On Thu, Jun 15, 2023 at 11:14:53AM +0200, Dmitry Vyukov wrote: > On Thu, 15 Jun 2023 at 01:38, Dave Chinner <david@xxxxxxxxxxxxx> wrote: > > Hence the claim that "because syzbot doesn't run we don't have > > visibility of code bugs" is naive, conceited, incredibly > > narcissistic and demonstratable false. It also indicates a very > > poor understanding of where syzbot actually fits into the overall > > engineering processes. > > Hi Dave, Ted, > > We are currently looking into options of how to satisfy all parties. > > I am not saying that all of these bugs need to be fixed, nor that they > are more important than bugs in supported parts. And we are very much > interested in testing supported parts as well as we can do. > > By CONFIG_INSECURE I just meant something similar to kernel taint > bits. How is that any better? Who gets to decide what sets this taint? Subsystem maintainers? > A user is free to continue after any bad thing has happened/they > did, but some warranties are void. And if a kernel developer receives > a bug report on a tainted kernel, they will take it with a grain of > salt. So it's important to note the fact and inform about it. > Something similar here: bugs in deprecated parts do not need to be > fixed, and distros are still free to enable them, but this fact is > acknowledged by distros and made visible to users. "Deprecated" does not mean *unmaintained*. They are two completely different things, and conflating the two does not help anyone. You are talking about marking unmaintained code with a taint, not deprecated code. reiserfs is unmaintained code. ntfs3 is unmaintained code. hfs is unmaintained code. There are several other unmaintained filesystems that don't have active maintainers. Bug reports never, ever get looked at, etc. Sure, there's an argument to taint them. But whilst XFS v4 is deprecated, it is still very much maintained. We encourage people to move off V4, we focus less on it, but if bug reports from users come in, we still fix them. So even if you introduce some "unmaintained" taint for the kernel, you aren't going to see it get set for the XFS V4 format. > But we are looking into other options that won't require even CONFIG_INSECURE. Who is this nebulous "we"? -Dave. -- Dave Chinner david@xxxxxxxxxxxxx
