netdev - Re: Issue accessing task_struct from BPF due to 4.16 stack-protector changes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 2 Mar 2018 14:24:03 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Gianluca Borello <g.borello@...il.com>
Cc:     Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Linux Networking Development Mailing List 
        <netdev@...r.kernel.org>, Daniel Borkmann <daniel@...earbox.net>
Subject: Re: Issue accessing task_struct from BPF due to 4.16 stack-protector changes

On Fri, Mar 2, 2018 at 2:04 PM, Gianluca Borello <g.borello@...il.com> wrote:
> On Fri, Mar 2, 2018 at 12:42 PM, Alexei Starovoitov
> <alexei.starovoitov@...il.com> wrote:
>>
>> good catch!
>> I wonder why sched.h is using this flag insead of relying on #defines from autoconf.h
>> It could have been using CONFIG_HAVE_CC_STACKPROTECTOR
>> instead of CONFIG_CC_STACKPROTECTOR, no ?
>>
>
> Thanks for your reply Alexei. I think switching to
> HAVE_CC_STACKPROTECTOR could indeed solve this particular BPF issue in
> a cleaner way (I tested it), at the cost of having that struct member
> always present for the supported architectures even if the stack
> protector is actually disabled (e.g. CONFIG_CC_STACKPROTECTOR_NONE=y).
>
> Not sure if this could be frowned upon by someone considering how
> critical task_struct is, but on the other hand is really just 8 bytes.

That structure is huge, and I think it's proper to leave this as is.

Adding KBUILD_CPPFLAGS (for now) seems like the right way to go;
though in the future stack protector will be changed around again (to
be purely Kconfig again). There are a number of issues with its logic
in detecting and enabling, and another draft at solving it is under
development.

-Kees

-- 
Kees Cook
Pixel Security