Signed-off-by: Thomas Graf
---
lib/dpif-linux.c | 3 +-
lib/netlink-protocol.h | 39 ++
lib/netlink-socket.c | 345 -
lib/netlink-socket.h | 10 +-
4 files changed, 361 insertions(+), 36 deletions(-)
diff --git a/lib/dpif-linux.c
On 05/23/14 at 01:15am, Thomas Graf wrote:
> Test: * Avg pps as reported by pktgen for a single pktgen thread
> * qemu, vcpu=2, mem=2GB, megaflows disabled
>
> Orig [0] ZC [1]mmap [2] shared-ofpbuf [3]
> 64B (linear)624K 564K 962K
On 05/22/14 at 05:20pm, Ben Pfaff wrote:
> On Fri, May 23, 2014 at 01:15:58AM +0200, Thomas Graf wrote:
> > Signed-off-by: Thomas Graf
>
> Thanks for the patches.
Thanks for the prompt review.
>
> Definitions in netlink-protocol.h have to be written from the
> perspecti
On 05/23/14 at 08:47am, Ben Pfaff wrote:
> On Fri, May 23, 2014 at 01:15:58AM +0200, Thomas Graf wrote:
> > Signed-off-by: Thomas Graf
>
> nl_sock_recv__() compares its 'events' member against EPOLLERR with ==,
> but EPOLL* are bitmasks so should it use & to test
On 05/23/14 at 08:51am, Ben Pfaff wrote:
> On Fri, May 23, 2014 at 09:03:54AM +0100, Thomas Graf wrote:
> > > Using nl_sock_send_linear() can cause messages to be reordered. Do we
> > > need to wait for the tx ring to empty before calling it?
> >
> > Are y
On 05/23/14 at 09:57am, Ben Pfaff wrote:
> On Fri, May 23, 2014 at 01:16:01AM +0200, Thomas Graf wrote:
> > If the message was received via a shared memory ring the frame
> > is usable as ofpbuf base directly without copying the buffer
> > unnecessarily. A new destructor() c
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
for prompting me to investigate this.
>
> Signed-off-by: Simon Horman
> Acked-by: YAMAMOTO Takashi
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:01pm, Ben Pfaff wrote:
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:01pm, Ben Pfaff wrote:
> This has always been supported, but the documentation was missing.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:01pm, Ben Pfaff wrote:
> This flag was overlooked when support for the "learn" action was added.
> (It was always supported in the OpenFlow code, just not in the ovs-ofctl
> interface.)
>
> Signed-off-by: Ben Pfaf
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:01pm, Ben Pfaff wrote:
> Found by valgrind.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> recirc_id_pool_create() allocates memory but recirc_id_pool_destroy() did
> not destroy it.
>
> Found by valgrind.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
d
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> Each of these functions had only a single caller, and breaking them out as
> separate functions didn't seem to many the code clearer.
>
> I added a new function meter_insert_rule(): oftable_insert_rule() had used
> members of struct meter directly, which
ations anyway. It might
> change behavior of some hardware implementation, but I don't actually know
> of any.
>
> ONF-JIRA: EXT-362.
> Signed-off-by: Ben Pfaff
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> I'd like to use this from connmgr.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> +/* By default, criteria initialized by rule_criteria_init() will match flows
> + * that are read-only, on the assumption that the collected flows won't be
> + * modified. This function to match only flows that are be modifiable.
gned-off-by: Ben Pfaff
Looks sane. Not verified line by line ;-)
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/04/14 at 01:53pm, Andy Zhou wrote:
> Signed-off-by: Andy Zhou
I would personally value a commit message here since this looks like
more than just a simplification ;-)
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo
On 05/22/14 at 01:39pm, Justin Pettit wrote:
> Below, is a first cut of the design document I wrote for integrating with the
> connection tracker. As I mentioned at my OpenStack presentation, I have a
> prototype that (largely) implements this, but it's not ready to be shared
> yet. The goal i
On 06/06/14 at 01:59pm, Andy Zhou wrote:
> On Fri, Jun 6, 2014 at 3:46 AM, Thomas Graf wrote:
> > On 06/04/14 at 01:53pm, Andy Zhou wrote:
> >> Signed-off-by: Andy Zhou
> >
> > I would personally value a commit message here since this looks like
> > more tha
shall have to fix this. Also we ought to annotate 'rules'
> with OVS_GUARDED_BY.)
