Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Reiser4 to gentoo-dev-sources. Poll!
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  

Do you need Reiser4?
It would be great to have it with gentoo-dev-sources and/or gentoo-sources
65%
 65%  [ 367 ]
It is enough if it works with other sources.
15%
 15%  [ 88 ]
I don't need Reiser4 at all, so I don't care.
18%
 18%  [ 103 ]
What is Reiser4?
0%
 0%  [ 2 ]
Total Votes : 560

Author Message
ciaranm
Retired Dev
Retired Dev


Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding

PostPosted: Wed Mar 02, 2005 11:43 pm    Post subject: Reply with quote

Mystilleef wrote:
Yes, and Hans responded to those issues. What's the real deal?

He didn't respond to them at all. He posted some rhetoric back that totally ignored said issues. If you've ever tried reiser4 on non-x86 it becomes rather obvious that the whole thing relies far too much upon a set of invalid assumptions.
Back to top
View user's profile Send private message
Jake
Veteran
Veteran


Joined: 31 Jul 2003
Posts: 1132

PostPosted: Thu Mar 03, 2005 12:02 am    Post subject: Reply with quote

ciaranm wrote:
Mystilleef wrote:
Yes, and Hans responded to those issues. What's the real deal?

He didn't respond to them at all. He posted some rhetoric back that totally ignored said issues. If you've ever tried reiser4 on non-x86 it becomes rather obvious that the whole thing relies far too much upon a set of invalid assumptions.

It has nothing to do with invalid assumptions. The problems are a lack of experience writing 64-bit clean code and not having test machines.
Back to top
View user's profile Send private message
codergeek42
Bodhisattva
Bodhisattva


Joined: 05 Apr 2004
Posts: 5142
Location: Anaheim, CA (USA)

PostPosted: Thu Mar 03, 2005 12:05 am    Post subject: Reply with quote

Jake wrote:
It has nothing to do with invalid assumptions. The problems are a lack of experience writing 64-bit clean code and not having test machines.
From what I've read, It's not just x86-64. It's _anything_ non-x86 (PPC, Sparc, etc.) that Reiser4 doesn't play nicely with yet.
_________________
~~ Peter: Programmer, Mathematician, STEM & Free Software Advocate, Enlightened Agent, Transhumanist, Fedora contributor
Who am I? :: EFF & FSF
Back to top
View user's profile Send private message
ciaranm
Retired Dev
Retired Dev


Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding

PostPosted: Thu Mar 03, 2005 12:12 am    Post subject: Reply with quote

Jake wrote:
It has nothing to do with invalid assumptions. The problems are a lack of experience writing 64-bit clean code and not having test machines.

Oh so wrong. The problem is that their code sucks and they don't understand what they're doing -- a lot of the code was just "keep on changing stuff until it looks like it works". As for test machines, they've turned down all kinds of offers of kit.
Back to top
View user's profile Send private message
Lokheed
Veteran
Veteran


Joined: 12 Jul 2004
Posts: 1295
Location: /usr/src/linux

PostPosted: Thu Mar 03, 2005 2:17 am    Post subject: Reply with quote

ciaranm wrote:
it's not going to happen.


They also said that about computers getting the size of toasters, that you would NEVER need more than a 32MB hardrive, and that while telephones are great for the US, Britain would never need them as they have enough messenger boys...

Are you always such a ray of sunshine?
_________________
You're not afraid of the dark are you?
Back to top
View user's profile Send private message
gentoo_lan
l33t
l33t


Joined: 08 Sep 2004
Posts: 891
Location: Charles Town, WV

PostPosted: Thu Mar 03, 2005 7:08 am    Post subject: Reply with quote

I was just wondering ciaranm when you thought Reiser 3.6 will be production ready if ever. I know you have a very low opinion of Reiser 4 is your opinion just as low of Reiser 3.6?
Back to top
View user's profile Send private message
gcasillo
l33t
l33t


Joined: 23 Sep 2003
Posts: 739
Location: Cincinnati, Ohio, USA

PostPosted: Thu Mar 03, 2005 7:16 am    Post subject: Reply with quote

codergeek42 has it right. When the vanilla kernel supports it, its good to go. Until then, forget it.

