linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Eric Paris <eparis@parisplace.org>
To: Christoph Lameter <cl@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	"Larry H." <research@subreption.com>,
	Alan Cox <alan@lxorguk.ukuu.org.uk>,
	linux-mm@kvack.org, Rik van Riel <riel@redhat.com>,
	linux-kernel@vger.kernel.org, pageexec@freemail.hu,
	jmorris@namei.org
Subject: Re: Security fix for remapping of page 0 (was [PATCH] Change ZERO_SIZE_PTR to point at unmapped space)
Date: Wed, 3 Jun 2009 16:16:02 -0400	[thread overview]
Message-ID: <7e0fb38c0906031316n7aeed974xf15f8af5a3b04f63@mail.gmail.com> (raw)
In-Reply-To: <alpine.DEB.1.10.0906031602250.20254@gentwo.org>

On Wed, Jun 3, 2009 at 4:04 PM, Christoph Lameter
<cl@linux-foundation.org> wrote:
> On Wed, 3 Jun 2009, Eric Paris wrote:
>
>> The 'right'est fix is as Alan suggested, duplicate the code
>>
>> from security/capability.c::cap_file_mmap()
>> to include/linux/security.h::securitry_file_mmap()
>
> Thats easy to do but isnt it a bit weird now to configure mmap_min_addr?

??

> A security model may give it a different interpretation?

Not sure what you mean.  Yes, each security model is allowed to decide
what permissions are needed to pass a given security check.  SELinux
decided that CAP_SYS_RAWIO was not needed, but the selinux permission
mmap_zero was.  Had there been a more specific capability to use
SELinux might have been happy using a capability...

> What about round_hint_to_min()?

not sure what you mean....

>
> Use mmap_min_addr indepedently of security models
>
> This patch removes the dependency of mmap_min_addr on CONFIG_SECURITY.
> It also sets a default mmap_min_addr of 4096.
>
> mmapping of addresses below 4096 will only be possible for processes
> with CAP_SYS_RAWIO.

<pedantic nit> "or the appropriate permission for the given LSM </pedantic nit>

> Signed-off-by: Christoph Lameter <cl@linux-foundation.org>

Clearly lots more cleanup can be done between CONFIG_SECURITY and
!CONFIG_SECURITY like Linus suggested, but