>
> * The eviction groups are only used during flow_mods, which
> hold ofproto_mutex anyway.
>
>
On 06/06/14 at 03:25pm, Ben Pfaff wrote:
> On Fri, Jun 06, 2014 at 11:11:57AM +0100, Thomas Graf wrote:
> > On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> > > +/* By default, criteria initialized by rule_criteria_init() will match
> > > flows
> > > + * that are r
On 06/06/14 at 02:37pm, Andy Zhou wrote:
> +static void tbl_mask_array_delete_mask(struct mask_array *ma,
> +const struct sw_flow_mask *mask)
> +{
> + int i = 0;
> +
> + /* Delete a mask pointer from the valid section.
> + *
> + * Also move the
s removal into two
> stages will make it easier to understand and review. If not, the commits
> could be squashed.
>
> Signed-off-by: Ben Pfaff
Looks straight forward
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/09/14 at 03:26pm, Andy Zhou wrote:
> On Mon, Jun 9, 2014 at 2:58 PM, Thomas Graf wrote:
> > On 06/06/14 at 02:37pm, Andy Zhou wrote:
> >> +static void tbl_mask_array_delete_mask(struct mask_array *ma,
> >> +cons
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> This commit finishes the removal of asynchronous flow table operations
> begun in the previous commit, by removing ofoperation and ofopgroup
> entirely and all of the code that depended on them. Following this commit,
> all the internal documentation and
On 06/05/14 at 10:02pm, Ben Pfaff wrote:
> Nothing actually used it.
>
> An upcoming commit will use part of the hole for a new member.
>
> Signed-off-by: Ben Pfaff
LGTM in general
The commit description is slightly misleading as 'hole' suggests that
none of struct rule_action's member offsets
r fail.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
ut it.
Code looks good aside from the race condition. Happy to provide my
ACK for the code as-is
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/10/14 at 07:02pm, Avinash wrote:
> Hello,
As a general remark. It would be nice to have such a mail formatted as
the cover letter of the patches series and all the patches as reply to
it. That way all patches and are neatly bundled in a single email
thread. git format-patch --cover and git s
On 06/10/14 at 07:03pm, Avinash wrote:
> Kernel recognizes and accepts ethertype of provider VLANs (ETH_P_8021AD)
> along with normal VLAN (ETH_P_8021Q). Also in flow key formation,
> only the outermost VLAN is considered. Remaining stacked VLANs are skipped.
>
> Signed-off-by: Avinash
Something
zeof(u32), 0);
> + else
> + hash = jhash2((const u32 *)&pkt_key->ipv6.addr,
> +sizeof(pkt_key->ipv6.addr) / sizeof(u32), 0);
Perhaps check for ETH_P_IPV6 and WARN() on the else branch? OTOH,
unlikely that is ever to
non-error situations.
>
> Removing the memset() reduces the profile of ovs_flow_extract()
> from 0.64% to 0.56% when tested with large packets on a 10G link.
>
> Suggested-by: Pravin Shelar
> Signed-off-by: Jesse Gross
Nice!
Acked-by: Thomas Graf
__
> Signed-off-by: Jesse Gross
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
/* No argument, OAM frame. */
> __OVS_TUNNEL_KEY_ATTR_MAX
> };
Depending on how long you want to keep this out of tree, it might be
wise to push the above bits upstream as a placeholder to reserve the
bit.
