Re: SEUL: "Core" proof-of-concept

[Neilen Marais <brick@cryogen.com>]

>> Sounds fine.  2.0.29 was very stable... Why they still include
>2.0.30 I
>> don't know, as its the only 2.0.x kernel thats ever made me doubt
>> linux's stability (but when I moved back to 2.0.29 the faith was
>> restored :)
>
>I used 2.0.30 for about a year, I gues and didn't see any problems. 
>What did you observe?  Do I need to check for any contamination?

No Filesystem contamination AFAIK.  It did crash for a number of
people, and on my one friend's system it would cause ls to coredump
after a couple days uptime... It would also crash every week or two on
the ISP I'm using now's system.  In both cases moving back to 2.0.29
solved the problem (this was before 2.0.31+ came out).  In fact the
only times I ever heard of linux crashes that were not related to
Svgalib or X windows happened with 2.0.30...
>
> Anyway, I won't be surprised if in the future 2.0.34 becomes the
>> base in the end, because were it not for the inode security bug in
>> 2.0.33 it might have been the last of the 2.0.x's.  (If you find my
>> description of the bug somewhat lacking check www.slashdot.org's
>back
>> articles for a link to a better description :)
>
>It was my understanding that the odd numbered kernels denoted unstable
>and were changed to the next even number once the bugs were ironed
>out. 
>Is this not the case?

this is the case for the middle digit.  Therefore 2.0.x are supposed to
be stable kernals... But the first few were not really, as it was a
subversive <g> attempt by Linus to get more people to test 1.3.x. 
Also now and then a new stable kernel just happens to have a bug... 
Best to wait a coupla days after a new kernel release if you'r weary.

----------------------------------
E-Mail: Neilen Marais <brick@cryogen.com>
Date: 19-Jan-98
Time: 23:49:43

This message was sent by XFMail
----------------------------------