Re: [PATCH] proc: prevent a task from writing on its own /proc/*/mem

From: Salvatore Mesoraca
Date: Mon May 28 2018 - 05:32:37 EST


2018-05-27 3:33 GMT+02:00 Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>:
> On Sat, May 26, 2018 at 5:32 PM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>
>> I went through some old threads from 2012 when e268337dfe26 was
>> introduced, and later when things got looked at during DirtyCOW. There
>> was discussion about removing FOLL_FORCE (in order to block writes on
>> a read-only memory region).
>
> Side note, we did that for /dev/mem, and things broke.
>
> Thus commit f511c0b17b08 "Yes, people use FOLL_FORCE ;)"
>
> Side note, that very sam ecommit f511c0b17b08 is also the explanation for
> why the patch under discussion now seems broken.
>
> People really do use "write to /proc/self/mem" as a way to keep the
> mappings read-only, but have a way to change them when required.

Oh, I didn't expect this, interesting...
A configurable LSM is probably the right way to do this.

Thank you for your time,

Salvatore