From: Mathieu Desnoyers Date: Sun, 30 Aug 2009 14:50:04 +0000 (-0400) Subject: Move test programs to tests/ subdir X-Git-Tag: v0.1~101 X-Git-Url: https://git.lttng.org./?a=commitdiff_plain;h=833dbdb633ba44bfb71b4b965a83096d8e4e827b;p=urcu.git Move test programs to tests/ subdir Signed-off-by: Mathieu Desnoyers --- diff --git a/Makefile b/Makefile index 98f9c4a..ab19097 100644 --- a/Makefile +++ b/Makefile @@ -1,170 +1 @@ - -CFLAGS=-Wall -O2 -g -I. -LDFLAGS=-lpthread - -HOSTTYPE=$(shell uname -m) - -ifeq ("${HOSTTYPE}","x86_64") -ARCHTYPE=x86 -endif -ifeq ("${HOSTTYPE}","i586") -ARCHTYPE=x86 -endif -ifeq ("${HOSTTYPE}","i686") -ARCHTYPE=x86 -endif -ifeq ("${HOSTTYPE}","powerpc") -ARCHTYPE=ppc -endif -ifeq ("${HOSTTYPE}","ppc64") -ARCHTYPE=ppc -endif -ifeq ("${HOSTTYPE}","s390") -ARCHTYPE=s390 -endif -ifeq ("${HOSTTYPE}","s390x") -ARCHTYPE=s390 -endif - -ifeq ($(findstring ${ARCHTYPE},"x86 ppc"),) -APIHEADER=api_gcc.h -else -APIHEADER=api_${ARCHTYPE}.h -endif - -#debug -#CFLAGS=-Wall -g - -#Build the library without using signals. Slower read-side. -#CFLAGS+=-DCONFIG_URCU_AVOID_SIGNALS - -#Changing the signal number used by the library. SIGUSR1 by default. -#CFLAGS+=-DSIGURCU=SIGUSR2 - -SRC_DEP=`echo $^ | sed 's/[^ ]*\.h//g'` - -all: checkarch test_urcu test_urcu_dynamic_link test_urcu_timing \ - test_rwlock_timing test_rwlock test_perthreadlock_timing \ - test_perthreadlock test_urcu_yield test_urcu_mb \ - urcu-asm.S test_qsbr_timing test_qsbr urcu-asm.o urcutorture \ - urcutorture-yield liburcu.so test_mutex test_looplen test_urcu_gc \ - test_urcu_gc_mb test_qsbr_gc test_qsbr_lgc test_urcu_lgc \ - test_urcu_lgc_mb - -checkarch: -ifeq (${ARCHTYPE},) - @echo "Architecture ${HOSTTYPE} is currently unsupported by liburcu" - @exit 1 -endif - -arch.h: arch_${ARCHTYPE}.h - cp -f arch_${ARCHTYPE}.h arch.h - -api.h: ${APIHEADER} - cp -f ${APIHEADER} api.h - -arch_atomic.h: arch_atomic_${ARCHTYPE}.h - cp -f arch_atomic_${ARCHTYPE}.h arch_atomic.h - -urcu.h: arch.h arch_atomic.h - -urcu-qsbr.h: arch.h arch_atomic.h - -test_urcu: urcu.o test_urcu.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_looplen: test_looplen.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_mb: urcu-mb.o test_urcu.c urcu.h - $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_gc: urcu.o test_urcu_gc.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_gc_mb: urcu-mb.o test_urcu_gc.c urcu.h - $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_lgc: urcu.o test_urcu_gc.c urcu.h - $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_lgc_mb: urcu-mb.o test_urcu_gc.c urcu.h - $(CC) -DTEST_LOCAL_GC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr: urcu-qsbr.o test_qsbr.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_gc: urcu-qsbr.o test_qsbr_gc.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_lgc: urcu-qsbr.o test_qsbr_gc.c urcu-qsbr.h - $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_rwlock: urcu.o test_rwlock.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_perthreadlock: urcu.o test_perthreadlock.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_mutex: urcu.o test_mutex.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_dynamic_link: urcu.o test_urcu.c urcu.h - $(CC) ${CFLAGS} -DDYNAMIC_LINK_TEST $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_yield: urcu-yield.o test_urcu.c urcu.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_timing: urcu.o test_urcu_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_timing: urcu-qsbr.o test_qsbr_timing.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_rwlock_timing: urcu.o test_rwlock_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_perthreadlock_timing: urcu.o test_perthreadlock_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -urcu.o: urcu.c urcu.h - $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-mb.o: urcu.c urcu.h - $(CC) -fPIC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-qsbr.o: urcu-qsbr.c urcu-qsbr.h - $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -liburcu.so: urcu.o - $(CC) -fPIC -shared -o $@ $< - -urcu-yield.o: urcu.c urcu.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-asm.S: urcu-asm.c urcu.h - $(CC) ${CFLAGS} -S -o $@ $(SRC_DEP) - -urcu-asm.o: urcu-asm.c urcu.h - $(CC) ${CFLAGS} -c -o $@ $(SRC_DEP) - -urcutorture: urcutorture.c urcu.o urcu.h rcutorture.h api.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -urcutorture-yield: urcutorture.c urcu-yield.o urcu.h rcutorture.h api.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -.PHONY: clean install checkarch - -install: liburcu.so - cp -f liburcu.so /usr/lib/ - cp -f arch.h arch_atomic.h compiler.h urcu.h urcu-static.h /usr/include/ - -clean: - rm -f *.o test_urcu test_urcu_dynamic_link test_urcu_timing \ - test_rwlock_timing test_rwlock test_perthreadlock_timing \ - test_perthreadlock test_urcu_yield test_urcu_mb \ - urcu-asm.S test_qsbr_timing test_qsbr urcutorture \ - urcutorture-yield liburcu.so api.h arch.h arch_atomic.h \ - test_mutex test_urcu_gc test_urcu_gc_mb - +include Makefile.inc diff --git a/Makefile.inc b/Makefile.inc new file mode 100644 index 0000000..870bde8 --- /dev/null +++ b/Makefile.inc @@ -0,0 +1,50 @@ + +include Makefile.build.inc + +DIRS=tests + +all: checkarch liburcu.so urcu.o urcu-yield.o urcu-qsbr.o urcu-mb.o subdirs + +checkarch: +ifeq (${ARCHTYPE},) + @echo "Architecture ${HOSTTYPE} is currently unsupported by liburcu" + @exit 1 +endif + +arch.h: arch_${ARCHTYPE}.h + cp -f arch_${ARCHTYPE}.h arch.h + +arch_atomic.h: arch_atomic_${ARCHTYPE}.h + cp -f arch_atomic_${ARCHTYPE}.h arch_atomic.h + +urcu.h: arch.h arch_atomic.h + +urcu-qsbr.h: arch.h arch_atomic.h + +urcu.o: urcu.c urcu.h + $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) + +urcu-mb.o: urcu.c urcu.h + $(CC) -fPIC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) + +urcu-qsbr.o: urcu-qsbr.c urcu-qsbr.h + $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) + +liburcu.so: urcu.o + $(CC) -fPIC -shared -o $@ $< + +urcu-yield.o: urcu.c urcu.h + $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) + +.PHONY: clean install checkarch tests + +subdirs: + -for d in ${DIRS}; do cd $${d}; ${MAKE} ${MFLAGS}; done + +install: liburcu.so + cp -f liburcu.so /usr/lib/ + cp -f arch.h arch_atomic.h compiler.h urcu.h urcu-static.h /usr/include/ + +clean: + rm -f *.o *.so arch.h arch_atomic.h + -for d in ${DIRS}; do cd $${d}; ${MAKE} clean; done diff --git a/Makefile64 b/Makefile64 index ba36fec..cca1cdc 100644 --- a/Makefile64 +++ b/Makefile64 @@ -1,153 +1,3 @@ +include Makefile.inc -#CFLAGS=-m64 -Wall -O2 -g -I. -CFLAGS=-m64 -O2 -g -I. -LDFLAGS=-lpthread - -HOSTTYPE=$(shell uname -m) - -ifeq ("${HOSTTYPE}","x86_64") -ARCHTYPE=x86 -endif -ifeq ("${HOSTTYPE}","powerpc") -ARCHTYPE=ppc -endif -ifeq ("${HOSTTYPE}","ppc64") -ARCHTYPE=ppc -endif - -#debug -#CFLAGS=-m64 -Wall -g - -#Build the library without using signals. Slower read-side. -#CFLAGS+=-DCONFIG_URCU_AVOID_SIGNALS - -#Changing the signal number used by the library. SIGUSR1 by default. -#CFLAGS+=-DSIGURCU=SIGUSR2 - -SRC_DEP=`echo $^ | sed 's/[^ ]*\.h//g'` - -all: checkarch test_urcu test_urcu_dynamic_link test_urcu_timing \ - test_rwlock_timing test_rwlock test_perthreadlock_timing \ - test_perthreadlock test_urcu_yield test_urcu_mb \ - urcu-asm.S test_qsbr_timing test_qsbr urcu-asm.o urcutorture \ - urcutorture-yield liburcu.so test_mutex test_looplen test_urcu_gc \ - test_urcu_gc_mb test_qsbr_gc test_qsbr_lgc test_urcu_lgc \ - test_urcu_lgc_mb - -checkarch: -ifeq (${ARCHTYPE},) - @echo "Architecture ${HOSTTYPE} is currently unsupported by liburcu" - @exit 1 -endif - -arch.h: arch_${ARCHTYPE}.h - cp -f arch_${ARCHTYPE}.h arch.h - -api.h: api_${ARCHTYPE}.h - cp -f api_${ARCHTYPE}.h api.h - -arch_atomic.h: arch_atomic_${ARCHTYPE}.h - cp -f arch_atomic_${ARCHTYPE}.h arch_atomic.h - -urcu.h: arch.h api.h arch_atomic.h - -urcu-qsbr.h: arch.h api.h arch_atomic.h - -test_urcu: urcu.o test_urcu.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_looplen: test_looplen.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_mb: urcu-mb.o test_urcu.c urcu.h - $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_gc: urcu.o test_urcu_gc.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_gc_mb: urcu-mb.o test_urcu_gc.c urcu.h - $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_lgc: urcu.o test_urcu_gc.c urcu.h - $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_lgc_mb: urcu-mb.o test_urcu_gc.c urcu.h - $(CC) -DTEST_LOCAL_GC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr: urcu-qsbr.o test_qsbr.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_gc: urcu-qsbr.o test_qsbr_gc.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_lgc: urcu-qsbr.o test_qsbr_gc.c urcu-qsbr.h - $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_rwlock: urcu.o test_rwlock.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_perthreadlock: urcu.o test_perthreadlock.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_mutex: urcu.o test_mutex.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_dynamic_link: urcu.o test_urcu.c urcu.h - $(CC) ${CFLAGS} -DDYNAMIC_LINK_TEST $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_yield: urcu-yield.o test_urcu.c urcu.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_urcu_timing: urcu.o test_urcu_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_qsbr_timing: urcu-qsbr.o test_qsbr_timing.c urcu-qsbr.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_rwlock_timing: urcu.o test_rwlock_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -test_perthreadlock_timing: urcu.o test_perthreadlock_timing.c urcu.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -urcu.o: urcu.c urcu.h - $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-mb.o: urcu.c urcu.h - $(CC) -fPIC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-qsbr.o: urcu-qsbr.c urcu-qsbr.h - $(CC) -fPIC ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -liburcu.so: urcu.o - $(CC) -m64 -fPIC -shared -o $@ $< - -urcu-yield.o: urcu.c urcu.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -c -o $@ $(SRC_DEP) - -urcu-asm.S: urcu-asm.c urcu.h - $(CC) ${CFLAGS} -S -o $@ $(SRC_DEP) - -urcu-asm.o: urcu-asm.c urcu.h - $(CC) ${CFLAGS} -c -o $@ $(SRC_DEP) - -urcutorture: urcutorture.c urcu.o urcu.h rcutorture.h - $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -urcutorture-yield: urcutorture.c urcu-yield.o urcu.h rcutorture.h - $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) - -.PHONY: clean install checkarch - -install: liburcu.so - cp -f liburcu.so /usr/lib/ - cp -f arch.h arch_atomic.h compiler.h urcu.h urcu-static.h /usr/include/ - -clean: - rm -f *.o test_urcu test_urcu_dynamic_link test_urcu_timing \ - test_rwlock_timing test_rwlock test_perthreadlock_timing \ - test_perthreadlock test_urcu_yield test_urcu_mb \ - urcu-asm.S test_qsbr_timing test_qsbr urcutorture \ - urcutorture-yield liburcu.so api.h arch.h arch_atomic.h \ - test_mutex test_urcu_gc test_urcu_gc_mb - +CFLAGS+=-m64 diff --git a/api_gcc.h b/api_gcc.h deleted file mode 100644 index 1ad7345..0000000 --- a/api_gcc.h +++ /dev/null @@ -1,1328 +0,0 @@ -#define _INCLUDE_API_H - -/* - * common.h: Common Linux kernel-isms. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; but version 2 of the License only due - * to code included from the Linux kernel. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -#ifndef __always_inline -#define __always_inline inline -#endif - -#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) -#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) - -#ifdef __ASSEMBLY__ -# define stringify_in_c(...) __VA_ARGS__ -# define ASM_CONST(x) x -#else -/* This version of stringify will deal with commas... */ -# define __stringify_in_c(...) #__VA_ARGS__ -# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " -# define __ASM_CONST(x) x##UL -# define ASM_CONST(x) __ASM_CONST(x) -#endif - - -/* - * arch-i386.h: Expose x86 atomic instructions. 80486 and better only. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation, but version 2 only due to inclusion - * of Linux-kernel code. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -/* - * Machine parameters. - */ - -#define CACHE_LINE_SIZE 64 -#define ____cacheline_internodealigned_in_smp \ - __attribute__((__aligned__(1 << 6))) - -#define LOCK_PREFIX "lock ; " - -/* - * Atomic data structure, initialization, and access. - */ - -typedef struct { volatile int counter; } atomic_t; - -#define ATOMIC_INIT(i) { (i) } - -#define atomic_read(v) ((v)->counter) -#define atomic_set(v, i) (((v)->counter) = (i)) - -/* - * Atomic operations. - */ - -/** - * atomic_add - add integer to atomic variable - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v. - */ - -static __inline__ void atomic_add(int i, atomic_t *v) -{ - (void)__sync_fetch_and_add(&v->counter, i); -} - -/** - * atomic_sub - subtract the atomic variable - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v. - */ -static __inline__ void atomic_sub(int i, atomic_t *v) -{ - (void)__sync_fetch_and_add(&v->counter, -i); -} - -/** - * atomic_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_sub_and_test(int i, atomic_t *v) -{ - return __sync_add_and_fetch(&v->counter, -i) == 0; -} - -/** - * atomic_inc - increment atomic variable - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1. - */ -static __inline__ void atomic_inc(atomic_t *v) -{ - (void)__sync_fetch_and_add(&v->counter, 1); -} - -/** - * atomic_dec - decrement atomic variable - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1. - */ -static __inline__ void atomic_dec(atomic_t *v) -{ - (void)__sync_fetch_and_add(&v->counter, -1); -} - -/** - * atomic_dec_and_test - decrement and test - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static __inline__ int atomic_dec_and_test(atomic_t *v) -{ - return __sync_add_and_fetch(&v->counter, -1) == 0; -} - -/** - * atomic_inc_and_test - increment and test - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_inc_and_test(atomic_t *v) -{ - return __sync_add_and_fetch(&v->counter, 1) == 0; -} - -/** - * atomic_add_negative - add and test if negative - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static __inline__ int atomic_add_negative(int i, atomic_t *v) -{ - return __sync_add_and_fetch(&v->counter, i) < 0; -} - -/** - * atomic_add_return - add and return - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns @i + @v - */ -static __inline__ int atomic_add_return(int i, atomic_t *v) -{ - return __sync_add_and_fetch(&v->counter, i); -} - -static __inline__ int atomic_sub_return(int i, atomic_t *v) -{ - return atomic_add_return(-i,v); -} - -static inline unsigned int -cmpxchg(volatile long *ptr, long oldval, long newval) -{ - return __sync_val_compare_and_swap(ptr, oldval, newval); -} - -#define atomic_cmpxchg(v, old, new) ((int)cmpxchg(&((v)->counter), old, new)) -#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) - -/** - * atomic_add_unless - add unless the number is a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as it was not @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -#define atomic_add_unless(v, a, u) \ -({ \ - int c, old; \ - c = atomic_read(v); \ - for (;;) { \ - if (unlikely(c == (u))) \ - break; \ - old = atomic_cmpxchg((v), c, c + (a)); \ - if (likely(old == c)) \ - break; \ - c = old; \ - } \ - c != (u); \ -}) -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) - -#define atomic_inc_return(v) (atomic_add_return(1,v)) -#define atomic_dec_return(v) (atomic_sub_return(1,v)) - -/* Atomic operations are already serializing on x86 */ -#define smp_mb__before_atomic_dec() barrier() -#define smp_mb__after_atomic_dec() barrier() -#define smp_mb__before_atomic_inc() barrier() -#define smp_mb__after_atomic_inc() barrier() - -/* - * api_pthreads.h: API mapping to pthreads environment. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. However, please note that much - * of the code in this file derives from the Linux kernel, and that such - * code may not be available except under GPLv2. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - */ - -#include -#include -#include -#include -#include -#define __USE_GNU -#include -#include -#include -/* #include "atomic.h" */ - -/* - * Compiler magic. - */ -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Default machine parameters. - */ - -#ifndef CACHE_LINE_SIZE -#define CACHE_LINE_SIZE 128 -#endif /* #ifndef CACHE_LINE_SIZE */ - -/* - * Exclusive locking primitives. - */ - -typedef pthread_mutex_t spinlock_t; - -#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; -#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER - -static void spin_lock_init(spinlock_t *sp) -{ - if (pthread_mutex_init(sp, NULL) != 0) { - perror("spin_lock_init:pthread_mutex_init"); - exit(-1); - } -} - -static void spin_lock(spinlock_t *sp) -{ - if (pthread_mutex_lock(sp) != 0) { - perror("spin_lock:pthread_mutex_lock"); - exit(-1); - } -} - -static void spin_unlock(spinlock_t *sp) -{ - if (pthread_mutex_unlock(sp) != 0) { - perror("spin_unlock:pthread_mutex_unlock"); - exit(-1); - } -} - -#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) -#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) - -/* - * Thread creation/destruction primitives. - */ - -typedef pthread_t thread_id_t; - -#define NR_THREADS 128 - -#define __THREAD_ID_MAP_EMPTY 0 -#define __THREAD_ID_MAP_WAITING 1 -thread_id_t __thread_id_map[NR_THREADS]; -spinlock_t __thread_id_map_mutex; - -#define for_each_thread(t) \ - for (t = 0; t < NR_THREADS; t++) - -#define for_each_running_thread(t) \ - for (t = 0; t < NR_THREADS; t++) \ - if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ - (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) - -pthread_key_t thread_id_key; - -static int __smp_thread_id(void) -{ - int i; - thread_id_t tid = pthread_self(); - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) { - long v = i + 1; /* must be non-NULL. */ - - if (pthread_setspecific(thread_id_key, (void *)v) != 0) { - perror("pthread_setspecific"); - exit(-1); - } - return i; - } - } - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - spin_unlock(&__thread_id_map_mutex); - return i; - } - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); -} - -static int smp_thread_id(void) -{ - void *id; - - id = pthread_getspecific(thread_id_key); - if (id == NULL) - return __smp_thread_id(); - return (long)(id - 1); -} - -static thread_id_t create_thread(void *(*func)(void *), void *arg) -{ - thread_id_t tid; - int i; - - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) - break; - } - if (i >= NR_THREADS) { - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_WAITING; - spin_unlock(&__thread_id_map_mutex); - if (pthread_create(&tid, NULL, func, arg) != 0) { - perror("create_thread:pthread_create"); - exit(-1); - } - __thread_id_map[i] = tid; - return tid; -} - -static void *wait_thread(thread_id_t tid) -{ - int i; - void *vp; - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - break; - } - if (i >= NR_THREADS){ - fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); - } - if (pthread_join(tid, &vp) != 0) { - perror("wait_thread:pthread_join"); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - return vp; -} - -static void wait_all_threads(void) -{ - int i; - thread_id_t tid; - - for (i = 1; i < NR_THREADS; i++) { - tid = __thread_id_map[i]; - if (tid != __THREAD_ID_MAP_EMPTY && - tid != __THREAD_ID_MAP_WAITING) - (void)wait_thread(tid); - } -} - -static void run_on(int cpu) -{ - cpu_set_t mask; - - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * timekeeping -- very crude -- should use MONOTONIC... - */ - -long long get_microseconds(void) -{ - struct timeval tv; - - if (gettimeofday(&tv, NULL) != 0) - abort(); - return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; -} - -/* - * Per-thread variables. - */ - -#define DEFINE_PER_THREAD(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_thread_##name[NR_THREADS]; -#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) - -#define per_thread(name, thread) __per_thread_##name[thread].v -#define __get_thread_var(name) per_thread(name, smp_thread_id()) - -#define init_per_thread(name, v) \ - do { \ - int __i_p_t_i; \ - for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ - per_thread(name, __i_p_t_i) = v; \ - } while (0) - -/* - * CPU traversal primitives. - */ - -#ifndef NR_CPUS -#define NR_CPUS 16 -#endif /* #ifndef NR_CPUS */ - -#define for_each_possible_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) -#define for_each_online_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) - -/* - * Per-CPU variables. - */ - -#define DEFINE_PER_CPU(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_cpu_##name[NR_CPUS] -#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) - -DEFINE_PER_THREAD(int, smp_processor_id); - -#define per_cpu(name, thread) __per_cpu_##name[thread].v -#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) - -#define init_per_cpu(name, v) \ - do { \ - int __i_p_c_i; \ - for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ - per_cpu(name, __i_p_c_i) = v; \ - } while (0) - -/* - * CPU state checking (crowbarred). - */ - -#define idle_cpu(cpu) 0 -#define in_softirq() 1 -#define hardirq_count() 0 -#define PREEMPT_SHIFT 0 -#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) -#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) -#define PREEMPT_BITS 8 -#define SOFTIRQ_BITS 8 - -/* - * CPU hotplug. - */ - -struct notifier_block { - int (*notifier_call)(struct notifier_block *, unsigned long, void *); - struct notifier_block *next; - int priority; -}; - -#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ -#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ -#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ -#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ -#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ -#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ -#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, - * not handling interrupts, soon dead */ -#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug - * lock is dropped */ - -/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend - * operation in progress - */ -#define CPU_TASKS_FROZEN 0x0010 - -#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) -#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) -#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) -#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) -#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) -#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) -#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) - -/* Hibernation and suspend events */ -#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ -#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ -#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ -#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ -#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ -#define PM_POST_RESTORE 0x0006 /* Restore failed */ - -#define NOTIFY_DONE 0x0000 /* Don't care */ -#define NOTIFY_OK 0x0001 /* Suits me */ -#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ -#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) - /* Bad/Veto action */ -/* - * Clean way to return from the notifier and stop further calls. - */ -#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) - -/* - * Bug checks. - */ - -#define BUG_ON(c) do { if (!(c)) abort(); } while (0) - -/* - * Initialization -- Must be called before calling any primitives. - */ - -static void smp_init(void) -{ - int i; - - spin_lock_init(&__thread_id_map_mutex); - __thread_id_map[0] = pthread_self(); - for (i = 1; i < NR_THREADS; i++) - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - init_per_thread(smp_processor_id, 0); - if (pthread_key_create(&thread_id_key, NULL) != 0) { - perror("pthread_key_create"); - exit(-1); - } -} - -/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ - -#ifndef _LINUX_LIST_H -#define _LINUX_LIST_H - -#define LIST_POISON1 ((void *) 0x00100100) -#define LIST_POISON2 ((void *) 0x00200200) - -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Simple doubly linked list implementation. - * - * Some of the internal functions ("__xxx") are useful when - * manipulating whole lists rather than single entries, as - * sometimes we already know the next/prev entries and we can - * generate better code by using them directly rather than - * using the generic single-entry routines. - */ - -struct list_head { - struct list_head *next, *prev; -}; - -#define LIST_HEAD_INIT(name) { &(name), &(name) } - -#define LIST_HEAD(name) \ - struct list_head name = LIST_HEAD_INIT(name) - -static inline void INIT_LIST_HEAD(struct list_head *list) -{ - list->next = list; - list->prev = list; -} - -/* - * Insert a new entry between two known consecutive entries. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -#ifndef CONFIG_DEBUG_LIST -static inline void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next) -{ - next->prev = new; - new->next = next; - new->prev = prev; - prev->next = new; -} -#else -extern void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next); -#endif - -/** - * list_add - add a new entry - * @new: new entry to be added - * @head: list head to add it after - * - * Insert a new entry after the specified head. - * This is good for implementing stacks. - */ -static inline void list_add(struct list_head *new, struct list_head *head) -{ - __list_add(new, head, head->next); -} - - -/** - * list_add_tail - add a new entry - * @new: new entry to be added - * @head: list head to add it before - * - * Insert a new entry before the specified head. - * This is useful for implementing queues. - */ -static inline void list_add_tail(struct list_head *new, struct list_head *head) -{ - __list_add(new, head->prev, head); -} - -/* - * Delete a list entry by making the prev/next entries - * point to each other. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -static inline void __list_del(struct list_head * prev, struct list_head * next) -{ - next->prev = prev; - prev->next = next; -} - -/** - * list_del - deletes entry from list. - * @entry: the element to delete from the list. - * Note: list_empty() on entry does not return true after this, the entry is - * in an undefined state. - */ -#ifndef CONFIG_DEBUG_LIST -static inline void list_del(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - entry->next = LIST_POISON1; - entry->prev = LIST_POISON2; -} -#else -extern void list_del(struct list_head *entry); -#endif - -/** - * list_replace - replace old entry by new one - * @old : the element to be replaced - * @new : the new element to insert - * - * If @old was empty, it will be overwritten. - */ -static inline void list_replace(struct list_head *old, - struct list_head *new) -{ - new->next = old->next; - new->next->prev = new; - new->prev = old->prev; - new->prev->next = new; -} - -static inline void list_replace_init(struct list_head *old, - struct list_head *new) -{ - list_replace(old, new); - INIT_LIST_HEAD(old); -} - -/** - * list_del_init - deletes entry from list and reinitialize it. - * @entry: the element to delete from the list. - */ -static inline void list_del_init(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - INIT_LIST_HEAD(entry); -} - -/** - * list_move - delete from one list and add as another's head - * @list: the entry to move - * @head: the head that will precede our entry - */ -static inline void list_move(struct list_head *list, struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add(list, head); -} - -/** - * list_move_tail - delete from one list and add as another's tail - * @list: the entry to move - * @head: the head that will follow our entry - */ -static inline void list_move_tail(struct list_head *list, - struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add_tail(list, head); -} - -/** - * list_is_last - tests whether @list is the last entry in list @head - * @list: the entry to test - * @head: the head of the list - */ -static inline int list_is_last(const struct list_head *list, - const struct list_head *head) -{ - return list->next == head; -} - -/** - * list_empty - tests whether a list is empty - * @head: the list to test. - */ -static inline int list_empty(const struct list_head *head) -{ - return head->next == head; -} - -/** - * list_empty_careful - tests whether a list is empty and not being modified - * @head: the list to test - * - * Description: - * tests whether a list is empty _and_ checks that no other CPU might be - * in the process of modifying either member (next or prev) - * - * NOTE: using list_empty_careful() without synchronization - * can only be safe if the only activity that can happen - * to the list entry is list_del_init(). Eg. it cannot be used - * if another CPU could re-list_add() it. - */ -static inline int list_empty_careful(const struct list_head *head) -{ - struct list_head *next = head->next; - return (next == head) && (next == head->prev); -} - -/** - * list_is_singular - tests whether a list has just one entry. - * @head: the list to test. - */ -static inline int list_is_singular(const struct list_head *head) -{ - return !list_empty(head) && (head->next == head->prev); -} - -static inline void __list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - struct list_head *new_first = entry->next; - list->next = head->next; - list->next->prev = list; - list->prev = entry; - entry->next = list; - head->next = new_first; - new_first->prev = head; -} - -/** - * list_cut_position - cut a list into two - * @list: a new list to add all removed entries - * @head: a list with entries - * @entry: an entry within head, could be the head itself - * and if so we won't cut the list - * - * This helper moves the initial part of @head, up to and - * including @entry, from @head to @list. You should - * pass on @entry an element you know is on @head. @list - * should be an empty list or a list you do not care about - * losing its data. - * - */ -static inline void list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - if (list_empty(head)) - return; - if (list_is_singular(head) && - (head->next != entry && head != entry)) - return; - if (entry == head) - INIT_LIST_HEAD(list); - else - __list_cut_position(list, head, entry); -} - -static inline void __list_splice(const struct list_head *list, - struct list_head *prev, - struct list_head *next) -{ - struct list_head *first = list->next; - struct list_head *last = list->prev; - - first->prev = prev; - prev->next = first; - - last->next = next; - next->prev = last; -} - -/** - * list_splice - join two lists, this is designed for stacks - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice(const struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head, head->next); -} - -/** - * list_splice_tail - join two lists, each list being a queue - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice_tail(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head->prev, head); -} - -/** - * list_splice_init - join two lists and reinitialise the emptied list. - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * The list at @list is reinitialised - */ -static inline void list_splice_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head, head->next); - INIT_LIST_HEAD(list); - } -} - -/** - * list_splice_tail_init - join two lists and reinitialise the emptied list - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * Each of the lists is a queue. - * The list at @list is reinitialised - */ -static inline void list_splice_tail_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head->prev, head); - INIT_LIST_HEAD(list); - } -} - -/** - * list_entry - get the struct for this entry - * @ptr: the &struct list_head pointer. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - */ -#define list_entry(ptr, type, member) \ - container_of(ptr, type, member) - -/** - * list_first_entry - get the first element from a list - * @ptr: the list head to take the element from. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - * - * Note, that list is expected to be not empty. - */ -#define list_first_entry(ptr, type, member) \ - list_entry((ptr)->next, type, member) - -/** - * list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each(pos, head) \ - for (pos = (head)->next; prefetch(pos->next), pos != (head); \ - pos = pos->next) - -/** - * __list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - * - * This variant differs from list_for_each() in that it's the - * simplest possible list iteration code, no prefetching is done. - * Use this for code that knows the list to be very short (empty - * or 1 entry) most of the time. - */ -#define __list_for_each(pos, head) \ - for (pos = (head)->next; pos != (head); pos = pos->next) - -/** - * list_for_each_prev - iterate over a list backwards - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each_prev(pos, head) \ - for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ - pos = pos->prev) - -/** - * list_for_each_safe - iterate over a list safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_safe(pos, n, head) \ - for (pos = (head)->next, n = pos->next; pos != (head); \ - pos = n, n = pos->next) - -/** - * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_prev_safe(pos, n, head) \ - for (pos = (head)->prev, n = pos->prev; \ - prefetch(pos->prev), pos != (head); \ - pos = n, n = pos->prev) - -/** - * list_for_each_entry - iterate over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry(pos, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_reverse - iterate backwards over list of given type. - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_reverse(pos, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() - * @pos: the type * to use as a start point - * @head: the head of the list - * @member: the name of the list_struct within the struct. - * - * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). - */ -#define list_prepare_entry(pos, head, member) \ - ((pos) ? : list_entry(head, typeof(*pos), member)) - -/** - * list_for_each_entry_continue - continue iteration over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Continue to iterate over list of given type, continuing after - * the current position. - */ -#define list_for_each_entry_continue(pos, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_continue_reverse - iterate backwards from the given point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Start to iterate over list of given type backwards, continuing after - * the current position. - */ -#define list_for_each_entry_continue_reverse(pos, head, member) \ - for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_for_each_entry_from - iterate over list of given type from the current point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing from current position. - */ -#define list_for_each_entry_from(pos, head, member) \ - for (; prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_safe(pos, n, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_continue - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing after current point, - * safe against removal of list entry. - */ -#define list_for_each_entry_safe_continue(pos, n, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_from - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type from current point, safe against - * removal of list entry. - */ -#define list_for_each_entry_safe_from(pos, n, head, member) \ - for (n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_reverse - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate backwards over list of given type, safe against removal - * of list entry. - */ -#define list_for_each_entry_safe_reverse(pos, n, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member), \ - n = list_entry(pos->member.prev, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.prev, typeof(*n), member)) - -/* - * Double linked lists with a single pointer list head. - * Mostly useful for hash tables where the two pointer list head is - * too wasteful. - * You lose the ability to access the tail in O(1). - */ - -struct hlist_head { - struct hlist_node *first; -}; - -struct hlist_node { - struct hlist_node *next, **pprev; -}; - -#define HLIST_HEAD_INIT { .first = NULL } -#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } -#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) -static inline void INIT_HLIST_NODE(struct hlist_node *h) -{ - h->next = NULL; - h->pprev = NULL; -} - -static inline int hlist_unhashed(const struct hlist_node *h) -{ - return !h->pprev; -} - -static inline int hlist_empty(const struct hlist_head *h) -{ - return !h->first; -} - -static inline void __hlist_del(struct hlist_node *n) -{ - struct hlist_node *next = n->next; - struct hlist_node **pprev = n->pprev; - *pprev = next; - if (next) - next->pprev = pprev; -} - -static inline void hlist_del(struct hlist_node *n) -{ - __hlist_del(n); - n->next = LIST_POISON1; - n->pprev = LIST_POISON2; -} - -static inline void hlist_del_init(struct hlist_node *n) -{ - if (!hlist_unhashed(n)) { - __hlist_del(n); - INIT_HLIST_NODE(n); - } -} - -static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) -{ - struct hlist_node *first = h->first; - n->next = first; - if (first) - first->pprev = &n->next; - h->first = n; - n->pprev = &h->first; -} - -/* next must be != NULL */ -static inline void hlist_add_before(struct hlist_node *n, - struct hlist_node *next) -{ - n->pprev = next->pprev; - n->next = next; - next->pprev = &n->next; - *(n->pprev) = n; -} - -static inline void hlist_add_after(struct hlist_node *n, - struct hlist_node *next) -{ - next->next = n->next; - n->next = next; - next->pprev = &n->next; - - if(next->next) - next->next->pprev = &next->next; -} - -/* - * Move a list from one list head to another. Fixup the pprev - * reference of the first entry if it exists. - */ -static inline void hlist_move_list(struct hlist_head *old, - struct hlist_head *new) -{ - new->first = old->first; - if (new->first) - new->first->pprev = &new->first; - old->first = NULL; -} - -#define hlist_entry(ptr, type, member) container_of(ptr,type,member) - -#define hlist_for_each(pos, head) \ - for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ - pos = pos->next) - -#define hlist_for_each_safe(pos, n, head) \ - for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ - pos = n) - -/** - * hlist_for_each_entry - iterate over list of given type - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry(tpos, pos, head, member) \ - for (pos = (head)->first; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_continue - iterate over a hlist continuing after current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_continue(tpos, pos, member) \ - for (pos = (pos)->next; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_from - iterate over a hlist continuing from current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_from(tpos, pos, member) \ - for (; pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @n: another &struct hlist_node to use as temporary storage - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ - for (pos = (head)->first; \ - pos && ({ n = pos->next; 1; }) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = n) - -#endif diff --git a/api_ppc.h b/api_ppc.h deleted file mode 100644 index 8a03faa..0000000 --- a/api_ppc.h +++ /dev/null @@ -1,1699 +0,0 @@ -/* MECHANICALLY GENERATED, DO NOT EDIT!!! */ - -#define _INCLUDE_API_H - -/* - * common.h: Common Linux kernel-isms. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; but version 2 of the License only due - * to code included from the Linux kernel. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -#ifndef __always_inline -#define __always_inline inline -#endif - -#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) -#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) - -#ifdef __ASSEMBLY__ -# define stringify_in_c(...) __VA_ARGS__ -# define ASM_CONST(x) x -#else -/* This version of stringify will deal with commas... */ -# define __stringify_in_c(...) #__VA_ARGS__ -# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " -# define __ASM_CONST(x) x##UL -# define ASM_CONST(x) __ASM_CONST(x) -#endif - - -/* - * arch-ppc64.h: Expose PowerPC atomic instructions. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; but version 2 of the License only due - * to code included from the Linux kernel. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -/* - * Machine parameters. - */ - -#define CONFIG_PPC64 - -#define CACHE_LINE_SIZE 128 -#define ____cacheline_internodealigned_in_smp \ - __attribute__((__aligned__(1 << 7))) - -/* - * Atomic data structure, initialization, and access. - */ - -typedef struct { volatile int counter; } atomic_t; - -#define ATOMIC_INIT(i) { (i) } - -#define atomic_read(v) ((v)->counter) -#define atomic_set(v, i) (((v)->counter) = (i)) - -/* - * Atomic operations. - */ - -#define LWSYNC lwsync -#define PPC405_ERR77(ra,rb) -#ifdef CONFIG_SMP -# define LWSYNC_ON_SMP stringify_in_c(LWSYNC) "\n" -# define ISYNC_ON_SMP "\n\tisync\n" -#else -# define LWSYNC_ON_SMP -# define ISYNC_ON_SMP -#endif - - -/* - * Atomic exchange - * - * Changes the memory location '*ptr' to be val and returns - * the previous value stored there. - */ -static __always_inline unsigned long -__xchg_u32(volatile void *p, unsigned long val) -{ - unsigned long prev; - - __asm__ __volatile__( - LWSYNC_ON_SMP -"1: lwarx %0,0,%2 \n" - PPC405_ERR77(0,%2) -" stwcx. %3,0,%2 \n\ - bne- 1b" - ISYNC_ON_SMP - : "=&r" (prev), "+m" (*(volatile unsigned int *)p) - : "r" (p), "r" (val) - : "cc", "memory"); - - return prev; -} - -/* - * Atomic exchange - * - * Changes the memory location '*ptr' to be val and returns - * the previous value stored there. - */ -static __always_inline unsigned long -__xchg_u32_local(volatile void *p, unsigned long val) -{ - unsigned long prev; - - __asm__ __volatile__( -"1: lwarx %0,0,%2 \n" - PPC405_ERR77(0,%2) -" stwcx. %3,0,%2 \n\ - bne- 1b" - : "=&r" (prev), "+m" (*(volatile unsigned int *)p) - : "r" (p), "r" (val) - : "cc", "memory"); - - return prev; -} - -#ifdef CONFIG_PPC64 -static __always_inline unsigned long -__xchg_u64(volatile void *p, unsigned long val) -{ - unsigned long prev; - - __asm__ __volatile__( - LWSYNC_ON_SMP -"1: ldarx %0,0,%2 \n" - PPC405_ERR77(0,%2) -" stdcx. %3,0,%2 \n\ - bne- 1b" - ISYNC_ON_SMP - : "=&r" (prev), "+m" (*(volatile unsigned long *)p) - : "r" (p), "r" (val) - : "cc", "memory"); - - return prev; -} - -static __always_inline unsigned long -__xchg_u64_local(volatile void *p, unsigned long val) -{ - unsigned long prev; - - __asm__ __volatile__( -"1: ldarx %0,0,%2 \n" - PPC405_ERR77(0,%2) -" stdcx. %3,0,%2 \n\ - bne- 1b" - : "=&r" (prev), "+m" (*(volatile unsigned long *)p) - : "r" (p), "r" (val) - : "cc", "memory"); - - return prev; -} -#endif - -/* - * This function doesn't exist, so you'll get a linker error - * if something tries to do an invalid xchg(). - */ -extern void __xchg_called_with_bad_pointer(void); - -static __always_inline unsigned long -__xchg(volatile void *ptr, unsigned long x, unsigned int size) -{ - switch (size) { - case 4: - return __xchg_u32(ptr, x); -#ifdef CONFIG_PPC64 - case 8: - return __xchg_u64(ptr, x); -#endif - } - __xchg_called_with_bad_pointer(); - return x; -} - -static __always_inline unsigned long -__xchg_local(volatile void *ptr, unsigned long x, unsigned int size) -{ - switch (size) { - case 4: - return __xchg_u32_local(ptr, x); -#ifdef CONFIG_PPC64 - case 8: - return __xchg_u64_local(ptr, x); -#endif - } - __xchg_called_with_bad_pointer(); - return x; -} -#define xchg(ptr,x) \ - ({ \ - __typeof__(*(ptr)) _x_ = (x); \ - (__typeof__(*(ptr))) __xchg((ptr), (unsigned long)_x_, sizeof(*(ptr))); \ - }) - -#define xchg_local(ptr,x) \ - ({ \ - __typeof__(*(ptr)) _x_ = (x); \ - (__typeof__(*(ptr))) __xchg_local((ptr), \ - (unsigned long)_x_, sizeof(*(ptr))); \ - }) - -/* - * Compare and exchange - if *p == old, set it to new, - * and return the old value of *p. - */ -#define __HAVE_ARCH_CMPXCHG 1 - -static __always_inline unsigned long -__cmpxchg_u32(volatile unsigned int *p, unsigned long old, unsigned long new) -{ - unsigned int prev; - - __asm__ __volatile__ ( - LWSYNC_ON_SMP -"1: lwarx %0,0,%2 # __cmpxchg_u32\n\ - cmpw 0,%0,%3\n\ - bne- 2f\n" - PPC405_ERR77(0,%2) -" stwcx. %4,0,%2\n\ - bne- 1b" - ISYNC_ON_SMP - "\n\ -2:" - : "=&r" (prev), "+m" (*p) - : "r" (p), "r" (old), "r" (new) - : "cc", "memory"); - - return prev; -} - -static __always_inline unsigned long -__cmpxchg_u32_local(volatile unsigned int *p, unsigned long old, - unsigned long new) -{ - unsigned int prev; - - __asm__ __volatile__ ( -"1: lwarx %0,0,%2 # __cmpxchg_u32\n\ - cmpw 0,%0,%3\n\ - bne- 2f\n" - PPC405_ERR77(0,%2) -" stwcx. %4,0,%2\n\ - bne- 1b" - "\n\ -2:" - : "=&r" (prev), "+m" (*p) - : "r" (p), "r" (old), "r" (new) - : "cc", "memory"); - - return prev; -} - -#ifdef CONFIG_PPC64 -static __always_inline unsigned long -__cmpxchg_u64(volatile unsigned long *p, unsigned long old, unsigned long new) -{ - unsigned long prev; - - __asm__ __volatile__ ( - LWSYNC_ON_SMP -"1: ldarx %0,0,%2 # __cmpxchg_u64\n\ - cmpd 0,%0,%3\n\ - bne- 2f\n\ - stdcx. %4,0,%2\n\ - bne- 1b" - ISYNC_ON_SMP - "\n\ -2:" - : "=&r" (prev), "+m" (*p) - : "r" (p), "r" (old), "r" (new) - : "cc", "memory"); - - return prev; -} - -static __always_inline unsigned long -__cmpxchg_u64_local(volatile unsigned long *p, unsigned long old, - unsigned long new) -{ - unsigned long prev; - - __asm__ __volatile__ ( -"1: ldarx %0,0,%2 # __cmpxchg_u64\n\ - cmpd 0,%0,%3\n\ - bne- 2f\n\ - stdcx. %4,0,%2\n\ - bne- 1b" - "\n\ -2:" - : "=&r" (prev), "+m" (*p) - : "r" (p), "r" (old), "r" (new) - : "cc", "memory"); - - return prev; -} -#endif - -/* This function doesn't exist, so you'll get a linker error - if something tries to do an invalid cmpxchg(). */ -extern void __cmpxchg_called_with_bad_pointer(void); - -static __always_inline unsigned long -__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, - unsigned int size) -{ - switch (size) { - case 4: - return __cmpxchg_u32(ptr, old, new); -#ifdef CONFIG_PPC64 - case 8: - return __cmpxchg_u64(ptr, old, new); -#endif - } - __cmpxchg_called_with_bad_pointer(); - return old; -} - -static __always_inline unsigned long -__cmpxchg_local(volatile void *ptr, unsigned long old, unsigned long new, - unsigned int size) -{ - switch (size) { - case 4: - return __cmpxchg_u32_local(ptr, old, new); -#ifdef CONFIG_PPC64 - case 8: - return __cmpxchg_u64_local(ptr, old, new); -#endif - } - __cmpxchg_called_with_bad_pointer(); - return old; -} - -#define cmpxchg(ptr, o, n) \ - ({ \ - __typeof__(*(ptr)) _o_ = (o); \ - __typeof__(*(ptr)) _n_ = (n); \ - (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_, \ - (unsigned long)_n_, sizeof(*(ptr))); \ - }) - - -#define cmpxchg_local(ptr, o, n) \ - ({ \ - __typeof__(*(ptr)) _o_ = (o); \ - __typeof__(*(ptr)) _n_ = (n); \ - (__typeof__(*(ptr))) __cmpxchg_local((ptr), (unsigned long)_o_, \ - (unsigned long)_n_, sizeof(*(ptr))); \ - }) - -#ifdef CONFIG_PPC64 -/* - * We handle most unaligned accesses in hardware. On the other hand - * unaligned DMA can be very expensive on some ppc64 IO chips (it does - * powers of 2 writes until it reaches sufficient alignment). - * - * Based on this we disable the IP header alignment in network drivers. - * We also modify NET_SKB_PAD to be a cacheline in size, thus maintaining - * cacheline alignment of buffers. - */ -#define NET_IP_ALIGN 0 -#define NET_SKB_PAD L1_CACHE_BYTES - -#define cmpxchg64(ptr, o, n) \ - ({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ - cmpxchg((ptr), (o), (n)); \ - }) -#define cmpxchg64_local(ptr, o, n) \ - ({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ - cmpxchg_local((ptr), (o), (n)); \ - }) -#endif - -#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n))) -#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) - -/** - * atomic_add - add integer to atomic variable - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @a to @v. - */ -static __inline__ void atomic_add(int a, atomic_t *v) -{ - int t; - - __asm__ __volatile__( - "1: lwarx %0,0,%3 # atomic_add\n\ - add %0,%2,%0 \n\ - stwcx. %0,0,%3 \n\ - bne- 1b" - : "=&r" (t), "+m" (v->counter) - : "r" (a), "r" (&v->counter) - : "cc"); -} - -/** - * atomic_sub - subtract the atomic variable - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @a from @v. - */ -static __inline__ void atomic_sub(int a, atomic_t *v) -{ - int t; - - __asm__ __volatile__( - "1: lwarx %0,0,%3 # atomic_sub \n\ - subf %0,%2,%0 \n\ - stwcx. %0,0,%3 \n\ - bne- 1b" - : "=&r" (t), "+m" (v->counter) - : "r" (a), "r" (&v->counter) - : "cc"); -} - -static __inline__ atomic_sub_return(int a, atomic_t *v) -{ - int t; - - __asm__ __volatile__( - "lwsync\n\ - 1: lwarx %0,0,%2 # atomic_sub_return\n\ - subf %0,%1,%0\n\ - stwcx. %0,0,%2 \n\ - bne- 1b \n\ - isync" - : "=&r" (t) - : "r" (a), "r" (&v->counter) - : "cc", "memory"); - - return t; -} - -/** - * atomic_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_sub_and_test(int a, atomic_t *v) -{ - return atomic_sub_return(a, v) == 0; -} - -/** - * atomic_inc - increment atomic variable - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1. - */ -static __inline__ void atomic_inc(atomic_t *v) -{ - atomic_add(1, v); -} - -/** - * atomic_dec - decrement atomic variable - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1. - */ -static __inline__ void atomic_dec(atomic_t *v) -{ - atomic_sub(1, v); -} - -/** - * atomic_dec_and_test - decrement and test - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static __inline__ int atomic_dec_and_test(atomic_t *v) -{ - return atomic_sub_and_test(1, v); -} - -/** - * atomic_inc_and_test - increment and test - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_inc_and_test(atomic_t *v) -{ - return atomic_inc_return(v); -} - -/** - * atomic_add_return - add and return - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns @i + @v - */ -static __inline__ int atomic_add_return(int a, atomic_t *v) -{ - int t; - - __asm__ __volatile__( - "lwsync \n\ - 1: lwarx %0,0,%2 # atomic_add_return \n\ - add %0,%1,%0 \n\ - stwcx. %0,0,%2 \n\ - bne- 1b \n\ - isync" - : "=&r" (t) - : "r" (a), "r" (&v->counter) - : "cc", "memory"); - - return t; -} - -/** - * atomic_add_negative - add and test if negative - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static __inline__ int atomic_add_negative(int a, atomic_t *v) -{ - return atomic_add_return(a, v) < 0; -} - -/** - * atomic_add_unless - add unless the number is a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as it was not @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -static __inline__ int atomic_add_unless(atomic_t *v, int a, int u) -{ - int t; - - __asm__ __volatile__( - "lwsync \n\ - 1: lwarx %0,0,%1 # atomic_add_unless\n\ - cmpd 0,%0,%3 \n\ - beq- 2f \n\ - add %0,%2,%0 \n\ - stwcx. %0,0,%1 \n\ - bne- 1b \n\ - isync \n\ - subf %0,%2,%0 \n\ - 2:" - : "=&r" (t) - : "r" (&v->counter), "r" (a), "r" (u) - : "cc", "memory"); - - return t != u; -} - -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) - -#define atomic_inc_return(v) (atomic_add_return(1,v)) -#define atomic_dec_return(v) (atomic_sub_return(1,v)) - -/* Atomic operations are already serializing on x86 */ -#define smp_mb__before_atomic_dec() smp_mb() -#define smp_mb__after_atomic_dec() smp_mb() -#define smp_mb__before_atomic_inc() smp_mb() -#define smp_mb__after_atomic_inc() smp_mb() - -/* - * api_pthreads.h: API mapping to pthreads environment. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. However, please note that much - * of the code in this file derives from the Linux kernel, and that such - * code may not be available except under GPLv2. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - */ - -#include -#include -#include -#include -#include -#define __USE_GNU -#include -#include -#include -/* #include "atomic.h" */ - -/* - * Compiler magic. - */ -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Default machine parameters. - */ - -#ifndef CACHE_LINE_SIZE -#define CACHE_LINE_SIZE 128 -#endif /* #ifndef CACHE_LINE_SIZE */ - -/* - * Exclusive locking primitives. - */ - -typedef pthread_mutex_t spinlock_t; - -#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; -#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER - -static void spin_lock_init(spinlock_t *sp) -{ - if (pthread_mutex_init(sp, NULL) != 0) { - perror("spin_lock_init:pthread_mutex_init"); - exit(-1); - } -} - -static void spin_lock(spinlock_t *sp) -{ - if (pthread_mutex_lock(sp) != 0) { - perror("spin_lock:pthread_mutex_lock"); - exit(-1); - } -} - -static void spin_unlock(spinlock_t *sp) -{ - if (pthread_mutex_unlock(sp) != 0) { - perror("spin_unlock:pthread_mutex_unlock"); - exit(-1); - } -} - -#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) -#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) - -/* - * Thread creation/destruction primitives. - */ - -typedef pthread_t thread_id_t; - -#define NR_THREADS 128 - -#define __THREAD_ID_MAP_EMPTY 0 -#define __THREAD_ID_MAP_WAITING 1 -thread_id_t __thread_id_map[NR_THREADS]; -spinlock_t __thread_id_map_mutex; - -#define for_each_thread(t) \ - for (t = 0; t < NR_THREADS; t++) - -#define for_each_running_thread(t) \ - for (t = 0; t < NR_THREADS; t++) \ - if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ - (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) - -#define for_each_tid(t, tid) \ - for (t = 0; t < NR_THREADS; t++) \ - if ((((tid) = __thread_id_map[t]) != __THREAD_ID_MAP_EMPTY) && \ - ((tid) != __THREAD_ID_MAP_WAITING)) - -pthread_key_t thread_id_key; - -static int __smp_thread_id(void) -{ - int i; - thread_id_t tid = pthread_self(); - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) { - long v = i + 1; /* must be non-NULL. */ - - if (pthread_setspecific(thread_id_key, (void *)v) != 0) { - perror("pthread_setspecific"); - exit(-1); - } - return i; - } - } - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - spin_unlock(&__thread_id_map_mutex); - return i; - } - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); -} - -static int smp_thread_id(void) -{ - void *id; - - id = pthread_getspecific(thread_id_key); - if (id == NULL) - return __smp_thread_id(); - return (long)(id - 1); -} - -static thread_id_t create_thread(void *(*func)(void *), void *arg) -{ - thread_id_t tid; - int i; - - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) - break; - } - if (i >= NR_THREADS) { - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_WAITING; - spin_unlock(&__thread_id_map_mutex); - if (pthread_create(&tid, NULL, func, arg) != 0) { - perror("create_thread:pthread_create"); - exit(-1); - } - __thread_id_map[i] = tid; - return tid; -} - -static void *wait_thread(thread_id_t tid) -{ - int i; - void *vp; - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - break; - } - if (i >= NR_THREADS){ - fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); - } - if (pthread_join(tid, &vp) != 0) { - perror("wait_thread:pthread_join"); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - return vp; -} - -static void wait_all_threads(void) -{ - int i; - thread_id_t tid; - - for (i = 1; i < NR_THREADS; i++) { - tid = __thread_id_map[i]; - if (tid != __THREAD_ID_MAP_EMPTY && - tid != __THREAD_ID_MAP_WAITING) - (void)wait_thread(tid); - } -} - -static void run_on(int cpu) -{ - cpu_set_t mask; - - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * timekeeping -- very crude -- should use MONOTONIC... - */ - -long long get_microseconds(void) -{ - struct timeval tv; - - if (gettimeofday(&tv, NULL) != 0) - abort(); - return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; -} - -/* - * Per-thread variables. - */ - -#define DEFINE_PER_THREAD(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_thread_##name[NR_THREADS]; -#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) - -#define per_thread(name, thread) __per_thread_##name[thread].v -#define __get_thread_var(name) per_thread(name, smp_thread_id()) - -#define init_per_thread(name, v) \ - do { \ - int __i_p_t_i; \ - for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ - per_thread(name, __i_p_t_i) = v; \ - } while (0) - -/* - * CPU traversal primitives. - */ - -#ifndef NR_CPUS -#define NR_CPUS 16 -#endif /* #ifndef NR_CPUS */ - -#define for_each_possible_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) -#define for_each_online_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) - -/* - * Per-CPU variables. - */ - -#define DEFINE_PER_CPU(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_cpu_##name[NR_CPUS] -#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) - -DEFINE_PER_THREAD(int, smp_processor_id); - -#define per_cpu(name, thread) __per_cpu_##name[thread].v -#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) - -#define init_per_cpu(name, v) \ - do { \ - int __i_p_c_i; \ - for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ - per_cpu(name, __i_p_c_i) = v; \ - } while (0) - -/* - * CPU state checking (crowbarred). - */ - -#define idle_cpu(cpu) 0 -#define in_softirq() 1 -#define hardirq_count() 0 -#define PREEMPT_SHIFT 0 -#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) -#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) -#define PREEMPT_BITS 8 -#define SOFTIRQ_BITS 8 - -/* - * CPU hotplug. - */ - -struct notifier_block { - int (*notifier_call)(struct notifier_block *, unsigned long, void *); - struct notifier_block *next; - int priority; -}; - -#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ -#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ -#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ -#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ -#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ -#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ -#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, - * not handling interrupts, soon dead */ -#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug - * lock is dropped */ - -/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend - * operation in progress - */ -#define CPU_TASKS_FROZEN 0x0010 - -#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) -#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) -#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) -#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) -#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) -#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) -#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) - -/* Hibernation and suspend events */ -#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ -#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ -#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ -#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ -#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ -#define PM_POST_RESTORE 0x0006 /* Restore failed */ - -#define NOTIFY_DONE 0x0000 /* Don't care */ -#define NOTIFY_OK 0x0001 /* Suits me */ -#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ -#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) - /* Bad/Veto action */ -/* - * Clean way to return from the notifier and stop further calls. - */ -#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) - -/* - * Bug checks. - */ - -#define BUG_ON(c) do { if (!(c)) abort(); } while (0) - -/* - * Initialization -- Must be called before calling any primitives. - */ - -static void smp_init(void) -{ - int i; - - spin_lock_init(&__thread_id_map_mutex); - __thread_id_map[0] = pthread_self(); - for (i = 1; i < NR_THREADS; i++) - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - init_per_thread(smp_processor_id, 0); - if (pthread_key_create(&thread_id_key, NULL) != 0) { - perror("pthread_key_create"); - exit(-1); - } -} - -/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ - -#ifndef _LINUX_LIST_H -#define _LINUX_LIST_H - -#define LIST_POISON1 ((void *) 0x00100100) -#define LIST_POISON2 ((void *) 0x00200200) - -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Simple doubly linked list implementation. - * - * Some of the internal functions ("__xxx") are useful when - * manipulating whole lists rather than single entries, as - * sometimes we already know the next/prev entries and we can - * generate better code by using them directly rather than - * using the generic single-entry routines. - */ - -struct list_head { - struct list_head *next, *prev; -}; - -#define LIST_HEAD_INIT(name) { &(name), &(name) } - -#define LIST_HEAD(name) \ - struct list_head name = LIST_HEAD_INIT(name) - -static inline void INIT_LIST_HEAD(struct list_head *list) -{ - list->next = list; - list->prev = list; -} - -/* - * Insert a new entry between two known consecutive entries. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -#ifndef CONFIG_DEBUG_LIST -static inline void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next) -{ - next->prev = new; - new->next = next; - new->prev = prev; - prev->next = new; -} -#else -extern void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next); -#endif - -/** - * list_add - add a new entry - * @new: new entry to be added - * @head: list head to add it after - * - * Insert a new entry after the specified head. - * This is good for implementing stacks. - */ -static inline void list_add(struct list_head *new, struct list_head *head) -{ - __list_add(new, head, head->next); -} - - -/** - * list_add_tail - add a new entry - * @new: new entry to be added - * @head: list head to add it before - * - * Insert a new entry before the specified head. - * This is useful for implementing queues. - */ -static inline void list_add_tail(struct list_head *new, struct list_head *head) -{ - __list_add(new, head->prev, head); -} - -/* - * Delete a list entry by making the prev/next entries - * point to each other. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -static inline void __list_del(struct list_head * prev, struct list_head * next) -{ - next->prev = prev; - prev->next = next; -} - -/** - * list_del - deletes entry from list. - * @entry: the element to delete from the list. - * Note: list_empty() on entry does not return true after this, the entry is - * in an undefined state. - */ -#ifndef CONFIG_DEBUG_LIST -static inline void list_del(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - entry->next = LIST_POISON1; - entry->prev = LIST_POISON2; -} -#else -extern void list_del(struct list_head *entry); -#endif - -/** - * list_replace - replace old entry by new one - * @old : the element to be replaced - * @new : the new element to insert - * - * If @old was empty, it will be overwritten. - */ -static inline void list_replace(struct list_head *old, - struct list_head *new) -{ - new->next = old->next; - new->next->prev = new; - new->prev = old->prev; - new->prev->next = new; -} - -static inline void list_replace_init(struct list_head *old, - struct list_head *new) -{ - list_replace(old, new); - INIT_LIST_HEAD(old); -} - -/** - * list_del_init - deletes entry from list and reinitialize it. - * @entry: the element to delete from the list. - */ -static inline void list_del_init(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - INIT_LIST_HEAD(entry); -} - -/** - * list_move - delete from one list and add as another's head - * @list: the entry to move - * @head: the head that will precede our entry - */ -static inline void list_move(struct list_head *list, struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add(list, head); -} - -/** - * list_move_tail - delete from one list and add as another's tail - * @list: the entry to move - * @head: the head that will follow our entry - */ -static inline void list_move_tail(struct list_head *list, - struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add_tail(list, head); -} - -/** - * list_is_last - tests whether @list is the last entry in list @head - * @list: the entry to test - * @head: the head of the list - */ -static inline int list_is_last(const struct list_head *list, - const struct list_head *head) -{ - return list->next == head; -} - -/** - * list_empty - tests whether a list is empty - * @head: the list to test. - */ -static inline int list_empty(const struct list_head *head) -{ - return head->next == head; -} - -/** - * list_empty_careful - tests whether a list is empty and not being modified - * @head: the list to test - * - * Description: - * tests whether a list is empty _and_ checks that no other CPU might be - * in the process of modifying either member (next or prev) - * - * NOTE: using list_empty_careful() without synchronization - * can only be safe if the only activity that can happen - * to the list entry is list_del_init(). Eg. it cannot be used - * if another CPU could re-list_add() it. - */ -static inline int list_empty_careful(const struct list_head *head) -{ - struct list_head *next = head->next; - return (next == head) && (next == head->prev); -} - -/** - * list_is_singular - tests whether a list has just one entry. - * @head: the list to test. - */ -static inline int list_is_singular(const struct list_head *head) -{ - return !list_empty(head) && (head->next == head->prev); -} - -static inline void __list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - struct list_head *new_first = entry->next; - list->next = head->next; - list->next->prev = list; - list->prev = entry; - entry->next = list; - head->next = new_first; - new_first->prev = head; -} - -/** - * list_cut_position - cut a list into two - * @list: a new list to add all removed entries - * @head: a list with entries - * @entry: an entry within head, could be the head itself - * and if so we won't cut the list - * - * This helper moves the initial part of @head, up to and - * including @entry, from @head to @list. You should - * pass on @entry an element you know is on @head. @list - * should be an empty list or a list you do not care about - * losing its data. - * - */ -static inline void list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - if (list_empty(head)) - return; - if (list_is_singular(head) && - (head->next != entry && head != entry)) - return; - if (entry == head) - INIT_LIST_HEAD(list); - else - __list_cut_position(list, head, entry); -} - -static inline void __list_splice(const struct list_head *list, - struct list_head *prev, - struct list_head *next) -{ - struct list_head *first = list->next; - struct list_head *last = list->prev; - - first->prev = prev; - prev->next = first; - - last->next = next; - next->prev = last; -} - -/** - * list_splice - join two lists, this is designed for stacks - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice(const struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head, head->next); -} - -/** - * list_splice_tail - join two lists, each list being a queue - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice_tail(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head->prev, head); -} - -/** - * list_splice_init - join two lists and reinitialise the emptied list. - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * The list at @list is reinitialised - */ -static inline void list_splice_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head, head->next); - INIT_LIST_HEAD(list); - } -} - -/** - * list_splice_tail_init - join two lists and reinitialise the emptied list - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * Each of the lists is a queue. - * The list at @list is reinitialised - */ -static inline void list_splice_tail_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head->prev, head); - INIT_LIST_HEAD(list); - } -} - -/** - * list_entry - get the struct for this entry - * @ptr: the &struct list_head pointer. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - */ -#define list_entry(ptr, type, member) \ - container_of(ptr, type, member) - -/** - * list_first_entry - get the first element from a list - * @ptr: the list head to take the element from. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - * - * Note, that list is expected to be not empty. - */ -#define list_first_entry(ptr, type, member) \ - list_entry((ptr)->next, type, member) - -/** - * list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each(pos, head) \ - for (pos = (head)->next; prefetch(pos->next), pos != (head); \ - pos = pos->next) - -/** - * __list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - * - * This variant differs from list_for_each() in that it's the - * simplest possible list iteration code, no prefetching is done. - * Use this for code that knows the list to be very short (empty - * or 1 entry) most of the time. - */ -#define __list_for_each(pos, head) \ - for (pos = (head)->next; pos != (head); pos = pos->next) - -/** - * list_for_each_prev - iterate over a list backwards - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each_prev(pos, head) \ - for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ - pos = pos->prev) - -/** - * list_for_each_safe - iterate over a list safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_safe(pos, n, head) \ - for (pos = (head)->next, n = pos->next; pos != (head); \ - pos = n, n = pos->next) - -/** - * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_prev_safe(pos, n, head) \ - for (pos = (head)->prev, n = pos->prev; \ - prefetch(pos->prev), pos != (head); \ - pos = n, n = pos->prev) - -/** - * list_for_each_entry - iterate over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry(pos, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_reverse - iterate backwards over list of given type. - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_reverse(pos, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() - * @pos: the type * to use as a start point - * @head: the head of the list - * @member: the name of the list_struct within the struct. - * - * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). - */ -#define list_prepare_entry(pos, head, member) \ - ((pos) ? : list_entry(head, typeof(*pos), member)) - -/** - * list_for_each_entry_continue - continue iteration over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Continue to iterate over list of given type, continuing after - * the current position. - */ -#define list_for_each_entry_continue(pos, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_continue_reverse - iterate backwards from the given point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Start to iterate over list of given type backwards, continuing after - * the current position. - */ -#define list_for_each_entry_continue_reverse(pos, head, member) \ - for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_for_each_entry_from - iterate over list of given type from the current point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing from current position. - */ -#define list_for_each_entry_from(pos, head, member) \ - for (; prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_safe(pos, n, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_continue - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing after current point, - * safe against removal of list entry. - */ -#define list_for_each_entry_safe_continue(pos, n, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_from - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type from current point, safe against - * removal of list entry. - */ -#define list_for_each_entry_safe_from(pos, n, head, member) \ - for (n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_reverse - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate backwards over list of given type, safe against removal - * of list entry. - */ -#define list_for_each_entry_safe_reverse(pos, n, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member), \ - n = list_entry(pos->member.prev, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.prev, typeof(*n), member)) - -/* - * Double linked lists with a single pointer list head. - * Mostly useful for hash tables where the two pointer list head is - * too wasteful. - * You lose the ability to access the tail in O(1). - */ - -struct hlist_head { - struct hlist_node *first; -}; - -struct hlist_node { - struct hlist_node *next, **pprev; -}; - -#define HLIST_HEAD_INIT { .first = NULL } -#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } -#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) -static inline void INIT_HLIST_NODE(struct hlist_node *h) -{ - h->next = NULL; - h->pprev = NULL; -} - -static inline int hlist_unhashed(const struct hlist_node *h) -{ - return !h->pprev; -} - -static inline int hlist_empty(const struct hlist_head *h) -{ - return !h->first; -} - -static inline void __hlist_del(struct hlist_node *n) -{ - struct hlist_node *next = n->next; - struct hlist_node **pprev = n->pprev; - *pprev = next; - if (next) - next->pprev = pprev; -} - -static inline void hlist_del(struct hlist_node *n) -{ - __hlist_del(n); - n->next = LIST_POISON1; - n->pprev = LIST_POISON2; -} - -static inline void hlist_del_init(struct hlist_node *n) -{ - if (!hlist_unhashed(n)) { - __hlist_del(n); - INIT_HLIST_NODE(n); - } -} - -static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) -{ - struct hlist_node *first = h->first; - n->next = first; - if (first) - first->pprev = &n->next; - h->first = n; - n->pprev = &h->first; -} - -/* next must be != NULL */ -static inline void hlist_add_before(struct hlist_node *n, - struct hlist_node *next) -{ - n->pprev = next->pprev; - n->next = next; - next->pprev = &n->next; - *(n->pprev) = n; -} - -static inline void hlist_add_after(struct hlist_node *n, - struct hlist_node *next) -{ - next->next = n->next; - n->next = next; - next->pprev = &n->next; - - if(next->next) - next->next->pprev = &next->next; -} - -/* - * Move a list from one list head to another. Fixup the pprev - * reference of the first entry if it exists. - */ -static inline void hlist_move_list(struct hlist_head *old, - struct hlist_head *new) -{ - new->first = old->first; - if (new->first) - new->first->pprev = &new->first; - old->first = NULL; -} - -#define hlist_entry(ptr, type, member) container_of(ptr,type,member) - -#define hlist_for_each(pos, head) \ - for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ - pos = pos->next) - -#define hlist_for_each_safe(pos, n, head) \ - for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ - pos = n) - -/** - * hlist_for_each_entry - iterate over list of given type - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry(tpos, pos, head, member) \ - for (pos = (head)->first; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_continue - iterate over a hlist continuing after current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_continue(tpos, pos, member) \ - for (pos = (pos)->next; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_from - iterate over a hlist continuing from current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_from(tpos, pos, member) \ - for (; pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @n: another &struct hlist_node to use as temporary storage - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ - for (pos = (head)->first; \ - pos && ({ n = pos->next; 1; }) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = n) - -#endif diff --git a/api_x86.h b/api_x86.h deleted file mode 100644 index f48fce9..0000000 --- a/api_x86.h +++ /dev/null @@ -1,1387 +0,0 @@ -/* MECHANICALLY GENERATED, DO NOT EDIT!!! */ - -#define _INCLUDE_API_H - -/* - * common.h: Common Linux kernel-isms. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; but version 2 of the License only due - * to code included from the Linux kernel. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -#ifndef __always_inline -#define __always_inline inline -#endif - -#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) -#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) - -#ifdef __ASSEMBLY__ -# define stringify_in_c(...) __VA_ARGS__ -# define ASM_CONST(x) x -#else -/* This version of stringify will deal with commas... */ -# define __stringify_in_c(...) #__VA_ARGS__ -# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " -# define __ASM_CONST(x) x##UL -# define ASM_CONST(x) __ASM_CONST(x) -#endif - - -/* - * arch-i386.h: Expose x86 atomic instructions. 80486 and better only. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation, but version 2 only due to inclusion - * of Linux-kernel code. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - * - * Much code taken from the Linux kernel. For such code, the option - * to redistribute under later versions of GPL might not be available. - */ - -/* - * Machine parameters. - */ - -#define CACHE_LINE_SIZE 64 -#define ____cacheline_internodealigned_in_smp \ - __attribute__((__aligned__(1 << 6))) - -#define LOCK_PREFIX "lock ; " - -/* - * Atomic data structure, initialization, and access. - */ - -typedef struct { volatile int counter; } atomic_t; - -#define ATOMIC_INIT(i) { (i) } - -#define atomic_read(v) ((v)->counter) -#define atomic_set(v, i) (((v)->counter) = (i)) - -/* - * Atomic operations. - */ - -/** - * atomic_add - add integer to atomic variable - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v. - */ -static __inline__ void atomic_add(int i, atomic_t *v) -{ - __asm__ __volatile__( - LOCK_PREFIX "addl %1,%0" - :"+m" (v->counter) - :"ir" (i)); -} - -/** - * atomic_sub - subtract the atomic variable - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v. - */ -static __inline__ void atomic_sub(int i, atomic_t *v) -{ - __asm__ __volatile__( - LOCK_PREFIX "subl %1,%0" - :"+m" (v->counter) - :"ir" (i)); -} - -/** - * atomic_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_sub_and_test(int i, atomic_t *v) -{ - unsigned char c; - - __asm__ __volatile__( - LOCK_PREFIX "subl %2,%0; sete %1" - :"+m" (v->counter), "=qm" (c) - :"ir" (i) : "memory"); - return c; -} - -/** - * atomic_inc - increment atomic variable - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1. - */ -static __inline__ void atomic_inc(atomic_t *v) -{ - __asm__ __volatile__( - LOCK_PREFIX "incl %0" - :"+m" (v->counter)); -} - -/** - * atomic_dec - decrement atomic variable - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1. - */ -static __inline__ void atomic_dec(atomic_t *v) -{ - __asm__ __volatile__( - LOCK_PREFIX "decl %0" - :"+m" (v->counter)); -} - -/** - * atomic_dec_and_test - decrement and test - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static __inline__ int atomic_dec_and_test(atomic_t *v) -{ - unsigned char c; - - __asm__ __volatile__( - LOCK_PREFIX "decl %0; sete %1" - :"+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_inc_and_test - increment and test - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static __inline__ int atomic_inc_and_test(atomic_t *v) -{ - unsigned char c; - - __asm__ __volatile__( - LOCK_PREFIX "incl %0; sete %1" - :"+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_add_negative - add and test if negative - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static __inline__ int atomic_add_negative(int i, atomic_t *v) -{ - unsigned char c; - - __asm__ __volatile__( - LOCK_PREFIX "addl %2,%0; sets %1" - :"+m" (v->counter), "=qm" (c) - :"ir" (i) : "memory"); - return c; -} - -/** - * atomic_add_return - add and return - * @v: pointer of type atomic_t - * @i: integer value to add - * - * Atomically adds @i to @v and returns @i + @v - */ -static __inline__ int atomic_add_return(int i, atomic_t *v) -{ - int __i; - - __i = i; - __asm__ __volatile__( - LOCK_PREFIX "xaddl %0, %1;" - :"=r"(i) - :"m"(v->counter), "0"(i)); - return i + __i; -} - -static __inline__ int atomic_sub_return(int i, atomic_t *v) -{ - return atomic_add_return(-i,v); -} - -static inline unsigned int -cmpxchg(volatile long *ptr, long oldval, long newval) -{ - unsigned long retval; - - asm("# cmpxchg\n" - "lock; cmpxchgl %4,(%2)\n" - "# end atomic_cmpxchg4" - : "=a" (retval), "=m" (*ptr) - : "r" (ptr), "0" (oldval), "r" (newval), "m" (*ptr) - : "cc"); - return (retval); -} - -#define atomic_cmpxchg(v, old, new) ((int)cmpxchg(&((v)->counter), old, new)) -#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) - -/** - * atomic_add_unless - add unless the number is a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as it was not @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -#define atomic_add_unless(v, a, u) \ -({ \ - int c, old; \ - c = atomic_read(v); \ - for (;;) { \ - if (unlikely(c == (u))) \ - break; \ - old = atomic_cmpxchg((v), c, c + (a)); \ - if (likely(old == c)) \ - break; \ - c = old; \ - } \ - c != (u); \ -}) -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) - -#define atomic_inc_return(v) (atomic_add_return(1,v)) -#define atomic_dec_return(v) (atomic_sub_return(1,v)) - -/* These are x86-specific, used by some header files */ -#define atomic_clear_mask(mask, addr) \ -__asm__ __volatile__(LOCK_PREFIX "andl %0,%1" \ -: : "r" (~(mask)),"m" (*addr) : "memory") - -#define atomic_set_mask(mask, addr) \ -__asm__ __volatile__(LOCK_PREFIX "orl %0,%1" \ -: : "r" (mask),"m" (*(addr)) : "memory") - -/* Atomic operations are already serializing on x86 */ -#define smp_mb__before_atomic_dec() barrier() -#define smp_mb__after_atomic_dec() barrier() -#define smp_mb__before_atomic_inc() barrier() -#define smp_mb__after_atomic_inc() barrier() - -/* - * api_pthreads.h: API mapping to pthreads environment. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. However, please note that much - * of the code in this file derives from the Linux kernel, and that such - * code may not be available except under GPLv2. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2006 Paul E. McKenney, IBM. - */ - -#include -#include -#include -#include -#include -#define __USE_GNU -#include -#include -#include -/* #include "atomic.h" */ - -/* - * Compiler magic. - */ -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Default machine parameters. - */ - -#ifndef CACHE_LINE_SIZE -#define CACHE_LINE_SIZE 128 -#endif /* #ifndef CACHE_LINE_SIZE */ - -/* - * Exclusive locking primitives. - */ - -typedef pthread_mutex_t spinlock_t; - -#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; -#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER - -static void spin_lock_init(spinlock_t *sp) -{ - if (pthread_mutex_init(sp, NULL) != 0) { - perror("spin_lock_init:pthread_mutex_init"); - exit(-1); - } -} - -static void spin_lock(spinlock_t *sp) -{ - if (pthread_mutex_lock(sp) != 0) { - perror("spin_lock:pthread_mutex_lock"); - exit(-1); - } -} - -static void spin_unlock(spinlock_t *sp) -{ - if (pthread_mutex_unlock(sp) != 0) { - perror("spin_unlock:pthread_mutex_unlock"); - exit(-1); - } -} - -#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) -#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) - -/* - * Thread creation/destruction primitives. - */ - -typedef pthread_t thread_id_t; - -#define NR_THREADS 128 - -#define __THREAD_ID_MAP_EMPTY 0 -#define __THREAD_ID_MAP_WAITING 1 -thread_id_t __thread_id_map[NR_THREADS]; -spinlock_t __thread_id_map_mutex; - -#define for_each_thread(t) \ - for (t = 0; t < NR_THREADS; t++) - -#define for_each_running_thread(t) \ - for (t = 0; t < NR_THREADS; t++) \ - if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ - (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) - -pthread_key_t thread_id_key; - -static int __smp_thread_id(void) -{ - int i; - thread_id_t tid = pthread_self(); - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) { - long v = i + 1; /* must be non-NULL. */ - - if (pthread_setspecific(thread_id_key, (void *)v) != 0) { - perror("pthread_setspecific"); - exit(-1); - } - return i; - } - } - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - spin_unlock(&__thread_id_map_mutex); - return i; - } - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); -} - -static int smp_thread_id(void) -{ - void *id; - - id = pthread_getspecific(thread_id_key); - if (id == NULL) - return __smp_thread_id(); - return (long)(id - 1); -} - -static thread_id_t create_thread(void *(*func)(void *), void *arg) -{ - thread_id_t tid; - int i; - - spin_lock(&__thread_id_map_mutex); - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) - break; - } - if (i >= NR_THREADS) { - spin_unlock(&__thread_id_map_mutex); - fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_WAITING; - spin_unlock(&__thread_id_map_mutex); - if (pthread_create(&tid, NULL, func, arg) != 0) { - perror("create_thread:pthread_create"); - exit(-1); - } - __thread_id_map[i] = tid; - return tid; -} - -static void *wait_thread(thread_id_t tid) -{ - int i; - void *vp; - - for (i = 0; i < NR_THREADS; i++) { - if (__thread_id_map[i] == tid) - break; - } - if (i >= NR_THREADS){ - fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", - (int)tid, (int)tid); - exit(-1); - } - if (pthread_join(tid, &vp) != 0) { - perror("wait_thread:pthread_join"); - exit(-1); - } - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - return vp; -} - -static void wait_all_threads(void) -{ - int i; - thread_id_t tid; - - for (i = 1; i < NR_THREADS; i++) { - tid = __thread_id_map[i]; - if (tid != __THREAD_ID_MAP_EMPTY && - tid != __THREAD_ID_MAP_WAITING) - (void)wait_thread(tid); - } -} - -static void run_on(int cpu) -{ - cpu_set_t mask; - - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * timekeeping -- very crude -- should use MONOTONIC... - */ - -long long get_microseconds(void) -{ - struct timeval tv; - - if (gettimeofday(&tv, NULL) != 0) - abort(); - return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; -} - -/* - * Per-thread variables. - */ - -#define DEFINE_PER_THREAD(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_thread_##name[NR_THREADS]; -#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) - -#define per_thread(name, thread) __per_thread_##name[thread].v -#define __get_thread_var(name) per_thread(name, smp_thread_id()) - -#define init_per_thread(name, v) \ - do { \ - int __i_p_t_i; \ - for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ - per_thread(name, __i_p_t_i) = v; \ - } while (0) - -/* - * CPU traversal primitives. - */ - -#ifndef NR_CPUS -#define NR_CPUS 16 -#endif /* #ifndef NR_CPUS */ - -#define for_each_possible_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) -#define for_each_online_cpu(cpu) \ - for (cpu = 0; cpu < NR_CPUS; cpu++) - -/* - * Per-CPU variables. - */ - -#define DEFINE_PER_CPU(type, name) \ - struct { \ - __typeof__(type) v \ - __attribute__((__aligned__(CACHE_LINE_SIZE))); \ - } __per_cpu_##name[NR_CPUS] -#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) - -DEFINE_PER_THREAD(int, smp_processor_id); - -#define per_cpu(name, thread) __per_cpu_##name[thread].v -#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) - -#define init_per_cpu(name, v) \ - do { \ - int __i_p_c_i; \ - for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ - per_cpu(name, __i_p_c_i) = v; \ - } while (0) - -/* - * CPU state checking (crowbarred). - */ - -#define idle_cpu(cpu) 0 -#define in_softirq() 1 -#define hardirq_count() 0 -#define PREEMPT_SHIFT 0 -#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) -#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) -#define PREEMPT_BITS 8 -#define SOFTIRQ_BITS 8 - -/* - * CPU hotplug. - */ - -struct notifier_block { - int (*notifier_call)(struct notifier_block *, unsigned long, void *); - struct notifier_block *next; - int priority; -}; - -#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ -#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ -#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ -#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ -#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ -#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ -#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, - * not handling interrupts, soon dead */ -#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug - * lock is dropped */ - -/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend - * operation in progress - */ -#define CPU_TASKS_FROZEN 0x0010 - -#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) -#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) -#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) -#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) -#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) -#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) -#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) - -/* Hibernation and suspend events */ -#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ -#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ -#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ -#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ -#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ -#define PM_POST_RESTORE 0x0006 /* Restore failed */ - -#define NOTIFY_DONE 0x0000 /* Don't care */ -#define NOTIFY_OK 0x0001 /* Suits me */ -#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ -#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) - /* Bad/Veto action */ -/* - * Clean way to return from the notifier and stop further calls. - */ -#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) - -/* - * Bug checks. - */ - -#define BUG_ON(c) do { if (!(c)) abort(); } while (0) - -/* - * Initialization -- Must be called before calling any primitives. - */ - -static void smp_init(void) -{ - int i; - - spin_lock_init(&__thread_id_map_mutex); - __thread_id_map[0] = pthread_self(); - for (i = 1; i < NR_THREADS; i++) - __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; - init_per_thread(smp_processor_id, 0); - if (pthread_key_create(&thread_id_key, NULL) != 0) { - perror("pthread_key_create"); - exit(-1); - } -} - -/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ - -#ifndef _LINUX_LIST_H -#define _LINUX_LIST_H - -#define LIST_POISON1 ((void *) 0x00100100) -#define LIST_POISON2 ((void *) 0x00200200) - -#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) -#define container_of(ptr, type, member) ({ \ - const typeof( ((type *)0)->member ) *__mptr = (ptr); \ - (type *)( (char *)__mptr - offsetof(type,member) );}) - -/* - * Simple doubly linked list implementation. - * - * Some of the internal functions ("__xxx") are useful when - * manipulating whole lists rather than single entries, as - * sometimes we already know the next/prev entries and we can - * generate better code by using them directly rather than - * using the generic single-entry routines. - */ - -struct list_head { - struct list_head *next, *prev; -}; - -#define LIST_HEAD_INIT(name) { &(name), &(name) } - -#define LIST_HEAD(name) \ - struct list_head name = LIST_HEAD_INIT(name) - -static inline void INIT_LIST_HEAD(struct list_head *list) -{ - list->next = list; - list->prev = list; -} - -/* - * Insert a new entry between two known consecutive entries. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -#ifndef CONFIG_DEBUG_LIST -static inline void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next) -{ - next->prev = new; - new->next = next; - new->prev = prev; - prev->next = new; -} -#else -extern void __list_add(struct list_head *new, - struct list_head *prev, - struct list_head *next); -#endif - -/** - * list_add - add a new entry - * @new: new entry to be added - * @head: list head to add it after - * - * Insert a new entry after the specified head. - * This is good for implementing stacks. - */ -static inline void list_add(struct list_head *new, struct list_head *head) -{ - __list_add(new, head, head->next); -} - - -/** - * list_add_tail - add a new entry - * @new: new entry to be added - * @head: list head to add it before - * - * Insert a new entry before the specified head. - * This is useful for implementing queues. - */ -static inline void list_add_tail(struct list_head *new, struct list_head *head) -{ - __list_add(new, head->prev, head); -} - -/* - * Delete a list entry by making the prev/next entries - * point to each other. - * - * This is only for internal list manipulation where we know - * the prev/next entries already! - */ -static inline void __list_del(struct list_head * prev, struct list_head * next) -{ - next->prev = prev; - prev->next = next; -} - -/** - * list_del - deletes entry from list. - * @entry: the element to delete from the list. - * Note: list_empty() on entry does not return true after this, the entry is - * in an undefined state. - */ -#ifndef CONFIG_DEBUG_LIST -static inline void list_del(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - entry->next = LIST_POISON1; - entry->prev = LIST_POISON2; -} -#else -extern void list_del(struct list_head *entry); -#endif - -/** - * list_replace - replace old entry by new one - * @old : the element to be replaced - * @new : the new element to insert - * - * If @old was empty, it will be overwritten. - */ -static inline void list_replace(struct list_head *old, - struct list_head *new) -{ - new->next = old->next; - new->next->prev = new; - new->prev = old->prev; - new->prev->next = new; -} - -static inline void list_replace_init(struct list_head *old, - struct list_head *new) -{ - list_replace(old, new); - INIT_LIST_HEAD(old); -} - -/** - * list_del_init - deletes entry from list and reinitialize it. - * @entry: the element to delete from the list. - */ -static inline void list_del_init(struct list_head *entry) -{ - __list_del(entry->prev, entry->next); - INIT_LIST_HEAD(entry); -} - -/** - * list_move - delete from one list and add as another's head - * @list: the entry to move - * @head: the head that will precede our entry - */ -static inline void list_move(struct list_head *list, struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add(list, head); -} - -/** - * list_move_tail - delete from one list and add as another's tail - * @list: the entry to move - * @head: the head that will follow our entry - */ -static inline void list_move_tail(struct list_head *list, - struct list_head *head) -{ - __list_del(list->prev, list->next); - list_add_tail(list, head); -} - -/** - * list_is_last - tests whether @list is the last entry in list @head - * @list: the entry to test - * @head: the head of the list - */ -static inline int list_is_last(const struct list_head *list, - const struct list_head *head) -{ - return list->next == head; -} - -/** - * list_empty - tests whether a list is empty - * @head: the list to test. - */ -static inline int list_empty(const struct list_head *head) -{ - return head->next == head; -} - -/** - * list_empty_careful - tests whether a list is empty and not being modified - * @head: the list to test - * - * Description: - * tests whether a list is empty _and_ checks that no other CPU might be - * in the process of modifying either member (next or prev) - * - * NOTE: using list_empty_careful() without synchronization - * can only be safe if the only activity that can happen - * to the list entry is list_del_init(). Eg. it cannot be used - * if another CPU could re-list_add() it. - */ -static inline int list_empty_careful(const struct list_head *head) -{ - struct list_head *next = head->next; - return (next == head) && (next == head->prev); -} - -/** - * list_is_singular - tests whether a list has just one entry. - * @head: the list to test. - */ -static inline int list_is_singular(const struct list_head *head) -{ - return !list_empty(head) && (head->next == head->prev); -} - -static inline void __list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - struct list_head *new_first = entry->next; - list->next = head->next; - list->next->prev = list; - list->prev = entry; - entry->next = list; - head->next = new_first; - new_first->prev = head; -} - -/** - * list_cut_position - cut a list into two - * @list: a new list to add all removed entries - * @head: a list with entries - * @entry: an entry within head, could be the head itself - * and if so we won't cut the list - * - * This helper moves the initial part of @head, up to and - * including @entry, from @head to @list. You should - * pass on @entry an element you know is on @head. @list - * should be an empty list or a list you do not care about - * losing its data. - * - */ -static inline void list_cut_position(struct list_head *list, - struct list_head *head, struct list_head *entry) -{ - if (list_empty(head)) - return; - if (list_is_singular(head) && - (head->next != entry && head != entry)) - return; - if (entry == head) - INIT_LIST_HEAD(list); - else - __list_cut_position(list, head, entry); -} - -static inline void __list_splice(const struct list_head *list, - struct list_head *prev, - struct list_head *next) -{ - struct list_head *first = list->next; - struct list_head *last = list->prev; - - first->prev = prev; - prev->next = first; - - last->next = next; - next->prev = last; -} - -/** - * list_splice - join two lists, this is designed for stacks - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice(const struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head, head->next); -} - -/** - * list_splice_tail - join two lists, each list being a queue - * @list: the new list to add. - * @head: the place to add it in the first list. - */ -static inline void list_splice_tail(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) - __list_splice(list, head->prev, head); -} - -/** - * list_splice_init - join two lists and reinitialise the emptied list. - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * The list at @list is reinitialised - */ -static inline void list_splice_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head, head->next); - INIT_LIST_HEAD(list); - } -} - -/** - * list_splice_tail_init - join two lists and reinitialise the emptied list - * @list: the new list to add. - * @head: the place to add it in the first list. - * - * Each of the lists is a queue. - * The list at @list is reinitialised - */ -static inline void list_splice_tail_init(struct list_head *list, - struct list_head *head) -{ - if (!list_empty(list)) { - __list_splice(list, head->prev, head); - INIT_LIST_HEAD(list); - } -} - -/** - * list_entry - get the struct for this entry - * @ptr: the &struct list_head pointer. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - */ -#define list_entry(ptr, type, member) \ - container_of(ptr, type, member) - -/** - * list_first_entry - get the first element from a list - * @ptr: the list head to take the element from. - * @type: the type of the struct this is embedded in. - * @member: the name of the list_struct within the struct. - * - * Note, that list is expected to be not empty. - */ -#define list_first_entry(ptr, type, member) \ - list_entry((ptr)->next, type, member) - -/** - * list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each(pos, head) \ - for (pos = (head)->next; prefetch(pos->next), pos != (head); \ - pos = pos->next) - -/** - * __list_for_each - iterate over a list - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - * - * This variant differs from list_for_each() in that it's the - * simplest possible list iteration code, no prefetching is done. - * Use this for code that knows the list to be very short (empty - * or 1 entry) most of the time. - */ -#define __list_for_each(pos, head) \ - for (pos = (head)->next; pos != (head); pos = pos->next) - -/** - * list_for_each_prev - iterate over a list backwards - * @pos: the &struct list_head to use as a loop cursor. - * @head: the head for your list. - */ -#define list_for_each_prev(pos, head) \ - for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ - pos = pos->prev) - -/** - * list_for_each_safe - iterate over a list safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_safe(pos, n, head) \ - for (pos = (head)->next, n = pos->next; pos != (head); \ - pos = n, n = pos->next) - -/** - * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry - * @pos: the &struct list_head to use as a loop cursor. - * @n: another &struct list_head to use as temporary storage - * @head: the head for your list. - */ -#define list_for_each_prev_safe(pos, n, head) \ - for (pos = (head)->prev, n = pos->prev; \ - prefetch(pos->prev), pos != (head); \ - pos = n, n = pos->prev) - -/** - * list_for_each_entry - iterate over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry(pos, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_reverse - iterate backwards over list of given type. - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_reverse(pos, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() - * @pos: the type * to use as a start point - * @head: the head of the list - * @member: the name of the list_struct within the struct. - * - * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). - */ -#define list_prepare_entry(pos, head, member) \ - ((pos) ? : list_entry(head, typeof(*pos), member)) - -/** - * list_for_each_entry_continue - continue iteration over list of given type - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Continue to iterate over list of given type, continuing after - * the current position. - */ -#define list_for_each_entry_continue(pos, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member); \ - prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_continue_reverse - iterate backwards from the given point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Start to iterate over list of given type backwards, continuing after - * the current position. - */ -#define list_for_each_entry_continue_reverse(pos, head, member) \ - for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ - prefetch(pos->member.prev), &pos->member != (head); \ - pos = list_entry(pos->member.prev, typeof(*pos), member)) - -/** - * list_for_each_entry_from - iterate over list of given type from the current point - * @pos: the type * to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing from current position. - */ -#define list_for_each_entry_from(pos, head, member) \ - for (; prefetch(pos->member.next), &pos->member != (head); \ - pos = list_entry(pos->member.next, typeof(*pos), member)) - -/** - * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - */ -#define list_for_each_entry_safe(pos, n, head, member) \ - for (pos = list_entry((head)->next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_continue - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type, continuing after current point, - * safe against removal of list entry. - */ -#define list_for_each_entry_safe_continue(pos, n, head, member) \ - for (pos = list_entry(pos->member.next, typeof(*pos), member), \ - n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_from - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate over list of given type from current point, safe against - * removal of list entry. - */ -#define list_for_each_entry_safe_from(pos, n, head, member) \ - for (n = list_entry(pos->member.next, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.next, typeof(*n), member)) - -/** - * list_for_each_entry_safe_reverse - * @pos: the type * to use as a loop cursor. - * @n: another type * to use as temporary storage - * @head: the head for your list. - * @member: the name of the list_struct within the struct. - * - * Iterate backwards over list of given type, safe against removal - * of list entry. - */ -#define list_for_each_entry_safe_reverse(pos, n, head, member) \ - for (pos = list_entry((head)->prev, typeof(*pos), member), \ - n = list_entry(pos->member.prev, typeof(*pos), member); \ - &pos->member != (head); \ - pos = n, n = list_entry(n->member.prev, typeof(*n), member)) - -/* - * Double linked lists with a single pointer list head. - * Mostly useful for hash tables where the two pointer list head is - * too wasteful. - * You lose the ability to access the tail in O(1). - */ - -struct hlist_head { - struct hlist_node *first; -}; - -struct hlist_node { - struct hlist_node *next, **pprev; -}; - -#define HLIST_HEAD_INIT { .first = NULL } -#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } -#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) -static inline void INIT_HLIST_NODE(struct hlist_node *h) -{ - h->next = NULL; - h->pprev = NULL; -} - -static inline int hlist_unhashed(const struct hlist_node *h) -{ - return !h->pprev; -} - -static inline int hlist_empty(const struct hlist_head *h) -{ - return !h->first; -} - -static inline void __hlist_del(struct hlist_node *n) -{ - struct hlist_node *next = n->next; - struct hlist_node **pprev = n->pprev; - *pprev = next; - if (next) - next->pprev = pprev; -} - -static inline void hlist_del(struct hlist_node *n) -{ - __hlist_del(n); - n->next = LIST_POISON1; - n->pprev = LIST_POISON2; -} - -static inline void hlist_del_init(struct hlist_node *n) -{ - if (!hlist_unhashed(n)) { - __hlist_del(n); - INIT_HLIST_NODE(n); - } -} - -static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) -{ - struct hlist_node *first = h->first; - n->next = first; - if (first) - first->pprev = &n->next; - h->first = n; - n->pprev = &h->first; -} - -/* next must be != NULL */ -static inline void hlist_add_before(struct hlist_node *n, - struct hlist_node *next) -{ - n->pprev = next->pprev; - n->next = next; - next->pprev = &n->next; - *(n->pprev) = n; -} - -static inline void hlist_add_after(struct hlist_node *n, - struct hlist_node *next) -{ - next->next = n->next; - n->next = next; - next->pprev = &n->next; - - if(next->next) - next->next->pprev = &next->next; -} - -/* - * Move a list from one list head to another. Fixup the pprev - * reference of the first entry if it exists. - */ -static inline void hlist_move_list(struct hlist_head *old, - struct hlist_head *new) -{ - new->first = old->first; - if (new->first) - new->first->pprev = &new->first; - old->first = NULL; -} - -#define hlist_entry(ptr, type, member) container_of(ptr,type,member) - -#define hlist_for_each(pos, head) \ - for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ - pos = pos->next) - -#define hlist_for_each_safe(pos, n, head) \ - for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ - pos = n) - -/** - * hlist_for_each_entry - iterate over list of given type - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry(tpos, pos, head, member) \ - for (pos = (head)->first; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_continue - iterate over a hlist continuing after current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_continue(tpos, pos, member) \ - for (pos = (pos)->next; \ - pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_from - iterate over a hlist continuing from current point - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_from(tpos, pos, member) \ - for (; pos && ({ prefetch(pos->next); 1;}) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = pos->next) - -/** - * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry - * @tpos: the type * to use as a loop cursor. - * @pos: the &struct hlist_node to use as a loop cursor. - * @n: another &struct hlist_node to use as temporary storage - * @head: the head for your list. - * @member: the name of the hlist_node within the struct. - */ -#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ - for (pos = (head)->first; \ - pos && ({ n = pos->next; 1; }) && \ - ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ - pos = n) - -#endif diff --git a/debug_yield.patch b/debug_yield.patch deleted file mode 100644 index 08c04aa..0000000 --- a/debug_yield.patch +++ /dev/null @@ -1,139 +0,0 @@ -diff --git a/urcu.c b/urcu.c -index 162ce00..018e09b 100644 ---- a/urcu.c -+++ b/urcu.c -@@ -91,24 +91,17 @@ static void force_mb_all_threads(void) - */ - if (!reader_data) - return; -- debug_yield_write(); - sig_done = 0; -- debug_yield_write(); - smp_mb(); /* write sig_done before sending the signals */ -- debug_yield_write(); -- for (index = reader_data; index < reader_data + num_readers; index++) { -+ for (index = reader_data; index < reader_data + num_readers; index++) - pthread_kill(index->tid, SIGURCU); -- debug_yield_write(); -- } - /* - * Wait for sighandler (and thus mb()) to execute on every thread. - * BUSY-LOOP. - */ - while (sig_done < num_readers) - barrier(); -- debug_yield_write(); - smp_mb(); /* read sig_done before ending the barrier */ -- debug_yield_write(); - } - #endif - -@@ -135,13 +128,10 @@ void synchronize_rcu(void) - * where new ptr points to. */ - /* Write new ptr before changing the qparity */ - force_mb_all_threads(); -- debug_yield_write(); - - internal_urcu_lock(); -- debug_yield_write(); - - switch_next_urcu_qparity(); /* 0 -> 1 */ -- debug_yield_write(); - - /* - * Must commit qparity update to memory before waiting for parity -@@ -155,7 +145,6 @@ void synchronize_rcu(void) - * Wait for previous parity to be empty of readers. - */ - wait_for_quiescent_state(); /* Wait readers in parity 0 */ -- debug_yield_write(); - - /* - * Must finish waiting for quiescent state for parity 0 before -@@ -166,7 +155,6 @@ void synchronize_rcu(void) - smp_mb(); - - switch_next_urcu_qparity(); /* 1 -> 0 */ -- debug_yield_write(); - - /* - * Must commit qparity update to memory before waiting for parity -@@ -180,17 +168,14 @@ void synchronize_rcu(void) - * Wait for previous parity to be empty of readers. - */ - wait_for_quiescent_state(); /* Wait readers in parity 1 */ -- debug_yield_write(); - - internal_urcu_unlock(); -- debug_yield_write(); - - /* All threads should finish using the data referred to by old ptr - * before decrementing their urcu_active_readers count */ - /* Finish waiting for reader threads before letting the old ptr being - * freed. */ - force_mb_all_threads(); -- debug_yield_write(); - } - - void urcu_add_reader(pthread_t id) -diff --git a/urcu.h b/urcu.h -index 92b31df..1b663c7 100644 ---- a/urcu.h -+++ b/urcu.h -@@ -219,13 +219,11 @@ static inline int rcu_old_gp_ongoing(long *value) - - if (value == NULL) - return 0; -- debug_yield_write(); - /* - * Make sure both tests below are done on the same version of *value - * to insure consistency. - */ - v = ACCESS_ONCE(*value); -- debug_yield_write(); - return (v & RCU_GP_CTR_NEST_MASK) && - ((v ^ urcu_gp_ctr) & RCU_GP_CTR_BIT); - } -@@ -234,34 +232,27 @@ static inline void rcu_read_lock(void) - { - long tmp; - -- debug_yield_read(); - tmp = urcu_active_readers; -- debug_yield_read(); - /* urcu_gp_ctr = RCU_GP_COUNT | (~RCU_GP_CTR_BIT or RCU_GP_CTR_BIT) */ - if (likely(!(tmp & RCU_GP_CTR_NEST_MASK))) - urcu_active_readers = urcu_gp_ctr; - else - urcu_active_readers = tmp + RCU_GP_COUNT; -- debug_yield_read(); - /* - * Increment active readers count before accessing the pointer. - * See force_mb_all_threads(). - */ - read_barrier(); -- debug_yield_read(); - } - - static inline void rcu_read_unlock(void) - { -- debug_yield_read(); - read_barrier(); -- debug_yield_read(); - /* - * Finish using rcu before decrementing the pointer. - * See force_mb_all_threads(). - */ - urcu_active_readers -= RCU_GP_COUNT; -- debug_yield_read(); - } - - /** -@@ -302,7 +293,6 @@ extern void synchronize_rcu(void); - #define urcu_publish_content(p, v) \ - ({ \ - void *oldptr; \ -- debug_yield_write(); \ - oldptr = rcu_xchg_pointer(p, v); \ - synchronize_rcu(); \ - oldptr; \ diff --git a/rcutorture.h b/rcutorture.h deleted file mode 100644 index 00b0b31..0000000 --- a/rcutorture.h +++ /dev/null @@ -1,417 +0,0 @@ -/* - * rcutorture.h: simple user-level performance/stress test of RCU. - * - * Usage: - * ./rcu rperf [ ] - * Run a read-side performance test with the specified - * number of readers spaced by . - * Thus "./rcu 16 rperf 2" would run 16 readers on even-numbered - * CPUs from 0 to 30. - * ./rcu uperf [ ] - * Run an update-side performance test with the specified - * number of updaters and specified CPU spacing. - * ./rcu perf [ ] - * Run a combined read/update performance test with the specified - * number of readers and one updater and specified CPU spacing. - * The readers run on the low-numbered CPUs and the updater - * of the highest-numbered CPU. - * - * The above tests produce output as follows: - * - * n_reads: 46008000 n_updates: 146026 nreaders: 2 nupdaters: 1 duration: 1 - * ns/read: 43.4707 ns/update: 6848.1 - * - * The first line lists the total number of RCU reads and updates executed - * during the test, the number of reader threads, the number of updater - * threads, and the duration of the test in seconds. The second line - * lists the average duration of each type of operation in nanoseconds, - * or "nan" if the corresponding type of operation was not performed. - * - * ./rcu stress - * Run a stress test with the specified number of readers and - * one updater. None of the threads are affinitied to any - * particular CPU. - * - * This test produces output as follows: - * - * n_reads: 114633217 n_updates: 3903415 n_mberror: 0 - * rcu_stress_count: 114618391 14826 0 0 0 0 0 0 0 0 0 - * - * The first line lists the number of RCU read and update operations - * executed, followed by the number of memory-ordering violations - * (which will be zero in a correct RCU implementation). The second - * line lists the number of readers observing progressively more stale - * data. A correct RCU implementation will have all but the first two - * numbers non-zero. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright (c) 2008 Paul E. McKenney, IBM Corporation. - */ - -/* - * Test variables. - */ - -DEFINE_PER_THREAD(long long, n_reads_pt); -DEFINE_PER_THREAD(long long, n_updates_pt); - -long long n_reads = 0LL; -long n_updates = 0L; -atomic_t nthreadsrunning; -char argsbuf[64]; - -#define GOFLAG_INIT 0 -#define GOFLAG_RUN 1 -#define GOFLAG_STOP 2 - -int goflag __attribute__((__aligned__(CACHE_LINE_SIZE))) = GOFLAG_INIT; - -#define RCU_READ_RUN 1000 - -//MD -#define RCU_READ_NESTABLE - -#ifdef RCU_READ_NESTABLE -#define rcu_read_lock_nest() rcu_read_lock() -#define rcu_read_unlock_nest() rcu_read_unlock() -#else /* #ifdef RCU_READ_NESTABLE */ -#define rcu_read_lock_nest() -#define rcu_read_unlock_nest() -#endif /* #else #ifdef RCU_READ_NESTABLE */ - -#ifndef mark_rcu_quiescent_state -#define mark_rcu_quiescent_state() do ; while (0) -#endif /* #ifdef mark_rcu_quiescent_state */ - -#ifndef put_thread_offline -#define put_thread_offline() do ; while (0) -#define put_thread_online() do ; while (0) -#define put_thread_online_delay() do ; while (0) -#else /* #ifndef put_thread_offline */ -#define put_thread_online_delay() synchronize_rcu() -#endif /* #else #ifndef put_thread_offline */ - -/* - * Performance test. - */ - -void *rcu_read_perf_test(void *arg) -{ - int i; - int me = (long)arg; - long long n_reads_local = 0; - - rcu_register_thread(); - run_on(me); - atomic_inc(&nthreadsrunning); - while (goflag == GOFLAG_INIT) - poll(NULL, 0, 1); - mark_rcu_quiescent_state(); - while (goflag == GOFLAG_RUN) { - for (i = 0; i < RCU_READ_RUN; i++) { - rcu_read_lock(); - /* rcu_read_lock_nest(); */ - /* rcu_read_unlock_nest(); */ - rcu_read_unlock(); - } - n_reads_local += RCU_READ_RUN; - mark_rcu_quiescent_state(); - } - __get_thread_var(n_reads_pt) += n_reads_local; - put_thread_offline(); - rcu_unregister_thread(); - - return (NULL); -} - -void *rcu_update_perf_test(void *arg) -{ - long long n_updates_local = 0; - - atomic_inc(&nthreadsrunning); - while (goflag == GOFLAG_INIT) - poll(NULL, 0, 1); - while (goflag == GOFLAG_RUN) { - synchronize_rcu(); - n_updates_local++; - } - __get_thread_var(n_updates_pt) += n_updates_local; - return NULL; -} - -void perftestinit(void) -{ - init_per_thread(n_reads_pt, 0LL); - init_per_thread(n_updates_pt, 0LL); - atomic_set(&nthreadsrunning, 0); -} - -void perftestrun(int nthreads, int nreaders, int nupdaters) -{ - int t; - int duration = 1; - - smp_mb(); - while (atomic_read(&nthreadsrunning) < nthreads) - poll(NULL, 0, 1); - goflag = GOFLAG_RUN; - smp_mb(); - sleep(duration); - smp_mb(); - goflag = GOFLAG_STOP; - smp_mb(); - wait_all_threads(); - for_each_thread(t) { - n_reads += per_thread(n_reads_pt, t); - n_updates += per_thread(n_updates_pt, t); - } - printf("n_reads: %lld n_updates: %ld nreaders: %d nupdaters: %d duration: %d\n", - n_reads, n_updates, nreaders, nupdaters, duration); - printf("ns/read: %g ns/update: %g\n", - ((duration * 1000*1000*1000.*(double)nreaders) / - (double)n_reads), - ((duration * 1000*1000*1000.*(double)nupdaters) / - (double)n_updates)); - exit(0); -} - -void perftest(int nreaders, int cpustride) -{ - int i; - long arg; - - perftestinit(); - for (i = 0; i < nreaders; i++) { - arg = (long)(i * cpustride); - create_thread(rcu_read_perf_test, (void *)arg); - } - arg = (long)(i * cpustride); - create_thread(rcu_update_perf_test, (void *)arg); - perftestrun(i + 1, nreaders, 1); -} - -void rperftest(int nreaders, int cpustride) -{ - int i; - long arg; - - perftestinit(); - init_per_thread(n_reads_pt, 0LL); - for (i = 0; i < nreaders; i++) { - arg = (long)(i * cpustride); - create_thread(rcu_read_perf_test, (void *)arg); - } - perftestrun(i, nreaders, 0); -} - -void uperftest(int nupdaters, int cpustride) -{ - int i; - long arg; - - perftestinit(); - init_per_thread(n_reads_pt, 0LL); - for (i = 0; i < nupdaters; i++) { - arg = (long)(i * cpustride); - create_thread(rcu_update_perf_test, (void *)arg); - } - perftestrun(i, 0, nupdaters); -} - -/* - * Stress test. - */ - -#define RCU_STRESS_PIPE_LEN 10 - -struct rcu_stress { - int pipe_count; - int mbtest; -}; - -struct rcu_stress rcu_stress_array[RCU_STRESS_PIPE_LEN] = { { 0 } }; -struct rcu_stress *rcu_stress_current; -int rcu_stress_idx = 0; - -int n_mberror = 0; -DEFINE_PER_THREAD(long long [RCU_STRESS_PIPE_LEN + 1], rcu_stress_count); - -int garbage = 0; - -void *rcu_read_stress_test(void *arg) -{ - int i; - int itercnt = 0; - struct rcu_stress *p; - int pc; - - rcu_register_thread(); - while (goflag == GOFLAG_INIT) - poll(NULL, 0, 1); - mark_rcu_quiescent_state(); - while (goflag == GOFLAG_RUN) { - rcu_read_lock(); - p = rcu_dereference(rcu_stress_current); - if (p->mbtest == 0) - n_mberror++; - rcu_read_lock_nest(); - for (i = 0; i < 100; i++) - garbage++; - rcu_read_unlock_nest(); - pc = p->pipe_count; - rcu_read_unlock(); - if ((pc > RCU_STRESS_PIPE_LEN) || (pc < 0)) - pc = RCU_STRESS_PIPE_LEN; - __get_thread_var(rcu_stress_count)[pc]++; - __get_thread_var(n_reads_pt)++; - mark_rcu_quiescent_state(); - if ((++itercnt % 0x1000) == 0) { - put_thread_offline(); - put_thread_online_delay(); - put_thread_online(); - } - } - put_thread_offline(); - rcu_unregister_thread(); - - return (NULL); -} - -void *rcu_update_stress_test(void *arg) -{ - int i; - struct rcu_stress *p; - - while (goflag == GOFLAG_INIT) - poll(NULL, 0, 1); - while (goflag == GOFLAG_RUN) { - i = rcu_stress_idx + 1; - if (i >= RCU_STRESS_PIPE_LEN) - i = 0; - p = &rcu_stress_array[i]; - p->mbtest = 0; - smp_mb(); - p->pipe_count = 0; - p->mbtest = 1; - rcu_assign_pointer(rcu_stress_current, p); - rcu_stress_idx = i; - for (i = 0; i < RCU_STRESS_PIPE_LEN; i++) - if (i != rcu_stress_idx) - rcu_stress_array[i].pipe_count++; - synchronize_rcu(); - n_updates++; - } - return NULL; -} - -void *rcu_fake_update_stress_test(void *arg) -{ - while (goflag == GOFLAG_INIT) - poll(NULL, 0, 1); - while (goflag == GOFLAG_RUN) { - synchronize_rcu(); - poll(NULL, 0, 1); - } - return NULL; -} - -void stresstest(int nreaders) -{ - int i; - int t; - long long *p; - long long sum; - - init_per_thread(n_reads_pt, 0LL); - for_each_thread(t) { - p = &per_thread(rcu_stress_count,t)[0]; - for (i = 0; i <= RCU_STRESS_PIPE_LEN; i++) - p[i] = 0LL; - } - rcu_stress_current = &rcu_stress_array[0]; - rcu_stress_current->pipe_count = 0; - rcu_stress_current->mbtest = 1; - for (i = 0; i < nreaders; i++) - create_thread(rcu_read_stress_test, NULL); - create_thread(rcu_update_stress_test, NULL); - for (i = 0; i < 5; i++) - create_thread(rcu_fake_update_stress_test, NULL); - smp_mb(); - goflag = GOFLAG_RUN; - smp_mb(); - sleep(10); - smp_mb(); - goflag = GOFLAG_STOP; - smp_mb(); - wait_all_threads(); - for_each_thread(t) - n_reads += per_thread(n_reads_pt, t); - printf("n_reads: %lld n_updates: %ld n_mberror: %d\n", - n_reads, n_updates, n_mberror); - printf("rcu_stress_count:"); - for (i = 0; i <= RCU_STRESS_PIPE_LEN; i++) { - sum = 0LL; - for_each_thread(t) { - sum += per_thread(rcu_stress_count, t)[i]; - } - printf(" %lld", sum); - } - printf("\n"); - exit(0); -} - -/* - * Mainprogram. - */ - -void usage(int argc, char *argv[]) -{ - fprintf(stderr, "Usage: %s [nreaders [ perf | stress ] ]\n", argv[0]); - exit(-1); -} - -int main(int argc, char *argv[]) -{ - int nreaders = 1; - int cpustride = 1; - - smp_init(); - //rcu_init(); - -#ifdef DEBUG_YIELD - yield_active |= YIELD_READ; - yield_active |= YIELD_WRITE; -#endif - - if (argc > 1) { - nreaders = strtoul(argv[1], NULL, 0); - if (argc == 2) - perftest(nreaders, cpustride); - if (argc > 3) - cpustride = strtoul(argv[3], NULL, 0); - if (strcmp(argv[2], "perf") == 0) - perftest(nreaders, cpustride); - else if (strcmp(argv[2], "rperf") == 0) - rperftest(nreaders, cpustride); - else if (strcmp(argv[2], "uperf") == 0) - uperftest(nreaders, cpustride); - else if (strcmp(argv[2], "stress") == 0) - stresstest(nreaders); - usage(argc, argv); - } - perftest(nreaders, cpustride); - return 0; -} diff --git a/runall.sh b/runall.sh deleted file mode 100755 index e2b47dc..0000000 --- a/runall.sh +++ /dev/null @@ -1,101 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -NUM_CPUS=8 - -#extra options, e.g. for setting affinity on even CPUs : -#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -#stride 1 -#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -#stride 2 -#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -#stride 4 -#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -#stride 8 -#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) - -#Vary update fraction -#x: vary update fraction from 0 to 0.0001 - #fix number of readers and reader C.S. length, vary delay between updates -#y: ops/s - -rm -f runall.log -rm -fr runall.detail.log - - -echo Executing batch RCU test - -DURATION=10 -BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 - 131072 262144" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f batch-rcu.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for BATCH_SIZE in ${BATCH_ARRAY}; do - echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log - ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log -done - -#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON TEST ABOVE ** -EXTRA_OPTS+="-b 32768" - -echo Executing update fraction test - -DURATION=10 -WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 - 65536 131072 262144 524288 1048576 2097152 4194304 8388608 - 16777216 33554432 67108864 134217728" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f update-fraction.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for WDELAY in ${WDELAY_ARRAY}; do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log -done - -#Test scalability : -# x: vary number of readers from 0 to num cpus -# y: ops/s -# 0 writer. - -echo Executing scalability test - -NR_WRITERS=0 -DURATION=10 - -rm -f scalability.log - -for NR_READERS in $(seq 1 ${NUM_CPUS}); do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log -done - - -# x: Vary reader C.S. length from 0 to 100 us -# y: ops/s -# 8 readers -# 0 writers - -echo Executing reader C.S. length test - -NR_READERS=${NUM_CPUS} -NR_WRITERS=0 -DURATION=10 -#in loops. -READERCSLEN_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152" - -rm -f readercslen.log - -for READERCSLEN in ${READERCSLEN_ARRAY}; do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log -done diff --git a/runpaul-phase1.sh b/runpaul-phase1.sh deleted file mode 100755 index d2c8649..0000000 --- a/runpaul-phase1.sh +++ /dev/null @@ -1,44 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -NUM_CPUS=64 - -#extra options, e.g. for setting affinity on even CPUs : -EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -#stride 1 -#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -#stride 2 -#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -#stride 4 -#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -#stride 8 -#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) - -#Vary update fraction -#x: vary update fraction from 0 to 0.0001 - #fix number of readers and reader C.S. length, vary delay between updates -#y: ops/s - -rm -f runall.log -rm -fr runall.detail.log - - -echo Executing batch RCU test - -DURATION=10 -BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 - 131072 262144" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f batch-rcu.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for BATCH_SIZE in ${BATCH_ARRAY}; do - echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log - ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log -done diff --git a/runpaul-phase2.sh b/runpaul-phase2.sh deleted file mode 100755 index 0f6bd91..0000000 --- a/runpaul-phase2.sh +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -NUM_CPUS=64 - -#extra options, e.g. for setting affinity on even CPUs : -EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -#stride 1 -#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -#stride 2 -#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -#stride 4 -#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -#stride 8 -#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) - -#Vary update fraction -#x: vary update fraction from 0 to 0.0001 - #fix number of readers and reader C.S. length, vary delay between updates -#y: ops/s - -rm -f runall.log -rm -fr runall.detail.log - -#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON PHASE 1 RESULT ** -EXTRA_OPTS+="-b 32768" - -echo Executing update fraction test - -DURATION=10 -WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 - 65536 131072 262144 524288 1048576 2097152 4194304 8388608 - 16777216 33554432 67108864 134217728" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f update-fraction.