threads. All not affected
threads will forward traffic without any additional latencies.
id-pool created for static tx queue ids to keep them sequential
in a flexible way. non-PMD thread will always have
static_tx_qid = 0 as it was before.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
This simple change makes id-pool to always allocate the
lowest possible id from the pool. No any other code affected
because, actually, there is no users of 'id_pool_free_id' in
OVS.
This behaviour of id-pool will be used in the next patch.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
number
of wanted tx queues to avoid unnecessary reconfigurations.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dpif-netdev.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 6e575ab..e2b4f39 100644
--- a/li
Ilya Maximets (4):
id-pool: Allocate the lowest available ids.
dpif-netdev: Incremental addition/deletion of PMD threads.
dpif-netdev: Avoid port's reconfiguration on pmd-cpu-mask changes.
dpif-netdev: Don't uninit emc on reload.
lib/dpif-netdev.c | 129
f this values from 0 to 4294967295 are
useful?
What do you think?
Best regards, Ilya Maximets.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
Not a complete review. This code is full of races.
See details inline.
Best regards, Ilya Maximets.
On 17.01.2017 18:37, antonio.fische...@intel.com wrote:
> This patch implements the intermediate Tx queues on 'dpdk' type ports.
>
> Test results:
> * In worst case scenario with f
This should be helpful for have all the logs in one place.
'ovs-appctl vlog' commands for 'dpdk' module can be used
to configure the log level. Lower bound for DPDK logging
(--log-level) still can be passed through 'dpdk-extra' field.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
4/libc.so.6
#8 _start ()
Fix that by not calling free for the memory passed to DPDK.
CC: Aaron Conole <acon...@redhat.com>
Fixes: bab694097133 ("netdev-dpdk: Convert initialization from cmdline to db")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/d
On 28.11.2016 21:55, Aaron Conole wrote:
> Ilya Maximets <i.maxim...@samsung.com> writes:
>
>> According to DPDK API (lib/librte_eal/common/include/rte_eal.h):
>>
>> "After the call to rte_eal_init(), all arguments argv[x]
>> with x < re
threads.
Also, few comments inline.
Best regards, Ilya Maximets.
On 16.11.2016 03:46, Daniele Di Proietto wrote:
> Currently we have three different code paths that deal with pmd threads
> and queues, in response to different input
>
> 1. When a port is added
> 2. When a port is de
On 16.11.2016 03:45, Daniele Di Proietto wrote:
> Tunnel devices have 0 txqs and don't support netdev_send(). While
> netdev_send() simply returns EOPNOTSUPP, the XPS logic is still executed
> on output, and that might be confused by devices with no txqs.
>
> It seems better to have different
quot;dpif-netdev: Allow different numbers of
rx queues for different ports.")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/netdev-dpdk.c | 2 +-
lib/netdev-dummy.c | 4 ++--
tests/pmd.at | 7 +++
3 files changed, 10 insertions(+), 3 del
Hi, Daniele.
Thanks for v3.
Acked-by: Ilya Maximets <i.maxim...@samsung.com>
On 09.01.2017 06:14, Daniele Di Proietto wrote:
> The first two commits of the series are trivial bugfixes for dpif-netdev.
>
> Then the series fixes a long standing bug that caused a crash when the
>
The copy should be used here.
Additionally, 'strlen' changed to the faster check.
Fixes: 821b86649a90 ("netdev-dpdk: Don't try to unregister empty vhost_id.")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
Version 2:
* 'strlen' --> '[0]' (Suggested by Ben
> How about "if (!vhost_id[0])", to avoid a useless strlen call?
LGTM, I've posted v2 with that fix:
https://mail.openvswitch.org/pipermail/ovs-dev/2016-December/325807.html
Best regards, Ilya Maximets.
___
dev mailing list
d...@openvswi
The copy should be used here.
