pam_limits again...

[Michal Moskal]

On Fri, Feb 02, 2001 at 03:35:17PM +0100, Jan Rekorajski wrote:
> [piątek, 02 luty 2001], Arkadiusz Miskiewicz napisał(a):
> 
> > On/Dnia Fri, Feb 02, 2001 at 03:11:59PM +0100, Jan Rekorajski wrote/napisał(a)
> > > > I don't remember his posting about how to change (fix?) glibc. But why
> > > > we need to fix glibc if it isn't broken?
> > > 
> > > It is. Glibc declares non existing limit - RLIMIT_LOCKS, it should
> > > either not declare it when compiled/used with linux 2.2,
> > see below
> > 
> > > or take care for it in get/setlimits().
> > What is should do? check what version of kernel you are running?
> 
> In headers, see below.
> 
> > But from another point of view. If we change glibc behaviour to not declare
> > RLIMIT_LOCKS when compiled with 2.2 kernel then we will have non-working
> > pam_limits on linux 2.4.
> 
> Glibc uses kernel headers. Right?
> Glibc can #ifdef KERNEL_VERSION in its own headers, Right?

No! What if i got 2.2.18 headers and running 2.4.0 kernel. Or even
worse, builder has 2.2.18 and I 2.4.0 ?

> 
> Then in limits.h should be:
> #ifdef 2.2 || 2.0
> #define NLIMITS 10
> else
> #define NLIMITS 11
> #endif
> 
> Got it?
> 
> Janek
> -- 
> Jan Rękorajski            |  ALL SUSPECTS ARE GUILTY. PERIOD!
> baggins<at>mimuw.edu.pl   |  OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
> BOFH, MANIAC              |                   -- TROOPS by Kevin Rubio
> 
> ________________________
> http://lists.pld.org.pl/

-- 
                        Michał Moskal <malekith at pld.org.pl>

$p=q{print "\$p=q{$p};eval \$p;\n"};eval $p;

NORAD Marxist Ft. Meade Honduras Ft. Bragg Nazi 
supercomputer Peking nuclear strategic AK-47 explosion BATF 
Uzi ammunition genetic SDI Nazi Serbian CIA 
--