Bug 125517

Summary: CVE-2005-4889 rpm: Updates leave hardlinked copies untouched.
Product: [Fedora] Fedora Reporter: Michael Schröder <mls>
Component: rpmAssignee: Paul Nasrat <nobody+pnasrat>
Status: CLOSED UPSTREAM QA Contact: Mike McLean <mikem>
Severity: medium Docs Contact:
Priority: medium    
Version: 3CC: jlieskov, mattdm, n3npq, pmatilai, vdanen
Target Milestone: ---Keywords: Security
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2005-11-04 13:38:47 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 150222, 625756    
Attachments:
Description Flags
Proposed patch
none
cvs diff none

Description Michael Schröder 2004-06-08 13:57:07 UTC
If a malicious creates a hardlink to a buggy s-bit program the 
system is still compromised even after a fixed version has been 
installed. The attached fix removes the s-bits from files that 
get updated.

Comment 1 Michael Schröder 2004-06-08 13:57:40 UTC
Created attachment 100965 [details]
Proposed patch

Comment 2 Matthew Miller 2005-04-26 16:37:56 UTC
Fedora Core 2 is now maintained by the Fedora Legacy project for
security updates only. If this problem is a security issue, please
reopen and reassign to the Fedora Legacy product. If it is not a
security issue and hasn't been resolved in the current FC3 updates or
in the FC4 test release, reopen and change the version to match.

Comment 3 Michael Schröder 2005-04-26 16:48:43 UTC
Hmm, good question. I leave it to your security group to decide if it's a 
security issue or not. For now, I've changed the product to FC3. 

Comment 4 Matthew Miller 2005-04-26 17:59:40 UTC
Hmmm. I can see how that might be an issue. I'm going to clone this into Fedora
Legacy as well as leaving in FC3. Thanks.

Comment 5 Paul Nasrat 2005-04-26 22:13:47 UTC
Adding security group and flag - will look at thanks.

Comment 7 Jeff Johnson 2005-11-04 13:38:47 UTC
This patch should probably be done in other ways, but is ok for what the patch does.

Patch added to rpm cvs, should be in rpm-4.4.3-0.35 when built.

Comment 8 Michael Schröder 2010-06-02 13:00:58 UTC
(Argh, you didn't apply the chunk that removes the bits in the FSM_RENAME case! See bug #598775)

Comment 9 Jeff Johnson 2010-06-02 13:31:35 UTC
Created attachment 419033 [details]
cvs diff

Patch to FSM_RENAME case was applied on 15 Nov 2005 20:06:53 exactly as stated.

Comment 10 Michael Schröder 2010-06-02 14:22:33 UTC
Hmm, must have been a different branch. I just see the commit by Paul done Apr 16 13:26:12 2007, which doesn't contain it. rpm-4.4.2.3 also doesn't contain the FSM_RENAME part.
Anyway, what's done is done. I don't want to do any finger pointing.

Comment 11 Jeff Johnson 2010-06-02 14:29:25 UTC
You chose to add the comment here with "you" attached.

And not a different branch, 4.4.2 (on which @rpm.org is based) != 4.4.3.

Comment 12 Michael Schröder 2010-06-04 09:49:17 UTC
(Turns out that maybe I'm the one to blame. I seem to have dropped the chunk by accident when porting the patch from 4.1.1 to 4.4.2, and Paul took it from out version. Hard to tell after such a long time.)

Comment 13 Jeff Johnson 2010-06-04 12:54:07 UTC
Yes its been a long time and removing setuid/setgid (and capabilities and acls and ...)
is a tedious issue which (if deserving of a CVE firedrill) also means that this rpmbuild issue
also needs a CVE (imho its a far more serious issue):

    Name: tag with malicious syntax in spec files can be used to remove home directories:

             Name: foo;~

    can trick rpmbuild into removing home directories and worse.

Then there's %verifyscript that is run multiple times, and the flaw saving/restoring
the chroot directory using embedded lua (also from you and incorrect).

So it goes ... *shrug* ... Have fun!

Comment 14 Vincent Danen 2010-06-14 19:52:56 UTC
MITRE assigned the name CVE-2005-4889 to this issue.  Applying the alias here as per bug #598775 comment #27.