Fixes: 821b86649a90 ("netdev-dpdk: Don't try to unregister empty vhost_id.")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/netdev-dpdk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/netdev-dpdk.c b/lib/net
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dpif-netdev.c | 14 --
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 8e9a623..8af2811 100644
--- a/lib/dpif-netdev.c
+++ b/lib/dpif-netdev.c
@@ -3317,13 +3
e prepared on top of my previous
incremental + your change for PMD/ports resetting. I hope, it'll
be easy to apply them on top of your current version of code.
Also, I'd like to see v2 for the path-set to be able to review or
prepare new changes for it.
Best regards, Ilya Maximets.
Ilya Maximets (
On 03.12.2016 05:14, Daniele Di Proietto wrote:
> +/* Assign pmds to queues. If 'pinned' is true, assign pmds to pinned
> + * queues and marks the pmds as isolated. Otherwise, assign non isolated
> + * pmds to unpinned queues.
> + *
> + * The function doesn't touch the pmd threads, it just
[---]
Using of 'active_backup' algorithm will help to avoid such issues.
CC: Thadeu Lima de Souza Cascardo <casca...@redhat.com>
Fixes: 63460a30c53e ("tests/bundle: test bundle action with ports up and down")
Signed-off-by: Ilya M
On 19.12.2016 21:05, Bodireddy, Bhanuprakash wrote:
> Thanks Ilya and Aaron for reviewing this patch and providing your comments,
> my reply inline.
>
>> -Original Message-
>> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
>> Sent: Monday, December 19
too.
Best regards, Ilya Maximets.
On 16.12.2016 22:24, Aaron Conole wrote:
> Hi Bhanu,
>
> Bhanuprakash Bodireddy <bhanuprakash.bodire...@intel.com> writes:
>
>> In exact match cache processing on an EMC hit, packets are queued in to
>> batches matching the flow. The
On 20.12.2016 15:19, Bodireddy, Bhanuprakash wrote:
>> -Original Message-
>> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
>> Sent: Tuesday, December 20, 2016 8:09 AM
>> To: Bodireddy, Bhanuprakash <bhanuprakash.bodire...@intel.com>; Aaron
>>
ocumentation builds at all before
lowering the minimal sphinx version, because Ubuntu 12.04
contains sphinx-1.1.3.
My RHEL7 installation has python-sphinx-1.1.3 and python-docutils-0.11
with no build issues.
[1] http://docutils.sourceforge.net/HISTORY.html
[2] http://pac
better to
use same type as in dpdk library for consistency.
Also, magic number '-1' replaced with DPDK_ETH_PORT_ID_INVALID
macro.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/netdev-dpdk.c | 61 +++
1 file changed, 35 inse
Ilya Maximets (3):
netdev-dpdk: Fix double attaching of virtual devices.
netdev-dpdk: Fix device leak on port deletion.
netdev-dpdk: Use uint8_t for port_id.
Documentation/howto/dpdk.rst | 5 +-
lib/netdev-dpdk.c| 137 +++
2 files
()' to avoid double attaching of
the same device.
CC: Ciara Loftus <ciara.lof...@intel.com>
Fixes: 69876ed78611 ("netdev-dpdk: Add support for virtual DPDK PMDs (vdevs)")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/netdev-dpdk.c | 10 +-
1 file chang
'netdev-dpdk/detach' removed because not needed anymore.
CC: Ciara Loftus <ciara.lof...@intel.com>
Fixes: 55e075e65ef9 ("netdev-dpdk: Arbitrary 'dpdk' port naming")
Fixes: 69876ed78611 ("netdev-dpdk: Add support for virtual DPDK PMDs (vdevs)")
Signed-off-by: Ilya Maxime
Hi Sundar.
> Hi,
> I have an OVS bridge br0 with no NICs and 1 vhost user port which is
> connected to a VM. But ping fails between the VM and the br0 port, either
> way. The stats show zero all the time. Inside the VM, tcpdump shows nothing.
