On Tue, Sep 17, 2013 at 1:40 AM, David Laight wrote:
>
> > Patch 1/2: generic BPF extension
> > Original A and X 32-bit BPF registers are replaced with ten 64-bit
> > registers.
> > bpf opcode encoding kept the same. load/store were generalized to access
> > stack,
> > bpf_tables and bpf_context
On Tue, Sep 17, 2013 at 1:20 PM, Pravin Shelar wrote:
> On Fri, Sep 13, 2013 at 2:16 PM, Jesse Gross wrote:
>> The symbol HAVE_NET_DEVICE_OPS was changed in 3.1 but code the that
>> it was protecting dates back to before 2.6.32. Therefore, we can
>> just assume that net_device_ops exists in all c
On Tue, Sep 17, 2013 at 1:15 PM, Pravin Shelar wrote:
> On Fri, Sep 13, 2013 at 2:37 PM, Jesse Gross wrote:
>> The symbol HAVE_NET_DEVICE_OPS was removed in 3.1 even though these ops
>> still exist, so on newer kernels we define the symbol ourselves. However,
>> SLES 11 SP3 appears to have backpo
On Sep 17, 2013, at 4:08 PM, Ben Pfaff wrote:
> When a flow is slow-pathed, every packet needs to translated separately
> and treated accordingly. The code here (and the code that preceded it)
> did translate every packet separately but ignored the actions, instead
> using the actions from the
I like this! Much clearer, especially when comparing to the upcall processing
in the past monolithic ofproto-dpif.
One small comment about a comment below,
Jarno
Acked-by: Jarno Rajahalme
On Sep 17, 2013, at 4:08 PM, Ben Pfaff wrote:
>
> -/* Process each element in the to-do list, co
On Tue, Sep 17, 2013 at 06:47:22PM -0500, Simon Horman wrote:
> On Mon, Sep 16, 2013 at 01:46:19PM -0700, Ben Pfaff wrote:
> > On Mon, Sep 16, 2013 at 03:38:21PM -0500, Jesse Gross wrote:
> > > - Ben, do you want to take over for the userspace portions of the
> > > series on the assumption that th
On Mon, Sep 16, 2013 at 01:46:19PM -0700, Ben Pfaff wrote:
> On Mon, Sep 16, 2013 at 03:38:21PM -0500, Jesse Gross wrote:
> > - Ben, do you want to take over for the userspace portions of the
> > series on the assumption that the above comments can be fixed fairly
> > easily?
>
> Yes, that's fine
On Tue, Aug 27, 2013 at 10:25:55AM -0700, Ben Pfaff wrote:
> These seem slightly nicer than e.g. htons(UINT16_MAX).
>
> Signed-off-by: Ben Pfaff
I applied this patch to master.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/lis
v1->v2:
- First 5 patches applied and dropped.
- Bug fix in what is now patch 1.
Ben Pfaff (2):
ofproto-dpif-upcall: Forward packets in order of arrival.
ofproto-dpif-upcall: Use correct actions for slow-pathed flows.
ofproto/ofproto-dpif-upcall.c | 361 ++---
When a flow is slow-pathed, every packet needs to translated separately
and treated accordingly. The code here (and the code that preceded it)
did translate every packet separately but ignored the actions, instead
using the actions from the initial translation that determined that the
flow needed
Until now, the code in ofproto-dpif-upcall (and the code that preceded it
in ofproto-dpif) obtained a batch of incoming packets, inserted them into
a hash table based on hashes of their flows, processed them, and then
forwarded them in hash order. Usually this maintains order within a single
netwo
Thanks, applied.
On Tue, Sep 17, 2013 at 02:44:28PM -0700, Justin Pettit wrote:
> Looks good to me.
>
> Acked-by: Justin Pettit
>
> --Justin
>
>
> On Sep 17, 2013, at 2:36 PM, Ben Pfaff wrote:
>
> > I backported millisecond timestamps to branch-2.0 upon request from Paul
> > Ingram, so we s
Found by valgrind.
