Re: PowerPC64 DT_RELR

[Alan Modra <amodra@gmail.com>]

On Fri, Jun 10, 2022 at 08:27:25AM +0930, Alan Modra wrote:
> On Thu, Jun 09, 2022 at 12:57:32PM +0200, Florian Weimer wrote:
> > * Alan Modra via Binutils:
> > 
> > > PowerPC64 takes a more traditional approach to DT_RELR than x86.  Count
> > > relative relocs in check_relocs, allocate space for them and output in
> > > the usual places but not doing so when enable_dt_relr.  DT_RELR is
> > > sized in the existing ppc stub relaxation machinery, run via the
> > > linker's ldemul_after_allocation hook.  DT_RELR is output in the same
> > > function that writes ppc stubs, run via ldemul_finish.
> > >
> > > This support should be considered experimental.
> > 
> > Upstream glibc enables DT_RELR for its ELF files by default (including
> > tests), and we seem to hit something DT_RELR-related:
> > 
> >   Some glibc tests crash in DT_RELR relocation on powerpc64le
> >   <https://sourceware.org/bugzilla/show_bug.cgi?id=29235>
> > 
> > (I'm working on attaching the test binary to the bug.)
> 
> I'll investigate.  My guess would be .eh_frame editing is the trigger.
> 
> > Maybe we should not automatically DT_RELR for glibc itself if BFD ld
> > supports it?
> 
> Yes, at least until I can fix this.

binutils-2.38-14.fc37 is missing binutils commit e4a35c731962.

-- 
Alan Modra
Australia Development Lab, IBM