On linux-next breakage

Posted Dec 15, 2025 9:39 UTC (Mon) by fratti-co (subscriber, #175548)
Parent article: Best practices for linux-next

> Torvalds repeated that he would like to see consequences when buggy commits break linux-next, that the guilty repository needs to be kicked out — at least temporarily. Once the relevant maintainer has acknowledged and fixed the problem, the repository would be allowed back in. If the problem is not fixed, though, then the repository should be kept out and not pulled during the next merge window.

I'm with Torvalds here. There seems to be somewhat of an indifference by some subsystem maintainers towards them having broken linux-next, and adding some consequences for their actions would probably help with that. I have come across broken linux-next builds before because someone pulled in their co-workers patch series at v1 without compile-testing it. I've also seen people apply broken code, be told it's broken, and then say they'll fix it "next week" instead of backing out of the change immediately so that next day's linux-next is no longer causing massive amounts of kernel console spam that will make any system logging to anything but memory unhappy.

It's not that I expect production-level quality from linux-next, but it's somewhat annoying when every next tag that contains a change you want to base on also happens to not build or run because some other subsystem decided to make linux-next completely unusable for several days.

to post comments

On linux-next breakage

Posted Dec 15, 2025 10:04 UTC (Mon) by sfr (subscriber, #7761) [Link] (2 responses)

There are already consequences. I use the version of the offending branch that was in the previous day's
linux-next. And keep doing that until it is fixed. I used to just drop it, but occassionally that had roll on effects
to later merges. It is very unusual for such a breakage to last more that a couple of days, and I can't remember
off the top of my head of such a breakage not being fixed before the merge window opens. I do not publish
a linux-next tree with a known build failure in it. That does not mean that there are no unknown build failures
as it is impossible to build for all configs on all architectures.

I have even, on occasson, reverted to using an old version of a branch when a failure is reported from other's
testing.

On linux-next breakage

Posted Dec 15, 2025 10:20 UTC (Mon) by fratti-co (subscriber, #175548) [Link] (1 responses)

What's the best way to report breakage to linux-next, as opposed to the subsystem maintainer? The build failure I have in mind in particular was likely overlooked because it was on arm64+clang. I don't recall how many days it took to disappear. Specifically, next-20251023 was broken because of the OPP subsystem, but I don't recall which precise commit it was.

On linux-next breakage

Posted Dec 16, 2025 1:37 UTC (Tue) by sfr (subscriber, #7761) [Link]

Email me (or Mark after January 16) and cc linux-next at vger.kernel.org