Bugzilla Bug 77094

From: 	Brad Spengler
Subject: 	[grsec] grsecurity 2.1.0 release / 5 Linux kernel advisories
Date: 	Fri, 7 Jan 2005 12:20:49 -0500	
grsecurity 2.1.0 release / Linux Kernel advisories
--------------------------------------------------------------------

Table Of Contents:
1) grsecurity 2.1.0 announcement and changelog
2) Linux Kernel advisory introduction
3) 2.4/2.6 random poolsize sysctl handler integer overflow
4) 2.6 scsi ioctl integer overflow and information leak
5) 2.2/2.4/2.6 moxa serial driver bss overflow
6) 2.4/2.6 RLIMIT_MEMLOCK bypass and (2.6) unprivileged user DoS
7) Attachments, including patches for all vulns, a POC for #3, and a
   working exploit for #6

1) grsecurity 2.1.0 announcement and changelog

I'm happy to announce the release of grsecurity 2.1.0.  It is being
released initially for the 2.4.28 and 2.6.10 kernels and will be ported
immediately to the next kernel versions when released.  It can be
downloaded at http://www.grsecurity.net.  We are still actively seeking
sponsorship, so if you benefit from using grsecurity and like the
changes you see in 2.1.0, please consider sponsoring the future
development and maintenance of the project.
Changes in this release include:

* New configuration file for full learning: /etc/grsec/learn_config
* Learning heuristics have been optimized to better detect temporary
  file usage and reduce appropriately.
* Learning heuristics have been modified to weight against reducing
  certain additional important directories.
* User/group ID transitions have been added to the learning system.
  Any subject transitioning to less than 3 different users or 3
  different groups that has CAP_SETUID or CAP_SETGID will have ID
  transitions added.  This is useful to automatically secure
  applications that only transition to one or few users/groups like
  nobody/nogroup.
* /proc/<pid>/* accesses are automatically rewritten as /proc by grlearn
  before being cached and written to disk
* The inherit-based learning usable through the learning configuration
  file is usable through a regular policy as well simply by adding "i"
  instead of "l" to a subject for learning.
* Inheritance is preserved whenever possible across uid/gid changes when
  the role resulting from the uid/gid change is no different from that
  before the change.
* A complete ~95-99% efficient LFU-hash hybrid caching system has been
  added that greatly reduces the number of full object lookups by
  caching the result.  The cache essentially mimics the filesystem
  around where applications are operating: nearly equivalent to having
  an object for every file and directory on the system, but without the
  wasted memory.  The cache is invalidated on creates and deletes that
  cause a change in policy (through policy re-creation) and on renames
  of directories or symlinks.
* Memory usage for non-full learning has been significantly reduced and
  all memory leaks have been plugged.
* A new object mode has been added for hardlinks for more fine-grained
  permissions.  See the sample policy file for information on what
  permissions are required to create a hardlink.  Its corresponding
  audit flag has been added as well.
* Destruction of unused shared memory feature added and included in
  the sysctl framework of grsecurity.  This feature was ported from
  Openwall (http://www.openwall.com/linux).
* A new option was added to the sysctl feature that enables at boot all
  features enabled in the kernel configuration, while allowing them to
  be changed via the sysctl interface until grsec_lock is set.
* Policy statistics have been added to gradm that provide useful,
  security-relevant information on the policy you are loading into the
  kernel.  You can view these statistics when enabling the system by
  running gradm -V -E.
* Interactive performance of full-learning has improved by ~15% by
  reducing the number of context switches caused by grlearn doing small
  disk writes by using a write buffer (writing more once instead of
  less 1000 times) and keeping track of log entry lengths for quicker
  string matching.  A signal handler was added to grlearn so that when
  learning is stopped, the write buffer is flushed to disk.
* Kernel headers are no longer used for gradm
* Updates from the PaX project (http://pax.grsecurity.net)
* Bugfixes for things mentioned on the list, etc

When patching your kernel for the 2.4.28 and 2.6.10 kernels, since they
contain several vulnerabilities, make sure to also apply the secfix
patches located on the website.