>
> This is with OVS 2.7.0 and DPDK 17.02.
hared between OVS and VM.
To make vhostuser work you must use 'share' option for qemu memory backing.
Please, refer the Documentation/topics/dpdk/vhost-user.rst for libvirt xml
example. "memAccess='shared'" - is what you need.
QEMU cmdline should contain something like this:
-obj
On 04.04.2017 11:29, Nadathur, Sundar wrote:
>> -Original Message-
>> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
>> Sent: Tuesday, April 4, 2017 12:07 AM
>> To: ovs-dev@openvswitch.org; Nadathur, Sundar
>> <sundar.nadat...@intel.com>
>>
and allows to remove painful external
dependency.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
Documentation/conf.py | 3 -
.../internals/contributing/documentation-style.rst | 10 ++-
Documentation/intro/install/windows.rst
regards, Ilya Maximets.
On 10.03.2017 10:47, Ilya Maximets wrote:
> On 10.03.2017 02:27, Ben Pfaff wrote:
>> On Thu, Mar 09, 2017 at 06:15:13PM +0300, Ilya Maximets wrote:
>>> On 07.03.2017 21:54, Ben Pfaff wrote:
>>>> Signed-off-by: Ben Pfaff <b...@ovn.org>
&
On 10.03.2017 02:27, Ben Pfaff wrote:
> On Thu, Mar 09, 2017 at 06:15:13PM +0300, Ilya Maximets wrote:
>> On 07.03.2017 21:54, Ben Pfaff wrote:
>>> Signed-off-by: Ben Pfaff <b...@ovn.org>
>>> Acked-by: Stephen Finucane <step...@that.guru>
>>> ---
0644
> index ..a67339bf2758
> --- /dev/null
> +++ b/Documentation/sphinx-version-blacklist
> @@ -0,0 +1,2 @@
> +code-block:: *ps1con
> +code-block:: *doscon
I don't feel this patch is fully correct, because it's not the features of
sphinx. And its version not really conne
-build v1.1.3' doesn't
support '--version' option.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
Note:
Marked as RFC because I don't know sphinx well.
Documentation/conf.py | 2 +-
Documentation/intro/install/windows.rs
On 02.03.2017 22:22, Aaron Conole wrote:
> Ilya Maximets <i.maxim...@samsung.com> writes:
>
>> This should be helpful for have all the logs in one place.
>> 'ovs-appctl vlog' commands for 'dpdk' module can be used
>> to configure the log level. Lower bound for DPDK
This should be helpful for have all the logs in one place.
'ovs-appctl vlog' commands for 'dpdk' module can be used
to configure the log level. Lower bound for DPDK logging
(--log-level) still can be passed through 'dpdk-extra' field.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
with
4096 flows. With higher number of flows there is no significant
difference with current implementation.
For the reference, number of EMC misses in current master is
around 20% for the case with 2048 active flows.
Testing performed with 100% EMC insert probability.
Signed-off-by: Ilya Maximets
appctl netdev-dpdk/detach to handle
> cases when del-port will not work.
>
> Fixes: 5dcde09c80a8 ("netdev-dpdk: Fix device leak on port deletion.")
> CC: Ilya Maximets <i.maxim...@samsung.com>
> Signed-off-by: Darrell Ball <dlu...@gmail.com&
tae82@samsung.com>, Ben Pfaff <b...@ovn.org>, Antonio
> Fischetti <antonio.fische...@intel.com>, Eelco Chaudron
> <echau...@redhat.com>, Ciara Loftus <ciara.lof...@intel.com>, Kevin Traynor
> <ktray...@redhat.com>, Darrell Ball <db...@vmware.co
to say if something else should be backported, but this
patch should fix described issue without any additional changes.
> On Wed, Jul 26, 2017 at 03:28:12PM +0300, Ilya Maximets wrote:
>> Hi.
