public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
To: Palmer Dabbelt <palmer@dabbelt.com>, ishitatsuyuki@gmail.com
Cc: libc-alpha@sourceware.org, rui314@gmail.com,
	ruiu@bluewhale.systems, schwab@linux-m68k.org,
	Andrew Waterman <andrew@sifive.com>,
	fweimer@redhat.com
Subject: Re: [PATCH v6 0/3] RISC-V: Implement TLS Descriptors.
Date: Tue, 30 Apr 2024 15:33:43 -0300	[thread overview]
Message-ID: <b0c7d88a-307f-4d94-9e47-1b1487c6f7c9@linaro.org> (raw)
In-Reply-To: <mhng-74f76a97-606b-4118-9439-35e3ca42052b@palmer-ri-x1c9a>



On 30/04/24 14:05, Palmer Dabbelt wrote:
> On Thu, 28 Mar 2024 23:18:31 PDT (-0700), ishitatsuyuki@gmail.com wrote:
>> This patchset implements TLS Descriptors (TLSDESC) for RISC-V targets, per
>> the ratified specification at [1].
>>
>> Partial support for TLSDESC is in binutils trunk. In [2] there is
>> additional support for relaxation which is required to pass
>> tst-tls1-static. GCC needs to be patched with [3].
>>
>> Passes binutils and gcc tests for rv32gcv and rv64gcv. For glibc I only
>> tested elf/ as other subsystems sometimes doesn't work well under qemu.
>>
>> This contribution is made on behalf of Blue Whale Systems, which has
>> copyright assignment on file with the FSF.
>>
>> v2: Fix end-of-file newlines.
>> v3: Fix segfaulting on the slow path of TLSDESC resolver.
>>     Fix handling of lazy relocations.
>> v4: Fix compiler warnings.
>>     Fix fast path stack alignment pointed out by Andrew.
>>     Fix style issues pointed out by Adhemerval.
>>     Include 2 missing prerequisite commits.
>>     Update localplt list.
>> v5: Add stubs for save/restore of vector registers.
>>     Add comments and reordered sections for readability in the TLSDESC assembly.
>>     Add clobber test sequence for tst-gnu2-tls2.
>> v6: Fix R_RISCV_NUM.
>>     Elaborate on the include guard commit (thanks Fangrui).
>>
>> [1]: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/373
>> [2]: https://inbox.sourceware.org/binutils/20240220175556.304692-1-ishitatsuyuki@gmail.com/
>> [3]: https://inbox.sourceware.org/gcc-patches/20240329055239.29719-1-ishitatsuyuki@gmail.com/
>>
>> Tatsuyuki Ishi (3):
>>   RISC-V: Add include guard for dl-tls.h.
>>   RISC-V: Add TLSDESC reloc definitions.
>>   RISC-V: Implement TLS Descriptors.
>>
>>  elf/elf.h                                   |   7 +-
>>  sysdeps/riscv/Makefile                      |  10 +
>>  sysdeps/riscv/dl-lookupcfg.h                |  27 ++
>>  sysdeps/riscv/dl-machine.h                  |  50 +++-
>>  sysdeps/riscv/dl-tls.h                      |   4 +
>>  sysdeps/riscv/dl-tlsdesc.S                  | 269 ++++++++++++++++++++
>>  sysdeps/riscv/dl-tlsdesc.h                  |  48 ++++
>>  sysdeps/riscv/linkmap.h                     |   1 +
>>  sysdeps/riscv/preconfigure                  |   1 +
>>  sysdeps/riscv/tlsdesc.c                     |  38 +++
>>  sysdeps/riscv/tlsdesc.sym                   |  19 ++
>>  sysdeps/riscv/tst-gnu2-tls2.c               |  33 +++
>>  sysdeps/unix/sysv/linux/riscv/localplt.data |   2 +
>>  13 files changed, 507 insertions(+), 2 deletions(-)
>>  create mode 100644 sysdeps/riscv/dl-lookupcfg.h
>>  create mode 100644 sysdeps/riscv/dl-tlsdesc.S
>>  create mode 100644 sysdeps/riscv/dl-tlsdesc.h
>>  create mode 100644 sysdeps/riscv/tlsdesc.c
>>  create mode 100644 sysdeps/riscv/tlsdesc.sym
>>  create mode 100644 sysdeps/riscv/tst-gnu2-tls2.c
> 
> This generally LGTM, but IIUC we're waiting on a glibc-wide decision as to what the stable ABIs should be around TLS before moving forward.

