Linux 2.6.17.4

Linux 2.6.17.4

[Greg KH]

We (the -stable team) are announcing the release of the 2.6.17.4 kernel.

I'll also be replying to this message with a copy of the patch between
2.6.17.3 and 2.6.17.4, as it is small enough to do so.

The updated 2.6.17.y git tree can be found at:
 	git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.17.y.git
and can be browsed at the normal kernel.org git web browser:
	www.kernel.org/git/

thanks,

greg k-h

--------

 Makefile     |    2 +-
 kernel/sys.c |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

Summary of changes from v2.6.17.3 to v2.6.17.4
==============================================

Greg Kroah-Hartman:
      fix prctl privilege escalation and suid_dumpable (CVE-2006-2451)
      Linux 2.6.17.4

Re: Linux 2.6.17.4

[Greg KH]

diff --git a/Makefile b/Makefile
index 8c72521..abcf2d7 100644
--- a/Makefile
+++ b/Makefile
@@ -1,7 +1,7 @@
 VERSION = 2
 PATCHLEVEL = 6
 SUBLEVEL = 17
-EXTRAVERSION = .3
+EXTRAVERSION = .4
 NAME=Crazed Snow-Weasel
 
 # *DOCUMENTATION*
diff --git a/kernel/sys.c b/kernel/sys.c
index 0b6ec0e..59273f7 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
 			error = current->mm->dumpable;
 			break;
 		case PR_SET_DUMPABLE:
-			if (arg2 < 0 || arg2 > 2) {
+			if (arg2 < 0 || arg2 > 1) {
 				error = -EINVAL;
 				break;
 			}

Re: Linux 2.6.17.4

[Scott J. Harmon]

Greg KH wrote:
> diff --git a/Makefile b/Makefile
> index 8c72521..abcf2d7 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -1,7 +1,7 @@
>  VERSION = 2
>  PATCHLEVEL = 6
>  SUBLEVEL = 17
> -EXTRAVERSION = .3
> +EXTRAVERSION = .4
>  NAME=Crazed Snow-Weasel
>  
>  # *DOCUMENTATION*
> diff --git a/kernel/sys.c b/kernel/sys.c
> index 0b6ec0e..59273f7 100644
> --- a/kernel/sys.c
> +++ b/kernel/sys.c
> @@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
>  			error = current->mm->dumpable;
>  			break;
>  		case PR_SET_DUMPABLE:
> -			if (arg2 < 0 || arg2 > 2) {
> +			if (arg2 < 0 || arg2 > 1) {
>  				error = -EINVAL;
>  				break;
>  			}
Just curious as to why this isn't just
...
			if (arg2 != 1) {
...

Thanks,

Scott.

Re: Linux 2.6.17.4

[Greg KH]

On Thu, Jul 06, 2006 at 05:43:53PM -0500, Scott J. Harmon wrote:
> Greg KH wrote:
> > diff --git a/Makefile b/Makefile
> > index 8c72521..abcf2d7 100644
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -1,7 +1,7 @@
> >  VERSION = 2
> >  PATCHLEVEL = 6
> >  SUBLEVEL = 17
> > -EXTRAVERSION = .3
> > +EXTRAVERSION = .4
> >  NAME=Crazed Snow-Weasel
> >  
> >  # *DOCUMENTATION*
> > diff --git a/kernel/sys.c b/kernel/sys.c
> > index 0b6ec0e..59273f7 100644
> > --- a/kernel/sys.c
> > +++ b/kernel/sys.c
> > @@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
> >  			error = current->mm->dumpable;
> >  			break;
> >  		case PR_SET_DUMPABLE:
> > -			if (arg2 < 0 || arg2 > 2) {
> > +			if (arg2 < 0 || arg2 > 1) {
> >  				error = -EINVAL;
> >  				break;
> >  			}
> Just curious as to why this isn't just
> ...
> 			if (arg2 != 1) {
> ...

Because that would be incorrect :)

thanks,

greg k-h

Re: Linux 2.6.17.4

[Scott J. Harmon]

Greg KH wrote:
> On Thu, Jul 06, 2006 at 05:43:53PM -0500, Scott J. Harmon wrote:
>> Greg KH wrote:
>>> diff --git a/Makefile b/Makefile
>>> index 8c72521..abcf2d7 100644
>>> --- a/Makefile
>>> +++ b/Makefile
>>> @@ -1,7 +1,7 @@
>>>  VERSION = 2
>>>  PATCHLEVEL = 6
>>>  SUBLEVEL = 17
>>> -EXTRAVERSION = .3
>>> +EXTRAVERSION = .4
>>>  NAME=Crazed Snow-Weasel
>>>  
>>>  # *DOCUMENTATION*
>>> diff --git a/kernel/sys.c b/kernel/sys.c
>>> index 0b6ec0e..59273f7 100644
>>> --- a/kernel/sys.c
>>> +++ b/kernel/sys.c
>>> @@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
>>>  			error = current->mm->dumpable;
>>>  			break;
>>>  		case PR_SET_DUMPABLE:
>>> -			if (arg2 < 0 || arg2 > 2) {
>>> +			if (arg2 < 0 || arg2 > 1) {
>>>  				error = -EINVAL;
>>>  				break;
>>>  			}
>> Just curious as to why this isn't just
>> ...
>> 			if (arg2 != 1) {
>> ...
> 
> Because that would be incorrect :)
> 
> thanks,
> 
> greg k-h

DOH!

