public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
To: daichengrong@iscas.ac.cn, libc-alpha@sourceware.org
Cc: aswaterman@gmail.com, palmer@rivosinc.com, darius@bluespec.com,
	jiageng08@iscas.ac.cn
Subject: Re: [PATCH v4 3/4] add rvv memcpy in ifunc-impl-list
Date: Wed, 22 Jan 2025 09:19:23 -0300	[thread overview]
Message-ID: <9a2a7235-cce6-4161-9c1d-0d9ce7f4fa2c@linaro.org> (raw)
In-Reply-To: <20250122020401.187099-4-daichengrong@iscas.ac.cn>



On 21/01/25 23:04, daichengrong@iscas.ac.cn wrote:
> From: daichengrong <daichengrong@iscas.ac.cn>

This patch should be squashed with the last one, otherwise this will break
glibc build.

> 
> ---
>  .../unix/sysv/linux/riscv/multiarch/ifunc-impl-list.c    | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/sysdeps/unix/sysv/linux/riscv/multiarch/ifunc-impl-list.c b/sysdeps/unix/sysv/linux/riscv/multiarch/ifunc-impl-list.c
> index 1c1deca8f6..01846a7414 100644
> --- a/sysdeps/unix/sysv/linux/riscv/multiarch/ifunc-impl-list.c
> +++ b/sysdeps/unix/sysv/linux/riscv/multiarch/ifunc-impl-list.c
> @@ -27,6 +27,7 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
>    size_t i = max;
>  
>    bool fast_unaligned = false;
> +  bool rvv_ext = false;
>  
>    struct riscv_hwprobe pair = { .key = RISCV_HWPROBE_KEY_CPUPERF_0 };
>    if (__riscv_hwprobe (&pair, 1, 0, NULL, 0) == 0
> @@ -34,7 +35,15 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
>            == RISCV_HWPROBE_MISALIGNED_FAST)
>      fast_unaligned = true;
>  
> +  pair.key = RISCV_HWPROBE_KEY_IMA_EXT_0;
> +  pair.value = 0;
> +  if (__riscv_hwprobe (&pair, 1, 0, NULL, 0) == 0
> +      && (pair.value & RISCV_HWPROBE_IMA_V) == RISCV_HWPROBE_IMA_V)
> +    rvv_ext = true;

Maybe issue only one __riscv_hwprobe? Afaiu its idea is to allow query multiple
hardware support with only one call.

> +
>    IFUNC_IMPL (i, name, memcpy,
> +	      IFUNC_IMPL_ADD (array, i, memcpy, rvv_ext,
> +			      __memcpy_vector)
>  	      IFUNC_IMPL_ADD (array, i, memcpy, fast_unaligned,
>  			      __memcpy_noalignment)
>  	      IFUNC_IMPL_ADD (array, i, memcpy, 1, __memcpy_generic))


  reply	other threads:[~2025-01-22 12:19 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-22  2:03 [PATCH v4 0/4] RISC-V: add multiarch RVV support for memcpy using FMV IFUNC daichengrong
2025-01-22  2:03 ` [PATCH v4 1/4] add a configure test to check if the compiler supports rvv daichengrong
2025-01-22 12:16   ` Adhemerval Zanella Netto
2025-01-22  2:03 ` [PATCH v4 2/4] add macro support for dl_hwcap daichengrong
2025-01-22  2:04 ` [PATCH v4 3/4] add rvv memcpy in ifunc-impl-list daichengrong
2025-01-22 12:19   ` Adhemerval Zanella Netto [this message]
2025-01-22 16:34   ` Darius Rad
2025-01-22  2:04 ` [PATCH v4 4/4] add riscv vector support for memcpy daichengrong
2025-01-22 12:30   ` Adhemerval Zanella Netto

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=9a2a7235-cce6-4161-9c1d-0d9ce7f4fa2c@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --cc=aswaterman@gmail.com \
    --cc=daichengrong@iscas.ac.cn \
    --cc=darius@bluespec.com \
    --cc=jiageng08@iscas.ac.cn \
    --cc=libc-alpha@sourceware.org \
    --cc=palmer@rivosinc.com \
    /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).