linux-riscv.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Atish Patra <atishp@atishpatra.org>
To: Jisheng Zhang <jszhang@kernel.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	 Albert Ou <aou@eecs.berkeley.edu>,
	Anup Patel <anup@brainfault.org>,
	 linux-riscv <linux-riscv@lists.infradead.org>,
	 "linux-kernel@vger.kernel.org List"
	<linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/2] riscv: introduce unified static key mechanism for ISA extensions
Date: Sun, 22 May 2022 00:33:12 -0700	[thread overview]
Message-ID: <CAOnJCU+aRoLRUjbum3o8N8J63vW+Y2974TjBueE__7PQAeJCQw@mail.gmail.com> (raw)
In-Reply-To: <20220517184453.3558-2-jszhang@kernel.org>

On Tue, May 17, 2022 at 11:53 AM Jisheng Zhang <jszhang@kernel.org> wrote:
>
> Currently, riscv has several extensions which may not be supported on
> all riscv platforms, for example, FPU and so on. To support unified
> kernel Image style, we need to check whether the feature is supported

/s/suportted/supported


> or not. If the check sits at hot code path, then performance will be
> impacted a lot. static key can be used to solve the issue. In the past
> FPU support has been converted to use static key mechanism. I believe
> we will have similar cases in the future.
>
> this patch tries to add an unified mechanism to use static keys for
> some ISA extensions by implementing an array of default-false static keys
> and enabling them when detected.
>
> Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
> ---
>  arch/riscv/include/asm/hwcap.h | 40 ++++++++++++++++++++++++++++++++++
>  arch/riscv/kernel/cpufeature.c |  7 ++++++
>  2 files changed, 47 insertions(+)
>
> diff --git a/arch/riscv/include/asm/hwcap.h b/arch/riscv/include/asm/hwcap.h
> index 0734e42f74f2..b0433d2b880d 100644
> --- a/arch/riscv/include/asm/hwcap.h
> +++ b/arch/riscv/include/asm/hwcap.h
> @@ -12,6 +12,7 @@
>  #include <uapi/asm/hwcap.h>
>
>  #ifndef __ASSEMBLY__
> +#include <linux/jump_label.h>
>  /*
>   * This yields a mask that user programs can use to figure out what
>   * instruction set this cpu supports.
> @@ -55,6 +56,16 @@ enum riscv_isa_ext_id {
>         RISCV_ISA_EXT_ID_MAX = RISCV_ISA_EXT_MAX,
>  };
>
> +/*
> + * This enum represents the logical ID for each RISC-V ISA extension static
> + * keys. We can use static key to optimize code path if some ISA extensions
> + * are available.
> + */
> +enum riscv_isa_ext_key {
> +       RISCV_ISA_EXT_KEY_FPU,          /* For 'F' and 'D' */
> +       RISCV_ISA_EXT_KEY_MAX,
> +};
> +
>  struct riscv_isa_ext_data {
>         /* Name of the extension displayed to userspace via /proc/cpuinfo */
>         char uprop[RISCV_ISA_EXT_NAME_LEN_MAX];
> @@ -62,6 +73,35 @@ struct riscv_isa_ext_data {
>         unsigned int isa_ext_id;
>  };
>
> +extern struct static_key_false riscv_isa_ext_keys[RISCV_ISA_EXT_KEY_MAX];
> +
> +static __always_inline int riscv_isa_ext2key(int num)
> +{
> +       switch (num) {
> +       case RISCV_ISA_EXT_f:
> +               return RISCV_ISA_EXT_KEY_FPU;
> +       case RISCV_ISA_EXT_d:
> +               return RISCV_ISA_EXT_KEY_FPU;
> +       default:
> +               return -EINVAL;
> +       }
> +}
> +
> +/*
> + * @num must be a compile-time constant.
> + */
> +static __always_inline bool riscv_isa_have_key_extension(int num)
> +{
> +       if (RISCV_ISA_EXT_ID_MAX <= num)
> +               return false;
> +
> +       num = riscv_isa_ext2key(num);
 > +       if (RISCV_ISA_EXT_KEY_MAX <= num || num < 0)
> +               return false;
> +

Why do you need the additional check in the hot path ?
riscv_isa_ext_keys array can be directly accessed at the caller
instead of calling this function.

> +       return static_branch_likely(&riscv_isa_ext_keys[num]);
> +}
> +



>  unsigned long riscv_isa_extension_base(const unsigned long *isa_bitmap);
>
>  #define riscv_isa_extension_mask(ext) BIT_MASK(RISCV_ISA_EXT_##ext)
> diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
> index 1b2d42d7f589..89f886b35357 100644
> --- a/arch/riscv/kernel/cpufeature.c
> +++ b/arch/riscv/kernel/cpufeature.c
> @@ -24,6 +24,8 @@ static DECLARE_BITMAP(riscv_isa, RISCV_ISA_EXT_MAX) __read_mostly;
>  #ifdef CONFIG_FPU
>  __ro_after_init DEFINE_STATIC_KEY_FALSE(cpu_hwcap_fpu);
>  #endif
> +__ro_after_init DEFINE_STATIC_KEY_ARRAY_FALSE(riscv_isa_ext_keys, RISCV_ISA_EXT_KEY_MAX);
> +EXPORT_SYMBOL(riscv_isa_ext_keys);
>
>  /**
>   * riscv_isa_extension_base() - Get base extension word
> @@ -232,6 +234,11 @@ void __init riscv_fill_hwcap(void)
>                         print_str[j++] = (char)('a' + i);
>         pr_info("riscv: ELF capabilities %s\n", print_str);
>
> +       for_each_set_bit(i, riscv_isa, RISCV_ISA_EXT_MAX) {
> +               j = riscv_isa_ext2key(i);
> +               if (j >= 0)
> +                       static_branch_enable(&riscv_isa_ext_keys[j]);
> +       }
>  #ifdef CONFIG_FPU
>         if (elf_hwcap & (COMPAT_HWCAP_ISA_F | COMPAT_HWCAP_ISA_D))
>                 static_branch_enable(&cpu_hwcap_fpu);
> --
> 2.34.1
>


--
Regards,
Atish

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

  reply	other threads:[~2022-05-22  7:33 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-17 18:44 [PATCH 0/2] introduce unified static key mechanism for ISA ext Jisheng Zhang
2022-05-17 18:44 ` [PATCH 1/2] riscv: introduce unified static key mechanism for ISA extensions Jisheng Zhang
2022-05-22  7:33   ` Atish Patra [this message]
2022-05-22 14:56     ` Jisheng Zhang
2022-05-17 18:44 ` [PATCH 2/2] riscv: switch has_fpu() to the unified static key mechanism Jisheng Zhang

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=CAOnJCU+aRoLRUjbum3o8N8J63vW+Y2974TjBueE__7PQAeJCQw@mail.gmail.com \
    --to=atishp@atishpatra.org \
    --cc=anup@brainfault.org \
    --cc=aou@eecs.berkeley.edu \
    --cc=jszhang@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.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 NNTP newsgroup(s).