>>
>> You need to backport at least following patch:
>>
>> commit 3
Ilya Maximets (2):
dpif-netdev: Decrease range of values for EMC probability.
dpif-netdev: Fix emc replacement policy.
lib/dpif-netdev.c| 33 +++--
vswitchd/vswitch.xml | 3 ++-
2 files changed, 25 insertions(+), 11 deletions(-)
--
2.7.4
On 31.07.2017 04:41, Darrell Ball wrote:
>
>
> -Original Message-
> From: <ovs-dev-boun...@openvswitch.org> on behalf of "Wang, Yipeng1"
> <yipeng1.w...@intel.com>
> Date: Friday, July 28, 2017 at 11:04 AM
> To: Ilya Maximets <i.maxim...@s
issue, because same uint32_t can be used to
store 2**20.
This patch implements solution #4.
CC: Ciara Loftus <ciara.lof...@intel.com>
Fixes: 4c30b24602c3 ("dpif-netdev: Conditional EMC insert")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
Infrastructure and
Hi Bhanuprakash,
Thanks for testing. Comments inline.
Best regards, Ilya Maximets.
On 01.08.2017 18:33, Bodireddy, Bhanuprakash wrote:
>> This patch-set inspired by [1] from Bhanuprakash Bodireddy.
>> Implementation of [1] looks very complex and introduces many pitfalls for
&
elated features in mail-list like rx queue
>>prioritization which are implemented in generic way on dpif-netdev layer.
>
> If you are referring to rxq prioritization work by Billy
> (https://mail.openvswitch.org/pipermail/ovs-dev/2017-July/336001.html),
> this feature is more
oper's Certificate of Origin.
2. I'm a fan of plain git history. Could we use 'Rebase and merge' policy
without merge commits ?
https://github.com/blog/2243-rebase-and-merge-pull-requests
PS: Ben, I'm sorry for the typo in your name in my previous letter.
Best regards, Ilya Maximets.
___
descriptions fixed to have consistent
and pretty log.
Best regards, Ilya Maximets.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
This allows to collect packets from more than one RX burst
and send them together with a configurable maximum latency.
'other_config:output-max-latency' can be used to configure
time that a packet can wait in output batch for sending.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
Cutlen already applied while processing OVS_ACTION_ATTR_OUTPUT.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/netdev-dpdk.c | 5 -
1 file changed, 5 deletions(-)
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index 8e3158f..ddcc574 100644
--- a/lib/netdev-dpdk.c
ded time based batching RFC patch.
* Fixed mixing packets with different sources in same batch.
Ilya Maximets (4):
dpif-netdev: Output packet batching.
netdev: Remove unused may_steal.
netdev-dpdk: Remove useless cutlen.
dpif-netdev: Time based output batching.
lib/dpif-netde
local and located in send port cache.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dpif-netdev.c | 104 ++
1 file changed, 82 insertions(+), 22 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index e2cd931..a
or example, 'emc_insert_min' which I'm introducing
in
the next patch is actually the characteristics of the last received packet
(maybe port)
but not the thread itself.
>
> BR, Jan
>
>> -Original Message-
>> From: ovs-dev-boun...@openvswitch.org
>> [mailto:ovs-d
On 10.08.2017 21:52, Bodireddy, Bhanuprakash wrote:
>>>
> } else {
> +/* If the queue is disabled in the guest, the
> corresponding qid
> + * map shall be set to OVS_VHOST_QUEUE_DISABLED(-2).
> + *
> +
On 11.08.2017 16:11, Bodireddy, Bhanuprakash wrote:
>> On 09.08.2017 15:35, Bodireddy, Bhanuprakash wrote:
>
> +static int
> +netdev_dpdk_vhost_tx_burst(struct netdev_dpdk *dev, int qid) {
> +struct dpdk_tx_queue *txq = >tx_q[qid];
> +struct rte_mbuf **cur_pkts =
dpif-netdev: dpcls per in_port with sorted subtables")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dpif-netdev.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index e2cd931..17e1666 100644
--- a/lib/d
-by: Ilya Maximets <i.maxim...@samsung.com>
---
Documentation/howto/dpdk.rst | 4 +-
NEWS | 2 +-
lib/dpif-netdev.c| 106 +--
tests/pmd.at | 7 ++-
vswitchd/vswitch.xml | 42 ++
possible modifications
in current logic.
