This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [Patch, AArch64] Fix ld ifunc test failures in aarch64-*-elf
- From: Will Newton <will dot newton at linaro dot org>
- To: Yufeng Zhang <Yufeng dot Zhang at arm dot com>
- Cc: "binutils at sourceware dot org" <binutils at sourceware dot org>, Marcus Shawcroft <Marcus dot Shawcroft at arm dot com>
- Date: Fri, 14 Jun 2013 15:32:21 +0100
- Subject: Re: [Patch, AArch64] Fix ld ifunc test failures in aarch64-*-elf
- References: <51BB1972 dot 6060609 at arm dot com> <CANu=DmjCun5CHSp9qaehBz72n7SPVGhJri-pfzsoNJ=b5eBt-Q at mail dot gmail dot com> <51BB2725 dot 6090103 at arm dot com>
On 14 June 2013 15:22, Yufeng Zhang <Yufeng.Zhang@arm.com> wrote:
> On 06/14/13 15:05, Will Newton wrote:
>>
>> On 14 June 2013 14:24, Yufeng Zhang<Yufeng.Zhang@arm.com> wrote:
>>>
>>> The patch fixes the failures of the ld ifunc tests in aarch64-*-elf.
>>>
>>> The failures are due to two reasons:
>>>
>>> 1. the expected dump contains hard-coded immediate offsets that seem to
>>> be
>>> from the linux-gnu environment.
>>> 2. the change to emulparams/aarch64elf.sh of adding IREL_IN_PLT seems to
>>> be
>>> forgotten to commit; this causes the mismatch of relocation section name
>>> (.rela.plt versus .rela.dyn). This missing change was mentioned in the
>>> changelog entry for the original ifunc patch.
>>
>>
>> Oh dear, this suggests I applied an out of date patch. My apologies. :-(
>>
>> I don't understand the ifunc-2-local.s change however.
>
>
> I noticed that ifunc-2-local.s and ifunc-2.s were the same so I updated the
> former according to the difference between their ifunc1 peers. Correct me if
> it was intended.
That's fine.
In my original fix for the the differing offsets between linux and ELF
(that seems to have vanished from the patch I actually checked in, hey
ho) I used (0x2c0|0x3a0) style regex rather than 0x[0-9a-f]+ which
keeps things slightly tighter but YMMV.
--
Will Newton
Toolchain Working Group, Linaro