Re: HEADS-UP: starting to commit linuxolator (SoC 2006) changes...

From: Alexander Leidinger <Alexander_at_Leidinger.net>
Date: Wed, 16 Aug 2006 13:25:39 +0200
Quoting Peter Jeremy <peterjeremy_at_optushome.com.au> (from Wed, 16 Aug  
2006 19:06:53 +1000):

> On Wed, 2006-Aug-16 00:23:28 +0200, Alexander Leidinger wrote:
>> For the curious ones: the code is "activated" by changing osrelease,
>> because the glibc behaves differently depending on the osrelease. It
>> makes different syscalls based upon the linux kernel version. So if
>> there's a program which tries if a syscall does not return ENOSYS
>> instead of checking the osrelease, it may fail. If you are aware of
>> such a program, please tell us about it.
>
> I don't think this was mentioned before.  IMHO, this is worth

It's mentioned in /usr/ports/UPDATING (as in: make sure you don't  
change the default osrelease).

> mentioning in the linuxolator documentation because it's not
> immediately obvious and it's likely to bite someone in a non-
> obvious way.

It should be mentioned in the official docs that a non default value  
may change the behavior of linux programs (in -stable you may see  
coredumps when you change the sysctl) and that bug reports with a non  
default value will not be taken care of.

The intend is to change the default value to 2.6.x when the code is  
stable enough. So we don't need to name specific values for osrelease,  
I think.

Bye,
Alexander.

-- 
does your DRESSING ROOM have enough ASPARAGUS?

http://www.Leidinger.net    Alexander _at_ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org       netchild _at_ FreeBSD.org  : PGP ID = 72077137
Received on Wed Aug 16 2006 - 09:26:06 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:59 UTC