOW-1.1+, Open vSwitch is missing only one feature
> that is required for OpenFlow 1.1, 1.2, or 1.3, which is:
>
> * VLANs tagged with 88a8 Ethertype. This requires kernel work for
> reasonable performance
The Open vSwitch software switch now supports all the required features of
OpenFlow 1.0 through 1.3, with one known trivial exception[*]. Enable them
by default in ovs-vswitchd.
For now, ovs-ofctl only enables OpenFlow 1.0 by default. This is
because ovs-ofctl implements command such as "add-flo
Thanks, applied to master.
On Tue, Apr 29, 2014 at 01:49:15PM -0700, Justin Pettit wrote:
> Acked-by: Justin Pettit
>
>
> On April 29, 2014 at 1:48:23 PM, Ben Pfaff (b...@nicira.com) wrote:
> > "OpenFlow" is one word.
> > "Version" isn't a proper noun.
> >
> > Signed-off-by: Ben Pfaff
> > ---
"OpenFlow" is one word.
"Version" isn't a proper noun.
Signed-off-by: Ben Pfaff
---
lib/ofp-version-opt.c |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/ofp-version-opt.c b/lib/ofp-version-opt.c
index 10784fc..46fb45a 100644
--- a/lib/ofp-version-opt.c
+++ b/lib/o
Acked-by: Justin Pettit
On April 29, 2014 at 1:48:23 PM, Ben Pfaff (b...@nicira.com) wrote:
> "OpenFlow" is one word.
> "Version" isn't a proper noun.
>
> Signed-off-by: Ben Pfaff
> ---
> lib/ofp-version-opt.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/lib/ofp
On Apr 28, 2014, at 7:07 PM, Ethan Jackson wrote:
> The comment of miniflow_and_mask_matches_flow() need to be reworded.
>
> Do we really need that function? Is it really that much faster?
> Having two functions which do the exact same thing is a bit confusing.
The almost equivalent function
Caught by clang-3.5.
Reported-by: Simon Horman
Signed-off-by: Andy Zhou
---
ofproto/bond.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/ofproto/bond.c b/ofproto/bond.c
index f5a9d47..c1a8448 100644
--- a/ofproto/bond.c
+++ b/ofproto/bond.c
@@ -1180,10 +1180,11 @@ bond_r
I upgraded my clang from 3.4 to 3.5. Now I can also reproduce the
warning message.
On Tue, Apr 29, 2014 at 2:32 AM, Simon Horman wrote:
> Thanks.
>
> On Tue, Apr 29, 2014 at 02:30:26AM -0700, Andy Zhou wrote:
>> Good catch. Thanks for pointing it out. I will send out patch to fix this.
>>
>> On T
On Apr 28, 2014, at 3:55 PM, Ethan Jackson wrote:
> Acked-by: Ethan Jackson
>
> I don't really like the API we've ended up with here. But I think
> it's fine for now. As discussed offline, I intend to change it later.
>
You mean you don’t like what happens beneath the API, as it did not ch
On Apr 24, 2014, at 6:15 PM, Ethan Jackson wrote:
> In the cache_push_back function, you might consider using the
> x2nrealloc() function.
>
I did not know this existed, thanks!
> Does this actually help? I've found we spend most of our time getting
> the memory for the rule, not the subtabl
Along with flow-table rehashing OVS can compact masks array. This
allows us to calculate highest index for mask array.
Signed-off-by: Pravin B Shelar
---
datapath/flow_table.c | 40 ++--
datapath/flow_table.h |2 +-
2 files changed, 31 insertions(+), 11
On Apr 24, 2014, at 5:55 PM, Ethan Jackson wrote:
> I think there should be a comment explaining when to use prefetch vs
> prefetch_write. I certainly don't know off the top of my head.
>
I wrote this based on GCC documentation:
/* OVS_PREFETCH() can be used to instruct the CPU to fetch the
> The common use case is a flow expiring during the lifetime of a TCP
> connection. It will result in multiple data packets being sent upwards.
> It's much less likely in the megaflows era though.
I think this is pretty unlikely except in some extreme circumstances.
For a flow to be removed from t
Fine with me go ahead and push it.
