Re: [PATCH] sgi-xp: use designated initializers

From: Kees Cook
Date: Sat May 06 2017 - 03:16:17 EST

On Wed, Mar 29, 2017 at 1:48 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> On Tue, Jan 3, 2017 at 3:19 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>> On Wed, Dec 21, 2016 at 8:24 AM, Robin Holt <robinmholt@xxxxxxxxx> wrote:
>>> On Fri, Dec 16, 2016 at 7:01 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>>>> Prepare to mark sensitive kernel structures for randomization by making
>>>> sure they're using designated initializers. These were identified during
>>>> allyesconfig builds of x86, arm, and arm64, with most initializer fixes
>>>> extracted from grsecurity.
>>> I guess I don't understand the context enough here to give you a
>>> Signed-off-by. Can you give us more background on this randomization?
>> Sure thing! The randomization is on the order of function pointers in
>> all-pointer structures (like struct xpc_interface). As long as the
>> memory containing the structure isn't shared externally, this
>> randomization should have no operational effect. The reason explicit
>> no-op functions were added was to avoid ugly casts, etc.
> Friendly ping ... any chance this can land in -next soon?
>>> From what I see in the code here, I can see you are providing
>>> equivalent functionality and I would give it a signed-off-by, but I am
>>> not sure this randomization of which you speak is not going to cause
>>> problems for XP, XPC, XPNET, and XPMEM (out of tree GPL kernel
>>> module).
>> Ah, hm, does this module share the structure without being built
>> against the kernel? (If built with the kernel, the randomization
>> plugin will keep things in the right order.)
> Regardless of your answer, this randomization can be turned off.
> Switching to designated initializers here is mainly just a clean up.

Any thoughts on this? I'd love to get this landed, and I'm happy to
carry the patch myself, if it helps. Anyone able to Ack?



Kees Cook
Pixel Security