lkml.org 
[lkml]   [2000]   [Nov]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [PATCH] no RLIMIT_NPROC for root, please
On Tue, 28 Nov 2000, Tigran Aivazian wrote:

> On Tue, 28 Nov 2000, Jan Rekorajski wrote:
> > --- linux/kernel/fork.c~ Tue Sep 5 23:48:59 2000
> > +++ linux/kernel/fork.c Sun Nov 26 20:22:20 2000
> > @@ -560,7 +560,8 @@
> > *p = *current;
> >
> > retval = -EAGAIN;
> > - if (atomic_read(&p->user->processes) >= p->rlim[RLIMIT_NPROC].rlim_cur)
> > + if (p->user->uid &&
> > + (atomic_read(&p->user->processes) >= p->rlim[RLIMIT_NPROC].rlim_cur))
>
> Jan,
>
> Hardcoding things signifying special treatment of uid=0 is almost always a
> bad idea. If you _really_ think that superuser (whatever entity that might
> be) should be exempt from RLIMIT_NPROC and can prove that (SuSv2 seems to
> be silent so you may be right), then you should use capable() to do proper
> capability test and not that horrible explicit uid test as in your patch
> above.

Ok, how about setting limits on login? When this looks like:

--- uid = 0 here
setrlimit(RLIMIT_NPROC, n)
fork() <- this will fail if root has >n processes
setuid(user)

and it is hard to change this sequence, all PAM enabled apps depend on it :(

The other scenario may be when root has already n processes,
call them system stuff, and cannot login or even kill anything
because of limits.

Is this patch better? I put CAP_SYS_RESOURCE, but maybe it should be
CAP_SYS_ADMIN.

--- linux/kernel/fork.c~ Fri Nov 17 15:09:26 2000
+++ linux/kernel/fork.c Tue Nov 28 22:02:17 2000
@@ -562,7 +562,8 @@
*p = *current;

retval = -EAGAIN;
- if (atomic_read(&p->user->processes) >= p->rlim[RLIMIT_NPROC].rlim_cur)
+ if ((atomic_read(&p->user->processes) >= p->rlim[RLIMIT_NPROC].rlim_cur) &&
+ !capable(CAP_SYS_RESOURCE))
goto bad_fork_free;
atomic_inc(&p->user->__count);
atomic_inc(&p->user->processes);
Jan
--
Jan Rękorajski | ALL SUSPECTS ARE GUILTY. PERIOD!
baggins<at>mimuw.edu.pl | OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
BOFH, type MANIAC | -- TROOPS by Kevin Rubio
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
\
 
 \ /
  Last update: 2005-03-22 12:47    [W:0.157 / U:0.048 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site