A vulnerability has been identified in the Performance Co-Pilot (PCP) package, stemming from the mixed privilege levels utilized by systemd services associated with PCP. While certain services operate within the confines of limited PCP user/group privileges, others are granted full root privileges. This disparity in privilege levels poses a risk when privileged root processes interact with directories or directory trees owned by unprivileged PCP users. Specifically, this vulnerability may lead to the compromise of PCP user isolation and facilitate local PCP-to-root exploits, particularly through symlink attacks. These vulnerabilities underscore the importance of maintaining robust privilege separation mechanisms within PCP to mitigate the potential for unauthorized privilege escalation.

A flaw was found in the way the xfsda3fixhashpath() function of the Linux kernel's XFS file system implementation ordered directory hashes under certain conditions. A local attacker could use this flaw to corrupt the file system by creating directories, potentially resulting in kernel panic.

Upstream fix:

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c88547a8119e3b581318ab65e9b72f27f23e641d

Reproducer:

http://oss.sgi.com/cgi-bin/gitweb.cgi?p=xfs/cmds/xfstests.git;a=blob;f=src/dirhashcollide.c;h=55cec872d5061ac2ca0f56d1f11e6bf349d5bb97;hb=947ee8bd4b59770534297572b14c695e9c6e001e

References:

http://seclists.org/oss-sec/2014/q4/28 http://marc.info/?l=linux-xfs&m=139590613002926&w=2

A security flaw was found in the way Performance Co-Pilot (PCP), a framework and services to support system-level performance monitoring and performance management, performed management of its temporary files used by various services from the suite. A local attacker could use this flaw to conduct symbolic link attacks (alter or remove different system files, accessible with the privileges of the user running the PCP suite, than it was originally intended).

References: [1] https://bugzilla.novell.com/showbug.cgi?id=782967 (private)

Florian Weimer of the Red Hat Product Security Team discovered a denial of service flaw in pmcd (the PCP (Performance Co-Pilot) performance metrics collector daemon) due to incorrect event-driven programming. Because the pduread() function in libpcp performs a select locally, waiting for more client data, an unauthenticated remote attacker could send individual bytes one by one, avoiding the timeout, and blocking pmcd in order to prevent it from responding to other legitimate requests.

Florian Weimer of the Red Hat Product Security Team discovered two memory leaks in libpcp that can be abused by an unauthenticated remote attacker to crash pmcd (the PCP (Performance Co-Pilot) performance metrics collector daemon) or to consume enough memory to trigger the OOM killer, which may have impact on other processes.

Florian Weimer of the Red Hat Product Security Team discovered that pmcd (the PCP (Performance Co-Pilot) performance metrics collector daemon) exports part of the /proc file system, including privileged information that could be used to aid in bypassing ASLR, as well as full commandline information on running programs.

Florian Weimer of the Red Hat Product Security Team discovered multiple integer and heap-based buffer overflow flaws in PCP (Performance Co-Pilot) libpcp protocol decoding functions. These flaws could lead to daemon crashes or the execution of arbitrary code with root privileges. Many of these flaws can be exploited without requiring the attacker to be authenticated.

Gabriel Vlasiu reported that xfsmetadump, part of the xfsprogs suite of tools for the XFS filesystem, did not properly obfuscate data. xfsmetadump properly obfuscates active metadata, but the rest of the space within that fs block comes through in the clear. This could lead to exposure of stale disk data via the produced metadump image.

The expectation of xfsmetadump is to obfuscate all but the shortest names in the metadata, as noted in the manpage:

By default, xfsmetadump obfuscates most file (regular file, directory and symbolic link) names and extended attribute names to allow the dumps to be sent without revealing confidential information. Extended attribute values are zeroed and no data is copied. The only exceptions are file or attribute names that are 4 or less characters in length. Also file names that span extents (this can only occur with the mkfs.xfs(8) options where -n size > -b size) are not obfuscated. Names between 5 and 8 characters in length inclusively are partially obfuscated.

