Re: [PATCH 4.9 46/92] x86/alternative: Print unadorned pointers

[Kees Cook]

On Fri, Feb 9, 2018 at 11:23 PM, Greg Kroah-Hartman
<gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> On Fri, Feb 09, 2018 at 02:01:32PM -0800, Kees Cook wrote:
>> On Fri, Feb 9, 2018 at 5:39 AM, Greg Kroah-Hartman
>> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>> > 4.9-stable review patch.  If anyone has any objections, please let me know.
>> >
>> > ------------------
>> >
>> > From: Borislav Petkov <bp@xxxxxxx>
>> >
>> > (cherry picked from commit 0e6c16c652cadaffd25a6bb326ec10da5bcec6b4)
>> >
>> > After commit ad67b74d2469 ("printk: hash addresses printed with %p")
>> > pointers are being hashed when printed. However, this makes the alternative
>> > debug output completely useless. Switch to %px in order to see the
>> > unadorned kernel pointers.
>>
>> This missed a "Fixes:" tag so probably missed automated checking on
>> how far back to port this. It shouldn't go back beyond 4.15 (where
>> ad67b74d2469 first appeared).
>
> Good point.  Should we instead be using %pK for this change instead?  Or
> should we just backport commit ad67b74d2469 to 4.14?  :)

ad67b74d2469 is pretty disruptive, I can't recommend putting it in
4.14. I mean, I think it'd be great, but I suspect other people would
find it very surprising. :)

I would just revert this commit from kernels earlier than 4.15; this
%p usage is in debugging, so %pK doesn't make much sense here.

-Kees

-- 
Kees Cook
Pixel Security