reiser4 in 2.6.15?

[wetter]

Hi there,
i just found this in the changelog for Linux v2.6.15-rc1

Author: Hans Reiser <reiser@namesys.com>
Date: Mon Nov 7 00:59:29 2005 -0800

[PATCH] reiser4: add radix_tree_lookup_slot()

Reiser4 uses radix trees to ...

does this mean that reiser4 is going to go vanilla with 2.6.15? i didn't see any big fuss on kerneltraffic or elsewhere. hmm.

[1U]

Wow! So it's finally making it's way into the vanilla sources? It's about time. I hope this isn't some kind of a mixup. Now I won't have to make custom LiveCDs and use unstable sources to get the advantages of this great file system. Can anyone else confirm if this is actually going to happen?

This will unleash a lot of flaming from all the neophobic developers (lovechild and most of the devs here for example).

[Voltago]

Wow! So it's finally making it's way into the vanilla sources?

No. Read the whole thing:

Code: Select all

commit a43313668f62a06e14c915b8c8994fc8a1257394
Author: Hans Reiser <reiser@namesys.com>
Date:   Mon Nov 7 00:59:29 2005 -0800

    [PATCH] reiser4: add radix_tree_lookup_slot()
    
    Reiser4 uses radix trees to solve a trouble reiser4_readdir has serving nfs
    requests.
    
    Unfortunately, radix tree api lacks an operation suitable for modifying
    existing entry.  This patch adds radix_tree_lookup_slot which returns pointer
    to found item within the tree.  That location can be then updated.
    
    Both Nick and Christoph Lameter have patches which need this as well.
    
    Signed-off-by: Andrew Morton <akpm@osdl.org>
    Signed-off-by: Linus Torvalds <torvalds@osdl.org>

It is at best a preparation for reiser4, not the thing itself.

[1U]

Well that sucks, it should have been added a long time ago. At this pace, Microsoft will support reiser4 sooner than the high wizard neophobic developers which make these decisions.

[Voltago]

What's the problem of patching your kernel if you want reiser4?

[1U]

What's the problem of simply including it in kernels? If people want to use it or not that's their choice to make, not the developers. The problem is they are creating inconveniences for everyone who wants to use it based on their own opinions and hatred of new things. The needs of the many outweigh the needs of the few (ignorant neophobics).

[Conan]

Yea... neophobic... that's it!

More like sane.

ricer4 (note the misspelling) has had more problems with it in the beginning of it's inception than any other file system ever. This creates a bias.

Hans reiser (note the no misspelling) has fought kernel coding standards forever or so. This is why it has taken so long to get included, he has to keep rewriting things to make the kernel guru's happy. Blame him, not the " cainotophobic developers" (see! I can use dictionary.com also)

[amne]

So, you called the kernel and Gentoo devs neophobics 3 times. Any other reason to include reiser4 in the kernel?

The needs of the many outweigh the needs of the few (ignorant neophobics).

Maybe the many need a stable and reliable file system that will surely be maintained by someone. The current situation doesn't seem to guarantee that the kernel devs seem to be willing/able maintain the current version of reiser 4. Hans Reiser otoh doesn't seem to maintain reiser3 any more but the kernel devs. Scary enough to me to switch to ext3 on future file systems even though i always was happy with reiser3.

[1U]

Doesn't Hans Reiser maintain Reiser4? I don't see any real maintenance work that has to be done by the kernel developers. Look at the nitro-sources patchset, just simply include the Reiser4 patches in the regular kernel and that's all there's to it.

[curtis119]

What's the problem of simply including it in kernels? If people want to use it or not that's their choice to make, not the developers. The problem is they are creating inconveniences for everyone who wants to use it based on their own opinions and hatred of new things. The needs of the many outweigh the needs of the few (ignorant neophobics).

This is why it is possible to patch your kernel in the first place. So that you DO have a choice. The kernel devs could just as easily make it impossible (or really really difficult) to apply third party patches if they so desired but they don't. They don't inlude it in the kernel proper yet because to many people are still reporting errors with it. Besides, Linus and the kernel crew are working with Hans Reiser as we speak to clean it up and get it included. If you search the LKML you will find a whole list of things they have given to Hans to fix so that he can get his code included.

[j-m]

What's the problem of simply including it in kernels? If people want to use it or not that's their choice to make, not the developers. The problem is they are creating inconveniences for everyone who wants to use it based on their own opinions and hatred of new things. The needs of the many outweigh the needs of the few (ignorant neophobics).