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for WDELAY in ${WDELAY_ARRAY}; do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log -done diff --git a/runpaul-phase3.sh b/runpaul-phase3.sh deleted file mode 100755 index 1956062..0000000 --- a/runpaul-phase3.sh +++ /dev/null @@ -1,50 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -NUM_CPUS=64 - -#extra options, e.g. for setting affinity on even CPUs : -EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -#stride 1 -#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -#stride 2 -#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -#stride 4 -#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -#stride 8 -#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) - -#Vary update fraction -#x: vary update fraction from 0 to 0.0001 - #fix number of readers and reader C.S. length, vary delay between updates -#y: ops/s - -rm -f runall.log -rm -fr runall.detail.log - -#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON PHASE 1 RESULT ** -EXTRA_OPTS+="-b 32768" - -#Test scalability : -# x: vary number of readers from 0 to num cpus -# y: ops/s -# 0 writer. - -echo Executing scalability test - -NR_WRITERS=0 -DURATION=10 - -rm -f scalability.log - -for NR_READERS in $(seq 1 ${NUM_CPUS}); do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log -done - - diff --git a/runpaul-phase4.sh b/runpaul-phase4.sh deleted file mode 100755 index ede402c..0000000 --- a/runpaul-phase4.sh +++ /dev/null @@ -1,40 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -export NUM_CPUS=8 - -#extra options, e.g. for setting affinity on even CPUs : -#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -rm -f *.log - -#stride 1 -export EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -sh subphase4.sh $* -mkdir ppc64-8cores-stride1 -mv *.log ppc64-8cores-stride1/ - - -#stride 2 -export EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -sh subphase4.sh $* -mkdir ppc64-8cores-stride2 -mv *.log ppc64-8cores-stride2/ - - -#stride 4 -export EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -sh subphase4.sh $* -mkdir ppc64-8cores-stride4 -mv *.log ppc64-8cores-stride4/ - - -#stride 8 -export EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) -sh subphase4.sh $* -mkdir ppc64-8cores-stride8 -mv *.log ppc64-8cores-stride8/ diff --git a/runpaul-phase5.sh b/runpaul-phase5.sh deleted file mode 100644 index bb4bfe7..0000000 --- a/runpaul-phase5.sh +++ /dev/null @@ -1,11 +0,0 @@ -# test run after write-size update - -sh runpaul-phase1.sh -mkdir runpaul-phase1 -mv *.log runpaul-phase1/ - -sh runpaul-phase2.sh -mkdir runpaul-phase2 -mv *.log runpaul-phase2/ - -sh runpaul-phase4.sh diff --git a/runpaul-phase6.sh b/runpaul-phase6.sh deleted file mode 100644 index 5f65072..0000000 --- a/runpaul-phase6.sh +++ /dev/null @@ -1,7 +0,0 @@ -sh runpaul-phase1.sh -mkdir runpaul-phase1 -mv *.log runpaul-phase1/ - -sh runpaul-phase2.sh -mkdir runpaul-phase2 -mv *.log runpaul-phase2/ diff --git a/runtests-batch.sh b/runtests-batch.sh deleted file mode 100755 index 67e019b..0000000 --- a/runtests-batch.sh +++ /dev/null @@ -1,8 +0,0 @@ -#!/bin/sh - -#for a in test_urcu_gc test_urcu_gc_mb test_qsbr_gc; do -for a in test_urcu_gc; do - echo "./${a} $*" | tee -a runall.detail.log - /usr/bin/time --append --output runall.detail.log ./${a} $* -done - diff --git a/runtests.sh b/runtests.sh deleted file mode 100755 index 981aef1..0000000 --- a/runtests.sh +++ /dev/null @@ -1,10 +0,0 @@ -#!/bin/sh - -for a in test_urcu_gc test_urcu_gc_mb test_urcu test_urcu_mb \ - test_urcu_lgc test_qsbr_lgc test_urcu_lgc_mb \ - test_qsbr test_qsbr_gc test_rwlock test_perthreadlock \ - test_mutex; do - echo "./${a} $*" | tee -a runall.detail.log - /usr/bin/time --append --output runall.detail.log ./${a} $* -done - diff --git a/subphase4.sh b/subphase4.sh deleted file mode 100755 index b69ffbf..0000000 --- a/subphase4.sh +++ /dev/null @@ -1,101 +0,0 @@ -#!/bin/sh - -#run all tests - -#set to number of active CPUS -#NUM_CPUS=8 - -#extra options, e.g. for setting affinity on even CPUs : -#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) - -#ppc64 striding, use with NUM_CPUS=8 - -#stride 1 -#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) -#stride 2 -#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) -#stride 4 -#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) -#stride 8 -#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) - -#Vary update fraction -#x: vary update fraction from 0 to 0.0001 - #fix number of readers and reader C.S. length, vary delay between updates -#y: ops/s - -rm -f runall.log -rm -fr runall.detail.log - - -echo Executing batch RCU test - -DURATION=10 -BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 - 131072 262144" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f batch-rcu.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for BATCH_SIZE in ${BATCH_ARRAY}; do - echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log - ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log -done - -#setting gc each 4096. ** UPDATE FOR YOUR ARCHITECTURE BASED ON TEST ABOVE ** -EXTRA_OPTS+="-b 32768" - -echo Executing update fraction test - -DURATION=10 -WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 - 65536 131072 262144 524288 1048576 2097152 4194304 8388608 - 16777216 33554432 67108864 134217728" -NR_WRITERS=$((${NUM_CPUS} / 2)) - -rm -f update-fraction.log - -NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) -for WDELAY in ${WDELAY_ARRAY}; do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log -done - -#Test scalability : -# x: vary number of readers from 0 to num cpus -# y: ops/s -# 0 writer. - -echo Executing scalability test - -NR_WRITERS=0 -DURATION=10 - -rm -f scalability.log - -for NR_READERS in $(seq 1 ${NUM_CPUS}); do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log -done - - -# x: Vary reader C.S. length from 0 to 100 us -# y: ops/s -# 8 readers -# 0 writers - -echo Executing reader C.S. length test - -NR_READERS=${NUM_CPUS} -NR_WRITERS=0 -DURATION=10 -#in loops. -READERCSLEN_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152" - -rm -f readercslen.log - -for READERCSLEN in ${READERCSLEN_ARRAY}; do - echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log" >> runall.log - ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log -done diff --git a/test_looplen.c b/test_looplen.c deleted file mode 100644 index e07409a..0000000 --- a/test_looplen.c +++ /dev/null @@ -1,89 +0,0 @@ -/* - * test_looplen.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -#define LOOPS 1048576 -#define TESTS 10 - -int main(int argc, char **argv) -{ - unsigned long i; - cycles_t time1, time2; - cycles_t time_tot = 0; - double cpl; - - for (i = 0; i < TESTS; i++) { - time1 = get_cycles(); - loop_sleep(LOOPS); - time2 = get_cycles(); - time_tot += time2 - time1; - } - cpl = ((double)time_tot) / (double)TESTS / (double)LOOPS; - - printf("CALIBRATION : %g cycles per loop\n", cpl); - printf("time_tot = %llu, LOOPS = %d, TESTS = %d\n", - time_tot, LOOPS, TESTS); - - return 0; -} diff --git a/test_mutex.c b/test_mutex.c deleted file mode 100644 index e824980..0000000 --- a/test_mutex.c +++ /dev/null @@ -1,392 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -struct test_array { - int a; -}; - -static pthread_mutex_t lock; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static volatile struct test_array test_array = { 8 }; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -void *thr_reader(void *data) -{ - unsigned long tidx = (unsigned long)data; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - - for (;;) { - pthread_mutex_lock(&lock); - assert(test_array.a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - pthread_mutex_unlock(&lock); - nr_reads++; - if (unlikely(!test_duration_read())) - break; - } - - tot_nr_reads[tidx] = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *data) -{ - unsigned long wtidx = (unsigned long)data; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - pthread_mutex_lock(&lock); - test_array.a = 0; - test_array.a = 8; - pthread_mutex_unlock(&lock); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - tot_nr_writes[wtidx] = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader, *count_writer; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - smp_mb(); - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - count_writer = malloc(sizeof(*count_writer) * nr_writers); - tot_nr_reads = malloc(sizeof(*tot_nr_reads) * nr_readers); - tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += tot_nr_reads[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += tot_nr_writes[i]; - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes); - - free(tid_reader); - free(tid_writer); - free(count_reader); - free(count_writer); - free(tot_nr_reads); - free(tot_nr_writes); - return 0; -} diff --git a/test_perthreadlock.c b/test_perthreadlock.c deleted file mode 100644 index 31b7721..0000000 --- a/test_perthreadlock.c +++ /dev/null @@ -1,403 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -struct test_array { - int a; -}; - -struct per_thread_lock { - pthread_mutex_t lock; -} __attribute__((aligned(CACHE_LINE_SIZE))); /* cache-line aligned */ - -static struct per_thread_lock *per_thread_lock; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static volatile struct test_array test_array = { 8 }; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -void *thr_reader(void *data) -{ - unsigned long tidx = (unsigned long)data; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - - for (;;) { - pthread_mutex_lock(&per_thread_lock[tidx].lock); - assert(test_array.a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - pthread_mutex_unlock(&per_thread_lock[tidx].lock); - nr_reads++; - if (unlikely(!test_duration_read())) - break; - } - - tot_nr_reads[tidx] = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *data) -{ - unsigned long wtidx = (unsigned long)data; - long tidx; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - for (tidx = 0; tidx < nr_readers; tidx++) { - pthread_mutex_lock(&per_thread_lock[tidx].lock); - } - test_array.a = 0; - test_array.a = 8; - for (tidx = (long)nr_readers - 1; tidx >= 0; tidx--) { - pthread_mutex_unlock(&per_thread_lock[tidx].lock); - } - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - tot_nr_writes[wtidx] = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader, *count_writer; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - smp_mb(); - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - count_writer = malloc(sizeof(*count_writer) * nr_writers); - tot_nr_reads = malloc(sizeof(*tot_nr_reads) * nr_readers); - tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); - per_thread_lock = malloc(sizeof(*per_thread_lock) * nr_readers); - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += tot_nr_reads[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += tot_nr_writes[i]; - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes); - - free(tid_reader); - free(tid_writer); - free(count_reader); - free(count_writer); - free(tot_nr_reads); - free(tot_nr_writes); - free(per_thread_lock); - return 0; -} diff --git a/test_perthreadlock_timing.c b/test_perthreadlock_timing.c deleted file mode 100644 index 3e8268b..0000000 --- a/test_perthreadlock_timing.c +++ /dev/null @@ -1,213 +0,0 @@ -/* - * test_perthreadloc_timing.c - * - * Per thread locks - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#include "urcu.h" - -struct test_array { - int a; -}; - -static struct test_array test_array = { 8 }; - -struct per_thread_lock { - pthread_mutex_t lock; -} __attribute__((aligned(CACHE_LINE_SIZE))); /* cache-line aligned */ - -static struct per_thread_lock *per_thread_lock; - -#define OUTER_READ_LOOP 200U -#define INNER_READ_LOOP 100000U -#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) - -#define OUTER_WRITE_LOOP 10U -#define INNER_WRITE_LOOP 200U -#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) - -static int num_read; -static int num_write; - -#define NR_READ num_read -#define NR_WRITE num_write - -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; - -void *thr_reader(void *arg) -{ - int i, j; - cycles_t time1, time2; - long tidx = (long)arg; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - sleep(2); - - time1 = get_cycles(); - for (i = 0; i < OUTER_READ_LOOP; i++) { - for (j = 0; j < INNER_READ_LOOP; j++) { - pthread_mutex_lock(&per_thread_lock[tidx].lock); - assert(test_array.a == 8); - pthread_mutex_unlock(&per_thread_lock[tidx].lock); - } - } - time2 = get_cycles(); - - reader_time[tidx] = time2 - time1; - - sleep(2); - printf("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *arg) -{ - int i, j; - long tidx; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - sleep(2); - - for (i = 0; i < OUTER_WRITE_LOOP; i++) { - for (j = 0; j < INNER_WRITE_LOOP; j++) { - time1 = get_cycles(); - for (tidx = 0; tidx < NR_READ; tidx++) { - pthread_mutex_lock(&per_thread_lock[tidx].lock); - } - test_array.a = 8; - for (tidx = NR_READ - 1; tidx >= 0; tidx--) { - pthread_mutex_unlock(&per_thread_lock[tidx].lock); - } - time2 = get_cycles(); - writer_time[(unsigned long)arg] += time2 - time1; - usleep(1); - } - } - - printf("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - return ((void*)2); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - int i; - cycles_t tot_rtime = 0; - cycles_t tot_wtime = 0; - - if (argc < 2) { - printf("Usage : %s nr_readers nr_writers\n", argv[0]); - exit(-1); - } - num_read = atoi(argv[1]); - num_write = atoi(argv[2]); - - reader_time = malloc(sizeof(*reader_time) * num_read); - writer_time = malloc(sizeof(*writer_time) * num_write); - tid_reader = malloc(sizeof(*tid_reader) * num_read); - tid_writer = malloc(sizeof(*tid_writer) * num_write); - - printf("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - per_thread_lock = malloc(sizeof(struct per_thread_lock) * NR_READ); - - for (i = 0; i < NR_READ; i++) { - pthread_mutex_init(&per_thread_lock[i].lock, NULL); - } - for (i = 0; i < NR_READ; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - sleep(10); - - for (i = 0; i < NR_READ; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_rtime += reader_time[i]; - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_wtime += writer_time[i]; - } - printf("Time per read : %g cycles\n", - (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); - printf("Time per write : %g cycles\n", - (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); - free(per_thread_lock); - - free(reader_time); - free(writer_time); - free(tid_reader); - free(tid_writer); - - return 0; -} diff --git a/test_qsbr.c b/test_qsbr.c deleted file mode 100644 index 97b918d..0000000 --- a/test_qsbr.c +++ /dev/null @@ -1,430 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#define _LGPL_SOURCE -#include "urcu-qsbr.h" - -struct test_array { - int a; -}; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static struct test_array *test_rcu_pointer; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -/* - * malloc/free are reusing memory areas too quickly, which does not let us - * test races appropriately. Use a large circular array for allocations. - * ARRAY_SIZE is larger than nr_writers, which insures we never run over our tail. - */ -#define ARRAY_SIZE (1048576 * nr_writers) -#define ARRAY_POISON 0xDEADBEEF -static int array_index; -static struct test_array *test_array; - -static struct test_array *test_array_alloc(void) -{ - struct test_array *ret; - int index; - - rcu_copy_mutex_lock(); - index = array_index % ARRAY_SIZE; - assert(test_array[index].a == ARRAY_POISON || - test_array[index].a == 0); - ret = &test_array[index]; - array_index++; - if (array_index == ARRAY_SIZE) - array_index = 0; - rcu_copy_mutex_unlock(); - return ret; -} - -static void test_array_free(struct test_array *ptr) -{ - if (!ptr) - return; - rcu_copy_mutex_lock(); - ptr->a = ARRAY_POISON; - rcu_copy_mutex_unlock(); -} - -void *thr_reader(void *_count) -{ - unsigned long long *count = _count; - struct test_array *local_ptr; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - rcu_register_thread(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - _rcu_read_lock(); - local_ptr = _rcu_dereference(test_rcu_pointer); - debug_yield_read(); - if (local_ptr) - assert(local_ptr->a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - _rcu_read_unlock(); - nr_reads++; - /* QS each 1024 reads */ - if (unlikely((nr_reads & ((1 << 10) - 1)) == 0)) - _rcu_quiescent_state(); - if (unlikely(!test_duration_read())) - break; - } - - rcu_unregister_thread(); - - *count = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *_count) -{ - unsigned long long *count = _count; - struct test_array *new, *old; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - new = test_array_alloc(); - new->a = 8; - old = _rcu_publish_content(&test_rcu_pointer, new); - /* can be done after unlock */ - if (old) - old->a = 0; - test_array_free(old); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - *count = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader, *count_writer; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - test_array = malloc(sizeof(*test_array) * ARRAY_SIZE); - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - count_writer = malloc(sizeof(*count_writer) * nr_writers); - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - &count_reader[i]); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - &count_writer[i]); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += count_reader[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += count_writer[i]; - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes); - test_array_free(test_rcu_pointer); - free(test_array); - free(tid_reader); - free(tid_writer); - free(count_reader); - free(count_writer); - return 0; -} diff --git a/test_qsbr_gc.c b/test_qsbr_gc.c deleted file mode 100644 index 3a66a84..0000000 --- a/test_qsbr_gc.c +++ /dev/null @@ -1,461 +0,0 @@ -/* - * test_urcu_gc.c - * - * Userspace RCU library - test program (with baatch reclamation) - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#define _LGPL_SOURCE -#include "urcu-qsbr.h" - -struct test_array { - int a; -}; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static struct test_array *test_rcu_pointer; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; -static unsigned int reclaim_batch = 1; - -struct reclaim_queue { - void **queue; /* Beginning of queue */ - void **head; /* Insert position */ -}; - -static struct reclaim_queue *pending_reclaims; - - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; - - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -void *thr_reader(void *_count) -{ - unsigned long long *count = _count; - struct test_array *local_ptr; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - rcu_register_thread(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - _rcu_read_lock(); - local_ptr = _rcu_dereference(test_rcu_pointer); - debug_yield_read(); - if (local_ptr) - assert(local_ptr->a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - _rcu_read_unlock(); - nr_reads++; - /* QS each 1024 reads */ - if (unlikely((nr_reads & ((1 << 10) - 1)) == 0)) - _rcu_quiescent_state(); - if (unlikely(!test_duration_read())) - break; - } - - rcu_unregister_thread(); - - *count = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -static void rcu_gc_clear_queue(unsigned long wtidx) -{ - void **p; - - /* Wait for Q.S and empty queue */ - synchronize_rcu(); - - for (p = pending_reclaims[wtidx].queue; - p < pending_reclaims[wtidx].head; p++) { - /* poison */ - if (*p) - ((struct test_array *)*p)->a = 0; - free(*p); - } - pending_reclaims[wtidx].head = pending_reclaims[wtidx].queue; -} - -/* Using per-thread queue */ -static void rcu_gc_reclaim(unsigned long wtidx, void *old) -{ - /* Queue pointer */ - *pending_reclaims[wtidx].head = old; - pending_reclaims[wtidx].head++; - - if (likely(pending_reclaims[wtidx].head - pending_reclaims[wtidx].queue - < reclaim_batch)) - return; - - rcu_gc_clear_queue(wtidx); -} - -void *thr_writer(void *data) -{ - unsigned long wtidx = (unsigned long)data; -#ifdef TEST_LOCAL_GC - struct test_array *old = NULL; -#else - struct test_array *new, *old; -#endif - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { -#ifndef TEST_LOCAL_GC - new = malloc(sizeof(*new)); - new->a = 8; - old = _rcu_xchg_pointer(&test_rcu_pointer, new); -#endif - rcu_gc_reclaim(wtidx, old); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - tot_nr_writes[wtidx] = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'b': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - reclaim_batch = atol(argv[++i]); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); - pending_reclaims = malloc(sizeof(*pending_reclaims) * nr_writers); - if (reclaim_batch * sizeof(*pending_reclaims[i].queue) - < CACHE_LINE_SIZE) - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].queue = calloc(1, CACHE_LINE_SIZE); - else - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].queue = calloc(reclaim_batch, - sizeof(*pending_reclaims[i].queue)); - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].head = pending_reclaims[i].queue; - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - &count_reader[i]); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += count_reader[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += tot_nr_writes[i]; - rcu_gc_clear_queue(i); - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu " - "batch %u\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes, reclaim_batch); - free(tid_reader); - free(tid_writer); - free(count_reader); - free(tot_nr_writes); - for (i = 0; i < nr_writers; i++) - free(pending_reclaims[i].queue); - free(pending_reclaims); - - return 0; -} diff --git a/test_qsbr_timing.c b/test_qsbr_timing.c deleted file mode 100644 index e46a435..0000000 --- a/test_qsbr_timing.c +++ /dev/null @@ -1,240 +0,0 @@ -/* - * test_qsbr_timing.c - * - * Userspace QSBR - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#define _LGPL_SOURCE -#include "urcu-qsbr.h" - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -struct test_array { - int a; -}; - -static struct test_array *test_rcu_pointer; - -#define OUTER_READ_LOOP 2000U -#define INNER_READ_LOOP 100000U -#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) - -#define OUTER_WRITE_LOOP 10U -#define INNER_WRITE_LOOP 200U -#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) - -static int num_read; -static int num_write; - -#define NR_READ num_read -#define NR_WRITE num_write - -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; - -void *thr_reader(void *arg) -{ - int i, j; - struct test_array *local_ptr; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - sleep(2); - - rcu_register_thread(); - - time1 = get_cycles(); - for (i = 0; i < OUTER_READ_LOOP; i++) { - for (j = 0; j < INNER_READ_LOOP; j++) { - _rcu_read_lock(); - local_ptr = _rcu_dereference(test_rcu_pointer); - if (local_ptr) { - assert(local_ptr->a == 8); - } - _rcu_read_unlock(); - } - _rcu_quiescent_state(); - } - time2 = get_cycles(); - - rcu_unregister_thread(); - - reader_time[(unsigned long)arg] = time2 - time1; - - sleep(2); - printf("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *arg) -{ - int i, j; - struct test_array *new, *old; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - sleep(2); - - for (i = 0; i < OUTER_WRITE_LOOP; i++) { - for (j = 0; j < INNER_WRITE_LOOP; j++) { - time1 = get_cycles(); - new = malloc(sizeof(struct test_array)); - rcu_copy_mutex_lock(); - old = test_rcu_pointer; - if (old) { - assert(old->a == 8); - } - new->a = 8; - old = _rcu_publish_content(&test_rcu_pointer, new); - rcu_copy_mutex_unlock(); - /* can be done after unlock */ - if (old) { - old->a = 0; - } - free(old); - time2 = get_cycles(); - writer_time[(unsigned long)arg] += time2 - time1; - usleep(1); - } - } - - printf("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - return ((void*)2); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - int i; - cycles_t tot_rtime = 0; - cycles_t tot_wtime = 0; - - if (argc < 2) { - printf("Usage : %s nr_readers nr_writers\n", argv[0]); - exit(-1); - } - num_read = atoi(argv[1]); - num_write = atoi(argv[2]); - - reader_time = malloc(sizeof(*reader_time) * num_read); - writer_time = malloc(sizeof(*writer_time) * num_write); - tid_reader = malloc(sizeof(*tid_reader) * num_read); - tid_writer = malloc(sizeof(*tid_writer) * num_write); - - printf("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - for (i = 0; i < NR_READ; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - sleep(10); - - for (i = 0; i < NR_READ; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_rtime += reader_time[i]; - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_wtime += writer_time[i]; - } - free(test_rcu_pointer); - printf("Time per read : %g cycles\n", - (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); - printf("Time per write : %g cycles\n", - (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); - - free(reader_time); - free(writer_time); - free(tid_reader); - free(tid_writer); - - return 0; -} diff --git a/test_rwlock.c b/test_rwlock.c deleted file mode 100644 index 4cc4c07..0000000 --- a/test_rwlock.c +++ /dev/null @@ -1,383 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -struct test_array { - int a; -}; - -pthread_rwlock_t lock = PTHREAD_RWLOCK_INITIALIZER; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static volatile struct test_array test_array = { 8 }; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -void *thr_reader(void *_count) -{ - unsigned long long *count = _count; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - - for (;;) { - pthread_rwlock_rdlock(&lock); - assert(test_array.a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - pthread_rwlock_unlock(&lock); - nr_reads++; - if (unlikely(!test_duration_read())) - break; - } - - *count = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *_count) -{ - unsigned long long *count = _count; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - pthread_rwlock_wrlock(&lock); - test_array.a = 0; - test_array.a = 8; - pthread_rwlock_unlock(&lock); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - *count = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader, *count_writer; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - smp_mb(); - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - count_writer = malloc(sizeof(*count_writer) * nr_writers); - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - &count_reader[i]); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - &count_writer[i]); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += count_reader[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += count_writer[i]; - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes); - - free(tid_reader); - free(tid_writer); - free(count_reader); - free(count_writer); - return 0; -} diff --git a/test_rwlock_timing.c b/test_rwlock_timing.c deleted file mode 100644 index 4884e4d..0000000 --- a/test_rwlock_timing.c +++ /dev/null @@ -1,197 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#include "urcu.h" - -struct test_array { - int a; -}; - -pthread_rwlock_t lock = PTHREAD_RWLOCK_INITIALIZER; - -static struct test_array test_array = { 8 }; - -#define OUTER_READ_LOOP 200U -#define INNER_READ_LOOP 100000U -#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) - -#define OUTER_WRITE_LOOP 10U -#define INNER_WRITE_LOOP 200U -#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) - -static int num_read; -static int num_write; - -#define NR_READ num_read -#define NR_WRITE num_write - -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; - -void *thr_reader(void *arg) -{ - int i, j; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - sleep(2); - - time1 = get_cycles(); - for (i = 0; i < OUTER_READ_LOOP; i++) { - for (j = 0; j < INNER_READ_LOOP; j++) { - pthread_rwlock_rdlock(&lock); - assert(test_array.a == 8); - pthread_rwlock_unlock(&lock); - } - } - time2 = get_cycles(); - - reader_time[(unsigned long)arg] = time2 - time1; - - sleep(2); - printf("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *arg) -{ - int i, j; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - sleep(2); - - for (i = 0; i < OUTER_WRITE_LOOP; i++) { - for (j = 0; j < INNER_WRITE_LOOP; j++) { - time1 = get_cycles(); - pthread_rwlock_wrlock(&lock); - test_array.a = 8; - pthread_rwlock_unlock(&lock); - time2 = get_cycles(); - writer_time[(unsigned long)arg] += time2 - time1; - usleep(1); - } - } - - printf("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - return ((void*)2); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - int i; - cycles_t tot_rtime = 0; - cycles_t tot_wtime = 0; - - if (argc < 2) { - printf("Usage : %s nr_readers nr_writers\n", argv[0]); - exit(-1); - } - num_read = atoi(argv[1]); - num_write = atoi(argv[2]); - - reader_time = malloc(sizeof(*reader_time) * num_read); - writer_time = malloc(sizeof(*writer_time) * num_write); - tid_reader = malloc(sizeof(*tid_reader) * num_read); - tid_writer = malloc(sizeof(*tid_writer) * num_write); - - printf("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - for (i = 0; i < NR_READ; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - sleep(10); - - for (i = 0; i < NR_READ; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_rtime += reader_time[i]; - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_wtime += writer_time[i]; - } - printf("Time per read : %g cycles\n", - (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); - printf("Time per write : %g cycles\n", - (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); - - free(reader_time); - free(writer_time); - free(tid_reader); - free(tid_writer); - - return 0; -} diff --git a/test_urcu.c b/test_urcu.c deleted file mode 100644 index 3feda89..0000000 --- a/test_urcu.c +++ /dev/null @@ -1,430 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -struct test_array { - int a; -}; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static struct test_array *test_rcu_pointer; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -/* - * malloc/free are reusing memory areas too quickly, which does not let us - * test races appropriately. Use a large circular array for allocations. - * ARRAY_SIZE is larger than nr_writers, which insures we never run over our tail. - */ -#define ARRAY_SIZE (1048576 * nr_writers) -#define ARRAY_POISON 0xDEADBEEF -static int array_index; -static struct test_array *test_array; - -static struct test_array *test_array_alloc(void) -{ - struct test_array *ret; - int index; - - rcu_copy_mutex_lock(); - index = array_index % ARRAY_SIZE; - assert(test_array[index].a == ARRAY_POISON || - test_array[index].a == 0); - ret = &test_array[index]; - array_index++; - if (array_index == ARRAY_SIZE) - array_index = 0; - rcu_copy_mutex_unlock(); - return ret; -} - -static void test_array_free(struct test_array *ptr) -{ - if (!ptr) - return; - rcu_copy_mutex_lock(); - ptr->a = ARRAY_POISON; - rcu_copy_mutex_unlock(); -} - -void *thr_reader(void *_count) -{ - unsigned long long *count = _count; - struct test_array *local_ptr; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - rcu_register_thread(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - rcu_read_lock(); - local_ptr = rcu_dereference(test_rcu_pointer); - debug_yield_read(); - if (local_ptr) - assert(local_ptr->a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - rcu_read_unlock(); - nr_reads++; - if (unlikely(!test_duration_read())) - break; - } - - rcu_unregister_thread(); - - *count = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *_count) -{ - unsigned long long *count = _count; - struct test_array *new, *old; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - new = test_array_alloc(); - new->a = 8; - old = rcu_publish_content(&test_rcu_pointer, new); - if (old) - old->a = 0; - test_array_free(old); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - *count = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader, *count_writer; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - test_array = malloc(sizeof(*test_array) * ARRAY_SIZE); - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - count_writer = malloc(sizeof(*count_writer) * nr_writers); - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - &count_reader[i]); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - &count_writer[i]); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += count_reader[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += count_writer[i]; - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes); - test_array_free(test_rcu_pointer); - free(test_array); - free(tid_reader); - free(tid_writer); - free(count_reader); - free(count_writer); - return 0; -} diff --git a/test_urcu_gc.c b/test_urcu_gc.c deleted file mode 100644 index 12ea181..0000000 --- a/test_urcu_gc.c +++ /dev/null @@ -1,462 +0,0 @@ -/* - * test_urcu_gc.c - * - * Userspace RCU library - test program (with baatch reclamation) - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#define _GNU_SOURCE -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "arch.h" - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -/* hardcoded number of CPUs */ -#define NR_CPUS 16384 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#ifndef DYNAMIC_LINK_TEST -#define _LGPL_SOURCE -#else -#define debug_yield_read() -#endif -#include "urcu.h" - -struct test_array { - int a; -}; - -static volatile int test_go, test_stop; - -static unsigned long wdelay; - -static struct test_array *test_rcu_pointer; - -static unsigned int reclaim_batch = 1; - -struct reclaim_queue { - void **queue; /* Beginning of queue */ - void **head; /* Insert position */ -}; - -static struct reclaim_queue *pending_reclaims; - -static unsigned long duration; - -/* read-side C.S. duration, in loops */ -static unsigned long rduration; - -static inline void loop_sleep(unsigned long l) -{ - while(l-- != 0) - cpu_relax(); -} - -static int verbose_mode; - -#define printf_verbose(fmt, args...) \ - do { \ - if (verbose_mode) \ - printf(fmt, args); \ - } while (0) - -static unsigned int cpu_affinities[NR_CPUS]; -static unsigned int next_aff = 0; -static int use_affinity = 0; - -pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; - -static void set_affinity(void) -{ - cpu_set_t mask; - int cpu; - int ret; - - if (!use_affinity) - return; - - ret = pthread_mutex_lock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } - cpu = cpu_affinities[next_aff++]; - ret = pthread_mutex_unlock(&affinity_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } - CPU_ZERO(&mask); - CPU_SET(cpu, &mask); - sched_setaffinity(0, sizeof(mask), &mask); -} - -/* - * returns 0 if test should end. - */ -static int test_duration_write(void) -{ - return !test_stop; -} - -static int test_duration_read(void) -{ - return !test_stop; -} - -static unsigned long long __thread nr_writes; -static unsigned long long __thread nr_reads; - -static -unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; - -static unsigned int nr_readers; -static unsigned int nr_writers; - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -void *thr_reader(void *_count) -{ - unsigned long long *count = _count; - struct test_array *local_ptr; - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - rcu_register_thread(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { - rcu_read_lock(); - local_ptr = rcu_dereference(test_rcu_pointer); - debug_yield_read(); - if (local_ptr) - assert(local_ptr->a == 8); - if (unlikely(rduration)) - loop_sleep(rduration); - rcu_read_unlock(); - nr_reads++; - if (unlikely(!test_duration_read())) - break; - } - - rcu_unregister_thread(); - - *count = nr_reads; - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -static void rcu_gc_clear_queue(unsigned long wtidx) -{ - void **p; - - /* Wait for Q.S and empty queue */ - synchronize_rcu(); - - for (p = pending_reclaims[wtidx].queue; - p < pending_reclaims[wtidx].head; p++) { - /* poison */ - if (*p) - ((struct test_array *)*p)->a = 0; - free(*p); - } - pending_reclaims[wtidx].head = pending_reclaims[wtidx].queue; -} - -/* Using per-thread queue */ -static void rcu_gc_reclaim(unsigned long wtidx, void *old) -{ - /* Queue pointer */ - *pending_reclaims[wtidx].head = old; - pending_reclaims[wtidx].head++; - - if (likely(pending_reclaims[wtidx].head - pending_reclaims[wtidx].queue - < reclaim_batch)) - return; - - rcu_gc_clear_queue(wtidx); -} - -void *thr_writer(void *data) -{ - unsigned long wtidx = (unsigned long)data; -#ifdef TEST_LOCAL_GC - struct test_array *old = NULL; -#else - struct test_array *new, *old; -#endif - - printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - - set_affinity(); - - while (!test_go) - { - } - smp_mb(); - - for (;;) { -#ifndef TEST_LOCAL_GC - new = malloc(sizeof(*new)); - new->a = 8; - old = rcu_xchg_pointer(&test_rcu_pointer, new); -#endif - rcu_gc_reclaim(wtidx, old); - nr_writes++; - if (unlikely(!test_duration_write())) - break; - if (unlikely(wdelay)) - loop_sleep(wdelay); - } - - printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - tot_nr_writes[wtidx] = nr_writes; - return ((void*)2); -} - -void show_usage(int argc, char **argv) -{ - printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); -#ifdef DEBUG_YIELD - printf(" [-r] [-w] (yield reader and/or writer)"); -#endif - printf(" [-d delay] (writer period (us))"); - printf(" [-c duration] (reader C.S. duration (in loops))"); - printf(" [-v] (verbose output)"); - printf(" [-a cpu#] [-a cpu#]... (affinity)"); - printf("\n"); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - unsigned long long *count_reader; - unsigned long long tot_reads = 0, tot_writes = 0; - int i, a; - - if (argc < 4) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[1], "%u", &nr_readers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[2], "%u", &nr_writers); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - err = sscanf(argv[3], "%lu", &duration); - if (err != 1) { - show_usage(argc, argv); - return -1; - } - - for (i = 4; i < argc; i++) { - if (argv[i][0] != '-') - continue; - switch (argv[i][1]) { -#ifdef DEBUG_YIELD - case 'r': - yield_active |= YIELD_READ; - break; - case 'w': - yield_active |= YIELD_WRITE; - break; -#endif - case 'a': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - a = atoi(argv[++i]); - cpu_affinities[next_aff++] = a; - use_affinity = 1; - printf_verbose("Adding CPU %d affinity\n", a); - break; - case 'b': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - reclaim_batch = atol(argv[++i]); - break; - case 'c': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - rduration = atol(argv[++i]); - break; - case 'd': - if (argc < i + 2) { - show_usage(argc, argv); - return -1; - } - wdelay = atol(argv[++i]); - break; - case 'v': - verbose_mode = 1; - break; - } - } - - printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", - duration, nr_readers, nr_writers); - printf_verbose("Writer delay : %lu loops.\n", wdelay); - printf_verbose("Reader duration : %lu loops.\n", rduration); - printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - tid_reader = malloc(sizeof(*tid_reader) * nr_readers); - tid_writer = malloc(sizeof(*tid_writer) * nr_writers); - count_reader = malloc(sizeof(*count_reader) * nr_readers); - tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); - pending_reclaims = malloc(sizeof(*pending_reclaims) * nr_writers); - if (reclaim_batch * sizeof(*pending_reclaims[i].queue) - < CACHE_LINE_SIZE) - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].queue = calloc(1, CACHE_LINE_SIZE); - else - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].queue = calloc(reclaim_batch, - sizeof(*pending_reclaims[i].queue)); - for (i = 0; i < nr_writers; i++) - pending_reclaims[i].head = pending_reclaims[i].queue; - - next_aff = 0; - - for (i = 0; i < nr_readers; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - &count_reader[i]); - if (err != 0) - exit(1); - } - for (i = 0; i < nr_writers; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - smp_mb(); - - test_go = 1; - - sleep(duration); - - test_stop = 1; - - for (i = 0; i < nr_readers; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_reads += count_reader[i]; - } - for (i = 0; i < nr_writers; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_writes += tot_nr_writes[i]; - rcu_gc_clear_queue(i); - } - - printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, - tot_writes); - printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " - "nr_writers %3u " - "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu " - "batch %u\n", - argv[0], duration, nr_readers, rduration, - nr_writers, wdelay, tot_reads, tot_writes, - tot_reads + tot_writes, reclaim_batch); - free(tid_reader); - free(tid_writer); - free(count_reader); - free(tot_nr_writes); - for (i = 0; i < nr_writers; i++) - free(pending_reclaims[i].queue); - free(pending_reclaims); - - return 0; -} diff --git a/test_urcu_timing.c b/test_urcu_timing.c deleted file mode 100644 index 8577b8f..0000000 --- a/test_urcu_timing.c +++ /dev/null @@ -1,239 +0,0 @@ -/* - * test_urcu.c - * - * Userspace RCU library - test program - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ -#define CACHE_LINE_SIZE 4096 - -#if defined(_syscall0) -_syscall0(pid_t, gettid) -#elif defined(__NR_gettid) -static inline pid_t gettid(void) -{ - return syscall(__NR_gettid); -} -#else -#warning "use pid as tid" -static inline pid_t gettid(void) -{ - return getpid(); -} -#endif - -#define _LGPL_SOURCE -#include "urcu.h" - -pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; - -void rcu_copy_mutex_lock(void) -{ - int ret; - ret = pthread_mutex_lock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex lock"); - exit(-1); - } -} - -void rcu_copy_mutex_unlock(void) -{ - int ret; - - ret = pthread_mutex_unlock(&rcu_copy_mutex); - if (ret) { - perror("Error in pthread mutex unlock"); - exit(-1); - } -} - -struct test_array { - int a; -}; - -static struct test_array *test_rcu_pointer; - -#define OUTER_READ_LOOP 2000U -#define INNER_READ_LOOP 100000U -#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) - -#define OUTER_WRITE_LOOP 10U -#define INNER_WRITE_LOOP 200U -#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) - -static int num_read; -static int num_write; - -#define NR_READ num_read -#define NR_WRITE num_write - -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; -static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; - -void *thr_reader(void *arg) -{ - int i, j; - struct test_array *local_ptr; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - sleep(2); - - rcu_register_thread(); - - time1 = get_cycles(); - for (i = 0; i < OUTER_READ_LOOP; i++) { - for (j = 0; j < INNER_READ_LOOP; j++) { - rcu_read_lock(); - local_ptr = rcu_dereference(test_rcu_pointer); - if (local_ptr) { - assert(local_ptr->a == 8); - } - rcu_read_unlock(); - } - } - time2 = get_cycles(); - - rcu_unregister_thread(); - - reader_time[(unsigned long)arg] = time2 - time1; - - sleep(2); - printf("thread_end %s, thread id : %lx, tid %lu\n", - "reader", pthread_self(), (unsigned long)gettid()); - return ((void*)1); - -} - -void *thr_writer(void *arg) -{ - int i, j; - struct test_array *new, *old; - cycles_t time1, time2; - - printf("thread_begin %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - sleep(2); - - for (i = 0; i < OUTER_WRITE_LOOP; i++) { - for (j = 0; j < INNER_WRITE_LOOP; j++) { - time1 = get_cycles(); - new = malloc(sizeof(struct test_array)); - rcu_copy_mutex_lock(); - old = test_rcu_pointer; - if (old) { - assert(old->a == 8); - } - new->a = 8; - old = rcu_publish_content(&test_rcu_pointer, new); - rcu_copy_mutex_unlock(); - /* can be done after unlock */ - if (old) { - old->a = 0; - } - free(old); - time2 = get_cycles(); - writer_time[(unsigned long)arg] += time2 - time1; - usleep(1); - } - } - - printf("thread_end %s, thread id : %lx, tid %lu\n", - "writer", pthread_self(), (unsigned long)gettid()); - return ((void*)2); -} - -int main(int argc, char **argv) -{ - int err; - pthread_t *tid_reader, *tid_writer; - void *tret; - int i; - cycles_t tot_rtime = 0; - cycles_t tot_wtime = 0; - - if (argc < 2) { - printf("Usage : %s nr_readers nr_writers\n", argv[0]); - exit(-1); - } - num_read = atoi(argv[1]); - num_write = atoi(argv[2]); - - reader_time = malloc(sizeof(*reader_time) * num_read); - writer_time = malloc(sizeof(*writer_time) * num_write); - tid_reader = malloc(sizeof(*tid_reader) * num_read); - tid_writer = malloc(sizeof(*tid_writer) * num_write); - - printf("thread %-6s, thread id : %lx, tid %lu\n", - "main", pthread_self(), (unsigned long)gettid()); - - for (i = 0; i < NR_READ; i++) { - err = pthread_create(&tid_reader[i], NULL, thr_reader, - (void *)(long)i); - if (err != 0) - exit(1); - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_create(&tid_writer[i], NULL, thr_writer, - (void *)(long)i); - if (err != 0) - exit(1); - } - - sleep(10); - - for (i = 0; i < NR_READ; i++) { - err = pthread_join(tid_reader[i], &tret); - if (err != 0) - exit(1); - tot_rtime += reader_time[i]; - } - for (i = 0; i < NR_WRITE; i++) { - err = pthread_join(tid_writer[i], &tret); - if (err != 0) - exit(1); - tot_wtime += writer_time[i]; - } - free(test_rcu_pointer); - printf("Time per read : %g cycles\n", - (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); - printf("Time per write : %g cycles\n", - (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); - - free(reader_time); - free(writer_time); - free(tid_reader); - free(tid_writer); - - return 0; -} diff --git a/tests/Makefile b/tests/Makefile new file mode 100644 index 0000000..ab19097 --- /dev/null +++ b/tests/Makefile @@ -0,0 +1 @@ +include Makefile.inc diff --git a/tests/Makefile.inc b/tests/Makefile.inc new file mode 100644 index 0000000..5fac462 --- /dev/null +++ b/tests/Makefile.inc @@ -0,0 +1,114 @@ + +include ../Makefile.build.inc + +ifeq ($(findstring ${ARCHTYPE},"x86 ppc"),) +APIHEADER=api_gcc.h +else +APIHEADER=api_${ARCHTYPE}.h +endif + +LIBDIR=.. + +CFLAGS+=-I ${LIBDIR} + +URCU_SIGNAL=${LIBDIR}/urcu.o ${LIBDIR}/urcu.h +URCU_SIGNAL_YIELD=${LIBDIR}/urcu-yield.o ${LIBDIR}/urcu.h +URCU_MB=${LIBDIR}/urcu-mb.o ${LIBDIR}/urcu.h +URCU_QSBR=${LIBDIR}/urcu-qsbr.o ${LIBDIR}/urcu-qsbr.h + +all: test_urcu test_urcu_dynamic_link test_urcu_timing \ + test_rwlock_timing test_rwlock test_perthreadlock_timing \ + test_perthreadlock test_urcu_yield test_urcu_mb \ + urcu-asm.S test_qsbr_timing test_qsbr urcu-asm.o urcutorture \ + urcutorture-yield test_mutex test_looplen test_urcu_gc \ + test_urcu_gc_mb test_qsbr_gc test_qsbr_lgc test_urcu_lgc \ + test_urcu_lgc_mb + +api.h: ${APIHEADER} + cp -f ${APIHEADER} api.h + +test_looplen: test_looplen.c ${LIBDIR}/urcu.h + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + + +test_urcu: test_urcu.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_gc: test_urcu_gc.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_lgc: test_urcu_gc.c ${URCU_SIGNAL} + $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + + +test_urcu_mb: test_urcu.c ${URCU_MB} + $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_gc_mb: test_urcu_gc.c ${URCU_MB} + $(CC) -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_lgc_mb: test_urcu_gc.c ${URCU_MB} + $(CC) -DTEST_LOCAL_GC -DCONFIG_URCU_AVOID_SIGNALS ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + + +test_qsbr: test_qsbr.c ${URCU_QSBR} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_qsbr_gc: test_qsbr_gc.c ${URCU_QSBR} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_qsbr_lgc: test_qsbr_gc.c ${URCU_QSBR} + $(CC) -DTEST_LOCAL_GC ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + + +test_rwlock: test_rwlock.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_perthreadlock: test_perthreadlock.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_mutex: test_mutex.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_dynamic_link: test_urcu.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} -DDYNAMIC_LINK_TEST $(LDFLAGS) -o $@ $(SRC_DEP) + + +test_urcu_yield: test_urcu.c ${URCU_SIGNAL_YIELD} + $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_urcu_timing: test_urcu_timing.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_qsbr_timing: test_qsbr_timing.c ${URCU_QSBR} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_rwlock_timing: test_rwlock_timing.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +test_perthreadlock_timing: test_perthreadlock_timing.c ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +urcu-asm.S: urcu-asm.c ${LIBDIR}/urcu.h + $(CC) ${CFLAGS} -S -o $@ $(SRC_DEP) + +urcu-asm.o: urcu-asm.c ${LIBDIR}/urcu.h + $(CC) ${CFLAGS} -c -o $@ $(SRC_DEP) + +urcutorture: urcutorture.c rcutorture.h api.h ${URCU_SIGNAL} + $(CC) ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +urcutorture-yield: urcutorture.c ${URCU_SIGNAL_YIELD} rcutorture.h api.h + $(CC) -DDEBUG_YIELD ${CFLAGS} $(LDFLAGS) -o $@ $(SRC_DEP) + +,PHONY: clean + +clean: + rm -f *.o test_urcu test_urcu_dynamic_link test_urcu_timing \ + test_rwlock_timing test_rwlock test_perthreadlock_timing \ + test_perthreadlock test_urcu_yield test_urcu_mb \ + urcu-asm.S test_qsbr_timing test_qsbr urcutorture \ + urcutorture-yield liburcu.so api.h \ + test_mutex test_urcu_gc test_urcu_gc_mb urcu-asm-1.S \ + test_qsbr_lgc test_qsbr_gc test_looplen test_urcu_lgc \ + test_urcu_lgc_mb diff --git a/tests/Makefile64 b/tests/Makefile64 new file mode 100644 index 0000000..cca1cdc --- /dev/null +++ b/tests/Makefile64 @@ -0,0 +1,3 @@ +include Makefile.inc + +CFLAGS+=-m64 diff --git a/tests/api_gcc.h b/tests/api_gcc.h new file mode 100644 index 0000000..1ad7345 --- /dev/null +++ b/tests/api_gcc.h @@ -0,0 +1,1328 @@ +#define _INCLUDE_API_H + +/* + * common.h: Common Linux kernel-isms. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; but version 2 of the License only due + * to code included from the Linux kernel. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +#ifndef __always_inline +#define __always_inline inline +#endif + +#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) +#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) + +#ifdef __ASSEMBLY__ +# define stringify_in_c(...) __VA_ARGS__ +# define ASM_CONST(x) x +#else +/* This version of stringify will deal with commas... */ +# define __stringify_in_c(...) #__VA_ARGS__ +# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " +# define __ASM_CONST(x) x##UL +# define ASM_CONST(x) __ASM_CONST(x) +#endif + + +/* + * arch-i386.h: Expose x86 atomic instructions. 80486 and better only. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, but version 2 only due to inclusion + * of Linux-kernel code. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +/* + * Machine parameters. + */ + +#define CACHE_LINE_SIZE 64 +#define ____cacheline_internodealigned_in_smp \ + __attribute__((__aligned__(1 << 6))) + +#define LOCK_PREFIX "lock ; " + +/* + * Atomic data structure, initialization, and access. + */ + +typedef struct { volatile int counter; } atomic_t; + +#define ATOMIC_INIT(i) { (i) } + +#define atomic_read(v) ((v)->counter) +#define atomic_set(v, i) (((v)->counter) = (i)) + +/* + * Atomic operations. + */ + +/** + * atomic_add - add integer to atomic variable + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @i to @v. + */ + +static __inline__ void atomic_add(int i, atomic_t *v) +{ + (void)__sync_fetch_and_add(&v->counter, i); +} + +/** + * atomic_sub - subtract the atomic variable + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v. + */ +static __inline__ void atomic_sub(int i, atomic_t *v) +{ + (void)__sync_fetch_and_add(&v->counter, -i); +} + +/** + * atomic_sub_and_test - subtract value from variable and test result + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v and returns + * true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_sub_and_test(int i, atomic_t *v) +{ + return __sync_add_and_fetch(&v->counter, -i) == 0; +} + +/** + * atomic_inc - increment atomic variable + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1. + */ +static __inline__ void atomic_inc(atomic_t *v) +{ + (void)__sync_fetch_and_add(&v->counter, 1); +} + +/** + * atomic_dec - decrement atomic variable + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1. + */ +static __inline__ void atomic_dec(atomic_t *v) +{ + (void)__sync_fetch_and_add(&v->counter, -1); +} + +/** + * atomic_dec_and_test - decrement and test + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +static __inline__ int atomic_dec_and_test(atomic_t *v) +{ + return __sync_add_and_fetch(&v->counter, -1) == 0; +} + +/** + * atomic_inc_and_test - increment and test + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_inc_and_test(atomic_t *v) +{ + return __sync_add_and_fetch(&v->counter, 1) == 0; +} + +/** + * atomic_add_negative - add and test if negative + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +static __inline__ int atomic_add_negative(int i, atomic_t *v) +{ + return __sync_add_and_fetch(&v->counter, i) < 0; +} + +/** + * atomic_add_return - add and return + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns @i + @v + */ +static __inline__ int atomic_add_return(int i, atomic_t *v) +{ + return __sync_add_and_fetch(&v->counter, i); +} + +static __inline__ int atomic_sub_return(int i, atomic_t *v) +{ + return atomic_add_return(-i,v); +} + +static inline unsigned int +cmpxchg(volatile long *ptr, long oldval, long newval) +{ + return __sync_val_compare_and_swap(ptr, oldval, newval); +} + +#define atomic_cmpxchg(v, old, new) ((int)cmpxchg(&((v)->counter), old, new)) +#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) + +/** + * atomic_add_unless - add unless the number is a given value + * @v: pointer of type atomic_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @v, so long as it was not @u. + * Returns non-zero if @v was not @u, and zero otherwise. + */ +#define atomic_add_unless(v, a, u) \ +({ \ + int c, old; \ + c = atomic_read(v); \ + for (;;) { \ + if (unlikely(c == (u))) \ + break; \ + old = atomic_cmpxchg((v), c, c + (a)); \ + if (likely(old == c)) \ + break; \ + c = old; \ + } \ + c != (u); \ +}) +#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) + +#define atomic_inc_return(v) (atomic_add_return(1,v)) +#define atomic_dec_return(v) (atomic_sub_return(1,v)) + +/* Atomic operations are already serializing on x86 */ +#define smp_mb__before_atomic_dec() barrier() +#define smp_mb__after_atomic_dec() barrier() +#define smp_mb__before_atomic_inc() barrier() +#define smp_mb__after_atomic_inc() barrier() + +/* + * api_pthreads.h: API mapping to pthreads environment. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. However, please note that much + * of the code in this file derives from the Linux kernel, and that such + * code may not be available except under GPLv2. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + */ + +#include +#include +#include +#include +#include +#define __USE_GNU +#include +#include +#include +/* #include "atomic.h" */ + +/* + * Compiler magic. + */ +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Default machine parameters. + */ + +#ifndef CACHE_LINE_SIZE +#define CACHE_LINE_SIZE 128 +#endif /* #ifndef CACHE_LINE_SIZE */ + +/* + * Exclusive locking primitives. + */ + +typedef pthread_mutex_t spinlock_t; + +#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; +#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER + +static void spin_lock_init(spinlock_t *sp) +{ + if (pthread_mutex_init(sp, NULL) != 0) { + perror("spin_lock_init:pthread_mutex_init"); + exit(-1); + } +} + +static void spin_lock(spinlock_t *sp) +{ + if (pthread_mutex_lock(sp) != 0) { + perror("spin_lock:pthread_mutex_lock"); + exit(-1); + } +} + +static void spin_unlock(spinlock_t *sp) +{ + if (pthread_mutex_unlock(sp) != 0) { + perror("spin_unlock:pthread_mutex_unlock"); + exit(-1); + } +} + +#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) +#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) + +/* + * Thread creation/destruction primitives. + */ + +typedef pthread_t thread_id_t; + +#define NR_THREADS 128 + +#define __THREAD_ID_MAP_EMPTY 0 +#define __THREAD_ID_MAP_WAITING 1 +thread_id_t __thread_id_map[NR_THREADS]; +spinlock_t __thread_id_map_mutex; + +#define for_each_thread(t) \ + for (t = 0; t < NR_THREADS; t++) + +#define for_each_running_thread(t) \ + for (t = 0; t < NR_THREADS; t++) \ + if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ + (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) + +pthread_key_t thread_id_key; + +static int __smp_thread_id(void) +{ + int i; + thread_id_t tid = pthread_self(); + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) { + long v = i + 1; /* must be non-NULL. */ + + if (pthread_setspecific(thread_id_key, (void *)v) != 0) { + perror("pthread_setspecific"); + exit(-1); + } + return i; + } + } + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + spin_unlock(&__thread_id_map_mutex); + return i; + } + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); +} + +static int smp_thread_id(void) +{ + void *id; + + id = pthread_getspecific(thread_id_key); + if (id == NULL) + return __smp_thread_id(); + return (long)(id - 1); +} + +static thread_id_t create_thread(void *(*func)(void *), void *arg) +{ + thread_id_t tid; + int i; + + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) + break; + } + if (i >= NR_THREADS) { + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_WAITING; + spin_unlock(&__thread_id_map_mutex); + if (pthread_create(&tid, NULL, func, arg) != 0) { + perror("create_thread:pthread_create"); + exit(-1); + } + __thread_id_map[i] = tid; + return tid; +} + +static void *wait_thread(thread_id_t tid) +{ + int i; + void *vp; + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + break; + } + if (i >= NR_THREADS){ + fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); + } + if (pthread_join(tid, &vp) != 0) { + perror("wait_thread:pthread_join"); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + return vp; +} + +static void wait_all_threads(void) +{ + int i; + thread_id_t tid; + + for (i = 1; i < NR_THREADS; i++) { + tid = __thread_id_map[i]; + if (tid != __THREAD_ID_MAP_EMPTY && + tid != __THREAD_ID_MAP_WAITING) + (void)wait_thread(tid); + } +} + +static void run_on(int cpu) +{ + cpu_set_t mask; + + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * timekeeping -- very crude -- should use MONOTONIC... + */ + +long long get_microseconds(void) +{ + struct timeval tv; + + if (gettimeofday(&tv, NULL) != 0) + abort(); + return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; +} + +/* + * Per-thread variables. + */ + +#define DEFINE_PER_THREAD(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_thread_##name[NR_THREADS]; +#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) + +#define per_thread(name, thread) __per_thread_##name[thread].v +#define __get_thread_var(name) per_thread(name, smp_thread_id()) + +#define init_per_thread(name, v) \ + do { \ + int __i_p_t_i; \ + for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ + per_thread(name, __i_p_t_i) = v; \ + } while (0) + +/* + * CPU traversal primitives. + */ + +#ifndef NR_CPUS +#define NR_CPUS 16 +#endif /* #ifndef NR_CPUS */ + +#define for_each_possible_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) +#define for_each_online_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) + +/* + * Per-CPU variables. + */ + +#define DEFINE_PER_CPU(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_cpu_##name[NR_CPUS] +#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) + +DEFINE_PER_THREAD(int, smp_processor_id); + +#define per_cpu(name, thread) __per_cpu_##name[thread].v +#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) + +#define init_per_cpu(name, v) \ + do { \ + int __i_p_c_i; \ + for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ + per_cpu(name, __i_p_c_i) = v; \ + } while (0) + +/* + * CPU state checking (crowbarred). + */ + +#define idle_cpu(cpu) 0 +#define in_softirq() 1 +#define hardirq_count() 0 +#define PREEMPT_SHIFT 0 +#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) +#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) +#define PREEMPT_BITS 8 +#define SOFTIRQ_BITS 8 + +/* + * CPU hotplug. + */ + +struct notifier_block { + int (*notifier_call)(struct notifier_block *, unsigned long, void *); + struct notifier_block *next; + int priority; +}; + +#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ +#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ +#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ +#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ +#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ +#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ +#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, + * not handling interrupts, soon dead */ +#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug + * lock is dropped */ + +/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend + * operation in progress + */ +#define CPU_TASKS_FROZEN 0x0010 + +#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) +#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) +#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) +#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) +#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) +#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) +#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) + +/* Hibernation and suspend events */ +#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ +#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ +#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ +#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ +#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ +#define PM_POST_RESTORE 0x0006 /* Restore failed */ + +#define NOTIFY_DONE 0x0000 /* Don't care */ +#define NOTIFY_OK 0x0001 /* Suits me */ +#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ +#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) + /* Bad/Veto action */ +/* + * Clean way to return from the notifier and stop further calls. + */ +#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) + +/* + * Bug checks. + */ + +#define BUG_ON(c) do { if (!(c)) abort(); } while (0) + +/* + * Initialization -- Must be called before calling any primitives. + */ + +static void smp_init(void) +{ + int i; + + spin_lock_init(&__thread_id_map_mutex); + __thread_id_map[0] = pthread_self(); + for (i = 1; i < NR_THREADS; i++) + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + init_per_thread(smp_processor_id, 0); + if (pthread_key_create(&thread_id_key, NULL) != 0) { + perror("pthread_key_create"); + exit(-1); + } +} + +/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ + +#ifndef _LINUX_LIST_H +#define _LINUX_LIST_H + +#define LIST_POISON1 ((void *) 0x00100100) +#define LIST_POISON2 ((void *) 0x00200200) + +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Simple doubly linked list implementation. + * + * Some of the internal functions ("__xxx") are useful when + * manipulating whole lists rather than single entries, as + * sometimes we already know the next/prev entries and we can + * generate better code by using them directly rather than + * using the generic single-entry routines. + */ + +struct list_head { + struct list_head *next, *prev; +}; + +#define LIST_HEAD_INIT(name) { &(name), &(name) } + +#define LIST_HEAD(name) \ + struct list_head name = LIST_HEAD_INIT(name) + +static inline void INIT_LIST_HEAD(struct list_head *list) +{ + list->next = list; + list->prev = list; +} + +/* + * Insert a new entry between two known consecutive entries. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +#ifndef CONFIG_DEBUG_LIST +static inline void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next) +{ + next->prev = new; + new->next = next; + new->prev = prev; + prev->next = new; +} +#else +extern void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next); +#endif + +/** + * list_add - add a new entry + * @new: new entry to be added + * @head: list head to add it after + * + * Insert a new entry after the specified head. + * This is good for implementing stacks. + */ +static inline void list_add(struct list_head *new, struct list_head *head) +{ + __list_add(new, head, head->next); +} + + +/** + * list_add_tail - add a new entry + * @new: new entry to be added + * @head: list head to add it before + * + * Insert a new entry before the specified head. + * This is useful for implementing queues. + */ +static inline void list_add_tail(struct list_head *new, struct list_head *head) +{ + __list_add(new, head->prev, head); +} + +/* + * Delete a list entry by making the prev/next entries + * point to each other. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +static inline void __list_del(struct list_head * prev, struct list_head * next) +{ + next->prev = prev; + prev->next = next; +} + +/** + * list_del - deletes entry from list. + * @entry: the element to delete from the list. + * Note: list_empty() on entry does not return true after this, the entry is + * in an undefined state. + */ +#ifndef CONFIG_DEBUG_LIST +static inline void list_del(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + entry->next = LIST_POISON1; + entry->prev = LIST_POISON2; +} +#else +extern void list_del(struct list_head *entry); +#endif + +/** + * list_replace - replace old entry by new one + * @old : the element to be replaced + * @new : the new element to insert + * + * If @old was empty, it will be overwritten. + */ +static inline void list_replace(struct list_head *old, + struct list_head *new) +{ + new->next = old->next; + new->next->prev = new; + new->prev = old->prev; + new->prev->next = new; +} + +static inline void list_replace_init(struct list_head *old, + struct list_head *new) +{ + list_replace(old, new); + INIT_LIST_HEAD(old); +} + +/** + * list_del_init - deletes entry from list and reinitialize it. + * @entry: the element to delete from the list. + */ +static inline void list_del_init(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + INIT_LIST_HEAD(entry); +} + +/** + * list_move - delete from one list and add as another's head + * @list: the entry to move + * @head: the head that will precede our entry + */ +static inline void list_move(struct list_head *list, struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add(list, head); +} + +/** + * list_move_tail - delete from one list and add as another's tail + * @list: the entry to move + * @head: the head that will follow our entry + */ +static inline void list_move_tail(struct list_head *list, + struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add_tail(list, head); +} + +/** + * list_is_last - tests whether @list is the last entry in list @head + * @list: the entry to test + * @head: the head of the list + */ +static inline int list_is_last(const struct list_head *list, + const struct list_head *head) +{ + return list->next == head; +} + +/** + * list_empty - tests whether a list is empty + * @head: the list to test. + */ +static inline int list_empty(const struct list_head *head) +{ + return head->next == head; +} + +/** + * list_empty_careful - tests whether a list is empty and not being modified + * @head: the list to test + * + * Description: + * tests whether a list is empty _and_ checks that no other CPU might be + * in the process of modifying either member (next or prev) + * + * NOTE: using list_empty_careful() without synchronization + * can only be safe if the only activity that can happen + * to the list entry is list_del_init(). Eg. it cannot be used + * if another CPU could re-list_add() it. + */ +static inline int list_empty_careful(const struct list_head *head) +{ + struct list_head *next = head->next; + return (next == head) && (next == head->prev); +} + +/** + * list_is_singular - tests whether a list has just one entry. + * @head: the list to test. + */ +static inline int list_is_singular(const struct list_head *head) +{ + return !list_empty(head) && (head->next == head->prev); +} + +static inline void __list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + struct list_head *new_first = entry->next; + list->next = head->next; + list->next->prev = list; + list->prev = entry; + entry->next = list; + head->next = new_first; + new_first->prev = head; +} + +/** + * list_cut_position - cut a list into two + * @list: a new list to add all removed entries + * @head: a list with entries + * @entry: an entry within head, could be the head itself + * and if so we won't cut the list + * + * This helper moves the initial part of @head, up to and + * including @entry, from @head to @list. You should + * pass on @entry an element you know is on @head. @list + * should be an empty list or a list you do not care about + * losing its data. + * + */ +static inline void list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + if (list_empty(head)) + return; + if (list_is_singular(head) && + (head->next != entry && head != entry)) + return; + if (entry == head) + INIT_LIST_HEAD(list); + else + __list_cut_position(list, head, entry); +} + +static inline void __list_splice(const struct list_head *list, + struct list_head *prev, + struct list_head *next) +{ + struct list_head *first = list->next; + struct list_head *last = list->prev; + + first->prev = prev; + prev->next = first; + + last->next = next; + next->prev = last; +} + +/** + * list_splice - join two lists, this is designed for stacks + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice(const struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head, head->next); +} + +/** + * list_splice_tail - join two lists, each list being a queue + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice_tail(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head->prev, head); +} + +/** + * list_splice_init - join two lists and reinitialise the emptied list. + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * The list at @list is reinitialised + */ +static inline void list_splice_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head, head->next); + INIT_LIST_HEAD(list); + } +} + +/** + * list_splice_tail_init - join two lists and reinitialise the emptied list + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * Each of the lists is a queue. + * The list at @list is reinitialised + */ +static inline void list_splice_tail_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head->prev, head); + INIT_LIST_HEAD(list); + } +} + +/** + * list_entry - get the struct for this entry + * @ptr: the &struct list_head pointer. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + */ +#define list_entry(ptr, type, member) \ + container_of(ptr, type, member) + +/** + * list_first_entry - get the first element from a list + * @ptr: the list head to take the element from. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + * + * Note, that list is expected to be not empty. + */ +#define list_first_entry(ptr, type, member) \ + list_entry((ptr)->next, type, member) + +/** + * list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each(pos, head) \ + for (pos = (head)->next; prefetch(pos->next), pos != (head); \ + pos = pos->next) + +/** + * __list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + * + * This variant differs from list_for_each() in that it's the + * simplest possible list iteration code, no prefetching is done. + * Use this for code that knows the list to be very short (empty + * or 1 entry) most of the time. + */ +#define __list_for_each(pos, head) \ + for (pos = (head)->next; pos != (head); pos = pos->next) + +/** + * list_for_each_prev - iterate over a list backwards + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each_prev(pos, head) \ + for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ + pos = pos->prev) + +/** + * list_for_each_safe - iterate over a list safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_safe(pos, n, head) \ + for (pos = (head)->next, n = pos->next; pos != (head); \ + pos = n, n = pos->next) + +/** + * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_prev_safe(pos, n, head) \ + for (pos = (head)->prev, n = pos->prev; \ + prefetch(pos->prev), pos != (head); \ + pos = n, n = pos->prev) + +/** + * list_for_each_entry - iterate over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry(pos, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_reverse - iterate backwards over list of given type. + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_reverse(pos, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() + * @pos: the type * to use as a start point + * @head: the head of the list + * @member: the name of the list_struct within the struct. + * + * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). + */ +#define list_prepare_entry(pos, head, member) \ + ((pos) ? : list_entry(head, typeof(*pos), member)) + +/** + * list_for_each_entry_continue - continue iteration over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Continue to iterate over list of given type, continuing after + * the current position. + */ +#define list_for_each_entry_continue(pos, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_continue_reverse - iterate backwards from the given point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Start to iterate over list of given type backwards, continuing after + * the current position. + */ +#define list_for_each_entry_continue_reverse(pos, head, member) \ + for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_for_each_entry_from - iterate over list of given type from the current point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing from current position. + */ +#define list_for_each_entry_from(pos, head, member) \ + for (; prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_safe(pos, n, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_continue + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing after current point, + * safe against removal of list entry. + */ +#define list_for_each_entry_safe_continue(pos, n, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_from + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type from current point, safe against + * removal of list entry. + */ +#define list_for_each_entry_safe_from(pos, n, head, member) \ + for (n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_reverse + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate backwards over list of given type, safe against removal + * of list entry. + */ +#define list_for_each_entry_safe_reverse(pos, n, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member), \ + n = list_entry(pos->member.prev, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.prev, typeof(*n), member)) + +/* + * Double linked lists with a single pointer list head. + * Mostly useful for hash tables where the two pointer list head is + * too wasteful. + * You lose the ability to access the tail in O(1). + */ + +struct hlist_head { + struct hlist_node *first; +}; + +struct hlist_node { + struct hlist_node *next, **pprev; +}; + +#define HLIST_HEAD_INIT { .first = NULL } +#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } +#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) +static inline void INIT_HLIST_NODE(struct hlist_node *h) +{ + h->next = NULL; + h->pprev = NULL; +} + +static inline int hlist_unhashed(const struct hlist_node *h) +{ + return !h->pprev; +} + +static inline int hlist_empty(const struct hlist_head *h) +{ + return !h->first; +} + +static inline void __hlist_del(struct hlist_node *n) +{ + struct hlist_node *next = n->next; + struct hlist_node **pprev = n->pprev; + *pprev = next; + if (next) + next->pprev = pprev; +} + +static inline void hlist_del(struct hlist_node *n) +{ + __hlist_del(n); + n->next = LIST_POISON1; + n->pprev = LIST_POISON2; +} + +static inline void hlist_del_init(struct hlist_node *n) +{ + if (!hlist_unhashed(n)) { + __hlist_del(n); + INIT_HLIST_NODE(n); + } +} + +static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) +{ + struct hlist_node *first = h->first; + n->next = first; + if (first) + first->pprev = &n->next; + h->first = n; + n->pprev = &h->first; +} + +/* next must be != NULL */ +static inline void hlist_add_before(struct hlist_node *n, + struct hlist_node *next) +{ + n->pprev = next->pprev; + n->next = next; + next->pprev = &n->next; + *(n->pprev) = n; +} + +static inline void hlist_add_after(struct hlist_node *n, + struct hlist_node *next) +{ + next->next = n->next; + n->next = next; + next->pprev = &n->next; + + if(next->next) + next->next->pprev = &next->next; +} + +/* + * Move a list from one list head to another. Fixup the pprev + * reference of the first entry if it exists. + */ +static inline void hlist_move_list(struct hlist_head *old, + struct hlist_head *new) +{ + new->first = old->first; + if (new->first) + new->first->pprev = &new->first; + old->first = NULL; +} + +#define hlist_entry(ptr, type, member) container_of(ptr,type,member) + +#define hlist_for_each(pos, head) \ + for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ + pos = pos->next) + +#define hlist_for_each_safe(pos, n, head) \ + for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ + pos = n) + +/** + * hlist_for_each_entry - iterate over list of given type + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry(tpos, pos, head, member) \ + for (pos = (head)->first; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_continue - iterate over a hlist continuing after current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_continue(tpos, pos, member) \ + for (pos = (pos)->next; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_from - iterate over a hlist continuing from current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_from(tpos, pos, member) \ + for (; pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @n: another &struct hlist_node to use as temporary storage + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ + for (pos = (head)->first; \ + pos && ({ n = pos->next; 1; }) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = n) + +#endif diff --git a/tests/api_ppc.h b/tests/api_ppc.h new file mode 100644 index 0000000..8a03faa --- /dev/null +++ b/tests/api_ppc.h @@ -0,0 +1,1699 @@ +/* MECHANICALLY GENERATED, DO NOT EDIT!!! */ + +#define _INCLUDE_API_H + +/* + * common.h: Common Linux kernel-isms. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; but version 2 of the License only due + * to code included from the Linux kernel. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +#ifndef __always_inline +#define __always_inline inline +#endif + +#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) +#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) + +#ifdef __ASSEMBLY__ +# define stringify_in_c(...) __VA_ARGS__ +# define ASM_CONST(x) x +#else +/* This version of stringify will deal with commas... */ +# define __stringify_in_c(...) #__VA_ARGS__ +# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " +# define __ASM_CONST(x) x##UL +# define ASM_CONST(x) __ASM_CONST(x) +#endif + + +/* + * arch-ppc64.h: Expose PowerPC atomic instructions. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; but version 2 of the License only due + * to code included from the Linux kernel. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +/* + * Machine parameters. + */ + +#define CONFIG_PPC64 + +#define CACHE_LINE_SIZE 128 +#define ____cacheline_internodealigned_in_smp \ + __attribute__((__aligned__(1 << 7))) + +/* + * Atomic data structure, initialization, and access. + */ + +typedef struct { volatile int counter; } atomic_t; + +#define ATOMIC_INIT(i) { (i) } + +#define atomic_read(v) ((v)->counter) +#define atomic_set(v, i) (((v)->counter) = (i)) + +/* + * Atomic operations. + */ + +#define LWSYNC lwsync +#define PPC405_ERR77(ra,rb) +#ifdef CONFIG_SMP +# define LWSYNC_ON_SMP stringify_in_c(LWSYNC) "\n" +# define ISYNC_ON_SMP "\n\tisync\n" +#else +# define LWSYNC_ON_SMP +# define ISYNC_ON_SMP +#endif + + +/* + * Atomic exchange + * + * Changes the memory location '*ptr' to be val and returns + * the previous value stored there. + */ +static __always_inline unsigned long +__xchg_u32(volatile void *p, unsigned long val) +{ + unsigned long prev; + + __asm__ __volatile__( + LWSYNC_ON_SMP +"1: lwarx %0,0,%2 \n" + PPC405_ERR77(0,%2) +" stwcx. %3,0,%2 \n\ + bne- 1b" + ISYNC_ON_SMP + : "=&r" (prev), "+m" (*(volatile unsigned int *)p) + : "r" (p), "r" (val) + : "cc", "memory"); + + return prev; +} + +/* + * Atomic exchange + * + * Changes the memory location '*ptr' to be val and returns + * the previous value stored there. + */ +static __always_inline unsigned long +__xchg_u32_local(volatile void *p, unsigned long val) +{ + unsigned long prev; + + __asm__ __volatile__( +"1: lwarx %0,0,%2 \n" + PPC405_ERR77(0,%2) +" stwcx. %3,0,%2 \n\ + bne- 1b" + : "=&r" (prev), "+m" (*(volatile unsigned int *)p) + : "r" (p), "r" (val) + : "cc", "memory"); + + return prev; +} + +#ifdef CONFIG_PPC64 +static __always_inline unsigned long +__xchg_u64(volatile void *p, unsigned long val) +{ + unsigned long prev; + + __asm__ __volatile__( + LWSYNC_ON_SMP +"1: ldarx %0,0,%2 \n" + PPC405_ERR77(0,%2) +" stdcx. %3,0,%2 \n\ + bne- 1b" + ISYNC_ON_SMP + : "=&r" (prev), "+m" (*(volatile unsigned long *)p) + : "r" (p), "r" (val) + : "cc", "memory"); + + return prev; +} + +static __always_inline unsigned long +__xchg_u64_local(volatile void *p, unsigned long val) +{ + unsigned long prev; + + __asm__ __volatile__( +"1: ldarx %0,0,%2 \n" + PPC405_ERR77(0,%2) +" stdcx. %3,0,%2 \n\ + bne- 1b" + : "=&r" (prev), "+m" (*(volatile unsigned long *)p) + : "r" (p), "r" (val) + : "cc", "memory"); + + return prev; +} +#endif + +/* + * This function doesn't exist, so you'll get a linker error + * if something tries to do an invalid xchg(). + */ +extern void __xchg_called_with_bad_pointer(void); + +static __always_inline unsigned long +__xchg(volatile void *ptr, unsigned long x, unsigned int size) +{ + switch (size) { + case 4: + return __xchg_u32(ptr, x); +#ifdef CONFIG_PPC64 + case 8: + return __xchg_u64(ptr, x); +#endif + } + __xchg_called_with_bad_pointer(); + return x; +} + +static __always_inline unsigned long +__xchg_local(volatile void *ptr, unsigned long x, unsigned int size) +{ + switch (size) { + case 4: + return __xchg_u32_local(ptr, x); +#ifdef CONFIG_PPC64 + case 8: + return __xchg_u64_local(ptr, x); +#endif + } + __xchg_called_with_bad_pointer(); + return x; +} +#define xchg(ptr,x) \ + ({ \ + __typeof__(*(ptr)) _x_ = (x); \ + (__typeof__(*(ptr))) __xchg((ptr), (unsigned long)_x_, sizeof(*(ptr))); \ + }) + +#define xchg_local(ptr,x) \ + ({ \ + __typeof__(*(ptr)) _x_ = (x); \ + (__typeof__(*(ptr))) __xchg_local((ptr), \ + (unsigned long)_x_, sizeof(*(ptr))); \ + }) + +/* + * Compare and exchange - if *p == old, set it to new, + * and return the old value of *p. + */ +#define __HAVE_ARCH_CMPXCHG 1 + +static __always_inline unsigned long +__cmpxchg_u32(volatile unsigned int *p, unsigned long old, unsigned long new) +{ + unsigned int prev; + + __asm__ __volatile__ ( + LWSYNC_ON_SMP +"1: lwarx %0,0,%2 # __cmpxchg_u32\n\ + cmpw 0,%0,%3\n\ + bne- 2f\n" + PPC405_ERR77(0,%2) +" stwcx. %4,0,%2\n\ + bne- 1b" + ISYNC_ON_SMP + "\n\ +2:" + : "=&r" (prev), "+m" (*p) + : "r" (p), "r" (old), "r" (new) + : "cc", "memory"); + + return prev; +} + +static __always_inline unsigned long +__cmpxchg_u32_local(volatile unsigned int *p, unsigned long old, + unsigned long new) +{ + unsigned int prev; + + __asm__ __volatile__ ( +"1: lwarx %0,0,%2 # __cmpxchg_u32\n\ + cmpw 0,%0,%3\n\ + bne- 2f\n" + PPC405_ERR77(0,%2) +" stwcx. %4,0,%2\n\ + bne- 1b" + "\n\ +2:" + : "=&r" (prev), "+m" (*p) + : "r" (p), "r" (old), "r" (new) + : "cc", "memory"); + + return prev; +} + +#ifdef CONFIG_PPC64 +static __always_inline unsigned long +__cmpxchg_u64(volatile unsigned long *p, unsigned long old, unsigned long new) +{ + unsigned long prev; + + __asm__ __volatile__ ( + LWSYNC_ON_SMP +"1: ldarx %0,0,%2 # __cmpxchg_u64\n\ + cmpd 0,%0,%3\n\ + bne- 2f\n\ + stdcx. %4,0,%2\n\ + bne- 1b" + ISYNC_ON_SMP + "\n\ +2:" + : "=&r" (prev), "+m" (*p) + : "r" (p), "r" (old), "r" (new) + : "cc", "memory"); + + return prev; +} + +static __always_inline unsigned long +__cmpxchg_u64_local(volatile unsigned long *p, unsigned long old, + unsigned long new) +{ + unsigned long prev; + + __asm__ __volatile__ ( +"1: ldarx %0,0,%2 # __cmpxchg_u64\n\ + cmpd 0,%0,%3\n\ + bne- 2f\n\ + stdcx. %4,0,%2\n\ + bne- 1b" + "\n\ +2:" + : "=&r" (prev), "+m" (*p) + : "r" (p), "r" (old), "r" (new) + : "cc", "memory"); + + return prev; +} +#endif + +/* This function doesn't exist, so you'll get a linker error + if something tries to do an invalid cmpxchg(). */ +extern void __cmpxchg_called_with_bad_pointer(void); + +static __always_inline unsigned long +__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, + unsigned int size) +{ + switch (size) { + case 4: + return __cmpxchg_u32(ptr, old, new); +#ifdef CONFIG_PPC64 + case 8: + return __cmpxchg_u64(ptr, old, new); +#endif + } + __cmpxchg_called_with_bad_pointer(); + return old; +} + +static __always_inline unsigned long +__cmpxchg_local(volatile void *ptr, unsigned long old, unsigned long new, + unsigned int size) +{ + switch (size) { + case 4: + return __cmpxchg_u32_local(ptr, old, new); +#ifdef CONFIG_PPC64 + case 8: + return __cmpxchg_u64_local(ptr, old, new); +#endif + } + __cmpxchg_called_with_bad_pointer(); + return old; +} + +#define cmpxchg(ptr, o, n) \ + ({ \ + __typeof__(*(ptr)) _o_ = (o); \ + __typeof__(*(ptr)) _n_ = (n); \ + (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_, \ + (unsigned long)_n_, sizeof(*(ptr))); \ + }) + + +#define cmpxchg_local(ptr, o, n) \ + ({ \ + __typeof__(*(ptr)) _o_ = (o); \ + __typeof__(*(ptr)) _n_ = (n); \ + (__typeof__(*(ptr))) __cmpxchg_local((ptr), (unsigned long)_o_, \ + (unsigned long)_n_, sizeof(*(ptr))); \ + }) + +#ifdef CONFIG_PPC64 +/* + * We handle most unaligned accesses in hardware. On the other hand + * unaligned DMA can be very expensive on some ppc64 IO chips (it does + * powers of 2 writes until it reaches sufficient alignment). + * + * Based on this we disable the IP header alignment in network drivers. + * We also modify NET_SKB_PAD to be a cacheline in size, thus maintaining + * cacheline alignment of buffers. + */ +#define NET_IP_ALIGN 0 +#define NET_SKB_PAD L1_CACHE_BYTES + +#define cmpxchg64(ptr, o, n) \ + ({ \ + BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ + cmpxchg((ptr), (o), (n)); \ + }) +#define cmpxchg64_local(ptr, o, n) \ + ({ \ + BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ + cmpxchg_local((ptr), (o), (n)); \ + }) +#endif + +#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n))) +#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) + +/** + * atomic_add - add integer to atomic variable + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @a to @v. + */ +static __inline__ void atomic_add(int a, atomic_t *v) +{ + int t; + + __asm__ __volatile__( + "1: lwarx %0,0,%3 # atomic_add\n\ + add %0,%2,%0 \n\ + stwcx. %0,0,%3 \n\ + bne- 1b" + : "=&r" (t), "+m" (v->counter) + : "r" (a), "r" (&v->counter) + : "cc"); +} + +/** + * atomic_sub - subtract the atomic variable + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @a from @v. + */ +static __inline__ void atomic_sub(int a, atomic_t *v) +{ + int t; + + __asm__ __volatile__( + "1: lwarx %0,0,%3 # atomic_sub \n\ + subf %0,%2,%0 \n\ + stwcx. %0,0,%3 \n\ + bne- 1b" + : "=&r" (t), "+m" (v->counter) + : "r" (a), "r" (&v->counter) + : "cc"); +} + +static __inline__ atomic_sub_return(int a, atomic_t *v) +{ + int t; + + __asm__ __volatile__( + "lwsync\n\ + 1: lwarx %0,0,%2 # atomic_sub_return\n\ + subf %0,%1,%0\n\ + stwcx. %0,0,%2 \n\ + bne- 1b \n\ + isync" + : "=&r" (t) + : "r" (a), "r" (&v->counter) + : "cc", "memory"); + + return t; +} + +/** + * atomic_sub_and_test - subtract value from variable and test result + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v and returns + * true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_sub_and_test(int a, atomic_t *v) +{ + return atomic_sub_return(a, v) == 0; +} + +/** + * atomic_inc - increment atomic variable + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1. + */ +static __inline__ void atomic_inc(atomic_t *v) +{ + atomic_add(1, v); +} + +/** + * atomic_dec - decrement atomic variable + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1. + */ +static __inline__ void atomic_dec(atomic_t *v) +{ + atomic_sub(1, v); +} + +/** + * atomic_dec_and_test - decrement and test + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +static __inline__ int atomic_dec_and_test(atomic_t *v) +{ + return atomic_sub_and_test(1, v); +} + +/** + * atomic_inc_and_test - increment and test + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_inc_and_test(atomic_t *v) +{ + return atomic_inc_return(v); +} + +/** + * atomic_add_return - add and return + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns @i + @v + */ +static __inline__ int atomic_add_return(int a, atomic_t *v) +{ + int t; + + __asm__ __volatile__( + "lwsync \n\ + 1: lwarx %0,0,%2 # atomic_add_return \n\ + add %0,%1,%0 \n\ + stwcx. %0,0,%2 \n\ + bne- 1b \n\ + isync" + : "=&r" (t) + : "r" (a), "r" (&v->counter) + : "cc", "memory"); + + return t; +} + +/** + * atomic_add_negative - add and test if negative + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +static __inline__ int atomic_add_negative(int a, atomic_t *v) +{ + return atomic_add_return(a, v) < 0; +} + +/** + * atomic_add_unless - add unless the number is a given value + * @v: pointer of type atomic_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @v, so long as it was not @u. + * Returns non-zero if @v was not @u, and zero otherwise. + */ +static __inline__ int atomic_add_unless(atomic_t *v, int a, int u) +{ + int t; + + __asm__ __volatile__( + "lwsync \n\ + 1: lwarx %0,0,%1 # atomic_add_unless\n\ + cmpd 0,%0,%3 \n\ + beq- 2f \n\ + add %0,%2,%0 \n\ + stwcx. %0,0,%1 \n\ + bne- 1b \n\ + isync \n\ + subf %0,%2,%0 \n\ + 2:" + : "=&r" (t) + : "r" (&v->counter), "r" (a), "r" (u) + : "cc", "memory"); + + return t != u; +} + +#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) + +#define atomic_inc_return(v) (atomic_add_return(1,v)) +#define atomic_dec_return(v) (atomic_sub_return(1,v)) + +/* Atomic operations are already serializing on x86 */ +#define smp_mb__before_atomic_dec() smp_mb() +#define smp_mb__after_atomic_dec() smp_mb() +#define smp_mb__before_atomic_inc() smp_mb() +#define smp_mb__after_atomic_inc() smp_mb() + +/* + * api_pthreads.h: API mapping to pthreads environment. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. However, please note that much + * of the code in this file derives from the Linux kernel, and that such + * code may not be available except under GPLv2. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + */ + +#include +#include +#include +#include +#include +#define __USE_GNU +#include +#include +#include +/* #include "atomic.h" */ + +/* + * Compiler magic. + */ +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Default machine parameters. + */ + +#ifndef CACHE_LINE_SIZE +#define CACHE_LINE_SIZE 128 +#endif /* #ifndef CACHE_LINE_SIZE */ + +/* + * Exclusive locking primitives. + */ + +typedef pthread_mutex_t spinlock_t; + +#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; +#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER + +static void spin_lock_init(spinlock_t *sp) +{ + if (pthread_mutex_init(sp, NULL) != 0) { + perror("spin_lock_init:pthread_mutex_init"); + exit(-1); + } +} + +static void spin_lock(spinlock_t *sp) +{ + if (pthread_mutex_lock(sp) != 0) { + perror("spin_lock:pthread_mutex_lock"); + exit(-1); + } +} + +static void spin_unlock(spinlock_t *sp) +{ + if (pthread_mutex_unlock(sp) != 0) { + perror("spin_unlock:pthread_mutex_unlock"); + exit(-1); + } +} + +#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) +#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) + +/* + * Thread creation/destruction primitives. + */ + +typedef pthread_t thread_id_t; + +#define NR_THREADS 128 + +#define __THREAD_ID_MAP_EMPTY 0 +#define __THREAD_ID_MAP_WAITING 1 +thread_id_t __thread_id_map[NR_THREADS]; +spinlock_t __thread_id_map_mutex; + +#define for_each_thread(t) \ + for (t = 0; t < NR_THREADS; t++) + +#define for_each_running_thread(t) \ + for (t = 0; t < NR_THREADS; t++) \ + if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ + (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) + +#define for_each_tid(t, tid) \ + for (t = 0; t < NR_THREADS; t++) \ + if ((((tid) = __thread_id_map[t]) != __THREAD_ID_MAP_EMPTY) && \ + ((tid) != __THREAD_ID_MAP_WAITING)) + +pthread_key_t thread_id_key; + +static int __smp_thread_id(void) +{ + int i; + thread_id_t tid = pthread_self(); + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) { + long v = i + 1; /* must be non-NULL. */ + + if (pthread_setspecific(thread_id_key, (void *)v) != 0) { + perror("pthread_setspecific"); + exit(-1); + } + return i; + } + } + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + spin_unlock(&__thread_id_map_mutex); + return i; + } + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); +} + +static int smp_thread_id(void) +{ + void *id; + + id = pthread_getspecific(thread_id_key); + if (id == NULL) + return __smp_thread_id(); + return (long)(id - 1); +} + +static thread_id_t create_thread(void *(*func)(void *), void *arg) +{ + thread_id_t tid; + int i; + + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) + break; + } + if (i >= NR_THREADS) { + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_WAITING; + spin_unlock(&__thread_id_map_mutex); + if (pthread_create(&tid, NULL, func, arg) != 0) { + perror("create_thread:pthread_create"); + exit(-1); + } + __thread_id_map[i] = tid; + return tid; +} + +static void *wait_thread(thread_id_t tid) +{ + int i; + void *vp; + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + break; + } + if (i >= NR_THREADS){ + fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); + } + if (pthread_join(tid, &vp) != 0) { + perror("wait_thread:pthread_join"); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + return vp; +} + +static void wait_all_threads(void) +{ + int i; + thread_id_t tid; + + for (i = 1; i < NR_THREADS; i++) { + tid = __thread_id_map[i]; + if (tid != __THREAD_ID_MAP_EMPTY && + tid != __THREAD_ID_MAP_WAITING) + (void)wait_thread(tid); + } +} + +static void run_on(int cpu) +{ + cpu_set_t mask; + + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * timekeeping -- very crude -- should use MONOTONIC... + */ + +long long get_microseconds(void) +{ + struct timeval tv; + + if (gettimeofday(&tv, NULL) != 0) + abort(); + return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; +} + +/* + * Per-thread variables. + */ + +#define DEFINE_PER_THREAD(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_thread_##name[NR_THREADS]; +#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) + +#define per_thread(name, thread) __per_thread_##name[thread].v +#define __get_thread_var(name) per_thread(name, smp_thread_id()) + +#define init_per_thread(name, v) \ + do { \ + int __i_p_t_i; \ + for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ + per_thread(name, __i_p_t_i) = v; \ + } while (0) + +/* + * CPU traversal primitives. + */ + +#ifndef NR_CPUS +#define NR_CPUS 16 +#endif /* #ifndef NR_CPUS */ + +#define for_each_possible_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) +#define for_each_online_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) + +/* + * Per-CPU variables. + */ + +#define DEFINE_PER_CPU(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_cpu_##name[NR_CPUS] +#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) + +DEFINE_PER_THREAD(int, smp_processor_id); + +#define per_cpu(name, thread) __per_cpu_##name[thread].v +#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) + +#define init_per_cpu(name, v) \ + do { \ + int __i_p_c_i; \ + for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ + per_cpu(name, __i_p_c_i) = v; \ + } while (0) + +/* + * CPU state checking (crowbarred). + */ + +#define idle_cpu(cpu) 0 +#define in_softirq() 1 +#define hardirq_count() 0 +#define PREEMPT_SHIFT 0 +#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) +#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) +#define PREEMPT_BITS 8 +#define SOFTIRQ_BITS 8 + +/* + * CPU hotplug. + */ + +struct notifier_block { + int (*notifier_call)(struct notifier_block *, unsigned long, void *); + struct notifier_block *next; + int priority; +}; + +#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ +#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ +#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ +#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ +#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ +#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ +#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, + * not handling interrupts, soon dead */ +#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug + * lock is dropped */ + +/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend + * operation in progress + */ +#define CPU_TASKS_FROZEN 0x0010 + +#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) +#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) +#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) +#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) +#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) +#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) +#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) + +/* Hibernation and suspend events */ +#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ +#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ +#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ +#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ +#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ +#define PM_POST_RESTORE 0x0006 /* Restore failed */ + +#define NOTIFY_DONE 0x0000 /* Don't care */ +#define NOTIFY_OK 0x0001 /* Suits me */ +#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ +#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) + /* Bad/Veto action */ +/* + * Clean way to return from the notifier and stop further calls. + */ +#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) + +/* + * Bug checks. + */ + +#define BUG_ON(c) do { if (!(c)) abort(); } while (0) + +/* + * Initialization -- Must be called before calling any primitives. + */ + +static void smp_init(void) +{ + int i; + + spin_lock_init(&__thread_id_map_mutex); + __thread_id_map[0] = pthread_self(); + for (i = 1; i < NR_THREADS; i++) + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + init_per_thread(smp_processor_id, 0); + if (pthread_key_create(&thread_id_key, NULL) != 0) { + perror("pthread_key_create"); + exit(-1); + } +} + +/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ + +#ifndef _LINUX_LIST_H +#define _LINUX_LIST_H + +#define LIST_POISON1 ((void *) 0x00100100) +#define LIST_POISON2 ((void *) 0x00200200) + +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Simple doubly linked list implementation. + * + * Some of the internal functions ("__xxx") are useful when + * manipulating whole lists rather than single entries, as + * sometimes we already know the next/prev entries and we can + * generate better code by using them directly rather than + * using the generic single-entry routines. + */ + +struct list_head { + struct list_head *next, *prev; +}; + +#define LIST_HEAD_INIT(name) { &(name), &(name) } + +#define LIST_HEAD(name) \ + struct list_head name = LIST_HEAD_INIT(name) + +static inline void INIT_LIST_HEAD(struct list_head *list) +{ + list->next = list; + list->prev = list; +} + +/* + * Insert a new entry between two known consecutive entries. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +#ifndef CONFIG_DEBUG_LIST +static inline void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next) +{ + next->prev = new; + new->next = next; + new->prev = prev; + prev->next = new; +} +#else +extern void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next); +#endif + +/** + * list_add - add a new entry + * @new: new entry to be added + * @head: list head to add it after + * + * Insert a new entry after the specified head. + * This is good for implementing stacks. + */ +static inline void list_add(struct list_head *new, struct list_head *head) +{ + __list_add(new, head, head->next); +} + + +/** + * list_add_tail - add a new entry + * @new: new entry to be added + * @head: list head to add it before + * + * Insert a new entry before the specified head. + * This is useful for implementing queues. + */ +static inline void list_add_tail(struct list_head *new, struct list_head *head) +{ + __list_add(new, head->prev, head); +} + +/* + * Delete a list entry by making the prev/next entries + * point to each other. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +static inline void __list_del(struct list_head * prev, struct list_head * next) +{ + next->prev = prev; + prev->next = next; +} + +/** + * list_del - deletes entry from list. + * @entry: the element to delete from the list. + * Note: list_empty() on entry does not return true after this, the entry is + * in an undefined state. + */ +#ifndef CONFIG_DEBUG_LIST +static inline void list_del(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + entry->next = LIST_POISON1; + entry->prev = LIST_POISON2; +} +#else +extern void list_del(struct list_head *entry); +#endif + +/** + * list_replace - replace old entry by new one + * @old : the element to be replaced + * @new : the new element to insert + * + * If @old was empty, it will be overwritten. + */ +static inline void list_replace(struct list_head *old, + struct list_head *new) +{ + new->next = old->next; + new->next->prev = new; + new->prev = old->prev; + new->prev->next = new; +} + +static inline void list_replace_init(struct list_head *old, + struct list_head *new) +{ + list_replace(old, new); + INIT_LIST_HEAD(old); +} + +/** + * list_del_init - deletes entry from list and reinitialize it. + * @entry: the element to delete from the list. + */ +static inline void list_del_init(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + INIT_LIST_HEAD(entry); +} + +/** + * list_move - delete from one list and add as another's head + * @list: the entry to move + * @head: the head that will precede our entry + */ +static inline void list_move(struct list_head *list, struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add(list, head); +} + +/** + * list_move_tail - delete from one list and add as another's tail + * @list: the entry to move + * @head: the head that will follow our entry + */ +static inline void list_move_tail(struct list_head *list, + struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add_tail(list, head); +} + +/** + * list_is_last - tests whether @list is the last entry in list @head + * @list: the entry to test + * @head: the head of the list + */ +static inline int list_is_last(const struct list_head *list, + const struct list_head *head) +{ + return list->next == head; +} + +/** + * list_empty - tests whether a list is empty + * @head: the list to test. + */ +static inline int list_empty(const struct list_head *head) +{ + return head->next == head; +} + +/** + * list_empty_careful - tests whether a list is empty and not being modified + * @head: the list to test + * + * Description: + * tests whether a list is empty _and_ checks that no other CPU might be + * in the process of modifying either member (next or prev) + * + * NOTE: using list_empty_careful() without synchronization + * can only be safe if the only activity that can happen + * to the list entry is list_del_init(). Eg. it cannot be used + * if another CPU could re-list_add() it. + */ +static inline int list_empty_careful(const struct list_head *head) +{ + struct list_head *next = head->next; + return (next == head) && (next == head->prev); +} + +/** + * list_is_singular - tests whether a list has just one entry. + * @head: the list to test. + */ +static inline int list_is_singular(const struct list_head *head) +{ + return !list_empty(head) && (head->next == head->prev); +} + +static inline void __list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + struct list_head *new_first = entry->next; + list->next = head->next; + list->next->prev = list; + list->prev = entry; + entry->next = list; + head->next = new_first; + new_first->prev = head; +} + +/** + * list_cut_position - cut a list into two + * @list: a new list to add all removed entries + * @head: a list with entries + * @entry: an entry within head, could be the head itself + * and if so we won't cut the list + * + * This helper moves the initial part of @head, up to and + * including @entry, from @head to @list. You should + * pass on @entry an element you know is on @head. @list + * should be an empty list or a list you do not care about + * losing its data. + * + */ +static inline void list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + if (list_empty(head)) + return; + if (list_is_singular(head) && + (head->next != entry && head != entry)) + return; + if (entry == head) + INIT_LIST_HEAD(list); + else + __list_cut_position(list, head, entry); +} + +static inline void __list_splice(const struct list_head *list, + struct list_head *prev, + struct list_head *next) +{ + struct list_head *first = list->next; + struct list_head *last = list->prev; + + first->prev = prev; + prev->next = first; + + last->next = next; + next->prev = last; +} + +/** + * list_splice - join two lists, this is designed for stacks + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice(const struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head, head->next); +} + +/** + * list_splice_tail - join two lists, each list being a queue + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice_tail(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head->prev, head); +} + +/** + * list_splice_init - join two lists and reinitialise the emptied list. + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * The list at @list is reinitialised + */ +static inline void list_splice_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head, head->next); + INIT_LIST_HEAD(list); + } +} + +/** + * list_splice_tail_init - join two lists and reinitialise the emptied list + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * Each of the lists is a queue. + * The list at @list is reinitialised + */ +static inline void list_splice_tail_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head->prev, head); + INIT_LIST_HEAD(list); + } +} + +/** + * list_entry - get the struct for this entry + * @ptr: the &struct list_head pointer. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + */ +#define list_entry(ptr, type, member) \ + container_of(ptr, type, member) + +/** + * list_first_entry - get the first element from a list + * @ptr: the list head to take the element from. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + * + * Note, that list is expected to be not empty. + */ +#define list_first_entry(ptr, type, member) \ + list_entry((ptr)->next, type, member) + +/** + * list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each(pos, head) \ + for (pos = (head)->next; prefetch(pos->next), pos != (head); \ + pos = pos->next) + +/** + * __list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + * + * This variant differs from list_for_each() in that it's the + * simplest possible list iteration code, no prefetching is done. + * Use this for code that knows the list to be very short (empty + * or 1 entry) most of the time. + */ +#define __list_for_each(pos, head) \ + for (pos = (head)->next; pos != (head); pos = pos->next) + +/** + * list_for_each_prev - iterate over a list backwards + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each_prev(pos, head) \ + for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ + pos = pos->prev) + +/** + * list_for_each_safe - iterate over a list safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_safe(pos, n, head) \ + for (pos = (head)->next, n = pos->next; pos != (head); \ + pos = n, n = pos->next) + +/** + * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_prev_safe(pos, n, head) \ + for (pos = (head)->prev, n = pos->prev; \ + prefetch(pos->prev), pos != (head); \ + pos = n, n = pos->prev) + +/** + * list_for_each_entry - iterate over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry(pos, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_reverse - iterate backwards over list of given type. + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_reverse(pos, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() + * @pos: the type * to use as a start point + * @head: the head of the list + * @member: the name of the list_struct within the struct. + * + * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). + */ +#define list_prepare_entry(pos, head, member) \ + ((pos) ? : list_entry(head, typeof(*pos), member)) + +/** + * list_for_each_entry_continue - continue iteration over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Continue to iterate over list of given type, continuing after + * the current position. + */ +#define list_for_each_entry_continue(pos, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_continue_reverse - iterate backwards from the given point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Start to iterate over list of given type backwards, continuing after + * the current position. + */ +#define list_for_each_entry_continue_reverse(pos, head, member) \ + for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_for_each_entry_from - iterate over list of given type from the current point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing from current position. + */ +#define list_for_each_entry_from(pos, head, member) \ + for (; prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_safe(pos, n, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_continue + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing after current point, + * safe against removal of list entry. + */ +#define list_for_each_entry_safe_continue(pos, n, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_from + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type from current point, safe against + * removal of list entry. + */ +#define list_for_each_entry_safe_from(pos, n, head, member) \ + for (n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_reverse + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate backwards over list of given type, safe against removal + * of list entry. + */ +#define list_for_each_entry_safe_reverse(pos, n, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member), \ + n = list_entry(pos->member.prev, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.prev, typeof(*n), member)) + +/* + * Double linked lists with a single pointer list head. + * Mostly useful for hash tables where the two pointer list head is + * too wasteful. + * You lose the ability to access the tail in O(1). + */ + +struct hlist_head { + struct hlist_node *first; +}; + +struct hlist_node { + struct hlist_node *next, **pprev; +}; + +#define HLIST_HEAD_INIT { .first = NULL } +#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } +#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) +static inline void INIT_HLIST_NODE(struct hlist_node *h) +{ + h->next = NULL; + h->pprev = NULL; +} + +static inline int hlist_unhashed(const struct hlist_node *h) +{ + return !h->pprev; +} + +static inline int hlist_empty(const struct hlist_head *h) +{ + return !h->first; +} + +static inline void __hlist_del(struct hlist_node *n) +{ + struct hlist_node *next = n->next; + struct hlist_node **pprev = n->pprev; + *pprev = next; + if (next) + next->pprev = pprev; +} + +static inline void hlist_del(struct hlist_node *n) +{ + __hlist_del(n); + n->next = LIST_POISON1; + n->pprev = LIST_POISON2; +} + +static inline void hlist_del_init(struct hlist_node *n) +{ + if (!hlist_unhashed(n)) { + __hlist_del(n); + INIT_HLIST_NODE(n); + } +} + +static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) +{ + struct hlist_node *first = h->first; + n->next = first; + if (first) + first->pprev = &n->next; + h->first = n; + n->pprev = &h->first; +} + +/* next must be != NULL */ +static inline void hlist_add_before(struct hlist_node *n, + struct hlist_node *next) +{ + n->pprev = next->pprev; + n->next = next; + next->pprev = &n->next; + *(n->pprev) = n; +} + +static inline void hlist_add_after(struct hlist_node *n, + struct hlist_node *next) +{ + next->next = n->next; + n->next = next; + next->pprev = &n->next; + + if(next->next) + next->next->pprev = &next->next; +} + +/* + * Move a list from one list head to another. Fixup the pprev + * reference of the first entry if it exists. + */ +static inline void hlist_move_list(struct hlist_head *old, + struct hlist_head *new) +{ + new->first = old->first; + if (new->first) + new->first->pprev = &new->first; + old->first = NULL; +} + +#define hlist_entry(ptr, type, member) container_of(ptr,type,member) + +#define hlist_for_each(pos, head) \ + for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ + pos = pos->next) + +#define hlist_for_each_safe(pos, n, head) \ + for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ + pos = n) + +/** + * hlist_for_each_entry - iterate over list of given type + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry(tpos, pos, head, member) \ + for (pos = (head)->first; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_continue - iterate over a hlist continuing after current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_continue(tpos, pos, member) \ + for (pos = (pos)->next; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_from - iterate over a hlist continuing from current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_from(tpos, pos, member) \ + for (; pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @n: another &struct hlist_node to use as temporary storage + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ + for (pos = (head)->first; \ + pos && ({ n = pos->next; 1; }) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = n) + +#endif diff --git a/tests/api_x86.h b/tests/api_x86.h new file mode 100644 index 0000000..f48fce9 --- /dev/null +++ b/tests/api_x86.h @@ -0,0 +1,1387 @@ +/* MECHANICALLY GENERATED, DO NOT EDIT!!! */ + +#define _INCLUDE_API_H + +/* + * common.h: Common Linux kernel-isms. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; but version 2 of the License only due + * to code included from the Linux kernel. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +#ifndef __always_inline +#define __always_inline inline +#endif + +#define BUILD_BUG_ON(condition) ((void)sizeof(char[1 - 2*!!(condition)])) +#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1) + +#ifdef __ASSEMBLY__ +# define stringify_in_c(...) __VA_ARGS__ +# define ASM_CONST(x) x +#else +/* This version of stringify will deal with commas... */ +# define __stringify_in_c(...) #__VA_ARGS__ +# define stringify_in_c(...) __stringify_in_c(__VA_ARGS__) " " +# define __ASM_CONST(x) x##UL +# define ASM_CONST(x) __ASM_CONST(x) +#endif + + +/* + * arch-i386.h: Expose x86 atomic instructions. 80486 and better only. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, but version 2 only due to inclusion + * of Linux-kernel code. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + * + * Much code taken from the Linux kernel. For such code, the option + * to redistribute under later versions of GPL might not be available. + */ + +/* + * Machine parameters. + */ + +#define CACHE_LINE_SIZE 64 +#define ____cacheline_internodealigned_in_smp \ + __attribute__((__aligned__(1 << 6))) + +#define LOCK_PREFIX "lock ; " + +/* + * Atomic data structure, initialization, and access. + */ + +typedef struct { volatile int counter; } atomic_t; + +#define ATOMIC_INIT(i) { (i) } + +#define atomic_read(v) ((v)->counter) +#define atomic_set(v, i) (((v)->counter) = (i)) + +/* + * Atomic operations. + */ + +/** + * atomic_add - add integer to atomic variable + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @i to @v. + */ +static __inline__ void atomic_add(int i, atomic_t *v) +{ + __asm__ __volatile__( + LOCK_PREFIX "addl %1,%0" + :"+m" (v->counter) + :"ir" (i)); +} + +/** + * atomic_sub - subtract the atomic variable + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v. + */ +static __inline__ void atomic_sub(int i, atomic_t *v) +{ + __asm__ __volatile__( + LOCK_PREFIX "subl %1,%0" + :"+m" (v->counter) + :"ir" (i)); +} + +/** + * atomic_sub_and_test - subtract value from variable and test result + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v and returns + * true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_sub_and_test(int i, atomic_t *v) +{ + unsigned char c; + + __asm__ __volatile__( + LOCK_PREFIX "subl %2,%0; sete %1" + :"+m" (v->counter), "=qm" (c) + :"ir" (i) : "memory"); + return c; +} + +/** + * atomic_inc - increment atomic variable + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1. + */ +static __inline__ void atomic_inc(atomic_t *v) +{ + __asm__ __volatile__( + LOCK_PREFIX "incl %0" + :"+m" (v->counter)); +} + +/** + * atomic_dec - decrement atomic variable + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1. + */ +static __inline__ void atomic_dec(atomic_t *v) +{ + __asm__ __volatile__( + LOCK_PREFIX "decl %0" + :"+m" (v->counter)); +} + +/** + * atomic_dec_and_test - decrement and test + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +static __inline__ int atomic_dec_and_test(atomic_t *v) +{ + unsigned char c; + + __asm__ __volatile__( + LOCK_PREFIX "decl %0; sete %1" + :"+m" (v->counter), "=qm" (c) + : : "memory"); + return c != 0; +} + +/** + * atomic_inc_and_test - increment and test + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +static __inline__ int atomic_inc_and_test(atomic_t *v) +{ + unsigned char c; + + __asm__ __volatile__( + LOCK_PREFIX "incl %0; sete %1" + :"+m" (v->counter), "=qm" (c) + : : "memory"); + return c != 0; +} + +/** + * atomic_add_negative - add and test if negative + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +static __inline__ int atomic_add_negative(int i, atomic_t *v) +{ + unsigned char c; + + __asm__ __volatile__( + LOCK_PREFIX "addl %2,%0; sets %1" + :"+m" (v->counter), "=qm" (c) + :"ir" (i) : "memory"); + return c; +} + +/** + * atomic_add_return - add and return + * @v: pointer of type atomic_t + * @i: integer value to add + * + * Atomically adds @i to @v and returns @i + @v + */ +static __inline__ int atomic_add_return(int i, atomic_t *v) +{ + int __i; + + __i = i; + __asm__ __volatile__( + LOCK_PREFIX "xaddl %0, %1;" + :"=r"(i) + :"m"(v->counter), "0"(i)); + return i + __i; +} + +static __inline__ int atomic_sub_return(int i, atomic_t *v) +{ + return atomic_add_return(-i,v); +} + +static inline unsigned int +cmpxchg(volatile long *ptr, long oldval, long newval) +{ + unsigned long retval; + + asm("# cmpxchg\n" + "lock; cmpxchgl %4,(%2)\n" + "# end atomic_cmpxchg4" + : "=a" (retval), "=m" (*ptr) + : "r" (ptr), "0" (oldval), "r" (newval), "m" (*ptr) + : "cc"); + return (retval); +} + +#define atomic_cmpxchg(v, old, new) ((int)cmpxchg(&((v)->counter), old, new)) +#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) + +/** + * atomic_add_unless - add unless the number is a given value + * @v: pointer of type atomic_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @v, so long as it was not @u. + * Returns non-zero if @v was not @u, and zero otherwise. + */ +#define atomic_add_unless(v, a, u) \ +({ \ + int c, old; \ + c = atomic_read(v); \ + for (;;) { \ + if (unlikely(c == (u))) \ + break; \ + old = atomic_cmpxchg((v), c, c + (a)); \ + if (likely(old == c)) \ + break; \ + c = old; \ + } \ + c != (u); \ +}) +#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) + +#define atomic_inc_return(v) (atomic_add_return(1,v)) +#define atomic_dec_return(v) (atomic_sub_return(1,v)) + +/* These are x86-specific, used by some header files */ +#define atomic_clear_mask(mask, addr) \ +__asm__ __volatile__(LOCK_PREFIX "andl %0,%1" \ +: : "r" (~(mask)),"m" (*addr) : "memory") + +#define atomic_set_mask(mask, addr) \ +__asm__ __volatile__(LOCK_PREFIX "orl %0,%1" \ +: : "r" (mask),"m" (*(addr)) : "memory") + +/* Atomic operations are already serializing on x86 */ +#define smp_mb__before_atomic_dec() barrier() +#define smp_mb__after_atomic_dec() barrier() +#define smp_mb__before_atomic_inc() barrier() +#define smp_mb__after_atomic_inc() barrier() + +/* + * api_pthreads.h: API mapping to pthreads environment. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. However, please note that much + * of the code in this file derives from the Linux kernel, and that such + * code may not be available except under GPLv2. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2006 Paul E. McKenney, IBM. + */ + +#include +#include +#include +#include +#include +#define __USE_GNU +#include +#include +#include +/* #include "atomic.h" */ + +/* + * Compiler magic. + */ +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Default machine parameters. + */ + +#ifndef CACHE_LINE_SIZE +#define CACHE_LINE_SIZE 128 +#endif /* #ifndef CACHE_LINE_SIZE */ + +/* + * Exclusive locking primitives. + */ + +typedef pthread_mutex_t spinlock_t; + +#define DEFINE_SPINLOCK(lock) spinlock_t lock = PTHREAD_MUTEX_INITIALIZER; +#define __SPIN_LOCK_UNLOCKED(lockp) PTHREAD_MUTEX_INITIALIZER + +static void spin_lock_init(spinlock_t *sp) +{ + if (pthread_mutex_init(sp, NULL) != 0) { + perror("spin_lock_init:pthread_mutex_init"); + exit(-1); + } +} + +static void spin_lock(spinlock_t *sp) +{ + if (pthread_mutex_lock(sp) != 0) { + perror("spin_lock:pthread_mutex_lock"); + exit(-1); + } +} + +static void spin_unlock(spinlock_t *sp) +{ + if (pthread_mutex_unlock(sp) != 0) { + perror("spin_unlock:pthread_mutex_unlock"); + exit(-1); + } +} + +#define spin_lock_irqsave(l, f) do { f = 1; spin_lock(l); } while (0) +#define spin_unlock_irqrestore(l, f) do { f = 0; spin_unlock(l); } while (0) + +/* + * Thread creation/destruction primitives. + */ + +typedef pthread_t thread_id_t; + +#define NR_THREADS 128 + +#define __THREAD_ID_MAP_EMPTY 0 +#define __THREAD_ID_MAP_WAITING 1 +thread_id_t __thread_id_map[NR_THREADS]; +spinlock_t __thread_id_map_mutex; + +#define for_each_thread(t) \ + for (t = 0; t < NR_THREADS; t++) + +#define for_each_running_thread(t) \ + for (t = 0; t < NR_THREADS; t++) \ + if ((__thread_id_map[t] != __THREAD_ID_MAP_EMPTY) && \ + (__thread_id_map[t] != __THREAD_ID_MAP_WAITING)) + +pthread_key_t thread_id_key; + +static int __smp_thread_id(void) +{ + int i; + thread_id_t tid = pthread_self(); + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) { + long v = i + 1; /* must be non-NULL. */ + + if (pthread_setspecific(thread_id_key, (void *)v) != 0) { + perror("pthread_setspecific"); + exit(-1); + } + return i; + } + } + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + spin_unlock(&__thread_id_map_mutex); + return i; + } + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "smp_thread_id: Rogue thread, id: %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); +} + +static int smp_thread_id(void) +{ + void *id; + + id = pthread_getspecific(thread_id_key); + if (id == NULL) + return __smp_thread_id(); + return (long)(id - 1); +} + +static thread_id_t create_thread(void *(*func)(void *), void *arg) +{ + thread_id_t tid; + int i; + + spin_lock(&__thread_id_map_mutex); + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == __THREAD_ID_MAP_EMPTY) + break; + } + if (i >= NR_THREADS) { + spin_unlock(&__thread_id_map_mutex); + fprintf(stderr, "Thread limit of %d exceeded!