This is the mail archive of the
mailing list for the binutils project.
Re: [RFC][PATCH][ld][testsuite] Add -ffat-lto-objects option to some ld tests
- From: Kyrill Tkachov <kyrylo dot tkachov at arm dot com>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: "binutils at sourceware dot org" <binutils at sourceware dot org>
- Date: Tue, 26 Nov 2013 14:55:58 +0000
- Subject: Re: [RFC][PATCH][ld][testsuite] Add -ffat-lto-objects option to some ld tests
- Authentication-results: sourceware.org; auth=none
- References: <5294B0F2 dot 8000006 at arm dot com> <CAMe9rOoi6PS6SWOJBsJHw=hFFTBeWohbsb_U+iBzTbRF=ft5Fw at mail dot gmail dot com>
On 26/11/13 14:36, H.J. Lu wrote:
On Tue, Nov 26, 2013 at 6:32 AM, Kyrill Tkachov <email@example.com> wrote:
I'm seeing some ld testsuite failures on arm-none-linux-gnueabi and
aarch64-none-linux-gnu after a recent gcc commit (r205065) that set
-fno-fat-lto-objects as the default:
PASS->FAIL: LTO 3 symbol
PASS->FAIL: PR ld/12758
PASS->FAIL: PR ld/12760
PASS->FAIL: PR ld/13183
It seems that these tests need -ffat-lto-objects, at least on arm and
aarch64, but I'm not sure I understand the structure of these ld tests well.
Could someone please comment on whether this is the correct approach?
CC'ing H.J. since he wrote the lto.exp file.
Does -ffat-lto-objects always work when -flto is supported?
If yes, OK for trunk.
The gcc manual says about -ffat-lto-objects: " This option is effective only
when compiling with -flto and is ignored at link time".
I'd think that it's supposed to always work with -flto.