There are things I can live without: experimental filesystems are among them. I recently downgraded all of my boxes (three at home, seven at work) from ~x86 keywords to non-keywords. Nothing but stable packages for me. Sanity has been restored, and productivity has risen.

Nothing against those on the bleeding edge. Just that I'd rather not have to manage issues like flaky filesystems because of mob rule. If those of you who desire it need the speed kick that reiserfs-4, you can always go your own way. Gentoo probably isn't what you're looking for.
Back to top
View user's profile Send private message
Jake
Veteran
Veteran


Joined: 31 Jul 2003
Posts: 1132

PostPosted: Thu Mar 03, 2005 8:11 am    Post subject: Reply with quote

gentoo_lan wrote:
I was just wondering ciaranm when you thought Reiser 3.6 will be production ready if ever. I know you have a very low opinion of Reiser 4 is your opinion just as low of Reiser 3.6?

ciaranm wrote:
Yes it can. Concensus amongst those who actually know what they're talking about is that reiser4 should be taken out and shot, and that reiser3 still isn't production ready. Of course, the screaming fanboy crowd disagrees, but most of them haven't even heard of the vfs layer, let alone know how to look at filesystem code.
Back to top
View user's profile Send private message
TheCoop
Veteran
Veteran


Joined: 15 Jun 2002
Posts: 1814
Location: Where you least expect it

PostPosted: Thu Mar 03, 2005 2:24 pm    Post subject: Reply with quote

well, from what I can tell, the major thing about reiser4 is that its Different to the previous filesystems, and those differences break a few programs/kernel things. Do you think that the current filesystem way of doing things will last forever? Yes, reiser4 has a few bugs, but what software doesnt? (I cant comment on the non-x86 issues, as I dont have the necessary technical knowledge to do so). Just because its Different doesnt mean its Wrong or Broken.
_________________
95% of all computer errors occur between chair and keyboard (TM)

"One World, One web, One program" - Microsoft Promo ad.
"Ein Volk, Ein Reich, Ein Führer" - Adolf Hitler

Change the world - move a rock
Back to top
View user's profile Send private message
ciaranm
Retired Dev
Retired Dev


Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding

PostPosted: Thu Mar 03, 2005 2:28 pm    Post subject: Reply with quote

TheCoop wrote:
well, from what I can tell, the major thing about reiser4 is that its Different to the previous filesystems, and those differences break a few programs/kernel things. Do you think that the current filesystem way of doing things will last forever? Yes, reiser4 has a few bugs, but what software doesnt? (I cant comment on the non-x86 issues, as I dont have the necessary technical knowledge to do so). Just because its Different doesnt mean its Wrong or Broken.

No no, it's utterly broken, and the aforementioned lkml thread explains why this is the case. As for the bugs, I'd seriously hope that my filesystem doesn't have any, at least of the "randomly lose all data, kill the box and generally screw things up hideously" kind.
Back to top
View user's profile Send private message
TheCoop
Veteran
Veteran


Joined: 15 Jun 2002
Posts: 1814
Location: Where you least expect it

PostPosted: Thu Mar 03, 2005 2:43 pm    Post subject: Reply with quote

I wouldnt say utterly broken, reiser4 has a lot of new ideas that break some things. Maybe they would be better as a general fs-plugin (eg .tar-traversal, files sharing bits between themselves), but no other filesystem has the right hooks therefore they're reiser4-only. There are problems that can be fixes (O_DIRECTORY fr instance) and some more insidious, but overall Hans has some very nice and revolutionary ideas for filesystems. They could have been done better, but how else do you get change? Again, just because its Different doesnt mean its Wrong. You werent expecting the current design to last forever did you?
_________________
95% of all computer errors occur between chair and keyboard (TM)

"One World, One web, One program" - Microsoft Promo ad.
"Ein Volk, Ein Reich, Ein Führer" - Adolf Hitler

Change the world - move a rock
Back to top
View user's profile Send private message
ciaranm
Retired Dev
Retired Dev


Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding

PostPosted: Thu Mar 03, 2005 2:48 pm    Post subject: Reply with quote

TheCoop wrote:
Again, just because its Different doesnt mean its Wrong. You werent expecting the current design to last forever did you?

