status of reiser4

status of reiser4

[Hans Reiser]

Vladimir is working on one last bug that occurs if copy from user page 
faults on the user space process's buffer, and then we are going to 
bundle a set of patches together, send them to andrew morton, and then 
start a discussion on the changes to core kernel code that Hellweg and 
company object to that currently prevent our inclusion.

hans

Re: status of reiser4

[Jake Maciejewski]

Have there been any reports other than Isaac's of reiser4 working on
AMD64? It still doesn't work for me, and I'd like to know if I'm the
only one having problems.

On Fri, 2004-12-17 at 10:18 -0800, Hans Reiser wrote:
> Vladimir is working on one last bug that occurs if copy from user page 
> faults on the user space process's buffer, and then we are going to 
> bundle a set of patches together, send them to andrew morton, and then 
> start a discussion on the changes to core kernel code that Hellweg and 
> company object to that currently prevent our inclusion.
> 
> hans
-- 
Jake Maciejewski <maciejej@msoe.edu>

Re: status of reiser4

[Maciej Soltysiak]

Hi,

> bundle a set of patches together, send them to andrew morton, and then 
> start a discussion on the changes to core kernel code that Hellweg and 
> company object to that currently prevent our inclusion.
That is nice.
Are we going to have patches for vanilla 2.6.10 for non -mm users?
I mean patches trivial to apply. Because the -mm tree contains lots of
r4 patches. And applying them in correct order may be troublesome.
Users really like one patch to be applied without unknown dependencies.


Maciej Soltysiak

Re: status of reiser4

[Vladimir Saveliev]

Hello

On Fri, 2004-12-17 at 21:54, Jake Maciejewski wrote:
> Have there been any reports other than Isaac's of reiser4 working on
> AMD64? It still doesn't work for me, and I'd like to know if I'm the
> only one having problems.
> 

What kernel are you using?

> On Fri, 2004-12-17 at 10:18 -0800, Hans Reiser wrote:
> > Vladimir is working on one last bug that occurs if copy from user page 
> > faults on the user space process's buffer, and then we are going to 
> > bundle a set of patches together, send them to andrew morton, and then 
> > start a discussion on the changes to core kernel code that Hellweg and 
> > company object to that currently prevent our inclusion.
> > 
> > hans

Re: status of reiser4

[maciejej]

I tried 2.6.10-rc2-mm4 with Zam's bitmap.c patch and 2.6.9 with the
version -3 patch. I also tried compiling with GCC 3.3 rather than 3.4,
using reiser4progs 1.0.3, and running in loopback mode, but nothing has
worked yet.

> Hello
>
> On Fri, 2004-12-17 at 21:54, Jake Maciejewski wrote:
>> Have there been any reports other than Isaac's of reiser4 working on
>> AMD64? It still doesn't work for me, and I'd like to know if I'm the
>> only one having problems.
>>
>
> What kernel are you using?
>
>> On Fri, 2004-12-17 at 10:18 -0800, Hans Reiser wrote:
>> > Vladimir is working on one last bug that occurs if copy from user page
>> > faults on the user space process's buffer, and then we are going to
>> > bundle a set of patches together, send them to andrew morton, and then
>> > start a discussion on the changes to core kernel code that Hellweg and
>> > company object to that currently prevent our inclusion.
>> >
>> > hans
>
>

Re: status of reiser4

[mjt]

On Fri, Dec 17, 2004 at 12:54:08PM -0600, Jake Maciejewski wrote:
>Have there been any reports other than Isaac's of reiser4 working on
>AMD64? It still doesn't work for me, and I'd like to know if I'm the
>only one having problems.

I recently purchased an AMD64 setup.
Here's the catch; I did so because my old box went belly-up and now
it seems the GFX card is broken (The box shut itself down and doesn't
boot if the GFX card has its extra power cable connected, regardless
of the amount of other hardware. I hate hardware.)

Anyway, due to severe space limitations, I couldn't back my home up
so I'm hoping it'll work long enough for me to wiggle some stuff
somewhere else.

My plan for testing is to use Reiser4 in (at least) an x86 chroot,
which should hit amd64 bugs nonetheless, because the only difference
is how the software is compiled, yes?

Anything specific I should look out for? Nothing I have in my home
directory is _that_ critical for me to get out _now_, but I would
like to get it some day :>

Thanks guys!

-- 
mjt

Re: status of reiser4

[Sander]

Jake Maciejewski wrote (ao):
> Have there been any reports other than Isaac's of reiser4 working on
> AMD64? It still doesn't work for me, and I'd like to know if I'm the
> only one having problems.

It sort of works for me on a dual opteron on top of raid10 on top op
scsi. This is 64bit and kernel 2.6.10-rc3-mm1. It survives bonnie++ but
I re-installed one of the two identical servers with reiser3 because I
got fs corruption (for whatever reason) which fsck could not fix. It
said the fs was just fine after several fsck runs with different
parameters (which it told me to run with), but at least several libs
were corrupted.

I'm sorry I don't have more info on this.

status of reiser4

[Hans Reiser]

Ron Joffe wrote:

>On Tuesday 25 October 2005 14:13, Hans Reiser wrote:
>  
>
>>If yes, what steps did you take to do it?
>>    
>>
>
>Not quite there yet, but would like to test it out.
>
>Hans, could you give a status of Reiser4. I have seen allot of different 
>issues on the maillist, but not a "status" of where the overall project 
>stands.
>
>Thanks,
>
>Ron
>
>  
>
vs is working on a major update that he will release any day now.  It
contains many bug fixes.  It should deturkey our latest reiser4 release,
sorry about that guys.

Nate did a mongo benchmark on an older reiser4, -mm1 or some such, and
the surprising thing we discovered was that reiser4 is now more cpu
efficient than ext3 for mongo.  I guess all those memory copy avoiding
balancing algorithms that made the  code so hairy and added so much to
the coding time are now paying off.;-)  We should do some detailed
profiling, and careful benchmark comparisons over time, and see if
memory copies are now the big remaining hogs and reiser4 is being more
space efficient in RAM thus going faster, or if ext3 is just doing
something wasteful.

We should do fresh benchmarks on the latest reiser4 also.