From: mengqinggang <mengqinggang@loongson.cn>
To: Xi Ruoyao <xry111@xry111.site>, binutils@sourceware.org
Cc: liuzhensong@loongson.cn, i.swmail@xen0n.name, maskray@google.com,
hejinyang@loongson.cn, cailulu@loongson.cn,
chenglulu@loongson.cn, xuchenghua@loongson.cn,
luweining@loongson.cn
Subject: Re: [PATCH] LoongArch: Adapt R_LARCH_{PCALA,GOT,TLS_IE,TLS_DESC}64_* handling per psABI v2.30
Date: Tue, 16 Jan 2024 15:02:18 +0800 [thread overview]
Message-ID: <1076e362-dff9-4730-f3d1-3b79a9cc1c6c@loongson.cn> (raw)
In-Reply-To: <a84ee942debf7637876462360ff0d7a48fbf3a32.camel@xry111.site>
在 2024/1/16 下午2:48, Xi Ruoyao 写道:
> On Tue, 2024-01-16 at 14:38 +0800, mengqinggang wrote:
>> 在 2024/1/15 下午8:44, Xi Ruoyao 写道:
>>> In LoongArch psABI v2.30, an offset (-8 for LO20 and -12 for HI12)
>>> should be applied on PC for these reloc types to avoid wrong relocation
>>> when the instruction sequence crosses a page boundary.
>>>
>>> The lld linker has already adapted the change. Make it for the bfd
>>> linker too.
>>>
>>> Link: https://github.com/loongson/la-abi-specs/releases/v2.30
>>> Link: https://github.com/loongson-community/discussions/issues/17
>>> Link: https://github.com/llvm/llvm-project/pull/73387
>>> Signed-off-by: Xi Ruoyao <xry111@xry111.site>
>>> ---
>>> bfd/elfnn-loongarch.c | 29 +++++++++++--------
>>> .../ld-loongarch-elf/ld-loongarch-elf.exp | 1 +
>>> ld/testsuite/ld-loongarch-elf/pcala64.d | 15 ++++++++++
>>> ld/testsuite/ld-loongarch-elf/pcala64.s | 8 +++++
>>> 4 files changed, 41 insertions(+), 12 deletions(-)
>>> create mode 100644 ld/testsuite/ld-loongarch-elf/pcala64.d
>>> create mode 100644 ld/testsuite/ld-loongarch-elf/pcala64.s
>>>
>>> diff --git a/bfd/elfnn-loongarch.c b/bfd/elfnn-loongarch.c
>>> index 3858a3179fd..e4f4da7a0dc 100644
>>> --- a/bfd/elfnn-loongarch.c
>>> +++ b/bfd/elfnn-loongarch.c
>>> @@ -3544,14 +3544,16 @@ loongarch_elf_relocate_section (bfd *output_bfd, struct bfd_link_info *info,
>>> }
>>> break;
>>>
>>> - case R_LARCH_PCALA64_LO20:
>>> case R_LARCH_PCALA64_HI12:
>>> + pc -= 4;
>>> + /* Fall through. */
>>> + case R_LARCH_PCALA64_LO20:
>>> if (h && h->plt.offset != MINUS_ONE)
>>> relocation = sec_addr (plt) + h->plt.offset;
>>> else
>>> relocation += rel->r_addend;
>>>
>>> - RELOCATE_CALC_PC64_HI32 (relocation, pc);
>>> + RELOCATE_CALC_PC64_HI32 (relocation, pc - 8);
>>
>> The 'pc - 8' triggered an error because the pc parameter is not enclosed
>> in parentheses in macro definition.
>>
>> /home/mengqinggang/toolchains/src/binutils-gdb/bfd/elfnn-loongarch.c:3546:51:
>> error: suggest parentheses around '-' in operand of '&'
>> [-Werror=parentheses]
>> 3546 | RELOCATE_CALC_PC64_HI32 (relocation, pc - 8);
>> | ~~~^~~
>> /home/mengqinggang/toolchains/src/binutils-gdb/bfd/elfnn-loongarch.c:2532:22:
>> note: in definition of macro 'RELOCATE_CALC_PC64_HI32'
>> 2532 | - (pc & ~(bfd_vma)0xfff); \
>> | ^~
> Interesting, why didn't I see this error... And in this case it happens
> to work because - has a higher precedence. But I'll add parentheses in
> V2 anyway.
>
> /* snip */
Maybe you configured with --disable-werror?
>>> diff --git a/ld/testsuite/ld-loongarch-elf/pcala64.d b/ld/testsuite/ld-loongarch-elf/pcala64.d
>>> new file mode 100644
>>> index 00000000000..6b1411577db
>>> --- /dev/null
>>> +++ b/ld/testsuite/ld-loongarch-elf/pcala64.d
>>> @@ -0,0 +1,15 @@
>>> +#ld: -Ttext=0x180000ff8 -Tdata=0x1000000000 -shared
>>
>> The test need to be disabled on loongarch*-elf target because -shared
>> option is not supported on loongarch*-elf target.
>>
>> It can be disabled by adding a line "#skip: loongarch*-elf" after
>> "#objdump -d".
> I'll just remove the -shared option. I changed the function name to
> "_start" so it should work without -shared, but I forgot to actually
> remove -shared.
>
>
next prev parent reply other threads:[~2024-01-16 7:02 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-15 12:44 Xi Ruoyao
2024-01-16 6:38 ` mengqinggang
2024-01-16 6:48 ` Xi Ruoyao
2024-01-16 7:02 ` mengqinggang [this message]
2024-01-16 7:07 ` Xi Ruoyao
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=1076e362-dff9-4730-f3d1-3b79a9cc1c6c@loongson.cn \
--to=mengqinggang@loongson.cn \
--cc=binutils@sourceware.org \
--cc=cailulu@loongson.cn \
--cc=chenglulu@loongson.cn \
--cc=hejinyang@loongson.cn \
--cc=i.swmail@xen0n.name \
--cc=liuzhensong@loongson.cn \
--cc=luweining@loongson.cn \
--cc=maskray@google.com \
--cc=xry111@xry111.site \
--cc=xuchenghua@loongson.cn \
/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).