I'm not saying it's wrong because it's different. I'm saying it's wrong for the reasons that have already been explained in great depth on lkml.
Back to top
View user's profile Send private message
nife
Tux's lil' helper
Tux's lil' helper


Joined: 24 Apr 2003
Posts: 87

PostPosted: Thu Mar 03, 2005 3:56 pm    Post subject: Reply with quote

ciaranm wrote:
TheCoop wrote:
Again, just because its Different doesnt mean its Wrong. You werent expecting the current design to last forever did you?

I'm not saying it's wrong because it's different. I'm saying it's wrong for the reasons that have already been explained in great depth on lkml.


But most of that thread focuses on where to put xattrs, either in the VFS or in the file system. It doesn't deal with it being bad code or loosing data. And it seems to be a very two sided argument. Yes Hans seems to be stooping to some childish stuff, but the technical merits are still debated as the main topic. I haven't read the entire thing yet, but still nothing has been posted about non-x86 arch. Since it looks like its moving to -mm soon I think its time to stop trashing it and time to start getting used to the fact that its comming.

http://kerneltrap.org/node/2761

I'm sure there are problems with other arches, but lets get real. x86 is the most used arch, so its easiest to fix bugs in that architecture. Other architectures will become stable later.

With all of that said I don't think it should be in the sources yet. So I vote no, not yet.
Back to top
View user's profile Send private message
firephoto
Veteran
Veteran


Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"

PostPosted: Thu Mar 03, 2005 4:22 pm    Post subject: Reply with quote

What gets me is that it can't go into the kernel because it's "broke", but I can go through my kernel config and find a dozen or more options that are "broke" if one was to turn them on and I suspect this is why this options have a big "EXPERIMENTAL" next to them. Just because it's there doesn't mean you have to use it.

My feeling is that all the lack of love for reiser4 is from people who don't user reiserfs and in some way or another don't believe or support what Hans and namesys are doing.

It's pretty obvious that the gentoo-sources will never have a reiser4 patch and almost certainly there will never be a kernel in the portage tree that has a reiser4 patch. If the patches are buggy, and they don't apply cleanly, or it causes problems to the system even if you were to never enable any reiser4 options in the kernel then these would be good reasons to not have support but I don't think these are issues on newer kernels.
Back to top
View user's profile Send private message
spb
Retired Dev
Retired Dev


Joined: 02 Jan 2004
Posts: 2135
Location: Cambridge, UK

PostPosted: Thu Mar 03, 2005 4:50 pm    Post subject: Reply with quote

firephoto wrote:
My feeling is that all the lack of love for reiser4 is from people who don't user reiserfs and in some way or another don't believe or support what Hans and namesys are doing.
Not to mention the people who use stuff that gets completely broken by certain braindead design decisions that ReiserFS made.
Back to top
View user's profile Send private message
firephoto
Veteran
Veteran


Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"

PostPosted: Thu Mar 03, 2005 4:55 pm    Post subject: Reply with quote

spb wrote:
firephoto wrote:
My feeling is that all the lack of love for reiser4 is from people who don't user reiserfs and in some way or another don't believe or support what Hans and namesys are doing.
Not to mention the people who use stuff that gets completely broken by certain braindead design decisions that ReiserFS made.


The whole point is that just because it's there you don't have to use it. No one is holding gun to your head telling you to compile reiser4 support into your kernel just because it might be there.....

I've never used ntfs write support because it says right in the kernel that it will probably screw up your ntfs file system if you're not careful. I read the description, I understood the possible consequences, and I chose not to enable ntfs write support.
Back to top
View user's profile Send private message
IWBCMAN
Guru
Guru


Joined: 25 Jun 2002
Posts: 474

PostPosted: Thu Mar 03, 2005 9:35 pm    Post subject: Reply with quote

For what it's worth I have been using Reiser4 for almost a year on my main machine and 8 months on my laptop. Have never had any problems whatsoever-in my experience it belongs to the most stable fs's I have ever used. I am somewhat frustrated that it still does not support ACL's -and am eagerly awaiting this.

I do not run it on the production LTSP server that I administer. There has been a history of strange problems using Reiser3 with LTSP which is why I use ext3 for that server.

