[PATCH] D28272: ELF: Reserve space for copy relocations of read-only symbols in relro.

[Peter Collingbourne via llvm-commits]

On Thu, Jan 5, 2017 at 1:49 PM, Rafael Avila de Espindola <
rafael.espindola at gmail.com> wrote:

> Peter Collingbourne <peter at pcc.me.uk> writes:
>
> > On Thu, Jan 5, 2017 at 1:17 PM, Rafael Avila de Espindola <
> > rafael.espindola at gmail.com> wrote:
> >
> >> Peter Collingbourne via Phabricator <reviews at reviews.llvm.org> writes:
> >>
> >> > Relatedly, I discovered a bug in this patch. We can't use SHT_NOBITS
> for
> >> the new section because we need to lay out all NOBITS sections
> contiguously
> >> at the end of the LOAD. RELRO also needs to be contiguous (there can
> only
> >> be one RELRO header), so we can't have another RELRO covering part of
> the
> >> bss.
> >>
> >> You could move the RELRO to the end of the regular RW PT_LOAD, no? That
> >> way it
> >> will be just before the nobits sections and the nobits sections could
> >> start with RELRO ones. That way the RELRO segment would span the end of
> >> the regular (non-nobits) sections and the start of the nobits ones.
> >>
> >
> > Yes, I think that would work, thanks. So the layout would be:
> >
> > non-RELRO non-NOBITS
> > [page alignment]
> > RELRO non-NOBITS
> > RELRO NOBITS
> > [page alignment]
> > non-RELRO NOBITS
> >
> > The downside is that there would be an extra page alignment, but that
> > doesn't seem too important.
>
> We already have two alignments. What we currently have is:
>
>    [page alignment]
>    RELRO (always non-NOBITS)
>    [page alignment]
>    non-RELRO non-NOBITS
>    non-RELRO NOBITS
>

Yes, but we'd still need a page alignment at the start of the r/w segment,
no?

Thanks,
-- 
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170105/e10f3a35/attachment.html>