| Message ID | 1391846481-31491-2-git-send-email-ak@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 02/08/2014 12:01 AM, Andi Kleen wrote: > The fancy x86 hweight uses different compiler options for the > hweight file. This does not work with LTO. Just disable the optimization > with LTO No, I'm going to NAK this. This means not using the POPCNT instruction if LTO is enabled, and that really isn't an acceptable option. -hpa -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sat, Feb 08, 2014 at 10:52:07AM -0800, H. Peter Anvin wrote: > On 02/08/2014 12:01 AM, Andi Kleen wrote: > > The fancy x86 hweight uses different compiler options for the > > hweight file. This does not work with LTO. Just disable the optimization > > with LTO > > No, I'm going to NAK this. This means not using the POPCNT instruction > if LTO is enabled, and that really isn't an acceptable option. I thought the use was obscure? Ok, suppose can just disable LTO for the file. The only drawback is that the functions will not be optimized away when not used, as they'll need to be __visible. -Andi
That's fine. On February 8, 2014 12:21:24 PM PST, Andi Kleen <ak@linux.intel.com> wrote: >On Sat, Feb 08, 2014 at 10:52:07AM -0800, H. Peter Anvin wrote: >> On 02/08/2014 12:01 AM, Andi Kleen wrote: >> > The fancy x86 hweight uses different compiler options for the >> > hweight file. This does not work with LTO. Just disable the >optimization >> > with LTO >> >> No, I'm going to NAK this. This means not using the POPCNT >instruction >> if LTO is enabled, and that really isn't an acceptable option. > >I thought the use was obscure? > >Ok, suppose can just disable LTO for the file. >The only drawback is that the functions will not be optimized away when >not used, as they'll need to be __visible. > >-Andi
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 940e50e..f125c5f 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -255,6 +255,7 @@ config X86_32_LAZY_GS config ARCH_HWEIGHT_CFLAGS string + default "" if LTO default "-fcall-saved-ecx -fcall-saved-edx" if X86_32 default "-fcall-saved-rdi -fcall-saved-rsi -fcall-saved-rdx -fcall-saved-rcx -fcall-saved-r8 -fcall-saved-r9 -fcall-saved-r10 -fcall-saved-r11" if X86_64 diff --git a/arch/x86/include/asm/arch_hweight.h b/arch/x86/include/asm/arch_hweight.h index 9686c3d..ca80549 100644 --- a/arch/x86/include/asm/arch_hweight.h +++ b/arch/x86/include/asm/arch_hweight.h @@ -25,9 +25,14 @@ static inline unsigned int __arch_hweight32(unsigned int w) { unsigned int res = 0; +#ifdef CONFIG_LTO + res = __sw_hweight32(w); +#else + asm (ALTERNATIVE("call __sw_hweight32", POPCNT32, X86_FEATURE_POPCNT) : "="REG_OUT (res) : REG_IN (w)); +#endif return res; } @@ -46,6 +51,9 @@ static inline unsigned long __arch_hweight64(__u64 w) { unsigned long res = 0; +#ifdef CONFIG_LTO + res = __sw_hweight64(w); +#else #ifdef CONFIG_X86_32 return __arch_hweight32((u32)w) + __arch_hweight32((u32)(w >> 32)); @@ -54,6 +62,7 @@ static inline unsigned long __arch_hweight64(__u64 w) : "="REG_OUT (res) : REG_IN (w)); #endif /* CONFIG_X86_32 */ +#endif return res; }
The fancy x86 hweight uses different compiler options for the hweight file. This does not work with LTO. Just disable the optimization with LTO v2: Simplify Kconfig checks (Jan Beulich) Cc: x86@kernel.org Signed-off-by: Andi Kleen <ak@linux.intel.com> --- arch/x86/Kconfig | 1 + arch/x86/include/asm/arch_hweight.h | 9 +++++++++ 2 files changed, 10 insertions(+)