Bug (?) with 2.6.15-mm4

Bug (?) with 2.6.15-mm4

[Andrea Gelmini]

Hi all,
	a few line, without logs and in depth info, to describe a problem
	(if it's not already known, I will do a real bug report).
	
	Well, I've got a Satellite P20, with /dev/md1 crypto-loopbacked
	with /dev/loop0 (losetup -e aes), and /dev/loop0 mount on / with
	reiser4.
	With
Linux cogno 2.6.15-mm2 #1 PREEMPT Sun Jan 8 14:27:42 CET 2006 i686 GNU/Linux
	everything works fine since the day I compiled it.
	Same config, but -mm4, I've got freeze. Sometimes after minutes,
	sometimes after hours.
	After last crash I booted from a rescue partition. Mounting
	partition where system was writing gives no warning, but checking
	it with fsck.reiser4 (Debian testing, package reiser4progs version 1.0.5-1)
	two inconsistency were founded (and fixed).

Thanks a lot for your time and your work,
gelma

Re: Bug (?) with 2.6.15-mm4

[Joe Feise]

Andrea Gelmini wrote on 01/16/06 10:22:

> Hi all,
> 	a few line, without logs and in depth info, to describe a problem
> 	(if it's not already known, I will do a real bug report).
> 	
> 	Well, I've got a Satellite P20, with /dev/md1 crypto-loopbacked
> 	with /dev/loop0 (losetup -e aes), and /dev/loop0 mount on / with
> 	reiser4.
> 	With
> Linux cogno 2.6.15-mm2 #1 PREEMPT Sun Jan 8 14:27:42 CET 2006 i686 GNU/Linux
> 	everything works fine since the day I compiled it.
> 	Same config, but -mm4, I've got freeze. Sometimes after minutes,
> 	sometimes after hours.
> 	After last crash I booted from a rescue partition. Mounting
> 	partition where system was writing gives no warning, but checking
> 	it with fsck.reiser4 (Debian testing, package reiser4progs version 1.0.5-1)
> 	two inconsistency were founded (and fixed).
> 
> Thanks a lot for your time and your work,
> gelma
> 


I also see freezes, since 2.6.15-mm2. I did not attribute that to reiser4,
though, since there is more in the -mm patches.
I have noticed, though, that at least the last -mm4 freeze resulted in serious
reiser4 corruption. I had to use the --build-fs option to fsck.reiser4.

-Joe

Re: Bug (?) with 2.6.15-mm4

[Heiko Przybyl]

[-- Attachment #1: Type: text/plain, Size: 499 bytes --]

just a quick dirty and possible bad thought:


-mm4:

+i386-enable-4k-stacks-by-default.patch

 Test 4k stacks some more


reiser4:

Note: please make sure that you have checked in

    Code maturity level options
    ---> Prompt for development and/or incomplete code/drivers

and NOT checked in

    Kernel hacking ---> Use 4Kb for kernel stacks instead of 8Kb



maybe the default config option of 4k stacks kills reiser4? so i won't test
it for now :-)


cya Heiko


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 253 bytes --]

Re: Bug (?) with 2.6.15-mm4

[Joe Feise]

Heiko Przybyl wrote on 01/16/06 12:29:

> just a quick dirty and possible bad thought:
> 
> 
> -mm4:
> 
> +i386-enable-4k-stacks-by-default.patch
> 
>  Test 4k stacks some more
> 
> 
> reiser4:
> 
> Note: please make sure that you have checked in
> 
>     Code maturity level options
>     ---> Prompt for development and/or incomplete code/drivers
> 
> and NOT checked in
> 
>     Kernel hacking ---> Use 4Kb for kernel stacks instead of 8Kb
> 
> 
> 
> maybe the default config option of 4k stacks kills reiser4? so i won't test
> it for now :-)


Well, for my configuration, the "use 4kb+4kb for kernel stacks instead of 8kb"
option is turned off.
I followed some of the 4kb stack posts on LKML, so I made sure to stick to 8kb
for the time being.
Thanks for the suggestion, though.

-Joe

Re: Bug (?) with 2.6.15-mm4

[Hans Reiser]

  vs, can you comment and if it is a problem send a patch?

Hans

Heiko Przybyl wrote:

>just a quick dirty and possible bad thought:
>
>
>-mm4:
>
>+i386-enable-4k-stacks-by-default.patch
>
> Test 4k stacks some more
>
>
>reiser4:
>
>Note: please make sure that you have checked in
>
>    Code maturity level options
>    ---> Prompt for development and/or incomplete code/drivers
>
>and NOT checked in
>
>    Kernel hacking ---> Use 4Kb for kernel stacks instead of 8Kb
>
>
>
>maybe the default config option of 4k stacks kills reiser4? so i won't test
>it for now :-)
>
>
>cya Heiko
>
>  
>

Re: Bug (?) with 2.6.15-mm4

[Andrea Gelmini]

On Mon, Jan 16, 2006 at 11:30:20AM -0800, Joe Feise wrote:
> I also see freezes, since 2.6.15-mm2. I did not attribute that to reiser4,
> though, since there is more in the -mm patches.
I emailed to reiser list not for the freeze, but because it seems wrong to
me that filesystem starts without complain when there's corruption/unclean
situation.[1]


ciao,
gelma

---------
[1] no filesystem flame wars, please. I don't pay and don't develop
reiser, so I've go no expectations.