linux-kernel - Re: [PATCH] Capabilities: BUG when an invalid capability is requested

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20081029162817.GB14705@us.ibm.com>
Date:	Wed, 29 Oct 2008 11:28:17 -0500
From:	"Serge E. Hallyn" <serue@...ibm.com>
To:	Eric Paris <eparis@...hat.com>
Cc:	linux-kernel@...r.kernel.org, morgan@...nel.org,
	arjan@...radead.org
Subject: Re: [PATCH] Capabilities: BUG when an invalid capability is
	requested

Quoting Eric Paris (eparis@...hat.com):
> If an invalid (large) capability is requested the capabilities system
> may panic as it is dereferencing an array of fixed (short) length.  Its
> possible (and actually often happens) that the capability system
> accidentally stumbled into a valid memory region but it also regularly
> happens that it hits invalid memory and BUGs.  If such an operation does
> get past cap_capable then the selinux system is sure to have problems as
> it already does a (simple) validity check and BUG.  This is known to
> happen by the broken and buggy firegl driver.
> 
> This patch cleanly checks all capable calls and BUG if a call is for an
> invalid capability.  This will likely break the firegl driver for some
> situations, but it is the right thing to do.  Garbage into a security
> system gets you killed/bugged
> 
> Signed-off-by: Eric Paris <eparis@...hat.com>

I really don't like this, but I'm not sure we really have a choice.

Acked-by: Serge Hallyn <serue@...ibm.com>

I suppose we can think later about whether it's worthwhile (a) having a
separate capable() function exported, keeping one without the check for
compiled-in use only, and/or (b) changing the cap_valid() definition to
be "(((unsigned int)cap) <= CAP_LAST_CAP)" which seems to work and shave
one whopping instruction.  I suspect the answer will be no to both.

Thanks, Eric.

-serge

> 
> ---
> 
>  kernel/capability.c |    5 +++++
>  1 files changed, 5 insertions(+), 0 deletions(-)
> 
> diff --git a/kernel/capability.c b/kernel/capability.c
> index 33e51e7..50d9d99 100644
> --- a/kernel/capability.c
> +++ b/kernel/capability.c
> @@ -498,6 +498,11 @@ asmlinkage long sys_capset(cap_user_header_t header, const cap_user_data_t data)
>   */
>  int capable(int cap)
>  {
> +	if (unlikely(!cap_valid(cap))) {
> +		printk(KERN_CRIT "capable() called with invalid cap=%u\n", cap);
> +		BUG();
> +	}
> +
>  	if (has_capability(current, cap)) {
>  		current->flags |= PF_SUPERPRIV;
>  		return 1;
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/