This patch introduces new structure 'dp_netdev_pmd_thread_ctx'
contained by 'struct dp_netdev_pmd_thread' to store any processing
context of this PMD thread. For now, only time and cycles moved to
that structure. Can be extended in the future.
Signed-off-by: Ilya
Ilya Maximets (2):
dpif-netdev: Keep latest measured time for PMD thread.
dpif-netdev: Per-port conditional EMC insert.
Documentation/howto/dpdk.rst | 4 +-
NEWS | 2 +-
lib/dpif-netdev.c| 223 ++-
tests/pmd.at
Could you suggest another dp_packet_batch_XXX() name (which can be exposed
to the end user) for the function that doesn't check the boundaries instead
of 'add_unsafe', if you think it is not accurate?
Best regards, Ilya Maximets.
___
dev mailing l
it in output batch for sending.
>>
>> Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
>> ---
>> Notes:
>>
>> * This is an RFC and should not be used for performance testing.
>> * Millisecond granularity is used for now. Can be easily switched
&
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dpif-netdev.c| 16 +---
vswitchd/vswitch.xml | 5 +++--
2 files changed, 12 insertions(+), 9 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 0d78ae4..cf1591c 100644
--- a/lib/dpif-netdev.c
This function will provide monotonic time in microseconds.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/timeval.c | 22 ++
lib/timeval.h | 2 ++
2 files changed, 24 insertions(+)
diff --git a/lib/timeval.c b/lib/timeval.c
index dd63f03..be2eddc
> The per packets stats are presently overlapping in that
> miss stats include lost stats; make these stats non-overlapping
> for clarity and make this clear in the dp_stat_type enum. This
> also eliminates the need to increment two 'miss' stats for a
> single packet.
>
> The subtable lookup
On 14.08.2017 16:12, Jan Scheurich wrote:
>>> >From earlier in-house trials we know we need to target flush times of 50
>> us or less, so we clearly need better time resolution. Sub-ms timing in PMD
>> should be based on TSC cycles, which are already kept in the pmd struct.
>> Could you provide a
> The per packets stats are presently overlapping in that
> miss stats include lost stats; make these stats non-overlapping
> for clarity and make this clear in the dp_stat_type enum. This
> also eliminates the need to increment two 'miss' stats for a
> single packet.
>
> The subtable lookup
On 12.07.2017 10:15, Andy Zhou wrote:
> On Tue, Jul 11, 2017 at 7:30 AM, Ilya Maximets <i.maxim...@samsung.com> wrote:
>> If RSS hash exists in a packet it can be reused instead of
>> 5 tuple hash re-calculation in OVS_ACTION_ATTR_HASH. This
>> leads to increasing
method' to not depend on dp_hash value.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
Version 2:
* Removed assumption on hash_basis value.
* hash_finish replaced with hash_int as more appropriate.
* Fixed 'select group with dp_hash selection method' UT.
l
Fixes: 72c84bc2db23 ("dp-packet: Enhance packet batch APIs.")
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
lib/dp-packet.h | 7 ++-
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/lib/dp-packet.h b/lib/dp-packet.h
index 38282bd..c5fe32e 100644
--- a/lib/dp
Not a full review. And I most likely will not review the logic.
Just one comment inline.
Best regards, Ilya Maximets.
> One SNAT test is based on a single ping being successful;
> to make the result more predictable, static arp binding is now used.
> Occasionally, tracing shows the r
If RSS hash exists in a packet it can be reused instead of
5 tuple hash re-calculation in OVS_ACTION_ATTR_HASH. This
leads to increasing the performance of sending packets to
the OVS bonding in userspace datapath up to 10-15%.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
l
ux (not so hard). DPDK already mentioned
in commit message as the target and there is no need for misleading prefixes.
