diff options
Diffstat (limited to 'arch/x86/include/asm/percpu.h')
| -rw-r--r-- | arch/x86/include/asm/percpu.h | 343 |
1 files changed, 272 insertions, 71 deletions
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h index f899e01a8ac..851bcdc5db0 100644 --- a/arch/x86/include/asm/percpu.h +++ b/arch/x86/include/asm/percpu.h @@ -45,14 +45,14 @@ #include <linux/stringify.h> #ifdef CONFIG_SMP -#define __percpu_arg(x) "%%"__stringify(__percpu_seg)":%P" #x -#define __my_cpu_offset percpu_read(this_cpu_off) +#define __percpu_prefix "%%"__stringify(__percpu_seg)":" +#define __my_cpu_offset this_cpu_read(this_cpu_off) /* * Compared to the generic __my_cpu_offset version, the following * saves one instruction and avoids clobbering a temp register. */ -#define __this_cpu_ptr(ptr) \ +#define raw_cpu_ptr(ptr) \ ({ \ unsigned long tcp_ptr__; \ __verify_pcpu_ptr(ptr); \ @@ -62,9 +62,11 @@ (typeof(*(ptr)) __kernel __force *)tcp_ptr__; \ }) #else -#define __percpu_arg(x) "%P" #x +#define __percpu_prefix "" #endif +#define __percpu_arg(x) __percpu_prefix "%P" #x + /* * Initialized pointers to per-cpu variables needed for the boot * processor need to use these macros to get the proper address @@ -126,7 +128,8 @@ do { \ do { \ typedef typeof(var) pao_T__; \ const int pao_ID__ = (__builtin_constant_p(val) && \ - ((val) == 1 || (val) == -1)) ? (val) : 0; \ + ((val) == 1 || (val) == -1)) ? \ + (int)(val) : 0; \ if (0) { \ pao_T__ pao_tmp__; \ pao_tmp__ = (val); \ @@ -230,43 +233,154 @@ do { \ }) /* - * percpu_read() makes gcc load the percpu variable every time it is - * accessed while percpu_read_stable() allows the value to be cached. - * percpu_read_stable() is more efficient and can be used if its value + * Add return operation + */ +#define percpu_add_return_op(var, val) \ +({ \ + typeof(var) paro_ret__ = val; \ + switch (sizeof(var)) { \ + case 1: \ + asm("xaddb %0, "__percpu_arg(1) \ + : "+q" (paro_ret__), "+m" (var) \ + : : "memory"); \ + break; \ + case 2: \ + asm("xaddw %0, "__percpu_arg(1) \ + : "+r" (paro_ret__), "+m" (var) \ + : : "memory"); \ + break; \ + case 4: \ + asm("xaddl %0, "__percpu_arg(1) \ + : "+r" (paro_ret__), "+m" (var) \ + : : "memory"); \ + break; \ + case 8: \ + asm("xaddq %0, "__percpu_arg(1) \ + : "+re" (paro_ret__), "+m" (var) \ + : : "memory"); \ + break; \ + default: __bad_percpu_size(); \ + } \ + paro_ret__ += val; \ + paro_ret__; \ +}) + +/* + * xchg is implemented using cmpxchg without a lock prefix. xchg is + * expensive due to the implied lock prefix. The processor cannot prefetch + * cachelines if xchg is used. + */ +#define percpu_xchg_op(var, nval) \ +({ \ + typeof(var) pxo_ret__; \ + typeof(var) pxo_new__ = (nval); \ + switch (sizeof(var)) { \ + case 1: \ + asm("\n\tmov "__percpu_arg(1)",%%al" \ + "\n1:\tcmpxchgb %2, "__percpu_arg(1) \ + "\n\tjnz 1b" \ + : "=&a" (pxo_ret__), "+m" (var) \ + : "q" (pxo_new__) \ + : "memory"); \ + break; \ + case 2: \ + asm("\n\tmov "__percpu_arg(1)",%%ax" \ + "\n1:\tcmpxchgw %2, "__percpu_arg(1) \ + "\n\tjnz 1b" \ + : "=&a" (pxo_ret__), "+m" (var) \ + : "r" (pxo_new__) \ + : "memory"); \ + break; \ + case 4: \ + asm("\n\tmov "__percpu_arg(1)",%%eax" \ + "\n1:\tcmpxchgl %2, "__percpu_arg(1) \ + "\n\tjnz 1b" \ + : "=&a" (pxo_ret__), "+m" (var) \ + : "r" (pxo_new__) \ + : "memory"); \ + break; \ + case 8: \ + asm("\n\tmov "__percpu_arg(1)",%%rax" \ + "\n1:\tcmpxchgq %2, "__percpu_arg(1) \ + "\n\tjnz 1b" \ + : "=&a" (pxo_ret__), "+m" (var) \ + : "r" (pxo_new__) \ + : "memory"); \ + break; \ + default: __bad_percpu_size(); \ + } \ + pxo_ret__; \ +}) + +/* + * cmpxchg has no such implied lock semantics as a result it is much + * more efficient for cpu local operations. + */ +#define percpu_cmpxchg_op(var, oval, nval) \ +({ \ + typeof(var) pco_ret__; \ + typeof(var) pco_old__ = (oval); \ + typeof(var) pco_new__ = (nval); \ + switch (sizeof(var)) { \ + case 1: \ + asm("cmpxchgb %2, "__percpu_arg(1) \ + : "=a" (pco_ret__), "+m" (var) \ + : "q" (pco_new__), "0" (pco_old__) \ + : "memory"); \ + break; \ + case 2: \ + asm("cmpxchgw %2, "__percpu_arg(1) \ + : "=a" (pco_ret__), "+m" (var) \ + : "r" (pco_new__), "0" (pco_old__) \ + : "memory"); \ + break; \ + case 4: \ + asm("cmpxchgl %2, "__percpu_arg(1) \ + : "=a" (pco_ret__), "+m" (var) \ + : "r" (pco_new__), "0" (pco_old__) \ + : "memory"); \ + break; \ + case 8: \ + asm("cmpxchgq %2, "__percpu_arg(1) \ + : "=a" (pco_ret__), "+m" (var) \ + : "r" (pco_new__), "0" (pco_old__) \ + : "memory"); \ + break; \ + default: __bad_percpu_size(); \ + } \ + pco_ret__; \ +}) + +/* + * this_cpu_read() makes gcc load the percpu variable every time it is + * accessed while this_cpu_read_stable() allows the value to be cached. + * this_cpu_read_stable() is more efficient and can be used if its value * is guaranteed to be valid across cpus. The current users include * get_current() and get_thread_info() both of which are actually * per-thread variables implemented as per-cpu variables and thus * stable for the duration of the respective task. */ -#define percpu_read(var) percpu_from_op("mov", var, "m" (var)) -#define percpu_read_stable(var) percpu_from_op("mov", var, "p" (&(var))) -#define percpu_write(var, val) percpu_to_op("mov", var, val) -#define percpu_add(var, val) percpu_add_op(var, val) -#define percpu_sub(var, val) percpu_add_op(var, -(val)) -#define percpu_and(var, val) percpu_to_op("and", var, val) -#define percpu_or(var, val) percpu_to_op("or", var, val) -#define percpu_xor(var, val) percpu_to_op("xor", var, val) -#define percpu_inc(var) percpu_unary_op("inc", var) - -#define __this_cpu_read_1(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) -#define __this_cpu_read_2(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) -#define __this_cpu_read_4(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) - -#define __this_cpu_write_1(pcp, val) percpu_to_op("mov", (pcp), val) -#define __this_cpu_write_2(pcp, val) percpu_to_op("mov", (pcp), val) -#define __this_cpu_write_4(pcp, val) percpu_to_op("mov", (pcp), val) -#define __this_cpu_add_1(pcp, val) percpu_add_op((pcp), val) -#define __this_cpu_add_2(pcp, val) percpu_add_op((pcp), val) -#define __this_cpu_add_4(pcp, val) percpu_add_op((pcp), val) -#define __this_cpu_and_1(pcp, val) percpu_to_op("and", (pcp), val) -#define __this_cpu_and_2(pcp, val) percpu_to_op("and", (pcp), val) -#define __this_cpu_and_4(pcp, val) percpu_to_op("and", (pcp), val) -#define __this_cpu_or_1(pcp, val) percpu_to_op("or", (pcp), val) -#define __this_cpu_or_2(pcp, val) percpu_to_op("or", (pcp), val) -#define __this_cpu_or_4(pcp, val) percpu_to_op("or", (pcp), val) -#define __this_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val) -#define __this_cpu_xor_2(pcp, val) percpu_to_op("xor", (pcp), val) -#define __this_cpu_xor_4(pcp, val) percpu_to_op("xor", (pcp), val) +#define this_cpu_read_stable(var) percpu_from_op("mov", var, "p" (&(var))) + +#define raw_cpu_read_1(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) +#define raw_cpu_read_2(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) +#define raw_cpu_read_4(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) + +#define raw_cpu_write_1(pcp, val) percpu_to_op("mov", (pcp), val) +#define raw_cpu_write_2(pcp, val) percpu_to_op("mov", (pcp), val) +#define raw_cpu_write_4(pcp, val) percpu_to_op("mov", (pcp), val) +#define raw_cpu_add_1(pcp, val) percpu_add_op((pcp), val) +#define raw_cpu_add_2(pcp, val) percpu_add_op((pcp), val) +#define raw_cpu_add_4(pcp, val) percpu_add_op((pcp), val) +#define raw_cpu_and_1(pcp, val) percpu_to_op("and", (pcp), val) +#define raw_cpu_and_2(pcp, val) percpu_to_op("and", (pcp), val) +#define raw_cpu_and_4(pcp, val) percpu_to_op("and", (pcp), val) +#define raw_cpu_or_1(pcp, val) percpu_to_op("or", (pcp), val) +#define raw_cpu_or_2(pcp, val) percpu_to_op("or", (pcp), val) +#define raw_cpu_or_4(pcp, val) percpu_to_op("or", (pcp), val) +#define raw_cpu_xchg_1(pcp, val) percpu_xchg_op(pcp, val) +#define raw_cpu_xchg_2(pcp, val) percpu_xchg_op(pcp, val) +#define raw_cpu_xchg_4(pcp, val) percpu_xchg_op(pcp, val) #define this_cpu_read_1(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) #define this_cpu_read_2(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) @@ -283,46 +397,85 @@ do { \ #define this_cpu_or_1(pcp, val) percpu_to_op("or", (pcp), val) #define this_cpu_or_2(pcp, val) percpu_to_op("or", (pcp), val) #define this_cpu_or_4(pcp, val) percpu_to_op("or", (pcp), val) -#define this_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val) -#define this_cpu_xor_2(pcp, val) percpu_to_op("xor", (pcp), val) -#define this_cpu_xor_4(pcp, val) percpu_to_op("xor", (pcp), val) - -#define irqsafe_cpu_add_1(pcp, val) percpu_add_op((pcp), val) -#define irqsafe_cpu_add_2(pcp, val) percpu_add_op((pcp), val) -#define irqsafe_cpu_add_4(pcp, val) percpu_add_op((pcp), val) -#define irqsafe_cpu_and_1(pcp, val) percpu_to_op("and", (pcp), val) -#define irqsafe_cpu_and_2(pcp, val) percpu_to_op("and", (pcp), val) -#define irqsafe_cpu_and_4(pcp, val) percpu_to_op("and", (pcp), val) -#define irqsafe_cpu_or_1(pcp, val) percpu_to_op("or", (pcp), val) -#define irqsafe_cpu_or_2(pcp, val) percpu_to_op("or", (pcp), val) -#define irqsafe_cpu_or_4(pcp, val) percpu_to_op("or", (pcp), val) -#define irqsafe_cpu_xor_1(pcp, val) percpu_to_op("xor", (pcp), val) -#define irqsafe_cpu_xor_2(pcp, val) percpu_to_op("xor", (pcp), val) -#define irqsafe_cpu_xor_4(pcp, val) percpu_to_op("xor", (pcp), val) +#define this_cpu_xchg_1(pcp, nval) percpu_xchg_op(pcp, nval) +#define this_cpu_xchg_2(pcp, nval) percpu_xchg_op(pcp, nval) +#define this_cpu_xchg_4(pcp, nval) percpu_xchg_op(pcp, nval) + +#define raw_cpu_add_return_1(pcp, val) percpu_add_return_op(pcp, val) +#define raw_cpu_add_return_2(pcp, val) percpu_add_return_op(pcp, val) +#define raw_cpu_add_return_4(pcp, val) percpu_add_return_op(pcp, val) +#define raw_cpu_cmpxchg_1(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) +#define raw_cpu_cmpxchg_2(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) +#define raw_cpu_cmpxchg_4(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) + +#define this_cpu_add_return_1(pcp, val) percpu_add_return_op(pcp, val) +#define