From: liuzhensong <liuzhensong@loongson.cn>
To: Fangrui Song <maskray@google.com>
Cc: binutils@sourceware.org, i.swmail@xen0n.name, xry111@xry111.site,
chenhuacai@loongson.cn
Subject: Re: [PATCH] LoongArch:ld: Fix relocation error of pcrel.
Date: Sat, 3 Sep 2022 15:12:50 +0800 [thread overview]
Message-ID: <a2119648-5181-4896-6a3c-b68b62f67bb0@loongson.cn> (raw)
In-Reply-To: <CAFP8O3Jkvt==05z7=gPqLZw1UL6TcQGpcSn9dHOPK7cybnnLwQ@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 5761 bytes --]
在 2022/9/3 下午12:33, Fangrui Song 写道:
> On Fri, Sep 2, 2022 at 6:29 PM liuzhensong <liuzhensong@loongson.cn> wrote:
>> Patch for branch 2.39.
>> Need to reduce the address of pc when using
>> reloction R_LARCH_SOP_PUSH_PLT_PCREL.
>>
>> bfd/
>> * elfnn-loongarch.c
>> ---
>> bfd/elfnn-loongarch.c | 3 +-
>> ld/testsuite/ld-loongarch-elf/pcrel-const.d | 14 +++++++
>> ld/testsuite/ld-loongarch-elf/pcrel-const.lds | 14 +++++++
>> ld/testsuite/ld-loongarch-elf/pcrel-const.s | 11 ++++++
>> ld/testsuite/ld-loongarch-elf/pr.exp | 39 +++++++++++++++++++
>> 5 files changed, 80 insertions(+), 1 deletion(-)
>> create mode 100644 ld/testsuite/ld-loongarch-elf/pcrel-const.d
>> create mode 100644 ld/testsuite/ld-loongarch-elf/pcrel-const.lds
>> create mode 100644 ld/testsuite/ld-loongarch-elf/pcrel-const.s
>> create mode 100644 ld/testsuite/ld-loongarch-elf/pr.exp
>>
>> diff --git a/bfd/elfnn-loongarch.c b/bfd/elfnn-loongarch.c
>> index 5b44901b9e0..fafdc7c7458 100644
>> --- a/bfd/elfnn-loongarch.c
>> +++ b/bfd/elfnn-loongarch.c
>> @@ -2341,9 +2341,10 @@ loongarch_elf_relocate_section (bfd *output_bfd, struct bfd_link_info *info,
>> case R_LARCH_SOP_PUSH_PLT_PCREL:
>> unresolved_reloc = false;
>>
>> - if (resolved_to_const)
>> + if (!is_undefweak && resolved_to_const)
>> {
>> relocation += rel->r_addend;
>> + relocation -= pc;
>> break;
>> }
>> else if (is_undefweak)
>> diff --git a/ld/testsuite/ld-loongarch-elf/pcrel-const.d b/ld/testsuite/ld-loongarch-elf/pcrel-const.d
>> new file mode 100644
>> index 00000000000..72abb204122
>> --- /dev/null
>> +++ b/ld/testsuite/ld-loongarch-elf/pcrel-const.d
>> @@ -0,0 +1,14 @@
>> +#as: -mla-global-with-pcrel
>> +#objdump: -Drsz
>> +
>> +.*:[ ]+file format .*
>> +
>> +
>> +Disassembly of section .text:
>> +
>> +.* <foo>:
>> +#...
>> +[ ]+4:[ ]+02c03084[ ]+addi.d[ ]+\$a0,[ ]+\$a0,[ ]+12\(0xc\)
>> +#...
