Unspecified versions of the Linux kernel allow local users to cause a denial of service (unrecoverable zombie process) via a program with certain instructions that prevent init from properly reaping a child whose parent has died.

The remote Oracle Linux 5 host has packages installed that are affected by multiple vulnerabilities as referenced in the ELSA-2008-0154 advisory.

    [2.6.18-53.1.14.0.1.el5]
     - [NET] Add entropy support to e1000 and bnx2 (John Sobecki) [orabug      6045759]
     - [NET] Fix msi issue with kexec/kdump (Michael Chan) [orabug 6219364]
     - [MM] Fix alloc_pages_node() static `nid' race made kernel crash (Joe      Jin) [orabug 6187457]
     - [splice] Fix bad unlock_page() in error case  (Jens Axboe) [orabug      6263574]
     - [dio] fix error-path crashes (Linus Torvalds) [orabug 6242289]

     [2.6.18-53.1.14.el5]
     - merge from 2.6.18-53.1.13 to 2.6.18-53.1.12
     - [nfs] potential file corruption issue when writing (Jeff Layton ) [432078]
     - [ppc] chrp: fix possible strncmp NULL pointer usage (Vitaly Mayatskikh      ) [396821]
     - [isdn] i4l: fix memory overruns (Vitaly Mayatskikh ) [425171]
     - [isdn] fix possible isdn_net buffer overflows (Aristeu Rozanski )      [392151] {CVE-2007-6063}
     - [mm] hugepages: leak due to pagetable page sharing (Larry Woodman )      [431522]
     - [net] NULL dereference in iwl driver (Vitaly Mayatskikh ) [401421]      {CVE-2007-5938}
     - [misc] Denial of service with wedged processes (Jerome Marchand ) [221403]
     - [xen] ia64: hvm guest memory range checking (Jarod Wilson ) [408701]

Tenable has extracted the preceding description block directly from the Oracle Linux security advisory.

Note that Nessus has not tested for these issues but has instead relied only on the application's self-reported version number.

From Red Hat Security Advisory 2007:0939 :

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

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

The Linux kernel is the core of the operating system.

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

* A flaw was found in the handling of process death signals. This allowed a local user to send arbitrary signals to the suid-process executed by that user. A successful exploitation of this flaw depends on the structure of the suid-program and its signal handling.
(CVE-2007-3848, Important)

* A flaw was found in the CIFS file system. This could cause the umask values of a process to not be honored on CIFS file systems where UNIX extensions are supported. (CVE-2007-3740, Important)

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

* A flaw was found in the Advanced Linux Sound Architecture (ALSA). A local user who had the ability to read the /proc/driver/snd-page-alloc file could see portions of kernel memory. (CVE-2007-4571, Moderate)

* A flaw was found in the aacraid SCSI driver. This allowed a local user to make ioctl calls to the driver that should be restricted to privileged users. (CVE-2007-4308, Moderate)

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

* A flaw was found in the handling of zombie processes. A local user could create processes that would not be properly reaped which could lead to a denial of service. (CVE-2006-6921, Moderate)

* A flaw was found in the CIFS file system handling. The mount option 'sec=' did not enable integrity checking or produce an error message if used. (CVE-2007-3843, Low)

* A flaw was found in the random number generator implementation that allowed a local user to cause a denial of service or possibly gain privileges. This flaw could be exploited if the root user raised the default wakeup threshold over the size of the output pool.
(CVE-2007-3105, Low)

Additionally, the following bugs were fixed :

* A flaw was found in the kernel netpoll code, creating a potential deadlock condition. If the xmit_lock for a given network interface is held, and a subsequent netpoll event is generated from within the lock owning context (a console message for example), deadlock on that cpu will result, because the netpoll code will attempt to re-acquire the xmit_lock. The fix is to, in the netpoll code, only attempt to take the lock, and fail if it is already acquired (rather than block on it), and queue the message to be sent for later delivery. Any user of netpoll code in the kernel (netdump or netconsole services), is exposed to this problem, and should resolve the issue by upgrading to this kernel release immediately.

* A flaw was found where, under 64-bit mode (x86_64), AMD processors were not able to address greater than a 40-bit physical address space;
and Intel processors were only able to address up to a 36-bit physical address space. The fix is to increase the physical addressing for an AMD processor to 48 bits, and an Intel processor to 38 bits. Please see the Red Hat Knowledgebase for more detailed information.