If I recall correctly, my latest remarks for this patch was:

1. Missing gcc/binutils upstream support (usually we wait the patches 
   to be upstream on the required projects)

2. Call the internal __tls_get_addr on _dl_tlsdesc_dynamic to avoid
   the PLT call.

3. Decide how RISCV would like to handle different ABI constraint for
   the _dl_tlsdesc_dynamic, now that you are moving to support ifunc
   and thus allowing a more generic glibc build.  You can either make
   the _dl_tlsdesc_dynamic with multiple ifunc variants (as x86 does),
   or check the chip support through the riscv_hwprobe on the function
   itself and only save/restore the support ABI register.

  reply	other threads:[~2024-04-30 18:33 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-17 18:12 [PATCH] " Tatsuyuki Ishi
2023-08-17 18:35 ` Andreas Schwab
2023-09-08 10:55 ` [PATCH v2] " Tatsuyuki Ishi
2023-09-13 17:26 ` [PATCH v3] " Tatsuyuki Ishi
2023-09-13 19:14   ` Adhemerval Zanella Netto
2023-09-14  8:39     ` Tatsuyuki Ishi
2023-09-14 12:09       ` Adhemerval Zanella Netto
2024-01-27  2:22         ` Fangrui Song
2023-09-13 19:07 ` [PATCH] " Andrew Waterman
2023-09-14  8:40 ` [PATCH v4 0/3] " Tatsuyuki Ishi
2023-09-14  8:40   ` [PATCH v4 1/3] RISC-V: Add include guard for dl-tls.h Tatsuyuki Ishi
2024-01-27  1:14     ` Fangrui Song
2023-09-14  8:40   ` [PATCH v4 2/3] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2024-01-27  1:12     ` Fangrui Song
2023-09-14  8:40   ` [PATCH v4 3/3] RISC-V: Implement TLS Descriptors Tatsuyuki Ishi
2023-11-23 11:39     ` Florian Weimer
2024-03-29  5:55 ` [PATCH v5 0/3] " Tatsuyuki Ishi
2024-03-29  5:55   ` [PATCH v5 1/3] RISC-V: Add include guard for dl-tls.h Tatsuyuki Ishi
2024-03-29  5:55   ` [PATCH v5 2/3] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2024-03-29  5:55   ` [PATCH v5 3/3] RISC-V: Implement TLS Descriptors Tatsuyuki Ishi
2024-03-29  6:18 ` [PATCH v6 0/3] " Tatsuyuki Ishi
2024-03-29  6:18   ` [PATCH v6 1/3] RISC-V: Add include guard for dl-tls.h Tatsuyuki Ishi
2024-04-03 11:48     ` Adhemerval Zanella Netto
2024-03-29  6:18   ` [PATCH v6 2/3] RISC-V: Add TLSDESC reloc definitions Tatsuyuki Ishi
2024-04-03  5:10     ` Fangrui Song
2024-04-03  8:03     ` Andreas Schwab
2024-03-29  6:18   ` [PATCH v6 3/3] RISC-V: Implement TLS Descriptors Tatsuyuki Ishi
2024-04-01 13:23     ` Florian Weimer
2024-04-01 19:29     ` Adhemerval Zanella Netto
2024-04-02  3:36       ` Tatsuyuki Ishi
2024-04-02 13:35         ` Adhemerval Zanella Netto
2024-04-02 15:25           ` Palmer Dabbelt
2024-04-02 15:32             ` Adhemerval Zanella Netto
2024-04-02 16:37               ` Palmer Dabbelt
2024-04-30 17:05   ` [PATCH v6 0/3] " Palmer Dabbelt
2024-04-30 18:33     ` Adhemerval Zanella Netto [this message]
2024-05-01  1:36       ` 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=b0c7d88a-307f-4d94-9e47-1b1487c6f7c9@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --cc=andrew@sifive.com \
    --cc=fweimer@redhat.com \
    --cc=ishitatsuyuki@gmail.com \
    --cc=libc-alpha@sourceware.org \
    --cc=palmer@dabbelt.com \
    --cc=rui314@gmail.com \
    --cc=ruiu@bluewhale.systems \
    --cc=schwab@linux-m68k.org \
    /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).