[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAAhV-H5Xt+1c_gCOpudzizurEbVXYonf_wWg9sTaCKWcONEL-Q@mail.gmail.com>
Date: Fri, 26 Feb 2021 09:22:18 +0800
From: Huacai Chen <chenhuacai@...nel.org>
To: Yury Norov <yury.norov@...il.com>
Cc: Will Deacon <will@...nel.org>,
Catalin Marinas <catalin.marinas@....com>,
Thomas Bogendoerfer <tsbogend@...ha.franken.de>,
Alexander Lobakin <alobakin@...me>,
Alexey Klimov <aklimov@...hat.com>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
LKML <linux-kernel@...r.kernel.org>, linux-arch@...r.kernel.org,
"open list:MIPS" <linux-mips@...r.kernel.org>
Subject: Re: [PATCH 2/2] MIPS: enable GENERIC_FIND_FIRST_BIT
Hi, Yury,
On Thu, Feb 25, 2021 at 9:59 PM Yury Norov <yury.norov@...il.com> wrote:
>
> From: Alexander Lobakin <alobakin@...me>
>
> MIPS doesn't have architecture-optimized bitsearching functions,
> like find_{first,next}_bit() etc.
Emm, I think MIPS can use clo/clz to optimize bitsearching functions.
Huacai
> It's absolutely harmless to enable GENERIC_FIND_FIRST_BIT as this
> functionality is not new at all and well-tested. It provides more
> optimized code and saves some .text memory (32 R2):
>
> add/remove: 4/1 grow/shrink: 1/53 up/down: 216/-372 (-156)
>
> Users of for_each_set_bit() like hotpath gic_handle_shared_int()
> will also benefit from this.
>
> Suggested-by: Yury Norov <yury.norov@...il.com>
> Signed-off-by: Alexander Lobakin <alobakin@...me>
> Signed-off-by: Yury Norov <yury.norov@...il.com>
> ---
> arch/mips/Kconfig | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig
> index d89efba3d8a4..164bdd715d4b 100644
> --- a/arch/mips/Kconfig
> +++ b/arch/mips/Kconfig
> @@ -26,6 +26,7 @@ config MIPS
> select GENERIC_ATOMIC64 if !64BIT
> select GENERIC_CMOS_UPDATE
> select GENERIC_CPU_AUTOPROBE
> + select GENERIC_FIND_FIRST_BIT
> select GENERIC_GETTIMEOFDAY
> select GENERIC_IOMAP
> select GENERIC_IRQ_PROBE
> --
> 2.25.1
>
Powered by blists - more mailing lists