* A flaw was found in the xenU kernel that may prevent a paravirtualized guest with more than one CPU from starting when running under an Enterprise Linux 5.1 hypervisor. The fix is to allow your Enterprise Linux 4 Xen SMP guests to boot under a 5.1 hypervisor.
Please see the Red Hat Knowledgebase for more detailed information.

Red Hat Enterprise Linux 4 users are advised to upgrade to these updated packages, which contain backported patches to correct these issues.

These updated packages fix the following security issues :

  - a flaw in the hypervisor for hosts running on Itanium     architectures allowed an Intel VTi domain to read     arbitrary physical memory from other Intel VTi domains,     which could make information available to unauthorized     users. (CVE-2007-6207, Important)

  - two buffer overflow flaws were found in ISDN subsystem.
    A local unprivileged user could use these flaws to cause     a denial of service. (CVE-2007-5938: Important,     CVE-2007-6063: Moderate)

  - a possible NULL pointer dereference was found in the     subsystem used for showing CPU information, as used by     CHRP systems on PowerPC architectures. This may have     allowed a local unprivileged user to cause a denial of     service (crash). (CVE-2007-6694, Moderate)

  - a flaw was found in the handling of zombie processes. A     local user could create processes that would not be     properly reaped, possibly causing a denial of service.
    (CVE-2006-6921, Moderate)

As well, these updated packages fix the following bugs :

  - a bug was found in the Linux kernel audit subsystem.
    When the audit daemon was setup to log the execve system     call with a large number of arguments, the kernel could     run out of memory, causing a kernel panic.

  - on IBM System z architectures, using the IBM Hardware     Management Console to toggle IBM FICON channel path ids     (CHPID) caused a file ID miscompare, possibly causing     data corruption.

  - when running the IA-32 Execution Layer (IA-32EL) or a     Java VM on Itanium architectures, a bug in the address     translation in the hypervisor caused the wrong address     to be registered, causing Dom0 to hang.

  - on Itanium architectures, frequent Corrected Platform     Error errors may have caused the hypervisor to hang.

  - when enabling a CPU without hot plug support, routines     for checking the presence of the CPU were missing. The     CPU tried to access its own resources, causing a kernel     panic.

  - after updating to kernel-2.6.18-53.el5, a bug in the     CCISS driver caused the HP Array Configuration Utility     CLI to become unstable, possibly causing a system hang,     or a kernel panic.

  - a bug in NFS directory caching could have caused     different hosts to have different views of NFS     directories.

  - on Itanium architectures, the Corrected Machine Check     Interrupt masked hot-added CPUs as disabled.

  - when running Oracle database software on the Intel 64     and AMD64 architectures, if an SGA larger than 4GB was     created, and had hugepages allocated to it, the     hugepages were not freed after database shutdown.

  - in a clustered environment, when two or more NFS clients     had the same logical volume mounted, and one of them     modified a file on the volume, NULL characters may have     been inserted, possibly causing data corruption.

These updated packages resolve several severe issues in the lpfc driver :

  - a system hang after LUN discovery.

  - a general fault protection, a NULL pointer dereference,     or slab corruption could occur while running a debug on     the kernel.

  - the inability to handle kernel paging requests in     'lpfc_get_scsi_buf'.

  - erroneous structure references caused certain FC     discovery routines to reference and change     'lpfc_nodelist' structures, even after they were freed.

  - the lpfc driver failed to interpret certain fields     correctly, causing tape backup software to fail. Tape     drives reported 'Illegal Request'.

  - the lpfc driver did not clear structures correctly,     resulting in SCSI I/Os being rejected by targets, and     causing errors.

- A flaw was found in the handling of process death     signals. This allowed a local user to send arbitrary     signals to the suid-process executed by that user. A     successful exploitation of this flaw depends on the     structure of the suid-program and its signal handling.
    (CVE-2007-3848, Important)

  - A flaw was found in the CIFS file system. This could     cause the umask values of a process to not be honored on     CIFS file systems where UNIX extensions are supported.
    (CVE-2007-3740, Important)

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

  - A flaw was found in the Advanced Linux Sound     Architecture (ALSA). A local user who had the ability to     read the /proc/driver/snd-page-alloc file could see     portions of kernel memory. (CVE-2007-4571, Moderate)

  - A flaw was found in the aacraid SCSI driver. This     allowed a local user to make ioctl calls to the driver     that should be restricted to privileged users.
    (CVE-2007-4308, Moderate)

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

  - A flaw was found in the handling of zombie processes. A     local user could create processes that would not be     properly reaped which could lead to a denial of service.
    (CVE-2006-6921, Moderate)

  - A flaw was found in the CIFS file system handling. The     mount option 'sec=' did not enable integrity checking or     produce an error message if used. (CVE-2007-3843, Low)

