Looks good. Sorry for the delay.
--Justin
On Sep 29, 2011, at 10:40 AM, Ben Pfaff wrote:
> We spotted one failure of the "lock_timeout_runs_out" test several builds
> ago, but there weren't enough diagnostics to track it down and I couldn't
> reproduce it.
>
> This commit should make the fail
---
debian/automake.mk |3 +++
debian/control | 10 ++
debian/openvswitch-switch.install |1 -
debian/openvswitch-switch.manpages |1 -
debian/openvswitch-test.dirs |1 +
debian/openvswitch-test.install|1 +
debian/openvs
ovs-vlan-test runs through a number of tests to identify VLAN issues. This
is useful when trying to debug why a particular driver has issues, but it made
the testing environment a bit harder to set up. This commit adds an iperf
test to check basic functionality. It also useful in detecting perfo
> Do you habitually run "make check" as root? This causes tests to be
> skipped, so I don't recommend it.
I do the same thing, didn't realize it skips tests. I'll have to change it.
Ethan
> ___
> dev mailing list
> dev@openvswitch.org
> http://openvs
On Tue, Oct 18, 2011 at 05:50:58PM -0700, Pravin Shelar wrote:
> On Tue, Oct 18, 2011 at 3:47 PM, Jesse Gross wrote:
> > You also need to update test 387: ofproto-dpif.at:83 ofproto-dpif -
> > VLAN handling to account for the changed output.
> >
> right, vlan test was skipped in my make check.
Do
On Tue, Oct 18, 2011 at 3:47 PM, Jesse Gross wrote:
> On Thu, Oct 13, 2011 at 3:23 PM, Pravin B Shelar wrote:
>> Almost all current actions can be expressed in the form of
>> push/pop/set , where field is one of the match fields. We can
>> create three base actions and take a field. This has both
On Mon, Oct 17, 2011 at 3:14 PM, Ben Pfaff wrote:
> diff --git a/utilities/ovs-dpctl.c b/utilities/ovs-dpctl.c
> index 4d0d3c2..01bc1eb 100644
> --- a/utilities/ovs-dpctl.c
> +++ b/utilities/ovs-dpctl.c
> @@ -130,13 +130,17 @@ usage(void)
> "usage: %s [OPTIONS] COMMAND [ARG...]\n"
>
From: Justin Pettit
In general it is only possible to have a single tunnel between a pair
of end hosts with a given key. This allows one encrypted and one
unencrypted tunnel when using STT. They are distinguished by using
a unique port number on transmit and looking at the security path
on rece
From: Justin Pettit
The functions to retrieve the sec_path from a skb didn't exist
until 2.6.29, so this backports them.
Signed-off-by: Justin Pettit
Signed-off-by: Jesse Gross
Acked-by: Ben Pfaff
---
datapath/linux/compat/include/linux/skbuff.h | 14 ++
1 files changed, 14 ins
From: Justin Pettit
This backports the constants for IPsec protocols to the kernels
when they weren't available.
Signed-off-by: Justin Pettit
Signed-off-by: Jesse Gross
Acked-by: Ben Pfaff
---
datapath/linux/Modules.mk|1 +
datapath/linux/compat/include/net/xfrm.h | 14
From: Justin Pettit
This makes it possible to automatically set up IPsec encrypted VXLAN
tunnels through the OVS database.
Signed-off-by: Justin Pettit
Signed-off-by: Jesse Gross
[Ben modified this commit for VXLAN, added unit tests, and tweaked it]
Signed-off-by: Ben Pfaff
---
NEWS
This commit implements the VXLAN tunneling protocol described at
http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-00
Multicast support is not yet implemented, but it should not be too hard to
add.
In my simple VM-based test rig, "netperf" performance was the same as GRE.
Jesse really
This fixes a problem that Jesse pointed out, whereby as I sent it
out before VXLAN and VXLAN-over-IPsec weren't both supported to a
single host simultaneously.
This patch series takes advantage of patches that Justin wrote
earlier for a different purpose. I've described my modifications
in brief
On Oct 18, 2011, at 2:55 PM, Ben Pfaff wrote:
>> The spec didn't provide a concise description of what's allowed in
>> the states, so I've also added the following patch:
>
> It's nice. The N and Y would be more visually distinguished if you
> changed N to something like -, e.g.:
>
> *
On Oct 18, 2011, at 10:22 AM, Ben Pfaff wrote:
> I'd prefer to limit this just to ofproto providers, so that it should
> be declared in ofproto-provider.h instead of ofproto.h.
Makes sense.
> I'd make ofproto_port_set_state() take a "struct ofport *" directly.