LGTM otherwise
Acked-by: Thomas Graf
___
dev mail
On 06/10/14 at 04:47pm, Jesse Gross wrote:
> In some cases, the names of tunnel protocols are longer than the
> maximum allowed after concatenating port numbers. This truncates
> names to the correct size rather than assert-failing on the assumption
> that collisions between names of protocols are
rs it out to reduce stack pressure and improve readibility.
>
> Signed-off-by: Jesse Gross
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/10/14 at 04:47pm, Jesse Gross wrote:
> diff --git a/datapath/flow.c b/datapath/flow.c
> index f1bb95d..7b108ed 100644
> --- a/datapath/flow.c
> +++ b/datapath/flow.c
> @@ -455,6 +455,13 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port,
> struct sw_flow_key *key)
> struc
On 06/11/14 at 11:11am, Jesse Gross wrote:
> On Wed, Jun 11, 2014 at 3:04 AM, Thomas Graf wrote:
> > On 06/10/14 at 04:47pm, Jesse Gross wrote:
> >> In some cases, the names of tunnel protocols are longer than the
> >> maximum allowed after concatenating port numbers. T
On 06/11/14 at 11:10am, Jesse Gross wrote:
> On Wed, Jun 11, 2014 at 3:00 AM, Thomas Graf wrote:
> > On 06/10/14 at 04:47pm, Jesse Gross wrote:
> >> diff --git a/include/linux/openvswitch.h b/include/linux/openvswitch.h
> >> index 33423ad..57d40e3 100644
> >&g
> I actually did look into breaking the flow key up but the result was
> messy and arguably worse since it means that the optimizations that we
> already have might not just work.
Thanks for this excellent explanation.
Acked-by: Thomas Graf
_
On 06/12/14 at 09:04am, Jesse Gross wrote:
> On Wed, Jun 11, 2014 at 11:06 PM, Avinash wrote:
> > On 6/11/14, Jesse Gross wrote:
> >> On Wed, Jun 11, 2014 at 7:26 AM, Avinash wrote:
> >>> On 6/10/14, Thomas Graf wrote:
> >>>> On 06/10/14 at 07:03
Signed-off-by: Thomas Graf
---
utilities/ovs-vsctl.8.in | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/utilities/ovs-vsctl.8.in b/utilities/ovs-vsctl.8.in
index 1701b48..50d8879 100644
--- a/utilities/ovs-vsctl.8.in
+++ b/utilities/ovs-vsctl.8.in
@@ -973,7 +973,7
Signed-off-by: Thomas Graf
---
AUTHORS | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/AUTHORS b/AUTHORS
index c03c705..33de7b1 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -121,7 +121,7 @@ SUGYO Kazushi sugyo@gmail.com
Tadaaki Nagao na...@stratosphere.co.jp
On 06/17/14 at 09:19am, Justin Pettit wrote:
> I think the way it was worked, since we use strcasecmp(), but we may as
> well be consistent with how it's documented elsewhere. It's been applied.
Thanks!
I think it's the actual schema validation that enforces correct
capitalization:
$ sudo ovs-v
Signed-off-by: Thomas Graf
---
utilities/ovs-ofctl.8.in | 4
1 file changed, 4 insertions(+)
diff --git a/utilities/ovs-ofctl.8.in b/utilities/ovs-ofctl.8.in
index cc811ff..efd8c11 100644
--- a/utilities/ovs-ofctl.8.in
+++ b/utilities/ovs-ofctl.8.in
@@ -1144,6 +1144,10 @@ This form of
Signed-off-by: Thomas Graf
---
utilities/ovs-ofctl.8.in | 50
1 file changed, 25 insertions(+), 25 deletions(-)
diff --git a/utilities/ovs-ofctl.8.in b/utilities/ovs-ofctl.8.in
index efd8c11..4ebaa18 100644
--- a/utilities/ovs-ofctl.8.in
+++ b
Make the parser compatible with how we print group buckets and allow
actions= to be specified, e.g.:
group_id=1,type=all,bucket=actions=output:2,bucket=actions=output:3
Also, better document the supported syntax in the manual page.
