This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [GOLD] Disable script_test_12 for powerpc
- From: Alan Modra <amodra at gmail dot com>
- To: Cary Coutant <ccoutant at gmail dot com>
- Cc: Binutils <binutils at sourceware dot org>, Ian Lance Taylor <iant at google dot com>
- Date: Wed, 29 Jun 2016 09:03:36 +0930
- Subject: Re: [GOLD] Disable script_test_12 for powerpc
- Authentication-results: sourceware.org; auth=none
- References: <20160628060834 dot GE3665 at bubble dot grove dot modra dot org> <20160628061209 dot GF3665 at bubble dot grove dot modra dot org> <CAJimCsEy_GfLepjRUyBLHmFRv=pfruF3Layvuf0Boo9mhXubLA at mail dot gmail dot com>
On Tue, Jun 28, 2016 at 12:44:26PM -0700, Cary Coutant wrote:
> > Changes necessary to make this test pass on powerpc64. For interest
> > really. I'm hesitant to apply this change so close to a release in
> > case it blows up on ARM or some other target. Passes on x86_64.
> >
> > * script_test_12.t: Delete .plt, specify 64k page size.
> > * script_test_12i.t: Likewise.
> >
> > diff --git a/gold/testsuite/script_test_12.t b/gold/testsuite/script_test_12.t
> > index 4579c58..d415d04 100644
> > --- a/gold/testsuite/script_test_12.t
> > +++ b/gold/testsuite/script_test_12.t
> > @@ -33,14 +33,13 @@ SECTIONS
> > .rel.plt : { *(.rel.plt) }
> > .rela.plt : { *(.rela.plt) }
> > .init : { *(.init) }
> > - .plt : { *(.plt) }
>
> To be honest, I don't know why this is there. I've verified that it
> works fine on x86_64 without it, and I'm certain that I didn't add it
> specifically to make it work on any other target. I think it's safe to
> remove.
The .plt removal was the only change I was a little concerned about.
Since you think it is OK I've committed this version of the fix.
> > .text : { *(.text) }
> > .fini : { *(.fini) }
> > .rodata : { *(.rodata .rodata.*) }
> > .eh_frame_hdr : { *(.eh_frame_hdr) }
> > .eh_frame : { *(.eh_frame) }
> >
> > - . = DATA_SEGMENT_ALIGN(0x1000, 0x1000);
> > + . = DATA_SEGMENT_ALIGN(0x10000, 0x10000);
>
> This also seems safe.
>
> If you're still even the least bit nervous, I'd be OK with simply
> disabling the test on ppc for 2.27, then applying this patch after the
> branch has been made.
>
> -cary
--
Alan Modra
Australia Development Lab, IBM