  - A flaw was found in the random number generator     implementation that allowed a local user to cause a     denial of service or possibly gain privileges. This flaw     could be exploited if the root user raised the default     wakeup threshold over the size of the output pool.
    (CVE-2007-3105, Low)

Additionally, the following bugs were fixed :

  - A flaw was found in the kernel netpoll code, creating a     potential deadlock condition. If the xmit_lock for a     given network interface is held, and a subsequent     netpoll event is generated from within the lock owning     context (a console message for example), deadlock on     that cpu will result, because the netpoll code will     attempt to re-acquire the xmit_lock. The fix is to, in     the netpoll code, only attempt to take the lock, and     fail if it is already acquired (rather than block on     it), and queue the message to be sent for later     delivery. Any user of netpoll code in the kernel     (netdump or netconsole services), is exposed to this     problem, and should resolve the issue by upgrading to     this kernel release immediately.

  - A flaw was found where, under 64-bit mode (x86_64), AMD     processors were not able to address greater than a     40-bit physical address space; and Intel processors were     only able to address up to a 36-bit physical address     space. The fix is to increase the physical addressing     for an AMD processor to 48 bits, and an Intel processor     to 38 bits.

  - A flaw was found in the xenU kernel that may prevent a     paravirtualized guest with more than one CPU from     starting when running under an Scientific Linux 5.1     hypervisor. The fix is to allow your Scientific Linux 4     Xen SMP guests to boot under a 5.1 hypervisor.

Updated kernel packages that fix various security issues and several bugs are now available for Red Hat Enterprise Linux 5.

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

The kernel packages contain the Linux kernel, the core of any Linux operating system.

These updated packages fix the following security issues :

* a flaw in the hypervisor for hosts running on Itanium architectures allowed an Intel VTi domain to read arbitrary physical memory from other Intel VTi domains, which could make information available to unauthorized users. (CVE-2007-6207, Important)

* two buffer overflow flaws were found in ISDN subsystem. A local unprivileged user could use these flaws to cause a denial of service.
(CVE-2007-5938: Important, CVE-2007-6063: Moderate)

* a possible NULL pointer dereference was found in the subsystem used for showing CPU information, as used by CHRP systems on PowerPC architectures. This may have allowed a local unprivileged user to cause a denial of service (crash). (CVE-2007-6694, Moderate)

* a flaw was found in the handling of zombie processes. A local user could create processes that would not be properly reaped, possibly causing a denial of service. (CVE-2006-6921, Moderate)

As well, these updated packages fix the following bugs :

* a bug was found in the Linux kernel audit subsystem. When the audit daemon was setup to log the execve system call with a large number of arguments, the kernel could run out of memory, causing a kernel panic.

* on IBM System z architectures, using the IBM Hardware Management Console to toggle IBM FICON channel path ids (CHPID) caused a file ID miscompare, possibly causing data corruption.

* when running the IA-32 Execution Layer (IA-32EL) or a Java VM on Itanium architectures, a bug in the address translation in the hypervisor caused the wrong address to be registered, causing Dom0 to hang.

* on Itanium architectures, frequent Corrected Platform Error errors may have caused the hypervisor to hang.

* when enabling a CPU without hot plug support, routines for checking the presence of the CPU were missing. The CPU tried to access its own resources, causing a kernel panic.

* after updating to kernel-2.6.18-53.el5, a bug in the CCISS driver caused the HP Array Configuration Utility CLI to become unstable, possibly causing a system hang, or a kernel panic.

* a bug in NFS directory caching could have caused different hosts to have different views of NFS directories.

* on Itanium architectures, the Corrected Machine Check Interrupt masked hot-added CPUs as disabled.

* when running Oracle database software on the Intel 64 and AMD64 architectures, if an SGA larger than 4GB was created, and had hugepages allocated to it, the hugepages were not freed after database shutdown.

* in a clustered environment, when two or more NFS clients had the same logical volume mounted, and one of them modified a file on the volume, NULL characters may have been inserted, possibly causing data corruption.

These updated packages resolve several severe issues in the lpfc driver :

* a system hang after LUN discovery.

* a general fault protection, a NULL pointer dereference, or slab corruption could occur while running a debug on the kernel.

* the inability to handle kernel paging requests in 'lpfc_get_scsi_buf'.

* erroneous structure references caused certain FC discovery routines to reference and change 'lpfc_nodelist' structures, even after they were freed.

* the lpfc driver failed to interpret certain fields correctly, causing tape backup software to fail. Tape drives reported 'Illegal Request'.

