Re: [stable] [patch 42/94] [PATCH 11/44] [CVE-2009-0029] System call wrappers part 01

[Greg KH <greg@kroah.com>]

On Fri, Jan 16, 2009 at 03:43:04PM +0100, Pavel Machek wrote:
> On Fri 2009-01-16 12:24:35, Heiko Carstens wrote:
> > On Fri, Jan 16, 2009 at 12:00:29PM +0100, Pavel Machek wrote:
> > > 
> > > > 2.6.28-stable review patch.  If anyone has any objections, please let us know.
> > > > 
> > > > ------------------
> > > > 
> > > > From: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > > 
> > > > commit 58fd3aa288939d3097fa04505b25c2f5e6e144d1 upstream.
> > > > 
> > > > Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
> > > > Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
> > > 
> > > That does not make review exactly easy. Would it be possible to
> > > inline upstream comment in the commit message?
> > > 
> > > ...upstream commit message is empty :-(.
> > 
> > Yes, what should it contain anyway?
> > "This converts the first 10 system calls to the system call wrapper
> > infrastructure"? IMHO the subject says enough.
> 
> "This converts the first 10 system calls to the system call wrapper
> infrastructure. This is neccessary because of <description of security
> hole>. It is good idea to convert all syscalls and not only affected
> ones for uniformity."
> 
> ....at the very least.
> 
> > > > -asmlinkage long
> > > > -sys_nanosleep(struct timespec __user *rqtp, struct timespec __user *rmtp)
> > > > +SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp,
> > > > +		struct timespec __user *, rmtp)
> > > >  {
> > > >  	struct timespec tu;
> > > >  
> > > 
> > > Is it strictly neccessary to modify all the syscalls? 
> > 
> > Not strictly necessary, but much easier to maintain in the long term.
> > It's simply a just convert 'em all approach and never think again about
> > this.
> 
> I believe we should go for minimal patch for -stable. This is really
> huge.

Well, it's spread over 44 patches, but the end result is not that much.
And yes, based on the problem, it is required for -stable.

thanks,

greg k-h