public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Fangrui Song <i@maskray.me>
To: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
Cc: Nelson Chu <nelson@rivosinc.com>,
	binutils@sourceware.org, rui314@gmail.com,
	 ruiu@bluewhale.systems
Subject: Re: [PATCH 0/4] RISC-V: Implement TLS Descriptors.
Date: Fri, 18 Aug 2023 00:13:24 -0700	[thread overview]
Message-ID: <DS7PR12MB576522B309E27EBA26708A85CB1BA@DS7PR12MB5765.namprd12.prod.outlook.com> (raw)
In-Reply-To: <CAPpQWtCKU2pOd++NfRPPmTyj9zUsu=B7A9eMFtmB5HG643dRgw@mail.gmail.com>

On Thu, Aug 17, 2023 at 5:22 PM Nelson Chu <nelson@rivosinc.com> wrote:
>
> Hi,
>
> Before the reviewing, I think you still should have test cases for each
> patch to make sure everything is correct in the binutils, and helps people
> understand what the patches try to resolve what problem or support what
> features.  In general, the regressions in binutils, gcc and glibc are
> seperate.  Besides, each patch should at least have the ChangLogs in the
> commit message.
>
> Thanks
> Nelson

A relocatable object file can contain both TLSDESC and TLS GD
relocations (e.g. relocatable output).
The GOT use cases are different. It will be nice to have such a test.

We need tests for at least (a) -shared and (b) -no-pie or -pie (I
prefer just -no-pie).

Perhaps you have done this (as the glibc patch suggests), but I want
to make sure that the GOT entries are placed in .rela.dyn, not
.rela.plt
(https://sourceware.org/bugzilla/show_bug.cgi?id=28387  ld: Move
R_*_TLSDESC to .rela.dyn)
sysdeps/{aarch64,x86_64}/dl-machine.h unfortunately have to handle
TLSDESC in elf_machine_lazy_rel for compatibility.

> [PATCH 4/4] RISC-V: Initial ld.bfd support for TLSDESC.
> + /* TLSDESC needs one dynamic reloc and four GOT slots. *

Why is there 4? Other ports reserve 2 GOT slots for one symbol.

> On Fri, Aug 18, 2023 at 2:10 AM Tatsuyuki Ishi via Binutils <
> binutils@sourceware.org> wrote:
>
> > This patchset implements TLS Descriptors (TLSDESC) for RISC-V targets, per
> > the specification draft at [1].
> >
> > This patchset is based on top of [2].
> >
> > No regression in binutils and gcc tests for rv64gc, tested alongside the
> > gcc and glibc implementation (will be posted shortly).
> >
> > This contribution is made on behalf of Blue Whale Systems, which has
> > copyright assignment on file with the FSF.
> >
> > [1]: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/373
> > [2]: https://sourceware.org/pipermail/binutils/2023-August/129075.html
> >
> > Tatsuyuki Ishi (4):
> >   RISC-V: Add TLSDESC reloc definitions.
> >   RISC-V: Add assembly support for TLSDESC.
> >   RISC-V: Define and use GOT entry size constants for TLS.
> >   RISC-V: Initial ld.bfd support for TLSDESC.
> >
> >  bfd/bfd-in2.h                              |   4 +
> >  bfd/elfnn-riscv.c                          | 105 ++++++++++++++++++---
> >  bfd/elfxx-riscv.c                          |  75 ++++++++++++++-
> >  bfd/libbfd.h                               |   4 +
> >  bfd/reloc.c                                |   8 ++
> >  gas/config/tc-riscv.c                      |  18 +++-
> >  include/elf/riscv.h                        |   5 +
> >
>
>
> >  ld/testsuite/ld-riscv-elf/ld-riscv-elf.exp |   2 +
> >
>
> This doesn't work, seems still not adding any test cases.
>
>
> >  opcodes/riscv-opc.c                        |   1 +
> >  9 files changed, 201 insertions(+), 21 deletions(-)
> >
> > --
> > 2.41.0
> >
> >