\n", NR_THREADS); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_WAITING; + spin_unlock(&__thread_id_map_mutex); + if (pthread_create(&tid, NULL, func, arg) != 0) { + perror("create_thread:pthread_create"); + exit(-1); + } + __thread_id_map[i] = tid; + return tid; +} + +static void *wait_thread(thread_id_t tid) +{ + int i; + void *vp; + + for (i = 0; i < NR_THREADS; i++) { + if (__thread_id_map[i] == tid) + break; + } + if (i >= NR_THREADS){ + fprintf(stderr, "wait_thread: bad tid = %d(%#x)\n", + (int)tid, (int)tid); + exit(-1); + } + if (pthread_join(tid, &vp) != 0) { + perror("wait_thread:pthread_join"); + exit(-1); + } + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + return vp; +} + +static void wait_all_threads(void) +{ + int i; + thread_id_t tid; + + for (i = 1; i < NR_THREADS; i++) { + tid = __thread_id_map[i]; + if (tid != __THREAD_ID_MAP_EMPTY && + tid != __THREAD_ID_MAP_WAITING) + (void)wait_thread(tid); + } +} + +static void run_on(int cpu) +{ + cpu_set_t mask; + + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * timekeeping -- very crude -- should use MONOTONIC... + */ + +long long get_microseconds(void) +{ + struct timeval tv; + + if (gettimeofday(&tv, NULL) != 0) + abort(); + return ((long long)tv.tv_sec) * 1000000LL + (long long)tv.tv_usec; +} + +/* + * Per-thread variables. + */ + +#define DEFINE_PER_THREAD(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_thread_##name[NR_THREADS]; +#define DECLARE_PER_THREAD(type, name) extern DEFINE_PER_THREAD(type, name) + +#define per_thread(name, thread) __per_thread_##name[thread].v +#define __get_thread_var(name) per_thread(name, smp_thread_id()) + +#define init_per_thread(name, v) \ + do { \ + int __i_p_t_i; \ + for (__i_p_t_i = 0; __i_p_t_i < NR_THREADS; __i_p_t_i++) \ + per_thread(name, __i_p_t_i) = v; \ + } while (0) + +/* + * CPU traversal primitives. + */ + +#ifndef NR_CPUS +#define NR_CPUS 16 +#endif /* #ifndef NR_CPUS */ + +#define for_each_possible_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) +#define for_each_online_cpu(cpu) \ + for (cpu = 0; cpu < NR_CPUS; cpu++) + +/* + * Per-CPU variables. + */ + +#define DEFINE_PER_CPU(type, name) \ + struct { \ + __typeof__(type) v \ + __attribute__((__aligned__(CACHE_LINE_SIZE))); \ + } __per_cpu_##name[NR_CPUS] +#define DECLARE_PER_CPU(type, name) extern DEFINE_PER_CPU(type, name) + +DEFINE_PER_THREAD(int, smp_processor_id); + +#define per_cpu(name, thread) __per_cpu_##name[thread].v +#define __get_cpu_var(name) per_cpu(name, smp_processor_id()) + +#define init_per_cpu(name, v) \ + do { \ + int __i_p_c_i; \ + for (__i_p_c_i = 0; __i_p_c_i < NR_CPUS; __i_p_c_i++) \ + per_cpu(name, __i_p_c_i) = v; \ + } while (0) + +/* + * CPU state checking (crowbarred). + */ + +#define idle_cpu(cpu) 0 +#define in_softirq() 1 +#define hardirq_count() 0 +#define PREEMPT_SHIFT 0 +#define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) +#define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) +#define PREEMPT_BITS 8 +#define SOFTIRQ_BITS 8 + +/* + * CPU hotplug. + */ + +struct notifier_block { + int (*notifier_call)(struct notifier_block *, unsigned long, void *); + struct notifier_block *next; + int priority; +}; + +#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ +#define CPU_UP_PREPARE 0x0003 /* CPU (unsigned)v coming up */ +#define CPU_UP_CANCELED 0x0004 /* CPU (unsigned)v NOT coming up */ +#define CPU_DOWN_PREPARE 0x0005 /* CPU (unsigned)v going down */ +#define CPU_DOWN_FAILED 0x0006 /* CPU (unsigned)v NOT going down */ +#define CPU_DEAD 0x0007 /* CPU (unsigned)v dead */ +#define CPU_DYING 0x0008 /* CPU (unsigned)v not running any task, + * not handling interrupts, soon dead */ +#define CPU_POST_DEAD 0x0009 /* CPU (unsigned)v dead, cpu_hotplug + * lock is dropped */ + +/* Used for CPU hotplug events occuring while tasks are frozen due to a suspend + * operation in progress + */ +#define CPU_TASKS_FROZEN 0x0010 + +#define CPU_ONLINE_FROZEN (CPU_ONLINE | CPU_TASKS_FROZEN) +#define CPU_UP_PREPARE_FROZEN (CPU_UP_PREPARE | CPU_TASKS_FROZEN) +#define CPU_UP_CANCELED_FROZEN (CPU_UP_CANCELED | CPU_TASKS_FROZEN) +#define CPU_DOWN_PREPARE_FROZEN (CPU_DOWN_PREPARE | CPU_TASKS_FROZEN) +#define CPU_DOWN_FAILED_FROZEN (CPU_DOWN_FAILED | CPU_TASKS_FROZEN) +#define CPU_DEAD_FROZEN (CPU_DEAD | CPU_TASKS_FROZEN) +#define CPU_DYING_FROZEN (CPU_DYING | CPU_TASKS_FROZEN) + +/* Hibernation and suspend events */ +#define PM_HIBERNATION_PREPARE 0x0001 /* Going to hibernate */ +#define PM_POST_HIBERNATION 0x0002 /* Hibernation finished */ +#define PM_SUSPEND_PREPARE 0x0003 /* Going to suspend the system */ +#define PM_POST_SUSPEND 0x0004 /* Suspend finished */ +#define PM_RESTORE_PREPARE 0x0005 /* Going to restore a saved image */ +#define PM_POST_RESTORE 0x0006 /* Restore failed */ + +#define NOTIFY_DONE 0x0000 /* Don't care */ +#define NOTIFY_OK 0x0001 /* Suits me */ +#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */ +#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) + /* Bad/Veto action */ +/* + * Clean way to return from the notifier and stop further calls. + */ +#define NOTIFY_STOP (NOTIFY_OK|NOTIFY_STOP_MASK) + +/* + * Bug checks. + */ + +#define BUG_ON(c) do { if (!(c)) abort(); } while (0) + +/* + * Initialization -- Must be called before calling any primitives. + */ + +static void smp_init(void) +{ + int i; + + spin_lock_init(&__thread_id_map_mutex); + __thread_id_map[0] = pthread_self(); + for (i = 1; i < NR_THREADS; i++) + __thread_id_map[i] = __THREAD_ID_MAP_EMPTY; + init_per_thread(smp_processor_id, 0); + if (pthread_key_create(&thread_id_key, NULL) != 0) { + perror("pthread_key_create"); + exit(-1); + } +} + +/* Taken from the Linux kernel source tree, so GPLv2-only!!! */ + +#ifndef _LINUX_LIST_H +#define _LINUX_LIST_H + +#define LIST_POISON1 ((void *) 0x00100100) +#define LIST_POISON2 ((void *) 0x00200200) + +#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) +#define container_of(ptr, type, member) ({ \ + const typeof( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - offsetof(type,member) );}) + +/* + * Simple doubly linked list implementation. + * + * Some of the internal functions ("__xxx") are useful when + * manipulating whole lists rather than single entries, as + * sometimes we already know the next/prev entries and we can + * generate better code by using them directly rather than + * using the generic single-entry routines. + */ + +struct list_head { + struct list_head *next, *prev; +}; + +#define LIST_HEAD_INIT(name) { &(name), &(name) } + +#define LIST_HEAD(name) \ + struct list_head name = LIST_HEAD_INIT(name) + +static inline void INIT_LIST_HEAD(struct list_head *list) +{ + list->next = list; + list->prev = list; +} + +/* + * Insert a new entry between two known consecutive entries. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +#ifndef CONFIG_DEBUG_LIST +static inline void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next) +{ + next->prev = new; + new->next = next; + new->prev = prev; + prev->next = new; +} +#else +extern void __list_add(struct list_head *new, + struct list_head *prev, + struct list_head *next); +#endif + +/** + * list_add - add a new entry + * @new: new entry to be added + * @head: list head to add it after + * + * Insert a new entry after the specified head. + * This is good for implementing stacks. + */ +static inline void list_add(struct list_head *new, struct list_head *head) +{ + __list_add(new, head, head->next); +} + + +/** + * list_add_tail - add a new entry + * @new: new entry to be added + * @head: list head to add it before + * + * Insert a new entry before the specified head. + * This is useful for implementing queues. + */ +static inline void list_add_tail(struct list_head *new, struct list_head *head) +{ + __list_add(new, head->prev, head); +} + +/* + * Delete a list entry by making the prev/next entries + * point to each other. + * + * This is only for internal list manipulation where we know + * the prev/next entries already! + */ +static inline void __list_del(struct list_head * prev, struct list_head * next) +{ + next->prev = prev; + prev->next = next; +} + +/** + * list_del - deletes entry from list. + * @entry: the element to delete from the list. + * Note: list_empty() on entry does not return true after this, the entry is + * in an undefined state. + */ +#ifndef CONFIG_DEBUG_LIST +static inline void list_del(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + entry->next = LIST_POISON1; + entry->prev = LIST_POISON2; +} +#else +extern void list_del(struct list_head *entry); +#endif + +/** + * list_replace - replace old entry by new one + * @old : the element to be replaced + * @new : the new element to insert + * + * If @old was empty, it will be overwritten. + */ +static inline void list_replace(struct list_head *old, + struct list_head *new) +{ + new->next = old->next; + new->next->prev = new; + new->prev = old->prev; + new->prev->next = new; +} + +static inline void list_replace_init(struct list_head *old, + struct list_head *new) +{ + list_replace(old, new); + INIT_LIST_HEAD(old); +} + +/** + * list_del_init - deletes entry from list and reinitialize it. + * @entry: the element to delete from the list. + */ +static inline void list_del_init(struct list_head *entry) +{ + __list_del(entry->prev, entry->next); + INIT_LIST_HEAD(entry); +} + +/** + * list_move - delete from one list and add as another's head + * @list: the entry to move + * @head: the head that will precede our entry + */ +static inline void list_move(struct list_head *list, struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add(list, head); +} + +/** + * list_move_tail - delete from one list and add as another's tail + * @list: the entry to move + * @head: the head that will follow our entry + */ +static inline void list_move_tail(struct list_head *list, + struct list_head *head) +{ + __list_del(list->prev, list->next); + list_add_tail(list, head); +} + +/** + * list_is_last - tests whether @list is the last entry in list @head + * @list: the entry to test + * @head: the head of the list + */ +static inline int list_is_last(const struct list_head *list, + const struct list_head *head) +{ + return list->next == head; +} + +/** + * list_empty - tests whether a list is empty + * @head: the list to test. + */ +static inline int list_empty(const struct list_head *head) +{ + return head->next == head; +} + +/** + * list_empty_careful - tests whether a list is empty and not being modified + * @head: the list to test + * + * Description: + * tests whether a list is empty _and_ checks that no other CPU might be + * in the process of modifying either member (next or prev) + * + * NOTE: using list_empty_careful() without synchronization + * can only be safe if the only activity that can happen + * to the list entry is list_del_init(). Eg. it cannot be used + * if another CPU could re-list_add() it. + */ +static inline int list_empty_careful(const struct list_head *head) +{ + struct list_head *next = head->next; + return (next == head) && (next == head->prev); +} + +/** + * list_is_singular - tests whether a list has just one entry. + * @head: the list to test. + */ +static inline int list_is_singular(const struct list_head *head) +{ + return !list_empty(head) && (head->next == head->prev); +} + +static inline void __list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + struct list_head *new_first = entry->next; + list->next = head->next; + list->next->prev = list; + list->prev = entry; + entry->next = list; + head->next = new_first; + new_first->prev = head; +} + +/** + * list_cut_position - cut a list into two + * @list: a new list to add all removed entries + * @head: a list with entries + * @entry: an entry within head, could be the head itself + * and if so we won't cut the list + * + * This helper moves the initial part of @head, up to and + * including @entry, from @head to @list. You should + * pass on @entry an element you know is on @head. @list + * should be an empty list or a list you do not care about + * losing its data. + * + */ +static inline void list_cut_position(struct list_head *list, + struct list_head *head, struct list_head *entry) +{ + if (list_empty(head)) + return; + if (list_is_singular(head) && + (head->next != entry && head != entry)) + return; + if (entry == head) + INIT_LIST_HEAD(list); + else + __list_cut_position(list, head, entry); +} + +static inline void __list_splice(const struct list_head *list, + struct list_head *prev, + struct list_head *next) +{ + struct list_head *first = list->next; + struct list_head *last = list->prev; + + first->prev = prev; + prev->next = first; + + last->next = next; + next->prev = last; +} + +/** + * list_splice - join two lists, this is designed for stacks + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice(const struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head, head->next); +} + +/** + * list_splice_tail - join two lists, each list being a queue + * @list: the new list to add. + * @head: the place to add it in the first list. + */ +static inline void list_splice_tail(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) + __list_splice(list, head->prev, head); +} + +/** + * list_splice_init - join two lists and reinitialise the emptied list. + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * The list at @list is reinitialised + */ +static inline void list_splice_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head, head->next); + INIT_LIST_HEAD(list); + } +} + +/** + * list_splice_tail_init - join two lists and reinitialise the emptied list + * @list: the new list to add. + * @head: the place to add it in the first list. + * + * Each of the lists is a queue. + * The list at @list is reinitialised + */ +static inline void list_splice_tail_init(struct list_head *list, + struct list_head *head) +{ + if (!list_empty(list)) { + __list_splice(list, head->prev, head); + INIT_LIST_HEAD(list); + } +} + +/** + * list_entry - get the struct for this entry + * @ptr: the &struct list_head pointer. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + */ +#define list_entry(ptr, type, member) \ + container_of(ptr, type, member) + +/** + * list_first_entry - get the first element from a list + * @ptr: the list head to take the element from. + * @type: the type of the struct this is embedded in. + * @member: the name of the list_struct within the struct. + * + * Note, that list is expected to be not empty. + */ +#define list_first_entry(ptr, type, member) \ + list_entry((ptr)->next, type, member) + +/** + * list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each(pos, head) \ + for (pos = (head)->next; prefetch(pos->next), pos != (head); \ + pos = pos->next) + +/** + * __list_for_each - iterate over a list + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + * + * This variant differs from list_for_each() in that it's the + * simplest possible list iteration code, no prefetching is done. + * Use this for code that knows the list to be very short (empty + * or 1 entry) most of the time. + */ +#define __list_for_each(pos, head) \ + for (pos = (head)->next; pos != (head); pos = pos->next) + +/** + * list_for_each_prev - iterate over a list backwards + * @pos: the &struct list_head to use as a loop cursor. + * @head: the head for your list. + */ +#define list_for_each_prev(pos, head) \ + for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \ + pos = pos->prev) + +/** + * list_for_each_safe - iterate over a list safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_safe(pos, n, head) \ + for (pos = (head)->next, n = pos->next; pos != (head); \ + pos = n, n = pos->next) + +/** + * list_for_each_prev_safe - iterate over a list backwards safe against removal of list entry + * @pos: the &struct list_head to use as a loop cursor. + * @n: another &struct list_head to use as temporary storage + * @head: the head for your list. + */ +#define list_for_each_prev_safe(pos, n, head) \ + for (pos = (head)->prev, n = pos->prev; \ + prefetch(pos->prev), pos != (head); \ + pos = n, n = pos->prev) + +/** + * list_for_each_entry - iterate over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry(pos, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_reverse - iterate backwards over list of given type. + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_reverse(pos, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue() + * @pos: the type * to use as a start point + * @head: the head of the list + * @member: the name of the list_struct within the struct. + * + * Prepares a pos entry for use as a start point in list_for_each_entry_continue(). + */ +#define list_prepare_entry(pos, head, member) \ + ((pos) ? : list_entry(head, typeof(*pos), member)) + +/** + * list_for_each_entry_continue - continue iteration over list of given type + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Continue to iterate over list of given type, continuing after + * the current position. + */ +#define list_for_each_entry_continue(pos, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member); \ + prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_continue_reverse - iterate backwards from the given point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Start to iterate over list of given type backwards, continuing after + * the current position. + */ +#define list_for_each_entry_continue_reverse(pos, head, member) \ + for (pos = list_entry(pos->member.prev, typeof(*pos), member); \ + prefetch(pos->member.prev), &pos->member != (head); \ + pos = list_entry(pos->member.prev, typeof(*pos), member)) + +/** + * list_for_each_entry_from - iterate over list of given type from the current point + * @pos: the type * to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing from current position. + */ +#define list_for_each_entry_from(pos, head, member) \ + for (; prefetch(pos->member.next), &pos->member != (head); \ + pos = list_entry(pos->member.next, typeof(*pos), member)) + +/** + * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + */ +#define list_for_each_entry_safe(pos, n, head, member) \ + for (pos = list_entry((head)->next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_continue + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type, continuing after current point, + * safe against removal of list entry. + */ +#define list_for_each_entry_safe_continue(pos, n, head, member) \ + for (pos = list_entry(pos->member.next, typeof(*pos), member), \ + n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_from + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate over list of given type from current point, safe against + * removal of list entry. + */ +#define list_for_each_entry_safe_from(pos, n, head, member) \ + for (n = list_entry(pos->member.next, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.next, typeof(*n), member)) + +/** + * list_for_each_entry_safe_reverse + * @pos: the type * to use as a loop cursor. + * @n: another type * to use as temporary storage + * @head: the head for your list. + * @member: the name of the list_struct within the struct. + * + * Iterate backwards over list of given type, safe against removal + * of list entry. + */ +#define list_for_each_entry_safe_reverse(pos, n, head, member) \ + for (pos = list_entry((head)->prev, typeof(*pos), member), \ + n = list_entry(pos->member.prev, typeof(*pos), member); \ + &pos->member != (head); \ + pos = n, n = list_entry(n->member.prev, typeof(*n), member)) + +/* + * Double linked lists with a single pointer list head. + * Mostly useful for hash tables where the two pointer list head is + * too wasteful. + * You lose the ability to access the tail in O(1). + */ + +struct hlist_head { + struct hlist_node *first; +}; + +struct hlist_node { + struct hlist_node *next, **pprev; +}; + +#define HLIST_HEAD_INIT { .first = NULL } +#define HLIST_HEAD(name) struct hlist_head name = { .first = NULL } +#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL) +static inline void INIT_HLIST_NODE(struct hlist_node *h) +{ + h->next = NULL; + h->pprev = NULL; +} + +static inline int hlist_unhashed(const struct hlist_node *h) +{ + return !h->pprev; +} + +static inline int hlist_empty(const struct hlist_head *h) +{ + return !h->first; +} + +static inline void __hlist_del(struct hlist_node *n) +{ + struct hlist_node *next = n->next; + struct hlist_node **pprev = n->pprev; + *pprev = next; + if (next) + next->pprev = pprev; +} + +static inline void hlist_del(struct hlist_node *n) +{ + __hlist_del(n); + n->next = LIST_POISON1; + n->pprev = LIST_POISON2; +} + +static inline void hlist_del_init(struct hlist_node *n) +{ + if (!hlist_unhashed(n)) { + __hlist_del(n); + INIT_HLIST_NODE(n); + } +} + +static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) +{ + struct hlist_node *first = h->first; + n->next = first; + if (first) + first->pprev = &n->next; + h->first = n; + n->pprev = &h->first; +} + +/* next must be != NULL */ +static inline void hlist_add_before(struct hlist_node *n, + struct hlist_node *next) +{ + n->pprev = next->pprev; + n->next = next; + next->pprev = &n->next; + *(n->pprev) = n; +} + +static inline void hlist_add_after(struct hlist_node *n, + struct hlist_node *next) +{ + next->next = n->next; + n->next = next; + next->pprev = &n->next; + + if(next->next) + next->next->pprev = &next->next; +} + +/* + * Move a list from one list head to another. Fixup the pprev + * reference of the first entry if it exists. + */ +static inline void hlist_move_list(struct hlist_head *old, + struct hlist_head *new) +{ + new->first = old->first; + if (new->first) + new->first->pprev = &new->first; + old->first = NULL; +} + +#define hlist_entry(ptr, type, member) container_of(ptr,type,member) + +#define hlist_for_each(pos, head) \ + for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \ + pos = pos->next) + +#define hlist_for_each_safe(pos, n, head) \ + for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \ + pos = n) + +/** + * hlist_for_each_entry - iterate over list of given type + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry(tpos, pos, head, member) \ + for (pos = (head)->first; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_continue - iterate over a hlist continuing after current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_continue(tpos, pos, member) \ + for (pos = (pos)->next; \ + pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_from - iterate over a hlist continuing from current point + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_from(tpos, pos, member) \ + for (; pos && ({ prefetch(pos->next); 1;}) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = pos->next) + +/** + * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry + * @tpos: the type * to use as a loop cursor. + * @pos: the &struct hlist_node to use as a loop cursor. + * @n: another &struct hlist_node to use as temporary storage + * @head: the head for your list. + * @member: the name of the hlist_node within the struct. + */ +#define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ + for (pos = (head)->first; \ + pos && ({ n = pos->next; 1; }) && \ + ({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \ + pos = n) + +#endif diff --git a/tests/rcutorture.h b/tests/rcutorture.h new file mode 100644 index 0000000..00b0b31 --- /dev/null +++ b/tests/rcutorture.h @@ -0,0 +1,417 @@ +/* + * rcutorture.h: simple user-level performance/stress test of RCU. + * + * Usage: + * ./rcu rperf [ ] + * Run a read-side performance test with the specified + * number of readers spaced by . + * Thus "./rcu 16 rperf 2" would run 16 readers on even-numbered + * CPUs from 0 to 30. + * ./rcu uperf [ ] + * Run an update-side performance test with the specified + * number of updaters and specified CPU spacing. + * ./rcu perf [ ] + * Run a combined read/update performance test with the specified + * number of readers and one updater and specified CPU spacing. + * The readers run on the low-numbered CPUs and the updater + * of the highest-numbered CPU. + * + * The above tests produce output as follows: + * + * n_reads: 46008000 n_updates: 146026 nreaders: 2 nupdaters: 1 duration: 1 + * ns/read: 43.4707 ns/update: 6848.1 + * + * The first line lists the total number of RCU reads and updates executed + * during the test, the number of reader threads, the number of updater + * threads, and the duration of the test in seconds. The second line + * lists the average duration of each type of operation in nanoseconds, + * or "nan" if the corresponding type of operation was not performed. + * + * ./rcu stress + * Run a stress test with the specified number of readers and + * one updater. None of the threads are affinitied to any + * particular CPU. + * + * This test produces output as follows: + * + * n_reads: 114633217 n_updates: 3903415 n_mberror: 0 + * rcu_stress_count: 114618391 14826 0 0 0 0 0 0 0 0 0 + * + * The first line lists the number of RCU read and update operations + * executed, followed by the number of memory-ordering violations + * (which will be zero in a correct RCU implementation). The second + * line lists the number of readers observing progressively more stale + * data. A correct RCU implementation will have all but the first two + * numbers non-zero. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (c) 2008 Paul E. McKenney, IBM Corporation. + */ + +/* + * Test variables. + */ + +DEFINE_PER_THREAD(long long, n_reads_pt); +DEFINE_PER_THREAD(long long, n_updates_pt); + +long long n_reads = 0LL; +long n_updates = 0L; +atomic_t nthreadsrunning; +char argsbuf[64]; + +#define GOFLAG_INIT 0 +#define GOFLAG_RUN 1 +#define GOFLAG_STOP 2 + +int goflag __attribute__((__aligned__(CACHE_LINE_SIZE))) = GOFLAG_INIT; + +#define RCU_READ_RUN 1000 + +//MD +#define RCU_READ_NESTABLE + +#ifdef RCU_READ_NESTABLE +#define rcu_read_lock_nest() rcu_read_lock() +#define rcu_read_unlock_nest() rcu_read_unlock() +#else /* #ifdef RCU_READ_NESTABLE */ +#define rcu_read_lock_nest() +#define rcu_read_unlock_nest() +#endif /* #else #ifdef RCU_READ_NESTABLE */ + +#ifndef mark_rcu_quiescent_state +#define mark_rcu_quiescent_state() do ; while (0) +#endif /* #ifdef mark_rcu_quiescent_state */ + +#ifndef put_thread_offline +#define put_thread_offline() do ; while (0) +#define put_thread_online() do ; while (0) +#define put_thread_online_delay() do ; while (0) +#else /* #ifndef put_thread_offline */ +#define put_thread_online_delay() synchronize_rcu() +#endif /* #else #ifndef put_thread_offline */ + +/* + * Performance test. + */ + +void *rcu_read_perf_test(void *arg) +{ + int i; + int me = (long)arg; + long long n_reads_local = 0; + + rcu_register_thread(); + run_on(me); + atomic_inc(&nthreadsrunning); + while (goflag == GOFLAG_INIT) + poll(NULL, 0, 1); + mark_rcu_quiescent_state(); + while (goflag == GOFLAG_RUN) { + for (i = 0; i < RCU_READ_RUN; i++) { + rcu_read_lock(); + /* rcu_read_lock_nest(); */ + /* rcu_read_unlock_nest(); */ + rcu_read_unlock(); + } + n_reads_local += RCU_READ_RUN; + mark_rcu_quiescent_state(); + } + __get_thread_var(n_reads_pt) += n_reads_local; + put_thread_offline(); + rcu_unregister_thread(); + + return (NULL); +} + +void *rcu_update_perf_test(void *arg) +{ + long long n_updates_local = 0; + + atomic_inc(&nthreadsrunning); + while (goflag == GOFLAG_INIT) + poll(NULL, 0, 1); + while (goflag == GOFLAG_RUN) { + synchronize_rcu(); + n_updates_local++; + } + __get_thread_var(n_updates_pt) += n_updates_local; + return NULL; +} + +void perftestinit(void) +{ + init_per_thread(n_reads_pt, 0LL); + init_per_thread(n_updates_pt, 0LL); + atomic_set(&nthreadsrunning, 0); +} + +void perftestrun(int nthreads, int nreaders, int nupdaters) +{ + int t; + int duration = 1; + + smp_mb(); + while (atomic_read(&nthreadsrunning) < nthreads) + poll(NULL, 0, 1); + goflag = GOFLAG_RUN; + smp_mb(); + sleep(duration); + smp_mb(); + goflag = GOFLAG_STOP; + smp_mb(); + wait_all_threads(); + for_each_thread(t) { + n_reads += per_thread(n_reads_pt, t); + n_updates += per_thread(n_updates_pt, t); + } + printf("n_reads: %lld n_updates: %ld nreaders: %d nupdaters: %d duration: %d\n", + n_reads, n_updates, nreaders, nupdaters, duration); + printf("ns/read: %g ns/update: %g\n", + ((duration * 1000*1000*1000.*(double)nreaders) / + (double)n_reads), + ((duration * 1000*1000*1000.*(double)nupdaters) / + (double)n_updates)); + exit(0); +} + +void perftest(int nreaders, int cpustride) +{ + int i; + long arg; + + perftestinit(); + for (i = 0; i < nreaders; i++) { + arg = (long)(i * cpustride); + create_thread(rcu_read_perf_test, (void *)arg); + } + arg = (long)(i * cpustride); + create_thread(rcu_update_perf_test, (void *)arg); + perftestrun(i + 1, nreaders, 1); +} + +void rperftest(int nreaders, int cpustride) +{ + int i; + long arg; + + perftestinit(); + init_per_thread(n_reads_pt, 0LL); + for (i = 0; i < nreaders; i++) { + arg = (long)(i * cpustride); + create_thread(rcu_read_perf_test, (void *)arg); + } + perftestrun(i, nreaders, 0); +} + +void uperftest(int nupdaters, int cpustride) +{ + int i; + long arg; + + perftestinit(); + init_per_thread(n_reads_pt, 0LL); + for (i = 0; i < nupdaters; i++) { + arg = (long)(i * cpustride); + create_thread(rcu_update_perf_test, (void *)arg); + } + perftestrun(i, 0, nupdaters); +} + +/* + * Stress test. + */ + +#define RCU_STRESS_PIPE_LEN 10 + +struct rcu_stress { + int pipe_count; + int mbtest; +}; + +struct rcu_stress rcu_stress_array[RCU_STRESS_PIPE_LEN] = { { 0 } }; +struct rcu_stress *rcu_stress_current; +int rcu_stress_idx = 0; + +int n_mberror = 0; +DEFINE_PER_THREAD(long long [RCU_STRESS_PIPE_LEN + 1], rcu_stress_count); + +int garbage = 0; + +void *rcu_read_stress_test(void *arg) +{ + int i; + int itercnt = 0; + struct rcu_stress *p; + int pc; + + rcu_register_thread(); + while (goflag == GOFLAG_INIT) + poll(NULL, 0, 1); + mark_rcu_quiescent_state(); + while (goflag == GOFLAG_RUN) { + rcu_read_lock(); + p = rcu_dereference(rcu_stress_current); + if (p->mbtest == 0) + n_mberror++; + rcu_read_lock_nest(); + for (i = 0; i < 100; i++) + garbage++; + rcu_read_unlock_nest(); + pc = p->pipe_count; + rcu_read_unlock(); + if ((pc > RCU_STRESS_PIPE_LEN) || (pc < 0)) + pc = RCU_STRESS_PIPE_LEN; + __get_thread_var(rcu_stress_count)[pc]++; + __get_thread_var(n_reads_pt)++; + mark_rcu_quiescent_state(); + if ((++itercnt % 0x1000) == 0) { + put_thread_offline(); + put_thread_online_delay(); + put_thread_online(); + } + } + put_thread_offline(); + rcu_unregister_thread(); + + return (NULL); +} + +void *rcu_update_stress_test(void *arg) +{ + int i; + struct rcu_stress *p; + + while (goflag == GOFLAG_INIT) + poll(NULL, 0, 1); + while (goflag == GOFLAG_RUN) { + i = rcu_stress_idx + 1; + if (i >= RCU_STRESS_PIPE_LEN) + i = 0; + p = &rcu_stress_array[i]; + p->mbtest = 0; + smp_mb(); + p->pipe_count = 0; + p->mbtest = 1; + rcu_assign_pointer(rcu_stress_current, p); + rcu_stress_idx = i; + for (i = 0; i < RCU_STRESS_PIPE_LEN; i++) + if (i != rcu_stress_idx) + rcu_stress_array[i].pipe_count++; + synchronize_rcu(); + n_updates++; + } + return NULL; +} + +void *rcu_fake_update_stress_test(void *arg) +{ + while (goflag == GOFLAG_INIT) + poll(NULL, 0, 1); + while (goflag == GOFLAG_RUN) { + synchronize_rcu(); + poll(NULL, 0, 1); + } + return NULL; +} + +void stresstest(int nreaders) +{ + int i; + int t; + long long *p; + long long sum; + + init_per_thread(n_reads_pt, 0LL); + for_each_thread(t) { + p = &per_thread(rcu_stress_count,t)[0]; + for (i = 0; i <= RCU_STRESS_PIPE_LEN; i++) + p[i] = 0LL; + } + rcu_stress_current = &rcu_stress_array[0]; + rcu_stress_current->pipe_count = 0; + rcu_stress_current->mbtest = 1; + for (i = 0; i < nreaders; i++) + create_thread(rcu_read_stress_test, NULL); + create_thread(rcu_update_stress_test, NULL); + for (i = 0; i < 5; i++) + create_thread(rcu_fake_update_stress_test, NULL); + smp_mb(); + goflag = GOFLAG_RUN; + smp_mb(); + sleep(10); + smp_mb(); + goflag = GOFLAG_STOP; + smp_mb(); + wait_all_threads(); + for_each_thread(t) + n_reads += per_thread(n_reads_pt, t); + printf("n_reads: %lld n_updates: %ld n_mberror: %d\n", + n_reads, n_updates, n_mberror); + printf("rcu_stress_count:"); + for (i = 0; i <= RCU_STRESS_PIPE_LEN; i++) { + sum = 0LL; + for_each_thread(t) { + sum += per_thread(rcu_stress_count, t)[i]; + } + printf(" %lld", sum); + } + printf("\n"); + exit(0); +} + +/* + * Mainprogram. + */ + +void usage(int argc, char *argv[]) +{ + fprintf(stderr, "Usage: %s [nreaders [ perf | stress ] ]\n", argv[0]); + exit(-1); +} + +int main(int argc, char *argv[]) +{ + int nreaders = 1; + int cpustride = 1; + + smp_init(); + //rcu_init(); + +#ifdef DEBUG_YIELD + yield_active |= YIELD_READ; + yield_active |= YIELD_WRITE; +#endif + + if (argc > 1) { + nreaders = strtoul(argv[1], NULL, 0); + if (argc == 2) + perftest(nreaders, cpustride); + if (argc > 3) + cpustride = strtoul(argv[3], NULL, 0); + if (strcmp(argv[2], "perf") == 0) + perftest(nreaders, cpustride); + else if (strcmp(argv[2], "rperf") == 0) + rperftest(nreaders, cpustride); + else if (strcmp(argv[2], "uperf") == 0) + uperftest(nreaders, cpustride); + else if (strcmp(argv[2], "stress") == 0) + stresstest(nreaders); + usage(argc, argv); + } + perftest(nreaders, cpustride); + return 0; +} diff --git a/tests/runall.sh b/tests/runall.sh new file mode 100755 index 0000000..e2b47dc --- /dev/null +++ b/tests/runall.sh @@ -0,0 +1,101 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +NUM_CPUS=8 + +#extra options, e.g. for setting affinity on even CPUs : +#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +#stride 1 +#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +#stride 2 +#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +#stride 4 +#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +#stride 8 +#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) + +#Vary update fraction +#x: vary update fraction from 0 to 0.0001 + #fix number of readers and reader C.S. length, vary delay between updates +#y: ops/s + +rm -f runall.log +rm -fr runall.detail.log + + +echo Executing batch RCU test + +DURATION=10 +BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 + 131072 262144" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f batch-rcu.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for BATCH_SIZE in ${BATCH_ARRAY}; do + echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log + ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log +done + +#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON TEST ABOVE ** +EXTRA_OPTS+="-b 32768" + +echo Executing update fraction test + +DURATION=10 +WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 + 65536 131072 262144 524288 1048576 2097152 4194304 8388608 + 16777216 33554432 67108864 134217728" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f update-fraction.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for WDELAY in ${WDELAY_ARRAY}; do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log +done + +#Test scalability : +# x: vary number of readers from 0 to num cpus +# y: ops/s +# 0 writer. + +echo Executing scalability test + +NR_WRITERS=0 +DURATION=10 + +rm -f scalability.log + +for NR_READERS in $(seq 1 ${NUM_CPUS}); do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log +done + + +# x: Vary reader C.S. length from 0 to 100 us +# y: ops/s +# 8 readers +# 0 writers + +echo Executing reader C.S. length test + +NR_READERS=${NUM_CPUS} +NR_WRITERS=0 +DURATION=10 +#in loops. +READERCSLEN_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152" + +rm -f readercslen.log + +for READERCSLEN in ${READERCSLEN_ARRAY}; do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log +done diff --git a/tests/runpaul-phase1.sh b/tests/runpaul-phase1.sh new file mode 100755 index 0000000..d2c8649 --- /dev/null +++ b/tests/runpaul-phase1.sh @@ -0,0 +1,44 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +NUM_CPUS=64 + +#extra options, e.g. for setting affinity on even CPUs : +EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +#stride 1 +#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +#stride 2 +#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +#stride 4 +#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +#stride 8 +#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) + +#Vary update fraction +#x: vary update fraction from 0 to 0.0001 + #fix number of readers and reader C.S. length, vary delay between updates +#y: ops/s + +rm -f runall.log +rm -fr runall.detail.log + + +echo Executing batch RCU test + +DURATION=10 +BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 + 131072 262144" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f batch-rcu.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for BATCH_SIZE in ${BATCH_ARRAY}; do + echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log + ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log +done diff --git a/tests/runpaul-phase2.sh b/tests/runpaul-phase2.sh new file mode 100755 index 0000000..0f6bd91 --- /dev/null +++ b/tests/runpaul-phase2.sh @@ -0,0 +1,47 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +NUM_CPUS=64 + +#extra options, e.g. for setting affinity on even CPUs : +EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +#stride 1 +#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +#stride 2 +#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +#stride 4 +#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +#stride 8 +#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) + +#Vary update fraction +#x: vary update fraction from 0 to 0.0001 + #fix number of readers and reader C.S. length, vary delay between updates +#y: ops/s + +rm -f runall.log +rm -fr runall.detail.log + +#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON PHASE 1 RESULT ** +EXTRA_OPTS+="-b 32768" + +echo Executing update fraction test + +DURATION=10 +WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 + 65536 131072 262144 524288 1048576 2097152 4194304 8388608 + 16777216 33554432 67108864 134217728" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f update-fraction.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for WDELAY in ${WDELAY_ARRAY}; do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log +done diff --git a/tests/runpaul-phase3.sh b/tests/runpaul-phase3.sh new file mode 100755 index 0000000..1956062 --- /dev/null +++ b/tests/runpaul-phase3.sh @@ -0,0 +1,50 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +NUM_CPUS=64 + +#extra options, e.g. for setting affinity on even CPUs : +EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +#stride 1 +#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +#stride 2 +#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +#stride 4 +#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +#stride 8 +#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) + +#Vary update fraction +#x: vary update fraction from 0 to 0.0001 + #fix number of readers and reader C.S. length, vary delay between updates +#y: ops/s + +rm -f runall.log +rm -fr runall.detail.log + +#setting gc each 32768. ** UPDATE FOR YOUR ARCHITECTURE BASED ON PHASE 1 RESULT ** +EXTRA_OPTS+="-b 32768" + +#Test scalability : +# x: vary number of readers from 0 to num cpus +# y: ops/s +# 0 writer. + +echo Executing scalability test + +NR_WRITERS=0 +DURATION=10 + +rm -f scalability.log + +for NR_READERS in $(seq 1 ${NUM_CPUS}); do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log +done + + diff --git a/tests/runpaul-phase4.sh b/tests/runpaul-phase4.sh new file mode 100755 index 0000000..ede402c --- /dev/null +++ b/tests/runpaul-phase4.sh @@ -0,0 +1,40 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +export NUM_CPUS=8 + +#extra options, e.g. for setting affinity on even CPUs : +#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +rm -f *.log + +#stride 1 +export EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +sh subphase4.sh $* +mkdir ppc64-8cores-stride1 +mv *.log ppc64-8cores-stride1/ + + +#stride 2 +export EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +sh subphase4.sh $* +mkdir ppc64-8cores-stride2 +mv *.log ppc64-8cores-stride2/ + + +#stride 4 +export EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +sh subphase4.sh $* +mkdir ppc64-8cores-stride4 +mv *.log ppc64-8cores-stride4/ + + +#stride 8 +export EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) +sh subphase4.sh $* +mkdir ppc64-8cores-stride8 +mv *.log ppc64-8cores-stride8/ diff --git a/tests/runpaul-phase5.sh b/tests/runpaul-phase5.sh new file mode 100644 index 0000000..bb4bfe7 --- /dev/null +++ b/tests/runpaul-phase5.sh @@ -0,0 +1,11 @@ +# test run after write-size update + +sh runpaul-phase1.sh +mkdir runpaul-phase1 +mv *.log runpaul-phase1/ + +sh runpaul-phase2.sh +mkdir runpaul-phase2 +mv *.log runpaul-phase2/ + +sh runpaul-phase4.sh diff --git a/tests/runpaul-phase6.sh b/tests/runpaul-phase6.sh new file mode 100644 index 0000000..5f65072 --- /dev/null +++ b/tests/runpaul-phase6.sh @@ -0,0 +1,7 @@ +sh runpaul-phase1.sh +mkdir runpaul-phase1 +mv *.log runpaul-phase1/ + +sh runpaul-phase2.sh +mkdir runpaul-phase2 +mv *.log runpaul-phase2/ diff --git a/tests/runtests-batch.sh b/tests/runtests-batch.sh new file mode 100755 index 0000000..67e019b --- /dev/null +++ b/tests/runtests-batch.sh @@ -0,0 +1,8 @@ +#!/bin/sh + +#for a in test_urcu_gc test_urcu_gc_mb test_qsbr_gc; do +for a in test_urcu_gc; do + echo "./${a} $*" | tee -a runall.detail.log + /usr/bin/time --append --output runall.detail.log ./${a} $* +done + diff --git a/tests/runtests.sh b/tests/runtests.sh new file mode 100755 index 0000000..981aef1 --- /dev/null +++ b/tests/runtests.sh @@ -0,0 +1,10 @@ +#!/bin/sh + +for a in test_urcu_gc test_urcu_gc_mb test_urcu test_urcu_mb \ + test_urcu_lgc test_qsbr_lgc test_urcu_lgc_mb \ + test_qsbr test_qsbr_gc test_rwlock test_perthreadlock \ + test_mutex; do + echo "./${a} $*" | tee -a runall.detail.log + /usr/bin/time --append --output runall.detail.log ./${a} $* +done + diff --git a/tests/subphase4.sh b/tests/subphase4.sh new file mode 100755 index 0000000..b69ffbf --- /dev/null +++ b/tests/subphase4.sh @@ -0,0 +1,101 @@ +#!/bin/sh + +#run all tests + +#set to number of active CPUS +#NUM_CPUS=8 + +#extra options, e.g. for setting affinity on even CPUs : +#EXTRA_OPTS=$(for a in $(seq 0 2 127); do echo -n "-a ${a} "; done) + +#ppc64 striding, use with NUM_CPUS=8 + +#stride 1 +#EXTRA_OPTS=$(for a in $(seq 0 2 15); do echo -n "-a ${a} "; done) +#stride 2 +#EXTRA_OPTS=$(for a in $(seq 0 4 31); do echo -n "-a ${a} "; done) +#stride 4 +#EXTRA_OPTS=$(for a in $(seq 0 8 63); do echo -n "-a ${a} "; done) +#stride 8 +#EXTRA_OPTS=$(for a in $(seq 0 16 127); do echo -n "-a ${a} "; done) + +#Vary update fraction +#x: vary update fraction from 0 to 0.0001 + #fix number of readers and reader C.S. length, vary delay between updates +#y: ops/s + +rm -f runall.log +rm -fr runall.detail.log + + +echo Executing batch RCU test + +DURATION=10 +BATCH_ARRAY="1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 + 131072 262144" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f batch-rcu.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for BATCH_SIZE in ${BATCH_ARRAY}; do + echo "./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log" >> runall.log + ./runtests-batch.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d 0 -b ${BATCH_SIZE} ${EXTRA_OPTS} | tee -a batch-rcu.log +done + +#setting gc each 4096. ** UPDATE FOR YOUR ARCHITECTURE BASED ON TEST ABOVE ** +EXTRA_OPTS+="-b 32768" + +echo Executing update fraction test + +DURATION=10 +WDELAY_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 + 65536 131072 262144 524288 1048576 2097152 4194304 8388608 + 16777216 33554432 67108864 134217728" +NR_WRITERS=$((${NUM_CPUS} / 2)) + +rm -f update-fraction.log + +NR_READERS=$((${NUM_CPUS} - ${NR_WRITERS})) +for WDELAY in ${WDELAY_ARRAY}; do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} -d ${WDELAY} ${EXTRA_OPTS} | tee -a update-fraction.log +done + +#Test scalability : +# x: vary number of readers from 0 to num cpus +# y: ops/s +# 0 writer. + +echo Executing scalability test + +NR_WRITERS=0 +DURATION=10 + +rm -f scalability.log + +for NR_READERS in $(seq 1 ${NUM_CPUS}); do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS}| tee -a scalability.log +done + + +# x: Vary reader C.S. length from 0 to 100 us +# y: ops/s +# 8 readers +# 0 writers + +echo Executing reader C.S. length test + +NR_READERS=${NUM_CPUS} +NR_WRITERS=0 +DURATION=10 +#in loops. +READERCSLEN_ARRAY="0 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152" + +rm -f readercslen.log + +for READERCSLEN in ${READERCSLEN_ARRAY}; do + echo "./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log" >> runall.log + ./runtests.sh ${NR_READERS} ${NR_WRITERS} ${DURATION} ${EXTRA_OPTS} -c ${READERCSLEN} | tee -a readercslen.log +done diff --git a/tests/test_looplen.c b/tests/test_looplen.c new file mode 100644 index 0000000..5006951 --- /dev/null +++ b/tests/test_looplen.c @@ -0,0 +1,89 @@ +/* + * test_looplen.