Ethan
On Tue, Apr 29, 2014 at 12:40 PM, Jarno Rajahalme wrote:
>
> On Apr 24, 2014, at 5:40 PM, Ethan Jackson wrote:
>
>> So it seems to me that the only data needed in struct classifier by
>> callers is the fat_rwlock. What if we add a classifier_rdlock() an
On Apr 24, 2014, at 5:40 PM, Ethan Jackson wrote:
> So it seems to me that the only data needed in struct classifier by
> callers is the fat_rwlock. What if we add a classifier_rdlock() and a
> classifier_wrlock() function. That done, we could entirely hide
> struct classifier, and would need
On Tue, Apr 29, 2014 at 8:35 AM, Jarno Rajahalme wrote:
>
> On Apr 28, 2014, at 3:58 PM, Andy Zhou wrote:
>
>> Currently physical ports stats are collected from kernel datapath.
>> However, those counter do not reflect actual wire packet counters
>> when LSO or TSO are enabled by the NIC. In the
On Tue, Apr 29, 2014 at 7:33 AM, Ben Pfaff wrote:
> On Mon, Apr 28, 2014 at 03:58:51PM -0700, Andy Zhou wrote:
>> Currently physical ports stats are collected from kernel datapath.
>> However, those counter do not reflect actual wire packet counters
>> when LSO or TSO are enabled by the NIC. In th
On Apr 24, 2014, at 5:31 PM, Ethan Jackson wrote:
> The first line in the commit message needs a period.
>
Thanks.
> As Yamamoto asked, please verify that this actually helps. If it
> doesn't I'd prefer not to merge it.
>
This helps in the range of 1% in TCP_CRR performance test. However,
On Mon, Apr 28, 2014 at 5:13 PM, Simon Horman wrote:
> On Mon, Apr 28, 2014 at 02:37:47PM -0700, Jesse Gross wrote:
>> On Mon, Apr 28, 2014 at 12:00 AM, Simon Horman wrote:
>> > On Fri, Apr 25, 2014 at 12:57:20PM -0700, Jesse Gross wrote:
>> >> On Fri, Apr 25, 2014 at 1:06 AM, YAMAMOTO Takashi
>>
On Apr 24, 2014, at 5:15 PM, Ethan Jackson wrote:
> In the commit message s/suppoort/support/
>
> Out of curiosity why replace the miniflow_get inline functions with macros?
I got tired of writing (or rather copy&pasting) the “offset of(struct flow,
)”. The macro can take the field name as an
On Wed, Apr 23, 2014 at 3:38 PM, Thomas Graf wrote:
> On 04/23/2014 05:49 PM, Pravin B Shelar wrote:
>>
>> Along with flow-table rehashing OVS can compact masks array. This
>> allows us to calculate highest index for mask array.
>>
>> Signed-off-by: Pravin B Shelar
>
>
> Looks great in general,
Commit 6b59b543 (ovs-thread: Use fair (but nonrecursive)
rwlocks on glibc.) changed the rwlocks to nonrecursive,
writer-biased lock. It also made the following deadlock
possible.
Assume BFD is used on both end of a link. Consider the
following events:
1. Handler at one end received the BFD contr
On Apr 21, 2014, at 6:08 PM, YAMAMOTO Takashi wrote:
>>
>> On Apr 20, 2014, at 7:49 PM, YAMAMOTO Takashi wrote:
>>
+hash = mhash_finish(hash, 13); /* No need to match byte length
here. */
>>>
>>> is it worth being special?
>>>
>>
>> I知 not exactly sure what you ask here,
On Wed, Apr 23, 2014 at 3:26 PM, Thomas Graf wrote:
> On 04/23/2014 05:49 PM, Pravin B Shelar wrote:
>>
>> mask caches index of mask in mask_list. On packet recv OVS
>> need to traverse mask-list to get cached mask. Therefore array
>> is better for retrieving cached mask. This also allows bette
Thanks for the review. Pushed both patches.
On Mon, Apr 28, 2014 at 3:57 PM, Ben Pfaff wrote:
> On Thu, Apr 24, 2014 at 10:34:14PM -0700, Andy Zhou wrote:
>> When recirc pool is exhausted, a new bond won't be allocate a new
>> recirc_id. The bond->recirc_id will remain zero. This condition
>> sho
This commit adds a new configuration "stats-update-interval" in
"other_config" of Open_Vswitch table. So users can control the
statistics update frequency. A possible use case is that, users
can lower the update frequency to reduce the cpu consumption of
the ovs-vswitchd thread.
