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 Ryan for quick turn around. It is good to take care those
little things while it is fresh in our minds.
I will push it in a minute.
On Wed, May 21, 2014 at 9:56 PM, Ryan Wilson 76511 wrote:
> Thanks for this, Andy!
>
> Acked-by: Ryan Wilson
>
>
Thanks for this, Andy!
Acked-by: Ryan Wilson
From: dev on behalf of Andy Zhou
Sent: Wednesday, May 21, 2014 9:47 PM
To: d...@openvswitch.com
Subject: [ovs-dev] [PATCH] ofproto: Add const to immutable members of ofgroup.
Signed-off-by: Andy Zhou
---
Signed-off-by: Andy Zhou
---
ofproto/ofproto-provider.h | 12 ++--
ofproto/ofproto.c | 21 -
2 files changed, 18 insertions(+), 15 deletions(-)
diff --git a/ofproto/ofproto-provider.h b/ofproto/ofproto-provider.h
index e5f71c7..9f87ba9 100644
--- a/ofproto/of
Thanks for the review. Ordinarily I would spend some time refining the
commit message to better explain, because I like commit messages to be
clear to everyone, but I am feeling a lot of pressure to fix this bug,
so I pushed it as is.
On Thu, May 22, 2014 at 02:24:37PM +1200, Joe Stringer wrote:
On Wed, May 21, 2014 at 9:00 PM, Ben Pfaff wrote:
> On Wed, May 21, 2014 at 07:06:05PM -0700, Jesse Gross wrote:
>> Reported-by: Ben Pfaff
>> Signed-off-by: Jesse Gross
>
> Thanks Jesse, I think that this will help. (At the very least it's easy
> to find and post.)
Thanks, I applied this.
On Wed, May 21, 2014 at 08:11:09PM -0700, Ryan Wilson wrote:
> Signed-off-by: Ryan Wilson
Applied, thanks!
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Wed, May 21, 2014 at 07:06:05PM -0700, Jesse Gross wrote:
> Reported-by: Ben Pfaff
> Signed-off-by: Jesse Gross
Thanks Jesse, I think that this will help. (At the very least it's easy
to find and post.)
___
dev mailing list
dev@openvswitch.org
http
Thanks, applied and backported,~
On Wed, May 21, 2014 at 7:08 PM, Joe Stringer wrote:
> Acked-by: Joe Stringer
>
>
> On 22 May 2014 12:27, Alex Wang wrote:
>
>> Commit b256dc525c8 (ofproto-dpif-xlate: Cache xlate_actions() effects.)
>> caches the variables needed for refreshing mac-learning ta
On Wed, May 21, 2014 at 7:47 PM, Ryan Wilson wrote:
> Signed-off-by: Ryan Wilson
> Acked-by: Andy Zhou
>
> ---
> v2: Fixed bug with group stats all buckets, cleaned up ofgroup unref code,
> added Andy Zhou's test for more complete test coverage
> v3: Split group ref/unref, support for group
Signed-off-by: Ryan Wilson
---
TODO | 17 -
1 file changed, 17 deletions(-)
diff --git a/TODO b/TODO
index dabe49c..e11089a 100644
--- a/TODO
+++ b/TODO
@@ -52,23 +52,6 @@ These changes might have backward-compatibility
implications; one
would have to test the behavior of the
Signed-off-by: Ryan Wilson
Acked-by: Andy Zhou
---
v2: Fixed bug with group stats all buckets, cleaned up ofgroup unref code,
added Andy Zhou's test for more complete test coverage
v3: Split group ref/unref, support for group and bucket stats, and Andy Zhou's
tests into 3 patches.
v4: Fix re
Would you please make a separate patch for the todo list change, and
resend this one without it? Thanks.
On Wed, May 21, 2014 at 7:29 PM, Ryan Wilson wrote:
> Signed-off-by: Ryan Wilson
> Acked-by: Andy Zhou
>
> ---
> v2: Fixed bug with group stats all buckets, cleaned up ofgroup unref code,
>
Signed-off-by: Ryan Wilson
Acked-by: Andy Zhou
---
v2: Fixed bug with group stats all buckets, cleaned up ofgroup unref code,
added Andy Zhou's test for more complete test coverage
v3: Split group ref/unref, support for group and bucket stats, and Andy Zhou's
tests into 3 patches.
v4: Fix re
The commit message is a bit confusing to me, I don't see the relationship
to commit 1340ce0c175. I see it more as a discrepancy between the behaviour
of dpif_linux_flow_dump_next_may_destroy_keys() and
dpif_linux_flow_dump_next(). However I agree with the rest of the commit
message.
