Hello Jakub,
On Thu, Jul 24, 2025 at 06:24:27PM -0700, Jakub Kicinski wrote:
On Wed, 23 Jul 2025 10:35:06 -0700 Breno Leitao wrote:
Extend the `check_for_dependencies()` function in `lib_netcons.sh` to check whether IPv6 is enabled by verifying the existence of `/proc/net/if_inet6`. Having IPv6 is a now a dependency of netconsole tests. If the file does not exist, the script will skip the test with an appropriate message suggesting to verify if `CONFIG_IPV6` is enabled.
This prevents the test to misbehave if IPv6 is not configured.
IDK. I think this is related to some of the recent patches?
Yes, commit 3dc6c76391cbe (“selftests: net: Add IPv6 support to netconsole basic tests”) introduced IPv6 support to the netconsole basic tests.
Because the NIPA config enables IPv6, the tests pass in that environment. However, if the tests are run somewhere without IPv6 support such as in a test I was doing regarding another patch, they will fail, when it should be skipped.
The context would be helpful in the commit message.
Apologies for not including more context in the commit message.
Otherwise, as networking people, I think we are obligated to respond with hostility to "IPv6 may not be enabled"..
As for handling systems without IPv6, if IPv6 isn’t available, the intention is for the test to be skipped. That’s exactly what this patch addresses.
I did consider making the test adaptable so it would just run with whichever protocol (IPv4 or IPv6) is present, but rejected that approach. Allowing the test to “pass” in such cases doesn’t really demonstrate meaningful coverage, since the test isn’t actually being exercised as intended.
In short, it seems more appropriate to skip the test entirely if all conditions aren’t met, so, you know that your .config needs adjustment.
Thanks for your review, --breno