This is the mail archive of the
mailing list for the binutils project.
Re: Sort relocs output by ld -r
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: Andreas Krebbel <krebbel at linux dot vnet dot ibm dot com>, Binutils <binutils at sourceware dot org>
- Date: Wed, 18 Mar 2015 06:34:10 -0700
- Subject: Re: Sort relocs output by ld -r
- Authentication-results: sourceware.org; auth=none
- References: <20141204072210 dot GW20383 at bubble dot grove dot modra dot org> <20150313195125 dot GA29455 at maggie> <20150314040500 dot GC16488 at bubble dot grove dot modra dot org> <20150318132041 dot GF24573 at bubble dot grove dot modra dot org>
On Wed, Mar 18, 2015 at 6:20 AM, Alan Modra <email@example.com> wrote:
> On Sat, Mar 14, 2015 at 02:35:00PM +1030, Alan Modra wrote:
>> On Fri, Mar 13, 2015 at 08:51:25PM +0100, Andreas Krebbel wrote:
>> > The attached patch mimics what you did for Mips and disables the
>> > sorting for relocations against code section. Does that look like a
>> > reasonable fix for this case as well?
>> Looks fine to me.
> This should also go on the 2.25 branch, since I backported the reloc
> sorting patch. I'm applying the following cherry-picks from master.
> e3ee8ed4 Fix gold error: hidden symbol '...' is not defined locally
> ec86f434 PowerPC64 GOLD: complain on misaligned _DS relocs
> a127494f Use dynamic text relocs for protected vars
I believe you should fix linker and glibc properly instead.
But it is your call.
> bd4d2eaa PowerPC64 thread-safe stubs not needed for iplt
> f01f1741 Tweak ppc32 tls_get_addr optimisation
> 1079403c Fix ppc32 synthetic symbols when __tls_get_addr_opt stub is generated
> b9005ba7 S/390: Disable relocation sort against code sections