> The caller should already have it
On Tue, Oct 18, 2011 at 03:47:19PM -0700, Jesse Gross wrote:
> On Thu, Oct 13, 2011 at 3:23 PM, Pravin B Shelar wrote:
> > Almost all current actions can be expressed in the form of
> > push/pop/set , where field is one of the match fields. We can
> > create three base actions and take a field. Th
On Thu, Oct 13, 2011 at 3:23 PM, Pravin B Shelar wrote:
> Almost all current actions can be expressed in the form of
> push/pop/set , where field is one of the match fields. We can
> create three base actions and take a field. This has both a nice
> symmetry and avoids inconsistencies where we can
On Tue, Oct 18, 2011 at 02:52:41PM -0700, Justin Pettit wrote:
> On Oct 18, 2011, at 10:12 AM, Ben Pfaff wrote:
> > Can the STP_ROLE_* constants have some comments? Maybe "Port closest
> > to root bridge.", "Forwards frames toward/away from root.", "Not used
> > for forwarding." or similar.
>
> H
On Oct 18, 2011, at 10:12 AM, Ben Pfaff wrote:
> STP_ID_ARGS should parenthesize the name of its argument, in case
> someone passes in a complicated expression (unlikely as that is).
Whoops.
> But can't the ID formatting macros be simplified to:
>
> #define STP_ID_FMT "%04"PRIx16"."012"PRIx64
>
Thank you, I pushed this.
On Tue, Oct 18, 2011 at 12:56:20PM -0700, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Tue, Sep 27, 2011 at 16:59, Ben Pfaff wrote:
> > Otherwise, "table=257" (e.g.) was silently accepted but had a surprising
> > effect.
> >
> > Bug #7445.
> > Reported-by: Micha
On Oct 18, 2011, at 9:53 AM, Ben Pfaff wrote:
Thanks, I cleaned up the whitespace issues, the sparse warnings, the OVS endian
typedefs, and updated the copyright.
> and then we get some useful warnings from sparse (at least with the
> patch to include/linux/types.h that I posted separately):
>
Thank you, I pushed this to master.
On Tue, Oct 18, 2011 at 11:19:10AM -0700, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Fri, Oct 14, 2011 at 11:37, Ben Pfaff wrote:
> > Avoids redundant "closing log file"/"opened log file" messages in log files.
> >
> > Reported-by: Reid Price
> > Bu
Thanks, I pushed this to master, branch-1.3, and branch-1.2.
On Tue, Oct 18, 2011 at 11:20:47AM -0700, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Sun, Oct 9, 2011 at 15:52, Ben Pfaff wrote:
> > This makes OFPAT_ENQUEUE consistent with OFPAT_OUTPUT for the purpose of
> > sending a packe
Thank you for the reviews. I pushed this series.
On Tue, Oct 18, 2011 at 12:53:59PM -0700, Ethan Jackson wrote:
> Looks good.
>
> Ethan
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Maybe. The rest are much simpler, though, and I haven't had trouble
with them in practice the way I have had with "learn".
On Tue, Oct 18, 2011 at 12:52:53PM -0700, Ethan Jackson wrote:
> Does it make sense to do something like this for the other complex
> actions we have (bundle, autopath, multi
That's reasonable, but I don't think it's going to trigger very often,
so I think it's OK as-is.
Thanks,
Ben.
On Tue, Oct 18, 2011 at 11:33:10AM -0700, Ethan Jackson wrote:
> Looks good to me.
>
> Does it make sense to guard the log message with a call to
> vlog_is_enabled()? That way we could
Thanks, I pushed this to master.
On Tue, Oct 18, 2011 at 11:28:10AM -0700, Ethan Jackson wrote:
> Looks good.
>
> Ethan
>
> On Tue, Oct 18, 2011 at 09:30, Ben Pfaff wrote:
> > Initial versions of commit 8cdf034974 "python: Implement write support in
> > Python IDL for OVSDB" converted the OVSDB
On Tue, Oct 18, 2011 at 10:49:31AM -0700, Jesse Gross wrote:
> On Mon, Oct 17, 2011 at 3:13 PM, Ben Pfaff wrote:
> > Without this commit, a pair of commands like
> > ?? ??ovs-dpctl add-if br0 gre0,type=gre,remote_ip=192.168.5.2,csum=true
> > ?? ??ovs-dpctl set-if br0 gre0,csum=false
> > would resu
On Mon, Oct 17, 2011 at 10:31:49PM -0700, Justin Pettit wrote:
> Add 802.1D spanning tree protocol support to ovs-vswitchd. Still needs
> improvements in the following areas:
>
> * Does not work over bonds.
> * Does not expire entries generated by the learning action when STP
> stat
Looks good.
Ethan
On Tue, Sep 27, 2011 at 16:59, Ben Pfaff wrote:
> Otherwise, "table=257" (e.g.) was silently accepted but had a surprising
> effect.
