Re: [PATCH] KERN_SETUID_DUMPABLE in /proc/sys/fs/

[Kurt Garloff <garloff@suse.de>]

Hi Arjan, Eric,

On Sun, Mar 12, 2006 at 03:07:47PM -0700, Eric W. Biederman wrote:
> Arjan van de Ven <arjan@infradead.org> writes:
> 
> > diff -purN linux-2.6.16-rc5/include/linux/sysctl.h
> > linux-2.6.16-rc5-setuid/include/linux/sysctl.h
> > --- linux-2.6.16-rc5/include/linux/sysctl.h 2006-02-27 09:13:31.000000000 +0100
> > +++ linux-2.6.16-rc5-setuid/include/linux/sysctl.h 2006-03-11 09:02:13.000000000
> > +0100
> > @@ -144,7 +144,6 @@ enum
> >  	KERN_UNKNOWN_NMI_PANIC=66, /* int: unknown nmi panic flag */
> >  	KERN_BOOTLOADER_TYPE=67, /* int: boot loader type */
> >  	KERN_RANDOMIZE=68, /* int: randomize virtual address space */
> > -	KERN_SETUID_DUMPABLE=69, /* int: behaviour of dumps for setuid core */
> >  	KERN_SPIN_RETRY=70,	/* int: number of spinlock retries */
> >  	KERN_ACPI_VIDEO_FLAGS=71, /* int: flags for setting up video after ACPI
> > sleep */
> >  };
> > @@ -759,6 +758,9 @@ enum
> >  	FS_AIO_NR=18,	/* current system-wide number of aio requests */
> >  	FS_AIO_MAX_NR=19, /* system-wide maximum number of aio requests */
> >  	FS_INOTIFY=20,	/* inotify submenu */
> > + /* Note: the following got misplaced but this mistake is
> > +			   now part of the ABI */
> > +	FS_SETUID_DUMPABLE=21, /* int: behaviour of dumps for setuid core */

OK, this is the other possibility. It'll leave a hole in the KERN_
sysctl numbering, but that's not a problem AFAICT.
I would however at least annotate with FIXME and make a note that it's
gonna change in 2.7.

It's hard to tell how many people depend on suid_dumpable being at the
wrong location. The fact nobody noticed it being placed wrongly would
make me suspect it's not widely used yet (it only got introduced
2005-06-23 by Alan Cox). So I would prefer to move it to the correct
location now rather than in 2.7. By the time 2.7 comes, we may have so
many users, we may not want to correct this any more at all.

> This must be number 69 here.  Or else we break the sys_sysctl ABI.

I don't think we break anything we want to maintain. 
Kernel-internal defines? 
OK to change when going from 2.6.15 to 2.6.16 IMVHO. No module
compiled for 2.6.15 will work with 2.6.16 anyway.

Best,
-- 
Kurt Garloff, Head Architect Linux, Novell Inc.

Attachment: pgp0NrF210moh.pgp
Description: PGP signature