On Wed, Nov 22, 2023 at 2:28 PM Ales Musil wrote:
> Build the image before every job to allow more
> fine-grained dependency pinning. This is especially
> useful for stable branches as they might need to stay
> on specific distribution or Python version.
>
> Signed-off-by: Ales Musil
> ---
> .c
The DPDK unit test only runs if vfio or igb_uio kernel modules are loaded:
on systems with only mlx5, this test is always skipped.
Besides, the test tries to grab the first device listed by dpdk-devbind.py,
regardless of the PCI device status regarding kmod binding.
Remove dependency on this DPDK
Add more conditions to the image prepare process.
This allows us to test the prebuilt images for both
Fedora and Ubuntu.
Run the weekly image on main branch with the
use of Fedora for the weekly runs.
Reported-at: https://issues.redhat.com/browse/FDP-149
Signed-off-by: Ales Musil
---
.github/wo
Build the image before every job to allow more
fine-grained dependency pinning. This is especially
useful for stable branches as they might need to stay
on specific distribution or Python version.
Signed-off-by: Ales Musil
---
.cirrus.yml| 40 -
.g
Switch the Cirrus CI to use the Ubuntu image,
so it is unified with the GH actions. That is
useful for various dependency pinnings.
Signed-off-by: Ales Musil
---
.cirrus.yml | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/.cirrus.yml b/.cirrus.yml
index 2d3824972..71e9afbb6
This series allows different branches to use different images.
This is needed for stabikity reasons as it is hard to keep up
with new Ubuntu/Fedora releases for all stable branches.
With the last patch we will still use the prebuilt images
for main branch which is completely fine.
Ales Musil (3):
Sorry for the nosie, there is a wrong branch reference on 2/3 I'll post v2.
On Wed, Nov 22, 2023 at 2:25 PM Ales Musil wrote:
> This series allows different branches to use different images.
> This is needed for stabikity reasons as it is hard to keep up
> with new Ubuntu/Fedora releases for all
Build the image before every job to allow more
fine-grained dependency pinning. This is especially
useful for stable branches as they might need to stay
on specific distribution or Python version.
Signed-off-by: Ales Musil
---
.cirrus.yml| 40 +-
.gith
Switch the Cirrus CI to use the Ubuntu image,
so it is unified with the GH actions. That is
useful for various dependency pinnings.
Signed-off-by: Ales Musil
---
.cirrus.yml | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/.cirrus.yml b/.cirrus.yml
index 2d3824972..71e9afbb6
Add more conditions to the image prepare process.
This allows us to test the prebuilt images for both
Fedora and Ubuntu.
Run the weekly image on main branch with the
use of Fedora for the weekly runs.
Reported-at: https://issues.redhat.com/browse/FDP-149
Signed-off-by: Ales Musil
---
.github/wo
This series allows different branches to use different images.
This is needed for stabikity reasons as it is hard to keep up
with new Ubuntu/Fedora releases for all stable branches.
With the last patch we will still use the prebuilt images
for main branch which is completely fine.
Ales Musil (3):
Since DPDK v23.11, it is possible to select more easily which optional
library is enabled.
OVS needs the vhost library (and its dependencies).
The net/tap DPDK driver needs the gso library.
Other optional library can be disabled.
This reduces the cache entry for DPDK from ~7MB to ~4MB.
Signed-of
Rather than save the whole DPDK sources and build artefacts, checkout
sources in a separate directory and build DPDK there.
Only the installed artefacts are then going to the cache.
Example sources in the share/dpdk installed directory can be pruned too.
This makes a (v23.11-rc3) DPDK cache entry
On Tue, Nov 21, 2023 at 02:26:50PM -0500, Mike Pattrick wrote:
> Currently when userspace-tso is enabled, netdev-linux interfaces will
> indicate support for all offload flags regardless of interface
> configuration. This patch checks for which offload features are enabled
> during netdev construct
On Tue, Nov 21, 2023 at 02:26:51PM -0500, Mike Pattrick wrote:
> From: Flavio Leitner
>
> Currently OVS will calculate the segment size based on the
> MTU of the egress port. That usually happens to be correct
> when the ports share the same MTU, but that is not always true.
>
> Therefore, if th
On Tue, Nov 21, 2023 at 02:26:52PM -0500, Mike Pattrick wrote:
> From: Flavio Leitner
>
> This provides a software implementation in the case
> the egress netdev doesn't support segmentation in hardware.
>
> The challenge here is to guarantee packet ordering in the
> original batch that may be f
16 matches
Mail list logo