Signed-off-by: Thomas Graf
---
lib/ofp-parse.c | 6
On 06/23/14 at 12:37pm, Alex Wang wrote:
> Signed-off-by: Alex Wang
Some comments below
> ---
> lib/automake.mk |2 +
> lib/ovs-numa.c | 208
> +++
> lib/ovs-numa.h | 33
> tests/ofproto-macros.at |1 +
>
On 06/24/14 at 12:40am, Thomas Graf wrote:
> On 06/23/14 at 12:37pm, Alex Wang wrote:
> > +/* Returns the number of cpu cores. */
> > +uint32_t
> > +ovs_numa_get_n_cores(void)
> > +{
> > +return hmap_count(&all_cpu_cores);
> > +}
> > +
> &
g
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/23/14 at 12:37pm, Alex Wang wrote:
> This will be the default setting for dpdk iface.
>
> Signed-off-by: Alex Wang
I think you can remove the NR_QUEUE #define in lib/dpif-netdev.h
with this patch.
> ---
> lib/netdev-dpdk.c | 47 ---
> 1 file c
On 06/23/14 at 12:37pm, Alex Wang wrote:
> The pmd threads are pinned to available cpu cores on the
> corresponding cpu socket.
>
> Signed-off-by: Alex Wang
> ---
> lib/dpif-netdev.c | 172
> ++---
> lib/netdev-dpdk.c | 27 -
> lib/netd
On 06/23/14 at 05:08pm, Ben Pfaff wrote:
> Found by sparse.
>
> Signed-off-by: Ben Pfaff
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/24/14 at 08:56am, Ben Pfaff wrote:
> On Fri, Jun 20, 2014 at 02:24:50PM +0200, Thomas Graf wrote:
> > Make the parser compatible with how we print group buckets and allow
> > actions= to be specified, e.g.:
> >
> > group_id=1,type=all,bucket=actions=outpu
: Thomas Graf
---
v2: Add unit test upon suggestion of Ben Pfaff
lib/ofp-parse.c | 6 ++
tests/ofproto.at | 3 ++-
utilities/ovs-ofctl.8.in | 6 --
3 files changed, 12 insertions(+), 3 deletions(-)
diff --git a/lib/ofp-parse.c b/lib/ofp-parse.c
index 93bdbea..ff7b9ce
uch as VXLAN. On these kernels the source
> port is always zero when using VXLAN. To solve this problem while
> still taking advantage of the procomputed hash, this rehashes the
> hash so that the entropy is spread throughout 32 bits.
>
> Signe
n, because the Linux kernel datapath throws away tunnel information
> supplied in the OVS_PACKET_CMD_EXECUTE metadata argument.
>
> VMware-BZ: #1270110
> Reported-by: Srinivas Neginhal
> Signed-off-by: Ben Pfaff
> CC: Jesse Gross
Thanks Ben! I ran into this t
On 06/24/14 at 06:40pm, Alex Wang wrote:
> Signed-off-by: Alex Wang
>
> ---
> PATCH -> V2:
> - Use readdir_r() instead of readdir() for reentrency.
> - Address review comments from Thomas Graf.
> - Add dummy interface for WIN32 case.
One minor comment below.
> in
On 06/24/14 at 04:04pm, Daniele Di Proietto wrote:
> Commit f4fd623c4c25 introduced a bug in netdev_dpdk_send(): if multiple
> consecutive packets exceed MTU, only the first one is deleted and
> counted.