Tested briefly
On Wed, May 21, 2014 at 5:38 AM, wrote:
> Hi Jesse,
>
> 2014-05-19, Jesse Gross:
> [...snip...]
I actually have much less of a problem with including an EtherType
from a GRE tunnel. For one thing it actually exists in the packet
rather than being an artificial namespace. Another th
Yup, modify_group looks much cleaner! And sorry for the lock annotations;
I typically use GCC but its clear I need to use clang as well. You're
welcome to push with your incremental, thanks!
I'll also send out another version of my group and bucket stats patch,
there are some rebase errors with th
Acked-by: Joe Stringer
On 22 May 2014 12:27, Alex Wang wrote:
> Commit b256dc525c8 (ofproto-dpif-xlate: Cache xlate_actions() effects.)
> caches the variables needed for refreshing mac-learning table in
> xlate_normal(). Wherein, the cache entry always records reference to
> the original 'ofp
On Tue, May 20, 2014 at 8:26 AM, Ben Pfaff wrote:
> On Tue, May 20, 2014 at 08:21:10AM -0700, Jesse Gross wrote:
>> On Tue, May 20, 2014 at 4:26 AM, Ajay Singh (HM0001261)
>> wrote:
>> > Hi,
>> >
>> >
>> >
>> > I have compiled and installed openvswitch-2.1.2 source code on Fedora
>> > machine ver
Reported-by: Ben Pfaff
Signed-off-by: Jesse Gross
---
FAQ | 21 +
1 file changed, 21 insertions(+)
diff --git a/FAQ b/FAQ
index cac9fcb..79d905d 100644
--- a/FAQ
+++ b/FAQ
@@ -190,6 +190,27 @@ A: The kernel module in upstream Linux does not include
support for
Linux rel
On Wed, May 21, 2014 at 5:12 PM, Ryan Wilson wrote:
> When adding support for OpenFlow group and bucket stats, a group entry is
> added
> to the xlate_cache. If the main thread removes the group from an ofproto, we
> need to guarantee that the group remains accessible to users of
> xlate_group_ac
On Tue, May 20, 2014 at 9:47 PM, Ajay Singh (HM0001261)
wrote:
> Hi Jesse,
>
> Can you suggest the correct version of Fedora on which I can install
> openvswitch and make use of VXLAN.
> It is a blocker for me. Please give me some pointer.
I see, you are using the upstream module. You can use th
On Wed, May 21, 2014 at 5:01 PM, Ryan Wilson 76511 wrote:
> Some inline comments below. The only thing I'm a bit concerned about it is
> the constant cast issue, especially since created / modified are actually
> altered in modify_group().
>
> Also, the problem is the struct-defined variables are
Hey Joe,
Could you review this patch, when you are available?
Thanks,
Alex Wang,
On Wed, May 21, 2014 at 5:27 PM, Alex Wang wrote:
> Commit b256dc525c8 (ofproto-dpif-xlate: Cache xlate_actions() effects.)
> caches the variables needed for refreshing mac-learning table in
> xlate_normal(). Wh
Commit b256dc525c8 (ofproto-dpif-xlate: Cache xlate_actions() effects.)
caches the variables needed for refreshing mac-learning table in
xlate_normal(). Wherein, the cache entry always records reference to
the original 'ofproto'.
When patch port is used to connect two 'ofproto's, packet goes thro
When adding support for OpenFlow group and bucket stats, a group entry is added
to the xlate_cache. If the main thread removes the group from an ofproto, we
need to guarantee that the group remains accessible to users of
xlate_group_actions and the xlate_cache, as the xlate_cache will not be cleane
On Wed, May 21, 2014 at 01:41:14PM -0700, Ben Pfaff wrote:
> On Wed, May 21, 2014 at 01:37:38PM -0700, Jarno Rajahalme wrote:
> >
> > On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
> >
> > > Found it, I think. I'll send a series that includes fixes and then
> > > the dpif-netdev conversion.
> >
This will only catch core dumps in tests that start and stop ovs-vswitchd,
but that covers the ones most likely to have problems.
Suggested-by: Jarno Rajahalme
Signed-off-by: Ben Pfaff
---
tests/ofproto-macros.at | 20 +++-
1 file changed, 19 insertions(+), 1 deletion(-)
diff
Some inline comments below. The only thing I'm a bit concerned about it is
the constant cast issue, especially since created / modified are actually
altered in modify_group().
Also, the problem is the struct-defined variables are constant, so
CONST_CAST won't work here. (CONST_CAST works for conve
On Wednesday, May 21, 2014, Simon Horman wrote:
> On Wed, May 21, 2014 at 01:31:55PM -0700, Jesse Gross wrote:
> > On Wed, May 21, 2014 at 8:31 AM, Simon Horman
> > >
> wrote:
> > > On Tue, May 20, 2014 at 07:30:38PM -0700, Jesse Gross wrote:
> > >> On Thu, May 15, 2014 at 4:07 PM, Simon Horman
Thanks for the acks. I pushed these to master. I'll work on
backports for the bug fixes now.
On Wed, May 21, 2014 at 12:47:04PM -0700, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
>
> On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
>
> > Signed-off-by: Ben Pfaff
> > Acked-by: Jarno Raja
On Wed, May 21, 2014 at 12:39:11PM -0700, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
Thanks. Applied to master.
> I still worry about reordering of the non-atomic reads in between
> the read_counter() calls, as in principle the acquire barrier does
> not prevent loads before such a barr
On Wed, May 21, 2014 at 03:52:39PM -0700, Ben Pfaff wrote:
> On Thu, May 22, 2014 at 10:25:04AM +1200, Joe Stringer wrote:
> > So just to make sure I'm understanding correctly,
> >
> > 1) revalidator calls dpif_flow_dump_next_may_destroy_keys(), which
> > indicates that it will not destroy the buf
Commit 1340ce0c175 (ofproto-dpif-upcall: Avoid use-after-free in
revalidate() corner cases.) fixed one use-after-free error in revalidate(),
but missed one more subtle case, in which dpif_linux_flow_dump_next()
attempts to retrieve actions for a flow that didn't have them in the main
dump result.
On Wed, May 21, 2014 at 01:31:55PM -0700, Jesse Gross wrote:
> On Wed, May 21, 2014 at 8:31 AM, Simon Horman wrote:
> > On Tue, May 20, 2014 at 07:30:38PM -0700, Jesse Gross wrote:
> >> On Thu, May 15, 2014 at 4:07 PM, Simon Horman wrote:
> >> > diff --git a/datapath/linux/compat/include/linux/ne
On Thu, May 22, 2014 at 10:25:04AM +1200, Joe Stringer wrote:
> So just to make sure I'm understanding correctly,
>
> 1) revalidator calls dpif_flow_dump_next_may_destroy_keys(), which
> indicates that it will not destroy the buffer.
> 2) revalidator calls dpif_flow_dump_next(), which dumps a flow
On Wed, May 21, 2014 at 2:14 PM, Ryan Wilson wrote:
> When adding support for OpenFlow group and bucket stats, a group entry is
> added
> to the xlate_cache. If the main thread removes the group from an ofproto, we
> need to guarantee that the group remains accessible to users of
> xlate_group_ac
So just to make sure I'm understanding correctly,
1) revalidator calls dpif_flow_dump_next_may_destroy_keys(), which
indicates that it will not destroy the buffer.
2) revalidator calls dpif_flow_dump_next(), which dumps a flow without
actions. The most likely case is that this is the last flow in
I forgot to mention that this is for branch-2.3, since this is already
fixed on master.
On Wed, May 21, 2014 at 02:33:51PM -0700, Ben Pfaff wrote:
> Commit 1340ce0c175 (ofproto-dpif-upcall: Avoid use-after-free in
> revalidate() corner cases.) fixed one use-after-free error in revalidate(),
> but
Commit 1340ce0c175 (ofproto-dpif-upcall: Avoid use-after-free in
revalidate() corner cases.) fixed one use-after-free error in revalidate(),
but missed one more subtle case, in which dpif_linux_flow_dump_next()
attempts to retrieve actions for a flow that didn't have them in the main
dump result.
I agree, the NULL set line is wrong. I can remove that in a future version.
Ethan, would you mind taking a look at this patch?
On 21 May 2014 17:37, Alex Wang wrote:
> Based on the discussion offline, we should not set the
>
> "+ukey->xcache = NULL;"
>
> And this change deserves fu
Thanks, I applied this to master and branch-2.3.
On 21 May 2014 18:14, Alex Wang wrote:
> Tested, and no more flow_del warning,
>
>
> Acked-by: Alex Wang
>
>
> On Tue, May 20, 2014 at 8:45 PM, Joe Stringer wrote:
>
>> Commit 73a3c4757e59 (revalidator: Prevent handling the same flow twice.)
>>
On 22 May 2014 06:12, Ben Pfaff wrote:
> On Tue, May 20, 2014 at 12:49:09PM +1200, Joe Stringer wrote:
> > This was previously logging that there was no such function, but passing
> > the test anyway.
> >
> > Signed-off-by: Joe Stringer
>
> Good catch.
>
> Acked-by: Ben Pfaff
>
Thanks, I appli
When adding support for OpenFlow group and bucket stats, a group entry is added
to the xlate_cache. If the main thread removes the group from an ofproto, we
need to guarantee that the group remains accessible to users of
xlate_group_actions and the xlate_cache, as the xlate_cache will not be cleane
When adding support for OpenFlow group and bucket stats, a group entry is added
to the xlate_cache. If the main thread removes the group from an ofproto, we
need to guarantee that the group remains accessible to users of
xlate_group_actions and the xlate_cache, as the xlate_cache will not be cleane
On Wed, May 21, 2014 at 01:37:38PM -0700, Jarno Rajahalme wrote:
>
> On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
>
> > Found it, I think. I'll send a series that includes fixes and then
> > the dpif-netdev conversion.
>
> Btw. it would be nice if the test suite Test results section would
> l
On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
> Found it, I think. I'll send a series that includes fixes and then
> the dpif-netdev conversion.
Btw. it would be nice if the test suite Test results section would list any
core files found on the tests/testsuite.dir// directories, so they w
On Wed, May 21, 2014 at 8:31 AM, Simon Horman wrote:
> On Tue, May 20, 2014 at 07:30:38PM -0700, Jesse Gross wrote:
>> On Thu, May 15, 2014 at 4:07 PM, Simon Horman wrote:
>> > diff --git a/datapath/linux/compat/include/linux/netdevice.h
>> > b/datapath/linux/compat/include/linux/netdevice.h
>>
LGTM,
Jarno
Acked-by: Jarno Rajahalme
> On Apr 30, 2014, at 2:35 PM, Ben Pfaff wrote:
>
> The poll_loop code has a feature that, when turned on manually or
> automatically (due to high CPU use), logs the source file and line number
> of the code that caused a thread to wake up from poll().
On Wed, May 21, 2014 at 8:15 AM, Anoob Soman wrote:
> Hi,
>
> I encountered softlockup (in stats_read@datapath/flow.c), while playing with
> OVS-2.1.2. I used change set 4f647e0a3c37b8d5086214128614a136064110c3 from
> linux-stable to fix the problem. I was wondering, if this changeset can be
> ap
Acked-by: Jarno Rajahalme
On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> Acked-by: Jarno Rajahalme
> ---
> lib/dpif-netdev.c | 171 +++--
> 1 file changed, 87 insertions(+), 84 deletions(-)
>
> diff --git a/lib/dpif-
Acked-by: Jarno Rajahalme
On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> lib/dpif-netdev.c |1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
> index a2e3308..a8e6a55 100644
> --- a/lib/dpif-netdev.c
> +++ b/lib/d
Acked-by: Jarno Rajahalme
On May 20, 2014, at 5:11 PM, Ben Pfaff wrote:
> When the last rxq is closed (which releases the rxq's internal reference
> to its netdev) the next call to netdev_n_rxq() accesses freed memory.
>
> Found by valgrind.
>
> Signed-off-by: Ben Pfaff
> Reported-by: Jarno
Acked-by: Jarno Rajahalme
I still worry about reordering of the non-atomic reads in between the
read_counter() calls, as in principle the acquire barrier does not prevent
loads before such a barrier to be moved after it. Thus, this could happen:
for (;;) {
c = read_even_counter()
check
>>>How about we remove the rwlock and only use ref count?
>>
>> We need the ref count to determine when we can free the group / bucket.
>> The rwlock is to serialize access to the ofgroup's properties. For
>> example, 2 threads could be accessing ofgroup's properties, but a ref
>> count doesn't p
>>How about we remove the rwlock and only use ref count?
>
> We need the ref count to determine when we can free the group / bucket.
> The rwlock is to serialize access to the ofgroup's properties. For
> example, 2 threads could be accessing ofgroup's properties, but a ref
> count doesn't prevent t
On Wed, May 21, 2014 at 08:28:28AM -0700, Gurucharan Shetty wrote:
> On Wed, Apr 30, 2014 at 10:58 AM, Ben Pfaff wrote:
> > Signed-off-by: Ben Pfaff
> Acked-by: Gurucharan Shetty
Thanks, applied.
