The sctp_new function in (1) ip_conntrack_proto_sctp.c and (2) nf_conntrack_proto_sctp.c in Netfilter in Linux kernel 2.6 before 2.6.20.13, and 2.6.21.x before 2.6.21.4, allows remote attackers to cause a denial of service by causing certain invalid states that trigger a NULL pointer dereference.

From Red Hat Security Advisory 2007:0705 :

Updated kernel packages that fix various security issues in the Red Hat Enterprise Linux 5 kernel are now available.

This update has been rated as having important security impact by the Red Hat Security Response Team.

The Linux kernel handles the basic functions of the operating system.

These new kernel packages contain fixes for the following security issues :

* a flaw in the DRM driver for Intel graphics cards that allowed a local user to access any part of the main memory. To access the DRM functionality a user must have access to the X server which is granted through the graphical login. This also only affected systems with an Intel 965 or later graphic chipset. (CVE-2007-3851, Important)

* a flaw in the VFAT compat ioctl handling on 64-bit systems that allowed a local user to corrupt a kernel_dirent struct and cause a denial of service (system crash). (CVE-2007-2878, Important)

* a flaw in the connection tracking support for SCTP that allowed a remote user to cause a denial of service by dereferencing a NULL pointer. (CVE-2007-2876, Important)

* flaw in the CIFS filesystem which could cause the umask values of a process to not be honored. This affected CIFS filesystems where the Unix extensions are supported. (CVE-2007-3740, Important)

* a flaw in the stack expansion when using the hugetlb kernel on PowerPC systems that allowed a local user to cause a denial of service. (CVE-2007-3739, Moderate)

* a flaw in the ISDN CAPI subsystem that allowed a remote user to cause a denial of service or potential remote access. Exploitation would require the attacker to be able to send arbitrary frames over the ISDN network to the victim's machine. (CVE-2007-1217, Moderate)

* a flaw in the cpuset support that allowed a local user to obtain sensitive information from kernel memory. To exploit this the cpuset filesystem would have to already be mounted. (CVE-2007-2875, Moderate)

* a flaw in the CIFS handling of the mount option 'sec=' that didn't enable integrity checking and didn't produce any error message.
(CVE-2007-3843, Low)

Red Hat Enterprise Linux 5 users are advised to upgrade to these packages, which contain backported patches to correct these issues.

From Red Hat Security Advisory 2007:0488 :

Updated kernel packages that fix several security issues and bugs in the Red Hat Enterprise Linux 4 kernel are now available.

This security advisory has been rated as having important security impact by the Red Hat Security Response Team.

The Linux kernel handles the basic functions of the operating system.

These new kernel packages contain fixes for the security issues described below :

* a flaw in the connection tracking support for SCTP that allowed a remote user to cause a denial of service by dereferencing a NULL pointer. (CVE-2007-2876, Important)

* a flaw in the mount handling routine for 64-bit systems that allowed a local user to cause denial of service (crash). (CVE-2006-7203, Important)

* a flaw in the IPv4 forwarding base that allowed a local user to cause an out-of-bounds access. (CVE-2007-2172, Important)

* a flaw in the PPP over Ethernet implementation that allowed a local user to cause a denial of service (memory consumption) by creating a socket using connect and then releasing it before the PPPIOCGCHAN ioctl has been called. (CVE-2007-2525, Important)

* a flaw in the fput ioctl handling of 32-bit applications running on 64-bit platforms that allowed a local user to cause a denial of service (panic). (CVE-2007-0773, Important)

* a flaw in the NFS locking daemon that allowed a local user to cause denial of service (deadlock). (CVE-2006-5158, Moderate)

* a flaw in the sysfs_readdir function that allowed a local user to cause a denial of service by dereferencing a NULL pointer.
(CVE-2007-3104, Moderate)

* a flaw in the core-dump handling that allowed a local user to create core dumps from unreadable binaries via PT_INTERP. (CVE-2007-0958, Low)

* a flaw in the Bluetooth subsystem that allowed a local user to trigger an information leak. (CVE-2007-1353, Low)

In addition, the following bugs were addressed :

* the NFS could recurse on the same spinlock. Also, NFS, under certain conditions, did not completely clean up Posix locks on a file close, leading to mount failures.

* the 32bit compatibility didn't return to userspace correct values for the rt_sigtimedwait system call.

* the count for unused inodes could be incorrect at times, resulting in dirty data not being written to disk in a timely manner.

* the cciss driver had an incorrect disk size calculation (off-by-one error) which prevented disk dumps.

Red Hat would like to thank Ilja van Sprundel and the OpenVZ Linux kernel team for reporting issues fixed in this erratum.

All Red Hat Enterprise Linux 4 users are advised to upgrade their kernels to the packages associated with their machine architectures and configurations as listed in this erratum.