2) Linux Kernel advisory introduction

Let me begin by giving you a timeline:

December 15th: I send Linus a mail with a subject line of
"RLIMIT_MEMLOCK bypass with locked stack"
December 27th: The PaX team sends Linus a mail with a subject line of
"2.6.9+ mlockall/expand_down DoS by unprivileged users"
January 2nd: The PaX team resends the previous mail to Linux and Andrew
Morton

Between December 15th and today, Linus has committed many changes to
the kernel.  Between January 2nd and today, Andrew Morton has committed
several changes to the kernel.  3 weeks is a sufficient amount of time
to be able to expect even a reply about a given vulnerability.  A patch
for the vulnerability was attached to the mails, and in the PaX team's
mails, a working exploit as well.  Private notification of
vulnerabilities is a privilege, and when that privilege is abused by not
responding promptly, it deserves to be revoked.

Using 'advanced static analysis': "cd drivers; grep copy_from_user -r ./* |
grep -v sizeof", I discovered 4 exploitable vulnerabilities in a matter
of 15 minutes.  More vulnerabilities were found in 2.6 than in 2.4.
It's a pretty sad state of affairs for Linux security when someone can
find 4 exploitable vulnerabilities in a matter of minutes.  Since there
was no point in sending more vulnerability reports when the first hadn't
even been responded to, I'm including all four of them in this mail, as
well as a POC for the poolsize bug.  The other bugs can have POCs 
written
for just as trivially.  The poolsize bug requires uid 0, but not any
root capabilities.  The scsi and serial bugs depend on the permissions
of their respective devices, and thus can possibly be exploited as
non-root.  The scsi bug in particular has a couple different attack
vectors that I haven't even bothered to investigate.  Some of these bugs
have gone unfixed for several years.

The PaX team discovered the mlockall DoS. It has been fixed in PaX for
2 years.  I have attached their mail and exploit code.

I'd really like to know what's being done about this pitiful trend of
Linux security, where it's 10x as easy to find a vulnerability in the
kernel than it is in any app on the system, where isec releases at
least one critical vulnerability for each kernel version.  I don't see
that the 2.6 development model is doing anything to help this (as the
spectrum of these vulnerabilities demonstrate), by throwing
experimental code into the kernel and claiming it to be "stable".
Hopefully now these vulnerabilities will be fixed in a timely manner.

3) 2.4/2.6 random poolsize sysctl handler integer overflow

In drivers/char/random.c:

at poolsize_strategy():
>        int     len;
        ^ signed integer
>
>        sysctl_poolsize = random_state->poolinfo.POOLBYTES;
>
>        /*
>
>        /*
>         * We only handle the write case, since the read case gets
>         * handled by the default handler (and we don't care if the
>         * write case happens twice; it's harmless).
>         */
>        if (newval && newlen) {
>                len = newlen;
                ^ unsigned int converted to signed
>                if (len > table->maxlen)
                ^ comparison of two signed integers
>                        len = table->maxlen;
>                if (copy_from_user(table->data, newval, len))
                ^ copy_from_user with len possibly > table->maxlen

4) 2.6 scsi ioctl integer overflow and information leak

In drivers/block/scsi_ioctl.c:

at sg_scsi_ioctl():
>        struct request *rq;
>        int err, in_len, out_len, bytes, opcode, cmdlen;
        ^ in_len, out_len are signed int
>        char *buffer = NULL, sense[SCSI_SENSE_BUFFERSIZE];
>
>        /*
>         * get in an out lengths, verify they don't exceed a page worth of data
>         */
>        if (get_user(in_len, &sic->inlen))
        ^ in_len is user-controlled
>                return -EFAULT;
>        if (get_user(out_len, &sic->outlen))
        ^ out_len is user-controlled
>                return -EFAULT;
>        if (in_len > PAGE_SIZE || out_len > PAGE_SIZE)
        ^ signed int only has upper bound checked
>                return -EINVAL;
>        if (get_user(opcode, sic->data))
>                return -EFAULT;
>        bytes = max(in_len, out_len);
...
>        rq->cmd_len = cmdlen;
>        if (copy_from_user(rq->cmd, sic->data, cmdlen))
>                goto error;
>         
>        if (copy_from_user(buffer, sic->data + cmdlen, in_len))
                ^ copy_from_user with size possibly > PAGE_SIZE
>                goto error;
...
>                if (copy_to_user(sic->data, buffer, out_len))
                ^ copy_to_user with size possibly > PAGE_SIZE
>                        err = -EFAULT;

5) 2.2/2.4/2.6 moxa serial driver bss overflow

In drivers/char/moxa.c:

>static unsigned char moxaBuff[10240];

In MoxaDriverIoctl():

>        if(copy_from_user(&dltmp, argp, sizeof(struct dl_str)))
>                return -EFAULT;
                ^ dltmp.len is user-controlled
>        if(dltmp.cardno < 0 || dltmp.cardno >= MAX_BOARDS)
>                return -EINVAL;
>                
>        switch(cmd)
>        {
>        case MOXA_LOAD_BIOS:
>                i = moxaloadbios(dltmp.cardno, dltmp.buf, dltmp.len);
                ^ called with no length checking
>                return (i);
>        case MOXA_FIND_BOARD:
>                return moxafindcard(dltmp.cardno);
>        case MOXA_LOAD_C320B:
>                moxaload320b(dltmp.cardno, dltmp.buf, dltmp.len);
                ^ called with no length checking
>        default: /* to keep gcc happy */
>                return (0);
>        case MOXA_LOAD_CODE:
>                i = moxaloadcode(dltmp.cardno, dltmp.buf, dltmp.len);
                ^ called with no length checking

In moxaloadbios():

>static int moxaloadbios(int cardno, unsigned char __user *tmp, int len)
>{
>        void __iomem *baseAddr;
>        int i;
>
>        if(copy_from_user(moxaBuff, tmp, len))
                ^ copy_from_user with no length checking
>                return -EFAULT;

In moxaloadcode():

> static int moxaloadcode(int cardno, unsigned char __user *tmp, int len)
> {
>        void __iomem *baseAddr, *ofsAddr;
>        int retval, port, i;
>
>        if(copy_from_user(moxaBuff, tmp, len))
                ^ copy_from_user with no length checking
>                return -EFAULT;

In moxaload320b():

>static int moxaload320b(int cardno, unsigned char __user *tmp, int len)
>{
>        void __iomem *baseAddr;
>        int i;
>
>        if(len > sizeof(moxaBuff))
                ^ signed int has only upper-bound checked
>                return -EINVAL;
>        if(copy_from_user(moxaBuff, tmp, len))
                ^ copy_from_user with len possibly > sizeof(moxaBuff)
>                return -EFAULT;

6) 2.4/2.6 RLIMIT_MEMLOCK bypass and (2.6) unprivileged user DoS

Taken from the mail from the PaX team to Linus and Andrew Morton:

the 'culprit' patch is how the default RLIM_MEMLOCK and the privilege
to call mlockall have changed in 2.6.9. namely, the former has been
reduced to 32 pages while the latter has been relaxed to allow it for
otherwise unprivileged users if their RLIM_MEMLOCK is bigger than the
currently allocated vm. which is normally good enough, except as you
now know there's a path that can increase the allocated vm without
checking for RLIM_MEMLOCK.

i'm attaching a small i386-specific demonstration, use the makefile to
create the small self-contained executable, e.g., 'make alloc=0x100000'
to have it allocate 1MB of stack and lock all of it. for demonstrating
the full effect of locking down arbitrary amounts of memory, you'll have
to set your stack rlimit to infinity (ulimit -s unlimited) and allocate
as much memory as your memory overcommit policy allows (this may mean
that you'll have to run multiple instances, if you have lots of memory).

surprisingly, in my tests the kernel survived pretty well, it just crawled
to a snail's speed as every mapped page access required disk i/o ;-). i
didn't play with overcommit policies nor any special workloads, so there
may very well be worse effects with that much locked memory. in any case,
this may warrant 2.6.10.1 because as soon as the fix goes into -bk, anyone
reading the logs can easily figure it out and reproduce the 'exploit'.

the attached patch is the excerpt from PaX that survives the exploit, so
i think it's good to go.

http://grsecurity.net/linux-2.4.28-secfix-200501071141.patch

http://grsecurity.net/linux-2.6.10-secfix-200501071130.patch

06:40PM <spender> my patch won't work on processor's without an mmu
06:40PM <spender> for the isec bug
06:40PM <spender> err processors
06:40PM <spender> but it wasn't a problem for my case
06:40PM <spender> just for yours, you might want to add the missing function and export_symbol to mm/nommu.c
06:41PM <ms> no success with 2.6.8...
06:41PM <ms> however I notice 2.6.8 really sucks ;)
<solar> so this and this are no good? http://bugs.gentoo.org/attachment.cgi?id=47891&action=view
<solar> http://bugs.gentoo.org/attachment.cgi?id=47865&action=edit
06:42PM <spender> oh, no the 2.4 one is fine
06:42PM <spender> it's just the 2.6
06:43PM <spender> btw
06:43PM <spender> if grsec sources includes pax
06:43PM <spender> you need to use my patch for binfmt_elf.c
06:43PM <spender> because you need to hold the semaphore around both the do_brk and pax's do_mmap_pgoff
06:43PM <spender> can't just lock both separately
06:44PM <spender> but yea that patch you have there looks fine, just i think you have to export the do_brk_locked symbol
06:44PM <spender> well, if do_brk is exported in there already
<solar> can I add this as a comment to that bug?
06:44PM <spender> yea
06:45PM <spender> but that 2.6 patch is better than the one i saw on lkml at least
06:45PM <spender> it has the sparc64 change

grsec 2.4.28.2.0.1-200501051112 is in the tree, but will be -* masked for a
little while pending testing and waiting/watching for upstream changes.
Adds linux-2.4.28-CAN-2004-0814.patch & linux-2.4.28-random-poolsize.patch

06:44PM <spender> but yea that patch you have there looks fine, just i think
you have to export the do_brk_locked symbol

In mmap.c, do_brk is exported, and our patch adds do_brk_locked and exports it.

In nommu.c, do_brk is not exported, and our patch adds do_brk_locked and does
not export it. I think this is ok, see the next thing he said:
06:44PM <spender> well, if do_brk is exported in there already
...which its not.

3) 2.4/2.6 random poolsize sysctl handler integer overflow
#3 is fixed in grsec-sources and in the 2.6.x -ac branch.

4) 2.6 scsi ioctl integer overflow and information leak
should be fixed in -ac branch.

5) 2.2/2.4/2.6 moxa serial driver bss overflow
Brads Patch does not apply to 2.4.x (already in grsec I think) fixed in -ac

6) 2.4/2.6 RLIMIT_MEMLOCK bypass and (2.6) unprivileged user DoS
This is fixed directly by the PaX patch. (fixed in -ac)

Tim do you have any of this broken out and can attach here or other bugs?

Yeah it was fixed in -ac7

2.6.10-ac7
+	Fix failure at boot with some setups and ac6	(Alan Cox)
	| Dumb bug indeed
o	Fix random poolsize sysctl			(Brad Spengler)
o	Fix scsi_ioctl leak				(Brad Spengler)
o	Fix rlimit memlock				(Brad Spengler)
o	Fix Moxa serial					(Alan Cox)
	| While moxa won't actually even build on 2.6 the grsecurity fix
	| is wrong (for 2.2, 2.4 as well). Without it being CAP_SYS_RAWIO
	| a user can insert alternative bios firmware into the card.

Created an attachment (id=48020) [details]
patch-2.6.10-ac-6-7.interdiff

interdiff patch-2.6.10-ac6 patch-2.6.10-ac7

Here are the broken out patches:

http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1115_sys-uselib-fix.patch
http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1120_moxa-overflow.patch
http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1125_random-poolsize-overflow.patch
http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1130_rlimit-memlock-dos.patch
http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1135_scsi-ioctl-overflox.patch

gentoo-dev-sources is fixed, btw.

dsd thanks.
Can you please chmod 644 1130_rlimit-memlock-dos.patch I getting no permission to access atm.

@Solar, you can get the exact same patch from this tarball:

http://seclists.org/lists/bugtraq/2005/Jan/att-0068/exploits_and_patches.tgz.

or the genpatches tarball from a mirror, or from here (albeit as a patch to genpatches ;):

http://genpatches.bkbits.net:8080/genpatches/cset@41e038bcWgRXUZFr9okc0cOZhtwNNg?nav=index.html|ChangeSet@-2d

@dsd, I notice that you have asked for gentoo-dev-sources-2.6.10-r4 to be tested pending a move to stable. I'm about to reboot my production box with it so I'll post a works-for-me(tm) when I can.

Fixed permissions. 2.6.10-r4 is also stable already, hence me noting it as
fixed :)

> 2.6.10-r4 is also stable already, hence me noting it as fixed :)

Excellent.

Created an attachment (id=48045) [details]
2.6 Patch

Created an attachment (id=48046) [details]
2.4 Patch (*WARNING* Read comment below!)

For 2.4 kernels you also need to apply the updated CAN-2004-0814 patches in
addition to the 2.4 patch attached on this bug.

For 2.4.28, link to and use
http://dev.gentoo.org/~plasmaroo/patches/kernel/misc/security/linux-2.4.28-CAN-2004-0814.patch

For 2.4.27, link to and use 
http://dev.gentoo.org/~plasmaroo/patches/kernel/misc/security/linux-2.4.27-CAN-2004-0814.2.patch

For 2.4.26, link to and use 
http://dev.gentoo.org/~plasmaroo/patches/kernel/misc/security/linux-2.4.26-CAN-2004-0814.2.patch

All done, following externally maintained sources still need fixing:

hardened-(dev-)sources -- Adding hardened herd...
hppa-sources -- Adding GMSoft...
mips-sources -- Adding Kumba...
openmosix-sources -- Adding cluster herd...
pegasos-dev-sources -- Adding dholm...
rsbac-(dev-)sources -- Adding kang...
sparc-sources -- Adding Joker.

Done in hppa-sources-2.6.10_p10.

done for openMo6-sources.

dsd: Applying the  random poolsize overflow patch from:
http://dev.gentoo.org/~dsd/gentoo-dev-sources/release-10.04/dist/1125_random-poolsize-overflow.patch
generates Hunks because the line does not match with the 2.6.10 source. Albeit
this is nothing serious I did a proper patch wich is attaches. You may want to
apply it.

Created an attachment (id=48178) [details]
Random poolsize overflow patch without hunk

As described above

~x86 hardened-dev-sources-2.6.10 patched

rsbac-sources (2.6) patched

hardened-sources-2.4.28-r2 patched in ~x86

We got 2.6.9 and 2.6.8.1 versions of this patch available?  I've tweaked the
current 2.6 patch for 2.6.9, but 2.6.8.1 is more interesting, in that
net/ipv4/netfilter/ip_conntrack_proto_tcp.c is vastly different than what the
patch expects.

pegasos-sources fixed

fixed in sparc-sources-2.4.28-r6

gentoo-dev-sources is done

mips-sources patched

rsbac-sources 2.4 is also fixed in ~x86

CAN-2005-0504 is for :
Buffer overflow in the MoxaDriverIoctl function for the moxa serial driver (moxa.c) in Linux 2.2.x, 2.4.x, and 2.6.x allows local users to execute arbitrary code via a certain modified length value.

CAN-2005-0179 is for:
Linux kernel 2.4.x and 2.6.x allows local users to cause a denial of service (CPU and memory consumption) and bypass RLIM_MEMLOCK limits via the mlockall call.

CAN-2005-0180 is for:
Multiple integer signedness errors in the sg_scsi_ioctl function in scsi_ioctl.c for Linux 2.6.x allow local users to read or modify kernel memory via negative integers in arguments to the scsi ioctl, which bypass a maximum length check before calling the copy_from_user and copy_to_user functions.

The last one has no CVE apparently

Mass-Ccing kern-sec@gentoo.org to make sure Kernel Security guys know about all
of these...

kang: Have you patched 2.6? I can't see any reference in the ChangeLog to this
bug...

I wrote i patched it: see comment #22 and #29

Since it was fixed before I saw this bug or had any CAN number I didn't wrote any of theses numbers in the ChangeLog.
RSBAC is based on rsbacfixed which had latest -as and/or/additional/brad patches (first had brad patches, and since was integrated by rsbacfixed)
(Now rsbacfixed has latest 2.6.11.z, I copy changelog from this kernel and bugs/CAN numbers whenever I have them after the patch)

Anyway, 0180, 0179, 504 are ok.

Ref: http://fixed.rsbac.org/

All fixed then, closing bug.

http://git.kernel.org/?p=linux/kernel/git/tglx/history.git;a=commit;h=26eecbf3543b7a57699ce5b0bed82b84d3b61705