Bug 3362 - DoS in linux networking hash tables + local vuln
: DoS in linux networking hash tables + local vuln
Status: CLOSED FIXED
Product: Security
Classification: Unclassified
Component: General / Other Security Issue
: unspecified
: Other Linux
: P2 critical
Assigned To: SM Security List
http://rhn.redhat.com/errata/RHSA-200...
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2003-05-16 18:13 UTC by Seth Woolley
Modified: 2003-10-19 23:20 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Seth Woolley 2003-05-16 18:13:34 UTC
DoS in linux networking hash tables

local vuln as well.

W0rf, when you get that patched, post a comment and I'll integrate it up.  If
you need a patch put on a mirror, I can host it, just attach the patch, and put
http://smgl.positivism.org/filename.patch.bz2 as the source.

If you're busy let me know, too, and I'll just extract it from the srpms and
host it, and put the patch in.
Comment 1 Seth Woolley 2003-05-16 19:30:46 UTC
this one has a patch:

http://marc.theaimsgroup.com/?l=bk-commits-24&m=105217616607144&w=2

this one doesn't:

http://bugzilla.kernel.org/show_bug.cgi?id=703

let's patch the remote DoS at least.  I've got a patches/hashDoS file made up to
get the patch from within the bitkeeper url above ;)

testing the compile now.

Does require a very small change to PRE_BUILD to allow urls with & in them.

BTW, sorcery is borked for urls with & in them, so I did a nice hack job, where
I did \& to it.  Also note that the stupid download routine tries to FUZZ the
url to add a .tar.bz2 to the end of it, so I just faked it out with a dirty
hack...  but it's all working now.

hashDoS

#!/bin/bash

DESCRIPTION="security fix for hash table DoS"
SHORT="security fix for hash table DoS"
HELP="Vulnerability: CAN-2003-0244
Several hash table implementations in the networking were remotely exploitable.
 Remote attackers could launch attacks whereby, using carefully choosen forged
source addresses, make every routing cache entry get hashed into the same hash
chain.


Netfilter's IP conntrack module and the TCP syn-queue implementation had
identical vulnerabilities and have been fixed too.

The choosen solution to the problem involved using Bob's Jenkins hash along with
a randomly choosen input.  For the ipv4 routing cache we take things one step
further and periodically choose a new random secret.  By default this happens
every 10 minutes, but this is configurable by the user via sysctl knobs.

This patch is for 2.4.20
"
VERSIONS="2.4.20"

MD5S[0]=0a9d83f0fb0715c3cef97a66f86e9b69

if [[ "$1" == "CONFIG" ]]; then
echo "
SOURCE${COUNTER}='index.html?m=105217616607144\\&q=raw'
SOURCE${COUNTER}_URL='http://marc.theaimsgroup.com/?m=105217616607144\\&q=raw;FAKE=.tar.bz2'"
>> ${SPELL_CONFIG}.DETAILS
source $SCRIPT_DIRECTORY/MD5PATCH
fi

PRE_BUILD: change "$mFILE" to unquoted $mFILE in md5check()

I'm going to submit this once it's all tested.  Does anybody know where the
patch is for the local vuln?
Comment 2 Seth Woolley 2003-05-16 20:30:15 UTC
in devel/test/stable in perforce.

I wonder when it will get tarballed up with the whole ibiblio issue...

Also, for devel, zero added in his cross compiling support to the kernel.  both
he and I have tested it, seems to work without issue by default.  Haven't tried
to actually cross compile with it yet.
Comment 3 erics 2003-06-02 03:24:35 UTC
Patched so closing... 
 
Eric 
Comment 4 games 2003-10-19 23:20:34 UTC
if any of these still have issues outstanding then they can be reopened, but
most  have just been overlooked/forgotten
("these" refers to the 611 fixed but not closed bugs I just found in our database)