[patch] Minor fixes for IPC hooks

Previous message: Stephen Smalley: "Re: Secure Computing statement of assurance"
Next in thread: James Morris: "Re: [patch] Minor fixes for IPC hooks"
Reply: James Morris: "Re: [patch] Minor fixes for IPC hooks"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

sdsat_private

In reviewing the System V IPC hooks again prior to submitting them, I
noticed what appear to be a few oversights in calling free_security on
error return paths after a successful alloc_security call.  I also noticed
that load_msg needs to initialize the msg->security much earlier in the
function so that the error path will not end up calling free_security with
an uninitialized security field.  The attached patches fix these
oversights.  For load_msg, I only moved the clearing of the security
field, as this should be sufficient for free_security to safely perform a
test, and some modules may not want the alloc_security call to occur until
after the load has completed.  Any objections to committing these patches
to the lsm-2.4 and lsm-2.5 BitKeeper trees?

--
Stephen D. Smalley, NAI Labs
ssmalleyat_private