Re: Append vs. write distinctions

Previous message: jmjonesat_private: "Extending a Security Module"
In reply to: David Wagner: "Re: Append vs. write distinctions"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

chrisat_private

On 20 May 2001, David Wagner wrote:

> Chris Evans  wrote:
> >On Sat, 19 May 2001, Crispin Cowan wrote:
> >> Does the LSM call to the permissions function specify the calling process?
> >
> >What about the global "current"?
>
> Is this safe?  Is there any place in the kernel that calls sys_read()
> without being in a valid user context?  If it doesn't happen today,
> is there a chance it might happen tomorrow (or in some loadable kernel
> module outside our control)?  Come to think of it, does this issue
> come up anywhere else?

Most syscalls code paths are generally riddled with the assumption that
there is a valid user context. For example, capable() references current.

The read example you highlight is interesting. I don't know of any
sys_read() usages within the kernel, but some places definitely use the
LSM hooked fops->read(),write(). Examples would be quota handling and
process accounting.

Cheers
Chris

_______________________________________________
linux-security-module mailing list
linux-security-moduleat_private
http://mail.wirex.com/mailman/listinfo/linux-security-module