Thanks for the fix! Sorry, I don't know too much about the intricacies of
NetBSD. My first version seemed to work on my NetBSD instance though, so I
wrongly assumed it would work for most versions of NetBSD.
I tested this on FreeBSD and it works, but I don't know the intricacies of
FreeBSD either.
This patch also checks the system platform as clock_gettime
could exist on different platforms but with different values of
CLOCK_MONOTONIC and different definitions of 'struct timespec'.
In this case, the system call would be expected to catch the
error, which is dangerous.
This patch ensures Lin
> This patch also checks the system platform as clock_gettime
> could exist on different platforms but with different values of
> CLOCK_MONOTONIC and different definitions of 'struct timespec'.
> In this case, the system call would be expected to catch the
> error, which is dangerous.
>
> This pat
Hey Yamamoto,
You're right, this is too fragile. This patch I just posted checks the
system's platform, so we don't run into situations like this (defaults to
using time.time() if its not Linux, NetBSD or FreeBSD). It also adds
CLOCK_MONOTONIC compatibility for NetBSD and FreeBSD, as well as updat
On Tue, Jun 03, 2014 at 01:46:11PM +0900, YAMAMOTO Takashi wrote:
> > diff --git a/net/core/dev.c b/net/core/dev.c
> > index 0355ca5..7c063ac 100644
> > --- a/net/core/dev.c
> > +++ b/net/core/dev.c
> > @@ -2498,13 +2498,42 @@ static int dev_gso_segment(struct sk_buff *skb,
> > netdev_features_t f
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 0355ca5..7c063ac 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -2498,13 +2498,42 @@ static int dev_gso_segment(struct sk_buff *skb,
> netdev_features_t features)
> return 0;
> }
>
> +/* If MPLS offload request, verify we a
This patch also checks the system platform as clock_gettime
could exist on different platforms but with different values of
CLOCK_MONOTONIC and different definitions of 'struct timespec'.
In this case, the system call would be expected to catch the
error, which is dangerous.
This patch ensures Lin
Hi Eric,
On Mon, Jun 02, 2014 at 08:42:02PM -0700, Eric Dumazet wrote:
> Hi Simon
>
> On Tue, 2014-06-03 at 11:38 +0900, Simon Horman wrote:
> > +/* If MPLS offload request, verify we are testing hardware MPLS features
> > + * instead of standard features for the netdev.
> > + */
> > +#ifdef CONF
> ovsrcu_set() and ovsrcu_init() look like functions, so users are
> justified in expecting the arguments to be evalueated before any of
evaluated
> the body of ovsrcu_set or ovsrcu_init().
>
> With ovs-atomic-pthreads, a fallback ovs-atomics implementation used
> when no C11 atomics are availab
WORM_MYDOOM.GEN has been detected,and Replace with text/file has been
taken on 6/3/2014 9:47:37 AM.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Hello,
Do you need a loan at 2%??? to start up a business or have your own
company or buy a car please visit our company today and visit our website
http://unionmoneyservice.mw.lt/ Thanks
Sign
Henry Reginal
unionmoneyservi...@gmail.com
http://unionmoneyservice.mw.lt/
Copyright@2014
___
Hi Jesse,
thanks for your feedback.
On Mon, Jun 02, 2014 at 05:58:10PM -0700, Jesse Gross wrote:
> On Sun, May 25, 2014 at 5:22 PM, Simon Horman wrote:
> > diff --git a/datapath/flow_netlink.c b/datapath/flow_netlink.c
> > index 803a94c..8ce596c 100644
> > --- a/datapath/flow_netlink.c
> > +++ b
> ovsrcu_init() was named after the atomic_init(), but the semantics are
> different enough to warrant a different name. Basically C11
> atomic_init is defined in a way that allows the implementation to
> assign the value without any syncronization, so in theory stores via
> atomic_init could be s
> 'struct ovsrcu_pointer' is always used with the 'struct' keyword, so
> remove the unneeded 'typedef'.
