This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR41529
On Sat, 2 Jan 2010, Dave Korn wrote:
> Richard Guenther wrote:
>
> > Yes - well, sort-of. The ELFish stuff only starts at LTO phase
> > where we use libelf to read those ELF sections and in WPA phase
> > where we use libelf to write intermediate code for the ltrans
> > stage (that might even work on non-ELF targets). The issue
> > is probably that reading ELF sections from inside native object
> > formats using libelf doesn't work. Does it?
>
> Without having checked the code yet, I'll bet you've nailed it there. The
> test you pointed me at is indeed failing for just the reason it should fail:
> no cross-TU inlining takes place.
>
> I'll see if there's a nice way to extend the capabilities of libelf.
In fact the initial object code won't contain any ELF sections. The
LTO sections are simply passed through the assembler by cc1 like
.section .gnu_lto_XXX
.ascii '....
Which of course libelf can't parse either.
Richard.