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +#define LOOPS 1048576 +#define TESTS 10 + +int main(int argc, char **argv) +{ + unsigned long i; + cycles_t time1, time2; + cycles_t time_tot = 0; + double cpl; + + for (i = 0; i < TESTS; i++) { + time1 = get_cycles(); + loop_sleep(LOOPS); + time2 = get_cycles(); + time_tot += time2 - time1; + } + cpl = ((double)time_tot) / (double)TESTS / (double)LOOPS; + + printf("CALIBRATION : %g cycles per loop\n", cpl); + printf("time_tot = %llu, LOOPS = %d, TESTS = %d\n", + time_tot, LOOPS, TESTS); + + return 0; +} diff --git a/tests/test_mutex.c b/tests/test_mutex.c new file mode 100644 index 0000000..e94819a --- /dev/null +++ b/tests/test_mutex.c @@ -0,0 +1,392 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +struct test_array { + int a; +}; + +static pthread_mutex_t lock; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static volatile struct test_array test_array = { 8 }; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +void *thr_reader(void *data) +{ + unsigned long tidx = (unsigned long)data; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + + for (;;) { + pthread_mutex_lock(&lock); + assert(test_array.a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + pthread_mutex_unlock(&lock); + nr_reads++; + if (unlikely(!test_duration_read())) + break; + } + + tot_nr_reads[tidx] = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *data) +{ + unsigned long wtidx = (unsigned long)data; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + pthread_mutex_lock(&lock); + test_array.a = 0; + test_array.a = 8; + pthread_mutex_unlock(&lock); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + tot_nr_writes[wtidx] = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader, *count_writer; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + smp_mb(); + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + count_writer = malloc(sizeof(*count_writer) * nr_writers); + tot_nr_reads = malloc(sizeof(*tot_nr_reads) * nr_readers); + tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += tot_nr_reads[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += tot_nr_writes[i]; + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes); + + free(tid_reader); + free(tid_writer); + free(count_reader); + free(count_writer); + free(tot_nr_reads); + free(tot_nr_writes); + return 0; +} diff --git a/tests/test_perthreadlock.c b/tests/test_perthreadlock.c new file mode 100644 index 0000000..7402f01 --- /dev/null +++ b/tests/test_perthreadlock.c @@ -0,0 +1,403 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +struct test_array { + int a; +}; + +struct per_thread_lock { + pthread_mutex_t lock; +} __attribute__((aligned(CACHE_LINE_SIZE))); /* cache-line aligned */ + +static struct per_thread_lock *per_thread_lock; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static volatile struct test_array test_array = { 8 }; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +void *thr_reader(void *data) +{ + unsigned long tidx = (unsigned long)data; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + + for (;;) { + pthread_mutex_lock(&per_thread_lock[tidx].lock); + assert(test_array.a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + pthread_mutex_unlock(&per_thread_lock[tidx].lock); + nr_reads++; + if (unlikely(!test_duration_read())) + break; + } + + tot_nr_reads[tidx] = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *data) +{ + unsigned long wtidx = (unsigned long)data; + long tidx; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + for (tidx = 0; tidx < nr_readers; tidx++) { + pthread_mutex_lock(&per_thread_lock[tidx].lock); + } + test_array.a = 0; + test_array.a = 8; + for (tidx = (long)nr_readers - 1; tidx >= 0; tidx--) { + pthread_mutex_unlock(&per_thread_lock[tidx].lock); + } + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + tot_nr_writes[wtidx] = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader, *count_writer; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + smp_mb(); + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + count_writer = malloc(sizeof(*count_writer) * nr_writers); + tot_nr_reads = malloc(sizeof(*tot_nr_reads) * nr_readers); + tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); + per_thread_lock = malloc(sizeof(*per_thread_lock) * nr_readers); + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += tot_nr_reads[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += tot_nr_writes[i]; + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes); + + free(tid_reader); + free(tid_writer); + free(count_reader); + free(count_writer); + free(tot_nr_reads); + free(tot_nr_writes); + free(per_thread_lock); + return 0; +} diff --git a/tests/test_perthreadlock_timing.c b/tests/test_perthreadlock_timing.c new file mode 100644 index 0000000..d5bd912 --- /dev/null +++ b/tests/test_perthreadlock_timing.c @@ -0,0 +1,214 @@ +/* + * test_perthreadloc_timing.c + * + * Per thread locks - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#include "../urcu.h" + +struct test_array { + int a; +}; + +static struct test_array test_array = { 8 }; + +struct per_thread_lock { + pthread_mutex_t lock; +} __attribute__((aligned(CACHE_LINE_SIZE))); /* cache-line aligned */ + +static struct per_thread_lock *per_thread_lock; + +#define OUTER_READ_LOOP 200U +#define INNER_READ_LOOP 100000U +#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) + +#define OUTER_WRITE_LOOP 10U +#define INNER_WRITE_LOOP 200U +#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) + +static int num_read; +static int num_write; + +#define NR_READ num_read +#define NR_WRITE num_write + +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; + +void *thr_reader(void *arg) +{ + int i, j; + cycles_t time1, time2; + long tidx = (long)arg; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + sleep(2); + + time1 = get_cycles(); + for (i = 0; i < OUTER_READ_LOOP; i++) { + for (j = 0; j < INNER_READ_LOOP; j++) { + pthread_mutex_lock(&per_thread_lock[tidx].lock); + assert(test_array.a == 8); + pthread_mutex_unlock(&per_thread_lock[tidx].lock); + } + } + time2 = get_cycles(); + + reader_time[tidx] = time2 - time1; + + sleep(2); + printf("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *arg) +{ + int i, j; + long tidx; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + sleep(2); + + for (i = 0; i < OUTER_WRITE_LOOP; i++) { + for (j = 0; j < INNER_WRITE_LOOP; j++) { + time1 = get_cycles(); + for (tidx = 0; tidx < NR_READ; tidx++) { + pthread_mutex_lock(&per_thread_lock[tidx].lock); + } + test_array.a = 8; + for (tidx = NR_READ - 1; tidx >= 0; tidx--) { + pthread_mutex_unlock(&per_thread_lock[tidx].lock); + } + time2 = get_cycles(); + writer_time[(unsigned long)arg] += time2 - time1; + usleep(1); + } + } + + printf("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + return ((void*)2); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + int i; + cycles_t tot_rtime = 0; + cycles_t tot_wtime = 0; + + if (argc < 2) { + printf("Usage : %s nr_readers nr_writers\n", argv[0]); + exit(-1); + } + num_read = atoi(argv[1]); + num_write = atoi(argv[2]); + + reader_time = malloc(sizeof(*reader_time) * num_read); + writer_time = malloc(sizeof(*writer_time) * num_write); + tid_reader = malloc(sizeof(*tid_reader) * num_read); + tid_writer = malloc(sizeof(*tid_writer) * num_write); + + printf("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + per_thread_lock = malloc(sizeof(struct per_thread_lock) * NR_READ); + + for (i = 0; i < NR_READ; i++) { + pthread_mutex_init(&per_thread_lock[i].lock, NULL); + } + for (i = 0; i < NR_READ; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + sleep(10); + + for (i = 0; i < NR_READ; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_rtime += reader_time[i]; + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_wtime += writer_time[i]; + } + printf("Time per read : %g cycles\n", + (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); + printf("Time per write : %g cycles\n", + (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); + free(per_thread_lock); + + free(reader_time); + free(writer_time); + free(tid_reader); + free(tid_writer); + + return 0; +} diff --git a/tests/test_qsbr.c b/tests/test_qsbr.c new file mode 100644 index 0000000..89a9423 --- /dev/null +++ b/tests/test_qsbr.c @@ -0,0 +1,430 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#define _LGPL_SOURCE +#include "../urcu-qsbr.h" + +struct test_array { + int a; +}; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static struct test_array *test_rcu_pointer; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +/* + * malloc/free are reusing memory areas too quickly, which does not let us + * test races appropriately. Use a large circular array for allocations. + * ARRAY_SIZE is larger than nr_writers, which insures we never run over our tail. + */ +#define ARRAY_SIZE (1048576 * nr_writers) +#define ARRAY_POISON 0xDEADBEEF +static int array_index; +static struct test_array *test_array; + +static struct test_array *test_array_alloc(void) +{ + struct test_array *ret; + int index; + + rcu_copy_mutex_lock(); + index = array_index % ARRAY_SIZE; + assert(test_array[index].a == ARRAY_POISON || + test_array[index].a == 0); + ret = &test_array[index]; + array_index++; + if (array_index == ARRAY_SIZE) + array_index = 0; + rcu_copy_mutex_unlock(); + return ret; +} + +static void test_array_free(struct test_array *ptr) +{ + if (!ptr) + return; + rcu_copy_mutex_lock(); + ptr->a = ARRAY_POISON; + rcu_copy_mutex_unlock(); +} + +void *thr_reader(void *_count) +{ + unsigned long long *count = _count; + struct test_array *local_ptr; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + rcu_register_thread(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + _rcu_read_lock(); + local_ptr = _rcu_dereference(test_rcu_pointer); + debug_yield_read(); + if (local_ptr) + assert(local_ptr->a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + _rcu_read_unlock(); + nr_reads++; + /* QS each 1024 reads */ + if (unlikely((nr_reads & ((1 << 10) - 1)) == 0)) + _rcu_quiescent_state(); + if (unlikely(!test_duration_read())) + break; + } + + rcu_unregister_thread(); + + *count = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *_count) +{ + unsigned long long *count = _count; + struct test_array *new, *old; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + new = test_array_alloc(); + new->a = 8; + old = _rcu_publish_content(&test_rcu_pointer, new); + /* can be done after unlock */ + if (old) + old->a = 0; + test_array_free(old); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + *count = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader, *count_writer; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + test_array = malloc(sizeof(*test_array) * ARRAY_SIZE); + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + count_writer = malloc(sizeof(*count_writer) * nr_writers); + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + &count_reader[i]); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + &count_writer[i]); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += count_reader[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += count_writer[i]; + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes); + test_array_free(test_rcu_pointer); + free(test_array); + free(tid_reader); + free(tid_writer); + free(count_reader); + free(count_writer); + return 0; +} diff --git a/tests/test_qsbr_gc.c b/tests/test_qsbr_gc.c new file mode 100644 index 0000000..004672d --- /dev/null +++ b/tests/test_qsbr_gc.c @@ -0,0 +1,461 @@ +/* + * test_urcu_gc.c + * + * Userspace RCU library - test program (with baatch reclamation) + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#define _LGPL_SOURCE +#include "../urcu-qsbr.h" + +struct test_array { + int a; +}; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static struct test_array *test_rcu_pointer; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; +static unsigned int reclaim_batch = 1; + +struct reclaim_queue { + void **queue; /* Beginning of queue */ + void **head; /* Insert position */ +}; + +static struct reclaim_queue *pending_reclaims; + + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; + + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +void *thr_reader(void *_count) +{ + unsigned long long *count = _count; + struct test_array *local_ptr; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + rcu_register_thread(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + _rcu_read_lock(); + local_ptr = _rcu_dereference(test_rcu_pointer); + debug_yield_read(); + if (local_ptr) + assert(local_ptr->a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + _rcu_read_unlock(); + nr_reads++; + /* QS each 1024 reads */ + if (unlikely((nr_reads & ((1 << 10) - 1)) == 0)) + _rcu_quiescent_state(); + if (unlikely(!test_duration_read())) + break; + } + + rcu_unregister_thread(); + + *count = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +static void rcu_gc_clear_queue(unsigned long wtidx) +{ + void **p; + + /* Wait for Q.S and empty queue */ + synchronize_rcu(); + + for (p = pending_reclaims[wtidx].queue; + p < pending_reclaims[wtidx].head; p++) { + /* poison */ + if (*p) + ((struct test_array *)*p)->a = 0; + free(*p); + } + pending_reclaims[wtidx].head = pending_reclaims[wtidx].queue; +} + +/* Using per-thread queue */ +static void rcu_gc_reclaim(unsigned long wtidx, void *old) +{ + /* Queue pointer */ + *pending_reclaims[wtidx].head = old; + pending_reclaims[wtidx].head++; + + if (likely(pending_reclaims[wtidx].head - pending_reclaims[wtidx].queue + < reclaim_batch)) + return; + + rcu_gc_clear_queue(wtidx); +} + +void *thr_writer(void *data) +{ + unsigned long wtidx = (unsigned long)data; +#ifdef TEST_LOCAL_GC + struct test_array *old = NULL; +#else + struct test_array *new, *old; +#endif + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { +#ifndef TEST_LOCAL_GC + new = malloc(sizeof(*new)); + new->a = 8; + old = _rcu_xchg_pointer(&test_rcu_pointer, new); +#endif + rcu_gc_reclaim(wtidx, old); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + tot_nr_writes[wtidx] = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'b': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + reclaim_batch = atol(argv[++i]); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); + pending_reclaims = malloc(sizeof(*pending_reclaims) * nr_writers); + if (reclaim_batch * sizeof(*pending_reclaims[i].queue) + < CACHE_LINE_SIZE) + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].queue = calloc(1, CACHE_LINE_SIZE); + else + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].queue = calloc(reclaim_batch, + sizeof(*pending_reclaims[i].queue)); + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].head = pending_reclaims[i].queue; + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + &count_reader[i]); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += count_reader[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += tot_nr_writes[i]; + rcu_gc_clear_queue(i); + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu " + "batch %u\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes, reclaim_batch); + free(tid_reader); + free(tid_writer); + free(count_reader); + free(tot_nr_writes); + for (i = 0; i < nr_writers; i++) + free(pending_reclaims[i].queue); + free(pending_reclaims); + + return 0; +} diff --git a/tests/test_qsbr_timing.c b/tests/test_qsbr_timing.c new file mode 100644 index 0000000..2a8963a --- /dev/null +++ b/tests/test_qsbr_timing.c @@ -0,0 +1,240 @@ +/* + * test_qsbr_timing.c + * + * Userspace QSBR - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#define _LGPL_SOURCE +#include "../urcu-qsbr.h" + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +struct test_array { + int a; +}; + +static struct test_array *test_rcu_pointer; + +#define OUTER_READ_LOOP 2000U +#define INNER_READ_LOOP 100000U +#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) + +#define OUTER_WRITE_LOOP 10U +#define INNER_WRITE_LOOP 200U +#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) + +static int num_read; +static int num_write; + +#define NR_READ num_read +#define NR_WRITE num_write + +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; + +void *thr_reader(void *arg) +{ + int i, j; + struct test_array *local_ptr; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + sleep(2); + + rcu_register_thread(); + + time1 = get_cycles(); + for (i = 0; i < OUTER_READ_LOOP; i++) { + for (j = 0; j < INNER_READ_LOOP; j++) { + _rcu_read_lock(); + local_ptr = _rcu_dereference(test_rcu_pointer); + if (local_ptr) { + assert(local_ptr->a == 8); + } + _rcu_read_unlock(); + } + _rcu_quiescent_state(); + } + time2 = get_cycles(); + + rcu_unregister_thread(); + + reader_time[(unsigned long)arg] = time2 - time1; + + sleep(2); + printf("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *arg) +{ + int i, j; + struct test_array *new, *old; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + sleep(2); + + for (i = 0; i < OUTER_WRITE_LOOP; i++) { + for (j = 0; j < INNER_WRITE_LOOP; j++) { + time1 = get_cycles(); + new = malloc(sizeof(struct test_array)); + rcu_copy_mutex_lock(); + old = test_rcu_pointer; + if (old) { + assert(old->a == 8); + } + new->a = 8; + old = _rcu_publish_content(&test_rcu_pointer, new); + rcu_copy_mutex_unlock(); + /* can be done after unlock */ + if (old) { + old->a = 0; + } + free(old); + time2 = get_cycles(); + writer_time[(unsigned long)arg] += time2 - time1; + usleep(1); + } + } + + printf("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + return ((void*)2); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + int i; + cycles_t tot_rtime = 0; + cycles_t tot_wtime = 0; + + if (argc < 2) { + printf("Usage : %s nr_readers nr_writers\n", argv[0]); + exit(-1); + } + num_read = atoi(argv[1]); + num_write = atoi(argv[2]); + + reader_time = malloc(sizeof(*reader_time) * num_read); + writer_time = malloc(sizeof(*writer_time) * num_write); + tid_reader = malloc(sizeof(*tid_reader) * num_read); + tid_writer = malloc(sizeof(*tid_writer) * num_write); + + printf("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + for (i = 0; i < NR_READ; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + sleep(10); + + for (i = 0; i < NR_READ; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_rtime += reader_time[i]; + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_wtime += writer_time[i]; + } + free(test_rcu_pointer); + printf("Time per read : %g cycles\n", + (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); + printf("Time per write : %g cycles\n", + (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); + + free(reader_time); + free(writer_time); + free(tid_reader); + free(tid_writer); + + return 0; +} diff --git a/tests/test_rwlock.c b/tests/test_rwlock.c new file mode 100644 index 0000000..c7edd32 --- /dev/null +++ b/tests/test_rwlock.c @@ -0,0 +1,383 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +struct test_array { + int a; +}; + +pthread_rwlock_t lock = PTHREAD_RWLOCK_INITIALIZER; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static volatile struct test_array test_array = { 8 }; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +void *thr_reader(void *_count) +{ + unsigned long long *count = _count; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + + for (;;) { + pthread_rwlock_rdlock(&lock); + assert(test_array.a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + pthread_rwlock_unlock(&lock); + nr_reads++; + if (unlikely(!test_duration_read())) + break; + } + + *count = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *_count) +{ + unsigned long long *count = _count; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + pthread_rwlock_wrlock(&lock); + test_array.a = 0; + test_array.a = 8; + pthread_rwlock_unlock(&lock); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + *count = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader, *count_writer; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + smp_mb(); + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + count_writer = malloc(sizeof(*count_writer) * nr_writers); + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + &count_reader[i]); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + &count_writer[i]); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += count_reader[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += count_writer[i]; + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes); + + free(tid_reader); + free(tid_writer); + free(count_reader); + free(count_writer); + return 0; +} diff --git a/tests/test_rwlock_timing.c b/tests/test_rwlock_timing.c new file mode 100644 index 0000000..5bc93d3 --- /dev/null +++ b/tests/test_rwlock_timing.c @@ -0,0 +1,197 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#include "../urcu.h" + +struct test_array { + int a; +}; + +pthread_rwlock_t lock = PTHREAD_RWLOCK_INITIALIZER; + +static struct test_array test_array = { 8 }; + +#define OUTER_READ_LOOP 200U +#define INNER_READ_LOOP 100000U +#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) + +#define OUTER_WRITE_LOOP 10U +#define INNER_WRITE_LOOP 200U +#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) + +static int num_read; +static int num_write; + +#define NR_READ num_read +#define NR_WRITE num_write + +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; + +void *thr_reader(void *arg) +{ + int i, j; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + sleep(2); + + time1 = get_cycles(); + for (i = 0; i < OUTER_READ_LOOP; i++) { + for (j = 0; j < INNER_READ_LOOP; j++) { + pthread_rwlock_rdlock(&lock); + assert(test_array.a == 8); + pthread_rwlock_unlock(&lock); + } + } + time2 = get_cycles(); + + reader_time[(unsigned long)arg] = time2 - time1; + + sleep(2); + printf("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *arg) +{ + int i, j; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + sleep(2); + + for (i = 0; i < OUTER_WRITE_LOOP; i++) { + for (j = 0; j < INNER_WRITE_LOOP; j++) { + time1 = get_cycles(); + pthread_rwlock_wrlock(&lock); + test_array.a = 8; + pthread_rwlock_unlock(&lock); + time2 = get_cycles(); + writer_time[(unsigned long)arg] += time2 - time1; + usleep(1); + } + } + + printf("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + return ((void*)2); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + int i; + cycles_t tot_rtime = 0; + cycles_t tot_wtime = 0; + + if (argc < 2) { + printf("Usage : %s nr_readers nr_writers\n", argv[0]); + exit(-1); + } + num_read = atoi(argv[1]); + num_write = atoi(argv[2]); + + reader_time = malloc(sizeof(*reader_time) * num_read); + writer_time = malloc(sizeof(*writer_time) * num_write); + tid_reader = malloc(sizeof(*tid_reader) * num_read); + tid_writer = malloc(sizeof(*tid_writer) * num_write); + + printf("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + for (i = 0; i < NR_READ; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + sleep(10); + + for (i = 0; i < NR_READ; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_rtime += reader_time[i]; + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_wtime += writer_time[i]; + } + printf("Time per read : %g cycles\n", + (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); + printf("Time per write : %g cycles\n", + (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); + + free(reader_time); + free(writer_time); + free(tid_reader); + free(tid_writer); + + return 0; +} diff --git a/tests/test_urcu.c b/tests/test_urcu.c new file mode 100644 index 0000000..18683bf --- /dev/null +++ b/tests/test_urcu.c @@ -0,0 +1,430 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +struct test_array { + int a; +}; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static struct test_array *test_rcu_pointer; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +/* + * malloc/free are reusing memory areas too quickly, which does not let us + * test races appropriately. Use a large circular array for allocations. + * ARRAY_SIZE is larger than nr_writers, which insures we never run over our tail. + */ +#define ARRAY_SIZE (1048576 * nr_writers) +#define ARRAY_POISON 0xDEADBEEF +static int array_index; +static struct test_array *test_array; + +static struct test_array *test_array_alloc(void) +{ + struct test_array *ret; + int index; + + rcu_copy_mutex_lock(); + index = array_index % ARRAY_SIZE; + assert(test_array[index].a == ARRAY_POISON || + test_array[index].a == 0); + ret = &test_array[index]; + array_index++; + if (array_index == ARRAY_SIZE) + array_index = 0; + rcu_copy_mutex_unlock(); + return ret; +} + +static void test_array_free(struct test_array *ptr) +{ + if (!ptr) + return; + rcu_copy_mutex_lock(); + ptr->a = ARRAY_POISON; + rcu_copy_mutex_unlock(); +} + +void *thr_reader(void *_count) +{ + unsigned long long *count = _count; + struct test_array *local_ptr; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + rcu_register_thread(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + rcu_read_lock(); + local_ptr = rcu_dereference(test_rcu_pointer); + debug_yield_read(); + if (local_ptr) + assert(local_ptr->a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + rcu_read_unlock(); + nr_reads++; + if (unlikely(!test_duration_read())) + break; + } + + rcu_unregister_thread(); + + *count = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *_count) +{ + unsigned long long *count = _count; + struct test_array *new, *old; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + new = test_array_alloc(); + new->a = 8; + old = rcu_publish_content(&test_rcu_pointer, new); + if (old) + old->a = 0; + test_array_free(old); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + *count = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader, *count_writer; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + test_array = malloc(sizeof(*test_array) * ARRAY_SIZE); + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + count_writer = malloc(sizeof(*count_writer) * nr_writers); + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + &count_reader[i]); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + &count_writer[i]); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += count_reader[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += count_writer[i]; + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes); + test_array_free(test_rcu_pointer); + free(test_array); + free(tid_reader); + free(tid_writer); + free(count_reader); + free(count_writer); + return 0; +} diff --git a/tests/test_urcu_gc.c b/tests/test_urcu_gc.c new file mode 100644 index 0000000..213c68b --- /dev/null +++ b/tests/test_urcu_gc.c @@ -0,0 +1,462 @@ +/* + * test_urcu_gc.c + * + * Userspace RCU library - test program (with baatch reclamation) + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +/* hardcoded number of CPUs */ +#define NR_CPUS 16384 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#ifndef DYNAMIC_LINK_TEST +#define _LGPL_SOURCE +#else +#define debug_yield_read() +#endif +#include "../urcu.h" + +struct test_array { + int a; +}; + +static volatile int test_go, test_stop; + +static unsigned long wdelay; + +static struct test_array *test_rcu_pointer; + +static unsigned int reclaim_batch = 1; + +struct reclaim_queue { + void **queue; /* Beginning of queue */ + void **head; /* Insert position */ +}; + +static struct reclaim_queue *pending_reclaims; + +static unsigned long duration; + +/* read-side C.S. duration, in loops */ +static unsigned long rduration; + +static inline void loop_sleep(unsigned long l) +{ + while(l-- != 0) + cpu_relax(); +} + +static int verbose_mode; + +#define printf_verbose(fmt, args...) \ + do { \ + if (verbose_mode) \ + printf(fmt, args); \ + } while (0) + +static unsigned int cpu_affinities[NR_CPUS]; +static unsigned int next_aff = 0; +static int use_affinity = 0; + +pthread_mutex_t affinity_mutex = PTHREAD_MUTEX_INITIALIZER; + +static void set_affinity(void) +{ + cpu_set_t mask; + int cpu; + int ret; + + if (!use_affinity) + return; + + ret = pthread_mutex_lock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } + cpu = cpu_affinities[next_aff++]; + ret = pthread_mutex_unlock(&affinity_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } + CPU_ZERO(&mask); + CPU_SET(cpu, &mask); + sched_setaffinity(0, sizeof(mask), &mask); +} + +/* + * returns 0 if test should end. + */ +static int test_duration_write(void) +{ + return !test_stop; +} + +static int test_duration_read(void) +{ + return !test_stop; +} + +static unsigned long long __thread nr_writes; +static unsigned long long __thread nr_reads; + +static +unsigned long long __attribute__((aligned(CACHE_LINE_SIZE))) *tot_nr_writes; + +static unsigned int nr_readers; +static unsigned int nr_writers; + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +void *thr_reader(void *_count) +{ + unsigned long long *count = _count; + struct test_array *local_ptr; + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + rcu_register_thread(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { + rcu_read_lock(); + local_ptr = rcu_dereference(test_rcu_pointer); + debug_yield_read(); + if (local_ptr) + assert(local_ptr->a == 8); + if (unlikely(rduration)) + loop_sleep(rduration); + rcu_read_unlock(); + nr_reads++; + if (unlikely(!test_duration_read())) + break; + } + + rcu_unregister_thread(); + + *count = nr_reads; + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +static void rcu_gc_clear_queue(unsigned long wtidx) +{ + void **p; + + /* Wait for Q.S and empty queue */ + synchronize_rcu(); + + for (p = pending_reclaims[wtidx].queue; + p < pending_reclaims[wtidx].head; p++) { + /* poison */ + if (*p) + ((struct test_array *)*p)->a = 0; + free(*p); + } + pending_reclaims[wtidx].head = pending_reclaims[wtidx].queue; +} + +/* Using per-thread queue */ +static void rcu_gc_reclaim(unsigned long wtidx, void *old) +{ + /* Queue pointer */ + *pending_reclaims[wtidx].head = old; + pending_reclaims[wtidx].head++; + + if (likely(pending_reclaims[wtidx].head - pending_reclaims[wtidx].queue + < reclaim_batch)) + return; + + rcu_gc_clear_queue(wtidx); +} + +void *thr_writer(void *data) +{ + unsigned long wtidx = (unsigned long)data; +#ifdef TEST_LOCAL_GC + struct test_array *old = NULL; +#else + struct test_array *new, *old; +#endif + + printf_verbose("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + + set_affinity(); + + while (!test_go) + { + } + smp_mb(); + + for (;;) { +#ifndef TEST_LOCAL_GC + new = malloc(sizeof(*new)); + new->a = 8; + old = rcu_xchg_pointer(&test_rcu_pointer, new); +#endif + rcu_gc_reclaim(wtidx, old); + nr_writes++; + if (unlikely(!test_duration_write())) + break; + if (unlikely(wdelay)) + loop_sleep(wdelay); + } + + printf_verbose("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + tot_nr_writes[wtidx] = nr_writes; + return ((void*)2); +} + +void show_usage(int argc, char **argv) +{ + printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]); +#ifdef DEBUG_YIELD + printf(" [-r] [-w] (yield reader and/or writer)"); +#endif + printf(" [-d delay] (writer period (us))"); + printf(" [-c duration] (reader C.S. duration (in loops))"); + printf(" [-v] (verbose output)"); + printf(" [-a cpu#] [-a cpu#]... (affinity)"); + printf("\n"); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + unsigned long long *count_reader; + unsigned long long tot_reads = 0, tot_writes = 0; + int i, a; + + if (argc < 4) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[1], "%u", &nr_readers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[2], "%u", &nr_writers); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + err = sscanf(argv[3], "%lu", &duration); + if (err != 1) { + show_usage(argc, argv); + return -1; + } + + for (i = 4; i < argc; i++) { + if (argv[i][0] != '-') + continue; + switch (argv[i][1]) { +#ifdef DEBUG_YIELD + case 'r': + yield_active |= YIELD_READ; + break; + case 'w': + yield_active |= YIELD_WRITE; + break; +#endif + case 'a': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + a = atoi(argv[++i]); + cpu_affinities[next_aff++] = a; + use_affinity = 1; + printf_verbose("Adding CPU %d affinity\n", a); + break; + case 'b': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + reclaim_batch = atol(argv[++i]); + break; + case 'c': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + rduration = atol(argv[++i]); + break; + case 'd': + if (argc < i + 2) { + show_usage(argc, argv); + return -1; + } + wdelay = atol(argv[++i]); + break; + case 'v': + verbose_mode = 1; + break; + } + } + + printf_verbose("running test for %lu seconds, %u readers, %u writers.\n", + duration, nr_readers, nr_writers); + printf_verbose("Writer delay : %lu loops.\n", wdelay); + printf_verbose("Reader duration : %lu loops.\n", rduration); + printf_verbose("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + tid_reader = malloc(sizeof(*tid_reader) * nr_readers); + tid_writer = malloc(sizeof(*tid_writer) * nr_writers); + count_reader = malloc(sizeof(*count_reader) * nr_readers); + tot_nr_writes = malloc(sizeof(*tot_nr_writes) * nr_writers); + pending_reclaims = malloc(sizeof(*pending_reclaims) * nr_writers); + if (reclaim_batch * sizeof(*pending_reclaims[i].queue) + < CACHE_LINE_SIZE) + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].queue = calloc(1, CACHE_LINE_SIZE); + else + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].queue = calloc(reclaim_batch, + sizeof(*pending_reclaims[i].queue)); + for (i = 0; i < nr_writers; i++) + pending_reclaims[i].head = pending_reclaims[i].queue; + + next_aff = 0; + + for (i = 0; i < nr_readers; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + &count_reader[i]); + if (err != 0) + exit(1); + } + for (i = 0; i < nr_writers; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + smp_mb(); + + test_go = 1; + + sleep(duration); + + test_stop = 1; + + for (i = 0; i < nr_readers; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_reads += count_reader[i]; + } + for (i = 0; i < nr_writers; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_writes += tot_nr_writes[i]; + rcu_gc_clear_queue(i); + } + + printf_verbose("total number of reads : %llu, writes %llu\n", tot_reads, + tot_writes); + printf("SUMMARY %-25s testdur %4lu nr_readers %3u rdur %6lu " + "nr_writers %3u " + "wdelay %6lu nr_reads %12llu nr_writes %12llu nr_ops %12llu " + "batch %u\n", + argv[0], duration, nr_readers, rduration, + nr_writers, wdelay, tot_reads, tot_writes, + tot_reads + tot_writes, reclaim_batch); + free(tid_reader); + free(tid_writer); + free(count_reader); + free(tot_nr_writes); + for (i = 0; i < nr_writers; i++) + free(pending_reclaims[i].queue); + free(pending_reclaims); + + return 0; +} diff --git a/tests/test_urcu_timing.c b/tests/test_urcu_timing.c new file mode 100644 index 0000000..a3ca783 --- /dev/null +++ b/tests/test_urcu_timing.c @@ -0,0 +1,239 @@ +/* + * test_urcu.c + * + * Userspace RCU library - test program + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "../arch.h" + +/* Make this big enough to include the POWER5+ L3 cacheline size of 256B */ +#define CACHE_LINE_SIZE 4096 + +#if defined(_syscall0) +_syscall0(pid_t, gettid) +#elif defined(__NR_gettid) +static inline pid_t gettid(void) +{ + return syscall(__NR_gettid); +} +#else +#warning "use pid as tid" +static inline pid_t gettid(void) +{ + return getpid(); +} +#endif + +#define _LGPL_SOURCE +#include "../urcu.h" + +pthread_mutex_t rcu_copy_mutex = PTHREAD_MUTEX_INITIALIZER; + +void rcu_copy_mutex_lock(void) +{ + int ret; + ret = pthread_mutex_lock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex lock"); + exit(-1); + } +} + +void rcu_copy_mutex_unlock(void) +{ + int ret; + + ret = pthread_mutex_unlock(&rcu_copy_mutex); + if (ret) { + perror("Error in pthread mutex unlock"); + exit(-1); + } +} + +struct test_array { + int a; +}; + +static struct test_array *test_rcu_pointer; + +#define OUTER_READ_LOOP 2000U +#define INNER_READ_LOOP 100000U +#define READ_LOOP ((unsigned long long)OUTER_READ_LOOP * INNER_READ_LOOP) + +#define OUTER_WRITE_LOOP 10U +#define INNER_WRITE_LOOP 200U +#define WRITE_LOOP ((unsigned long long)OUTER_WRITE_LOOP * INNER_WRITE_LOOP) + +static int num_read; +static int num_write; + +#define NR_READ num_read +#define NR_WRITE num_write + +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *reader_time; +static cycles_t __attribute__((aligned(CACHE_LINE_SIZE))) *writer_time; + +void *thr_reader(void *arg) +{ + int i, j; + struct test_array *local_ptr; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + sleep(2); + + rcu_register_thread(); + + time1 = get_cycles(); + for (i = 0; i < OUTER_READ_LOOP; i++) { + for (j = 0; j < INNER_READ_LOOP; j++) { + rcu_read_lock(); + local_ptr = rcu_dereference(test_rcu_pointer); + if (local_ptr) { + assert(local_ptr->a == 8); + } + rcu_read_unlock(); + } + } + time2 = get_cycles(); + + rcu_unregister_thread(); + + reader_time[(unsigned long)arg] = time2 - time1; + + sleep(2); + printf("thread_end %s, thread id : %lx, tid %lu\n", + "reader", pthread_self(), (unsigned long)gettid()); + return ((void*)1); + +} + +void *thr_writer(void *arg) +{ + int i, j; + struct test_array *new, *old; + cycles_t time1, time2; + + printf("thread_begin %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + sleep(2); + + for (i = 0; i < OUTER_WRITE_LOOP; i++) { + for (j = 0; j < INNER_WRITE_LOOP; j++) { + time1 = get_cycles(); + new = malloc(sizeof(struct test_array)); + rcu_copy_mutex_lock(); + old = test_rcu_pointer; + if (old) { + assert(old->a == 8); + } + new->a = 8; + old = rcu_publish_content(&test_rcu_pointer, new); + rcu_copy_mutex_unlock(); + /* can be done after unlock */ + if (old) { + old->a = 0; + } + free(old); + time2 = get_cycles(); + writer_time[(unsigned long)arg] += time2 - time1; + usleep(1); + } + } + + printf("thread_end %s, thread id : %lx, tid %lu\n", + "writer", pthread_self(), (unsigned long)gettid()); + return ((void*)2); +} + +int main(int argc, char **argv) +{ + int err; + pthread_t *tid_reader, *tid_writer; + void *tret; + int i; + cycles_t tot_rtime = 0; + cycles_t tot_wtime = 0; + + if (argc < 2) { + printf("Usage : %s nr_readers nr_writers\n", argv[0]); + exit(-1); + } + num_read = atoi(argv[1]); + num_write = atoi(argv[2]); + + reader_time = malloc(sizeof(*reader_time) * num_read); + writer_time = malloc(sizeof(*writer_time) * num_write); + tid_reader = malloc(sizeof(*tid_reader) * num_read); + tid_writer = malloc(sizeof(*tid_writer) * num_write); + + printf("thread %-6s, thread id : %lx, tid %lu\n", + "main", pthread_self(), (unsigned long)gettid()); + + for (i = 0; i < NR_READ; i++) { + err = pthread_create(&tid_reader[i], NULL, thr_reader, + (void *)(long)i); + if (err != 0) + exit(1); + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_create(&tid_writer[i], NULL, thr_writer, + (void *)(long)i); + if (err != 0) + exit(1); + } + + sleep(10); + + for (i = 0; i < NR_READ; i++) { + err = pthread_join(tid_reader[i], &tret); + if (err != 0) + exit(1); + tot_rtime += reader_time[i]; + } + for (i = 0; i < NR_WRITE; i++) { + err = pthread_join(tid_writer[i], &tret); + if (err != 0) + exit(1); + tot_wtime += writer_time[i]; + } + free(test_rcu_pointer); + printf("Time per read : %g cycles\n", + (double)tot_rtime / ((double)NR_READ * (double)READ_LOOP)); + printf("Time per write : %g cycles\n", + (double)tot_wtime / ((double)NR_WRITE * (double)WRITE_LOOP)); + + free(reader_time); + free(writer_time); + free(tid_reader); + free(tid_writer); + + return 0; +} diff --git a/tests/urcu-asm.c b/tests/urcu-asm.c new file mode 100644 index 0000000..66be709 --- /dev/null +++ b/tests/urcu-asm.c @@ -0,0 +1,37 @@ +/* + * urcu-asm.c + * + * Userspace RCU library - assembly dump of primitives + * + * Copyright February 2009 - Mathieu Desnoyers + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * 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. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include "../urcu.h" + +void show_read_lock(void) +{ + asm volatile ("/* start */"); + rcu_read_lock(); + asm volatile ("/* end */"); +} + +void show_read_unlock(void) +{ + asm volatile ("/* start */"); + rcu_read_unlock(); + asm volatile ("/* end */"); +} diff --git a/tests/urcutorture.c b/tests/urcutorture.c new file mode 100644 index 0000000..70abc68 --- /dev/null +++ b/tests/urcutorture.c @@ -0,0 +1,9 @@ +#include +#include +#include +#include +#include +#include "api.h" +#define _LGPL_SOURCE +#include "../urcu.h" +#include "rcutorture.h" diff --git a/urcu-asm.c b/urcu-asm.c deleted file mode 100644 index 0e833f4..0000000 --- a/urcu-asm.c +++ /dev/null @@ -1,37 +0,0 @@ -/* - * urcu-asm.c - * - * Userspace RCU library - assembly dump of primitives - * - * Copyright February 2009 - Mathieu Desnoyers - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * 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. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - */ - -#include "urcu.h" - -void show_read_lock(void) -{ - asm volatile ("/* start */"); - rcu_read_lock(); - asm volatile ("/* end */"); -} - -void show_read_unlock(void) -{ - asm volatile ("/* start */"); - rcu_read_unlock(); - asm volatile ("/* end */"); -} diff --git a/urcutorture.c b/urcutorture.c deleted file mode 100644 index 75256f9..0000000 --- a/urcutorture.c +++ /dev/null @@ -1,9 +0,0 @@ -#include -#include -#include -#include -#include -#include "api.h" -#define _LGPL_SOURCE -#include "urcu.h" -#include "rcutorture.h"