These new kernel packages contain fixes for the security issues described below :

  - a flaw in the connection tracking support for SCTP that     allowed a remote user to cause a denial of service by     dereferencing a NULL pointer. (CVE-2007-2876, Important)

  - a flaw in the mount handling routine for 64-bit systems     that allowed a local user to cause denial of service     (crash). (CVE-2006-7203, Important)

  - a flaw in the IPv4 forwarding base that allowed a local     user to cause an out-of-bounds access. (CVE-2007-2172,     Important)

  - a flaw in the PPP over Ethernet implementation that     allowed a local user to cause a denial of service     (memory consumption) by creating a socket using connect     and then releasing it before the PPPIOCGCHAN ioctl has     been called. (CVE-2007-2525, Important)

  - a flaw in the fput ioctl handling of 32-bit applications     running on 64-bit platforms that allowed a local user to     cause a denial of service (panic). (CVE-2007-0773,     Important)

  - a flaw in the NFS locking daemon that allowed a local     user to cause denial of service (deadlock).
    (CVE-2006-5158, Moderate)

  - a flaw in the sysfs_readdir function that allowed a     local user to cause a denial of service by dereferencing     a NULL pointer. (CVE-2007-3104, Moderate)

  - a flaw in the core-dump handling that allowed a local     user to create core dumps from unreadable binaries via     PT_INTERP. (CVE-2007-0958, Low)

  - a flaw in the Bluetooth subsystem that allowed a local     user to trigger an information leak. (CVE-2007-1353,     Low)

In addition, the following bugs were addressed :

  - the NFS could recurse on the same spinlock. Also, NFS,     under certain conditions, did not completely clean up     Posix locks on a file close, leading to mount failures.

  - the 32bit compatibility didn't return to userspace     correct values for the rt_sigtimedwait system call.

  - the count for unused inodes could be incorrect at times,     resulting in dirty data not being written to disk in a     timely manner.

  - the cciss driver had an incorrect disk size calculation     (off-by-one error) which prevented disk dumps.

NOTE1: From The Upstream Vendors release notes 'During PCI probing, Red Hat Enterprise Linux 4 Update 5 attempts to use information obtained from MCFG (memory-mapped PCI configuration space). On AMD-systems, this type of access does not work on some buses, as the kernel cannot parse the MCFG table.

To work around this, add the parameter pci=conf1 or pci=nommconf on the kernel boot line in /etc/grub.conf. For example :

title Red Hat Enterprise Linux AS (2.6.9-42.0.2.EL) root (hd0,0) kernel /vmlinuz-2.6.9-42.0.2.EL ro root=/dev/VolGroup00/LogVol00 rhgb quiet pci=conf1 initrd /initrd-2.6.9-42.0.2.EL.img

Doing this instructs the kernel to use PCI Conf1 access instead of MCFG-based access.'

NOTE2: From The Upstream Vendors Knowledge Base 'Why did the ordering of my NIC devices change in Red Hat Enterprise Linux 4.5?

The 2.6.9-55 version of the Red Hat Enterprise Linux 4 kernel (Update 5) reverts to the 2.4 ordering of network interface cards (NICs) on certain systems. Note that if the 'HWADDR=MAC ADDRESS' line is present in the /etc/sysconfig/network-scripts/ifcfg-ethX files, the NIC ordering will not change.

To restore the original 2.6 ordering, which is different from the 2.4 ordering, boot with the option pci=nobfsort '

This kernel update fixes the following security problems :

  - The IPv6 protocol allows remote attackers to cause a     denial of service via crafted IPv6 type 0 route headers     (IPV6_RTHDR_TYPE_0) that create network amplification     between two routers. (CVE-2007-2242)

    The default is that RH0 is disabled now. To adjust this,     write to the file /proc/net/accept_source_route6.

  - The random number feature in the Linux kernel 2.6 (1)     did not properly seed pools when there is no entropy, or     (2) used an incorrect cast when extracting entropy,     which might have caused the random number generator to     provide the same values after reboots on systems without     an entropy source. (CVE-2007-2453)

  - A NULL pointer dereference in SCTP connection tracking     could be caused by a remote attacker by sending     specially crafted packets. Note that this requires SCTP     set-up and active to be exploitable. (CVE-2007-2876)

  - Stack-based buffer overflow in the random number     generator (RNG) implementation in the Linux kernel     before 2.6.22 might allow local root users to cause a     denial of service or gain privileges by setting the     default wakeup threshold to a value greater than the     output pool size, which triggers writing random numbers     to the stack by the pool transfer function involving     'bound check ordering'. (CVE-2007-3105)

    Since this value can only be changed by a root user,     exploitability is low.

  - The signal handling in the Linux kernel, when run on     PowerPC systems using HTX, allows local users to cause a     denial of service via unspecified vectors involving     floating point corruption and concurrency.
    (CVE-2007-3107)

  - Memory leak in the PPP over Ethernet (PPPoE) socket     implementation in the Linux kernel allowed local users     to cause a denial of service (memory consumption) by     creating a socket using connect, and releasing it before     the PPPIOCGCHAN ioctl is initialized. (CVE-2007-2525)

  - The lcd_write function in drivers/usb/misc/usblcd.c in     the Linux kernel did not limit the amount of memory used     by a caller, which allowed local users to cause a denial     of service (memory consumption). (CVE-2007-3513)

  - A local attacker could send a death signal to a setuid     root program under certain conditions, potentially     causing unwanted behaviour in this program.
    (CVE-2007-3848)

  - On machines with a Intel i965 based graphics card local     users with access to the direct rendering devicenode     could overwrite memory on the machine and so gain root     privileges. (CVE-2007-3851)

  - Fixed a denial of service possibility where a local     attacker with access to a pwc camera device could hang     the USB subsystem. [#302194]

and the following non security bugs :

