Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit f5967101 authored by Borislav Petkov's avatar Borislav Petkov Committed by Ingo Molnar
Browse files

x86/hweight: Get rid of the special calling convention 



People complained about ARCH_HWEIGHT_CFLAGS and how it throws a wrench
into kcov, lto, etc, experimentations.

Add asm versions for __sw_hweight{32,64}() and do explicit saving and
restoring of clobbered registers. This gets rid of the special calling
convention. We get to call those functions on !X86_FEATURE_POPCNT CPUs.

We still need to hardcode POPCNT and register operands as some old gas
versions which we support, do not know about POPCNT.

Btw, remove redundant REX prefix from 32-bit POPCNT because alternatives
can do padding now.

Suggested-by: default avatarH. Peter Anvin <hpa@zytor.com>
Signed-off-by: default avatarBorislav Petkov <bp@suse.de>
Acked-by: default avatarPeter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1464605787-20603-1-git-send-email-bp@alien8.de


Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
parent 08dd8cd0

arch/x86/Kconfig

+0 −5
Original line number Diff line number Diff line
@@ -294,11 +294,6 @@ config X86_32_LAZY_GS 
	def_bool y

	depends on X86_32 && !CC_STACKPROTECTOR



config ARCH_HWEIGHT_CFLAGS

	string

	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



config ARCH_SUPPORTS_UPROBES

	def_bool y

arch/x86/include/asm/arch_hweight.h

+10 −14
Original line number Diff line number Diff line
@@ -4,8 +4,8 @@ 
#include <asm/cpufeatures.h>



#ifdef CONFIG_64BIT

/* popcnt %edi, %eax -- redundant REX prefix for alignment */

#define POPCNT32 ".byte 0xf3,0x40,0x0f,0xb8,0xc7"

/* popcnt %edi, %eax */

#define POPCNT32 ".byte 0xf3,0x0f,0xb8,0xc7"

/* popcnt %rdi, %rax */

#define POPCNT64 ".byte 0xf3,0x48,0x0f,0xb8,0xc7"

#define REG_IN "D"
@@ -17,15 +17,11 @@ 
#define REG_OUT "a"

#endif



/*

 * __sw_hweightXX are called from within the alternatives below

 * and callee-clobbered registers need to be taken care of. See

 * ARCH_HWEIGHT_CFLAGS in <arch/x86/Kconfig> for the respective

 * compiler switches.

 */

#define __HAVE_ARCH_SW_HWEIGHT



static __always_inline unsigned int __arch_hweight32(unsigned int w)

{

	unsigned int res = 0;

	unsigned int res;



	asm (ALTERNATIVE("call __sw_hweight32", POPCNT32, X86_FEATURE_POPCNT)

			 : "="REG_OUT (res)
@@ -53,7 +49,7 @@ static inline unsigned long __arch_hweight64(__u64 w) 
#else

static __always_inline unsigned long __arch_hweight64(__u64 w)

{

	unsigned long res = 0;

	unsigned long res;



	asm (ALTERNATIVE("call __sw_hweight64", POPCNT64, X86_FEATURE_POPCNT)

			 : "="REG_OUT (res)

arch/x86/kernel/i386_ksyms_32.c

+2 −0
Original line number Diff line number Diff line
@@ -42,3 +42,5 @@ EXPORT_SYMBOL(empty_zero_page); 
EXPORT_SYMBOL(___preempt_schedule);

EXPORT_SYMBOL(___preempt_schedule_notrace);

#endif



EXPORT_SYMBOL(__sw_hweight32);

arch/x86/kernel/x8664_ksyms_64.c

+3 −0
Original line number Diff line number Diff line
@@ -44,6 +44,9 @@ EXPORT_SYMBOL(clear_page); 


EXPORT_SYMBOL(csum_partial);



EXPORT_SYMBOL(__sw_hweight32);

EXPORT_SYMBOL(__sw_hweight64);



/*

 * Export string functions. We normally rely on gcc builtin for most of these,

 * but gcc sometimes decides not to inline them.

arch/x86/lib/Makefile

+1 −1
Original line number Diff line number Diff line
@@ -25,7 +25,7 @@ lib-y += memcpy_$(BITS).o 
lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem.o

lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o



obj-y += msr.o msr-reg.o msr-reg-export.o

obj-y += msr.o msr-reg.o msr-reg-export.o hweight.o



ifeq ($(CONFIG_X86_32),y)

        obj-y += atomic64_32.o