this_cpu_add_return_2(pcp, val) percpu_add_return_op(pcp, val) +#define this_cpu_add_return_4(pcp, val) percpu_add_return_op(pcp, val) +#define this_cpu_cmpxchg_1(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) +#define this_cpu_cmpxchg_2(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) +#define this_cpu_cmpxchg_4(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) + +#ifdef CONFIG_X86_CMPXCHG64 +#define percpu_cmpxchg8b_double(pcp1, pcp2, o1, o2, n1, n2) \ +({ \ + bool __ret; \ + typeof(pcp1) __o1 = (o1), __n1 = (n1); \ + typeof(pcp2) __o2 = (o2), __n2 = (n2); \ + asm volatile("cmpxchg8b "__percpu_arg(1)"\n\tsetz %0\n\t" \ + : "=a" (__ret), "+m" (pcp1), "+m" (pcp2), "+d" (__o2) \ + : "b" (__n1), "c" (__n2), "a" (__o1)); \ + __ret; \ +}) + +#define raw_cpu_cmpxchg_double_4 percpu_cmpxchg8b_double +#define this_cpu_cmpxchg_double_4 percpu_cmpxchg8b_double +#endif /* CONFIG_X86_CMPXCHG64 */ /* * Per cpu atomic 64 bit operations are only available under 64 bit. * 32 bit must fall back to generic operations. */ #ifdef CONFIG_X86_64 -#define __this_cpu_read_8(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) -#define __this_cpu_write_8(pcp, val) percpu_to_op("mov", (pcp), val) -#define __this_cpu_add_8(pcp, val) percpu_add_op((pcp), val) -#define __this_cpu_and_8(pcp, val) percpu_to_op("and", (pcp), val) -#define __this_cpu_or_8(pcp, val) percpu_to_op("or", (pcp), val) -#define __this_cpu_xor_8(pcp, val) percpu_to_op("xor", (pcp), val) - -#define this_cpu_read_8(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) -#define this_cpu_write_8(pcp, val) percpu_to_op("mov", (pcp), val) -#define this_cpu_add_8(pcp, val) percpu_add_op((pcp), val) -#define this_cpu_and_8(pcp, val) percpu_to_op("and", (pcp), val) -#define this_cpu_or_8(pcp, val) percpu_to_op("or", (pcp), val) -#define this_cpu_xor_8(pcp, val) percpu_to_op("xor", (pcp), val) - -#define irqsafe_cpu_add_8(pcp, val) percpu_add_op((pcp), val) -#define irqsafe_cpu_and_8(pcp, val) percpu_to_op("and", (pcp), val) -#define irqsafe_cpu_or_8(pcp, val) percpu_to_op("or", (pcp), val) -#define irqsafe_cpu_xor_8(pcp, val) percpu_to_op("xor", (pcp), val) +#define raw_cpu_read_8(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) +#define raw_cpu_write_8(pcp, val) percpu_to_op("mov", (pcp), val) +#define raw_cpu_add_8(pcp, val) percpu_add_op((pcp), val) +#define raw_cpu_and_8(pcp, val) percpu_to_op("and", (pcp), val) +#define raw_cpu_or_8(pcp, val) percpu_to_op("or", (pcp), val) +#define raw_cpu_add_return_8(pcp, val) percpu_add_return_op(pcp, val) +#define raw_cpu_xchg_8(pcp, nval) percpu_xchg_op(pcp, nval) +#define raw_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) + +#define this_cpu_read_8(pcp) percpu_from_op("mov", (pcp), "m"(pcp)) +#define this_cpu_write_8(pcp, val) percpu_to_op("mov", (pcp), val) +#define this_cpu_add_8(pcp, val) percpu_add_op((pcp), val) +#define this_cpu_and_8(pcp, val) percpu_to_op("and", (pcp), val) +#define this_cpu_or_8(pcp, val) percpu_to_op("or", (pcp), val) +#define this_cpu_add_return_8(pcp, val) percpu_add_return_op(pcp, val) +#define this_cpu_xchg_8(pcp, nval) percpu_xchg_op(pcp, nval) +#define this_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(pcp, oval, nval) + +/* + * Pretty complex macro to generate cmpxchg16 instruction. The instruction + * is not supported on early AMD64 processors so we must be able