public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Fangrui Song <maskray@google.com>
To: liuzhensong <liuzhensong@loongson.cn>
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: Fri, 2 Sep 2022 21:33:33 -0700	[thread overview]
Message-ID: <CAFP8O3Jkvt==05z7=gPqLZw1UL6TcQGpcSn9dHOPK7cybnnLwQ@mail.gmail.com> (raw)
In-Reply-To: <20220903012848.968446-1-liuzhensong@loongson.cn>

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?

`__sec_size = __sec_end - 4` is unusual. Is `- 4` necessary to
reproduce the bug? If necessary, add a comment.

> 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
>

  reply	other threads:[~2022-09-03  4:33 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 [this message]
2022-09-03  7:12   ` liuzhensong

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='CAFP8O3Jkvt==05z7=gPqLZw1UL6TcQGpcSn9dHOPK7cybnnLwQ@mail.gmail.com' \
    --to=maskray@google.com \
    --cc=binutils@sourceware.org \
    --cc=chenhuacai@loongson.cn \
    --cc=i.swmail@xen0n.name \
    --cc=liuzhensong@loongson.cn \
    --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).