>
> Signed-off-by: Jarno Rajahalme
Acked-by: YAMAMOTO Takashi
> ---
> lib/ovs-rcu.h |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/ovs-rcu.h b/lib/ovs-rcu.
>
> On Jun 2, 2014, at 12:51 PM, Ben Pfaff wrote:
>
>> On Mon, Jun 02, 2014 at 12:50:45PM -0700, Jarno Rajahalme wrote:
>>>
>>> On Jun 2, 2014, at 11:45 AM, Ben Pfaff wrote:
>>>
I'd really hope that any serious OVS implementation would be able to
use C11 or GCC or Clang or other com
Hi Simon
On Tue, 2014-06-03 at 11:38 +0900, Simon Horman wrote:
> +/* If MPLS offload request, verify we are testing hardware MPLS features
> + * instead of standard features for the netdev.
> + */
> +#ifdef CONFIG_NET_MPLS_GSO
> +static netdev_features_t net_mpls_features(struct sk_buff *skb,
> +
> On Mon, Jun 02, 2014 at 11:28:24AM -0700, Ben Pfaff wrote:
>> This is partly documentation of how patches are tagged in practice in Open
>> vSwitch. The bits at the end about "Reported-at:" and "VMware-BZ:" are
>> an attempt to influence future practices; I cannot say how successful they
>> will
hi,
> Hey Yamamoto,
>
> Sorry for the delay, I finally set up my NetBSD environment and tested
> this script. You're right, but NetBSD will actually return an error on
> clock_gettime because of this argument difference. Thus, NetBSD will
> continue to use time.time(), which is as expected.
it r
On Mon, Jun 02, 2014 at 05:45:22PM -0700, Jesse Gross wrote:
> On Mon, Jun 2, 2014 at 5:16 PM, Simon Horman wrote:
> > On Mon, Jun 02, 2014 at 05:21:45PM +0100, Thomas Graf wrote:
> >> On 06/02/14 at 01:43pm, Simon Horman wrote:
> >> > +#ifdef CONFIG_NET_MPLS_GSO
> >> > +static netdev_features_t n
hi,
> Sorry for the delay, I swear I haven't been ignoring this email.
>
> Ok finally set up my NetBSD environment and master (with my rtbsd change)
> compiles and passes all unit tests.
thank you for taking care of netbsd port!
YAMAMOTO Takashi
>
> Ryan
>
> From: Ryan Wilson mailto:wr...@v
Hi Folks,
Using Open vSwitch I am setting up VXLAN tunnels between hosts. I intend to
make VXLAN tunnels "highly available" via Keepalived. Specifically, there
are one master node and one backup node for each host, and they share a
Virtual IP (VIP) via Keepalived, and the VXLAN tunnel will use the
On Sun, May 25, 2014 at 5:22 PM, Simon Horman wrote:
> diff --git a/datapath/flow_netlink.c b/datapath/flow_netlink.c
> index 803a94c..8ce596c 100644
> --- a/datapath/flow_netlink.c
> +++ b/datapath/flow_netlink.c
> + case OVS_ACTION_ATTR_POP_MPLS:
> + if (!eth_
On Mon, Jun 2, 2014 at 5:16 PM, Simon Horman wrote:
> On Mon, Jun 02, 2014 at 05:21:45PM +0100, Thomas Graf wrote:
>> On 06/02/14 at 01:43pm, Simon Horman wrote:
>> > +#ifdef CONFIG_NET_MPLS_GSO
>> > +static netdev_features_t net_mpls_features(struct sk_buff *skb,
>> > +
ovsrcu_init() was named after the atomic_init(), but the semantics are
different enough to warrant a different name. Basically C11
atomic_init is defined in a way that allows the implementation to
assign the value without any syncronization, so in theory stores via
atomic_init could be seen by oth
'struct ovsrcu_pointer' is always used with the 'struct' keyword, so
remove the unneeded 'typedef'.
Signed-off-by: Jarno Rajahalme
---
lib/ovs-rcu.h |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/ovs-rcu.h b/lib/ovs-rcu.h
index 775b829..c3e06ad 100644
--- a/lib/ovs-rc
ovsrcu_set() and ovsrcu_init() look like functions, so users are
justified in expecting the arguments to be evalueated before any of
the body of ovsrcu_set or ovsrcu_init().
With ovs-atomic-pthreads, a fallback ovs-atomics implementation used
when no C11 atomics are available or with GCC older tha
Thanks, applied.
I'll change the way I push to prevent this in future.
On 3 June 2014 12:26, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
>
> On Jun 2, 2014, at 5:23 PM, Joe Stringer wrote:
>
> > Commit acaa8dac490cc6c36026d45f3010ec75f38ee142 (revalidator: Eliminate
> > duplicate flow
Acked-by: Jarno Rajahalme
On Jun 2, 2014, at 5:23 PM, Joe Stringer wrote:
> Commit acaa8dac490cc6c36026d45f3010ec75f38ee142 (revalidator: Eliminate
> duplicate flow handling.) introduced a build error. This fixes the bug.
>
> Reported-by: Jarno Rajahalme
> Signed-off-by: Joe Stringer
> ---
>
Commit acaa8dac490cc6c36026d45f3010ec75f38ee142 (revalidator: Eliminate
duplicate flow handling.) introduced a build error. This fixes the bug.
Reported-by: Jarno Rajahalme
Signed-off-by: Joe Stringer
---
ofproto/ofproto-dpif-upcall.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
di
Sorry, I must have pushed the wrong version. I'll send a patch.
On 3 June 2014 12:19, Jarno Rajahalme wrote:
>
> On Jun 2, 2014, at 4:55 PM, Joe Stringer wrote:
>
> On 3 June 2014 08:41, Ethan Jackson wrote:
>
>> The comment over the trylock exceeds the line limit.
>>
>
> I'll fix that up.
>
On Jun 2, 2014, at 4:55 PM, Joe Stringer wrote:
> On 3 June 2014 08:41, Ethan Jackson wrote:
> The comment over the trylock exceeds the line limit.
>
> I'll fix that up.
>
> I'm not sure if this is the right patch to fix it or not, but I don't
> like how we're overloading the meaning of the
On Mon, Jun 02, 2014 at 05:21:45PM +0100, Thomas Graf wrote:
> On 06/02/14 at 01:43pm, Simon Horman wrote:
> > +#ifdef CONFIG_NET_MPLS_GSO
> > +static netdev_features_t net_mpls_features(struct sk_buff *skb,
> > + struct net_device *dev,
> > +
On 3 June 2014 08:41, Ethan Jackson wrote:
> The comment over the trylock exceeds the line limit.
>
I'll fix that up.
> I'm not sure if this is the right patch to fix it or not, but I don't
> like how we're overloading the meaning of the 'mark'. I think the
> code would be a lot clearer if we
On 31 May 2014 13:27, Ethan Jackson wrote:
> > I agree with the sentiment that it shouldn't be optional; that was my
> first
> > thought. Although, feedback from V1 was concerned about additional system
> > resources required to run it like that.
> >
> > Perhaps a compromise is to have it configu
When a bridge of datatype type netdev receives a packet, it copies the
packet from the NIC to a buffer in userspace. Currently, when making
an upcall, the packet is again copied to the upcall's buffer. However,
this extra copy is not necessary when the datapath exists in userspace
as the upcall can
On Jun 2, 2014, at 2:21 PM, Jarno Rajahalme wrote:
> Pushed to master with the following incremental. Backporting to branch-2.3 as
> well.
I backported the lib/match.c change and the test case to branch-2.1 as well.
Jarno
___
dev mailing list
dev
Thanks, Daniele!
Pushed to master with the following incremental. Backporting to branch-2.3 as
well.
Jarno
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index eb88f10..c14d671 100644
--- a/tests/ofproto-dpif.at
+++ b/tests/ofproto-dpif.at
@@ -4829,11 +4829,10 @@ AT_CAPTURE_FILE([
I'm wondering if it'd be cleaner to have ukey_create() return a locked
ukey, and to pull the try_lock() into the critical section for the
ukeys map? I.E.
lock_hmap()
ukey = ukey_lookup()
if (!ukey) {
//create and insert
} else {
// trylock
}
unlock_hmap()
What do you think?
On Thu, Ma
The comment over the trylock exceeds the line limit.
I'm not sure if this is the right patch to fix it or not, but I don't
like how we're overloading the meaning of the 'mark'. I think the
code would be a lot clearer if we had both a "already_dumped" flag and
a "delete" flag on every ukey. What
Sorry for the delay, I swear I haven't been ignoring this email.
Ok finally set up my NetBSD environment and master (with my rtbsd change)
compiles and passes all unit tests.
Ryan
From: Ryan Wilson mailto:wr...@vmware.com>>
Date: Thursday, May 29, 2014 10:47 PM
To: Ben Pfaff mailto:b...@nicira.
On Jun 2, 2014, at 12:51 PM, Ben Pfaff wrote:
> On Mon, Jun 02, 2014 at 12:50:45PM -0700, Jarno Rajahalme wrote:
>>
>> On Jun 2, 2014, at 11:45 AM, Ben Pfaff wrote:
>>
>>> I'd really hope that any serious OVS implementation would be able to
>>> use C11 or GCC or Clang or other compiler-specif
On Mon, Jun 02, 2014 at 12:50:45PM -0700, Jarno Rajahalme wrote:
>
> On Jun 2, 2014, at 11:45 AM, Ben Pfaff wrote:
>
> > I'd really hope that any serious OVS implementation would be able to
> > use C11 or GCC or Clang or other compiler-specific techniques to get a
> > "real" implementation of at
On Jun 2, 2014, at 11:45 AM, Ben Pfaff wrote:
> I'd really hope that any serious OVS implementation would be able to
> use C11 or GCC or Clang or other compiler-specific techniques to get a
> "real" implementation of atomics. Really the pthreads version is just
> to make porting easier.
>
Are
Hey Yamamoto,
Sorry for the delay, I finally set up my NetBSD environment and tested
this script. You're right, but NetBSD will actually return an error on
clock_gettime because of this argument difference. Thus, NetBSD will
continue to use time.time(), which is as expected.
I can add NetBSD func
On Mon, Jun 02, 2014 at 11:28:24AM -0700, Ben Pfaff wrote:
> This is partly documentation of how patches are tagged in practice in Open
> vSwitch. The bits at the end about "Reported-at:" and "VMware-BZ:" are
> an attempt to influence future practices; I cannot say how successful they
> will be.
>
I'd really hope that any serious OVS implementation would be able to
use C11 or GCC or Clang or other compiler-specific techniques to get a
"real" implementation of atomics. Really the pthreads version is just
to make porting easier.
On Mon, Jun 02, 2014 at 11:07:12AM -0700, Jarno Rajahalme wrote
This is partly documentation of how patches are tagged in practice in Open
vSwitch. The bits at the end about "Reported-at:" and "VMware-BZ:" are
an attempt to influence future practices; I cannot say how successful they
will be.
I am not sure whether these key-value pairs at the end of commit me
As I’m testing for this I notice that the cmap (or, ovs-rcu) with
ova-atomic-pthreads is painfully slow. The reason for this is the the generic
pthreads atomics have no memory barrier support, which is needed for efficient
RCU. As we make more use of ovs-rcu and cmap we need to address this in o
This patch addresses two bugs related to ICMPv6(NDP) packets:
- In miniflow_extract() push the words in the correct order
- In parse_icmpv6() use sizeof struct, not size of struct *
match_wc_init() has been modified, to include the nd_target field
when the transport layer protocol is ICMPv6
A te
On Mon, Jun 02, 2014 at 02:40:28PM -0300, Flavio Leitner wrote:
> On Mon, Jun 02, 2014 at 10:20:24AM -0700, Alex Wang wrote:
> > Quote from Ben,
> >
> > """
> > when VMware QA gives us a bug report, through the VMware bug tracker, we
> > include the number in the change log entry, and that others
Thanks for reporting this!