While the xfsmetadump tool can be run by unprivileged users, it requires appropriate permissions to access block devices (such as root) where the sensitive data might be dumped. An unprivileged user, without access to the block device, could not use this flaw to obtain sensitive data they would not otherwise have permission to access.

eEye Audit ID 2499 in eEye Digital Security Audits 2406 through 2423 for eEye Retina Network Security Scanner on HP-UX, IRIX, and Solaris allows local users to gain privileges via a Trojan horse gauntlet program in an arbitrary directory under /usr/local/.

Description of problem: This series closes a recently discovered problem in XFS filehandle conversion. On systems where inodes are dynamically deleted, XFS does not adequately verify the inode numbers in the filehandles, which results in reading stale inodes from disk and potentially returning them as valid files. Because these unlinked inodes were never zeroed out when the chunk was deallocated, some inodes in the chunk can still appear to have to data extents attached to them. This can lead to stale data exposure, exposure of active data and potentially overwriting of active data if the stale extents referenced in the unlinked inodes have been re-allocated.

Both NFS filehandles and local filehandles provided through libhandle have this same problem. libhandle requires root permissions to use the interface, so it is not exposing information that you can't get more easily with other means (e.g. xfsdb or reading directly form the block device), so there isn't really an issue here.

For NFS, we may incorrectly accept stale file handles for unlinked inodes after a server reboot if the unlinked inodes have not been overwritten leading to the above issues being triggered if multiple NFS clients are accessing the same files.

Christoph's make-bulkstat-coherent patch is the basis for this series as bulkstat can also expose unlinked inodes and information about them back to userspace as it makes the same assumptions about inode lookups as the file handle interfaces.

As a result, the first two patches of the series make up the real bug fix. The last two patches make it clear we are lookuping up untrusted inode numbers and clear away a shortcut that these interfaces used that we do not want used any more. Hence for backports to other kernels, only the first two patches are necessary.

The test program that demonstrates the issue via the openbyhandle interface can be found here:

http://oss.sgi.com/archives/xfs/2010-06/msg00191.html

Version 2: - removed useless ip->iimap.imblkno initialisation in xfsiread() - reworked a comment refering to bulkstat when it should refer to untrusted inodes. - removed typedefs from xfsimaplookup() - killed useless error logging from xfsimaplookup() - rearranged the logic flow of xfsimaplookup() to remove the gotos.

[PATCH 0/4, V2] xfs: validate inode numbers in file handles correctly http://article.gmane.org/gmane.comp.file-systems.xfs.general/33767

[PATCH 1/4] xfs: always use iget in bulkstat http://article.gmane.org/gmane.comp.file-systems.xfs.general/33770

[PATCH 2/4] xfs: validate untrusted inode numbers during lookup http://article.gmane.org/gmane.comp.file-systems.xfs.general/33771

[PATCH 3/4] xfs: rename XFSIGETBULKSTAT to XFSIGETUNTRUSTED http://article.gmane.org/gmane.comp.file-systems.xfs.general/33768

[PATCH 4/4] xfs: remove block number from inode lookup code http://article.gmane.org/gmane.comp.file-systems.xfs.general/33769

[PATCH] xfsqa: test openbyhandle() on unlinked and freed inode cluster http://oss.sgi.com/archives/xfs/2010-06/msg00191.html

Multiple cross-site request forgery (CSRF) vulnerabilities in the web management interface in InterSect Alliance Snare Agent 3.2.3 and earlier on Solaris, Snare Agent 3.1.7 and earlier on Windows, Snare Agent 1.5.0 and earlier on Linux and AIX, Snare Agent 1.4 and earlier on IRIX, Snare Epilog 1.5.3 and earlier on Windows, and Snare Epilog 1.2 and earlier on UNIX allow remote attackers to hijack the authentication of administrators for requests that (1) change the password or (2) change the listening port.