>
> Bug #7445.
> Reported-by: Michael Mao
> ---
> lib/ofp-parse.c | 30 ++
> 1 files changed, 26 insertions(+), 4
Looks good.
Ethan
On Tue, Sep 27, 2011 at 13:36, Ben Pfaff wrote:
> ---
> include/openflow/nicira-ext.h | 3 ++-
> tests/learn.at | 17 +
> 2 files changed, 19 insertions(+), 1 deletions(-)
>
> diff --git a/include/openflow/nicira-ext.h b/include/openflow/nic
Looks good.
Does it make sense to do something like this for the other complex
actions we have (bundle, autopath, multipath, etc)?
Ethan
On Tue, Sep 27, 2011 at 13:36, Ben Pfaff wrote:
> learn_parse() and learn_check() are supposed to apply the same checks, but
> they are implemented very diffe
Looks good.
Ethan
On Tue, Sep 27, 2011 at 13:36, Ben Pfaff wrote:
> nxm_dst_check() requires a writable field. Since NX_LEARN_DST_MATCH only
> matches on a field and doesn't write to a field, use nxm_src_check() for
> that case instead.
>
> Bug #7510.
> Reported-by: Michael Mao
> ---
> lib/le
Looks good to me.
Does it make sense to guard the log message with a call to
vlog_is_enabled()? That way we could avoid the memory allocation when
it isn't going to be used. May not matter though.
Ethan
On Tue, Sep 27, 2011 at 13:36, Ben Pfaff wrote:
> ---
> lib/nx-match.c | 7 +--
>
Looks good.
Ethan
On Tue, Oct 18, 2011 at 09:30, Ben Pfaff wrote:
> Initial versions of commit 8cdf034974 "python: Implement write support in
> Python IDL for OVSDB" converted the OVSDB schemas to Python code so that
> their Python clients could just import them. The final version instead
> rea
On Tue, Oct 18, 2011 at 11:13 AM, Ben Pfaff wrote:
> On Tue, Oct 18, 2011 at 11:08:06AM -0700, Pravin B Shelar wrote:
>> Use compose_output_action() API to generate OUTPUT action so
>> that sFlow can record output port.
>
> Looks good. Thank you for the bug fix.
>
Thanks, pushed to mast
Looks good.
Ethan
On Sun, Oct 9, 2011 at 15:52, Ben Pfaff wrote:
> This makes OFPAT_ENQUEUE consistent with OFPAT_OUTPUT for the purpose of
> sending a packet back out the input port: both only do it if the port is
> given as OFPP_IN_PORT.
>
> Found by inspection.
> ---
> ofproto/ofproto-dpif.c
Looks good.
Ethan
On Fri, Oct 14, 2011 at 11:37, Ben Pfaff wrote:
> Avoids redundant "closing log file"/"opened log file" messages in log files.
>
> Reported-by: Reid Price
> Bug #7750.
> ---
> lib/vlog.c | 21 -
> 1 files changed, 20 insertions(+), 1 deletions(-)
>
> dif
On Tue, Oct 18, 2011 at 11:08:06AM -0700, Pravin B Shelar wrote:
> Use compose_output_action() API to generate OUTPUT action so
> that sFlow can record output port.
Looks good. Thank you for the bug fix.
___
dev mailing list
dev@openvswitch.org
Use compose_output_action() API to generate OUTPUT action so
that sFlow can record output port.
---
ofproto/ofproto-dpif.c |5 ++---
1 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index ae76e98..9f73a8f 100644
--- a/ofprot
On Mon, Oct 17, 2011 at 3:13 PM, Ben Pfaff wrote:
> Without this commit, a pair of commands like
> ovs-dpctl add-if br0 gre0,type=gre,remote_ip=192.168.5.2,csum=true
> ovs-dpctl set-if br0 gre0,csum=false
> would result in a csum of "true" for gre0, that is, the second command
> would silent
On Mon, Oct 17, 2011 at 3:13 PM, Ben Pfaff wrote:
> Until now this has just silently failed, but it seems to me like we should
> actively reject it.
>
> Signed-off-by: Ben Pfaff
This makes sense but it seems to me that there are plenty of other
things that fall into the category of invalid but s
On Oct 18, 2011, at 9:35 AM, Ben Pfaff wrote:
> This should go to branch-1.2 and branch-1.3 as well as master.
Thanks. I'll push this and the other bug fix into branch-1.2 momentarily. As
previously mentioned, this whole series is going into branch-1.3.
--Justin
On Oct 18, 2011, at 9:32 AM, Ben Pfaff wrote:
> I guess that this should go to branch-1.2 and branch-1.3 also.
I'm pulling it into branch-1.2 right now. The plan is for this whole series to
go into branch-1.3, so it should be covered.
--Justin
___
Thanks for the review. I pushed this to master and branch-1.3.
On Tue, Oct 18, 2011 at 10:21:11AM -0700, Justin Pettit wrote:
> Looks good. Since there's still some stuff landing in branch-1.3
> (such as STP), I'm guessing this is worth including, too. If you
> agree, can you do that when you c
On Mon, Oct 17, 2011 at 10:31:48PM -0700, Justin Pettit wrote:
> This will be used in an upcoming commit.
I'd prefer to limit this just to ofproto providers, so that it should
be declared in ofproto-provider.h instead of ofproto.h.
I'd make ofproto_port_set_state() take a "struct ofport *" direct
Looks good. Since there's still some stuff landing in branch-1.3 (such as
STP), I'm guessing this is worth including, too. If you agree, can you do that
when you commit?
Thanks,
--Justin
On Oct 18, 2011, at 9:49 AM, Ben Pfaff wrote:
> The Linux headers only check endianness if __CHECK_ENDI
On Mon, Oct 17, 2011 at 10:31:47PM -0700, Justin Pettit wrote:
> No changes are made to 'ofproto', so it's safe to mark the argument as
> const. This will be useful in a later commit.
Looks good, thanks.
___
dev mailing list
dev@openvswitch.org
http://o
On Mon, Oct 17, 2011 at 10:31:46PM -0700, Justin Pettit wrote:
> - Don't apply endian conversions to flags, which are 8 bits.
> - Use #defines for default times for use outside library.
> - Clarify our behavior when in STP_DISABLED state.
> - Add "aux" member to STP port struct to b
On Mon, Oct 17, 2011 at 10:31:45PM -0700, Justin Pettit wrote:
> At one point, the OVS distribution contained an IEEE 802.1D Spanning
> Tree Protocol (STP) library written by Ben Pfaff and based on the
> 802.1D-1998 reference code. It was never integrated into ovs-vswitchd,
> so it was removed as
The Linux headers only check endianness if __CHECK_ENDIAN__ is declared.
We want that, so turn it on.
---
include/linux/types.h |5 +
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/include/linux/types.h b/include/linux/types.h
index 13112e8..fb05744 100644
--- a/include/lin
On Mon, Oct 17, 2011 at 10:31:44PM -0700, Justin Pettit wrote:
> static const uint8_t eth_addr_stp[ETH_ADDR_LEN] OVS_UNUSED
> -= { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x01 };
> += { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x00 };
Sheesh. That's embarrassing.
This should go to branch-1.2 and branch-1.3
Looks good, thank you.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Looks good, thank you.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On Mon, Oct 17, 2011 at 10:31:41PM -0700, Justin Pettit wrote:
> When the OFPPC_NO_FLOOD flag is toggled on the port, the "floodable"
> member of the bundle was not updated. This would cause OFPP_NORMAL to
> not include the proper ports when flooding. With this commit,
> OFPPC_NO_FLOOD changes wi
I sent out my fix.
On Mon, Oct 17, 2011 at 12:34:27PM -0700, Ethan Jackson wrote:
> Sounds good. I'll drop this patch.
>
> Ethan
>
> On Mon, Oct 17, 2011 at 12:32, Ben Pfaff wrote:
> > On Mon, Oct 17, 2011 at 12:10:18PM -0700, Ethan Jackson wrote:
> >> --- a/tests/.gitignore
> >> +++ b/tests/.
Initial versions of commit 8cdf034974 "python: Implement write support in
Python IDL for OVSDB" converted the OVSDB schemas to Python code so that
their Python clients could just import them. The final version instead
read the schemas from a file, but I forgot to remove some code that
generated th
On Tue, Oct 18, 2011 at 8:51 AM, Ben Pfaff wrote:
> On Mon, Oct 17, 2011 at 10:10:52PM -0700, Jesse Gross wrote:
>> Linux 3.1 drops the symbol HAVE_NET_DEVICE_OPS that lets us know
>> whether struct netdev_ops is present. As a result, we need to
>> replace it with an explicit version check.
>>
>>
On Mon, Oct 17, 2011 at 10:10:53PM -0700, Jesse Gross wrote:
> Linux 3.1 adds a flag to check whether it's OK for shared skbs to
> be transmitted on devices. This generally isn't a problem for
> hardware devices but software devices such as OVS that hold state
> in the skb need to clear the flag,
On Mon, Oct 17, 2011 at 10:10:52PM -0700, Jesse Gross wrote:
> Linux 3.1 drops the symbol HAVE_NET_DEVICE_OPS that lets us know
> whether struct netdev_ops is present. As a result, we need to
> replace it with an explicit version check.
>
> Signed-off-by: Jesse Gross
Acked-by: Ben Pfaff
__
60 matches
Mail list logo