Acked-by: Eric Paris <eparis@redhat.com>

> ---
>  include/linux/mm.h       |    2 --
>  include/linux/security.h |    2 ++
>  kernel/sysctl.c          |    2 --
>  mm/Kconfig               |   19 +++++++++++++++++++
>  mm/mmap.c                |    3 +++
>  security/Kconfig         |   20 --------------------
>  security/security.c      |    3 ---
>  7 files changed, 24 insertions(+), 27 deletions(-)
>
> Index: linux-2.6/include/linux/mm.h
> ===================================================================
> --- linux-2.6.orig/include/linux/mm.h   2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/include/linux/mm.h        2009-06-03 15:00:56.000000000 -0500
> @@ -580,12 +580,10 @@ static inline void set_page_links(struct
>  */
>  static inline unsigned long round_hint_to_min(unsigned long hint)
>  {
> -#ifdef CONFIG_SECURITY
>        hint &= PAGE_MASK;
>        if (((void *)hint != NULL) &&
>            (hint < mmap_min_addr))
>                return PAGE_ALIGN(mmap_min_addr);
> -#endif
>        return hint;
>  }
>
> Index: linux-2.6/kernel/sysctl.c
> ===================================================================
> --- linux-2.6.orig/kernel/sysctl.c      2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/kernel/sysctl.c   2009-06-03 15:00:56.000000000 -0500
> @@ -1225,7 +1225,6 @@ static struct ctl_table vm_table[] = {
>                .strategy       = &sysctl_jiffies,
>        },
>  #endif
> -#ifdef CONFIG_SECURITY
>        {
>                .ctl_name       = CTL_UNNUMBERED,
>                .procname       = "mmap_min_addr",
> @@ -1234,7 +1233,6 @@ static struct ctl_table vm_table[] = {
>                .mode           = 0644,
>                .proc_handler   = &proc_doulongvec_minmax,
>        },
> -#endif
>  #ifdef CONFIG_NUMA
>        {
>                .ctl_name       = CTL_UNNUMBERED,
> Index: linux-2.6/mm/mmap.c
> ===================================================================
> --- linux-2.6.orig/mm/mmap.c    2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/mm/mmap.c 2009-06-03 15:01:18.000000000 -0500
> @@ -87,6 +87,9 @@ int sysctl_overcommit_ratio = 50;     /* def
>  int sysctl_max_map_count __read_mostly = DEFAULT_MAX_MAP_COUNT;
>  struct percpu_counter vm_committed_as;
>
> +/* amount of vm to protect from userspace access */
> +unsigned long mmap_min_addr = CONFIG_DEFAULT_MMAP_MIN_ADDR;
> +
>  /*
>  * Check that a process has enough memory to allocate a new virtual
>  * mapping. 0 means there is enough memory for the allocation to
> Index: linux-2.6/security/security.c
> ===================================================================
> --- linux-2.6.orig/security/security.c  2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/security/security.c       2009-06-03 15:00:56.000000000 -0500
> @@ -26,9 +26,6 @@ extern void security_fixup_ops(struct se
>
>  struct security_operations *security_ops;      /* Initialized to NULL */
>
> -/* amount of vm to protect from userspace access */
> -unsigned long mmap_min_addr = CONFIG_SECURITY_DEFAULT_MMAP_MIN_ADDR;
> -
>  static inline int verify(struct security_operations *ops)
>  {
>        /* verify the security_operations structure exists */
> Index: linux-2.6/mm/Kconfig
> ===================================================================
> --- linux-2.6.orig/mm/Kconfig   2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/mm/Kconfig        2009-06-03 15:00:56.000000000 -0500
> @@ -226,6 +226,25 @@ config HAVE_MLOCKED_PAGE_BIT
>  config MMU_NOTIFIER
>        bool
>
> +config DEFAULT_MMAP_MIN_ADDR
> +        int "Low address space to protect from user allocation"
> +        default 4096
> +        help
> +         This is the portion of low virtual memory which should be protected
> +         from userspace allocation.  Keeping a user from writing to low pages
> +         can help reduce the impact of kernel NULL pointer bugs.
> +
> +         For most ia64, ppc64 and x86 users with lots of address space
> +         a value of 65536 is reasonable and should cause no problems.
> +         On arm and other archs it should not be higher than 32768.
> +         Programs which use vm86 functionality would either need additional
> +         permissions from either the LSM or the capabilities module or have
> +         this protection disabled.
> +
> +         This value can be changed after boot using the
> +         /proc/sys/vm/mmap_min_addr tunable.
> +
> +
>  config NOMMU_INITIAL_TRIM_EXCESS
>        int "Turn on mmap() excess space trimming before booting"
>        depends on !MMU
> Index: linux-2.6/security/Kconfig
> ===================================================================
> --- linux-2.6.orig/security/Kconfig     2009-06-03 15:00:54.000000000 -0500
> +++ linux-2.6/security/Kconfig  2009-06-03 15:00:56.000000000 -0500
> @@ -113,26 +113,6 @@ config SECURITY_ROOTPLUG
>
>          If you are unsure how to answer this question, answer N.
>
> -config SECURITY_DEFAULT_MMAP_MIN_ADDR
> -        int "Low address space to protect from user allocation"
> -        depends on SECURITY
> -        default 0
> -        help
> -         This is the portion of low virtual memory which should be protected