The needs of a couple of ricers don't make up for the support nightmare for the devs. Go patch the kernel yourself and don't file bugs when things break.

[1U]

Who says the devs are forced to support every feature in the kernel? If they don't want to they don't have to support other's code, such as Reiser4. I don't understand why there's this strict "If we include it we'll HAVE to help EVERY person who uses it and we're goign to have a nightmare!! The gentoo project and kernel will die all because of the insane support from this one little feature... ohh nooo" attitude towards experimental software.

Fyi, I don't file bugs when things break so don't accuse me of doing so. There is nothing "ricer" about wanting good performance out of your system. If performance was ricing, then all Linux users are just ricers. Let's all go back to Windows XP and enjoy the "efficiency" of it's design. And don't start with the instability problems either, I've never had any Reiser4 problems. It works as stable as any other fs I've tried in Linux while being much faster.

[spb]

Look at the nitro-sources patchset, just simply include the Reiser4 patches in the regular kernel and that's all there's to it.

No it's not. Someone has to take care of forward-porting all the in-kernel filesystem drivers when the (internal) filesystem API changes. That's a non-trivial workload, and every filesystem added without a good reason only increases it. Not to mention the amount of trouble that reiser4 caused the external patchset maintainers last time it changed -- IIRC they were without it for several months because noone could port the thing. It's not worth holding up new kernel releases for such a thing.

[lbrtuk]

What's the problem of simply including it in kernels? If people want to use it or not that's their choice to make, not the developers.

This argument could be used to include any piece of crap in the kernel. It's Linus' tight meritocracy over such things that has made the kernel so reliable over the years. Crap just simply won't go in. No matter how hard people complain or who it is doing the complaining.

[spb]

There is nothing "ricer" about wanting good performance out of your system.

No, the rice factor comes from trying to get better performance, and ending up with something that's both slower and less stable.

And yes, I've tried it. My current system with a complete hardened setup (pie/ssp, PaX, and SELinux, all of which add not insignificant overhead, especially to filesystem operations in the latter case) on exclusively ext3 filesystems is faster, more responsive, and more stable than the reiser4 non-hardened system was.

[amne]

Doesn't Hans Reiser maintain Reiser4? I don't see any real maintenance work that has to be done by the kernel developers.

I was talking about reiser3 and he does not really seem to actively maintain that any more.

[Conan]

At the risk of getting in trouble, you really need to learn what you are talking about.

Who says the devs are forced to support every feature in the kernel? If they don't want to they don't have to support other's code, such as Reiser4. I don't understand why there's this strict "If we include it we'll HAVE to help EVERY person who uses it and we're goign to have a nightmare!! The gentoo project and kernel will die all because of the insane support from this one little feature... ohh nooo" attitude towards experimental software.

There is in implication, that if one downloads software from a company, that company is responsible for maintaining said software. This implication holds true here. The kernel needs to know that software added will be maintained for a while to come, and based on Hans Reiser's attitude to reiser3 (screw it!); there is good chance that when ricer5 comes out (the OMGTHISISEVENFASTERTHANBEFOREITBOOTSINLIEK3SEKONDS) Version, reiser4 will be left in the dust. If this is to happen, the other developers need to feel comfortable enough about the features to maintain it.

And fyi, I don't file bugs when things break so don't accuse me of doing so.

<sarcastic>thanks for the help..</sarcastic>

There is nothing "ricer" about wanting good performance out of your system.

There is something "ricer" about sacrificing stability for performance.

If performance was ricing, then all Linux users are just ricers. Let's all go back to Windows XP and enjoy the "efficiency" of it's design.

Err, ya...

And don't start with the instability problems either, I've never had any Reiser4 problems. It works as stable as any other fs I've tried in Linux while being much faster.

Good for you. Now what about all the other people who have had complaints?

[Voltago]

If people want to use it or not that's their choice to make, not the developers. The problem is they are creating inconveniences for everyone who wants to use it based on their own opinions and hatred of new things.