  - patches.arch/ppc-oprofile-970mp.patch: enable ppc64/970     MP, requires oprofile 0.9.3 [#252696]

  - patches.arch/x86_64-no-tsc-with-C3: don't use TSC on     x86_64 Intel systems when CPU has C3 [#254061]

  - patches.arch/x86_64-hpet-lost-interrupts-fix.patch:
    backport x86_64 hpet lost interrupts code [#257035]

  - patches.fixes/fusion-nat-consumption-fix: handle a     potential race in mptbase. This fixes a NaT consumption     crash [#257412]

  - patches.arch/ia64-skip-clock-calibration: enabled     [#259501]

  - patches.fixes/md-raid1-handle-read-error: Correctly     handle read errors from a failed drive in raid1     [#261459]

  - patches.arch/ia64-fix-kdump-on-init: kdump on INIT needs     multi-nodes sync-up (v.2) [#265764]

  - patches.arch/ia64-perfmon-fix-2: race condition between     pfm_context_create and pfm_read [#268131]

  - patches.fixes/cpufreq_ppc_boot_option.patch: workaround     for _PPC (BIOS cpufreq limitations) [#269579]

  - patches.arch/acpi_package_object_support.patch: ACPI     package object as method parameter support (in AML)     [#270956]

  - patches.fixes/ia64_cpufreq_PDC.patch: correctly assign     as cpufreq capable driver (_PDC) to BIOS [#270973]

  - patches.arch/ia64-kdump-hpzx1-ioc-workaround: update to     latest upstream version of the patch [#271158]

  - patches.suse/delayacct_memleak.patch: Fix delayacct     memory leak [#271187]

  - patches.fixes/fc_transport-check-portstate-before-scan:
    check FC portstates before invoking target scan     [#271338]

  - patches.fixes/unusual14cd.patch: quirk for 14cd:6600     [#274087]

  -     patches.fixes/reiserfs-change_generation_on_update_sd.di     ff: fix assertion failure in reiserfs [#274288]

  -     patches.drivers/d-link-dge-530t-should-use-the-skge-driv     er.patch: D-Link DGE-530T should use the skge driver     [#275376]

  - patches.arch/ia64-dont-unwind-running-tasks.patch: Only     unwind non-running tasks [#275854]

  - patches.fixes/dm-mpath-rdac-avt-support: short circuit     RDAC hardware handler in AVT mode [#277834]

  - patches.fixes/lkcd-re-enable-valid_phys_addr_range:
    re-enable the valid_phys_addr_range() check [#279433]

  - patches.drivers/cciss-panic-on-reboot: when root     filesystem is xfs the server cannot do a second reboot     [#279436] Also resolves same issue in [#291759].

  - patches.drivers/ide-hpt366-fix-302n-oops: fix hpt302n     oops [#279705]

  - patches.fixes/serial-8250-backup-timer-2-deadlock-fix:
    fix possible deadlock [#280771]

  - patches.fixes/nfs-osync-error-return: ensure proper     error return from O_SYNC writes [#280833]

  - patches.fixes/acpi_pci_hotplug_poweroff.patch: ACPI PCI     hotplug driver acpiphp unable to power off PCI slot     [#281234]

  -     patches.drivers/pci-hotplug-acpiphp-remove-hot-plug-para     meter-write-to-pci-host-bridge.patch: remove hot plug     parameter write to PCI host bridge [#281239]

  - patches.fixes/scsi-set-correct-resid: Incorrect 'resid'     field values when using a tape device [#281640]

  - patches.drivers/usb-edgeport-epic-support.patch: USB:
    add EPIC support to the io_edgeport driver [#281921]

  - patches.fixes/usb-hid-ncr-no-init-reports.patch: HID:
    Don't initialize reports for NCR devices [#281921]

  - patches.drivers/ppc-power6-ehea.patch: use decimal     values in sysfs propery logical_port_id, fix panic when     adding / removing logical eHEA ports [#283070]

  - patches.arch/ppc-power6-ebus.patch: DLPAR Adapter     add/remove functionality for eHEA [#283239]

  - patches.fixes/nfs-enospc: Return ENOSPC and EDQUOT to     NFS write requests more promptly [#284042]

  -     patches.drivers/pci-hotplug-acpiphp-avoid-acpiphp-cannot
    -get-bridge-info-pci-hotplug-failure.patch: PCI:
    hotplug: acpiphp: avoid acpiphp 'cannot get bridge info'     PCI hotplug failure [#286193]

  - patches.drivers/lpfc-8.1.10.9-update: lpfc update to     8.1.10.9 [#286223]

  - patches.fixes/make-swappiness-safer-to-use.patch: Handle     low swappiness gracefully [#288799]

  - patches.arch/ppc-oprofile-power5plusplus.patch: oprofile     support for Power 5++ [#289223]

  - patches.drivers/ppc-power6-ehea.patch: Fixed possible     kernel panic on VLAN packet recv [#289301]

  - patches.fixes/igrab_should_check_for_i_clear.patch:
    igrab() should check for I_CLEAR [#289576]

  - patches.fixes/wait_for_sysfs_population.diff: Driver     core: bus device event delay [#289964]

  -     patches.drivers/scsi-throttle-SG_DXFER_TO_FROM_DEV-warni     ng-better: better throttling of SG_DXFER_TO_FROM_DEV     warning messages [#290117]