* the lpfc driver did not clear structures correctly, resulting in SCSI I/Os being rejected by targets, and causing errors.

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

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

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

The Linux kernel is the core of the operating system.

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

* A flaw was found in the handling of process death signals. This allowed a local user to send arbitrary signals to the suid-process executed by that user. A successful exploitation of this flaw depends on the structure of the suid-program and its signal handling.
(CVE-2007-3848, Important)

* A flaw was found in the CIFS file system. This could cause the umask values of a process to not be honored on CIFS file systems where UNIX extensions are supported. (CVE-2007-3740, Important)

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

* A flaw was found in the Advanced Linux Sound Architecture (ALSA). A local user who had the ability to read the /proc/driver/snd-page-alloc file could see portions of kernel memory. (CVE-2007-4571, Moderate)

* A flaw was found in the aacraid SCSI driver. This allowed a local user to make ioctl calls to the driver that should be restricted to privileged users. (CVE-2007-4308, Moderate)

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

* A flaw was found in the handling of zombie processes. A local user could create processes that would not be properly reaped which could lead to a denial of service. (CVE-2006-6921, Moderate)

* A flaw was found in the CIFS file system handling. The mount option 'sec=' did not enable integrity checking or produce an error message if used. (CVE-2007-3843, Low)

* A flaw was found in the random number generator implementation that allowed a local user to cause a denial of service or possibly gain privileges. This flaw could be exploited if the root user raised the default wakeup threshold over the size of the output pool.
(CVE-2007-3105, Low)

Additionally, the following bugs were fixed :

* A flaw was found in the kernel netpoll code, creating a potential deadlock condition. If the xmit_lock for a given network interface is held, and a subsequent netpoll event is generated from within the lock owning context (a console message for example), deadlock on that cpu will result, because the netpoll code will attempt to re-acquire the xmit_lock. The fix is to, in the netpoll code, only attempt to take the lock, and fail if it is already acquired (rather than block on it), and queue the message to be sent for later delivery. Any user of netpoll code in the kernel (netdump or netconsole services), is exposed to this problem, and should resolve the issue by upgrading to this kernel release immediately.

* A flaw was found where, under 64-bit mode (x86_64), AMD processors were not able to address greater than a 40-bit physical address space;
and Intel processors were only able to address up to a 36-bit physical address space. The fix is to increase the physical addressing for an AMD processor to 48 bits, and an Intel processor to 38 bits. Please see the Red Hat Knowledgebase for more detailed information.

* A flaw was found in the xenU kernel that may prevent a paravirtualized guest with more than one CPU from starting when running under an Enterprise Linux 5.1 hypervisor. The fix is to allow your Enterprise Linux 4 Xen SMP guests to boot under a 5.1 hypervisor.
Please see the Red Hat Knowledgebase for more detailed information.

Red Hat Enterprise Linux 4 users are advised to upgrade to these updated packages, which contain backported patches to correct these issues.

The remote CentOS system is missing a security update which has been documented in Red Hat advisory RHSA-2008-0154.

The remote CentOS system is missing a security update which has been documented in Red Hat advisory RHSA-2007-0939.

ID	Name	Product	Family	Severity
67659	Oracle Linux 5 : Important: / kernel (ELSA-2008-0154)	Nessus	Oracle Linux Local Security Checks	high
67580	Oracle Linux 4 : kernel (ELSA-2007-0939)	Nessus	Oracle Linux Local Security Checks	medium
60370	Scientific Linux Security Update : kernel on SL5.x i386/x86_64	Nessus	Scientific Linux Local Security Checks	high
60280	Scientific Linux Security Update : kernel on SL4.x i386/x86_64	Nessus	Scientific Linux Local Security Checks	medium
43674	CentOS 4 / 5 : kernel (CESA-2008:0154)	Nessus	CentOS Local Security Checks	high
37953	CentOS 4 : kernel (CESA-2007:0939)	Nessus	CentOS Local Security Checks	medium
31388	RHEL 5 : kernel (RHSA-2008:0154)	Nessus	Red Hat Local Security Checks	high
27616	RHEL 4 : kernel (RHSA-2007:0939)	Nessus	Red Hat Local Security Checks	medium
801447	CentOS RHSA-2008-0154 Security Check	Log Correlation Engine	Generic	high
801439	CentOS RHSA-2007-0939 Security Check	Log Correlation Engine	Generic	high

Detections

Analytics

CVE-2006-6921

medium

Tenable Plugins

high

medium

high

medium

high

medium

high

medium

high

high