diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2024-05-13 19:16:02 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2024-05-13 19:16:02 -0700 |
commit | e76f69b91e1ad11e9df68ba6ca4b4ad2bf39915c (patch) | |
tree | 1efdf68d60b4336b233b9bd570d73beed72d2f50 /arch/x86/include/asm | |
parent | eabb6297abd3115220c8072e463780efc549bc97 (diff) | |
parent | 93cfa544cf9e4771def159002304a2e366cd97af (diff) |
Merge tag 'x86-percpu-2024-05-13' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 percpu updates from Ingo Molnar:
- Expand the named address spaces optimizations down to
GCC 9.1+.
- Re-enable named address spaces with sanitizers for GCC 13.3+
- Generate better this_percpu_xchg_op() code
- Introduce raw_cpu_read_long() to reduce ifdeffery
- Simplify the x86_this_cpu_test_bit() et al macros
- Address Sparse warnings
- Misc cleanups & fixes
* tag 'x86-percpu-2024-05-13' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/percpu: Introduce raw_cpu_read_long() to reduce ifdeffery
x86/percpu: Rewrite x86_this_cpu_test_bit() and friends as macros
x86/percpu: Fix x86_this_cpu_variable_test_bit() asm template
x86/percpu: Re-enable named address spaces with sanitizers for GCC 13.3+
x86/percpu: Use __force to cast from __percpu address space
x86/percpu: Do not use this_cpu_read_stable_8() for 32-bit targets
x86/percpu: Unify arch_raw_cpu_ptr() defines
x86/percpu: Enable named address spaces for GCC 9.1+
x86/percpu: Re-enable named address spaces with KASAN for GCC 13.3+
x86/percpu: Move raw_percpu_xchg_op() to a better place
x86/percpu: Convert this_percpu_xchg_op() from asm() to C code, to generate better code
Diffstat (limited to 'arch/x86/include/asm')
-rw-r--r-- | arch/x86/include/asm/cpufeature.h | 3 | ||||
-rw-r--r-- | arch/x86/include/asm/percpu.h | 157 |
2 files changed, 69 insertions, 91 deletions
diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 0325424c87f3..b96f9cabea0c 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -129,8 +129,7 @@ extern const char * const x86_bug_flags[NBUGINTS*32]; #define this_cpu_has(bit) \ (__builtin_constant_p(bit) && REQUIRED_MASK_BIT_SET(bit) ? 1 : \ - x86_this_cpu_test_bit(bit, \ - (unsigned long __percpu *)&cpu_info.x86_capability)) + x86_this_cpu_test_bit(bit, cpu_info.x86_capability)) /* * This macro is for detection of features which need kernel diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h index 44958ebaf626..3bedee1801e2 100644 --- a/arch/x86/include/asm/percpu.h +++ b/arch/x86/include/asm/percpu.h @@ -59,36 +59,24 @@ #define __force_percpu_prefix "%%"__stringify(__percpu_seg)":" #define __my_cpu_offset this_cpu_read(this_cpu_off) -#ifdef CONFIG_USE_X86_SEG_SUPPORT -/* - * Efficient implementation for cases in which the compiler supports - * named address spaces. Allows the compiler to perform additional - * optimizations that can save more instructions. - */ -#define arch_raw_cpu_ptr(ptr) \ -({ \ - unsigned long tcp_ptr__; \ - tcp_ptr__ = __raw_cpu_read(, this_cpu_off); \ - \ - tcp_ptr__ += (unsigned long)(ptr); \ - (typeof(*(ptr)) __kernel __force *)tcp_ptr__; \ -}) -#else /* CONFIG_USE_X86_SEG_SUPPORT */ /* * Compared to the generic __my_cpu_offset version, the following * saves one instruction and avoids clobbering a temp register. + * + * arch_raw_cpu_ptr should not be used in 32-bit VDSO for a 64-bit + * kernel, because games are played with CONFIG_X86_64 there and + * sizeof(this_cpu_off) becames 4. */ -#define arch_raw_cpu_ptr(ptr) \ +#ifndef BUILD_VDSO32_64 +#define arch_raw_cpu_ptr(_ptr) \ ({ \ - unsigned long tcp_ptr__; \ - asm ("mov " __percpu_arg(1) ", %0" \ - : "=r" (tcp_ptr__) \ - : "m" (__my_cpu_var(this_cpu_off))); \ - \ - tcp_ptr__ += (unsigned long)(ptr); \ - (typeof(*(ptr)) __kernel __force *)tcp_ptr__; \ + unsigned long tcp_ptr__ = raw_cpu_read_long(this_cpu_off); \ + tcp_ptr__ += (__force unsigned long)(_ptr); \ + (typeof(*(_ptr)) __kernel __force *)tcp_ptr__; \ }) -#endif /* CONFIG_USE_X86_SEG_SUPPORT */ +#else +#define arch_raw_cpu_ptr(_ptr) ({ BUILD_BUG(); (typeof(_ptr))0; }) +#endif #define PER_CPU_VAR(var) %__percpu_seg:(var)__percpu_rel @@ -102,8 +90,8 @@ #endif /* CONFIG_SMP */ #define __my_cpu_type(var) typeof(var) __percpu_seg_override -#define __my_cpu_ptr(ptr) (__my_cpu_type(*ptr) *)(uintptr_t)(ptr) -#define __my_cpu_var(var) (*__my_cpu_ptr(&var)) +#define __my_cpu_ptr(ptr) (__my_cpu_type(*(ptr))*)(__force uintptr_t)(ptr) +#define __my_cpu_var(var) (*__my_cpu_ptr(&(var))) #define __percpu_arg(x) __percpu_prefix "%" #x #define __force_percpu_arg(x) __force_percpu_prefix "%" #x @@ -230,25 +218,26 @@ do { \ }) /* - * 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. + * raw_cpu_xchg() can use a load-store since + * it is not required to be IRQ-safe. */ -#define percpu_xchg_op(size, qual, _var, _nval) \ +#define raw_percpu_xchg_op(_var, _nval) \ ({ \ - __pcpu_type_##size pxo_old__; \ - __pcpu_type_##size pxo_new__ = __pcpu_cast_##size(_nval); \ - asm qual (__pcpu_op2_##size("mov", __percpu_arg([var]), \ - "%[oval]") \ - "\n1:\t" \ - __pcpu_op2_##size("cmpxchg", "%[nval]", \ - __percpu_arg([var])) \ - "\n\tjnz 1b" \ - : [oval] "=&a" (pxo_old__), \ - [var] "+m" (__my_cpu_var(_var)) \ - : [nval] __pcpu_reg_##size(, pxo_new__) \ - : "memory"); \ - (typeof(_var))(unsigned long) pxo_old__; \ + typeof(_var) pxo_old__ = raw_cpu_read(_var); \ + raw_cpu_write(_var, _nval); \ + pxo_old__; \ +}) + +/* + * this_cpu_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 this_percpu_xchg_op(_var, _nval) \ +({ \ + typeof(_var) pxo_old__ = this_cpu_read(_var); \ + do { } while (!this_cpu_try_cmpxchg(_var, &pxo_old__, _nval)); \ + pxo_old__; \ }) /* @@ -428,10 +417,6 @@ do { \ * actually per-thread variables implemented as per-CPU variables and * thus stable for the duration of the respective task. */ -#define this_cpu_read_stable_1(pcp) percpu_stable_op(1, "mov", pcp) -#define this_cpu_read_stable_2(pcp) percpu_stable_op(2, "mov", pcp) -#define this_cpu_read_stable_4(pcp) percpu_stable_op(4, "mov", pcp) -#define this_cpu_read_stable_8(pcp) percpu_stable_op(8, "mov", pcp) #define this_cpu_read_stable(pcp) __pcpu_size_call_return(this_cpu_read_stable_, pcp) #ifdef CONFIG_USE_X86_SEG_SUPPORT @@ -500,6 +485,10 @@ do { \ #define this_cpu_read_const(pcp) ({ BUILD_BUG(); (typeof(pcp))0; }) #endif /* CONFIG_USE_X86_SEG_SUPPORT */ +#define this_cpu_read_stable_1(pcp) percpu_stable_op(1, "mov", pcp) +#define this_cpu_read_stable_2(pcp) percpu_stable_op(2, "mov", pcp) +#define this_cpu_read_stable_4(pcp) percpu_stable_op(4, "mov", pcp) + #define raw_cpu_add_1(pcp, val) percpu_add_op(1, , (pcp), val) #define raw_cpu_add_2(pcp, val) percpu_add_op(2, , (pcp), val) #define raw_cpu_add_4(pcp, val) percpu_add_op(4, , (pcp), val) @@ -509,18 +498,6 @@ do { \ #define raw_cpu_or_1(pcp, val) percpu_to_op(1, , "or", (pcp), val) #define raw_cpu_or_2(pcp, val) percpu_to_op(2, , "or", (pcp), val) #define raw_cpu_or_4(pcp, val) percpu_to_op(4, , "or", (pcp), val) - -/* - * raw_cpu_xchg() can use a load-store since it is not required to be - * IRQ-safe. - */ -#define raw_percpu_xchg_op(var, nval) \ -({ \ - typeof(var) pxo_ret__ = raw_cpu_read(var); \ - raw_cpu_write(var, (nval)); \ - pxo_ret__; \ -}) - #define raw_cpu_xchg_1(pcp, val) raw_percpu_xchg_op(pcp, val) #define raw_cpu_xchg_2(pcp, val) raw_percpu_xchg_op(pcp, val) #define raw_cpu_xchg_4(pcp, val) raw_percpu_xchg_op(pcp, val) @@ -534,9 +511,9 @@ do { \ #define this_cpu_or_1(pcp, val) percpu_to_op(1, volatile, "or", (pcp), val) #define this_cpu_or_2(pcp, val) percpu_to_op(2, volatile, "or", (pcp), val) #define this_cpu_or_4(pcp, val) percpu_to_op(4, volatile, "or", (pcp), val) -#define this_cpu_xchg_1(pcp, nval) percpu_xchg_op(1, volatile, pcp, nval) -#define this_cpu_xchg_2(pcp, nval) percpu_xchg_op(2, volatile, pcp, nval) -#define this_cpu_xchg_4(pcp, nval) percpu_xchg_op(4, volatile, pcp, nval) +#define this_cpu_xchg_1(pcp, nval) this_percpu_xchg_op(pcp, nval) +#define this_cpu_xchg_2(pcp, nval) this_percpu_xchg_op(pcp, nval) +#define this_cpu_xchg_4(pcp, nval) this_percpu_xchg_op(pcp, nval) #define raw_cpu_add_return_1(pcp, val) percpu_add_return_op(1, , pcp, val) #define raw_cpu_add_return_2(pcp, val) percpu_add_return_op(2, , pcp, val) @@ -563,6 +540,8 @@ do { \ * 32 bit must fall back to generic operations. */ #ifdef CONFIG_X86_64 +#define this_cpu_read_stable_8(pcp) percpu_stable_op(8, "mov", pcp) + #define raw_cpu_add_8(pcp, val) percpu_add_op(8, , (pcp), val) #define raw_cpu_and_8(pcp, val) percpu_to_op(8, , "and", (pcp), val) #define raw_cpu_or_8(pcp, val) percpu_to_op(8, , "or", (pcp), val) @@ -575,41 +554,41 @@ do { \ #define this_cpu_and_8(pcp, val) percpu_to_op(8, volatile, "and", (pcp), val) #define this_cpu_or_8(pcp, val) percpu_to_op(8, volatile, "or", (pcp), val) #define this_cpu_add_return_8(pcp, val) percpu_add_return_op(8, volatile, pcp, val) -#define this_cpu_xchg_8(pcp, nval) percpu_xchg_op(8, volatile, pcp, nval) +#define this_cpu_xchg_8(pcp, nval) this_percpu_xchg_op(pcp, nval) #define this_cpu_cmpxchg_8(pcp, oval, nval) percpu_cmpxchg_op(8, volatile, pcp, oval, nval) #define this_cpu_try_cmpxchg_8(pcp, ovalp, nval) percpu_try_cmpxchg_op(8, volatile, pcp, ovalp, nval) -#endif - -static __always_inline bool x86_this_cpu_constant_test_bit(unsigned int nr, - const unsigned long __percpu *addr) -{ - unsigned long __percpu *a = - (unsigned long __percpu *)addr + nr / BITS_PER_LONG; -#ifdef CONFIG_X86_64 - return ((1UL << (nr % BITS_PER_LONG)) & raw_cpu_read_8(*a)) != 0; +#define raw_cpu_read_long(pcp) raw_cpu_read_8(pcp) #else - return ((1UL << (nr % BITS_PER_LONG)) & raw_cpu_read_4(*a)) != 0; -#endif -} +/* There is no generic 64 bit read stable operation for 32 bit targets. */ +#define this_cpu_read_stable_8(pcp) ({ BUILD_BUG(); (typeof(pcp))0; }) -static inline bool x86_this_cpu_variable_test_bit(int nr, - const unsigned long __percpu *addr) -{ - bool oldbit; +#define raw_cpu_read_long(pcp) raw_cpu_read_4(pcp) +#endif - asm volatile("btl "__percpu_arg(2)",%1" - CC_SET(c) - : CC_OUT(c) (oldbit) - : "m" (*__my_cpu_ptr((unsigned long __percpu *)(addr))), "Ir" (nr)); +#define x86_this_cpu_constant_test_bit(_nr, _var) \ +({ \ + unsigned long __percpu *addr__ = \ + (unsigned long __percpu *)&(_var) + ((_nr) / BITS_PER_LONG); \ + !!((1UL << ((_nr) % BITS_PER_LONG)) & raw_cpu_read(*addr__)); \ +}) - return oldbit; -} +#define x86_this_cpu_variable_test_bit(_nr, _var) \ +({ \ + bool oldbit; \ + \ + asm volatile("btl %[nr], " __percpu_arg([var]) \ + CC_SET(c) \ + : CC_OUT(c) (oldbit) \ + : [var] "m" (__my_cpu_var(_var)), \ + [nr] "rI" (_nr)); \ + 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))) +#define x86_this_cpu_test_bit(_nr, _var) \ + (__builtin_constant_p(_nr) \ + ? x86_this_cpu_constant_test_bit(_nr, _var) \ + : x86_this_cpu_variable_test_bit(_nr, _var)) #include <asm-generic/percpu.h> |