  -     patches.arch/mark-unwind-info-for-signal-trampolines-in-     vdsos.patch: Mark unwind info for signal trampolines in     vDSOs [#291421]

  - patches.fixes/hugetlbfs-stack-grows-fix.patch: don't     allow the stack to grow into hugetlb reserved regions     [#294021]

  - patches.drivers/alsa-post-sp1-hda-analog-update: add     support of of missing AD codecs [#294471]

  - patches.drivers/alsa-post-sp1-hda-conexant-fixes: fix     unterminated arrays [#294480]

  - patches.fixes/fix_hpet_init_race.patch: fix a race in     HPET initialization on x86_64 resulting in a lockup on     boot [#295115]

  - patches.drivers/alsa-post-sp1-hda-sigmatel-pin-fix: Fix     number of pin widgets with STAC codecs [#295653]

  -     patches.fixes/pci-pcieport-driver-remove-invalid-warning
    -message.patch: PCI: pcieport-driver: remove invalid     warning message [#297135] [#298561]

  - patches.kernel.org/patch-2.6.16.NN-$((NN+1)), NN =     18,...,52: update to Kernel 2.6.16.53; lots of bugfixes     [#298719] [#186582] [#186583] [#186584]

  - patches.fixes/ocfs2-1.2-svn-r3027.diff: proactive patch     [#298845]

  - patches.drivers/b44-phy-fix: Fix frequent PHY resets     under load on b44 [#301653]

  - dd patches.arch/ppc-eeh-node-status-okay.patch firmware     returns 'okay' instead of 'ok' for node status [#301788]

Updated kernel packages that fix various security issues in the Red Hat Enterprise Linux 5 kernel are now available.

This update has been rated as having important security impact by the Red Hat Security Response Team.

The Linux kernel handles the basic functions of the operating system.

These new kernel packages contain fixes for the following security issues :

* a flaw in the DRM driver for Intel graphics cards that allowed a local user to access any part of the main memory. To access the DRM functionality a user must have access to the X server which is granted through the graphical login. This also only affected systems with an Intel 965 or later graphic chipset. (CVE-2007-3851, Important)

* a flaw in the VFAT compat ioctl handling on 64-bit systems that allowed a local user to corrupt a kernel_dirent struct and cause a denial of service (system crash). (CVE-2007-2878, Important)

* a flaw in the connection tracking support for SCTP that allowed a remote user to cause a denial of service by dereferencing a NULL pointer. (CVE-2007-2876, Important)

* flaw in the CIFS filesystem which could cause the umask values of a process to not be honored. This affected CIFS filesystems where the Unix extensions are supported. (CVE-2007-3740, Important)

* a flaw in the stack expansion when using the hugetlb kernel on PowerPC systems that allowed a local user to cause a denial of service. (CVE-2007-3739, Moderate)

* a flaw in the ISDN CAPI subsystem that allowed a remote user to cause a denial of service or potential remote access. Exploitation would require the attacker to be able to send arbitrary frames over the ISDN network to the victim's machine. (CVE-2007-1217, Moderate)

* a flaw in the cpuset support that allowed a local user to obtain sensitive information from kernel memory. To exploit this the cpuset filesystem would have to already be mounted. (CVE-2007-2875, Moderate)

* a flaw in the CIFS handling of the mount option 'sec=' that didn't enable integrity checking and didn't produce any error message.
(CVE-2007-3843, Low)

Red Hat Enterprise Linux 5 users are advised to upgrade to these packages, which contain backported patches to correct these issues.

A flaw was discovered in the PPP over Ethernet implementation. Local attackers could manipulate ioctls and cause kernel memory consumption leading to a denial of service. (CVE-2007-2525)

An integer underflow was discovered in the cpuset filesystem. If mounted, local attackers could obtain kernel memory using large file offsets while reading the tasks file. This could disclose sensitive data. (CVE-2007-2875)

Vilmos Nebehaj discovered that the SCTP netfilter code did not correctly validate certain states. A remote attacker could send a specially crafted packet causing a denial of service. (CVE-2007-2876)

Luca Tettamanti discovered a flaw in the VFAT compat ioctls on 64-bit systems. A local attacker could corrupt a kernel_dirent struct and cause a denial of service. (CVE-2007-2878)

A flaw in the sysfs_readdir function allowed a local user to cause a denial of service by dereferencing a NULL pointer. (CVE-2007-3104)

A buffer overflow was discovered in the random number generator. In environments with granular assignment of root privileges, a local attacker could gain additional privileges. (CVE-2007-3105)

A flaw was discovered in the usblcd driver. A local attacker could cause large amounts of kernel memory consumption, leading to a denial of service. (CVE-2007-3513)

Zhongling Wen discovered that the h323 conntrack handler did not correctly handle certain bitfields. A remote attacker could send a specially crafted packet and cause a denial of service.
(CVE-2007-3642)

A flaw was discovered in the CIFS mount security checking. Remote attackers could spoof CIFS network traffic, which could lead a client to trust the connection. (CVE-2007-3843)

It was discovered that certain setuid-root processes did not correctly reset process death signal handlers. A local user could manipulate this to send signals to processes they would not normally have access to. (CVE-2007-3848)

The Direct Rendering Manager for the i915 driver could be made to write to arbitrary memory locations. An attacker with access to a running X11 session could send a specially crafted buffer and gain root privileges. (CVE-2007-3851)

It was discovered that the aacraid SCSI driver did not correctly check permissions on certain ioctls. A local attacker could cause a denial of service or gain privileges. (CVE-2007-4308).