Best regards, Ilya Maximets.
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
is preferable to operate with a pmd on the 'wrong' numa node albeit with
>> lower performance. Local pmds are still chosen when available.
>>
>> Signed-off-by: Billy O'Mahony <billy.o.mah...@intel.com>
>> Signed-off-by: Ilya Maximets <i.maxim...@samsung.com&
On 10.07.2017 13:42, O Mahony, Billy wrote:
>
>
>> -Original Message-
>> From: Stokes, Ian
>> Sent: Monday, July 10, 2017 10:41 AM
>> To: Ilya Maximets <i.maxim...@samsung.com>; O Mahony, Billy
>> <billy.o.mah...@intel.com>; d...@openv
On 11.07.2017 05:10, Darrell Ball wrote:
>
>
> On 7/10/17, 12:41 AM, "Ilya Maximets" <i.maxim...@samsung.com> wrote:
>
> On 07.07.2017 21:09, Darrell Ball wrote:
> >
> >
> > On 7/6/17, 11:11 PM, "Ilya Maximets" <i.m
Patches 1 and 3 are also reviewed by Darrell now.
So, maybe we can apply them before we finish our holy war about patch #2 ?
We can easily apply #3 without #2 because they have nothing in common.
Best regards, Ilya Maximets.
On 05.07.2017 17:30, Ilya Maximets wrote:
> Hi all.
>
Ilya Maximets (4):
checkpatch: Don't allow Gerrit Change-Ids.
checkpatch: Print results while checking HEAD and stdin.
checkpatch: Allow checking more than one file.
checkpatch: Print commit hashes and names.
utilities/checkpatch.py | 64
Local Gerrit Change-Ids are not welcome in common repository.
Inspired by checkpatch.pl from Linux Kernel.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
utilities/checkpatch.py | 6 ++
1 file changed, 6 insertions(+)
diff --git a/utilities/checkpatch.py b/utilities/checkpa
-by: Ilya Maximets <i.maxim...@samsung.com>
---
utilities/checkpatch.py | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/utilities/checkpatch.py b/utilities/checkpatch.py
index 4a92890..7ccec51 100755
--- a/utilities/checkpatch.py
+++ b/utilities/checkpa
Currently, result status printed only for patch files.
It'll be nice to have results for other checking types.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
utilities/checkpatch.py | 33 +
1 file changed, 25 insertions(+), 8 deletions(-)
diff
It's better to see real commits instead of 'HEAD~n'.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
utilities/checkpatch.py | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/utilities/checkpatch.py b/utilities/checkpatch.py
index 7ccec51..f61b8b7
On 18.07.2017 06:48, Darrell Ball wrote:
>
>
> On 7/13/17, 8:07 AM, "ovs-dev-boun...@openvswitch.org on behalf of Ilya
> Maximets" <ovs-dev-boun...@openvswitch.org on behalf of
> i.maxim...@samsung.com> wrote:
>
> If RSS hash exists in a packet it c
On 07.07.2017 21:09, Darrell Ball wrote:
>
>
> On 7/6/17, 11:11 PM, "Ilya Maximets" <i.maxim...@samsung.com> wrote:
>
> On 07.07.2017 08:08, Darrell Ball wrote:
> >
> >
> > On 5/30/17, 7:12 A
So, about current patch-set:
The first two patches are localized to lib/conntrack* files and should
have 'conntrack' prefix as a module name. 3rd and 4th patches should have
'system-userspace-macros' and 'system-traffic' module names respectively.
Such names will be more accurate and conformed t
Same thing here. Not a full review.
Please, use 'system-traffic' for these patches as a prefix.
Details are in previous e-mail:
https://mail.openvswitch.org/pipermail/ovs-dev/2017-July/335751.html
Best regards, Ilya Maximets.
___
dev mailing list
d
y to move few hashes simultaniously to fit
constraint #3.
Implementation includes sorting of 'entries' to be able to collect
entries with accumulated load close enough to ideal value.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
I guess, the following tag should be correct, but n
Add regression test which checks rebalancing of tcp balanced bonding
after link state changes of one of the ports.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
To work correctly this unit test requires fix for '--len' option of
netdev-dummy/receive appctl command:
*
.
https://mail.openvswitch.org/pipermail/ovs-dev/2017-July/335930.html
Ilya Maximets (2):
bond: Fix broken rebalancing after link state changes.
ofproto-dpif.at: Add bonding down/up rebalancing test.
ofproto/bond.c| 128 +++---
tests/ofproto
threads. All not affected
threads will forward traffic without any additional latencies.
id-pool created for static tx queue ids to keep them sequential
in a flexible way. non-PMD thread will always have
static_tx_qid = 0 as it was before.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.
the old/unwanted entries.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
Acked-by: Cian Ferriter <cian.ferri...@intel.com>
Tested-by: Cian Ferriter <cian.ferri...@intel.com>
---
lib/dpif-netdev.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/l
number
of wanted tx queues to avoid unnecessary reconfigurations.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
Tested-by: Ian Stokes <ian.sto...@intel.com>
Acked-by: Ian Stokes <ian.sto...@intel.com>
---
lib/dpif-netdev.c | 26 +-
1 file changed, 2
the new one?
If everybody will use their own preferable prefixes, git history
will become a total mess. And that is the main concern.
Once again, It's only my opinion and you're free to disagree.
Best regards, Ilya Maximets.
> /
>
> commit 9d3e0e5c196c0a91ea23d8d9254b14
will also speed up handling of the new flows and
statistics accounting.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
ofproto/bond.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/ofproto/bond.c b/ofproto/bond.c
index cb25a1d..72b373c 100644
--- a/ofproto/
On 24.07.2017 22:53, Andy Zhou wrote:
> On Mon, Jul 24, 2017 at 9:23 AM, Ilya Maximets <i.maxim...@samsung.com> wrote:
>> On 23.07.2017 00:02, Darrell Ball wrote:
>>>
>>>
>>> -Original Message-
>>> From: <ovs-dev-boun...@openvswitch.org
-by: Andy Zhou <az...@ovn.org>
Signed-off-by: Andy Zhou <az...@ovn.org>
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
ofproto/bond.c | 16 +---
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/ofproto/bond.c b/ofproto/bond.c
index cb25a1d..e4d4b65 10
az...@ovn.org>
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
ofproto/bond.c | 12 ++--
1 file changed, 2 insertions(+), 10 deletions(-)
diff --git a/ofproto/bond.c b/ofproto/bond.c
index e4d4b65..e09136e 100644
--- a/ofproto/bond.c
+++ b/ofproto/bond.c
@@ -175,8
e not used for flow
balancing.
Fix the documentation accordingly.
Signed-off-by: Ilya Maximets <i.maxim...@samsung.com>
---
I think, this should be applied to some stable branches too.
vswitchd/vswitch.xml | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/vswitc
Version 2:
* Became a patch-set.
v1:
https://mail.openvswitch.org/pipermail/ovs-dev/2017-July/336078.html
* Removed bond_hash_tcp.
* 2 new patches added:
* Documentation fix.
* Remove bond_hash_src.
Ilya Maximets (3
On 24.07.2017 22:40, Andy Zhou wrote:
> On Mon, Jul 24, 2017 at 6:33 AM, Ilya Maximets <i.maxim...@samsung.com> wrote:
>> On 22.07.2017 01:38, Andy Zhou wrote:
>>> On Wed, Jul 19, 2017 at 7:51 AM, Ilya Maximets <i.maxim...@samsung.com>
>>> wrote:
>>>
1 - 100 of 6783 matches
Mail list logo