LinuxLists.cc - [PATCH 00/24] Introduce credential record

2007-09-26 14:21:34

Subject: [PATCH 00/24] Introduce credential record

Hi Al, Christoph, Trond, Stephen, Casey,

Here's a set of patches that implement a very basic set of COW credentials. It
compiles, links and runs for x86_64 with EXT3, (V)FAT, NFS, AFS, SELinux and
keyrings all enabled. I've included a patch that should make most of the other
archs and filesystems work, but I haven't yet merged it into the primary
patches.

A tarball of the patches can be retrieved from:

http://people.redhat.com/~dhowells/nfs/nfs+fscache-24.tar.bz2

The cred struct contains the credentials that the kernel needs to act upon
something or to create something. Credentials that govern how a task may be
acted upon remain in the task struct.

In essence, the introduction of the cred struct separates a task's subjective
context (the authority with which it acts) from its objective context (the
authorisation required by others that want to act upon it), and permits
overriding of the subjective context by a kernel service so that the service
can act on the task's behalf to do something the task couldn't do on its own
authority.

Because keyrings and effective capabilities can be installed or changed in one
process by another process, they are shadowed by the cred structure rather than
residing there. Additionally, the session and process keyrings are shared
between all the threads of a process. The shadowing is performed by
update_current_cred() which is invoked on entry to any system call that might
need it.

A thread's cred struct may be read by that thread without any RCU precautions
as only that thread may replace the its own cred struct. To change a thread's
credentials, dup_cred() should be called to create a new copy, the copy should
be changed, and then set_current_cred() should be called to make it live. Once
live, it may not be changed as it may then be shared with file descriptors, RPC
calls and other threads. RCU will be used to dispose of the old structure.

The six patches are:

(1) Introduce struct cred and migrate fsuid, fsgid, the groups list and the
keyrings pointer to it.

(2) Introduce a security pointer into the cred struct and add LSM hooks to
duplicate the information pointed to thereby and to free it.

Make SELinux implement the hooks, splitting out some the task security
data to be associated with struct cred instead.

(3) Make the security functions that permit task SID retrieval return both the
objective and subjective SIDs as required.

(4) Migrate the effective capabilities mask into the cred struct.

(5) Fix up all the other archs and filesystems that I can manage to compile.
This should be merged into the preceding patches at some point.

(6) Provide a pair of LSM hooks so that a kernel service can (a) get a
credential record representing the authority with which it is permitted to
act, and (b) alter the file creation context in a credential record.

In addition, as this works with cachefiles, I've included all the FS-Cache,
CacheFiles, NFS and AFS patches.

To substitute a temporary set of credentials, the cred struct attached to the
task should be altered, like so:

int get_privileged_creds(...)
{
/* get special privileged creds */
my_special_cred = get_kernel_cred("cachefiles", current);
change_create_files_as(my_special_cred, my_cache_dir);
}

int do_stuff(...)
{
struct cred *cred;

/* rotate in the new creds, saving the old */
cred = __set_current_cred(get_cred(my_special_cred));

do_privileged_stuff();

/* restore the old creds */
set_current_cred(cred);
}

One thing I'm not certain about is how this should interact with /proc, which
can display some of the stuff in the cred struct. I think it may be necessary
to have a real cred pointer and an effective cred pointer, with the contents of
/proc coming from the real, but the effective governing what actually goes on.

Furthemore, I was thinking that it was a good idea to move the setting of i_uid
and i_gid to current->cred->i_[ug]id into new_inode(), but now I'm not so sure,
since the kernel special filesystems may assume that the i_uid and i_gid
default to 0. Any thoughts on this?

The NFS FS-Cache sharing patch still needs fixing up to correctly do the
sharing thing when local caching is enabled.

David

2007-09-26 14:21:59

Subject: [PATCH 00/24] Introduce credential record

Subject: [PATCH 03/24] CRED: Alter security_task_getsecid() and similar to return both task SIDs

Subject: [PATCH 02/24] CRED: Split the task security data and move part of it into struct cred

Subject: [PATCH 04/24] CRED: Move the effective capabilities into the cred struct

Subject: [PATCH 06/24] CRED: Request a credential record for a kernel service

Subject: [PATCH 08/24] FS-Cache: Recruit a couple of page flags for cache management

Subject: [PATCH 05/24] CRED: Fix up the other credentials references

Subject: [PATCH 07/24] FS-Cache: Release page->private after failed readahead

Subject: [PATCH 01/24] CRED: Introduce a COW credentials record

Subject: [PATCH 09/24] FS-Cache: Provide an add_wait_queue_tail() function

Subject: [PATCH 11/24] CacheFiles: Add missing copy_page export for ia64

Subject: [PATCH 13/24] CacheFiles: Permit the page lock state to be monitored

Subject: [PATCH 12/24] CacheFiles: Add a hook to write a single page of data to an inode

Subject: [PATCH 14/24] CacheFiles: Export things for CacheFiles

Subject: [PATCH 10/24] FS-Cache: Generic filesystem caching facility

Subject: [PATCH 15/24] CacheFiles: A cache that backs onto a mounted filesystem

Subject: [PATCH 19/24] AFS: Add TestSetPageError()

Subject: [PATCH 17/24] NFS: Configuration and mount option changes to enable local caching on NFS

Subject: [PATCH 16/24] NFS: Use local caching

Subject: [PATCH 18/24] NFS: Display local caching state

Subject: [PATCH 20/24] AFS: Add a function to excise a rejected write from the pagecache

Subject: [PATCH 21/24] AFS: Improve handling of a rejected writeback

Subject: [PATCH 23/24] AF_RXRPC: Save the operation ID for debugging

Subject: [PATCH 24/24] FS-Cache: Make kAFS use FS-Cache

Subject: [PATCH 22/24] AFS: Implement shared-writable mmap

Subject: Re: [PATCH 01/24] CRED: Introduce a COW credentials record

Subject: Re: [PATCH 01/24] CRED: Introduce a COW credentials record