Note that Tenable Network Security has extracted the preceding description block directly from the Ubuntu security advisory. Tenable has attempted to automatically clean and format it as much as possible without introducing additional issues.

A flaw was discovered in dvb ULE decapsulation. A remote attacker could send a specially crafted message and cause a denial of service.
(CVE-2006-4623)

The compat_sys_mount function allowed local users to cause a denial of service when mounting a smbfs filesystem in compatibility mode.
(CVE-2006-7203)

The Omnikey CardMan 4040 driver (cm4040_cs) did not limit the size of buffers passed to read() and write(). A local attacker could exploit this to execute arbitrary code with kernel privileges. (CVE-2007-0005)

Due to an variable handling flaw in the ipv6_getsockopt_sticky() function a local attacker could exploit the getsockopt() calls to read arbitrary kernel memory. This could disclose sensitive data.
(CVE-2007-1000)

Ilja van Sprundel discovered that Bluetooth setsockopt calls could leak kernel memory contents via an uninitialized stack buffer. A local attacker could exploit this flaw to view sensitive kernel information.
(CVE-2007-1353)

A flaw was discovered in the handling of netlink messages. Local attackers could cause infinite recursion leading to a denial of service. (CVE-2007-1861)

The random number generator was hashing a subset of the available entropy, leading to slightly less random numbers. Additionally, systems without an entropy source would be seeded with the same inputs at boot time, leading to a repeatable series of random numbers.
(CVE-2007-2453)

A flaw was discovered in the PPP over Ethernet implementation. Local attackers could manipulate ioctls and cause kernel memory consumption leading to a denial of service. (CVE-2007-2525)

An integer underflow was discovered in the cpuset filesystem. If mounted, local attackers could obtain kernel memory using large file offsets while reading the tasks file. This could disclose sensitive data. (CVE-2007-2875)

Vilmos Nebehaj discovered that the SCTP netfilter code did not correctly validate certain states. A remote attacker could send a specially crafted packet causing a denial of service. (CVE-2007-2876)

Luca Tettamanti discovered a flaw in the VFAT compat ioctls on 64-bit systems. A local attacker could corrupt a kernel_dirent struct and cause a denial of service. (CVE-2007-2878)

A flaw was discovered in the cluster manager. A remote attacker could connect to the DLM port and block further DLM operations.
(CVE-2007-3380)

A flaw was discovered in the usblcd driver. A local attacker could cause large amounts of kernel memory consumption, leading to a denial of service. (CVE-2007-3513).

Note that Tenable Network Security has extracted the preceding description block directly from the Ubuntu security advisory. Tenable has attempted to automatically clean and format it as much as possible without introducing additional issues.

The compat_sys_mount function allowed local users to cause a denial of service when mounting a smbfs filesystem in compatibility mode.
(CVE-2006-7203)

The Omnikey CardMan 4040 driver (cm4040_cs) did not limit the size of buffers passed to read() and write(). A local attacker could exploit this to execute arbitrary code with kernel privileges. (CVE-2007-0005)

Due to a variable handling flaw in the ipv6_getsockopt_sticky() function a local attacker could exploit the getsockopt() calls to read arbitrary kernel memory. This could disclose sensitive data.
(CVE-2007-1000)

Ilja van Sprundel discovered that Bluetooth setsockopt calls could leak kernel memory contents via an uninitialized stack buffer. A local attacker could exploit this flaw to view sensitive kernel information.
(CVE-2007-1353)

A flaw was discovered in the handling of netlink messages. Local attackers could cause infinite recursion leading to a denial of service. (CVE-2007-1861)

A flaw was discovered in the IPv6 stack's handling of type 0 route headers. By sending a specially crafted IPv6 packet, a remote attacker could cause a denial of service between two IPv6 hosts.
(CVE-2007-2242)

The random number generator was hashing a subset of the available entropy, leading to slightly less random numbers. Additionally, systems without an entropy source would be seeded with the same inputs at boot time, leading to a repeatable series of random numbers.
(CVE-2007-2453)

A flaw was discovered in the PPP over Ethernet implementation. Local attackers could manipulate ioctls and cause kernel memory consumption leading to a denial of service. (CVE-2007-2525)

An integer underflow was discovered in the cpuset filesystem. If mounted, local attackers could obtain kernel memory using large file offsets while reading the tasks file. This could disclose sensitive data. (CVE-2007-2875)

Vilmos Nebehaj discovered that the SCTP netfilter code did not correctly validate certain states. A remote attacker could send a specially crafted packet causing a denial of service. (CVE-2007-2876)

Luca Tettamanti discovered a flaw in the VFAT compat ioctls on 64-bit systems. A local attacker could corrupt a kernel_dirent struct and cause a denial of service. (CVE-2007-2878).

Note that Tenable Network Security has extracted the preceding description block directly from the Ubuntu security advisory. Tenable has attempted to automatically clean and format it as much as possible without introducing additional issues.

Rebase to upstream 2.6.21.5, including several security fixes (See CVEs for details).

