Re: [PATCH 1/3] riscv: word-at-a-time: improve find_zero() for !RISCV_ISA_ZBB

From: Jisheng Zhang <jszhang@kernel.org>
To: Nam Cao <namcao@linutronix.de>
Cc: Paul Walmsley <pjw@kernel.org>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Alexandre Ghiti <alex@ghiti.fr>,
	linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/3] riscv: word-at-a-time: improve find_zero() for !RISCV_ISA_ZBB
Date: Tue, 1 Sep 2026 12:34:11 +0800	[thread overview]
Message-ID: <apZVwzxBD8Ih493W@xhacker> (raw)
In-Reply-To: <87mruh5h1m.fsf@yellow.woof>

On Thu, Aug 20, 2026 at 10:31:49AM +0200, Nam Cao wrote:
> Jisheng Zhang <jszhang@kernel.org> writes:
> > +#if !(defined(CONFIG_RISCV_ISA_ZBB) && defined(CONFIG_TOOLCHAIN_HAS_ZBB))
> > +#include <asm-generic/word-at-a-time.h>
> > +#else
> 
> Instead of this #if, would it be better to do
> 
> static inline unsigned long find_zero(unsigned long mask)
> {
>         if (IS_ENABLED(CONFIG_RISCV_ISA_ZBB) && IS_ENABLED(CONFIG_TOOLCHAIN_HAS_ZBB) &&
> 	    riscv_has_extension_likely(RISCV_ISA_EXT_ZBB))
> 		return !mask ? 0 : ((__fls(mask) + 1) >> 3);
> 
> 	return count_masked_bytes(mask);
> }
> 
> and let compiler's dead code elimination does its job?

This is impossible because the generic word-at-a-time.h implements
the generic find_zero() itself, so there will be compile error.

Thanks

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