>
> This should fix the bug
>
> Signed-off-by: Daniele Di Proietto
On 06/24/14 at 04:05pm, Daniele Di Proietto wrote:
> Signed-off-by: Daniele Di Proietto
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
s the driver to use a simpler tx routine, resulting in throuput
> improvements (~7.5%)
>
> Signed-off-by: Daniele Di Proietto
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/24/14 at 06:40pm, Alex Wang wrote:
> Signed-off-by: Alex Wang
> ---
> PATCH -> V2:
> - Rebase.
LGTM, minor nitpick below.
Acked-by: Thomas Graf
> diff --git a/lib/netdev-dpdk.h b/lib/netdev-dpdk.h
> index f443a21..da507ce 100644
> --- a/lib/netdev-dpdk.h
&
On 06/25/14 at 08:54am, Alex Wang wrote:
> > index 85ecc5c..77b9b39 100644
>
> > > --- a/tests/ofproto-macros.at
> > > +++ b/tests/ofproto-macros.at
> > > @@ -76,6 +76,7 @@ m4_define([OVS_VSWITCHD_START],
> > > AT_CHECK([ovs-vswitchd --detach --no-chdir --pidfile
> > --enable-dummy$3 --disable
unfolded.
Tested with autoconf 2.63 and 2.69.
Cc: Scott Mann
Cc: Don Kehn
Signed-off-by: Thomas Graf
---
configure.ac | 2 +-
m4/compat.at | 32 ++
m4/openvswitch.m4 | 2 ++
tests/ovsdb-row.at| 62
unfolded.
The Debian control file is left untouched on purpose. The decision
whether to adjust the dependency is left to the respective maintainers.
Tested with autoconf 2.63 and 2.69.
Cc: Scott Mann
Cc: Don Kehn
Signed-off-by: Thomas Graf
---
v2: - Addressed Ben's comments:
- Up
On 06/26/14 at 06:04pm, Ben Pfaff wrote:
> On Fri, Jun 27, 2014 at 01:50:25AM +0200, Thomas Graf wrote:
> > Reduces the dependency on autoconf from 2.64 to 2.63 to ease building
> > on older platforms. There is only a few macros missing and they can
> > be provided easily.
On 07/01/14 at 05:43pm, Jesse Gross wrote:
> There are many possible ways that a flow can be invalid so we've
> added logging for most of them. This adds logs for the remaining
> possible cases so there isn't any ambiguity while debugging.
>
> CC: Federico Iezzi
> Signed-off-by: Jesse Gross
It'
; Signed-off-by: Jesse Gross
LGTM
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 06/27/14 at 04:21pm, Lorand Jakab wrote:
> diff --git a/lib/odp-execute.c b/lib/odp-execute.c
> index cb89e72..b5add78 100644
> --- a/lib/odp-execute.c
> +++ b/lib/odp-execute.c
> @@ -260,6 +260,25 @@ odp_execute_actions__(void *dp, struct dpif_packet
> **packets, int cnt,
> break;
On 07/10/14 at 05:29pm, Joe Stringer wrote:
> Split up ovs_flow_cmd_fill_info() to make it easier to cache parts of a
> dump reply. This will be used to streamline flow_dump in the next patch.
>
> Signed-off-by: Joe Stringer
LGTM
Acked-by: Thomas Graf
> @@ -734,9 +722,2
On 07/10/14 at 05:29pm, Joe Stringer wrote:
> Converting the flow key and mask back into netlink format during flow
> dump is fairly expensive. By caching the netlink versions of these
> during flow setup, and copying the memory directly during flow dump, we
> are able to support up to 33% more flo
On 07/09/14 at 03:22pm, Daniele Di Proietto wrote:
> match_validate() enforce that a mask matching on NDP attributes has also an
> exact match on ICMPv6 type.
> The ICMPv6 type, which is 8-bit wide, is stored in the 'tp.src' field of
> 'struct sw_flow_key', which is 16-bit wide.