/me hides under a rock

Thanks,

Scott.

Re: Linux 2.6.17.4

[Chase Venters]

On Thursday 06 July 2006 17:28, Greg KH wrote:
>  		case PR_SET_DUMPABLE:
> -			if (arg2 < 0 || arg2 > 2) {
> +			if (arg2 < 0 || arg2 > 1) {
>  				error = -EINVAL;
>  				break;
>  			}

Am I staring at this crooked, or not looking deep enough? My manual page for 
prctl says 2 is valid there. Specifically:

              Since  kernel 2.6.13, the value 2 is also permitted; this causes
              any binary which normally would not be dumped to be dumped read-
              able   by   root   only.    (See   also   the   description   of
              /proc/sys/fs/suid_dumpable in proc(5).)

...has something changed, and my manpages don't reflect it? Did I miss a 
conversation about this?

Thanks,
Chase

Re: Linux 2.6.17.4

[Willy Tarreau]

On Thu, Jul 06, 2006 at 03:46:14PM -0700, Greg KH wrote:
> On Thu, Jul 06, 2006 at 05:43:53PM -0500, Scott J. Harmon wrote:
> > Greg KH wrote:
> > > diff --git a/Makefile b/Makefile
> > > index 8c72521..abcf2d7 100644
> > > --- a/Makefile
> > > +++ b/Makefile
> > > @@ -1,7 +1,7 @@
> > >  VERSION = 2
> > >  PATCHLEVEL = 6
> > >  SUBLEVEL = 17
> > > -EXTRAVERSION = .3
> > > +EXTRAVERSION = .4
> > >  NAME=Crazed Snow-Weasel
> > >  
> > >  # *DOCUMENTATION*
> > > diff --git a/kernel/sys.c b/kernel/sys.c
> > > index 0b6ec0e..59273f7 100644
> > > --- a/kernel/sys.c
> > > +++ b/kernel/sys.c
> > > @@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
> > >  			error = current->mm->dumpable;
> > >  			break;
> > >  		case PR_SET_DUMPABLE:
> > > -			if (arg2 < 0 || arg2 > 2) {
> > > +			if (arg2 < 0 || arg2 > 1) {
> > >  				error = -EINVAL;
> > >  				break;
> > >  			}
> > Just curious as to why this isn't just
> > ...
> > 			if (arg2 != 1) {
> > ...
> 
> Because that would be incorrect :)

Interestingly, 2.4 tests (arg2 !=0 && arg2 != 1) so from the code changes
above, it looks like the value 2 was added on purpose, but for what ? Maybe
the fix is not really correct yet ?

Cheers,
Willy

> 
> thanks,
> 
> greg k-h
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [stable] Re: Linux 2.6.17.4

[Greg KH]

On Fri, Jul 07, 2006 at 01:49:18AM +0200, Willy Tarreau wrote:
> On Thu, Jul 06, 2006 at 03:46:14PM -0700, Greg KH wrote:
> > On Thu, Jul 06, 2006 at 05:43:53PM -0500, Scott J. Harmon wrote:
> > > Greg KH wrote:
> > > > diff --git a/Makefile b/Makefile
> > > > index 8c72521..abcf2d7 100644
> > > > --- a/Makefile
> > > > +++ b/Makefile
> > > > @@ -1,7 +1,7 @@
> > > >  VERSION = 2
> > > >  PATCHLEVEL = 6
> > > >  SUBLEVEL = 17
> > > > -EXTRAVERSION = .3
> > > > +EXTRAVERSION = .4
> > > >  NAME=Crazed Snow-Weasel
> > > >  
> > > >  # *DOCUMENTATION*
> > > > diff --git a/kernel/sys.c b/kernel/sys.c
> > > > index 0b6ec0e..59273f7 100644
> > > > --- a/kernel/sys.c
> > > > +++ b/kernel/sys.c
> > > > @@ -1991,7 +1991,7 @@ asmlinkage long sys_prctl(int option, un
> > > >  			error = current->mm->dumpable;
> > > >  			break;
> > > >  		case PR_SET_DUMPABLE:
> > > > -			if (arg2 < 0 || arg2 > 2) {
> > > > +			if (arg2 < 0 || arg2 > 1) {
> > > >  				error = -EINVAL;
> > > >  				break;
> > > >  			}
> > > Just curious as to why this isn't just
> > > ...
> > > 			if (arg2 != 1) {
> > > ...
> > 
> > Because that would be incorrect :)
> 
> Interestingly, 2.4 tests (arg2 !=0 && arg2 != 1) so from the code changes
> above, it looks like the value 2 was added on purpose, but for what ? Maybe
> the fix is not really correct yet ?

No, it's correct.  The change was incorrect.

thanks,

greg k-h

Re: [stable] Re: Linux 2.6.17.4

[Greg KH]

On Thu, Jul 06, 2006 at 06:24:23PM -0500, Chase Venters wrote:
> On Thursday 06 July 2006 17:28, Greg KH wrote:
> >  		case PR_SET_DUMPABLE:
> > -			if (arg2 < 0 || arg2 > 2) {
> > +			if (arg2 < 0 || arg2 > 1) {
> >  				error = -EINVAL;
> >  				break;
> >  			}
> 
> Am I staring at this crooked, or not looking deep enough? My manual page for 
> prctl says 2 is valid there. Specifically:
> 
>               Since  kernel 2.6.13, the value 2 is also permitted; this causes
>               any binary which normally would not be dumped to be dumped read-
>               able   by   root   only.    (See   also   the   description   of
>               /proc/sys/fs/suid_dumpable in proc(5).)
> 
> ...has something changed, and my manpages don't reflect it? Did I miss a 
> conversation about this?

Please refer to the CVE number for details.

thanks,

greg k-h

Re: [stable] Re: Linux 2.6.17.4

[Chris Wright]

* Willy Tarreau (w@1wt.eu) wrote:
> Interestingly, 2.4 tests (arg2 !=0 && arg2 != 1) so from the code changes
> above, it looks like the value 2 was added on purpose, but for what ? Maybe
> the fix is not really correct yet ?

The old code was changed to support a new feature (suid_dumpable
for debugging).  The sysctl should support 2 (the new value), but the
prctl can be abused and hence reducing the test to one analgous to 2.4.
This is a correct fix.

thanks,
-chris

Re: Linux 2.6.17.4

[Chase Venters]

On Thursday 06 July 2006 18:48, Willy Tarreau wrote:
> Interestingly, 2.4 tests (arg2 !=0 && arg2 != 1) so from the code changes
> above, it looks like the value 2 was added on purpose, but for what ? Maybe
> the fix is not really correct yet ?

Hence the source of my curiosity. My prctl() manpage says that 2 makes a core 
that is only readable by root.

       PR_SET_DUMPABLE
              (Since  Linux 2.4) Set the state of the flag determining whether
              core dumps are produced for this process upon delivery of a sig-
              nal  whose  default  behaviour is to produce a core dump.  (Nor-
              mally this flag is set for a  process  by  default,  but  it  is
              cleared  when  a set-user-ID or set-group-ID program is executed
              and also by various system calls that  manipulate  process  UIDs
              and  GIDs).  In kernels up to and including 2.6.12, arg2 must be
              either 0 (process is not dumpable) or 1 (process  is  dumpable).
              Since  kernel 2.6.13, the value 2 is also permitted; this causes
              any binary which normally would not be dumped to be dumped read-
              able   by   root   only.    (See   also   the   description   of
              /proc/sys/fs/suid_dumpable in proc(5).)

> Cheers,
> Willy

Thanks,
Chase

Re: Linux 2.6.17.4

[Jan Engelhardt]

>>>>  		case PR_SET_DUMPABLE:
>>>> -			if (arg2 < 0 || arg2 > 2) {
>>>> +			if (arg2 < 0 || arg2 > 1) {
>>>>  				error = -EINVAL;
>>>>  				break;
>>>>  			}
>>> Just curious as to why this isn't just
>>> ...
>>> 			if (arg2 != 1) {
>>> ...
>> 
>> Because that would be incorrect :)
>
>DOH!
>/me hides under a rock

That's quite a good one.


Jan Engelhardt
-- 

Re: Linux 2.6.17.4

[Marcel Holtmann]

Hi Chase,

> >  		case PR_SET_DUMPABLE:
> > -			if (arg2 < 0 || arg2 > 2) {
> > +			if (arg2 < 0 || arg2 > 1) {
> >  				error = -EINVAL;
> >  				break;
> >  			}
> 
> Am I staring at this crooked, or not looking deep enough? My manual page for 
> prctl says 2 is valid there. Specifically:
> 
>               Since  kernel 2.6.13, the value 2 is also permitted; this causes
>               any binary which normally would not be dumped to be dumped read-
>               able   by   root   only.    (See   also   the   description   of
>               /proc/sys/fs/suid_dumpable in proc(5).)
> 
> ...has something changed, and my manpages don't reflect it? Did I miss a 
> conversation about this?

the manual page must change. This paragraph must be removed.

Regards

Marcel