I think we should allow the illusion that ovsrcu_init is a function, and it’s
parameters are evaluated before the body. Otherwise problems like this are
bound to be reintroduced. I’ll post a patch soon.
Jarno
On Jun 2, 2014, at 3:34 AM, YAMAMOTO Takashi wrote:
>
On Mon, Jun 02, 2014 at 10:20:24AM -0700, Alex Wang wrote:
> Hey Flavio,
Hi Alex,
> Quote from Ben,
>
> """
> when VMware QA gives us a bug report, through the VMware bug tracker, we
> include the number in the change log entry, and that others are welcome to
> include their internal tracking ID
On Mon, Jun 02, 2014 at 09:36:28PM +0900, Simon Horman wrote:
> Add OFPERR_OFPBIC_DUP_INST (type = OFPET_BAD_INSTRUCTION, code = 9)
> and use it for OpenFlow1.4+.
>
> For OpenFlow1.1 - 1.3 map this error to ONFBIC_DUP_INSTRUCTION
> (experimenter = ONF, type = 2600) which is proposed in
> OpenFlow
Thanks for the review, Flavio~
> Since those are all a sequence of mutex_lock(), get some value,
> mutex_unlock, it might be a good optimization to have a function
> that get all the status at once, e.g.:
>
> cfm_get_status(cfm, &status)
> {
> mutex_lock()
> get all status
>
Hey Flavio,
Quote from Ben,
"""
when VMware QA gives us a bug report, through the VMware bug tracker, we
include the number in the change log entry, and that others are welcome to
include their internal tracking IDs when they submit changes.
"""
Thanks,
Alex Wang,
On Mon, Jun 2, 2014 at 7:32 A
On 06/02/14 at 01:43pm, Simon Horman wrote:
> +#ifdef CONFIG_NET_MPLS_GSO
> +static netdev_features_t net_mpls_features(struct sk_buff *skb,
> +struct net_device *dev,
> +netdev_features_t features)
> +{
> + /* Ther
Hi,
Would you be interested in acquiring 2014 Updated Contact Databases of below
specified titles that will add great value to your marketing programs?
We can segment Industry, Technology users & Healthcare database by C Level,
VP Level, Director Level and Manager Level as per your requirem
On Sun, Jun 01, 2014 at 10:04:30PM -0700, Alex Wang wrote:
> This commit adds logic that marks the 'struct iface's when their
> status need to be updated but the current status update transaction
> fails. So, the main thread can re-try the update in the next run.
>
> Bug #1256577
Where is that b
On Sun, Jun 01, 2014 at 10:04:28PM -0700, Alex Wang wrote:
> Currently, ofproto_port_get_bfd/cfm_status() is used to check the
> bfd/cfm status change and query the status change. Users decide
> what to do with the filled status struct based on the return value
> of the funciton. Such design is c
Add OFPERR_OFPBIC_DUP_INST (type = OFPET_BAD_INSTRUCTION, code = 9)
and use it for OpenFlow1.4+.
For OpenFlow1.1 - 1.3 map this error to ONFBIC_DUP_INSTRUCTION
(experimenter = ONF, type = 2600) which is proposed in
OpenFlow enhancement proposal EXT-260 "Add error
code for duplicate instruction.".
In the case of ovs-atomic-pthreads, the previous coding is expanded
into successive atomic_lock__ calls which ends up with deadlock.
Signed-off-by: YAMAMOTO Takashi
---
lib/cmap.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/lib/cmap.c b/lib/cmap.c
index a760235..7892d5
> On Thu, May 29, 2014 at 1:21 AM, YAMAMOTO Takashi
> wrote:
what would the new definition of DP_HASH action look like?
>>> flow_hash_5tuple
>>
>> DP_HASH was defined as "datapath specific hash of the packet".
>> my question is what it will be after your proposed change.
>
> It will be
62 matches
Mail list logo