  reply	other threads:[~2023-08-18  7:20 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-17 18:08 Tatsuyuki Ishi
2023-08-17 18:08 ` [PATCH 1/4] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2023-08-17 18:08 ` [PATCH 2/4] RISC-V: Add assembly support for TLSDESC Tatsuyuki Ishi
2023-08-17 18:08 ` [PATCH 3/4] RISC-V: Define and use GOT entry size constants for TLS Tatsuyuki Ishi
2023-08-17 18:08 ` [PATCH 4/4] RISC-V: Initial ld.bfd support for TLSDESC Tatsuyuki Ishi
2023-08-18  0:22 ` [PATCH 0/4] RISC-V: Implement TLS Descriptors Nelson Chu
2023-08-18  7:13   ` Fangrui Song [this message]
2023-08-31 17:13 ` [PATCH v2 0/5] " Tatsuyuki Ishi
2023-08-31 17:13   ` [PATCH v2 1/5] RISC-V: Fix local GOT and reloc size calculation for TLS Tatsuyuki Ishi
2024-02-21  0:49     ` Nelson Chu
2024-02-21  7:04       ` Nelson Chu
2023-08-31 17:13   ` [PATCH v2 2/5] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2023-08-31 17:13   ` [PATCH v2 3/5] RISC-V: Add assembly support for TLSDESC Tatsuyuki Ishi
2023-08-31 17:13   ` [PATCH v2 4/5] RISC-V: Define and use GOT entry size constants for TLS Tatsuyuki Ishi
2023-08-31 17:13   ` [PATCH v2 5/5] RISC-V: Initial ld.bfd support for TLSDESC Tatsuyuki Ishi
2023-11-28  8:51 ` [PATCH v3 0/9] RISC-V: Implement TLS Descriptors Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 1/9] RISC-V: Fix local GOT and reloc size calculation for TLS Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 2/9] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2024-02-19  0:49     ` Nelson Chu
2024-02-20 17:28       ` Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 3/9] RISC-V: Add assembly support for TLSDESC Tatsuyuki Ishi
2024-02-19  1:44     ` Nelson Chu
2024-02-20 17:29       ` Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 4/9] RISC-V: Define and use GOT entry size constants for TLS Tatsuyuki Ishi
2024-02-19  1:57     ` Nelson Chu
2024-02-20 17:32       ` Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 5/9] RISC-V: Initial ld.bfd support for TLSDESC Tatsuyuki Ishi
2024-02-19  4:33     ` Nelson Chu
2024-02-20 17:36       ` Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 6/9] RISC-V: Move STATIC_TLS handling into record_tls_type Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 7/9] RISC-V: Unify TLS handling in check_relocs Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 8/9] RISC-V: Add elf_link_hash_entry to relax_func args Tatsuyuki Ishi
2023-11-28  8:51   ` [PATCH v3 9/9] RISC-V: Introduce TLSDESC relaxation Tatsuyuki Ishi
2023-12-05 16:44   ` [PATCH v3 0/9] RISC-V: Implement TLS Descriptors Tatsuyuki Ishi
2023-12-06  0:33     ` Nelson Chu
2023-12-07  3:35       ` Fangrui Song
2023-12-13  0:27         ` Palmer Dabbelt
2023-12-13  1:53           ` Tatsuyuki Ishi
2024-01-27  0:57             ` Fangrui Song
2024-02-20 17:55 ` [PATCH v4 " Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 1/9] RISC-V: Fix local GOT and reloc size calculation for TLS Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 2/9] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 3/9] RISC-V: Add assembly support for TLSDESC Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 4/9] RISC-V: Define and use GOT entry size constants for TLS Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 5/9] RISC-V: Initial ld.bfd support for TLSDESC Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 6/9] RISC-V: Move STATIC_TLS handling into record_tls_type Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 7/9] RISC-V: Unify TLS handling in check_relocs Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 8/9] RISC-V: Add elf_link_hash_entry to relax_func args Tatsuyuki Ishi
2024-02-20 17:55   ` [PATCH v4 9/9] RISC-V: Introduce TLSDESC relaxation Tatsuyuki Ishi
2024-02-29  7:06   ` [PATCH v4 0/9] RISC-V: Implement TLS Descriptors Nelson Chu
2024-02-29  7:14     ` Tatsuyuki Ishi
2024-03-29  6:22       ` Tatsuyuki Ishi

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=DS7PR12MB576522B309E27EBA26708A85CB1BA@DS7PR12MB5765.namprd12.prod.outlook.com \
    --to=i@maskray.me \
    --cc=binutils@sourceware.org \
    --cc=ishitatsuyuki@gmail.com \
    --cc=nelson@rivosinc.com \
    --cc=rui314@gmail.com \
    --cc=ruiu@bluewhale.systems \
    /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).