> -         from userspace allocation.  Keeping a user from writing to low pages
> -         can help reduce the impact of kernel NULL pointer bugs.
> -
> -         For most ia64, ppc64 and x86 users with lots of address space
> -         a value of 65536 is reasonable and should cause no problems.
> -         On arm and other archs it should not be higher than 32768.
> -         Programs which use vm86 functionality would either need additional
> -         permissions from either the LSM or the capabilities module or have
> -         this protection disabled.
> -
> -         This value can be changed after boot using the
> -         /proc/sys/vm/mmap_min_addr tunable.
> -
> -
>  source security/selinux/Kconfig
>  source security/smack/Kconfig
>  source security/tomoyo/Kconfig
> Index: linux-2.6/include/linux/security.h
> ===================================================================
> --- linux-2.6.orig/include/linux/security.h     2009-06-03 15:01:28.000000000 -0500
> +++ linux-2.6/include/linux/security.h  2009-06-03 15:01:42.000000000 -0500
> @@ -2197,6 +2197,8 @@ static inline int security_file_mmap(str
>                                     unsigned long addr,
>                                     unsigned long addr_only)
>  {
> +       if ((addr < mmap_min_addr) && !capable(CAP_SYS_RAWIO))
> +               return -EACCES;
>        return 0;
>  }
>
>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2009-06-03 20:16 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-05-30 19:28 [PATCH] Change ZERO_SIZE_PTR to point at unmapped space Larry H.
2009-05-30 22:29 ` Linus Torvalds
2009-05-30 23:00   ` Larry H.
2009-05-31  2:02     ` Linus Torvalds
2009-05-31  2:21       ` Larry H.
2009-06-02 15:37         ` Christoph Lameter
2009-06-02 20:34           ` Larry H.
2009-06-03 14:50             ` Security fix for remapping of page 0 (was [PATCH] Change ZERO_SIZE_PTR to point at unmapped space) Christoph Lameter
2009-06-03 15:07               ` Linus Torvalds
2009-06-03 15:23                 ` Christoph Lameter
2009-06-03 15:38                   ` Linus Torvalds
2009-06-03 16:14                     ` Alan Cox
2009-06-03 16:19                       ` Linus Torvalds
2009-06-03 16:24                         ` Eric Paris
2009-06-03 16:22                     ` Eric Paris
2009-06-03 16:28                       ` Linus Torvalds
2009-06-03 16:32                         ` Eric Paris
2009-06-03 16:44                           ` Linus Torvalds
2009-06-03 15:11               ` Stephen Smalley
2009-06-03 15:41                 ` Christoph Lameter
2009-06-03 16:18                   ` Linus Torvalds
2009-06-03 16:28                   ` Larry H.
2009-06-03 16:36                     ` Rik van Riel
2009-06-03 16:47                       ` Linus Torvalds
2009-06-03 17:16                         ` Eric Paris
2009-06-03 17:28                           ` Linus Torvalds
2009-06-03 17:31                             ` Eric Paris
2009-06-03 17:24                         ` Larry H.
2009-06-03 17:21                       ` Larry H.
2009-06-03 22:52                         ` James Morris
2009-06-03 17:29               ` Alan Cox
2009-06-03 17:35                 ` Linus Torvalds
2009-06-03 18:00                   ` Larry H.
2009-06-03 18:12                     ` Linus Torvalds
2009-06-03 18:39                       ` Larry H.
2009-06-03 18:45                         ` Linus Torvalds
2009-06-03 18:50                           ` Linus Torvalds
2009-06-03 18:59                             ` Christoph Lameter
2009-06-03 19:11                               ` Rik van Riel
2009-06-03 19:14                               ` Eric Paris
2009-06-03 19:42                                 ` Christoph Lameter
2009-06-03 19:51                                   ` Eric Paris
2009-06-03 20:04                                     ` Christoph Lameter
2009-06-03 20:16                                       ` Eric Paris [this message]
2009-06-03 20:36                                         ` Christoph Lameter
2009-06-03 21:20                                       ` Linus Torvalds
2009-06-04  2:41                                       ` James Morris
2009-06-03 19:21                               ` Alan Cox
2009-06-03 19:45                                 ` Christoph Lameter
2009-06-03 21:07                                   ` Alan Cox
2009-06-03 19:27                               ` Linus Torvalds
2009-06-03 19:50                                 ` Christoph Lameter
2009-06-03 20:00                             ` pageexec
2009-06-03 19:41                           ` pageexec
2009-06-07 10:29               ` Pavel Machek
2009-05-30 22:32 ` [PATCH] Change ZERO_SIZE_PTR to point at unmapped space Peter Zijlstra
2009-05-30 22:51   ` Larry H.

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=7e0fb38c0906031316n7aeed974xf15f8af5a3b04f63@mail.gmail.com \
    --to=eparis@parisplace.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=cl@linux-foundation.org \
    --cc=jmorris@namei.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=pageexec@freemail.hu \
    --cc=research@subreption.com \
    --cc=riel@redhat.com \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox