Time to enable partial relro

[Warner Losh]

> On Aug 26, 2016, at 9:14 AM, Pedro Giffuni <pfg at FreeBSD.org> wrote:
> 
> Hello;
> 
> On 08/26/16 10:06, Warner Losh wrote:
>> On Fri, Aug 26, 2016 at 9:00 AM, Pedro Giffuni <pfg at freebsd.org> wrote:
>>> 
>>> 
>>> On 08/26/16 05:56, Konstantin Belousov wrote:
>>>> 
>>>> On Thu, Aug 25, 2016 at 05:50:31PM -0500, Pedro Giffuni wrote:
>>>>> 
>>>>> Hello;
>>>>> 
>>>>> GNU RELRO support was committed in r230784 (2012-01-30) but we never
>>>>> enabled it by default.
>>>>> 
>>>>> There was some discussion about it on
>>>>> https://reviews.freebsd.org/D3001
>>>>> 
>>>>> By now, all Linux distributions, NetBSD and DragonFly support it and
>>>>> it is the default for most systems in binutils 2.27.
>>>>> 
>>>>> This doesn't affect performance, I ran it through an exp-run last
>>>>> year, no other OS has had issues etc ... seems safe and can be
>>>>> disabled if needed when linking.
>>>> 
>>>> Exp-run does not test anything interesting about relro. If all testing
>>>> that was done is basically just an exp-run, then there was no useful
>>>> runtime testing done.
>>>> 
>>> 
>>> The exp-run does cover Java and other VM-type thingies that bootstrap.
>>> For upstream binutils this is now the default (at least for linux,
>>> they never ask us if we want to follow). So the change has been tested
>>> extensively but perhaps not on cases that are relevant to us.
>>> 
>>> Note that the "fix" for any port is ultimately trivial:
>>> LDFLAGS+= "-z norelro"
>>> 
>>>>> 
>>>>> I think it's time to enable it be default in our base binutils. If
>>>>> there are no objections, I will just commit the attached patch over
>>>>> the weekend.
>>>> 
>>>> 
>>>> There are objections, the change must be runtime tested on large and
>>>> representative set of real-world applications before turning the knob.
>>>> 
>>> 
>>> You are not giving any hint on what would be a "representative set of
>>> real-world applications". Given that you committed the initial support your
>>> objection stands very high and is a blocker. :(
>>> 
>>> As I see it committing it now would give ample time to test this in current
>>> before it hits any release. If you want more extensive testing merging it in
>>> -stable right after the 11-Release is guaranteed to help
>>> weed out any remaining update ports may need.
>> 
>> I'd say a minimum is 'buildworld' + a test boot on at least Intel (i386 and
>> amd64), armv6 and mips (both 32-bit and 64-bit) before we proceed. How
>> many of those have we done?
>> 
> 
> I have been running it my desktop (amd64) for a year now. I can test i386 in a VM but I doubt it will affect anything. The issue, and it's probably kib's worry are some rarely used but important ports. Stuff like erlang, or virtualbox maybe, but as I wrote, the fix (if needed)
> is trivial by adding a flag to the link command.
> 
> FWIW, but it is largely irrelevant to us, RELRO is the default on
> OpenBSD and there is no way out of it there.

What I’m worried about is that our run time linker may get the RELRO stuff wrong and that’s a very platform specific thing and needs to be validated. I also share Kib’s worry about different ports being broken, but that’s a different issue. Recent compilers have broken our run time linker on mips, for example, because they generate new / different relocations than those before. It’s easy enough to test to make sure that we’re good on at least the fairly active platforms (i386, amd64, armv6, mips and mips64) to make sure that nothing bad happens. The others can be tested in due time (though the powerpc ones likely can be tested easily enough by the powerpc guys since they are quite active).

I get very nervous when I see “should work” or “should be platform independent” for such a low-level thing.

Warner
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.freebsd.org/pipermail/freebsd-toolchain/attachments/20160826/9825ee7a/attachment.sig>