mbox series

[RFC,for-faizal,0/4] tc-taprio selftests

Message ID 20231221132521.2314811-1-vladimir.oltean@nxp.com (mailing list archive)
Headers show
Series tc-taprio selftests | expand

Message

Vladimir Oltean Dec. 21, 2023, 1:25 p.m. UTC
Unfortunately the isochron version from debian will not work on veth
pairs, I had to make some changes and add the --omit-hwts option to
the sender and receiver programs. The modified version is on this branch
here, I will merge the changes to 'master' once I have enough confidence
in them.
https://github.com/vladimiroltean/isochron/tree/omit-hwts

For testing the tc-taprio software scheduling path, we don't need PTP
synchronization or hardware timestamps anyway.

This is just a skeleton that I'm hoping Faizal can pick up and extend
with more test cases for dynamic schedule changes. It should run
primarily on the veth driver, but should behave the same on any network
driver as well, including those who also have tc-taprio offload.

Vladimir Oltean (4):
  selftests: net: forwarding: allow veth pairs to be created with
    multiple queues
  selftests: net: tsn: push --txtime out of isochron_do()
  selftests: net: tsn: allow isochron_do() to skip sync monitoring on
    sender too
  selftests: net: tsn: add tc-taprio test cases (WIP)

 .../selftests/drivers/net/ocelot/psfp.sh      |   1 +
 .../testing/selftests/net/forwarding/Makefile |   1 +
 tools/testing/selftests/net/forwarding/lib.sh |   3 +-
 .../selftests/net/forwarding/tc_taprio.sh     | 143 ++++++++++++++++++
 .../selftests/net/forwarding/tsn_lib.sh       |  70 +++++++--
 5 files changed, 205 insertions(+), 13 deletions(-)
 create mode 100755 tools/testing/selftests/net/forwarding/tc_taprio.sh

Comments

Hangbin Liu Dec. 22, 2023, 6:28 a.m. UTC | #1
On Thu, Dec 21, 2023 at 03:25:17PM +0200, Vladimir Oltean wrote:
> Unfortunately the isochron version from debian will not work on veth
> pairs, I had to make some changes and add the --omit-hwts option to
> the sender and receiver programs. The modified version is on this branch
> here, I will merge the changes to 'master' once I have enough confidence
> in them.
> https://github.com/vladimiroltean/isochron/tree/omit-hwts
> 
> For testing the tc-taprio software scheduling path, we don't need PTP
> synchronization or hardware timestamps anyway.
> 
> This is just a skeleton that I'm hoping Faizal can pick up and extend
> with more test cases for dynamic schedule changes. It should run
> primarily on the veth driver, but should behave the same on any network
> driver as well, including those who also have tc-taprio offload.

Tested-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>