Signed-off-by: A
On Tue, Apr 29, 2014 at 7:33 AM, Ben Pfaff wrote:
> On Mon, Apr 28, 2014 at 03:58:51PM -0700, Andy Zhou wrote:
>> Currently physical ports stats are collected from kernel datapath.
>> However, those counter do not reflect actual wire packet counters
>> when LSO or TSO are enabled by the NIC. In th
On Tue, Apr 29, 2014 at 12:48:49PM -0400, Thomas F Herbert wrote:
> On 4/29/2014 10:53 AM, Ben Pfaff wrote:
> >Looking through OPENFLOW-1.1+, Open vSwitch is missing only one feature
> >that is required for OpenFlow 1.1, 1.2, or 1.3, which is:
> >
> > * VLANs tagged with 88a8 Ethertype. This r
On 4/29/2014 10:53 AM, Ben Pfaff wrote:
Looking through OPENFLOW-1.1+, Open vSwitch is missing only one feature
that is required for OpenFlow 1.1, 1.2, or 1.3, which is:
* VLANs tagged with 88a8 Ethertype. This requires kernel work for
reasonable performance.
[required for OF
On Tue, Apr 29, 2014 at 05:17:07PM +0100, Zoltan Kiss wrote:
> On 23/04/14 22:56, Thomas Graf wrote:
> >On 04/23/2014 10:12 PM, Ethan Jackson wrote:
> >>The problem has actually gotten worse since we've gotten rid of the
> >>dispatcher thread. Now each thread has it's own channel per port.
> >>
>
On April 29, 2014 at 7:53:21 AM, Ben Pfaff (b...@nicira.com) wrote:
> Looking through OPENFLOW-1.1+, Open vSwitch is missing only one feature
> that is required for OpenFlow 1.1, 1.2, or 1.3, which is:
>
> * VLANs tagged with 88a8 Ethertype. This requires kernel work for
> reasonable performance.
On 23/04/14 22:56, Thomas Graf wrote:
On 04/23/2014 10:12 PM, Ethan Jackson wrote:
The problem has actually gotten worse since we've gotten rid of the
dispatcher thread. Now each thread has it's own channel per port.
I wonder if the right approach is to simply ditch the per-port
fairness in th
On Apr 19, 2014, at 10:09 PM, Kmindg G wrote:
> On Sat, Apr 19, 2014 at 3:42 AM, Jarno Rajahalme
> wrote:
>>
>> /* This is useful for initializing a miniflow for a miniflow_extract() call.
>> */
>> static inline void miniflow_initialize(struct miniflow *mf,
>>
On Apr 28, 2014, at 3:58 PM, Andy Zhou wrote:
> Currently physical ports stats are collected from kernel datapath.
> However, those counter do not reflect actual wire packet counters
> when LSO or TSO are enabled by the NIC. In the meantime, the stats
> collected form routing stack does. While b
On Apr 29, 2014, at 7:40 AM, Martino Fornasa wrote:
> Jarno Rajahalme wrote:
>>
>> Also, we have defined the type ‘stp_identifier’ in lib/stp.h. I’d like you
>> to use it (with the accompanying formatting tools) also for RSTP (maybe move
>> them to lib/util.h to use them both from STP and RST
On Tue, Apr 22, 2014 at 09:32:01PM +0300, Alexandru Copot wrote:
> Also fix some tests that can now properly print packets
> with the new protocol version.
>
> Signed-off-by: Alexandru Copot
> Cc: Daniel Baluta
Looks fine, thanks, I'll wait for the new version of the series.
___
On Tue, Apr 22, 2014 at 09:32:00PM +0300, Alexandru Copot wrote:
> Signed-off-by: Alexandru Copot
> Cc: Daniel Baluta
I would fold these into patch 1.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Wed, Apr 23, 2014 at 02:03:56PM +0300, Alexandru Copot wrote:
> This is only the communication part of the bundles functionality.
> The actual message pre-validation and commits are not implemented.
>
> Signed-off-by: Alexandru Copot
Thanks for the new revision.
I think that ofputil_decode_b
applied and backported,
thx,
On Tue, Apr 29, 2014 at 7:39 AM, Ben Pfaff wrote:
> On Mon, Apr 28, 2014 at 11:53:52PM -0700, Alex Wang wrote:
> > Commit 05bf6d3c62e1d (ovs-thread: Add checking for mutex and
> > rwlock initialization.) helps find an use of uninitialized
> > mutex (netdev_class_mu
Looking through OPENFLOW-1.1+, Open vSwitch is missing only one feature
that is required for OpenFlow 1.1, 1.2, or 1.3, which is:
* VLANs tagged with 88a8 Ethertype. This requires kernel work for
reasonable performance.
