Comment 22 for bug 1947099

Thank you for the detailed additional information. I think I understand the situation properly now, together with the need for the fix.

> That means that if something is currently accepting DHCP offers after the specified timeout has expired, then that is buggy behaviour.

We sometimes run into the case where we're concerned about users being regressed by the very fix for the buggy behaviour itself. I think we need to take a pragmatic view in these cases. But unless there's some specific scenario we're concerned about, I think it's OK to not worry about the general case. As you say, if there's a user who is specifying a 2 second timeout but then relying on a DHCP reply coming after 3 seconds, then we will regress them but I think this is reasonable under the circumstances. I think it's sufficient to have considered and documented this case carefully, as you now have done (thanks!)

We also need to be careful to ensure that the code that we ship in the archive isn't itself already making such an assumption. But I think you've sufficiently covered that already.

It might be worth verifying that this is exactly the result of the change being landed here though. During SRU verification, in addition to the test plan you've already specified, would it be practical and/or useful to verify that, with a timeout of 2s, a DHCP reply sent after 1.5s works, but a DHCP reply sent after 2.5s does not? This would help us ensure that we are introducing the exact change we intend to introduce.

I would also like you to verify your actual failing user story please - that your boot time does actually speed up with that hardware.

With those two requests added to your Test Plan I am in principle fine to proceed. I still need to review your patch in detail but I've unfortunately run out of time for that today. I'd also like Łukasz's ack please.

If another SRU team member looks again before I do, feel free to take over and accept without my further input if you feel that's appropriate.