[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20231222050403.GB52600@quark.localdomain>
Date: Thu, 21 Dec 2023 23:04:03 -0600
From: Eric Biggers <ebiggers@...nel.org>
To: Qingfang DENG <dqfext@...il.com>
Cc: Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>,
Conor Dooley <conor.dooley@...rochip.com>,
linux-riscv@...ts.infradead.org, linux-kernel@...r.kernel.org,
Jisheng Zhang <jszhang@...nel.org>
Subject: Re: [PATCH v2 1/2] riscv: introduce RISCV_EFFICIENT_UNALIGNED_ACCESS
On Tue, Dec 05, 2023 at 04:39:24PM +0800, Qingfang DENG wrote:
> Hi,
>
> You may as well remove the -mstrict-align CFLAGS in the Makefile, if
> this option is enabled:
>
> --- a/arch/riscv/Makefile
> +++ b/arch/riscv/Makefile
> @@ -108,7 +108,9 @@ KBUILD_AFLAGS_MODULE += $(call as-option,-Wa$(comma)-mno-relax)
> # unaligned accesses. While unaligned accesses are explicitly allowed in the
> # RISC-V ISA, they're emulated by machine mode traps on all extant
> # architectures. It's faster to have GCC emit only aligned accesses.
> +ifneq ($(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS),y)
> KBUILD_CFLAGS += $(call cc-option,-mstrict-align)
> +endif
>
Agreed. When CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y, we shouldn't use
-mstrict-align, so that the compiler can actually use unaligned memory accesses.
If I understand correctly, beyond the change requested above, people seem to be
happy with this patch. Jisheng, can you resend it with the above feedback
addressed? Thanks!
- Eric
Powered by blists - more mailing lists