Upstream changelog:
http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.21.3 http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.21.4 http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.21.5

Fedora specific changes detailed below :

Note that Tenable Network Security has extracted the preceding description block directly from the Fedora security advisory. Tenable has attempted to automatically clean and format it as much as possible without introducing additional issues.

This kernel update brings the kernel to the one shipped with SLES 10 Service Pack 1 and also fixes the following security problems:

- CVE-2007-2242: The IPv6 protocol allows remote attackers   to cause a denial of service via crafted IPv6 type 0   route headers (IPV6_RTHDR_TYPE_0) that create network   amplification between two routers. 

  The default is that RH0 is disabled now. To adjust this, write to the file /proc/net/accept_source_route6.

- CVE-2007-2453: The random number feature in the Linux   kernel 2.6 (1) did not properly seed pools when there is   no entropy, or (2) used an incorrect cast when extracting   entropy, which might have caused the random number   generator to provide the same values after reboots on   systems without an entropy source.

- CVE-2007-2876: A NULL pointer dereference in SCTP   connection tracking could be caused by a remote attacker   by sending specially crafted packets. Note that this   requires SCTP set-up and active to be exploitable.

- CVE-2007-3105: Stack-based buffer overflow in the random   number generator (RNG) implementation in the Linux kernel   before 2.6.22 might allow local root users to cause a   denial of service or gain privileges by setting the   default wakeup threshold to a value greater than the   output pool size, which triggers writing random numbers   to the stack by the pool transfer function involving   'bound check ordering'.

  Since this value can only be changed by a root user, exploitability is low.

- CVE-2007-3107: The signal handling in the Linux kernel,   when run on PowerPC systems using HTX, allows local users   to cause a denial of service via unspecified vectors   involving floating point corruption and concurrency.

- CVE-2007-2525: Memory leak in the PPP over Ethernet   (PPPoE) socket implementation in the Linux kernel allowed   local users to cause a denial of service (memory   consumption) by creating a socket using connect, and   releasing it before the PPPIOCGCHAN ioctl is initialized.

- CVE-2007-3513: The lcd_write function in   drivers/usb/misc/usblcd.c in the Linux kernel did not   limit the amount of memory used by a caller, which   allowed local users to cause a denial of service (memory   consumption).

- CVE-2007-3851: On machines with a Intel i965 based   graphics card local users with access to the direct   rendering devicenode could overwrite memory on the   machine and so gain root privileges.

This kernel is not compatible to the previous SUSE Linux 10.1 kernel, so the Kernel Module Packages will need to be updated.

This kernel update fixes the following security problems :

  - CVE-2007-1861: The nl_fib_lookup function in     net/ipv4/fib_frontend.c allows attackers to cause a     denial of service (kernel panic) via NETLINK_FIB_LOOKUP     replies, which trigger infinite recursion and a stack     overflow.

  - CVE-2007-1496: nfnetlink_log in netfilter allows     attackers to cause a denial of service (crash) via     unspecified vectors involving the (1) nfulnl_recv_config     function, (2) using 'multiple packets per netlink     message', and (3) bridged packets, which trigger a NULL     pointer dereference.

  - CVE-2007-1497: nf_conntrack in netfilter does not set     nfctinfo during reassembly of fragmented packets, which     leaves the default value as IP_CT_ESTABLISHED and might     allow remote attackers to bypass certain rulesets using     IPv6 fragments.

    Please note that the connection tracking option for IPv6     is not enabled in any currently shipping SUSE Linux     kernel, so it does not affect SUSE Linux default     kernels.

  - CVE-2007-2242: The IPv6 protocol allows remote attackers     to cause a denial of service via crafted IPv6 type 0     route headers (IPV6_RTHDR_TYPE_0) that create network     amplification between two routers.

    The behaviour has been disabled by default, and the     patch introduces a new sysctl with which the     administrator can reenable it again.

  - CVE-2006-7203: The compat_sys_mount function in     fs/compat.c allows local users to cause a denial of     service (NULL pointer dereference and oops) by mounting     a smbfs file system in compatibility mode ('mount -t     smbfs').

  - CVE-2007-2453: Seeding of the kernel random generator on     boot did not work correctly due to a programming mistake     and so the kernel might have more predictable random     numbers than assured.

  - CVE-2007-2876: A NULL pointer dereference in SCTP     connection tracking could be caused by a remote attacker     by sending specially crafted packets. Note that this     requires SCTP set-up and active to be exploitable.

and the following non security bugs :

    - patches.fixes/cpufreq_fix_limited_on_battery.patch:
      Fix limited freq when booted on battery. [#231107]

    - patches.fixes/usb-keyspan-regression-fix.patch: USB:
      keyspan regression fix [#240919]

  - -     patches.fixes/hpt366-dont-check-enablebits-for-hpt36x.pa     tch: hpt366: don't check enablebits for HPT36x [#278696]

Some vulnerabilities were discovered and corrected in the Linux 2.6 kernel :

The Linux kernel did not properly save or restore EFLAGS during a context switch, or reset the flags when creating new threads, which allowed local users to cause a denial of service (process crash) (CVE-2006-5755).

The compat_sys_mount function in fs/compat.c allowed local users to cause a denial of service (NULL pointer dereference and oops) by mounting a smbfs file system in compatibility mode (CVE-2006-7203).

