On Wed, Aug 23, 2023 at 5:35 PM David Marchand
wrote:
>
> Integrate system-traffic.at tests as part of check-dpdk.
CI (thinking of Intel robot) other than GHA might not be too happy
about this change.
It is hard to tell what fails in the report I received.
A safer approach is probably to define
On Thu, Aug 10, 2023 at 5:45 AM Dumitru Ceara wrote:
>
> It's similar to the processing we do for address sets. There's a bit
> more mechanics involved due to the fact that we need to split NB port
> groups per datapath.
>
> We currently only partially implement incremental processing of
> port_g
Add config option called "northd-backoff-interval-ms" that allows
to delay northd engine runs capped by the config option.
When the config option is set to 0 or unspecified, the engine
will run without any restrictions. If the value >0 we will delay
northd engine run by the previous run time capped
It doesn't affect the allocation size, but the number of elements
should go before the size of these elements.
Fixes: 74daa0607c7f ("ovn-northd: Introduce parallel lflow build")
Signed-off-by: Ilya Maximets
---
northd/northd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/n
We need to allocate memory for 'n_router_dps' pointers, but the
current code allocates space for 'n_router_dps' datapaths, which
is 762 byte structure. Having 5000 routers in the OpenStack
setup, we allocate 762 * 5000 * 5000 ~= 17 GB of memory instead
of 190 MB:
98.17% (19,962,746,247B) (heap a
Current implementation of 'ref_chassis' recompute is horribly slow.
For each port binding it finds LR group, then it adds a chassis of the
current port binding to each HA chassis group of this LR group.
The number of ports is much greater than number of chassis. And the
number of distinct LR grou
As reported by Ales, net/tap has broken support for checksum offloading.
Fixes have been sent to the DPDK side, but waiting for the fixes, simply
disable all conntrack related tests and some IPv6 tunnel tests.
Signed-off-by: David Marchand
---
tests/system-dpdk-macros.at | 9 +
1 file ch
Integrate system-traffic.at tests as part of check-dpdk.
DPDK net/tap specific logs are added to the log exclusions.
Some tests that can't work with the userspace datapath are skipped by
overriding some OVS_CHECK_* macros.
ADD_VETH is implemented with a tap interface in the kernel, directly
poll
Align system-dpdk existing helpers to other common OVS helpers so they
can accept some optional arguments.
Introduce a OVS_DPDK_STOP_VSWITCHD wrapper around OVS_VSWITCHD_STOP to
catch dpdk related logs in a centralised fashion.
Signed-off-by: David Marchand
---
tests/system-dpdk-macros.at | 21
Some unit tests expect that a OVS port has an associated netdevice on
which they can hook tc or tcpdump.
This will not be possible when testing the userspace datapath with DPDK.
Introduce two helpers (which will be overriden in system-dpdk tests) and
use them in the existing tests.
Signed-off-by:
dpdk-testpmd does not need hugetlbfs backing as we don't require
multiprocess support in OVS unit tests.
Plus, there is no need for explicitly reserving more memory than what
testpmd would actually need at runtime.
Switch to --in-memory, use dynamic allocations and remove the then
unneeded check o
Let's enhance our coverage in the CI and run DPDK system tests.
A few DPDK drivers are enabled in DPDK compilation.
Put DPDK build in $PATH for dpdk-testpmd to be available.
sudo drops PATH= updates and -E alone does not seem to preserve this
variable.
Pass PATH=$PATH when running the tests, as a
Rather than copy/paste everywhere, introduce helpers to control
testpmd runs.
Rely on --stats-period (which outputs port stats every n seconds) so that
testpmd keeps running without expecting any user input.
Signed-off-by: David Marchand
Acked-by: Aaron Conole
---
Changes since v1:
- fixed OVS_D
Quoting Paul Ganssle, setuptools maintainer:
* The setuptools project has stopped maintaining all direct
invocations of setup.py years ago, and distutils is deprecated.
There are undoubtedly many ways that your setup.py-based system is
broken today, even if it's not failing loudly or
setup.py upload is now deprecated. When used, pypi.org returns an error:
Upload failed (400): Invalid value for blake2_256_digest. Error: Use
a valid, hex-encoded, BLAKE2 message digest.
Use twine which is the recommended replacement tool to upload on
pypi.org.
Link: https://blog.ganssle.io/
v2:
* Keep code in the python dir but rename it so that it does not clash
with `python -m build`.
* Remove duplication in makefiles.
Robin Jarry (3):
python: Rename build related code to ovs_build_helpers.
python: Use twine to upload sdist package to pypi.org.
python: Use build to generat
The python/build folder contents are completely unrelated to the ovs
python bindings. These files are only used during the build for various
subsystems (docs, man pages, code generation, etc.).
Having that folder in that location prevents from running:
cd python && python3 -m build
Which
On 8/18/23 11:51, Nobuhiro MIKI wrote:
> Fixes: 30b79363b967 ("Makefile.am: Add clang static analysis support")
> Signed-off-by: Nobuhiro MIKI
> ---
> tests/.gitignore | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/tests/.gitignore b/tests/.gitignore
> index 83b1cb3b489e..3a8c45975645
On 8/21/23 15:53, Frode Nordahl wrote:
> From: Colin Watson
>
> If we omit this, groff 1.23.0 warns:
>
> tbl preprocessor failed, or it or soelim was not run; table(s) likely
> not rendered (TE macro called with TW register undefined)
>
> Reported-by: Lucas Nussbaum
> Reported-at: https://
On 8/21/23 15:53, Frode Nordahl wrote:
> Starting with groff 1.23.0 a warning is produced if the tbl
> preprocessor is not run. A side effect of enabling it is that
> new warnings on table formatting is printed.
>
> As requested during the review [0] of a series [1] attempting to
> address this,
On 8/22/23 15:34, Eelco Chaudron wrote:
>
>
> On 22 Aug 2023, at 12:35, Ilya Maximets wrote:
>
>> It should generally be possible to build OVS with just C99.
>> We do need some POSIX extensions, e.g. support for RW locks,
>> but we should not require any GNU extensions or features
>> from later
Bleep bloop. Greetings Paolo Valerio, I am a robot and I have tried out your
patch.
Thanks for your contribution.
I encountered some error that I wasn't expecting. See the details below.
checkpatch:
ERROR: Author hepeng needs to sign off.
WARNING: Unexpected sign-offs from developers who are
From: hepeng
The patch avoids the extra allocation for nat_conn.
Currently, when doing NAT, the userspace conntrack will use an extra
conn for the two directions in a flow. However, each conn has actually
the two keys for both orig and rev directions. This patch introduces a
key_node[CT_DIRS] mem
23 matches
Mail list logo