Just curious - Why is ipfire keeping ReiserFS over Reiser4

[sphericalumbra]

I have read some history and am well aware that ipfire has dropped support for Reiser4 in the past due to Hans Reiser’s murder crime but that was so many years ago already and I can see that Reiser4 is still actively being maintained so why not start to include it back?

I’m asking this cause it’s just that ReiserFS is really solid and I feel like I could benefit from using its more developed descendant - not unless I’m missing a big part of the picture here like really good reasons as to why I should stick with ReiserFS instead of going for Reiser4 (or Reiser5 in the future).

[arne_f]

ReiserFS is upstream in mainline kernel but we dont recommend it.

At the time we have dropped Reiser4 it was not really in development anymore and the patchset was buggy so we have decided to remove it. Also currently it is outdated. (5.8.1 is the latest supported kernel but we plan to update to 5.10 in the next.
If someone want Reiser4 back he had to bring it upstream into the mainline kernel.

[ms (Michael Tremer)]

Hello,

I appreciate your post, but I just need to get a couple of things straight:

@sphericalumbra

I have read some history and am well aware that ipfire has dropped support for Reiser4 in the past due to Hans Reiser’s murder crime but that was so many years ago already and I can see that Reiser4 is still actively being maintained so why not start to include it back?

We didn’t drop it because of his conviction of murder - actually, this technically does not really matter. In an open source project, we have code from thousands of people. Statistically some of them have committed crimes, but we cannot rip out everything those people wrote. Technology and those things don’t go together. I do of course not want to downplay at all what Hans Reiser did.

@sphericalumbra

I’m asking this cause it’s just that ReiserFS is really solid and I feel like I could benefit from using its more developed descendant - not unless I’m missing a big part of the picture here like really good reasons as to why I should stick with ReiserFS instead of going for Reiser4 (or Reiser5 in the future).

As Arne already said, Reiser4 was an amazing file system at that time. Robust and fast. What else do we want from a file system?

The biggest problem for us is, that it is being maintained out of tree. With every kernel update, we will have to wait until someone has ported Reiser4 and we rely on that patch being maintained for that release we are on.

In practise it hasn’t worked well for us to rely on other people like that (see grsecurity) and so I am not very tempted to merge another out-of-tree patch again.

Regarding Reiser5, I saw the announcement over a year ago. It is a bit sad, that they did not rename it to something more neutral to distance themselves more from Hans Reiser, but that is of course not my decision.

For political reasons, it will become quite hard again to get it merged into the mainline kernel I would predict. And secondly, it seems that the hunger for new file systems has disappeared. Underlying storage is fast enough that the FS usually isn’t the bottleneck any more. ext4 and XFS which I would regard as the most commonly used file systems are good enough for everybody.

ReiserFS is there for historic reasons. It is absolutely outdated, and again, nobody has requested more choice here in a while. I wouldn’t mind to drop support for it actually.

[sphericalumbra]

Thank you for this reply of yours that really helps me see the big picture as to why it isn’t included in IPFire and welp, I guess they’re keeping the name in respect to the original author regardless of the crimes that he has committed.

If anything, for a filesystem to be rejected from the mainline kernel merely because the name still connects to someone who made some mistake in the real world is a really childish stance in my opinion (specially when he absolutely has no connection to the FS development anymore) which is really dismaying if you ask me.