2010-04-26 10:58:57

by David Howells

Subject: [PATCH] CRED: Holding a spinlock does not imply the holding of RCU read lock

From: Serge E. Hallyn <[email protected]>

Change the credentials documentation to make it clear that the RCU read lock
must be explicitly held when accessing credentials pointers in some other task
than current. Holding a spinlock does not implicitly hold the RCU read lock.

Signed-off-by: Serge Hallyn <[email protected]>
Reviewed-by: Paul E. McKenney <[email protected]>
Signed-off-by: David Howells <[email protected]>
---

Documentation/credentials.txt | 14 +++++---------
1 files changed, 5 insertions(+), 9 deletions(-)

diff --git a/Documentation/credentials.txt b/Documentation/credentials.txt
index df03169..a2db352 100644
--- a/Documentation/credentials.txt
+++ b/Documentation/credentials.txt
@@ -408,9 +408,6 @@ This should be used inside the RCU read lock, as in the following example:
...
}

-A function need not get RCU read lock to use __task_cred() if it is holding a
-spinlock at the time as this implicitly holds the RCU read lock.
-
Should it be necessary to hold another task's credentials for a long period of
time, and possibly to sleep whilst doing so, then the caller should get a
reference on them using:
@@ -426,17 +423,16 @@ credentials, hiding the RCU magic from the caller:
uid_t task_uid(task) Task's real UID
uid_t task_euid(task) Task's effective UID

-If the caller is holding a spinlock or the RCU read lock at the time anyway,
-then:
+If the caller is holding the RCU read lock at the time anyway, then:

__task_cred(task)->uid
__task_cred(task)->euid

should be used instead. Similarly, if multiple aspects of a task's credentials
-need to be accessed, RCU read lock or a spinlock should be used, __task_cred()
-called, the result stored in a temporary pointer and then the credential
-aspects called from that before dropping the lock. This prevents the
-potentially expensive RCU magic from being invoked multiple times.
+need to be accessed, RCU read lock should be used, __task_cred() called, the
+result stored in a temporary pointer and then the credential aspects called
+from that before dropping the lock. This prevents the potentially expensive
+RCU magic from being invoked multiple times.

Should some other single aspect of another task's credentials need to be
accessed, then this can be used:

2010-04-26 22:50:21

by James Morris

Subject: Re: [PATCH] CRED: Holding a spinlock does not imply the holding of RCU read lock

On Mon, 26 Apr 2010, David Howells wrote:

> From: Serge E. Hallyn <[email protected]>
>
> Change the credentials documentation to make it clear that the RCU read lock
> must be explicitly held when accessing credentials pointers in some other task
> than current. Holding a spinlock does not implicitly hold the RCU read lock.
>
> Signed-off-by: Serge Hallyn <[email protected]>
> Reviewed-by: Paul E. McKenney <[email protected]>
> Signed-off-by: David Howells <[email protected]>


Applied to
git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6#next


--
James Morris
<[email protected]>