Signed-off-by: Ben Pfaff
---
ofproto/ofproto.c |1 +
1 file changed, 1 insertion(+)
diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c
index 0ceb85f..bd85104 100644
--- a/ofproto/ofproto.c
+++ b/ofproto/ofproto.c
@@ -2487,6 +2487,7 @@ rule_actions_unref(struct rule_action
On Tue, Sep 17, 2013 at 11:56:57AM -0700, Ben Pfaff wrote:
> On Tue, Sep 17, 2013 at 10:22:04AM -0700, Jarno Rajahalme wrote:
> > I get test failures after this patch, seems to relate to slow path
> > processing. Haven't looked at them, but I see failures on 656 and 657
> > come and go, so those ar
Looks good to me.
Acked-by: Justin Pettit
--Justin
On Sep 17, 2013, at 2:36 PM, Ben Pfaff wrote:
> I backported millisecond timestamps to branch-2.0 upon request from Paul
> Ingram, so we should mention it as part of that version.
>
> Signed-off-by: Ben Pfaff
> ---
> NEWS |6 +++---
> 1
I backported millisecond timestamps to branch-2.0 upon request from Paul
Ingram, so we should mention it as part of that version.
Signed-off-by: Ben Pfaff
---
NEWS |6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/NEWS b/NEWS
index c46c3cc..eae1146 100644
--- a/NEWS
++
Thanks for the reviews. I applied patches 1 to 5 to master and
backported patch 1 to branch-2.0.
I'll figure out what's up with patch 6 now.
On Tue, Sep 17, 2013 at 12:25:50PM -0700, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme
>
> On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
>
> > Si
On Mon, Sep 16, 2013 at 03:15:01PM -0700, Ben Pfaff wrote:
> Commit 2b31d8e713de7 (vlog: Report timestamps in millisecond resolution in
> log messages.) introduced milliseconds to log messages by default, but the
> Python version did not ensure that milliseconds were always formatted with
> 3 digit
Acked-by: Jarno Rajahalme
On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> ofproto/ofproto-dpif.c |3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
> index 19f69e2..01299e1 100644
> --- a/ofproto/ofprot
Acked-by: Jarno Rajahalme
On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> ofproto/ofproto-dpif.c |4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
> index 01299e1..80874b8 100644
> --
Acked-by: Jarno Rajahalme
On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
> Until now, struct flow_miss contained a list of packets and a list of
> upcalls. Each packet in the list of packets can be obtained from the
> corresponding upcall in the list of upcalls via upcall->dpif_upcall.packet,
>
Acked-by: Jarno Rajahalme
On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
> ---
> ofproto/ofproto-dpif.c |5 +
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
> index b41c179..28dc661 100644
> --
Acked-by: Jarno Rajahalme
On Sep 16, 2013, at 2:59 PM, Ben Pfaff wrote:
> valgrind reported:
>
>Invalid read of size 4
> at 0x806ADC1: odp_port_to_ofport (hmap.h:267)
> by 0x8077C05: xlate_receive (ofproto-dpif-xlate.c:523)
> by 0x8073994: handle_miss_upcalls (ofproto-dpi
On Tue, Sep 17, 2013 at 10:22:04AM -0700, Jarno Rajahalme wrote:
> I get test failures after this patch, seems to relate to slow path
> processing. Haven't looked at them, but I see failures on 656 and 657
> come and go, so those are most likely unrelated test timing issues.
Thanks for the report,
On Mon, Sep 9, 2013 at 12:20 AM, Simon Horman wrote:
> Allow datapath to recognize and extract MPLS labels into flow keys
> and execute actions which push, pop, and set labels on packets.
>
> Based heavily on work by Leo Alterman, Ravi K, Isaku Yamahata and Joe
> Stringer.
>
> Cc: Ravi K
> Cc: L
OK, done.
I reposted the series since this had been left dormant a while.
On Thu, Sep 05, 2013 at 01:49:58PM -0700, Ethan Jackson wrote:
> I think ovsthread_id_self() is the best choice. It might be nice to
> make that return the result of gettid() on linux, but I don't feel
> strongly about it.
This will be convenient in an upcoming commit.
I had to add -Wno-format-zero-length to suppress a GCC warning about a
zero-length format string in this monitor_daemon() call:
set_subprogram_name("");
I don't know what that warning is good for anyway, and I guess the Clang
developers don't eith
On Fri, Sep 13, 2013 at 2:16 PM, Jesse Gross wrote:
> The symbol HAVE_NET_DEVICE_OPS was changed in 3.1 but code the that
> it was protecting dates back to before 2.6.32. Therefore, we can
> just assume that net_device_ops exists in all cases and drop the
> compat code.
>
> Signed-off-by: Jesse Gr
On Fri, Sep 13, 2013 at 2:37 PM, Jesse Gross wrote:
> The symbol HAVE_NET_DEVICE_OPS was removed in 3.1 even though these ops
> still exist, so on newer kernels we define the symbol ourselves. However,
> SLES 11 SP3 appears to have backported this symbol removal. This simply
> defines the symbol o
I get test failures after this patch, seems to relate to slow path processing.
Haven't looked at them, but I see failures on 656 and 657 come and go, so those
are most likely unrelated test timing issues.
after 6/7:
632: ofproto-dpif - controller FAILED
(ofproto-dpif.at:2
Thanks guys.
:: psi
On Sep 17, 2013, at 10:05 AM, Ben Pfaff wrote:
> On Tue, Sep 17, 2013 at 09:59:58AM -0700, Gurucharan Shetty wrote:
>> On Mon, Sep 16, 2013 at 3:15 PM, Ben Pfaff wrote:
>>> Commit 2b31d8e713de7 (vlog: Report timestamps in millisecond resolution in
>>> log messages.) introdu
Your message dated Tue, 17 Sep 2013 09:48:02 -0700
with message-id <20130917164802.gj12...@nicira.com>
and subject line Re: Bug#723564: openvswitch-datapath-dkms: fails to build on
3.10
has caused the Debian Bug report #723564,
regarding openvswitch-datapath-dkms: fails to build on 3.10
to be mark
On Mon, Sep 16, 2013 at 3:15 PM, Ben Pfaff wrote:
> Commit 2b31d8e713de7 (vlog: Report timestamps in millisecond resolution in
> log messages.) introduced milliseconds to log messages by default, but the
> Python version did not ensure that milliseconds were always formatted with
> 3 digits, so 3.
From: Guolin Yang
There are two models in OVS in configure datapath port:
1. Datapath port created before user space bridge is configured.
In this model, datapath can be deleted or added independent of
user-space.
2. Datapath port created when ovsdb is requested to add the relevant
port.
On Tue, Sep 17, 2013 at 09:59:58AM -0700, Gurucharan Shetty wrote:
> On Mon, Sep 16, 2013 at 3:15 PM, Ben Pfaff wrote:
> > Commit 2b31d8e713de7 (vlog: Report timestamps in millisecond resolution in
> > log messages.) introduced milliseconds to log messages by default, but the
> > Python version di
This may occasionally make debugging easier.
Suggested-by: Keith Amidon
Signed-off-by: Ben Pfaff
---
ofproto/ofproto-dpif-upcall.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/ofproto/ofproto-dpif-upcall.c b/ofproto/ofproto-dpif-upcall.c
index 54f441b..c65c82b 100644
-
v1->v2: Use OVS thread ID instead of arbitrary index in thread names,
as Ethan suggested.
Ben Pfaff (2):
util: Allow set_subprogram_name() to take a printf() format string.
ofproto-dpif-upcall: Give each miss_handler thread a unique name.
configure.ac |1 +
lib/util.c
Package: openvswitch-datapath-dkms
Version: 1.4.2+git20120612-9.1
Severity: important
Dear Maintainer,
When doing an installation of xcp-xapi, openvswitch gets pulled in. This
package, fails during a build of its dkms build.
Setting up openvswitch-common (1.4.2+git20120612-9.1) ...
Setting up
> Patch 1/2: generic BPF extension
> Original A and X 32-bit BPF registers are replaced with ten 64-bit registers.
> bpf opcode encoding kept the same. load/store were generalized to access
> stack,
> bpf_tables and bpf_context.
> BPF program interfaces to outside world via tables that it can read
39 matches
Mail list logo