Although I am eagerly awaiting it's inclusion in the gentoo-sources kernel I expect a number of issues need to be ironed out prior to this happening. Andrew will likely put it into the kernel fairly soon-once this happens the *real* testing will begin-and with sufficent users the code should improve rapidly as it is exposed to the real world conditions. Hopefully once this transpires some of the realy extraodinary things which compose Reiser4 can be abstracted into the proper kernel vfs enabling all fs's to profit from the advances which it represents.

Even if it was in the kernel right now I would still hesitate to have it included in the gentoo-sources so long as it does not properly support ACL's and of course so long as there are outstanding bugs in it as regards x86-64 and other architectures.

I think it's a fair policy of gentoo-sources to demand that those fs's included properly support things like ACL's .

But then again the reason I like and exclusivley use Gentoo is that it gives me the option to do that which I want easily. There are several quality alternative kernels sources which do include it(Nitro, already mentioned)-I haven't used gentoo-sources for a long, long time and I can't honestly say I have missed it.
Back to top
View user's profile Send private message
ninjabadger
Apprentice
Apprentice


Joined: 23 Apr 2003
Posts: 270
Location: Scotland

PostPosted: Thu Mar 03, 2005 9:38 pm    Post subject: Reply with quote

There is debate as to how stable it is,

If the devs put it in gentoo-dev-sources they will have to support it and answer bug fixes on it etc.

This takes away their valuable time from doing other gentoo stuff

Its enough that other sources have it and you can patch it yourself anyway.

Once its been through Morton then maybe, but i think the gentoo devs hav the right idea (as well as the right) here
_________________
Is Microsoft Stronger?

No, no, no. Quicker, easier, more seductive.
Back to top
View user's profile Send private message
brjames
n00b
n00b


Joined: 30 Sep 2003
Posts: 52

PostPosted: Thu Mar 03, 2005 10:05 pm    Post subject: Reply with quote

nife wrote:
ciaranm wrote:
TheCoop wrote:
Again, just because its Different doesnt mean its Wrong. You werent expecting the current design to last forever did you?

I'm not saying it's wrong because it's different. I'm saying it's wrong for the reasons that have already been explained in great depth on lkml.


But most of that thread focuses on where to put xattrs, either in the VFS or in the file system. It doesn't deal with it being bad code or loosing data. And it seems to be a very two sided argument. Yes Hans seems to be stooping to some childish stuff, but the technical merits are still debated as the main topic. I haven't read the entire thing yet, but still nothing has been posted about non-x86 arch. Since it looks like its moving to -mm soon I think its time to stop trashing it and time to start getting used to the fact that its comming.

http://kerneltrap.org/node/2761

I'm sure there are problems with other arches, but lets get real. x86 is the most used arch, so its easiest to fix bugs in that architecture. Other architectures will become stable later.

With all of that said I don't think it should be in the sources yet. So I vote no, not yet.


I thought it was already in -mm? Did it get taken back out or something?

I also havent read through the entire reiser4 thread (that thing is HUGE), but from what I have read, Linus seems to actually be in favor of files as directories, which seems to be causing most of the problems.

ciaranm, your attitude of "No! Never! It will never work!" seems a little childish. I'm definitely not using Reiser4 and won't for some time, as it obviously has issues that need to worked out. But there is a lot of work going on to fix those issues and im sure they will get fixed. Maybe next week (Hahah!) maybe next year, maybe a few years down the road. Now maybe you've actually gone over the Reiser4 code and have personally seen such issues and how they're unfixable, but I won't buy it until somebody who is very intimately familiar with the Reiser4 code (i.e. a Reiser4 developer) says its impossible to fix.

Until then I think a more reasonable response would be "Gentoo has no interest in supporting Reiser4 and will not until it is accepted by the upstream kernel devs."
Back to top
View user's profile Send private message
Glorandar
n00b
n00b


Joined: 15 Jun 2003
Posts: 64
Location: Vancouver, BC, Canada

PostPosted: Fri Mar 04, 2005 12:02 am    Post subject: IMO the reasons why Reiser4 isn't supported by gentoo-source Reply with quote

Quote:
The question is where or what in Reiser4 is so horrible?
IMO, its not an issue about "horrible". Its just not supported by the mainline 2.6 kernel.

Filesystems are required to be very stable. That is because users (us) hate it when our data has been corrupted. We tend to yell at folks, like the developers, about such things.

Thus, in general, its a bad idea for the developers to commit themselves to supporting an unstable filesystem.

Why do I think reiser4 is unstable? Its not in the mainline kernel... ergo its unstable.

As others have stated, reiser4 is currently available in forms, other than gentoo-sources.

However, caveat emptor, its your data on your filesystem. Me? I'm going to be sticking to filesystems that are proven to be stable.

In closing, developers, please keep up the good work.

I fully support the decision to not support filesystems, such as reiser4, until they are incorporated into the mainline kernel. I'd be happy to wait until they are no longer marked as "experiemental". When it comes to my data, I'm paranoid. Maybe that is because I am a 16+ year software developing professional.
_________________
----- Glorandar
Back to top
View user's profile Send private message
spb
Retired Dev
Retired Dev


Joined: 02 Jan 2004
Posts: 2135
Location: Cambridge, UK

PostPosted: Fri Mar 04, 2005 12:20 am    Post subject: Re: IMO the reasons why Reiser4 isn't supported by gentoo-so Reply with quote

Glorandar wrote:
Why do I think reiser4 is unstable? Its not in the mainline kernel... ergo its unstable.
While I agree with the rest of your points, that one's wrong. ;) PaX/Grsecurity is a prime counterexample.
Back to top
View user's profile Send private message
firephoto
Veteran
Veteran


Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"

PostPosted: Fri Mar 04, 2005 12:22 am    Post subject: Reply with quote

A snippet of the ntfs write support help from the kernel gives a little history on what kind of file systems got included in the kernel. ;)
Code:

                                 NTFS write support
Note:  While write support is safe in this version (a rewrite from
scratch of the NTFS support), it should be noted that the old NTFS
write support, included in Linux 2.5.10 and before (since 1997),
is not safe.

Now that was the old ntfs write support it mentions and the new one is obviously better.

And also there's some not so friendly file systems that can be enabled in our current vanilla kernel.
Code:

<*> UFS file system support (read only)
 [ ]   UFS file system write support (DANGEROUS) (NEW)


Now it's pretty obvious one would have to pretty dense to enable a file system like that for their root partition and expect it to work and also expect to get help when things break or go awol. So it can reasonably expect that someone who enables a kernel option that has big words like EXPERIMENTAL or DANGEROUS next to them that they take the risks into their own hands. ;)
Back to top
View user's profile Send private message
fuji
Tux's lil' helper
Tux's lil' helper


Joined: 26 Apr 2002
Posts: 111

PostPosted: Fri Mar 04, 2005 4:37 am    Post subject: Reply with quote

I wouldn't touch Reiser4 at all... the only thing Reiser 3 touches is /usr/portage -- IE: something that is easily replaced.

If people want Reiser4, they're probably preformance guys so why would they use gentoo-sources which has now become a stable patchset as opposed to the preformance patchset which it once was. Those guys should be using some of the other sets like Nitro or love or patch their own.

There's no reason for the gentoo-dev-sources to support reiser4.
_________________
Came for the hype, stayed for Portage.
Back to top
View user's profile Send private message
petrjanda
Veteran
Veteran


Joined: 05 Sep 2003
Posts: 1557
Location: Brno, Czech Republic

PostPosted: Fri Mar 04, 2005 6:04 am    Post subject: Reply with quote

I dont care, but then Im not gonna touch Reiser4 again for at least a year. Its too unstable. Its probably not gonna make the vanilla source for another year and half.
_________________
There is, a not-born, a not-become, a not-made, a not-compounded. If that unborn, not-become, not-made, not-compounded were not, there would be no escape from this here that is born, become, made and compounded. - Gautama Siddharta
Back to top
View user's profile Send private message
firephoto
Veteran
Veteran


Joined: 29 Oct 2003
Posts: 1612
Location: +48° 5' 23.40", -119° 48' 30.00"

PostPosted: Mon Mar 14, 2005 5:11 am    Post subject: Reply with quote

Maybe someone should tell Eugenia over at OSNews about reiser4 so she can add it to her list of things that users want but deaf devs ignore. :D

... or not. ;)

just gettin' your goats. :lol:
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3, 4, 5, 6, 7  Next
Page 3 of 7

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum