Home » Mailing list archives » linux-kernel » 2008 » February » 10

Re: kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit

Previous thread: [PATCH] Silent compiler warning introduced by commit 801c135ce73d5df1caf3eca35b66a10824ae0707 (UBI: Unsorted Block Images) by S.Çağlar on Saturday, February 9, 2008 - 8:37 pm. (4 messages)

Next thread: /bin/sh: -c: line 0: syntax error near unexpected token `;' by Mr. James W. Laferriere on Saturday, February 9, 2008 - 11:08 pm. (1 message)
[view original message]
From: Niki Denev
Subject: kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Saturday, February 9, 2008 - 11:04 pm

Hi,

As the subject says the 2.6.24.1 is still vulnerable to the vmsplice
local root exploit.

[opa@test tmp]$ uname -a
Linux tester 2.6.24.1 #1 Sun Feb 10 00:06:49 EST 2008 i686 unknown
[opa@test tmp]$ ./vms

-----------------------------------
 Linux vmsplice Local Root Exploit
 By qaaz
-----------------------------------
[+] mmap: 0x0 .. 0x1000
[+] page: 0x0
[+] page: 0x20
[+] mmap: 0x4000 .. 0x5000
[+] page: 0x4000
[+] page: 0x4020
[+] mmap: 0x1000 .. 0x2000
[+] page: 0x1000
[+] mmap: 0xb7f56000 .. 0xb7f88000
[+] root
[root@test tmp]#
[root@test tmp]# id
uid=0(root) gid=0(root) groups=2033(opa)
[root@test tmp]# uname -a
Linux test 2.6.24.1 #1 Sun Feb 10 00:06:49 EST 2008 i686 unknown

Is there any known fix/patch for this?
--


[ message continues ]
[view original message]
From: Willy Tarreau
Subject: Re: kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Saturday, February 9, 2008 - 11:32 pm

Yes indeed, that's quite bad. 2.6.24-git is still vulnerable too, and
also contains the fix :-(

CC'd Jens as he worked on the fix.

Willy

--


[ message continues ]
[view original message]
From: Niki Denev
Subject: Re: kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Saturday, February 9, 2008 - 11:38 pm

I was unable to gain root on 2.6.24-git20
but after several segfaults when executing the exploit continously
the machine crashes.

--Niki
--


[ message continues ]
[view original message]
From: Niki Denev
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 2:40 am

this fixed the problem for me (kernel 2.6.24.1) :
It appears that the initial patch checked the input to vmsplice_to_user,
but the exploit used vmsplice_to_pipe which remained open to the attack.

--- fs/splice.c.orig	2008-02-08 21:55:30.000000000 +0200
+++ fs/splice.c	2008-02-10 11:32:50.000000000 +0200
@@ -1443,6 +1443,10 @@
 	struct pipe_inode_info *pipe;
 	struct page *pages[PIPE_BUFFERS];
 	struct partial_page partial[PIPE_BUFFERS];
+	int error;
+	long ret;
+	void __user *base;
+	size_t len;
 	struct splice_pipe_desc spd = {
 		.pages = pages,
 		.partial = partial,
@@ -1450,6 +1454,31 @@
 		.ops = &user_page_pipe_buf_ops,
 	};

+	error = ret = 0;
+
+	/*
+	 * Get user address base and length for this iovec.
+	 */
+	error = get_user(base, &iov->iov_base);
+	if (unlikely(error))
+		return error;
+	error = get_user(len, &iov->iov_len);
+	if (unlikely(error))
+		return error;
+
+	/*
+	 * Sanity check this iovec. 0 read succeeds.
+	 */
+	if (unlikely(!len))
+		return 0;
+	if (unlikely(!base)) {
+		return -EFAULT;	
+	}
+
+	if (unlikely(!access_ok(VERIFY_WRITE, base, len))) {
+		return -EFAULT;
+	}
+
 	pipe = pipe_info(file->f_path.dentry->d_inode);
 	if (!pipe)
 		return -EBADF;
--


[ message continues ]
[view original message]
From: Oliver Pinter
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 5:04 am

hmmm, with 2.6.22.y serie is too affected

Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law.
Last login: Sat Feb  9 15:34:35 2008
oliver@home:~$ ./2617_26241_root_exploit
-----------------------------------
 Linux vmsplice Local Root Exploit
 By qaaz
-----------------------------------
[+] mmap: 0x0 .. 0x1000
[+] page: 0x0
[+] page: 0x20
[+] mmap: 0x4000 .. 0x5000
[+] page: 0x4000
[+] page: 0x4020
[+] mmap: 0x1000 .. 0x2000
[+] page: 0x1000
[+] mmap: 0xb7f79000 .. 0xb7fab000
[+] root
root@home:~# uname -a
Linux home 2.6.22.17 #3 SMP PREEMPT Mon Feb 4 17:38:33 CET 2008 i686 GNU/Linux
root@home:~#

-- 
Thanks,
Oliver
--


[ message continues ]
[view original message]
From: Bastian Blank
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 5:22 am

Use VERIFY_READ and this only checks the first entry.

I checked the following patch and it at least fixes the known exploit.

diff --git a/fs/splice.c b/fs/splice.c
index 14e2262..80beb2b 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -1237,6 +1237,11 @@ static int get_iovec_page_array(const struct iovec __user *iov,
 		if (unlikely(!base))
 			break;
 
+		if (!access_ok(VERIFY_READ, base, len)) {
+			error = -EFAULT;
+			break;
+		}
+
 		/*
 		 * Get this base offset and number of pages, then map
 		 * in the user pages.
-- 
Even historians fail to learn from history -- they repeat the same mistakes.
		-- John Gill, "Patterns of Force", stardate 2534.7
--


[ message continues ]
[view original message]
From: Niki Denev
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 5:39 am

This patch is against 2.6.24.1 which has already the fix to vmsplice_to_user
With it i can't exploit the hole, and it is returns "invalid address"
--


[ message continues ]
[view original message]
From: Bastian Blank
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 5:47 am

This is the vmsplice_to_pipe path and I have many reports that it is not
fixed.

Bastian

-- 
If there are self-made purgatories, then we all have to live in them.
		-- Spock, "This Side of Paradise", stardate 3417.7
--


[ message continues ]
[view original message]
From: Niki Denev
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 5:54 am

Exactly, my patch is for the the vmsplice_to_pipe path.
I don't guarantee correctness, but it stops the exploit in my environment.

Niki
--


[ message continues ]
[view original message]
From: Oliver Pinter
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 6:02 am

commit f6e993b835393543bab2d917f9dea75218473edd
Author: Oliver Pinter <oliver.pntr@gmail.com>
Date:   Sun Feb 10 14:03:46 2008 +0100

    [PATCH] vm: splice local root exploit fix for 2.6.22.y

    Based on Bastian Blank's patch

    Fix for CVE_2008_0009 and CVE_2008-0010

    ----->8-----

    oliver@pancs:/tmp$ ./2617_26241_root_exploit
    -----------------------------------
     Linux vmsplice Local Root Exploit
      By qaaz
      -----------------------------------
      [+] mmap: 0x0 .. 0x1000
      [+] page: 0x0
      [+] page: 0x20
      [+] mmap: 0x4000 .. 0x5000
      [+] page: 0x4000
      [+] page: 0x4020
      [+] mmap: 0x1000 .. 0x2000
      [+] page: 0x1000
      [+] mmap: 0xb7f1a000 .. 0xb7f4c000
      [-] vmsplice: Bad address

    -----8<-----

    Signed-off-by: Oliver Pinter <oliver.pntr@gmail.com>

diff --git a/fs/splice.c b/fs/splice.c
index e263d3b..d8b106e 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -1182,6 +1182,12 @@ static int get_iovec_page_array(const struct
iovec __user *iov,
 		if (unlikely(!base))
 			break;

+		/* CVE-2008-0009, CVE-2008-0010 fix */
+		if(!access_ok(VERIFY_READ, base, len)) {
+			error = -EFAULT;
+			break;
+		}
+
 		/*
 		 * Get this base offset and number of pages, then map
 		 * in the user pages.


<<<<<<<


oliver@pancs:/tmp$ ./2617_26241_root_exploit
-----------------------------------
 Linux vmsplice Local Root Exploit
 By qaaz
-----------------------------------
[+] mmap: 0x0 .. 0x1000
[+] page: 0x0
[+] page: 0x20
[+] mmap: 0x4000 .. 0x5000
[+] page: 0x4000
[+] page: 0x4020
[+] mmap: 0x1000 .. 0x2000
[+] page: 0x1000
[+] mmap: 0xb7f1a000 .. 0xb7f4c000
[-] vmsplice: Bad addres

-- 
Thanks,
Oliver
--


[ message continues ]
[view original message]
From: Greg KH
Subject: Re: [stable] [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 10:05 am

No, this is a different CVE, as it is a different problem from the
original 09 and 10 report.

It has been given CVE-2008-0600 to address this issue (09 and 10 only

Hm, perhaps we should just properly check the len field instead?  That's
what is being overflowed here...

thanks,

greg k-h
--


[ message continues ]
[view original message]
From: Pekka Enberg
Subject: Re: [stable] [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 10:11 am

Sorry, I forgot to cc you on this one:

http://lkml.org/lkml/2008/2/10/153

I don't see where the current code is checking that base is
accessible. We just check that we can copy the struct iovecs, right?

                        Pekka
--


[ message continues ]
[view original message]
From: Oliver Pinter
Subject: Re: [stable] [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 10:44 am

simple len and base check is already in kernel:

2.6.22.17 @ 1176,2-16 - fs/splice.c

                /*
                 * Sanity check this iovec. 0 read succeeds.
                 */
                if (unlikely(!len))
                        break;
                error = -EFAULT;
                if (unlikely(!base))
                        break;





-- 
Thanks,
Oliver
--


[ message continues ]
[view original message]
From: Oliver Pinter
Subject: Re: [stable] [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 10:48 am

thanks the info



-- 
Thanks,
Oliver
--


[ message continues ]
[view original message]
From: Niki Denev
Subject: Re: [PATCH] kernel 2.6.24.1 still vulnerable to the vmsplice local root exploit
Date: Sunday, February 10, 2008 - 6:48 am

As far as i can see, at least on x86 and x86_64 the first argument to
access_ok : (VERIFY_READ|VERIFY_WRITE) is ignored.
Also even if it is used on different arch, using WRITE instead of READ
should be safe because WRITE is a superset of READ.

You are right that it only catches the first entry.

--Niki
--


[ message continues ]
Previous thread: [PATCH] Silent compiler warning introduced by commit 801c135ce73d5df1caf3eca35b66a10824ae0707 (UBI: Unsorted Block Images) by S.Çağlar on Saturday, February 9, 2008 - 8:37 pm. (4 messages)

Next thread: /bin/sh: -c: line 0: syntax error near unexpected token `;' by Mr. James W. Laferriere on Saturday, February 9, 2008 - 11:08 pm. (1 message)