> Therefore, an exa
On 07/11/14 at 11:22am, Joe Stringer wrote:
> Thanks for the review,
>
> On 10 July 2014 21:13, Thomas Graf wrote:
>
> > If I understand the code correctly the gain is only visible on
> > consecutive dumps of the same flow. How about constructing the cache
> > whe
On 07/10/14 at 09:32am, Daniele Di Proietto wrote:
> This cannot set tp.src to htons(0x), because icmpv6_key->icmpv6_type is
> __u8.
>
> The bug has always been there, since a1c564be (datapath: Mega flow
> implementation).
>
> datapath/flow.c:201
>
> if (match->mask && (match->mask->
On 07/11/14 at 11:29pm, Joe Stringer wrote:
> I'm skeptical of taking the ovs_lock(). The current patch performs this
> key/mask cache construction inside ovs_lock() as part of the critical
> section for flow install. If we perform this during flow_dump, but extend
> the ovs_lock to the entire flow
On 07/11/14 at 09:28am, Flavio Leitner wrote:
> On Fri, Jul 11, 2014 at 11:27:09AM +1200, Joe Stringer wrote:
> > On 11 July 2014 04:53, Flavio Leitner wrote:
> >
> > > Can we push the cache construction to later that it doesn't impact
> > > either flow setup or flow dump? I.e. like a workqueue?
Port datapath to work with kernrels up to 3.17 and use 3.16.2 as
the new kernel for CI testing.
Tested with 3.14, 3.16.2, and net-next (3.17).
Signed-off-by: Thomas Graf
Co-authored-by: Madhu Challa
---
v2:
- Swapped ignore_df local_df compat direction
- Provide alloc_netdev() compat macro
On 09/18/14 at 10:55am, Simon Horman wrote:
> Allow formatting of select group action. This is used
> when pretty-printing datapath flows. Subsequent patches
> will add support for the select group action to the datapath
> and ovs-vswtichd.
>
> Signed-off-by: Simon Horman
> ---
> lib/odp-util.c
On 09/18/14 at 10:55am, Simon Horman wrote:
> +const struct nlattr *bucket_actions(const struct nlattr *attr)
> +{
> + const struct nlattr *a;
> + int rem;
> +
> + for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
> + a = nla_next(a, &rem)) {
> + if (nla_type(a
On 09/18/14 at 10:55am, Simon Horman wrote:
> diff --git a/datapath/datapath.h b/datapath/datapath.h
> index c5d3c86..74a15e6 100644
> --- a/datapath/datapath.h
> +++ b/datapath/datapath.h
> @@ -209,4 +209,9 @@ do {
> \
> if (net_r
On 09/19/14 at 01:09pm, Ben Pfaff wrote:
> I don't see why the testsuite is supported only with GCC.
The motivation was to keep the total build time on a reasonable level
and below the < 1h limit for all build threads.
___
dev mailing list
dev@openvswitc
On 09/19/14 at 01:09pm, Ben Pfaff wrote:
> CC: Thomas Graf
> Signed-off-by: Ben Pfaff
Nice
Acked-by: Thomas Graf
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
On 09/19/14 at 10:47pm, Thomas Graf wrote:
> On 09/19/14 at 01:09pm, Ben Pfaff wrote:
> > I don't see why the testsuite is supported only with GCC.
>
> The motivation was to keep the total build time on a reasonable level
> and below the < 1h limit for all build thread
On 09/20/14 at 10:14am, Jiri Pirko wrote:
> Sat, Sep 20, 2014 at 12:12:12AM CEST, john.r.fastab...@intel.com wrote:
> >I was considering a slightly different approach where the
> >device would report via netlink the fields/actions it
> >supported rather than creating pre-defined enums for every
> >
On 09/20/14 at 06:19am, Jamal Hadi Salim wrote:
> The ovs guys are against this and now no *api exists*?
> Write a 15 tuple classifier tc classifier and use it. I will be more
> than happy to help you. I will get to it when we have basics L2 working
> on real devices.