Format string vulnerability in the msgout function in rpc.pcnfsd in IBM AIX 6.1, 5.3, and earlier; IBM VIOS 2.1, 1.5, and earlier; NFS/ONCplus B.11.3109 and earlier on HP HP-UX B.11.11, B.11.23, and B.11.31; and SGI IRIX 6.5 allows remote attackers to execute arbitrary code via an RPC request containing format string specifiers in an invalid directory name.

Cross-site scripting (XSS) vulnerability in index.php in FTP Admin 0.1.0 allows remote attackers to inject arbitrary web script or HTML via the error parameter in an error page action.

Heap-based buffer overflow in libmpdemux/aviheader.c in MPlayer 1.0rc1 and earlier allows remote attackers to cause a denial of service (application crash) or possibly execute arbitrary code via a .avi file with certain large "indx truck size" and nEntriesInuse values, and a certain wLongsPerEntry value.

SGI ProPack 3 SP6 kernel displays the frame buffer contents of the last session after a reboot, which might allow local users to obtain sensitive information.

Xpdf, as used in products such as gpdf, kpdf, pdftohtml, poppler, teTeX, CUPS, libextractor, and others, allows attackers to cause a denial of service (infinite loop) via streams that end prematurely, as demonstrated using the (1) CCITTFaxDecode and (2) DCTDecode streams, aka "Infinite CPU spins."

Xpdf, as used in products such as gpdf, kpdf, pdftohtml, poppler, teTeX, CUPS, libextractor, and others, allows attackers to cause a denial of service (crash) via a crafted FlateDecode stream that triggers a null dereference.

The CCITTFaxStream::CCITTFaxStream function in Stream.cc for xpdf, gpdf, kpdf, pdftohtml, poppler, teTeX, CUPS, libextractor, and others allows attackers to corrupt the heap via negative or large integers in a CCITTFaxDecode stream, which lead to integer overflows and integer underflows.

runpriv in SGI IRIX allows local users to bypass intended restrictions and execute arbitrary commands via shell metacharacters in a command line for a privileged binary in /usr/sysadm/privbin.

rpc.mountd in SGI IRIX 6.5.25, 6.5.26, and 6.5.27 does not correctly allow access to anonymous clients that connect from a system whose hostname can not be determined. NOTE: while this issue occurs in a security mechanism, there is no apparent attacker role and probably does not satisfy the CVE definition of a vulnerability.

Unknown vulnerability in rpc.mountd in SGI IRIX 6.5.25, 6.5.26, and 6.5.27 does not sufficiently restrict access rights for read-mostly exports, which allows attackers to conduct unauthorized activities.

Unknown vulnerability in arshell in the Array Service (arrayd) for SGI ProPack 3 with SP 5 and 6, and SGI ProPack 4, allows local users to execute arbitrary shells as root on other hosts in the cluster or array.

exif.c in PHP before 4.3.11 allows remote attackers to cause a denial of service (memory consumption and crash) via an EXIF header with a large IFD nesting level, which causes significant stack recursion.

grosview in SGI IRIX does not drop privileges before opening files, which allows local users to overwrite arbitrary files via the -s option.

grosview in SGI IRIX 6.5.22, and possibly other 6.5 versions, does not drop privileges when opening description files while in debug mode, which allows local users to read a line from arbitrary files via the -d and -D options, which prints the line as a formatting error.

Unknown vulnerability in ImageMagick before 6.1.8 allows remote attackers to cause a denial of service (application crash) via a crafted PSD file.

ImageMagick before 6.0 allows remote attackers to cause a denial of service (application crash) via a TIFF image with an invalid tag.

The KAME racoon daemon in ipsec-tools before 0.5 allows remote attackers to cause a denial of service (crash) via malformed ISAKMP packets.

scan.c for LibXPM may allow attackers to execute arbitrary code via a negative bitmapunit value that leads to a buffer overflow.

The patch for integer overflow vulnerabilities in Xpdf 2.0 and 3.0 (CVE-2004-0888) is incomplete for 64-bit architectures on certain Linux distributions such as Red Hat, which could leave Xpdf users exposed to the original vulnerabilities.