>> +0+10 <__sec_end>:
>> +#pass
>> diff --git a/ld/testsuite/ld-loongarch-elf/pcrel-const.lds b/ld/testsuite/ld-loongarch-elf/pcrel-const.lds
>> new file mode 100644
>> index 00000000000..064a6894b4d
>> --- /dev/null
>> +++ b/ld/testsuite/ld-loongarch-elf/pcrel-const.lds
>> @@ -0,0 +1,14 @@
>> +ENTRY(foo);
>> +SECTIONS
>> +{
>> + .text : ALIGN(4) {
>> + *(.text*)
>> + }
>> +
>> + .data : ALIGN(4) {
>> + __sec_start = .;
>> + *(.gzdata)
>> + __sec_end = .;
>> + }
>> +}
>> +PROVIDE(__sec_size = __sec_end - 4);
> I do not know the legacy relocation scheme, but for tests I would try
> removing noise like ALIGN(4). .data output section containing .gzdata
> input sections looks strange too.
> Can .data input sections be used instead?
The __sec_end segment cannot be dumped using .data.
t: file format elf64-loongarch
Disassembly of section .text:
0000000000000000 <foo>:
0: 03400000 andi $zero, $zero, 0x0
4: 1c000004 pcaddu12i $a0, 0
8: 02c06084 addi.d $a0, $a0, 24(0x18)
c: 24000084 ldptr.w $a0, $a0, 0
10: 4c000020 jirl $zero, $ra, 0
Disassembly of section .data:
0000000000000014 <__sec_start>:
14: 00000001 0x00000001
>
> `__sec_size = __sec_end - 4` is unusual. Is `- 4` necessary to
> reproduce the bug? If necessary, add a comment.
Unnecessary, using "nop" before pcaddu12i instead.
>
>> diff --git a/ld/testsuite/ld-loongarch-elf/pcrel-const.s b/ld/testsuite/ld-loongarch-elf/pcrel-const.s
>> new file mode 100644
>> index 00000000000..bc4fb979a80
>> --- /dev/null
>> +++ b/ld/testsuite/ld-loongarch-elf/pcrel-const.s
>> @@ -0,0 +1,11 @@
>> + .text
>> + .align 2
>> + .globl foo
>> + .type foo, @function
>> +foo:
>> + la.global $r4,__sec_size
>> + ldptr.w $r4,$r4,0
>> + jr $r1
>> + .size foo, .-foo
>> + .data
>> + .word 1
>> diff --git a/ld/testsuite/ld-loongarch-elf/pr.exp b/ld/testsuite/ld-loongarch-elf/pr.exp
>> new file mode 100644
>> index 00000000000..c0cc2fc30d2
>> --- /dev/null
>> +++ b/ld/testsuite/ld-loongarch-elf/pr.exp
>> @@ -0,0 +1,39 @@
>> +# Expect script for LoongArch ELF linker tests
>> +# Copyright (C) 2022 Free Software Foundation, Inc.
>> +#
>> +# This file is part of the GNU Binutils.
>> +#
>> +# This program is free software; you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation; either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program; if not, write to the Free Software
>> +# Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
>> +# MA 02110-1301, USA.
>> +#
>> +
>> +if ![istarget loongarch64-*-*] {
>> + return
>> +}
>> +
>> +set link_tests [list \
>> + [list \
>> + "pcrel const" \
>> + "-T pcrel-const.lds" "" \
>> + "-mla-global-with-pcrel" \
>> + { pcrel-const.s } \
>> + [list \
>> + [list objdump -D pcrel-const.d] \
>> + ] \
>> + "pcrel-const" \
>> + ] \
>> +]
>> +
>> +run_ld_link_tests $link_tests
>> --
>> 2.31.1
>>
prev parent reply other threads:[~2022-09-03 7:12 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-03 1:28 liuzhensong
2022-09-03 4:33 ` Fangrui Song
2022-09-03 7:12 ` liuzhensong [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a2119648-5181-4896-6a3c-b68b62f67bb0@loongson.cn \
--to=liuzhensong@loongson.cn \
--cc=binutils@sourceware.org \
--cc=chenhuacai@loongson.cn \
--cc=i.swmail@xen0n.name \
--cc=maskray@google.com \
--cc=xry111@xry111.site \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).