___
dev mailing list
dev@openvswitch.org
http://openvsw
On Tue, May 20, 2014 at 12:49:09PM +1200, Joe Stringer wrote:
> This was previously logging that there was no such function, but passing
> the test anyway.
>
> Signed-off-by: Joe Stringer
Good catch.
Acked-by: Ben Pfaff
___
dev mailing list
dev@openv
On Tue, May 20, 2014 at 07:30:38PM -0700, Jesse Gross wrote:
> On Thu, May 15, 2014 at 4:07 PM, Simon Horman wrote:
> > diff --git a/datapath/linux/compat/gso.c b/datapath/linux/compat/gso.c
> > index 9ded17c..d9b4485 100644
> > --- a/datapath/linux/compat/gso.c
> > +++ b/datapath/linux/compat/gso
On Wed, Apr 30, 2014 at 10:58 AM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Acked-by: Gurucharan Shetty
> ---
> utilities/ovs-vsctl.8.in |4
> 1 file changed, 4 insertions(+)
>
> diff --git a/utilities/ovs-vsctl.8.in b/utilities/ovs-vsctl.8.in
> index 43b00bf..1701b48 100644
> --- a/
On Tue, May 20, 2014 at 07:05:28PM -0700, Jesse Gross wrote:
> On Tue, May 20, 2014 at 3:48 AM, Simon Horman wrote:
> > On Mon, May 19, 2014 at 06:34:05PM -0700, Jesse Gross wrote:
> >> On Thu, May 15, 2014 at 4:07 PM, Simon Horman wrote:
> >> > diff --git a/datapath/actions.c b/datapath/actions.
Hi,
I encountered softlockup (in stats_read@datapath/flow.c), while playing
with OVS-2.1.2. I used change set
4f647e0a3c37b8d5086214128614a136064110c3 from linux-stable to fix the
problem. I was wondering, if this changeset can be applied to
ovs/branch-2.1. Other branches doesn't need this f
On Wed, Apr 30, 2014 at 02:35:13PM -0700, Ben Pfaff wrote:
> The poll_loop code has a feature that, when turned on manually or
> automatically (due to high CPU use), logs the source file and line number
> of the code that caused a thread to wake up from poll(). Until now, when
> a function calls s
On Wed, Apr 30, 2014 at 10:58:59AM -0700, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
This still needs a review.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Fri, Apr 25, 2014 at 10:24:25AM -0700, Ben Pfaff wrote:
> The test added in this commit would have caught the bug fixed by commit
> 96be8de595150 (bridge: When ports disappear from a datapath, add them
> back.). With that commit reverted, the new test fails.
>
> Signed-off-by: Ben Pfaff
This
On Tue, Mar 25, 2014 at 12:32:03PM +, Anoob Soman wrote:
> fake-iface, option to add-bond, was added for compatibility
> reasons and need not be used otherwise.
>
> Signed-off-by: Anoob Soman
Applied, thanks.
___
dev mailing list
dev@openvswitch.or
Release memory ordering only affects visibility of stores, and is not
allowed on a memory read. Some compilers enforce this, making this code
fail to compile.
Reported-by: Alex Wang
Reported-by: Kmindg G
CC: Jarno Rajahalme
Signed-off-by: Ben Pfaff
---
lib/cmap.c | 8
1 file changed
On Wed, May 21, 2014 at 02:20:08PM +0800, Kmindg G wrote:
> When I compile ovs src on the master branch with gcc 4.9, I get this error:
Thanks. I've received the same reports from one other, too. I'll send
out a patch.
___
dev mailing list
dev@openvswi
Gentile utente
La tua email è superato 2 GB creati dal webmaster, è attualmente in esecuzione
a 2.30GB, non que possibile inviare o ricevere nuovi messaggi Entro il prossimo
24 ore fino a che non Verificare email account.
Inserisci i tuoi dati qui sotto per verificare l'account:
(1) E-mail:
Hi Jesse,
2014-05-19, Jesse Gross:
[...snip...]
>>> I actually have much less of a problem with including an EtherType
>>> from a GRE tunnel. For one thing it actually exists in the packet
>>> rather than being an artificial namespace. Another thing is that it is
>>> metadata akin to the input por
I have published an android Email Scanner app for email hunter, completely free here: Amazon store: Amazon.com: Email Scanner: Appstore for AndroidSamsung store: Email Scanner - Business - Samsung Apps - Samsung Content & ServicesPlease rock!
___
dev mai
71 matches
Mail list logo
