| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20131108095842.GG28304@ns203013.ovh.net>
Date: Fri, 8 Nov 2013 10:58:42 +0100
From: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@...osoft.com>
To: Stephen Boyd <sboyd@...eaurora.org>
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ARM: Use udiv/sdiv for __aeabi_{u}idiv library functions
On 11:20 Thu 07 Nov , Stephen Boyd wrote:
> If we're running on a v7 ARM CPU, detect if the CPU supports the
> sdiv/udiv instructions and replace the signed and unsigned
> division library functions with an sdiv/udiv instruction.
>
> Running the perf messaging benchmark in pipe mode
>
> $ perf bench sched messaging -p
>
> shows a modest improvement on my v7 CPU.
>
> before:
> (5.060 + 5.960 + 5.971 + 5.643 + 6.029 + 5.665 + 6.050 + 5.870 + 6.117 + 5.683) / 10 = 5.805
>
> after:
> (4.884 + 5.549 + 5.749 + 6.001 + 5.460 + 5.103 + 5.956 + 6.112 + 5.468 + 5.093) / 10 = 5.538
>
> (5.805 - 5.538) / 5.805 = 4.6%
>
> Signed-off-by: Stephen Boyd <sboyd@...eaurora.org>
> ---
>
> Should we add in the __div0() call if the denominator is 0?
>
> arch/arm/kernel/setup.c | 10 +++++++++
> arch/arm/lib/Makefile | 3 +++
> arch/arm/lib/div-v7.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++++
> arch/arm/lib/lib1funcs.S | 16 +++++++++++++
> 4 files changed, 87 insertions(+)
> create mode 100644 arch/arm/lib/div-v7.c
>
> diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
> index 0e1e2b3..7d519f4 100644
> --- a/arch/arm/kernel/setup.c
> +++ b/arch/arm/kernel/setup.c
> @@ -30,6 +30,7 @@
> #include <linux/bug.h>
> #include <linux/compiler.h>
> #include <linux/sort.h>
> +#include <linux/static_key.h>
>
> #include <asm/unified.h>
> #include <asm/cp15.h>
> @@ -365,6 +366,8 @@ void __init early_print(const char *str, ...)
> printk("%s", buf);
> }
>
> +struct static_key cpu_has_idiv = STATIC_KEY_INIT_FALSE;
> +
> static void __init cpuid_init_hwcaps(void)
> {
> unsigned int divide_instrs, vmsa;
> @@ -381,6 +384,13 @@ static void __init cpuid_init_hwcaps(void)
> elf_hwcap |= HWCAP_IDIVT;
> }
>
> +#ifdef CONFIG_THUMB2_KERNEL
if (IS_ENABLED(CONFIG_THUMB2_KERNEL) && elf_hwcap & HWCAP_IDIVT)
> + if (elf_hwcap & HWCAP_IDIVT)
> +#else
> + if (elf_hwcap & HWCAP_IDIVA)
> +#endif
> + static_key_slow_inc(&cpu_has_idiv);
> +
> /* LPAE implies atomic ldrd/strd instructions */
> vmsa = (read_cpuid_ext(CPUID_EXT_MMFR0) & 0xf) >> 0;
> if (vmsa >= 5)
> diff --git a/arch/arm/lib/Makefile b/arch/arm/lib/Makefile
> index bd454b0..6ed6496 100644
> --- a/arch/arm/lib/Makefile
> +++ b/arch/arm/lib/Makefile
> @@ -15,6 +15,9 @@ lib-y := backtrace.o changebit.o csumipv6.o csumpartial.o \
> io-readsb.o io-writesb.o io-readsl.o io-writesl.o \
> call_with_stack.o
>
> +lib-$(CONFIG_CPU_V7) += div-v7.o
> +CFLAGS_div-v7.o := -march=armv7-a
> +
> mmu-y := clear_user.o copy_page.o getuser.o putuser.o
>
> # the code in uaccess.S is not preemption safe and
> diff --git a/arch/arm/lib/div-v7.c b/arch/arm/lib/div-v7.c
> new file mode 100644
> index 0000000..96ceb92
> --- /dev/null
> +++ b/arch/arm/lib/div-v7.c
> @@ -0,0 +1,58 @@
> +/* Copyright (c) 2013, The Linux Foundation. All rights reserved.