The nfnetlink_log function in netfilter allowed an attacker to cause a denial of service (crash) via unspecified vectors which would trigger a NULL pointer dereference (CVE-2007-1496).

The nf_conntrack function in netfilter did not set nfctinfo during reassembly of fragmented packets, which left the default value as IP_CT_ESTABLISHED and could allow remote attackers to bypass certain rulesets using IPv6 fragments (CVE-2007-1497).

The netlink functionality did not properly handle NETLINK_FIB_LOOKUP replies, which allowed a remote attacker to cause a denial of service (resource consumption) via unspecified vectors, probably related to infinite recursion (CVE-2007-1861).

A typo in the Linux kernel caused RTA_MAX to be used as an array size instead of RTN_MAX, which lead to an out of bounds access by certain functions (CVE-2007-2172).

The IPv6 protocol allowed remote attackers to cause a denial of service via crafted IPv6 type 0 route headers that create network amplification between two routers (CVE-2007-2242).

The random number feature did not properly seed pools when there was no entropy, or used an incorrect cast when extracting entropy, which could cause the random number generator to provide the same values after reboots on systems without an entropy source (CVE-2007-2453).

A memory leak in the PPPoE socket implementation allowed local users to cause a denial of service (memory consumption) by creating a socket using connect, and releasing it before the PPPIOCGCHAN ioctl is initialized (CVE-2007-2525).

An integer underflow in the cpuset_tasks_read function, when the cpuset filesystem is mounted, allowed local users to obtain kernel memory contents by using a large offset when reading the /dev/cpuset/tasks file (CVE-2007-2875).

The sctp_new function in netfilter allowed remote attackers to cause a denial of service by causing certain invalid states that triggered a NULL pointer dereference (CVE-2007-2876).

In addition to these security fixes, other fixes have been included such as :

  - Fix crash on netfilter when nfnetlink_log is used on     certain hooks on packets forwarded to or from a bridge

  - Fixed busy sleep on IPVS which caused high load averages

    - Fixed possible race condition on ext[34]_link

    - Fixed missing braces in condition block that led to       wrong behaviour in NFS

  - Fixed XFS lock deallocation that resulted in oops when     unmounting

To update your kernel, please follow the directions located at :

http://www.mandriva.com/en/security/kernelupdate

Several local and remote vulnerabilities have been discovered in the Linux kernel that may lead to a denial of service or the execution of arbitrary code. The Common Vulnerabilities and Exposures project identifies the following problems :

  - CVE-2007-1353     Ilja van Sprundel discovered that kernel memory could be     leaked via the Bluetooth setsockopt call due to an     uninitialized stack buffer. This could be used by local     attackers to read the contents of sensitive kernel     memory.

  - CVE-2007-2172     Thomas Graf reported a typo in the DECnet protocol     handler that could be used by a local attacker to     overrun an array via crafted packets, potentially     resulting in a Denial of Service (system crash). A     similar issue exists in the IPV4 protocol handler and     will be fixed in a subsequent update.

  - CVE-2007-2453     A couple of issues with random number generation were     discovered. Slightly less random numbers resulted from     hashing a subset of the available entropy. Zero-entropy     systems were seeded with the same inputs at boot time,     resulting in repeatable series of random numbers.

  - CVE-2007-2525     Florian Zumbiehl discovered a memory leak in the PPPOE     subsystem caused by releasing a socket before     PPPIOCGCHAN is called upon it. This could be used by a     local user to DoS a system by consuming all available     memory.

  - CVE-2007-2876     Vilmos Nebehaj discovered a NULL pointer dereference     condition in the netfilter subsystem. This allows remote     systems which communicate using the SCTP protocol to     crash a system by creating a connection with an unknown     chunk type.

  - CVE-2007-3513     Oliver Neukum reported an issue in the usblcd driver     which, by not limiting the size of write buffers,     permits local users with write access to trigger a DoS     by consuming all available memory.

  - CVE-2007-3642     Zhongling Wen reported an issue in nf_conntrack_h323     where the lack of range checking may lead to NULL     pointer dereferences. Remote attackers could exploit     this to create a DoS condition (system crash).

  - CVE-2007-3848     Wojciech Purczynski discovered that pdeath_signal was     not being reset properly under certain conditions which     may allow local users to gain privileges by sending     arbitrary signals to suid binaries.

  - CVE-2007-3851     Dave Airlie reported that Intel 965 and above chipsets     have relocated their batch buffer security bits. Local X     server users may exploit this to write user data to     arbitrary physical memory addresses.

These problems have been fixed in the stable distribution in version 2.6.18.dfsg.1-13etch1.

The following matrix lists additional packages that were rebuilt for compatibility with or to take advantage of this update :

                    Debian 4.0 (etch)    fai-kernels        1.17+etch4           user-mode-linux    2.6.18-1um-2etch3

Updated kernel packages that fix several security issues and bugs in the Red Hat Enterprise Linux 4 kernel are now available.

This security advisory has been rated as having important security impact by the Red Hat Security Response Team.

The Linux kernel handles the basic functions of the operating system.