Nothing speaks against having
On 09/20/14 at 07:32am, Jamal Hadi Salim wrote:
> I am sorry to have tied the two together. Maybe not OVS but the approach
> described is heaven for vendor SDKs.
I fail to see the connection. You can use switch vendor SDK no matter
how we define the kernel APIs. They already exist and have been
de
On 09/20/14 at 03:50pm, Alexei Starovoitov wrote:
> I think HW should not be limited by SW abstractions whether
> these abstractions are called flows, n-tuples, bridge or else.
> Really looking forward to see "device reporting the headers as
> header fields (len, offset) and the associated parse gr
On 09/20/14 at 07:54pm, Pravin Shelar wrote:
> > +#define NET_NAME_UNKNOWN 0
> > +#define alloc_netdev(sizeof_priv, name, name_assign_type, setup) \
> > +alloc_netdev_mqs(sizeof_priv, name_assign_type, setup, 1, 1)
> > +#endif
> > +
> I changed name_assign_type to name.
Thanks Pravin.
On 09/22/14 at 08:10am, Tom Herbert wrote:
> Thomas, can you (or someone else) quantify what the host case is. I
> suppose there may be merit in using a switch on NIC for kernel bypass
> scenarios, but I'm still having a hard time understanding how this
> could be integrated into the host stack wit
On 09/22/14 at 03:40pm, Tom Herbert wrote:
> On Mon, Sep 22, 2014 at 3:17 PM, Thomas Graf wrote:
> > What makes stateful offload interesting to me is that the final
> > desintation of a packet is known at RX and can be redirected to a
> > queue or VF. This allows to build pac
On 09/22/14 at 06:36pm, John Fastabend wrote:
> n-tuple has some deficiencies,
>
> - its not possible to get the capabilities to learn what
> fields are supported by the device, what actions, etc.
>
> - its ioctl based so we have to poll the device
>
> - only supports a
On 09/22/14 at 03:40pm, Tom Herbert wrote:
> On Mon, Sep 22, 2014 at 3:17 PM, Thomas Graf wrote:
> > What makes stateful offload interesting to me is that the final
> > desintation of a packet is known at RX and can be redirected to a
> > queue or VF. This allows to build pac
On 09/22/14 at 07:16pm, Tom Herbert wrote:
> Turn on UDP RSS on the device and I bet you'll see those differences
> go away! Once we moved to UDP encapsulation, there's really little
> value in looking at inner headers for RSS or ECMP, this should be
> sufficient. Sure someone might want to parse t
On 09/23/14 at 12:11am, Andy Gospodarek wrote:
> There are clearly some that are most interested in how an eSwitch on an
> SR-IOV capable NIC be controlled can provide traditional forwarding help
> as well as offload the various technologies they hope to terminate
> at/inside their endpoint (host/g
Help produce better optimized code.
Signed-off-by: Thomas Graf
---
datapath/actions.c | 2 +-
datapath/datapath.c | 4 ++--
datapath/datapath.h | 2 +-
datapath/flow_table.c | 2 +-
datapath/flow_table.h | 2 +-
datapath/vport-netdev.c | 2 +-
datapath/vport.c| 4
On 09/23/14 at 10:57am, Pravin Shelar wrote:
> On Tue, Sep 23, 2014 at 12:31 AM, Thomas Graf wrote:
> > On 09/17/14 at 02:02pm, Thomas Graf wrote:
> >> ovs_nla_copy_actions() already frees the allocated actions buffers,
> >> ovs_flow_cmd_new() will free it a
On 09/24/14 at 03:00pm, Simon Horman wrote:
> On Fri, Sep 19, 2014 at 03:06:38PM +0100, Thomas Graf wrote:
> > Can we rename & move this to instead?
>
> Sure, how about nla_is_last()?
Sounds great
___
dev mailing list
dev@o
201 - 300 of 1023 matches
Mail list logo