[required for OF1.1+]
I'm proposing that we ignore this for
On Tue, Apr 29, 2014 at 06:34:22PM +0900, Simon Horman wrote:
> Add per-table counters. This resolves some short-comings
> in the data provided in a table stats reply message.
>
> * Lookups and matches are calculated based on table
> accesses rather than datapath flow hits and misses.
>
> * Loo
Jarno Rajahalme wrote:
Another high-level thing is that in OVS we use the ‘ovs_beXX’ types to annotate
integers in network byte order, and the corresponding ‘uintXX_t’ types for the
host byte order equivalents. You would use ovs_beXX types to define the BPDUs,
and then convert each member to h
On Mon, Apr 28, 2014 at 11:53:52PM -0700, Alex Wang wrote:
> Commit 05bf6d3c62e1d (ovs-thread: Add checking for mutex and
> rwlock initialization.) helps find an use of uninitialized
> mutex (netdev_class_mutex) during upgrade. The assertion
> check aborts the ovs.
>
> This commit fixes the issue
On Mon, Apr 28, 2014 at 11:53:51PM -0700, Alex Wang wrote:
> Based on the policy of 'OVS_VPORT_ATTR_UPCALL_PID', if upcall should
> not be sent to userspace (i.e. the number of handler threads is zero),
> the netlink message for creating vport should be an one-element array of
> value 0. However,
On Mon, Apr 28, 2014 at 10:39:27PM -0700, Justin Pettit wrote:
> This reverts commit 82e413df because bug fix bf52ed4 (datapath: Check for
> backported skb_orphan_frags().) was required.
>
> Signed-off-by: Justin Pettit
Acked-by: Ben Pfaff
___
dev mai
On Mon, Apr 28, 2014 at 03:58:51PM -0700, Andy Zhou wrote:
> Currently physical ports stats are collected from kernel datapath.
> However, those counter do not reflect actual wire packet counters
> when LSO or TSO are enabled by the NIC. In the meantime, the stats
> collected form routing stack doe
Add per-table counters. This resolves some short-comings
in the data provided in a table stats reply message.
* Lookups and matches are calculated based on table
accesses rather than datapath flow hits and misses.
* Lookups and matches are credited to the table where they
occurred rather than
Thanks.
On Tue, Apr 29, 2014 at 02:30:26AM -0700, Andy Zhou wrote:
> Good catch. Thanks for pointing it out. I will send out patch to fix this.
>
> On Tue, Apr 29, 2014 at 2:14 AM, Simon Horman wrote:
> > On Thu, Apr 24, 2014 at 10:34:15PM -0700, Andy Zhou wrote:
> >> Bond rebalancing was disabl
Good catch. Thanks for pointing it out. I will send out patch to fix this.
On Tue, Apr 29, 2014 at 2:14 AM, Simon Horman wrote:
> On Thu, Apr 24, 2014 at 10:34:15PM -0700, Andy Zhou wrote:
>> Bond rebalancing was disabled for bonds not using recirculation. The
>> patch fixes this bug.
>>
>> While
On Thu, Apr 24, 2014 at 10:34:15PM -0700, Andy Zhou wrote:
> Bond rebalancing was disabled for bonds not using recirculation. The
> patch fixes this bug.
>
> While fixing the bug, the bond_rebalance() was also restructured
> slightly to move bond related logic back into ofproto/bond.c
>
> Signed-
On Tue, Apr 29, 2014 at 12:40:18AM -0700, Andy Zhou wrote:
> If p7 and p37 are now connected together as patch ports, would any
> traffic injected from p7 bypass (in the user space) the bond
> interfaces all-together?
I believe what happens is that the normal action forwards
the packet to all port
If p7 and p37 are now connected together as patch ports, would any
traffic injected from p7 bypass (in the user space) the bond
interfaces all-together?
On Sun, Apr 27, 2014 at 6:24 PM, Simon Horman wrote:
> On Fri, Apr 25, 2014 at 12:46:16AM -0700, Andy Zhou wrote:
>> Simon, I thought this test
86 matches
Mail list logo