to emulate + * it in software. The address used in the cmpxchg16 instruction must be + * aligned to a 16 byte boundary. + */ +#define percpu_cmpxchg16b_double(pcp1, pcp2, o1, o2, n1, n2) \ +({ \ + bool __ret; \ + typeof(pcp1) __o1 = (o1), __n1 = (n1); \ + typeof(pcp2) __o2 = (o2), __n2 = (n2); \ + alternative_io("leaq %P1,%%rsi\n\tcall this_cpu_cmpxchg16b_emu\n\t", \ + "cmpxchg16b " __percpu_arg(1) "\n\tsetz %0\n\t", \ + X86_FEATURE_CX16, \ + ASM_OUTPUT2("=a" (__ret), "+m" (pcp1), \ + "+m" (pcp2), "+d" (__o2)), \ + "b" (__n1), "c" (__n2), "a" (__o1) : "rsi"); \ + __ret; \ +}) + +#define raw_cpu_cmpxchg_double_8 percpu_cmpxchg16b_double +#define this_cpu_cmpxchg_double_8 percpu_cmpxchg16b_double #endif @@ -336,6 +489,37 @@ do { \ old__; \ }) +static __always_inline int x86_this_cpu_constant_test_bit(unsigned int nr, + const unsigned long __percpu *addr) +{ + unsigned long __percpu *a = (unsigned long *)addr + nr / BITS_PER_LONG; + +#ifdef CONFIG_X86_64 + return ((1UL << (nr % BITS_PER_LONG)) & raw_cpu_read_8(*a)) != 0; +#else + return ((1UL << (nr % BITS_PER_LONG)) & raw_cpu_read_4(*a)) != 0; +#endif +} + +static inline int x86_this_cpu_variable_test_bit(int nr, + const unsigned long __percpu *addr) +{ + int oldbit; + + asm volatile("bt "__percpu_arg(2)",%1\n\t" + "sbb %0,%0" + : "=r" (oldbit) + : "m" (*(unsigned long *)addr), "Ir" (nr)); + + return oldbit; +} + +#define x86_this_cpu_test_bit(nr, addr) \ + (__builtin_constant_p((nr)) \ + ? x86_this_cpu_constant_test_bit((nr), (addr)) \ + : x86_this_cpu_variable_test_bit((nr), (addr))) + + #include <asm-generic/percpu.h> /* We can use this directly for local CPU (faster). */ @@ -357,6 +541,12 @@ DECLARE_PER_CPU(unsigned long, this_cpu_off); { [0 ... NR_CPUS-1] = _initvalue }; \ __typeof__(_type) *_name##_early_ptr __refdata = _name##_early_map +#define DEFINE_EARLY_PER_CPU_READ_MOSTLY(_type, _name, _initvalue) \ + DEFINE_PER_CPU_READ_MOSTLY(_type, _name) = _initvalue; \ + __typeof__(_type) _name##_early_map[NR_CPUS] __initdata = \ + { [0 ... NR_CPUS-1] = _initvalue }; \ + __typeof__(_type) *_name##_early_ptr __refdata = _name##_early_map + #define EXPORT_EARLY_PER_CPU_SYMBOL(_name) \ EXPORT_PER_CPU_SYMBOL(_name) @@ -365,6 +555,11 @@ DECLARE_PER_CPU(unsigned long, this_cpu_off); extern __typeof__(_type) *_name##_early_ptr; \ extern __typeof__(_type) _name##_early_map[] +#define DECLARE_EARLY_PER_CPU_READ_MOSTLY(_type, _name) \ + DECLARE_PER_CPU_READ_MOSTLY(_type, _name); \ + extern __typeof__(_type) *_name##_early_ptr; \ + extern __typeof__(_type) _name##_early_map[] + #define early_per_cpu_ptr(_name) (_name##_early_ptr) #define early_per_cpu_map(_name, _idx) (_name##_early_map[_idx]) #define early_per_cpu(_name, _cpu) \ @@ -376,12 +571,18 @@ DECLARE_PER_CPU(unsigned long, this_cpu_off); #define DEFINE_EARLY_PER_CPU(_type, _name, _initvalue) \ DEFINE_PER_CPU(_type, _name) = _initvalue +#define DEFINE_EARLY_PER_CPU_READ_MOSTLY(_type, _name, _initvalue) \ + DEFINE_PER_CPU_READ_MOSTLY(_type, _name) = _initvalue + #define EXPORT_EARLY_PER_CPU_SYMBOL(_name) \ EXPORT_PER_CPU_SYMBOL(_name) #define DECLARE_EARLY_PER_CPU(_type, _name) \ DECLARE_PER_CPU(_type, _name) +#define DECLARE_EARLY_PER_CPU_READ_MOSTLY(_type, _name) \ + DECLARE_PER_CPU_READ_MOSTLY(_type, _name) + #define early_per_cpu(_name, _cpu) per_cpu(_name, _cpu) #define early_per_cpu_ptr(_name) NULL /* no early_per_cpu_map() */ |