On 2022-01-31 10:50, David Ahern wrote:

On 1/31/22 5:54 AM, Jamal Hadi Salim wrote:

quoted

quoted

quoted

Do you need a patch for that in some documentation?

How about adding some comments to README.devel?

Sure - but it wont be sufficient IMO.
Best course of action is for the maintainers to remind people to run
tests.

Above, you said the tests were meant for bots.

The invocation that tdc.sh makes is targeted for the bots.
It only tests actions and qdiscs.
If you run  ./tdc.py -h you'll see more options.

quoted

BTW: We found out that Stephen's patches still break the latest -next.

ugh. I committed them after running tdc.sh and not seeing a change in
output. We'll need fixup patches then.

Clearly some work is needed on getting the test suite usable by a wider
audience. I am new to running those tests as well and probably had some
pilot errors running them.

One of the issues is sometimes some of the tests take a while to
complete, so some tweaking of the timer may be needed depending on your
setup etc. Probably that is what you may be running into.

I do wait for ACKs from tc folks, but can't wait forever. Right now
there is the 'skip_hw' and 'skip_sw' patch that Victor sent a Tested-by.
When I apply that v2 patch, I see errors in the tdc.sh output so my
mileage varies from Victor's. There is also the v5 of this set which I
have no applied yet; it could use some acks and tdc testing as well.

ok. Punting to Victor for testing the v5 - was it posted on the list?
I will Ack Bowen's patch.

cheers,
jamal