Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220824012624.2826445-1-yury.norov@gmail.com> <20220824012624.2826445-2-yury.norov@gmail.com>
In-Reply-To: <20220824012624.2826445-2-yury.norov@gmail.com>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Wed, 24 Aug 2022 12:10:02 +0300
Message-ID: <CAHp75VcB08oTrB8R9Zyo4Ja=c_XqybqdCw46fY4_MNqvSSCtLQ@mail.gmail.com>
Subject: Re: [PATCH v2 1/3] lib/find_bit: introduce FIND_FIRST_BIT() macro
To:     Yury Norov <yury.norov@gmail.com>
Cc:     Linus Torvalds <torvalds@linux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Guenter Roeck <linux@roeck-us.net>,
        Dennis Zhou <dennis@kernel.org>,
        Russell King <linux@armlinux.org.uk>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Rasmus Villemoes <linux@rasmusvillemoes.dk>,
        Alexey Klimov <aklimov@redhat.com>,
        Kees Cook <keescook@chromium.org>,
        Andy Whitcroft <apw@canonical.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Aug 24, 2022 at 4:51 AM Yury Norov <yury.norov@gmail.com> wrote:
>
> Now that we have many flavors of find_first_bit(), and expect even more,
> it's better to have one macro that generates optimal code for all and makes
> maintaining of slightly different functions simpler.
>
> The logic common to all versions is moved to the new macro, and all the
> flavors are generated by providing an EXPRESSION macro-parameter, like
> in this example:
>
>   #define FIND_FIRST_BIT(EXPRESSION, size) ...
>
>   find_first_ornot_and_bit(addr1, addr2, addr3, size)
>   {
>         return FIND_NEXT_BIT(addr1[idx] | ~addr2[idx] & addr3[idx], size);
>   }
>
> The EXPRESSION may be of any complexity, as soon as it only refers
> the bitmap(s) and an iterator idx.
>
> The 'word_op' is here to allow the macro to generate code for _le
> versions on big-endian machines; used in the following patches.

...

> +#ifndef word_op
> +#define word_op
> +#endif

Not sure about the naming without namespace. Perhaps __ffs_word_op?

> +#define FIND_FIRST_BIT(EXPRESSION, size)                                       \
> +({                                                                             \
> +       unsigned long idx, val, sz = (size);                                    \
> +                                                                               \
> +       for (idx = 0; idx * BITS_PER_LONG < sz; idx++) {                        \

I think we can do slightly better:

for (unsigned long idx = 0; idx < sz; idx += BITS_PER_LONG) {
  unsigned long val;

> +               val = (EXPRESSION);                                             \
> +               if (val) {                                                      \
> +                       sz = min(idx * BITS_PER_LONG + __ffs(word_op(val)), sz);\

sz = min(idx + __ffs(...));

> +                       break;                                                  \
> +               }                                                               \
> +       }                                                                       \
> +                                                                               \
> +       sz;                                                                     \
> +})


-- 
With Best Regards,
Andy Shevchenko