These new kernel packages contain fixes for the security issues described below :

* a flaw in the connection tracking support for SCTP that allowed a remote user to cause a denial of service by dereferencing a NULL pointer. (CVE-2007-2876, Important)

* a flaw in the mount handling routine for 64-bit systems that allowed a local user to cause denial of service (crash). (CVE-2006-7203, Important)

* a flaw in the IPv4 forwarding base that allowed a local user to cause an out-of-bounds access. (CVE-2007-2172, Important)

* a flaw in the PPP over Ethernet implementation that allowed a local user to cause a denial of service (memory consumption) by creating a socket using connect and then releasing it before the PPPIOCGCHAN ioctl has been called. (CVE-2007-2525, Important)

* a flaw in the fput ioctl handling of 32-bit applications running on 64-bit platforms that allowed a local user to cause a denial of service (panic). (CVE-2007-0773, Important)

* a flaw in the NFS locking daemon that allowed a local user to cause denial of service (deadlock). (CVE-2006-5158, Moderate)

* a flaw in the sysfs_readdir function that allowed a local user to cause a denial of service by dereferencing a NULL pointer.
(CVE-2007-3104, Moderate)

* a flaw in the core-dump handling that allowed a local user to create core dumps from unreadable binaries via PT_INTERP. (CVE-2007-0958, Low)

* a flaw in the Bluetooth subsystem that allowed a local user to trigger an information leak. (CVE-2007-1353, Low)

In addition, the following bugs were addressed :

* the NFS could recurse on the same spinlock. Also, NFS, under certain conditions, did not completely clean up Posix locks on a file close, leading to mount failures.

* the 32bit compatibility didn't return to userspace correct values for the rt_sigtimedwait system call.

* the count for unused inodes could be incorrect at times, resulting in dirty data not being written to disk in a timely manner.

* the cciss driver had an incorrect disk size calculation (off-by-one error) which prevented disk dumps.

Red Hat would like to thank Ilja van Sprundel and the OpenVZ Linux kernel team for reporting issues fixed in this erratum.

All Red Hat Enterprise Linux 4 users are advised to upgrade their kernels to the packages associated with their machine architectures and configurations as listed in this erratum.

There is a flaw in the SCTP code included in Linux kernel versions before 2.6.21.4 that results in a kernel panic when an SCTP packet with an unknown chunk type is received. An attacker can leverage this flaw to crash the remote host with a single, possibly forged, packet.

The remote CentOS system is missing a security update which has been documented in Red Hat advisory RHSA-2007-0705.

The remote CentOS system is missing a security update which has been documented in Red Hat advisory RHSA-2007-0488.

ID	Name	Product	Family	Severity
67543	Oracle Linux 5 : kernel (ELSA-2007-0705)	Nessus	Oracle Linux Local Security Checks	medium
67520	Oracle Linux 4 : kernel (ELSA-2007-0488)	Nessus	Oracle Linux Local Security Checks	medium
60215	Scientific Linux Security Update : kernel on SL4.x i386/x86_64	Nessus	Scientific Linux Local Security Checks	medium
59123	SuSE 10 Security Update : Linux kernel (ZYPP Patch Number 4186)	Nessus	SuSE Local Security Checks	high
43648	CentOS 5 : kernel (CESA-2007:0705)	Nessus	CentOS Local Security Checks	medium
29487	SuSE 10 Security Update : Linux kernel (ZYPP Patch Number 4185)	Nessus	SuSE Local Security Checks	high
28114	Ubuntu 7.04 : linux-source-2.6.20 vulnerabilities (USN-510-1)	Nessus	Ubuntu Local Security Checks	high
28090	Ubuntu 6.06 LTS : linux-source-2.6.15 vulnerability (USN-489-1)	Nessus	Ubuntu Local Security Checks	high
28087	Ubuntu 6.10 : linux-source-2.6.17 vulnerabilities (USN-486-1)	Nessus	Ubuntu Local Security Checks	high
27664	Fedora 7 : kernel-2.6.21-1.3228.fc7 (2007-0409)	Nessus	Fedora Local Security Checks	medium
27296	SuSE Security Update: Kernel Update for SUSE Linux 10.1 (kernel-4193)	Nessus	SuSE Local Security Checks	high
27295	openSUSE 10 Security Update : kernel (kernel-3760)	Nessus	SuSE Local Security Checks	high
26050	RHEL 5 : kernel (RHSA-2007:0705)	Nessus	Red Hat Local Security Checks	medium
25968	Mandrake Linux Security Advisory : kernel (MDKSA-2007:171)	Nessus	Mandriva Local Security Checks	high
25909	Debian DSA-1356-1 : linux-2.6 - several vulnerabilities	Nessus	Debian Local Security Checks	high
25605	RHEL 4 : kernel (RHSA-2007:0488)	Nessus	Red Hat Local Security Checks	medium
25575	CentOS 4 : kernel (CESA-2007:0488)	Nessus	CentOS Local Security Checks	medium
25483	Linux Kernel Netfilter *_conntrack_proto_sctp.c sctp_new Function Unknown Chunk Type Remote DoS	Nessus	Denial of Service	medium
801434	CentOS RHSA-2007-0705 Security Check	Log Correlation Engine	Generic	high
801431	CentOS RHSA-2007-0488 Security Check	Log Correlation Engine	Generic	high

Detections

Analytics

CVE-2007-2876

high

Tenable Plugins

medium

medium

medium

high

medium

high

high

high

high

medium

high

high

medium

high

high

medium

medium

medium

high

high