Wolfram Schlich | 8 Feb 2008 20:12

plans/schedule for 2.6.24

Hey Brad and pipacs,

can you tell us about your plans/schedule for 2.6.24?

Thanks!
--

-- 
Regards,
Wolfram Schlich <wschlich@...>
Gentoo Linux * http://dev.gentoo.org/~wschlich/
bodik | 11 Feb 2008 15:05
Picon

http://www.securityfocus.com/bid/27704/exploit

Hi,

I've just one question. I've deployed PaX and a little from grsecurity
(no RSBAC features) on two different computers. Kernel configuration
differs only in:

--- config-2.6.19.2bgr3     2008-02-11 14:55:42.000000000 +0100
+++ config-2.6.19.2eryx-gr      2008-02-11 13:38:05.000000000 +0100
 <at>  <at>  -13,8 +13,8  <at>  <at> 
 # CONFIG_PAX_SOFTMODE is not set
 CONFIG_PAX_EI_PAX=y
 CONFIG_PAX_PT_PAX_FLAGS=y
-# CONFIG_PAX_NO_ACL_FLAGS is not set
-CONFIG_PAX_HAVE_ACL_FLAGS=y
+CONFIG_PAX_NO_ACL_FLAGS=y
+# CONFIG_PAX_HAVE_ACL_FLAGS is not set
 # CONFIG_PAX_HOOK_ACL_FLAGS is not set

 #
 <at>  <at>  -25,7 +25,7  <at>  <at> 
 CONFIG_PAX_SEGMEXEC=y
 CONFIG_PAX_EMUTRAMP=y
 CONFIG_PAX_MPROTECT=y
-# CONFIG_PAX_NOELFRELOCS is not set
+CONFIG_PAX_NOELFRELOCS=y
 CONFIG_PAX_KERNEXEC=y

 #
 <at>  <at>  -58,7 +58,7  <at>  <at> 
 # CONFIG_GRKERNSEC_IO is not set
(Continue reading)

bodik | 11 Feb 2008 15:40
Picon

http://www.securityfocus.com/bid/27704/exploit


> I've just one question. I've deployed PaX and a little from grsecurity
> (no RSBAC features) on two different computers. Kernel configuration
> differs only in:

well, to be precise I've to say that bgr2 end eryx are different HW:
* bgr2 is some 64b procesor running 32b OS
   Intel(R) Pentium(R) 4 CPU 3.20GHz
* eryx is native 32b processor
  Intel(R) Xeon(TM) CPU 2.00GHz

bodik
Wolfram Schlich | 13 Feb 2008 21:35

Re: plans/schedule for 2.6.24

* Bradley Spengler <spender@...> [2008-02-08 23:30]:
> On Fri, Feb 08, 2008 at 08:12:02PM +0100, Wolfram Schlich wrote:
> > Hey Brad and pipacs,
> > 
> > can you tell us about your plans/schedule for 2.6.24?
> 
> A stable 2.6.24 patch will be released on the site after the PaX team 
> finishes UDEREF support for 64-bit intel/amd kernels.  I may release 
> test patches in the near future that include the latest PaX patch 
> available but without the new UDEREF support.  Expect those within the 
> next week.

Thanks for the info!
--

-- 
Regards,
Wolfram Schlich <wschlich@...>
Gentoo Linux * http://dev.gentoo.org/~wschlich/
Marc Schiffbauer | 13 Feb 2008 21:46

grsec for vmsplice-bug-fixed Linux 2.6?

Hi all, hi Brad,

will there be a grsec patch for 2.6.24.2 which (hopefully) fixes the
recent vmsplice root-exploit bug?

TIA
-Marc
--

-- 
8AAC 5F46 83B4 DB70 8317  3723 296C 6CCA 35A6 4134
bon | 13 Feb 2008 22:03
Picon

Re: grsec for vmsplice-bug-fixed Linux 2.6?

in the meanwhile
try this sexy one-liner by sd :)

echo -e '\xc3' | dd of=/dev/kmem bs=1 count=1 seek=$((0x`cat 
/proc/kallsyms|grep sys_vmsplice | awk {'print $1'}`))

Marc Schiffbauer wrote:
> Hi all, hi Brad,
> 
> will there be a grsec patch for 2.6.24.2 which (hopefully) fixes the
> recent vmsplice root-exploit bug?
> 
> TIA
> -Marc
Wolfram Schlich | 13 Feb 2008 22:42

shared memory

Heyho,

I am trying to setup a Linux HA cluster using Heartbeat.
There is a component called stonithd which is using shared memory
to store some data. It uses 2 functions, one for storing the data and
one for reading the data.

Here is the source file:
http://hg.linux-ha.org/dev/file/c8d573589311/fencing/stonithd/stonithd.c

	hostlist2shmem() stores data
	shmem2hostlist() reads data

The problem is (happens only with grsecurity kernel), that
shmem2hostlist fails:

	ERROR: shmem2hostlist:3078: shmat failed: Invalid argument

I can see that some privilege dropping and child process stuff is
being done in that code, maybe it's relevant, because only the
read-function fails, the initial store-function works...

The kernel does not log anything (I am used to grsecurity logging
when it denies things)...

What could be the reason?

UPDATE: I found out *sigh*

	sysctl -w kernel.grsecurity.destroy_unused_shm=0
(Continue reading)

Lubomir Host | 13 Feb 2008 22:43
Picon

Re: grsec for vmsplice-bug-fixed Linux 2.6?

I like to use only awk instead of 'cat | grep | awk' combo:

  echo -e '\xc3' | dd of=/dev/kmem bs=1 count=1 seek=$((0x`awk
  '/sys_vmsplice/ { print $1; }' /proc/kallsyms`))

Here is another how to patch running linux kernel:

http://platon.sk/article.php?vmsplice-vulnerability-fast-fix

On Wed, Feb 13, 2008 at 10:03:49PM +0100, bon wrote:
> in the meanwhile
> try this sexy one-liner by sd :)
> 
> echo -e '\xc3' | dd of=/dev/kmem bs=1 count=1 seek=$((0x`cat 
> /proc/kallsyms|grep sys_vmsplice | awk {'print $1'}`))
> 
> Marc Schiffbauer wrote:
> > Hi all, hi Brad,
> > 
> > will there be a grsec patch for 2.6.24.2 which (hopefully) fixes the
> > recent vmsplice root-exploit bug?
> > 
> > TIA
> > -Marc
> _______________________________________________
> grsecurity mailing list
> grsecurity@...
> http://grsecurity.net/cgi-bin/mailman/listinfo/grsecurity

--

-- 
(Continue reading)

bon | 13 Feb 2008 22:55
Picon

Re: grsec for vmsplice-bug-fixed Linux 2.6?

that, however, requires you to have lkm support in the kernel.

Lubomir Host wrote:
> I like to use only awk instead of 'cat | grep | awk' combo:
> 
>   echo -e '\xc3' | dd of=/dev/kmem bs=1 count=1 seek=$((0x`awk
>   '/sys_vmsplice/ { print $1; }' /proc/kallsyms`))
> 
> Here is another how to patch running linux kernel:
> 
> http://platon.sk/article.php?vmsplice-vulnerability-fast-fix
> 
> On Wed, Feb 13, 2008 at 10:03:49PM +0100, bon wrote:
>> in the meanwhile
>> try this sexy one-liner by sd :)
>>
>> echo -e '\xc3' | dd of=/dev/kmem bs=1 count=1 seek=$((0x`cat 
>> /proc/kallsyms|grep sys_vmsplice | awk {'print $1'}`))
>>
>> Marc Schiffbauer wrote:
>>> Hi all, hi Brad,
>>>
>>> will there be a grsec patch for 2.6.24.2 which (hopefully) fixes the
>>> recent vmsplice root-exploit bug?
>>>
>>> TIA
>>> -Marc
>> _______________________________________________
>> grsecurity mailing list
>> grsecurity@...
(Continue reading)


Gmane