Well, there's a little more to it.
a) Reiser4 doesn't integrate with the Linux VFS layer. It needs a lot of extra care in its interaction with the rest of the kernel.
b) Hans Reiser stopped at some point to maintain reiserfs3. Other linux devs had to take over. There's concern it might happen again.
c) Mr. Reiser by all accounts isn't very cooperative at meeting the requirements for inclusion in the kernel.
d) If reiser4 is included into mainstream and stuff breaks and data is lost (like it did when reiserfs3 was included, for example), people will come whining to the kernel devs, not to namesys.com. Including not-quite-mature software could damage the kernels reputation for stability. Consider a noob. He doesn't know the next thing about the kernel, but wants to be a 1337 h4x0r and chooses use reiser4 for rootfs. Some time later: System crashes, data loss, emails lost, documents gone, everything lost. Now our noob might feel a bit annoyed if not positively pissed. But wait, there's helpful advice we could give him: It's his own fault because it was his own choice. Now how do you think our noob feels? Sometimes the clueless have to be protected from themselves.
e) The kernel devs carry a lot of responsibility nowadays and generally don't deserve to be trash-talked like this.

The needs of the many outweigh the needs of the few

Who are these may? Can you show them to me?

[1U]

This argument could be used to include any piece of crap in the kernel. It's Linus' tight meritocracy over such things that has made the kernel so reliable over the years. Crap just simply won't go in. No matter how hard people complain or who it is doing the complaining.

So you mean to say nitro-sources and others are just pieces of crap? It runs better than the regular kernel on my machine, and my friend even got his sound working in it when the stock kernel without the "pieces of crap" always had problems.

[1U]

I'm not sure what you people were doing with nitro-sources and reiser4, but it must have been your fault. It runs far better on my system and is very stable. Sure the releases aren't perfect and sometimes there are experimental versions, but once you get the final release it can keep running forever. I've had a nitro-sources .12 system run for 90 days and it could have kept going.

[Conan]

This argument could be used to include any piece of crap in the kernel. It's Linus' tight meritocracy over such things that has made the kernel so reliable over the years. Crap just simply won't go in. No matter how hard people complain or who it is doing the complaining.

So you mean to say nitro-sources and others are just pieces of crap? It runs better than the regular kernel on my machine, and my friend even got his sound working in it when the stock kernel without the "pieces of crap" always had problems.

I mean to say nitro-sources are less stable then vanilla kernels, and less supported when things break (whcih they do frequently...)

[j-m]

Who says the devs are forced to support every feature in the kernel? If they don't want to they don't have to support other's code, such as Reiser4. I don't understand why there's this strict "If we include it we'll HAVE to help EVERY person who uses it and we're goign to have a nightmare!! The gentoo project and kernel will die all because of the insane support from this one little feature... ohh nooo" attitude towards experimental software.

Won't be in supported kernels until it's accepted upstream. Use an unsupported kernel that's already patched or patch it yourself. Period.

And fyi, I don't file bugs when things break so don't accuse me of doing so. There is nothing "ricer" about wanting good performance out of your system. If performance was ricing, then all Linux users are just ricers. Let's all go back to Windows XP and enjoy the "efficiency" of it's design.

Ah sure, you definitely need a good performace on /boot (like in the example I used)... Also, it's much more important to have a "good performance" than to have data stored in a safe way, apparently.

[1U]

My data is as safe as yours, you're not the ones that use it so why keep making up false statements about how it works on MY machine? This is NOT ricing because I don't sacrifice any performance, stability, or anything. I simply get a BETTER system, not much to it.

What's wrong with Hans cutting off ReiserFS? If he wants to move onto newer projects then good for him. Things evolve, and there's no point of supporting old systems. It's not hard to update a partition. Crying about ReiserFS being left behind is like crying about having little support for 486 systems (though actually there's plenty of that useless junk still in the kernel). Nobody cares, it's junk, move on and quit wasting time/resources on maintaining useless things.

[j-m]

What's wrong with Hans cutting off ReiserFS? If he wants to move onto newer projects then good for him. Things evolve, and there's no point of supporting old systems.

Sure, and dump the support burden on kernel devs instead. Excellent!

It's not hard to update a partition.

Uh eh? There's no way to convert reiserfs to reiser4.

Crying about ReiserFS being left behind is like crying about having little support for 486 systems (though actually there's plenty of that useless junk still in the kernel). Nobody cares, it's junk, move on and quit wasting time/resources on maintaining useless things.

Proves the stupid ricers point, pretty much.

[Voltago]

My data is as safe as yours, you're not the ones that use it so why keep making up false statements about how it works on MY machine? This is NOT ricing because I don't sacrifice any performance, stability, or anything. I simply get a BETTER system, not much to it.

As long as you believe this, fine. But don't press your choice on other people. As you see, most people here object to the inclusion of reiser4 into mainstream. The needs of the many and so on...