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 and
> + * only version 2 as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + */
> +
> +#include <linux/static_key.h>
> +
> +extern int ___aeabi_idiv(int, int);
> +extern unsigned ___aeabi_uidiv(int, int);
> +
> +extern struct static_key cpu_has_idiv;
> +
> +int __aeabi_idiv(int numerator, int denominator)
> +{
> + if (static_key_false(&cpu_has_idiv)) {
> + int ret;
> +
> + asm volatile (
> + ".arch_extension idiv\n"
> + "sdiv %0, %1, %2"
> + : "=&r" (ret)
> + : "r" (numerator), "r" (denominator));
> +
> + return ret;
> + }
> +
> + return ___aeabi_idiv(numerator, denominator);
> +}
> +
> +int __divsi3(int numerator, int denominator)
> + __attribute__((alias("__aeabi_idiv")));
> +
> +unsigned __aeabi_uidiv(int numerator, int denominator)
> +{
> + if (static_key_false(&cpu_has_idiv)) {
> + int ret;
> +
> + asm volatile (
> + ".arch_extension idiv\n"
> + "udiv %0, %1, %2"
> + : "=&r" (ret)
> + : "r" (numerator), "r" (denominator));
> +
> + return ret;
> + }
> +
> + return ___aeabi_uidiv(numerator, denominator);
> +}
> +
> +unsigned __udivsi3(int numerator, int denominator)
> + __attribute__((alias("__aeabi_uidiv")));
> diff --git a/arch/arm/lib/lib1funcs.S b/arch/arm/lib/lib1funcs.S
> index c562f64..adea088 100644
> --- a/arch/arm/lib/lib1funcs.S
> +++ b/arch/arm/lib/lib1funcs.S
> @@ -205,8 +205,12 @@ Boston, MA 02111-1307, USA. */
> .endm
>
>
> +#if defined(ZIMAGE) || !defined(CONFIG_CPU_V7)
> ENTRY(__udivsi3)
> ENTRY(__aeabi_uidiv)
> +#else
> +ENTRY(___aeabi_uidiv)
> +#endif
> UNWIND(.fnstart)
>
> subs r2, r1, #1
> @@ -232,8 +236,12 @@ UNWIND(.fnstart)
> mov pc, lr
>
> UNWIND(.fnend)
> +#if defined(ZIMAGE) || !defined(CONFIG_CPU_V7)
> ENDPROC(__udivsi3)
> ENDPROC(__aeabi_uidiv)
> +#else
> +ENDPROC(___aeabi_uidiv)
> +#endif
>
> ENTRY(__umodsi3)
> UNWIND(.fnstart)
> @@ -253,8 +261,12 @@ UNWIND(.fnstart)
> UNWIND(.fnend)
> ENDPROC(__umodsi3)
>
> +#if defined(ZIMAGE) || !defined(CONFIG_CPU_V7)
> ENTRY(__divsi3)
> ENTRY(__aeabi_idiv)
> +#else
> +ENTRY(___aeabi_idiv)
> +#endif
> UNWIND(.fnstart)
>
> cmp r1, #0
> @@ -293,8 +305,12 @@ UNWIND(.fnstart)
> mov pc, lr
>
> UNWIND(.fnend)
> +#if defined(ZIMAGE) || !defined(CONFIG_CPU_V7)
> ENDPROC(__divsi3)
> ENDPROC(__aeabi_idiv)
> +#else
> +ENDPROC(___aeabi_idiv)
> +#endif
>
> ENTRY(__modsi3)
> UNWIND(.fnstart)
> --
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> hosted by The Linux Foundation
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@...ts.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists