diff options
author | Stephen Rothwell <sfr@canb.auug.org.au> | 2011-02-02 13:47:48 +1100 |
---|---|---|
committer | Stephen Rothwell <sfr@canb.auug.org.au> | 2011-02-02 13:47:48 +1100 |
commit | 47cf5f4e60b5d732dcbf0b9cfc99e772f7c136d6 (patch) | |
tree | 574089c30edd4a2ec888dc8cd20d73f657fdca19 | |
parent | a0eeb409249d2c6e943f739c14be60d93f980ea9 (diff) | |
parent | 4e70e1a1b8bf2425aaf3d25f6e65ce795d347723 (diff) |
Merge remote branch 'tip/auto-latest'
84 files changed, 2246 insertions, 1879 deletions
diff --git a/Documentation/spinlocks.txt b/Documentation/spinlocks.txt index 178c831b907d..2e3c64b1a6a5 100644 --- a/Documentation/spinlocks.txt +++ b/Documentation/spinlocks.txt @@ -86,7 +86,7 @@ to change the variables it has to get an exclusive write lock. The routines look the same as above: - rwlock_t xxx_lock = RW_LOCK_UNLOCKED; + rwlock_t xxx_lock = __RW_LOCK_UNLOCKED(xxx_lock); unsigned long flags; @@ -196,25 +196,3 @@ appropriate: For static initialization, use DEFINE_SPINLOCK() / DEFINE_RWLOCK() or __SPIN_LOCK_UNLOCKED() / __RW_LOCK_UNLOCKED() as appropriate. - -SPIN_LOCK_UNLOCKED and RW_LOCK_UNLOCKED are deprecated. These interfere -with lockdep state tracking. - -Most of the time, you can simply turn: - static spinlock_t xxx_lock = SPIN_LOCK_UNLOCKED; -into: - static DEFINE_SPINLOCK(xxx_lock); - -Static structure member variables go from: - - struct foo bar { - .lock = SPIN_LOCK_UNLOCKED; - }; - -to: - - struct foo bar { - .lock = __SPIN_LOCK_UNLOCKED(bar.lock); - }; - -Declaration of static rw_locks undergo a similar transformation. diff --git a/arch/alpha/include/asm/rwsem.h b/arch/alpha/include/asm/rwsem.h index 1570c0b54336..a83bbea62c67 100644 --- a/arch/alpha/include/asm/rwsem.h +++ b/arch/alpha/include/asm/rwsem.h @@ -13,44 +13,13 @@ #ifdef __KERNEL__ #include <linux/compiler.h> -#include <linux/list.h> -#include <linux/spinlock.h> -struct rwsem_waiter; - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -/* - * the semaphore definition - */ -struct rw_semaphore { - long count; #define RWSEM_UNLOCKED_VALUE 0x0000000000000000L #define RWSEM_ACTIVE_BIAS 0x0000000000000001L #define RWSEM_ACTIVE_MASK 0x00000000ffffffffL #define RWSEM_WAITING_BIAS (-0x0000000100000000L) #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) - spinlock_t wait_lock; - struct list_head wait_list; -}; - -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, SPIN_LOCK_UNLOCKED, \ - LIST_HEAD_INIT((name).wait_list) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -static inline void init_rwsem(struct rw_semaphore *sem) -{ - sem->count = RWSEM_UNLOCKED_VALUE; - spin_lock_init(&sem->wait_lock); - INIT_LIST_HEAD(&sem->wait_list); -} static inline void __down_read(struct rw_semaphore *sem) { @@ -250,10 +219,5 @@ static inline long rwsem_atomic_update(long val, struct rw_semaphore *sem) #endif } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* __KERNEL__ */ #endif /* _ALPHA_RWSEM_H */ diff --git a/arch/ia64/include/asm/rwsem.h b/arch/ia64/include/asm/rwsem.h index 215d5454c7d3..3027e7516d85 100644 --- a/arch/ia64/include/asm/rwsem.h +++ b/arch/ia64/include/asm/rwsem.h @@ -25,20 +25,8 @@ #error "Please don't include <asm/rwsem.h> directly, use <linux/rwsem.h> instead." #endif -#include <linux/list.h> -#include <linux/spinlock.h> - #include <asm/intrinsics.h> -/* - * the semaphore definition - */ -struct rw_semaphore { - signed long count; - spinlock_t wait_lock; - struct list_head wait_list; -}; - #define RWSEM_UNLOCKED_VALUE __IA64_UL_CONST(0x0000000000000000) #define RWSEM_ACTIVE_BIAS (1L) #define RWSEM_ACTIVE_MASK (0xffffffffL) @@ -46,26 +34,6 @@ struct rw_semaphore { #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED((name).wait_lock), \ - LIST_HEAD_INIT((name).wait_list) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -static inline void -init_rwsem (struct rw_semaphore *sem) -{ - sem->count = RWSEM_UNLOCKED_VALUE; - spin_lock_init(&sem->wait_lock); - INIT_LIST_HEAD(&sem->wait_list); -} - /* * lock for reading */ @@ -174,9 +142,4 @@ __downgrade_write (struct rw_semaphore *sem) #define rwsem_atomic_add(delta, sem) atomic64_add(delta, (atomic64_t *)(&(sem)->count)) #define rwsem_atomic_update(delta, sem) atomic64_add_return(delta, (atomic64_t *)(&(sem)->count)) -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* _ASM_IA64_RWSEM_H */ diff --git a/arch/mips/kernel/vpe.c b/arch/mips/kernel/vpe.c index 6a1fdfef8fde..ab52b7cf3b6b 100644 --- a/arch/mips/kernel/vpe.c +++ b/arch/mips/kernel/vpe.c @@ -148,9 +148,9 @@ struct { spinlock_t tc_list_lock; struct list_head tc_list; /* Thread contexts */ } vpecontrol = { - .vpe_list_lock = SPIN_LOCK_UNLOCKED, + .vpe_list_lock = __SPIN_LOCK_UNLOCKED(vpe_list_lock), .vpe_list = LIST_HEAD_INIT(vpecontrol.vpe_list), - .tc_list_lock = SPIN_LOCK_UNLOCKED, + .tc_list_lock = __SPIN_LOCK_UNLOCKED(tc_list_lock), .tc_list = LIST_HEAD_INIT(vpecontrol.tc_list) }; diff --git a/arch/powerpc/include/asm/rwsem.h b/arch/powerpc/include/asm/rwsem.h index 8447d89fbe72..bb1e2cdeb9bf 100644 --- a/arch/powerpc/include/asm/rwsem.h +++ b/arch/powerpc/include/asm/rwsem.h @@ -13,11 +13,6 @@ * by Paul Mackerras <paulus@samba.org>. */ -#include <linux/list.h> -#include <linux/spinlock.h> -#include <asm/atomic.h> -#include <asm/system.h> - /* * the semaphore definition */ @@ -33,47 +28,6 @@ #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) -struct rw_semaphore { - long count; - spinlock_t wait_lock; - struct list_head wait_list; -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -}; - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - -#define __RWSEM_INITIALIZER(name) \ -{ \ - RWSEM_UNLOCKED_VALUE, \ - __SPIN_LOCK_UNLOCKED((name).wait_lock), \ - LIST_HEAD_INIT((name).wait_list) \ - __RWSEM_DEP_MAP_INIT(name) \ -} - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ - do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ - } while (0) - /* * lock for reading */ @@ -174,10 +128,5 @@ static inline long rwsem_atomic_update(long delta, struct rw_semaphore *sem) return atomic_long_add_return(delta, (atomic_long_t *)&sem->count); } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return sem->count != 0; -} - #endif /* __KERNEL__ */ #endif /* _ASM_POWERPC_RWSEM_H */ diff --git a/arch/s390/include/asm/rwsem.h b/arch/s390/include/asm/rwsem.h index 423fdda2322d..d0eb4653cebd 100644 --- a/arch/s390/include/asm/rwsem.h +++ b/arch/s390/include/asm/rwsem.h @@ -43,29 +43,6 @@ #ifdef __KERNEL__ -#include <linux/list.h> -#include <linux/spinlock.h> - -struct rwsem_waiter; - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *); -extern struct rw_semaphore *rwsem_downgrade_write(struct rw_semaphore *); - -/* - * the semaphore definition - */ -struct rw_semaphore { - signed long count; - spinlock_t wait_lock; - struct list_head wait_list; -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -}; - #ifndef __s390x__ #define RWSEM_UNLOCKED_VALUE 0x00000000 #define RWSEM_ACTIVE_BIAS 0x00000001 @@ -81,41 +58,6 @@ struct rw_semaphore { #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) /* - * initialisation - */ - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED((name).wait.lock), \ - LIST_HEAD_INIT((name).wait_list) __RWSEM_DEP_MAP_INIT(name) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -static inline void init_rwsem(struct rw_semaphore *sem) -{ - sem->count = RWSEM_UNLOCKED_VALUE; - spin_lock_init(&sem->wait_lock); - INIT_LIST_HEAD(&sem->wait_list); -} - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ -do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ -} while (0) - - -/* * lock for reading */ static inline void __down_read(struct rw_semaphore *sem) @@ -377,10 +319,5 @@ static inline long rwsem_atomic_update(long delta, struct rw_semaphore *sem) return new; } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* __KERNEL__ */ #endif /* _S390_RWSEM_H */ diff --git a/arch/sh/include/asm/rwsem.h b/arch/sh/include/asm/rwsem.h index 06e2251a5e48..edab57265293 100644 --- a/arch/sh/include/asm/rwsem.h +++ b/arch/sh/include/asm/rwsem.h @@ -11,64 +11,13 @@ #endif #ifdef __KERNEL__ -#include <linux/list.h> -#include <linux/spinlock.h> -#include <asm/atomic.h> -#include <asm/system.h> -/* - * the semaphore definition - */ -struct rw_semaphore { - long count; #define RWSEM_UNLOCKED_VALUE 0x00000000 #define RWSEM_ACTIVE_BIAS 0x00000001 #define RWSEM_ACTIVE_MASK 0x0000ffff #define RWSEM_WAITING_BIAS (-0x00010000) #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) - spinlock_t wait_lock; - struct list_head wait_list; -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -}; - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED((name).wait_lock), \ - LIST_HEAD_INIT((name).wait_list) \ - __RWSEM_DEP_MAP_INIT(name) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ -do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ -} while (0) - -static inline void init_rwsem(struct rw_semaphore *sem) -{ - sem->count = RWSEM_UNLOCKED_VALUE; - spin_lock_init(&sem->wait_lock); - INIT_LIST_HEAD(&sem->wait_list); -} /* * lock for reading @@ -179,10 +128,5 @@ static inline int rwsem_atomic_update(int delta, struct rw_semaphore *sem) return atomic_add_return(delta, (atomic_t *)(&sem->count)); } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* __KERNEL__ */ #endif /* _ASM_SH_RWSEM_H */ diff --git a/arch/sparc/include/asm/rwsem.h b/arch/sparc/include/asm/rwsem.h index a2b4302869bc..069bf4d663a1 100644 --- a/arch/sparc/include/asm/rwsem.h +++ b/arch/sparc/include/asm/rwsem.h @@ -13,53 +13,12 @@ #ifdef __KERNEL__ -#include <linux/list.h> -#include <linux/spinlock.h> - -struct rwsem_waiter; - -struct rw_semaphore { - signed long count; #define RWSEM_UNLOCKED_VALUE 0x00000000L #define RWSEM_ACTIVE_BIAS 0x00000001L #define RWSEM_ACTIVE_MASK 0xffffffffL #define RWSEM_WAITING_BIAS (-RWSEM_ACTIVE_MASK-1) #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) - spinlock_t wait_lock; - struct list_head wait_list; -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -}; - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - -#define __RWSEM_INITIALIZER(name) \ -{ RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED((name).wait_lock), \ - LIST_HEAD_INIT((name).wait_list) __RWSEM_DEP_MAP_INIT(name) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ -do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ -} while (0) /* * lock for reading @@ -160,11 +119,6 @@ static inline long rwsem_atomic_update(long delta, struct rw_semaphore *sem) return atomic64_add_return(delta, (atomic64_t *)(&sem->count)); } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* __KERNEL__ */ #endif /* _SPARC64_RWSEM_H */ diff --git a/arch/sparc/lib/atomic32.c b/arch/sparc/lib/atomic32.c index cbddeb38ffda..d3c7a12ad879 100644 --- a/arch/sparc/lib/atomic32.c +++ b/arch/sparc/lib/atomic32.c @@ -16,7 +16,7 @@ #define ATOMIC_HASH(a) (&__atomic_hash[(((unsigned long)a)>>8) & (ATOMIC_HASH_SIZE-1)]) spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] = { - [0 ... (ATOMIC_HASH_SIZE-1)] = SPIN_LOCK_UNLOCKED + [0 ... (ATOMIC_HASH_SIZE-1)] = __SPIN_LOCK_UNLOCKED(__atomic_hash) }; #else /* SMP */ diff --git a/arch/um/drivers/ubd_kern.c b/arch/um/drivers/ubd_kern.c index ba4a98ba39c0..620f5b70957d 100644 --- a/arch/um/drivers/ubd_kern.c +++ b/arch/um/drivers/ubd_kern.c @@ -185,7 +185,7 @@ struct ubd { .no_cow = 0, \ .shared = 0, \ .cow = DEFAULT_COW, \ - .lock = SPIN_LOCK_UNLOCKED, \ + .lock = __SPIN_LOCK_UNLOCKED(ubd_devs.lock), \ .request = NULL, \ .start_sg = 0, \ .end_sg = 0, \ diff --git a/arch/x86/include/asm/amd_nb.h b/arch/x86/include/asm/amd_nb.h index 64dc82ee19f0..3e7070071d73 100644 --- a/arch/x86/include/asm/amd_nb.h +++ b/arch/x86/include/asm/amd_nb.h @@ -26,6 +26,7 @@ extern void amd_get_nodes(struct bootnode *nodes); struct amd_northbridge { struct pci_dev *misc; + struct pci_dev *link; }; struct amd_northbridge_info { diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 45636cefa186..4c25ab48257b 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -94,10 +94,6 @@ struct cpuinfo_x86 { int x86_cache_alignment; /* In bytes */ int x86_power; unsigned long loops_per_jiffy; -#ifdef CONFIG_SMP - /* cpus sharing the last level cache: */ - cpumask_var_t llc_shared_map; -#endif /* cpuid returned max cores value: */ u16 x86_max_cores; u16 apicid; diff --git a/arch/x86/include/asm/rwsem.h b/arch/x86/include/asm/rwsem.h index d1e41b0f9b60..df4cd32b4cc6 100644 --- a/arch/x86/include/asm/rwsem.h +++ b/arch/x86/include/asm/rwsem.h @@ -37,26 +37,9 @@ #endif #ifdef __KERNEL__ - -#include <linux/list.h> -#include <linux/spinlock.h> -#include <linux/lockdep.h> #include <asm/asm.h> -struct rwsem_waiter; - -extern asmregparm struct rw_semaphore * - rwsem_down_read_failed(struct rw_semaphore *sem); -extern asmregparm struct rw_semaphore * - rwsem_down_write_failed(struct rw_semaphore *sem); -extern asmregparm struct rw_semaphore * - rwsem_wake(struct rw_semaphore *); -extern asmregparm struct rw_semaphore * - rwsem_downgrade_wake(struct rw_semaphore *sem); - /* - * the semaphore definition - * * The bias values and the counter type limits the number of * potential readers/writers to 32767 for 32 bits and 2147483647 * for 64 bits. @@ -74,43 +57,6 @@ extern asmregparm struct rw_semaphore * #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) -typedef signed long rwsem_count_t; - -struct rw_semaphore { - rwsem_count_t count; - spinlock_t wait_lock; - struct list_head wait_list; -#ifdef CONFIG_DEBUG_LOCK_ALLOC - struct lockdep_map dep_map; -#endif -}; - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - - -#define __RWSEM_INITIALIZER(name) \ -{ \ - RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED((name).wait_lock), \ - LIST_HEAD_INIT((name).wait_list) __RWSEM_DEP_MAP_INIT(name) \ -} - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ -do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ -} while (0) - /* * lock for reading */ @@ -133,7 +79,7 @@ static inline void __down_read(struct rw_semaphore *sem) */ static inline int __down_read_trylock(struct rw_semaphore *sem) { - rwsem_count_t result, tmp; + long result, tmp; asm volatile("# beginning __down_read_trylock\n\t" " mov %0,%1\n\t" "1:\n\t" @@ -155,7 +101,7 @@ static inline int __down_read_trylock(struct rw_semaphore *sem) */ static inline void __down_write_nested(struct rw_semaphore *sem, int subclass) { - rwsem_count_t tmp; + long tmp; asm volatile("# beginning down_write\n\t" LOCK_PREFIX " xadd %1,(%2)\n\t" /* adds 0xffff0001, returns the old value */ @@ -180,9 +126,8 @@ static inline void __down_write(struct rw_semaphore *sem) */ static inline int __down_write_trylock(struct rw_semaphore *sem) { - rwsem_count_t ret = cmpxchg(&sem->count, - RWSEM_UNLOCKED_VALUE, - RWSEM_ACTIVE_WRITE_BIAS); + long ret = cmpxchg(&sem->count, RWSEM_UNLOCKED_VALUE, + RWSEM_ACTIVE_WRITE_BIAS); if (ret == RWSEM_UNLOCKED_VALUE) return 1; return 0; @@ -193,7 +138,7 @@ static inline int __down_write_trylock(struct rw_semaphore *sem) */ static inline void __up_read(struct rw_semaphore *sem) { - rwsem_count_t tmp; + long tmp; asm volatile("# beginning __up_read\n\t" LOCK_PREFIX " xadd %1,(%2)\n\t" /* subtracts 1, returns the old value */ @@ -211,7 +156,7 @@ static inline void __up_read(struct rw_semaphore *sem) */ static inline void __up_write(struct rw_semaphore *sem) { - rwsem_count_t tmp; + long tmp; asm volatile("# beginning __up_write\n\t" LOCK_PREFIX " xadd %1,(%2)\n\t" /* subtracts 0xffff0001, returns the old value */ @@ -247,8 +192,7 @@ static inline void __downgrade_write(struct rw_semaphore *sem) /* * implement atomic add functionality */ -static inline void rwsem_atomic_add(rwsem_count_t delta, - struct rw_semaphore *sem) +static inline void rwsem_atomic_add(long delta, struct rw_semaphore *sem) { asm volatile(LOCK_PREFIX _ASM_ADD "%1,%0" : "+m" (sem->count) @@ -258,10 +202,9 @@ static inline void rwsem_atomic_add(rwsem_count_t delta, /* * implement exchange and add functionality */ -static inline rwsem_count_t rwsem_atomic_update(rwsem_count_t delta, - struct rw_semaphore *sem) +static inline long rwsem_atomic_update(long delta, struct rw_semaphore *sem) { - rwsem_count_t tmp = delta; + long tmp = delta; asm volatile(LOCK_PREFIX "xadd %0,%1" : "+r" (tmp), "+m" (sem->count) @@ -270,10 +213,5 @@ static inline rwsem_count_t rwsem_atomic_update(rwsem_count_t delta, return tmp + delta; } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* __KERNEL__ */ #endif /* _ASM_X86_RWSEM_H */ diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index 4c2f63c7fc1b..3597825a3112 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -23,6 +23,8 @@ extern unsigned int num_processors; DECLARE_PER_CPU(cpumask_var_t, cpu_sibling_map); DECLARE_PER_CPU(cpumask_var_t, cpu_core_map); +/* cpus sharing the last level cache: */ +DECLARE_PER_CPU(cpumask_var_t, cpu_llc_shared_map); DECLARE_PER_CPU(u16, cpu_llc_id); DECLARE_PER_CPU(int, cpu_number); @@ -36,6 +38,11 @@ static inline struct cpumask *cpu_core_mask(int cpu) return per_cpu(cpu_core_map, cpu); } +static inline struct cpumask *cpu_llc_shared_mask(int cpu) +{ + return per_cpu(cpu_llc_shared_map, cpu); +} + DECLARE_EARLY_PER_CPU(u16, x86_cpu_to_apicid); DECLARE_EARLY_PER_CPU(u16, x86_bios_cpu_apicid); diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c index 0a99f7198bc3..4ae9a961c33c 100644 --- a/arch/x86/kernel/amd_nb.c +++ b/arch/x86/kernel/amd_nb.c @@ -20,6 +20,11 @@ struct pci_device_id amd_nb_misc_ids[] = { }; EXPORT_SYMBOL(amd_nb_misc_ids); +static struct pci_device_id amd_nb_link_ids[] = { + { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_15H_NB_LINK) }, + {} +}; + const struct amd_nb_bus_dev_range amd_nb_bus_dev_ranges[] __initconst = { { 0x00, 0x18, 0x20 }, { 0xff, 0x00, 0x20 }, @@ -45,7 +50,7 @@ int amd_cache_northbridges(void) { int i = 0; struct amd_northbridge *nb; - struct pci_dev *misc; + struct pci_dev *misc, *link; if (amd_nb_num()) return 0; @@ -64,10 +69,12 @@ int amd_cache_northbridges(void) amd_northbridges.nb = nb; amd_northbridges.num = i; - misc = NULL; + link = misc = NULL; for (i = 0; i != amd_nb_num(); i++) { node_to_amd_nb(i)->misc = misc = next_northbridge(misc, amd_nb_misc_ids); + node_to_amd_nb(i)->link = link = + next_northbridge(link, amd_nb_link_ids); } /* some CPU families (e.g. family 0x11) do not support GART */ @@ -85,6 +92,9 @@ int amd_cache_northbridges(void) boot_cpu_data.x86_mask >= 0x1)) amd_northbridges.flags |= AMD_NB_L3_INDEX_DISABLE; + if (boot_cpu_data.x86 == 0x15) + amd_northbridges.flags |= AMD_NB_L3_INDEX_DISABLE; + return 0; } EXPORT_SYMBOL_GPL(amd_cache_northbridges); diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 7c7bedb83c5a..990cc4861586 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -261,7 +261,7 @@ static int __cpuinit nearby_node(int apicid) #ifdef CONFIG_X86_HT static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) { - u32 nodes; + u32 nodes, cores_per_cu; u8 node_id; int cpu = smp_processor_id(); @@ -276,6 +276,7 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) /* get compute unit information */ smp_num_siblings = ((ebx >> 8) & 3) + 1; c->compute_unit_id = ebx & 0xff; + cores_per_cu = ((ebx >> 8) & 3) + 1; } else if (cpu_has(c, X86_FEATURE_NODEID_MSR)) { u64 value; @@ -288,15 +289,18 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c) /* fixup multi-node processor information */ if (nodes > 1) { u32 cores_per_node; + u32 cus_per_node; set_cpu_cap(c, X86_FEATURE_AMD_DCM); cores_per_node = c->x86_max_cores / nodes; + cus_per_node = cores_per_node / cores_per_cu; /* store NodeID, use llc_shared_map to store sibling info */ per_cpu(cpu_llc_id, cpu) = node_id; /* core id to be in range from 0 to (cores_per_node - 1) */ - c->cpu_core_id = c->cpu_core_id % cores_per_node; + c->cpu_core_id %= cores_per_node; + c->compute_unit_id %= cus_per_node; } } #endif diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index ec2c19a7b8ef..5419a263ebd1 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -732,11 +732,11 @@ static void __cpuinit cache_shared_cpu_map_setup(unsigned int cpu, int index) struct cpuinfo_x86 *c = &cpu_data(cpu); if ((index == 3) && (c->x86_vendor == X86_VENDOR_AMD)) { - for_each_cpu(i, c->llc_shared_map) { + for_each_cpu(i, cpu_llc_shared_mask(cpu)) { if (!per_cpu(ici_cpuid4_info, i)) continue; this_leaf = CPUID4_INFO_IDX(i, index); - for_each_cpu(sibling, c->llc_shared_map) { + for_each_cpu(sibling, cpu_llc_shared_mask(cpu)) { if (!cpu_online(sibling)) continue; set_bit(sibling, this_leaf->shared_cpu_map); diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c index 5bf2fac52aca..167f97b5596e 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c @@ -527,15 +527,12 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) int i, err = 0; struct threshold_bank *b = NULL; char name[32]; -#ifdef CONFIG_SMP - struct cpuinfo_x86 *c = &cpu_data(cpu); -#endif sprintf(name, "threshold_bank%i", bank); #ifdef CONFIG_SMP if (cpu_data(cpu).cpu_core_id && shared_bank[bank]) { /* symlink */ - i = cpumask_first(c->llc_shared_map); + i = cpumask_first(cpu_llc_shared_mask(cpu)); /* first core not up yet */ if (cpu_data(i).cpu_core_id) @@ -555,7 +552,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) if (err) goto out; - cpumask_copy(b->cpus, c->llc_shared_map); + cpumask_copy(b->cpus, cpu_llc_shared_mask(cpu)); per_cpu(threshold_banks, cpu)[bank] = b; goto out; diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 9d977a2ea693..4d98789b0664 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1389,7 +1389,7 @@ static void __init pmu_check_apic(void) pr_info("no hardware sampling interrupt available.\n"); } -int __init init_hw_perf_events(void) +static int __init init_hw_perf_events(void) { struct event_constraint *c; int err; @@ -1608,7 +1608,7 @@ out: return ret; } -int x86_pmu_event_init(struct perf_event *event) +static int x86_pmu_event_init(struct perf_event *event) { struct pmu *tmp; int err; diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index e56b9bfbabd1..f7a0993c1e7c 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -682,7 +682,7 @@ static int p4_validate_raw_event(struct perf_event *event) * if an event is shared accross the logical threads * the user needs special permissions to be able to use it */ - if (p4_event_bind_map[v].shared) { + if (p4_ht_active() && p4_event_bind_map[v].shared) { if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) return -EACCES; } @@ -727,7 +727,8 @@ static int p4_hw_config(struct perf_event *event) event->hw.config = p4_set_ht_bit(event->hw.config); if (event->attr.type == PERF_TYPE_RAW) { - + struct p4_event_bind *bind; + unsigned int esel; /* * Clear bits we reserve to be managed by kernel itself * and never allowed from a user space @@ -743,6 +744,13 @@ static int p4_hw_config(struct perf_event *event) * bits since we keep additional info here (for cache events and etc) */ event->hw.config |= event->attr.config; + bind = p4_config_get_bind(event->attr.config); + if (!bind) { + rc = -EINVAL; + goto out; + } + esel = P4_OPCODE_ESEL(bind->opcode); + event->hw.config |= p4_config_pack_cccr(P4_CCCR_ESEL(esel)); } rc = x86_setup_perfctr(event); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 0cbe8c0b35ed..7a29a00c172b 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -130,6 +130,8 @@ EXPORT_PER_CPU_SYMBOL(cpu_sibling_map); DEFINE_PER_CPU(cpumask_var_t, cpu_core_map); EXPORT_PER_CPU_SYMBOL(cpu_core_map); +DEFINE_PER_CPU(cpumask_var_t, cpu_llc_shared_map); + /* Per CPU bogomips and other parameters */ DEFINE_PER_CPU_SHARED_ALIGNED(struct cpuinfo_x86, cpu_info); EXPORT_PER_CPU_SYMBOL(cpu_info); @@ -355,23 +357,6 @@ notrace static void __cpuinit start_secondary(void *unused) cpu_idle(); } -#ifdef CONFIG_CPUMASK_OFFSTACK -/* In this case, llc_shared_map is a pointer to a cpumask. */ -static inline void copy_cpuinfo_x86(struct cpuinfo_x86 *dst, - const struct cpuinfo_x86 *src) -{ - struct cpumask *llc = dst->llc_shared_map; - *dst = *src; - dst->llc_shared_map = llc; -} -#else -static inline void copy_cpuinfo_x86(struct cpuinfo_x86 *dst, - const struct cpuinfo_x86 *src) -{ - *dst = *src; -} -#endif /* CONFIG_CPUMASK_OFFSTACK */ - /* * The bootstrap kernel entry code has set these up. Save them for * a given CPU @@ -381,7 +366,7 @@ void __cpuinit smp_store_cpu_info(int id) { struct cpuinfo_x86 *c = &cpu_data(id); - copy_cpuinfo_x86(c, &boot_cpu_data); + *c = boot_cpu_data; c->cpu_index = id; if (id != 0) identify_secondary_cpu(c); @@ -389,15 +374,12 @@ void __cpuinit smp_store_cpu_info(int id) static void __cpuinit link_thread_siblings(int cpu1, int cpu2) { - struct cpuinfo_x86 *c1 = &cpu_data(cpu1); - struct cpuinfo_x86 *c2 = &cpu_data(cpu2); - cpumask_set_cpu(cpu1, cpu_sibling_mask(cpu2)); cpumask_set_cpu(cpu2, cpu_sibling_mask(cpu1)); cpumask_set_cpu(cpu1, cpu_core_mask(cpu2)); cpumask_set_cpu(cpu2, cpu_core_mask(cpu1)); - cpumask_set_cpu(cpu1, c2->llc_shared_map); - cpumask_set_cpu(cpu2, c1->llc_shared_map); + cpumask_set_cpu(cpu1, cpu_llc_shared_mask(cpu2)); + cpumask_set_cpu(cpu2, cpu_llc_shared_mask(cpu1)); } @@ -414,6 +396,7 @@ void __cpuinit set_cpu_sibling_map(int cpu) if (cpu_has(c, X86_FEATURE_TOPOEXT)) { if (c->phys_proc_id == o->phys_proc_id && + per_cpu(cpu_llc_id, cpu) == per_cpu(cpu_llc_id, i) && c->compute_unit_id == o->compute_unit_id) link_thread_siblings(cpu, i); } else if (c->phys_proc_id == o->phys_proc_id && @@ -425,7 +408,7 @@ void __cpuinit set_cpu_sibling_map(int cpu) cpumask_set_cpu(cpu, cpu_sibling_mask(cpu)); } - cpumask_set_cpu(cpu, c->llc_shared_map); + cpumask_set_cpu(cpu, cpu_llc_shared_mask(cpu)); if (__this_cpu_read(cpu_info.x86_max_cores) == 1) { cpumask_copy(cpu_core_mask(cpu), cpu_sibling_mask(cpu)); @@ -436,8 +419,8 @@ void __cpuinit set_cpu_sibling_map(int cpu) for_each_cpu(i, cpu_sibling_setup_mask) { if (per_cpu(cpu_llc_id, cpu) != BAD_APICID && per_cpu(cpu_llc_id, cpu) == per_cpu(cpu_llc_id, i)) { - cpumask_set_cpu(i, c->llc_shared_map); - cpumask_set_cpu(cpu, cpu_data(i).llc_shared_map); + cpumask_set_cpu(i, cpu_llc_shared_mask(cpu)); + cpumask_set_cpu(cpu, cpu_llc_shared_mask(i)); } if (c->phys_proc_id == cpu_data(i).phys_proc_id) { cpumask_set_cpu(i, cpu_core_mask(cpu)); @@ -476,7 +459,7 @@ const struct cpumask *cpu_coregroup_mask(int cpu) !(cpu_has(c, X86_FEATURE_AMD_DCM))) return cpu_core_mask(cpu); else - return c->llc_shared_map; + return cpu_llc_shared_mask(cpu); } static void impress_friends(void) @@ -1089,13 +1072,13 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus) preempt_disable(); smp_cpu_index_default(); - memcpy(__this_cpu_ptr(&cpu_info), &boot_cpu_data, sizeof(cpu_info)); - cpumask_copy(cpu_callin_mask, cpumask_of(0)); - mb(); + /* * Setup boot CPU information */ smp_store_cpu_info(0); /* Final full version of the data */ + cpumask_copy(cpu_callin_mask, cpumask_of(0)); + mb(); #ifdef CONFIG_X86_32 boot_cpu_logical_apicid = logical_smp_processor_id(); #endif @@ -1103,7 +1086,7 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus) for_each_possible_cpu(i) { zalloc_cpumask_var(&per_cpu(cpu_sibling_map, i), GFP_KERNEL); zalloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL); - zalloc_cpumask_var(&cpu_data(i).llc_shared_map, GFP_KERNEL); + zalloc_cpumask_var(&per_cpu(cpu_llc_shared_map, i), GFP_KERNEL); } set_cpu_sibling_map(0); diff --git a/arch/x86/kernel/x8664_ksyms_64.c b/arch/x86/kernel/x8664_ksyms_64.c index 1b950d151e58..9796c2f3d074 100644 --- a/arch/x86/kernel/x8664_ksyms_64.c +++ b/arch/x86/kernel/x8664_ksyms_64.c @@ -52,6 +52,7 @@ extern void *__memcpy(void *, const void *, __kernel_size_t); EXPORT_SYMBOL(memset); EXPORT_SYMBOL(memcpy); EXPORT_SYMBOL(__memcpy); +EXPORT_SYMBOL(memmove); EXPORT_SYMBOL(empty_zero_page); #ifndef CONFIG_PARAVIRT diff --git a/arch/x86/lib/memmove_64.S b/arch/x86/lib/memmove_64.S new file mode 100644 index 000000000000..0ecb8433e5a8 --- /dev/null +++ b/arch/x86/lib/memmove_64.S @@ -0,0 +1,197 @@ +/* + * Normally compiler builtins are used, but sometimes the compiler calls out + * of line code. Based on asm-i386/string.h. + * + * This assembly file is re-written from memmove_64.c file. + * - Copyright 2011 Fenghua Yu <fenghua.yu@intel.com> + */ +#define _STRING_C +#include <linux/linkage.h> +#include <asm/dwarf2.h> + +#undef memmove + +/* + * Implement memmove(). This can handle overlap between src and dst. + * + * Input: + * rdi: dest + * rsi: src + * rdx: count + * + * Output: + * rax: dest + */ +ENTRY(memmove) + CFI_STARTPROC + /* Handle more 32bytes in loop */ + mov %rdi, %rax + cmp $0x20, %rdx + jb 1f + + /* Decide forward/backward copy mode */ + cmp %rdi, %rsi + jb 2f + + /* + * movsq instruction have many startup latency + * so we handle small size by general register. + */ + cmp $680, %rdx + jb 3f + /* + * movsq instruction is only good for aligned case. + */ + + cmpb %dil, %sil + je 4f +3: + sub $0x20, %rdx + /* + * We gobble 32byts forward in each loop. + */ +5: + sub $0x20, %rdx + movq 0*8(%rsi), %r11 + movq 1*8(%rsi), %r10 + movq 2*8(%rsi), %r9 + movq 3*8(%rsi), %r8 + leaq 4*8(%rsi), %rsi + + movq %r11, 0*8(%rdi) + movq %r10, 1*8(%rdi) + movq %r9, 2*8(%rdi) + movq %r8, 3*8(%rdi) + leaq 4*8(%rdi), %rdi + jae 5b + addq $0x20, %rdx + jmp 1f + /* + * Handle data forward by movsq. + */ + .p2align 4 +4: + movq %rdx, %rcx + movq -8(%rsi, %rdx), %r11 + lea -8(%rdi, %rdx), %r10 + shrq $3, %rcx + rep movsq + movq %r11, (%r10) + jmp 13f + /* + * Handle data backward by movsq. + */ + .p2align 4 +7: + movq %rdx, %rcx + movq (%rsi), %r11 + movq %rdi, %r10 + leaq -8(%rsi, %rdx), %rsi + leaq -8(%rdi, %rdx), %rdi + shrq $3, %rcx + std + rep movsq + cld + movq %r11, (%r10) + jmp 13f + + /* + * Start to prepare for backward copy. + */ + .p2align 4 +2: + cmp $680, %rdx + jb 6f + cmp %dil, %sil + je 7b +6: + /* + * Calculate copy position to tail. + */ + addq %rdx, %rsi + addq %rdx, %rdi + subq $0x20, %rdx + /* + * We gobble 32byts backward in each loop. + */ +8: + subq $0x20, %rdx + movq -1*8(%rsi), %r11 + movq -2*8(%rsi), %r10 + movq -3*8(%rsi), %r9 + movq -4*8(%rsi), %r8 + leaq -4*8(%rsi), %rsi + + movq %r11, -1*8(%rdi) + movq %r10, -2*8(%rdi) + movq %r9, -3*8(%rdi) + movq %r8, -4*8(%rdi) + leaq -4*8(%rdi), %rdi + jae 8b + /* + * Calculate copy position to head. + */ + addq $0x20, %rdx + subq %rdx, %rsi + subq %rdx, %rdi +1: + cmpq $16, %rdx + jb 9f + /* + * Move data from 16 bytes to 31 bytes. + */ + movq 0*8(%rsi), %r11 + movq 1*8(%rsi), %r10 + movq -2*8(%rsi, %rdx), %r9 + movq -1*8(%rsi, %rdx), %r8 + movq %r11, 0*8(%rdi) + movq %r10, 1*8(%rdi) + movq %r9, -2*8(%rdi, %rdx) + movq %r8, -1*8(%rdi, %rdx) + jmp 13f + .p2align 4 +9: + cmpq $8, %rdx + jb 10f + /* + * Move data from 8 bytes to 15 bytes. + */ + movq 0*8(%rsi), %r11 + movq -1*8(%rsi, %rdx), %r10 + movq %r11, 0*8(%rdi) + movq %r10, -1*8(%rdi, %rdx) + jmp 13f +10: + cmpq $4, %rdx + jb 11f + /* + * Move data from 4 bytes to 7 bytes. + */ + movl (%rsi), %r11d + movl -4(%rsi, %rdx), %r10d + movl %r11d, (%rdi) + movl %r10d, -4(%rdi, %rdx) + jmp 13f +11: + cmp $2, %rdx + jb 12f + /* + * Move data from 2 bytes to 3 bytes. + */ + movw (%rsi), %r11w + movw -2(%rsi, %rdx), %r10w + movw %r11w, (%rdi) + movw %r10w, -2(%rdi, %rdx) + jmp 13f +12: + cmp $1, %rdx + jb 13f + /* + * Move data for 1 byte. + */ + movb (%rsi), %r11b + movb %r11b, (%rdi) +13: + retq + CFI_ENDPROC +ENDPROC(memmove) diff --git a/arch/x86/lib/memmove_64.c b/arch/x86/lib/memmove_64.c deleted file mode 100644 index 6d0f0ec41b34..000000000000 --- a/arch/x86/lib/memmove_64.c +++ /dev/null @@ -1,192 +0,0 @@ -/* Normally compiler builtins are used, but sometimes the compiler calls out - of line code. Based on asm-i386/string.h. - */ -#define _STRING_C -#include <linux/string.h> -#include <linux/module.h> - -#undef memmove -void *memmove(void *dest, const void *src, size_t count) -{ - unsigned long d0,d1,d2,d3,d4,d5,d6,d7; - char *ret; - - __asm__ __volatile__( - /* Handle more 32bytes in loop */ - "mov %2, %3\n\t" - "cmp $0x20, %0\n\t" - "jb 1f\n\t" - - /* Decide forward/backward copy mode */ - "cmp %2, %1\n\t" - "jb 2f\n\t" - - /* - * movsq instruction have many startup latency - * so we handle small size by general register. - */ - "cmp $680, %0\n\t" - "jb 3f\n\t" - /* - * movsq instruction is only good for aligned case. - */ - "cmpb %%dil, %%sil\n\t" - "je 4f\n\t" - "3:\n\t" - "sub $0x20, %0\n\t" - /* - * We gobble 32byts forward in each loop. - */ - "5:\n\t" - "sub $0x20, %0\n\t" - "movq 0*8(%1), %4\n\t" - "movq 1*8(%1), %5\n\t" - "movq 2*8(%1), %6\n\t" - "movq 3*8(%1), %7\n\t" - "leaq 4*8(%1), %1\n\t" - - "movq %4, 0*8(%2)\n\t" - "movq %5, 1*8(%2)\n\t" - "movq %6, 2*8(%2)\n\t" - "movq %7, 3*8(%2)\n\t" - "leaq 4*8(%2), %2\n\t" - "jae 5b\n\t" - "addq $0x20, %0\n\t" - "jmp 1f\n\t" - /* - * Handle data forward by movsq. - */ - ".p2align 4\n\t" - "4:\n\t" - "movq %0, %8\n\t" - "movq -8(%1, %0), %4\n\t" - "lea -8(%2, %0), %5\n\t" - "shrq $3, %8\n\t" - "rep movsq\n\t" - "movq %4, (%5)\n\t" - "jmp 13f\n\t" - /* - * Handle data backward by movsq. - */ - ".p2align 4\n\t" - "7:\n\t" - "movq %0, %8\n\t" - "movq (%1), %4\n\t" - "movq %2, %5\n\t" - "leaq -8(%1, %0), %1\n\t" - "leaq -8(%2, %0), %2\n\t" - "shrq $3, %8\n\t" - "std\n\t" - "rep movsq\n\t" - "cld\n\t" - "movq %4, (%5)\n\t" - "jmp 13f\n\t" - - /* - * Start to prepare for backward copy. - */ - ".p2align 4\n\t" - "2:\n\t" - "cmp $680, %0\n\t" - "jb 6f \n\t" - "cmp %%dil, %%sil\n\t" - "je 7b \n\t" - "6:\n\t" - /* - * Calculate copy position to tail. - */ - "addq %0, %1\n\t" - "addq %0, %2\n\t" - "subq $0x20, %0\n\t" - /* - * We gobble 32byts backward in each loop. - */ - "8:\n\t" - "subq $0x20, %0\n\t" - "movq -1*8(%1), %4\n\t" - "movq -2*8(%1), %5\n\t" - "movq -3*8(%1), %6\n\t" - "movq -4*8(%1), %7\n\t" - "leaq -4*8(%1), %1\n\t" - - "movq %4, -1*8(%2)\n\t" - "movq %5, -2*8(%2)\n\t" - "movq %6, -3*8(%2)\n\t" - "movq %7, -4*8(%2)\n\t" - "leaq -4*8(%2), %2\n\t" - "jae 8b\n\t" - /* - * Calculate copy position to head. - */ - "addq $0x20, %0\n\t" - "subq %0, %1\n\t" - "subq %0, %2\n\t" - "1:\n\t" - "cmpq $16, %0\n\t" - "jb 9f\n\t" - /* - * Move data from 16 bytes to 31 bytes. - */ - "movq 0*8(%1), %4\n\t" - "movq 1*8(%1), %5\n\t" - "movq -2*8(%1, %0), %6\n\t" - "movq -1*8(%1, %0), %7\n\t" - "movq %4, 0*8(%2)\n\t" - "movq %5, 1*8(%2)\n\t" - "movq %6, -2*8(%2, %0)\n\t" - "movq %7, -1*8(%2, %0)\n\t" - "jmp 13f\n\t" - ".p2align 4\n\t" - "9:\n\t" - "cmpq $8, %0\n\t" - "jb 10f\n\t" - /* - * Move data from 8 bytes to 15 bytes. - */ - "movq 0*8(%1), %4\n\t" - "movq -1*8(%1, %0), %5\n\t" - "movq %4, 0*8(%2)\n\t" - "movq %5, -1*8(%2, %0)\n\t" - "jmp 13f\n\t" - "10:\n\t" - "cmpq $4, %0\n\t" - "jb 11f\n\t" - /* - * Move data from 4 bytes to 7 bytes. - */ - "movl (%1), %4d\n\t" - "movl -4(%1, %0), %5d\n\t" - "movl %4d, (%2)\n\t" - "movl %5d, -4(%2, %0)\n\t" - "jmp 13f\n\t" - "11:\n\t" - "cmp $2, %0\n\t" - "jb 12f\n\t" - /* - * Move data from 2 bytes to 3 bytes. - */ - "movw (%1), %4w\n\t" - "movw -2(%1, %0), %5w\n\t" - "movw %4w, (%2)\n\t" - "movw %5w, -2(%2, %0)\n\t" - "jmp 13f\n\t" - "12:\n\t" - "cmp $1, %0\n\t" - "jb 13f\n\t" - /* - * Move data for 1 byte. - */ - "movb (%1), %4b\n\t" - "movb %4b, (%2)\n\t" - "13:\n\t" - : "=&d" (d0), "=&S" (d1), "=&D" (d2), "=&a" (ret) , - "=r"(d3), "=r"(d4), "=r"(d5), "=r"(d6), "=&c" (d7) - :"0" (count), - "1" (src), - "2" (dest) - :"memory"); - - return ret; - -} -EXPORT_SYMBOL(memmove); diff --git a/arch/xtensa/include/asm/rwsem.h b/arch/xtensa/include/asm/rwsem.h index e39edf5c86f2..249619e7e7f2 100644 --- a/arch/xtensa/include/asm/rwsem.h +++ b/arch/xtensa/include/asm/rwsem.h @@ -17,44 +17,12 @@ #error "Please don't include <asm/rwsem.h> directly, use <linux/rwsem.h> instead." #endif -#include <linux/list.h> -#include <linux/spinlock.h> -#include <asm/atomic.h> -#include <asm/system.h> - -/* - * the semaphore definition - */ -struct rw_semaphore { - signed long count; #define RWSEM_UNLOCKED_VALUE 0x00000000 #define RWSEM_ACTIVE_BIAS 0x00000001 #define RWSEM_ACTIVE_MASK 0x0000ffff #define RWSEM_WAITING_BIAS (-0x00010000) #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) - spinlock_t wait_lock; - struct list_head wait_list; -}; - -#define __RWSEM_INITIALIZER(name) \ - { RWSEM_UNLOCKED_VALUE, SPIN_LOCK_UNLOCKED, \ - LIST_HEAD_INIT((name).wait_list) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem); -extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); - -static inline void init_rwsem(struct rw_semaphore *sem) -{ - sem->count = RWSEM_UNLOCKED_VALUE; - spin_lock_init(&sem->wait_lock); - INIT_LIST_HEAD(&sem->wait_list); -} /* * lock for reading @@ -160,9 +128,4 @@ static inline int rwsem_atomic_update(int delta, struct rw_semaphore *sem) return atomic_add_return(delta, (atomic_t *)(&sem->count)); } -static inline int rwsem_is_locked(struct rw_semaphore *sem) -{ - return (sem->count != 0); -} - #endif /* _XTENSA_RWSEM_H */ diff --git a/include/asm-generic/cputime.h b/include/asm-generic/cputime.h index 2bcc5c7c22a6..61e03dd7939e 100644 --- a/include/asm-generic/cputime.h +++ b/include/asm-generic/cputime.h @@ -30,6 +30,9 @@ typedef u64 cputime64_t; #define cputime64_to_jiffies64(__ct) (__ct) #define jiffies64_to_cputime64(__jif) (__jif) #define cputime_to_cputime64(__ct) ((u64) __ct) +#define cputime64_gt(__a, __b) ((__a) > (__b)) + +#define nsecs_to_cputime64(__ct) nsecs_to_jiffies64(__ct) /* diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 63c5ad78e37c..e89c5c735d1e 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -457,6 +457,13 @@ extern void raise_softirq(unsigned int nr); */ DECLARE_PER_CPU(struct list_head [NR_SOFTIRQS], softirq_work_list); +DECLARE_PER_CPU(struct task_struct *, ksoftirqd); + +static inline struct task_struct *this_cpu_ksoftirqd(void) +{ + return this_cpu_read(ksoftirqd); +} + /* Try to send a softirq to a remote cpu. If this cannot be done, the * work will be queued to the local cpu. */ diff --git a/include/linux/jiffies.h b/include/linux/jiffies.h index 6811f4bfc6e7..922aa313c9f9 100644 --- a/include/linux/jiffies.h +++ b/include/linux/jiffies.h @@ -307,6 +307,7 @@ extern clock_t jiffies_to_clock_t(long x); extern unsigned long clock_t_to_jiffies(unsigned long x); extern u64 jiffies_64_to_clock_t(u64 x); extern u64 nsec_to_clock_t(u64 x); +extern u64 nsecs_to_jiffies64(u64 n); extern unsigned long nsecs_to_jiffies(u64 n); #define TIMESTAMP_SIZE 30 diff --git a/include/linux/kthread.h b/include/linux/kthread.h index ce0775aa64c3..7ff16f7d3ed4 100644 --- a/include/linux/kthread.h +++ b/include/linux/kthread.h @@ -64,7 +64,7 @@ struct kthread_work { }; #define KTHREAD_WORKER_INIT(worker) { \ - .lock = SPIN_LOCK_UNLOCKED, \ + .lock = __SPIN_LOCK_UNLOCKED((worker).lock), \ .work_list = LIST_HEAD_INIT((worker).work_list), \ } diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 3adb06ebf841..580de67f318b 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -518,6 +518,7 @@ #define PCI_DEVICE_ID_AMD_11H_NB_MISC 0x1303 #define PCI_DEVICE_ID_AMD_11H_NB_LINK 0x1304 #define PCI_DEVICE_ID_AMD_15H_NB_MISC 0x1603 +#define PCI_DEVICE_ID_AMD_15H_NB_LINK 0x1604 #define PCI_DEVICE_ID_AMD_CNB17H_F3 0x1703 #define PCI_DEVICE_ID_AMD_LANCE 0x2000 #define PCI_DEVICE_ID_AMD_LANCE_HOME 0x2001 diff --git a/include/linux/rwlock_types.h b/include/linux/rwlock_types.h index bd31808c7d8e..cc0072e93e36 100644 --- a/include/linux/rwlock_types.h +++ b/include/linux/rwlock_types.h @@ -43,14 +43,6 @@ typedef struct { RW_DEP_MAP_INIT(lockname) } #endif -/* - * RW_LOCK_UNLOCKED defeat lockdep state tracking and is hence - * deprecated. - * - * Please use DEFINE_RWLOCK() or __RW_LOCK_UNLOCKED() as appropriate. - */ -#define RW_LOCK_UNLOCKED __RW_LOCK_UNLOCKED(old_style_rw_init) - #define DEFINE_RWLOCK(x) rwlock_t x = __RW_LOCK_UNLOCKED(x) #endif /* __LINUX_RWLOCK_TYPES_H */ diff --git a/include/linux/rwsem-spinlock.h b/include/linux/rwsem-spinlock.h index bdfcc2527970..34701241b673 100644 --- a/include/linux/rwsem-spinlock.h +++ b/include/linux/rwsem-spinlock.h @@ -12,15 +12,7 @@ #error "please don't include linux/rwsem-spinlock.h directly, use linux/rwsem.h instead" #endif -#include <linux/spinlock.h> -#include <linux/list.h> - #ifdef __KERNEL__ - -#include <linux/types.h> - -struct rwsem_waiter; - /* * the rw-semaphore definition * - if activity is 0 then there are no active readers or writers @@ -37,28 +29,7 @@ struct rw_semaphore { #endif }; -#ifdef CONFIG_DEBUG_LOCK_ALLOC -# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } -#else -# define __RWSEM_DEP_MAP_INIT(lockname) -#endif - -#define __RWSEM_INITIALIZER(name) \ -{ 0, __SPIN_LOCK_UNLOCKED(name.wait_lock), LIST_HEAD_INIT((name).wait_list) \ - __RWSEM_DEP_MAP_INIT(name) } - -#define DECLARE_RWSEM(name) \ - struct rw_semaphore name = __RWSEM_INITIALIZER(name) - -extern void __init_rwsem(struct rw_semaphore *sem, const char *name, - struct lock_class_key *key); - -#define init_rwsem(sem) \ -do { \ - static struct lock_class_key __key; \ - \ - __init_rwsem((sem), #sem, &__key); \ -} while (0) +#define RWSEM_UNLOCKED_VALUE 0x00000000 extern void __down_read(struct rw_semaphore *sem); extern int __down_read_trylock(struct rw_semaphore *sem); diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h index efd348fe8ca7..a8afe9cd000c 100644 --- a/include/linux/rwsem.h +++ b/include/linux/rwsem.h @@ -11,6 +11,9 @@ #include <linux/types.h> #include <linux/kernel.h> +#include <linux/list.h> +#include <linux/spinlock.h> + #include <asm/system.h> #include <asm/atomic.h> @@ -19,9 +22,57 @@ struct rw_semaphore; #ifdef CONFIG_RWSEM_GENERIC_SPINLOCK #include <linux/rwsem-spinlock.h> /* use a generic implementation */ #else -#include <asm/rwsem.h> /* use an arch-specific implementation */ +/* All arch specific implementations share the same struct */ +struct rw_semaphore { + long count; + spinlock_t wait_lock; + struct list_head wait_list; +#ifdef CONFIG_DEBUG_LOCK_ALLOC + struct lockdep_map dep_map; +#endif +}; + +extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem); +extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem); +extern struct rw_semaphore *rwsem_wake(struct rw_semaphore *); +extern struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem); + +/* Include the arch specific part */ +#include <asm/rwsem.h> + +/* In all implementations count != 0 means locked */ +static inline int rwsem_is_locked(struct rw_semaphore *sem) +{ + return sem->count != 0; +} + +#endif + +/* Common initializer macros and functions */ + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname } +#else +# define __RWSEM_DEP_MAP_INIT(lockname) #endif +#define __RWSEM_INITIALIZER(name) \ + { RWSEM_UNLOCKED_VALUE, __SPIN_LOCK_UNLOCKED(name.wait_lock), \ + LIST_HEAD_INIT((name).wait_list) __RWSEM_DEP_MAP_INIT(name) } + +#define DECLARE_RWSEM(name) \ + struct rw_semaphore name = __RWSEM_INITIALIZER(name) + +extern void __init_rwsem(struct rw_semaphore *sem, const char *name, + struct lock_class_key *key); + +#define init_rwsem(sem) \ +do { \ + static struct lock_class_key __key; \ + \ + __init_rwsem((sem), #sem, &__key); \ +} while (0) + /* * lock for reading */ diff --git a/include/linux/sched.h b/include/linux/sched.h index d747f948b34e..0542774914d4 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1084,12 +1084,10 @@ struct sched_class { void (*task_tick) (struct rq *rq, struct task_struct *p, int queued); void (*task_fork) (struct task_struct *p); - void (*switched_from) (struct rq *this_rq, struct task_struct *task, - int running); - void (*switched_to) (struct rq *this_rq, struct task_struct *task, - int running); + void (*switched_from) (struct rq *this_rq, struct task_struct *task); + void (*switched_to) (struct rq *this_rq, struct task_struct *task); void (*prio_changed) (struct rq *this_rq, struct task_struct *task, - int oldprio, int running); + int oldprio); unsigned int (*get_rr_interval) (struct rq *rq, struct task_struct *task); @@ -1715,7 +1713,6 @@ extern void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t * /* * Per process flags */ -#define PF_KSOFTIRQD 0x00000001 /* I am ksoftirqd */ #define PF_STARTING 0x00000002 /* being created */ #define PF_EXITING 0x00000004 /* getting shut down */ #define PF_EXITPIDONE 0x00000008 /* pi exit done on shut down */ diff --git a/include/linux/spinlock_types.h b/include/linux/spinlock_types.h index 851b7783720d..73548eb13a5d 100644 --- a/include/linux/spinlock_types.h +++ b/include/linux/spinlock_types.h @@ -81,14 +81,6 @@ typedef struct spinlock { #define __SPIN_LOCK_UNLOCKED(lockname) \ (spinlock_t ) __SPIN_LOCK_INITIALIZER(lockname) -/* - * SPIN_LOCK_UNLOCKED defeats lockdep state tracking and is hence - * deprecated. - * Please use DEFINE_SPINLOCK() or __SPIN_LOCK_UNLOCKED() as - * appropriate. - */ -#define SPIN_LOCK_UNLOCKED __SPIN_LOCK_UNLOCKED(old_style_spin_init) - #define DEFINE_SPINLOCK(x) spinlock_t x = __SPIN_LOCK_UNLOCKED(x) #include <linux/rwlock_types.h> diff --git a/kernel/cred.c b/kernel/cred.c index 6a1aa004e376..b5496e81b0f7 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -35,7 +35,7 @@ static struct kmem_cache *cred_jar; static struct thread_group_cred init_tgcred = { .usage = ATOMIC_INIT(2), .tgid = 0, - .lock = SPIN_LOCK_UNLOCKED, + .lock = __SPIN_LOCK_UNLOCKED(init_cred.tgcred.lock), }; #endif diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 0587c5ceaed8..538fce2db51c 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -179,7 +179,7 @@ static void irq_affinity_notify(struct work_struct *work) cpumask_copy(cpumask, desc->pending_mask); else #endif - cpumask_copy(cpumask, desc->affinity); + cpumask_copy(cpumask, desc->irq_data.affinity); raw_spin_unlock_irqrestore(&desc->lock, flags); notify->notify(notify, cpumask); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 126a302c481c..852ae8c66502 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1999,8 +1999,7 @@ static int alloc_callchain_buffers(void) * accessed from NMI. Use a temporary manual per cpu allocation * until that gets sorted out. */ - size = sizeof(*entries) + sizeof(struct perf_callchain_entry *) * - num_possible_cpus(); + size = offsetof(struct callchain_cpus_entries, cpu_entries[nr_cpu_ids]); entries = kzalloc(size, GFP_KERNEL); if (!entries) diff --git a/kernel/rtmutex-tester.c b/kernel/rtmutex-tester.c index 66cb89bc5ef1..d5b543506cbc 100644 --- a/kernel/rtmutex-tester.c +++ b/kernel/rtmutex-tester.c @@ -9,7 +9,6 @@ #include <linux/kthread.h> #include <linux/module.h> #include <linux/sched.h> -#include <linux/smp_lock.h> #include <linux/spinlock.h> #include <linux/sysdev.h> #include <linux/timer.h> @@ -27,7 +26,6 @@ struct test_thread_data { int opcode; int opdata; int mutexes[MAX_RT_TEST_MUTEXES]; - int bkl; int event; struct sys_device sysdev; }; @@ -46,8 +44,8 @@ enum test_opcodes { RTTEST_LOCKINTNOWAIT, /* 6 Lock interruptible no wait in wakeup, data = lockindex */ RTTEST_LOCKCONT, /* 7 Continue locking after the wakeup delay */ RTTEST_UNLOCK, /* 8 Unlock, data = lockindex */ - RTTEST_LOCKBKL, /* 9 Lock BKL */ - RTTEST_UNLOCKBKL, /* 10 Unlock BKL */ + RTTEST_LOCKBKL, /* 9 Was: Lock BKL */ + RTTEST_UNLOCKBKL, /* 10 Was: Unlock BKL */ RTTEST_SIGNAL, /* 11 Signal other test thread, data = thread id */ RTTEST_RESETEVENT = 98, /* 98 Reset event counter */ RTTEST_RESET = 99, /* 99 Reset all pending operations */ @@ -74,13 +72,6 @@ static int handle_op(struct test_thread_data *td, int lockwakeup) td->mutexes[i] = 0; } } - - if (!lockwakeup && td->bkl == 4) { -#ifdef CONFIG_LOCK_KERNEL - unlock_kernel(); -#endif - td->bkl = 0; - } return 0; case RTTEST_RESETEVENT: @@ -131,25 +122,6 @@ static int handle_op(struct test_thread_data *td, int lockwakeup) td->mutexes[id] = 0; return 0; - case RTTEST_LOCKBKL: - if (td->bkl) - return 0; - td->bkl = 1; -#ifdef CONFIG_LOCK_KERNEL - lock_kernel(); -#endif - td->bkl = 4; - return 0; - - case RTTEST_UNLOCKBKL: - if (td->bkl != 4) - break; -#ifdef CONFIG_LOCK_KERNEL - unlock_kernel(); -#endif - td->bkl = 0; - return 0; - default: break; } @@ -196,7 +168,6 @@ void schedule_rt_mutex_test(struct rt_mutex *mutex) td->event = atomic_add_return(1, &rttest_event); break; - case RTTEST_LOCKBKL: default: break; } @@ -229,8 +200,6 @@ void schedule_rt_mutex_test(struct rt_mutex *mutex) td->event = atomic_add_return(1, &rttest_event); return; - case RTTEST_LOCKBKL: - return; default: return; } @@ -380,11 +349,11 @@ static ssize_t sysfs_test_status(struct sys_device *dev, struct sysdev_attribute spin_lock(&rttest_lock); curr += sprintf(curr, - "O: %4d, E:%8d, S: 0x%08lx, P: %4d, N: %4d, B: %p, K: %d, M:", + "O: %4d, E:%8d, S: 0x%08lx, P: %4d, N: %4d, B: %p, M:", td->opcode, td->event, tsk->state, (MAX_RT_PRIO - 1) - tsk->prio, (MAX_RT_PRIO - 1) - tsk->normal_prio, - tsk->pi_blocked_on, td->bkl); + tsk->pi_blocked_on); for (i = MAX_RT_TEST_MUTEXES - 1; i >=0 ; i--) curr += sprintf(curr, "%d", td->mutexes[i]); diff --git a/kernel/sched.c b/kernel/sched.c index 18d38e4ec7ba..477e1bcc63f9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1880,7 +1880,7 @@ void account_system_vtime(struct task_struct *curr) */ if (hardirq_count()) __this_cpu_add(cpu_hardirq_time, delta); - else if (in_serving_softirq() && !(curr->flags & PF_KSOFTIRQD)) + else if (in_serving_softirq() && curr != this_cpu_ksoftirqd()) __this_cpu_add(cpu_softirq_time, delta); irq_time_write_end(); @@ -1920,8 +1920,40 @@ static void update_rq_clock_task(struct rq *rq, s64 delta) sched_rt_avg_update(rq, irq_delta); } +static int irqtime_account_hi_update(void) +{ + struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + unsigned long flags; + u64 latest_ns; + int ret = 0; + + local_irq_save(flags); + latest_ns = this_cpu_read(cpu_hardirq_time); + if (cputime64_gt(nsecs_to_cputime64(latest_ns), cpustat->irq)) + ret = 1; + local_irq_restore(flags); + return ret; +} + +static int irqtime_account_si_update(void) +{ + struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + unsigned long flags; + u64 latest_ns; + int ret = 0; + + local_irq_save(flags); + latest_ns = this_cpu_read(cpu_softirq_time); + if (cputime64_gt(nsecs_to_cputime64(latest_ns), cpustat->softirq)) + ret = 1; + local_irq_restore(flags); + return ret; +} + #else /* CONFIG_IRQ_TIME_ACCOUNTING */ +#define sched_clock_irqtime (0) + static void update_rq_clock_task(struct rq *rq, s64 delta) { rq->clock_task += delta; @@ -2025,14 +2057,14 @@ inline int task_curr(const struct task_struct *p) static inline void check_class_changed(struct rq *rq, struct task_struct *p, const struct sched_class *prev_class, - int oldprio, int running) + int oldprio) { if (prev_class != p->sched_class) { if (prev_class->switched_from) - prev_class->switched_from(rq, p, running); - p->sched_class->switched_to(rq, p, running); - } else - p->sched_class->prio_changed(rq, p, oldprio, running); + prev_class->switched_from(rq, p); + p->sched_class->switched_to(rq, p); + } else if (oldprio != p->prio) + p->sched_class->prio_changed(rq, p, oldprio); } static void check_preempt_curr(struct rq *rq, struct task_struct *p, int flags) @@ -2566,6 +2598,7 @@ static void __sched_fork(struct task_struct *p) p->se.sum_exec_runtime = 0; p->se.prev_sum_exec_runtime = 0; p->se.nr_migrations = 0; + p->se.vruntime = 0; #ifdef CONFIG_SCHEDSTATS memset(&p->se.statistics, 0, sizeof(p->se.statistics)); @@ -3568,6 +3601,32 @@ static void account_guest_time(struct task_struct *p, cputime_t cputime, } /* + * Account system cpu time to a process and desired cpustat field + * @p: the process that the cpu time gets accounted to + * @cputime: the cpu time spent in kernel space since the last update + * @cputime_scaled: cputime scaled by cpu frequency + * @target_cputime64: pointer to cpustat field that has to be updated + */ +static inline +void __account_system_time(struct task_struct *p, cputime_t cputime, + cputime_t cputime_scaled, cputime64_t *target_cputime64) +{ + cputime64_t tmp = cputime_to_cputime64(cputime); + + /* Add system time to process. */ + p->stime = cputime_add(p->stime, cputime); + p->stimescaled = cputime_add(p->stimescaled, cputime_scaled); + account_group_system_time(p, cputime); + + /* Add system time to cpustat. */ + *target_cputime64 = cputime64_add(*target_cputime64, tmp); + cpuacct_update_stats(p, CPUACCT_STAT_SYSTEM, cputime); + + /* Account for system time used */ + acct_update_integrals(p); +} + +/* * Account system cpu time to a process. * @p: the process that the cpu time gets accounted to * @hardirq_offset: the offset to subtract from hardirq_count() @@ -3578,33 +3637,90 @@ void account_system_time(struct task_struct *p, int hardirq_offset, cputime_t cputime, cputime_t cputime_scaled) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; - cputime64_t tmp; + cputime64_t *target_cputime64; if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) { account_guest_time(p, cputime, cputime_scaled); return; } - /* Add system time to process. */ - p->stime = cputime_add(p->stime, cputime); - p->stimescaled = cputime_add(p->stimescaled, cputime_scaled); - account_group_system_time(p, cputime); - - /* Add system time to cpustat. */ - tmp = cputime_to_cputime64(cputime); if (hardirq_count() - hardirq_offset) - cpustat->irq = cputime64_add(cpustat->irq, tmp); + target_cputime64 = &cpustat->irq; else if (in_serving_softirq()) - cpustat->softirq = cputime64_add(cpustat->softirq, tmp); + target_cputime64 = &cpustat->softirq; else - cpustat->system = cputime64_add(cpustat->system, tmp); + target_cputime64 = &cpustat->system; - cpuacct_update_stats(p, CPUACCT_STAT_SYSTEM, cputime); + __account_system_time(p, cputime, cputime_scaled, target_cputime64); +} - /* Account for system time used */ - acct_update_integrals(p); +#ifdef CONFIG_IRQ_TIME_ACCOUNTING +/* + * Account a tick to a process and cpustat + * @p: the process that the cpu time gets accounted to + * @user_tick: is the tick from userspace + * @rq: the pointer to rq + * + * Tick demultiplexing follows the order + * - pending hardirq update + * - pending softirq update + * - user_time + * - idle_time + * - system time + * - check for guest_time + * - else account as system_time + * + * Check for hardirq is done both for system and user time as there is + * no timer going off while we are on hardirq and hence we may never get an + * opportunity to update it solely in system time. + * p->stime and friends are only updated on system time and not on irq + * softirq as those do not count in task exec_runtime any more. + */ +static void irqtime_account_process_tick(struct task_struct *p, int user_tick, + struct rq *rq) +{ + cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy); + cputime64_t tmp = cputime_to_cputime64(cputime_one_jiffy); + struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + + if (irqtime_account_hi_update()) { + cpustat->irq = cputime64_add(cpustat->irq, tmp); + } else if (irqtime_account_si_update()) { + cpustat->softirq = cputime64_add(cpustat->softirq, tmp); + } else if (this_cpu_ksoftirqd() == p) { + /* + * ksoftirqd time do not get accounted in cpu_softirq_time. + * So, we have to handle it separately here. + * Also, p->stime needs to be updated for ksoftirqd. + */ + __account_system_time(p, cputime_one_jiffy, one_jiffy_scaled, + &cpustat->softirq); + } else if (user_tick) { + account_user_time(p, cputime_one_jiffy, one_jiffy_scaled); + } else if (p == rq->idle) { + account_idle_time(cputime_one_jiffy); + } else if (p->flags & PF_VCPU) { /* System time or guest time */ + account_guest_time(p, cputime_one_jiffy, one_jiffy_scaled); + } else { + __account_system_time(p, cputime_one_jiffy, one_jiffy_scaled, + &cpustat->system); + } } +static void irqtime_account_idle_ticks(int ticks) +{ + int i; + struct rq *rq = this_rq(); + + for (i = 0; i < ticks; i++) + irqtime_account_process_tick(current, 0, rq); +} +#else +static void irqtime_account_idle_ticks(int ticks) {} +static void irqtime_account_process_tick(struct task_struct *p, int user_tick, + struct rq *rq) {} +#endif + /* * Account for involuntary wait time. * @steal: the cpu time spent in involuntary wait @@ -3645,6 +3761,11 @@ void account_process_tick(struct task_struct *p, int user_tick) cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy); struct rq *rq = this_rq(); + if (sched_clock_irqtime) { + irqtime_account_process_tick(p, user_tick, rq); + return; + } + if (user_tick) account_user_time(p, cputime_one_jiffy, one_jiffy_scaled); else if ((p != rq->idle) || (irq_count() != HARDIRQ_OFFSET)) @@ -3670,6 +3791,12 @@ void account_steal_ticks(unsigned long ticks) */ void account_idle_ticks(unsigned long ticks) { + + if (sched_clock_irqtime) { + irqtime_account_idle_ticks(ticks); + return; + } + account_idle_time(jiffies_to_cputime(ticks)); } @@ -4570,11 +4697,10 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); - if (on_rq) { + if (on_rq) enqueue_task(rq, p, oldprio < prio ? ENQUEUE_HEAD : 0); - check_class_changed(rq, p, prev_class, oldprio, running); - } + check_class_changed(rq, p, prev_class, oldprio); task_rq_unlock(rq, &flags); } @@ -4902,11 +5028,10 @@ recheck: if (running) p->sched_class->set_curr_task(rq); - if (on_rq) { + if (on_rq) activate_task(rq, p, 0); - check_class_changed(rq, p, prev_class, oldprio, running); - } + check_class_changed(rq, p, prev_class, oldprio); __task_rq_unlock(rq); raw_spin_unlock_irqrestore(&p->pi_lock, flags); @@ -7796,6 +7921,10 @@ static void init_cfs_rq(struct cfs_rq *cfs_rq, struct rq *rq) INIT_LIST_HEAD(&cfs_rq->tasks); #ifdef CONFIG_FAIR_GROUP_SCHED cfs_rq->rq = rq; + /* allow initial update_cfs_load() to truncate */ +#ifdef CONFIG_SMP + cfs_rq->load_stamp = 1; +#endif #endif cfs_rq->min_vruntime = (u64)(-(1LL << 20)); } @@ -8109,6 +8238,8 @@ EXPORT_SYMBOL(__might_sleep); #ifdef CONFIG_MAGIC_SYSRQ static void normalize_task(struct rq *rq, struct task_struct *p) { + const struct sched_class *prev_class = p->sched_class; + int old_prio = p->prio; int on_rq; on_rq = p->se.on_rq; @@ -8119,6 +8250,8 @@ static void normalize_task(struct rq *rq, struct task_struct *p) activate_task(rq, p, 0); resched_task(rq->curr); } + + check_class_changed(rq, p, prev_class, old_prio); } void normalize_rt_tasks(void) @@ -8510,7 +8643,7 @@ int sched_group_set_shares(struct task_group *tg, unsigned long shares) /* Propagate contribution to hierarchy */ raw_spin_lock_irqsave(&rq->lock, flags); for_each_sched_entity(se) - update_cfs_shares(group_cfs_rq(se), 0); + update_cfs_shares(group_cfs_rq(se)); raw_spin_unlock_irqrestore(&rq->lock, flags); } diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 0c26e2df450e..55040f3938d8 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -540,7 +540,7 @@ static u64 sched_vslice(struct cfs_rq *cfs_rq, struct sched_entity *se) } static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update); -static void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta); +static void update_cfs_shares(struct cfs_rq *cfs_rq); /* * Update the current task's runtime statistics. Skip current tasks that @@ -733,6 +733,7 @@ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) now - cfs_rq->load_last > 4 * period) { cfs_rq->load_period = 0; cfs_rq->load_avg = 0; + delta = period - 1; } cfs_rq->load_stamp = now; @@ -763,16 +764,15 @@ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) list_del_leaf_cfs_rq(cfs_rq); } -static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg, - long weight_delta) +static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg) { long load_weight, load, shares; - load = cfs_rq->load.weight + weight_delta; + load = cfs_rq->load.weight; load_weight = atomic_read(&tg->load_weight); - load_weight -= cfs_rq->load_contribution; load_weight += load; + load_weight -= cfs_rq->load_contribution; shares = (tg->shares * load); if (load_weight) @@ -790,7 +790,7 @@ static void update_entity_shares_tick(struct cfs_rq *cfs_rq) { if (cfs_rq->load_unacc_exec_time > sysctl_sched_shares_window) { update_cfs_load(cfs_rq, 0); - update_cfs_shares(cfs_rq, 0); + update_cfs_shares(cfs_rq); } } # else /* CONFIG_SMP */ @@ -798,8 +798,7 @@ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) { } -static inline long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg, - long weight_delta) +static inline long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg) { return tg->shares; } @@ -824,7 +823,7 @@ static void reweight_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, account_entity_enqueue(cfs_rq, se); } -static void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta) +static void update_cfs_shares(struct cfs_rq *cfs_rq) { struct task_group *tg; struct sched_entity *se; @@ -838,7 +837,7 @@ static void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta) if (likely(se->load.weight == tg->shares)) return; #endif - shares = calc_cfs_shares(cfs_rq, tg, weight_delta); + shares = calc_cfs_shares(cfs_rq, tg); reweight_entity(cfs_rq_of(se), se, shares); } @@ -847,7 +846,7 @@ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) { } -static inline void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta) +static inline void update_cfs_shares(struct cfs_rq *cfs_rq) { } @@ -978,8 +977,8 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) */ update_curr(cfs_rq); update_cfs_load(cfs_rq, 0); - update_cfs_shares(cfs_rq, se->load.weight); account_entity_enqueue(cfs_rq, se); + update_cfs_shares(cfs_rq); if (flags & ENQUEUE_WAKEUP) { place_entity(cfs_rq, se, 0); @@ -1041,7 +1040,7 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) update_cfs_load(cfs_rq, 0); account_entity_dequeue(cfs_rq, se); update_min_vruntime(cfs_rq); - update_cfs_shares(cfs_rq, 0); + update_cfs_shares(cfs_rq); /* * Normalize the entity after updating the min_vruntime because the @@ -1282,7 +1281,7 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags) struct cfs_rq *cfs_rq = cfs_rq_of(se); update_cfs_load(cfs_rq, 0); - update_cfs_shares(cfs_rq, 0); + update_cfs_shares(cfs_rq); } hrtick_update(rq); @@ -1312,7 +1311,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) struct cfs_rq *cfs_rq = cfs_rq_of(se); update_cfs_load(cfs_rq, 0); - update_cfs_shares(cfs_rq, 0); + update_cfs_shares(cfs_rq); } hrtick_update(rq); @@ -2123,7 +2122,7 @@ static int update_shares_cpu(struct task_group *tg, int cpu) * We need to update shares after updating tg->load_weight in * order to adjust the weight of groups with long running tasks. */ - update_cfs_shares(cfs_rq, 0); + update_cfs_shares(cfs_rq); raw_spin_unlock_irqrestore(&rq->lock, flags); @@ -4079,33 +4078,62 @@ static void task_fork_fair(struct task_struct *p) * Priority of the task has changed. Check to see if we preempt * the current task. */ -static void prio_changed_fair(struct rq *rq, struct task_struct *p, - int oldprio, int running) +static void +prio_changed_fair(struct rq *rq, struct task_struct *p, int oldprio) { + if (!p->se.on_rq) + return; + /* * Reschedule if we are currently running on this runqueue and * our priority decreased, or if we are not currently running on * this runqueue and our priority is higher than the current's */ - if (running) { + if (rq->curr == p) { if (p->prio > oldprio) resched_task(rq->curr); } else check_preempt_curr(rq, p, 0); } +static void switched_from_fair(struct rq *rq, struct task_struct *p) +{ + struct sched_entity *se = &p->se; + struct cfs_rq *cfs_rq = cfs_rq_of(se); + + /* + * Ensure the task's vruntime is normalized, so that when its + * switched back to the fair class the enqueue_entity(.flags=0) will + * do the right thing. + * + * If it was on_rq, then the dequeue_entity(.flags=0) will already + * have normalized the vruntime, if it was !on_rq, then only when + * the task is sleeping will it still have non-normalized vruntime. + */ + if (!se->on_rq && p->state != TASK_RUNNING) { + /* + * Fix up our vruntime so that the current sleep doesn't + * cause 'unlimited' sleep bonus. + */ + place_entity(cfs_rq, se, 0); + se->vruntime -= cfs_rq->min_vruntime; + } +} + /* * We switched to the sched_fair class. */ -static void switched_to_fair(struct rq *rq, struct task_struct *p, - int running) +static void switched_to_fair(struct rq *rq, struct task_struct *p) { + if (!p->se.on_rq) + return; + /* * We were most likely switched from sched_rt, so * kick off the schedule if running, otherwise just see * if we can still preempt the current task. */ - if (running) + if (rq->curr == p) resched_task(rq->curr); else check_preempt_curr(rq, p, 0); @@ -4191,6 +4219,7 @@ static const struct sched_class fair_sched_class = { .task_fork = task_fork_fair, .prio_changed = prio_changed_fair, + .switched_from = switched_from_fair, .switched_to = switched_to_fair, .get_rr_interval = get_rr_interval_fair, diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 9fa0f402c87c..c82f26c1b7c3 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -52,31 +52,15 @@ static void set_curr_task_idle(struct rq *rq) { } -static void switched_to_idle(struct rq *rq, struct task_struct *p, - int running) +static void switched_to_idle(struct rq *rq, struct task_struct *p) { - /* Can this actually happen?? */ - if (running) - resched_task(rq->curr); - else - check_preempt_curr(rq, p, 0); + BUG(); } -static void prio_changed_idle(struct rq *rq, struct task_struct *p, - int oldprio, int running) +static void +prio_changed_idle(struct rq *rq, struct task_struct *p, int oldprio) { - /* This can happen for hot plug CPUS */ - - /* - * Reschedule if we are currently running on this runqueue and - * our priority decreased, or if we are not currently running on - * this runqueue and our priority is higher than the current's - */ - if (running) { - if (p->prio > oldprio) - resched_task(rq->curr); - } else - check_preempt_curr(rq, p, 0); + BUG(); } static unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index c914ec747ca6..c381fdc18c64 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1595,8 +1595,7 @@ static void rq_offline_rt(struct rq *rq) * When switch from the rt queue, we bring ourselves to a position * that we might want to pull RT tasks from other runqueues. */ -static void switched_from_rt(struct rq *rq, struct task_struct *p, - int running) +static void switched_from_rt(struct rq *rq, struct task_struct *p) { /* * If there are other RT tasks then we will reschedule @@ -1605,7 +1604,7 @@ static void switched_from_rt(struct rq *rq, struct task_struct *p, * we may need to handle the pulling of RT tasks * now. */ - if (!rq->rt.rt_nr_running) + if (p->se.on_rq && !rq->rt.rt_nr_running) pull_rt_task(rq); } @@ -1624,8 +1623,7 @@ static inline void init_sched_rt_class(void) * with RT tasks. In this case we try to push them off to * other runqueues. */ -static void switched_to_rt(struct rq *rq, struct task_struct *p, - int running) +static void switched_to_rt(struct rq *rq, struct task_struct *p) { int check_resched = 1; @@ -1636,7 +1634,7 @@ static void switched_to_rt(struct rq *rq, struct task_struct *p, * If that current running task is also an RT task * then see if we can move to another run queue. */ - if (!running) { + if (p->se.on_rq && rq->curr != p) { #ifdef CONFIG_SMP if (rq->rt.overloaded && push_rt_task(rq) && /* Don't resched if we changed runqueues */ @@ -1652,10 +1650,13 @@ static void switched_to_rt(struct rq *rq, struct task_struct *p, * Priority of the task has changed. This may cause * us to initiate a push or pull. */ -static void prio_changed_rt(struct rq *rq, struct task_struct *p, - int oldprio, int running) +static void +prio_changed_rt(struct rq *rq, struct task_struct *p, int oldprio) { - if (running) { + if (!p->se.on_rq) + return; + + if (rq->curr == p) { #ifdef CONFIG_SMP /* * If our priority decreases while running, we diff --git a/kernel/sched_stoptask.c b/kernel/sched_stoptask.c index 2bf6b47058c1..84ec9bcf82d9 100644 --- a/kernel/sched_stoptask.c +++ b/kernel/sched_stoptask.c @@ -59,14 +59,13 @@ static void set_curr_task_stop(struct rq *rq) { } -static void switched_to_stop(struct rq *rq, struct task_struct *p, - int running) +static void switched_to_stop(struct rq *rq, struct task_struct *p) { BUG(); /* its impossible to change to this class */ } -static void prio_changed_stop(struct rq *rq, struct task_struct *p, - int oldprio, int running) +static void +prio_changed_stop(struct rq *rq, struct task_struct *p, int oldprio) { BUG(); /* how!?, what priority? */ } diff --git a/kernel/softirq.c b/kernel/softirq.c index 68eb5efec388..0cee50487629 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -54,7 +54,7 @@ EXPORT_SYMBOL(irq_stat); static struct softirq_action softirq_vec[NR_SOFTIRQS] __cacheline_aligned_in_smp; -static DEFINE_PER_CPU(struct task_struct *, ksoftirqd); +DEFINE_PER_CPU(struct task_struct *, ksoftirqd); char *softirq_to_name[NR_SOFTIRQS] = { "HI", "TIMER", "NET_TX", "NET_RX", "BLOCK", "BLOCK_IOPOLL", @@ -721,7 +721,6 @@ static int run_ksoftirqd(void * __bind_cpu) { set_current_state(TASK_INTERRUPTIBLE); - current->flags |= PF_KSOFTIRQD; while (!kthread_should_stop()) { preempt_disable(); if (!local_softirq_pending()) { diff --git a/kernel/time.c b/kernel/time.c index 32174359576f..55337a816b20 100644 --- a/kernel/time.c +++ b/kernel/time.c @@ -645,7 +645,7 @@ u64 nsec_to_clock_t(u64 x) } /** - * nsecs_to_jiffies - Convert nsecs in u64 to jiffies + * nsecs_to_jiffies64 - Convert nsecs in u64 to jiffies64 * * @n: nsecs in u64 * @@ -657,7 +657,7 @@ u64 nsec_to_clock_t(u64 x) * NSEC_PER_SEC = 10^9 = (5^9 * 2^9) = (1953125 * 512) * ULLONG_MAX ns = 18446744073.709551615 secs = about 584 years */ -unsigned long nsecs_to_jiffies(u64 n) +u64 nsecs_to_jiffies64(u64 n) { #if (NSEC_PER_SEC % HZ) == 0 /* Common case, HZ = 100, 128, 200, 250, 256, 500, 512, 1000 etc. */ @@ -674,6 +674,25 @@ unsigned long nsecs_to_jiffies(u64 n) #endif } + +/** + * nsecs_to_jiffies - Convert nsecs in u64 to jiffies + * + * @n: nsecs in u64 + * + * Unlike {m,u}secs_to_jiffies, type of input is not unsigned int but u64. + * And this doesn't return MAX_JIFFY_OFFSET since this function is designed + * for scheduler, not for use in device drivers to calculate timeout value. + * + * note: + * NSEC_PER_SEC = 10^9 = (5^9 * 2^9) = (1953125 * 512) + * ULLONG_MAX ns = 18446744073.709551615 secs = about 584 years + */ +unsigned long nsecs_to_jiffies(u64 n) +{ + return (unsigned long)nsecs_to_jiffies64(n); +} + #if (BITS_PER_LONG < 64) u64 get_jiffies_64(void) { diff --git a/lib/rwsem.c b/lib/rwsem.c index f236d7cd5cf3..aa7c3052261f 100644 --- a/lib/rwsem.c +++ b/lib/rwsem.c @@ -222,8 +222,7 @@ rwsem_down_failed_common(struct rw_semaphore *sem, /* * wait for the read lock to be granted */ -asmregparm struct rw_semaphore __sched * -rwsem_down_read_failed(struct rw_semaphore *sem) +struct rw_semaphore __sched *rwsem_down_read_failed(struct rw_semaphore *sem) { return rwsem_down_failed_common(sem, RWSEM_WAITING_FOR_READ, -RWSEM_ACTIVE_READ_BIAS); @@ -232,8 +231,7 @@ rwsem_down_read_failed(struct rw_semaphore *sem) /* * wait for the write lock to be granted */ -asmregparm struct rw_semaphore __sched * -rwsem_down_write_failed(struct rw_semaphore *sem) +struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem) { return rwsem_down_failed_common(sem, RWSEM_WAITING_FOR_WRITE, -RWSEM_ACTIVE_WRITE_BIAS); @@ -243,7 +241,7 @@ rwsem_down_write_failed(struct rw_semaphore *sem) * handle waking up a waiter on the semaphore * - up_read/up_write has decremented the active part of count if we come here */ -asmregparm struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem) +struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem) { unsigned long flags; @@ -263,7 +261,7 @@ asmregparm struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem) * - caller incremented waiting part of count and discovered it still negative * - just wake up any readers at the front of the queue */ -asmregparm struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem) +struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem) { unsigned long flags; diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl index 4c0383da1c9a..58848e3e392c 100755 --- a/scripts/checkpatch.pl +++ b/scripts/checkpatch.pl @@ -2654,11 +2654,6 @@ sub process { WARN("Use of volatile is usually wrong: see Documentation/volatile-considered-harmful.txt\n" . $herecurr); } -# SPIN_LOCK_UNLOCKED & RW_LOCK_UNLOCKED are deprecated - if ($line =~ /\b(SPIN_LOCK_UNLOCKED|RW_LOCK_UNLOCKED)/) { - ERROR("Use of $1 is deprecated: see Documentation/spinlocks.txt\n" . $herecurr); - } - # warn about #if 0 if ($line =~ /^.\s*\#\s*if\s+0\b/) { CHK("if this code is redundant consider removing it\n" . diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt index 86b797a35aa6..fcc51fe0195c 100644 --- a/tools/perf/Documentation/perf-probe.txt +++ b/tools/perf/Documentation/perf-probe.txt @@ -73,6 +73,10 @@ OPTIONS (Only for --vars) Show external defined variables in addition to local variables. +-F:: +--funcs:: + Show available functions in given module or kernel. + -f:: --force:: Forcibly add events with existing name. diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 7141c42e1469..638e8e146bb9 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -402,6 +402,7 @@ LIB_H += util/debug.h LIB_H += util/debugfs.h LIB_H += util/event.h LIB_H += util/evsel.h +LIB_H += util/evlist.h LIB_H += util/exec_cmd.h LIB_H += util/types.h LIB_H += util/levenshtein.h @@ -425,6 +426,7 @@ LIB_H += util/values.h LIB_H += util/sort.h LIB_H += util/hist.h LIB_H += util/thread.h +LIB_H += util/thread_map.h LIB_H += util/trace-event.h LIB_H += util/probe-finder.h LIB_H += util/probe-event.h @@ -440,6 +442,7 @@ LIB_OBJS += $(OUTPUT)util/ctype.o LIB_OBJS += $(OUTPUT)util/debugfs.o LIB_OBJS += $(OUTPUT)util/environment.o LIB_OBJS += $(OUTPUT)util/event.o +LIB_OBJS += $(OUTPUT)util/evlist.o LIB_OBJS += $(OUTPUT)util/evsel.o LIB_OBJS += $(OUTPUT)util/exec_cmd.o LIB_OBJS += $(OUTPUT)util/help.o @@ -469,6 +472,7 @@ LIB_OBJS += $(OUTPUT)util/map.o LIB_OBJS += $(OUTPUT)util/pstack.o LIB_OBJS += $(OUTPUT)util/session.o LIB_OBJS += $(OUTPUT)util/thread.o +LIB_OBJS += $(OUTPUT)util/thread_map.o LIB_OBJS += $(OUTPUT)util/trace-event-parse.o LIB_OBJS += $(OUTPUT)util/trace-event-read.o LIB_OBJS += $(OUTPUT)util/trace-event-info.o diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index add163c9f0e7..6cf708aba7c9 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -52,6 +52,7 @@ static struct { bool show_lines; bool show_vars; bool show_ext_vars; + bool show_funcs; bool mod_events; int nevents; struct perf_probe_event events[MAX_PROBES]; @@ -221,6 +222,8 @@ static const struct option options[] = { OPT__DRY_RUN(&probe_event_dry_run), OPT_INTEGER('\0', "max-probes", ¶ms.max_probe_points, "Set how many probe points can be found for a probe."), + OPT_BOOLEAN('F', "funcs", ¶ms.show_funcs, + "Show potential probe-able functions."), OPT_END() }; @@ -246,7 +249,7 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) params.max_probe_points = MAX_PROBES; if ((!params.nevents && !params.dellist && !params.list_events && - !params.show_lines)) + !params.show_lines && !params.show_funcs)) usage_with_options(probe_usage, options); /* @@ -267,12 +270,36 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) pr_err(" Error: Don't use --list with --vars.\n"); usage_with_options(probe_usage, options); } + if (params.show_funcs) { + pr_err(" Error: Don't use --list with --funcs.\n"); + usage_with_options(probe_usage, options); + } ret = show_perf_probe_events(); if (ret < 0) pr_err(" Error: Failed to show event list. (%d)\n", ret); return ret; } + if (params.show_funcs) { + if (params.nevents != 0 || params.dellist) { + pr_err(" Error: Don't use --funcs with" + " --add/--del.\n"); + usage_with_options(probe_usage, options); + } + if (params.show_lines) { + pr_err(" Error: Don't use --funcs with --line.\n"); + usage_with_options(probe_usage, options); + } + if (params.show_vars) { + pr_err(" Error: Don't use --funcs with --vars.\n"); + usage_with_options(probe_usage, options); + } + ret = show_available_funcs(params.target_module); + if (ret < 0) + pr_err(" Error: Failed to show functions." + " (%d)\n", ret); + return ret; + } #ifdef DWARF_SUPPORT if (params.show_lines) { diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index b2f729fdb317..d7886307f6f4 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -18,17 +18,20 @@ #include "util/header.h" #include "util/event.h" +#include "util/evlist.h" #include "util/evsel.h" #include "util/debug.h" #include "util/session.h" #include "util/symbol.h" #include "util/cpumap.h" +#include "util/thread_map.h" #include <unistd.h> #include <sched.h> #include <sys/mman.h> #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y)) +#define SID(e, x, y) xyarray__entry(e->id, x, y) enum write_mode_t { WRITE_FORCE, @@ -46,7 +49,7 @@ static unsigned int user_freq = UINT_MAX; static int freq = 1000; static int output; static int pipe_output = 0; -static const char *output_name = "perf.data"; +static const char *output_name = NULL; static int group = 0; static int realtime_prio = 0; static bool nodelay = false; @@ -66,51 +69,17 @@ static bool sample_address = false; static bool sample_time = false; static bool no_buildid = false; static bool no_buildid_cache = false; +static struct perf_evlist *evsel_list; static long samples = 0; static u64 bytes_written = 0; -static struct pollfd *event_array; - -static int nr_poll = 0; -static int nr_cpu = 0; - static int file_new = 1; static off_t post_processing_offset; static struct perf_session *session; static const char *cpu_list; -struct mmap_data { - void *base; - unsigned int mask; - unsigned int prev; -}; - -static struct mmap_data mmap_array[MAX_NR_CPUS]; - -static unsigned long mmap_read_head(struct mmap_data *md) -{ - struct perf_event_mmap_page *pc = md->base; - long head; - - head = pc->data_head; - rmb(); - - return head; -} - -static void mmap_write_tail(struct mmap_data *md, unsigned long tail) -{ - struct perf_event_mmap_page *pc = md->base; - - /* - * ensure all reads are done before we write the tail out. - */ - /* mb(); */ - pc->data_tail = tail; -} - static void advance_output(size_t size) { bytes_written += size; @@ -139,9 +108,9 @@ static int process_synthesized_event(event_t *event, return 0; } -static void mmap_read(struct mmap_data *md) +static void mmap_read(struct perf_mmap *md) { - unsigned int head = mmap_read_head(md); + unsigned int head = perf_mmap__read_head(md); unsigned int old = md->prev; unsigned char *data = md->base + page_size; unsigned long size; @@ -185,7 +154,7 @@ static void mmap_read(struct mmap_data *md) write_output(buf, size); md->prev = old; - mmap_write_tail(md, old); + perf_mmap__write_tail(md, old); } static volatile int done = 0; @@ -209,8 +178,6 @@ static void sig_atexit(void) kill(getpid(), signr); } -static int group_fd; - static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int nr) { struct perf_header_attr *h_attr; @@ -234,28 +201,47 @@ static void create_counter(struct perf_evsel *evsel, int cpu) char *filter = evsel->filter; struct perf_event_attr *attr = &evsel->attr; struct perf_header_attr *h_attr; - int track = !evsel->idx; /* only the first counter needs these */ + struct perf_sample_id *sid; int thread_index; int ret; - struct { - u64 count; - u64 time_enabled; - u64 time_running; - u64 id; - } read_data; - /* - * Check if parse_single_tracepoint_event has already asked for - * PERF_SAMPLE_TIME. - * - * XXX this is kludgy but short term fix for problems introduced by - * eac23d1c that broke 'perf script' by having different sample_types - * when using multiple tracepoint events when we use a perf binary - * that tries to use sample_id_all on an older kernel. - * - * We need to move counter creation to perf_session, support - * different sample_types, etc. - */ - bool time_needed = attr->sample_type & PERF_SAMPLE_TIME; + + for (thread_index = 0; thread_index < threads->nr; thread_index++) { + h_attr = get_header_attr(attr, evsel->idx); + if (h_attr == NULL) + die("nomem\n"); + + if (!file_new) { + if (memcmp(&h_attr->attr, attr, sizeof(*attr))) { + fprintf(stderr, "incompatible append\n"); + exit(-1); + } + } + + sid = SID(evsel, cpu, thread_index); + if (perf_header_attr__add_id(h_attr, sid->id) < 0) { + pr_warning("Not enough memory to add id\n"); + exit(-1); + } + + if (filter != NULL) { + ret = ioctl(FD(evsel, cpu, thread_index), + PERF_EVENT_IOC_SET_FILTER, filter); + if (ret) { + error("failed to set filter with %d (%s)\n", errno, + strerror(errno)); + exit(-1); + } + } + } + + if (!sample_type) + sample_type = attr->sample_type; +} + +static void config_attr(struct perf_evsel *evsel, struct perf_evlist *evlist) +{ + struct perf_event_attr *attr = &evsel->attr; + int track = !evsel->idx; /* only the first counter needs these */ attr->read_format = PERF_FORMAT_TOTAL_TIME_ENABLED | PERF_FORMAT_TOTAL_TIME_RUNNING | @@ -263,7 +249,7 @@ static void create_counter(struct perf_evsel *evsel, int cpu) attr->sample_type |= PERF_SAMPLE_IP | PERF_SAMPLE_TID; - if (nr_counters > 1) + if (evlist->nr_entries > 1) attr->sample_type |= PERF_SAMPLE_ID; /* @@ -315,19 +301,39 @@ static void create_counter(struct perf_evsel *evsel, int cpu) attr->mmap = track; attr->comm = track; - attr->inherit = !no_inherit; + if (target_pid == -1 && target_tid == -1 && !system_wide) { attr->disabled = 1; attr->enable_on_exec = 1; } -retry_sample_id: - attr->sample_id_all = sample_id_all_avail ? 1 : 0; +} - for (thread_index = 0; thread_index < threads->nr; thread_index++) { -try_again: - FD(evsel, nr_cpu, thread_index) = sys_perf_event_open(attr, threads->map[thread_index], cpu, group_fd, 0); +static void open_counters(struct perf_evlist *evlist) +{ + struct perf_evsel *pos; + int cpu; - if (FD(evsel, nr_cpu, thread_index) < 0) { + list_for_each_entry(pos, &evlist->entries, node) { + struct perf_event_attr *attr = &pos->attr; + /* + * Check if parse_single_tracepoint_event has already asked for + * PERF_SAMPLE_TIME. + * + * XXX this is kludgy but short term fix for problems introduced by + * eac23d1c that broke 'perf script' by having different sample_types + * when using multiple tracepoint events when we use a perf binary + * that tries to use sample_id_all on an older kernel. + * + * We need to move counter creation to perf_session, support + * different sample_types, etc. + */ + bool time_needed = attr->sample_type & PERF_SAMPLE_TIME; + + config_attr(pos, evlist); +retry_sample_id: + attr->sample_id_all = sample_id_all_avail ? 1 : 0; +try_again: + if (perf_evsel__open(pos, cpus, threads, group, !no_inherit) < 0) { int err = errno; if (err == EPERM || err == EACCES) @@ -364,7 +370,7 @@ try_again: } printf("\n"); error("sys_perf_event_open() syscall returned with %d (%s). /bin/dmesg may provide additional information.\n", - FD(evsel, nr_cpu, thread_index), strerror(err)); + err, strerror(err)); #if defined(__i386__) || defined(__x86_64__) if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP) @@ -375,90 +381,16 @@ try_again: #endif die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); - exit(-1); - } - - h_attr = get_header_attr(attr, evsel->idx); - if (h_attr == NULL) - die("nomem\n"); - - if (!file_new) { - if (memcmp(&h_attr->attr, attr, sizeof(*attr))) { - fprintf(stderr, "incompatible append\n"); - exit(-1); - } - } - - if (read(FD(evsel, nr_cpu, thread_index), &read_data, sizeof(read_data)) == -1) { - perror("Unable to read perf file descriptor"); - exit(-1); - } - - if (perf_header_attr__add_id(h_attr, read_data.id) < 0) { - pr_warning("Not enough memory to add id\n"); - exit(-1); - } - - assert(FD(evsel, nr_cpu, thread_index) >= 0); - fcntl(FD(evsel, nr_cpu, thread_index), F_SETFL, O_NONBLOCK); - - /* - * First counter acts as the group leader: - */ - if (group && group_fd == -1) - group_fd = FD(evsel, nr_cpu, thread_index); - - if (evsel->idx || thread_index) { - struct perf_evsel *first; - first = list_entry(evsel_list.next, struct perf_evsel, node); - ret = ioctl(FD(evsel, nr_cpu, thread_index), - PERF_EVENT_IOC_SET_OUTPUT, - FD(first, nr_cpu, 0)); - if (ret) { - error("failed to set output: %d (%s)\n", errno, - strerror(errno)); - exit(-1); - } - } else { - mmap_array[nr_cpu].prev = 0; - mmap_array[nr_cpu].mask = mmap_pages*page_size - 1; - mmap_array[nr_cpu].base = mmap(NULL, (mmap_pages+1)*page_size, - PROT_READ | PROT_WRITE, MAP_SHARED, FD(evsel, nr_cpu, thread_index), 0); - if (mmap_array[nr_cpu].base == MAP_FAILED) { - error("failed to mmap with %d (%s)\n", errno, strerror(errno)); - exit(-1); - } - - event_array[nr_poll].fd = FD(evsel, nr_cpu, thread_index); - event_array[nr_poll].events = POLLIN; - nr_poll++; - } - - if (filter != NULL) { - ret = ioctl(FD(evsel, nr_cpu, thread_index), - PERF_EVENT_IOC_SET_FILTER, filter); - if (ret) { - error("failed to set filter with %d (%s)\n", errno, - strerror(errno)); - exit(-1); - } } } - if (!sample_type) - sample_type = attr->sample_type; -} - -static void open_counters(int cpu) -{ - struct perf_evsel *pos; - - group_fd = -1; - - list_for_each_entry(pos, &evsel_list, node) - create_counter(pos, cpu); + if (perf_evlist__mmap(evlist, cpus, threads, mmap_pages, false) < 0) + die("failed to mmap with %d (%s)\n", errno, strerror(errno)); - nr_cpu++; + for (cpu = 0; cpu < cpus->nr; ++cpu) { + list_for_each_entry(pos, &evlist->entries, node) + create_counter(pos, cpu); + } } static int process_buildids(void) @@ -481,9 +413,9 @@ static void atexit_header(void) if (!no_buildid) process_buildids(); - perf_header__write(&session->header, output, true); + perf_header__write(&session->header, evsel_list, output, true); perf_session__delete(session); - perf_evsel_list__delete(); + perf_evlist__delete(evsel_list); symbol__exit(); } } @@ -533,9 +465,9 @@ static void mmap_read_all(void) { int i; - for (i = 0; i < nr_cpu; i++) { - if (mmap_array[i].base) - mmap_read(&mmap_array[i]); + for (i = 0; i < cpus->nr; i++) { + if (evsel_list->mmap[i].base) + mmap_read(&evsel_list->mmap[i]); } if (perf_header__has_feat(&session->header, HEADER_TRACE_INFO)) @@ -566,18 +498,26 @@ static int __cmd_record(int argc, const char **argv) exit(-1); } - if (!strcmp(output_name, "-")) - pipe_output = 1; - else if (!stat(output_name, &st) && st.st_size) { - if (write_mode == WRITE_FORCE) { - char oldname[PATH_MAX]; - snprintf(oldname, sizeof(oldname), "%s.old", - output_name); - unlink(oldname); - rename(output_name, oldname); + if (!output_name) { + if (!fstat(STDOUT_FILENO, &st) && S_ISFIFO(st.st_mode)) + pipe_output = 1; + else + output_name = "perf.data"; + } + if (output_name) { + if (!strcmp(output_name, "-")) + pipe_output = 1; + else if (!stat(output_name, &st) && st.st_size) { + if (write_mode == WRITE_FORCE) { + char oldname[PATH_MAX]; + snprintf(oldname, sizeof(oldname), "%s.old", + output_name); + unlink(oldname); + rename(output_name, oldname); + } + } else if (write_mode == WRITE_APPEND) { + write_mode = WRITE_FORCE; } - } else if (write_mode == WRITE_APPEND) { - write_mode = WRITE_FORCE; } flags = O_CREAT|O_RDWR; @@ -611,7 +551,7 @@ static int __cmd_record(int argc, const char **argv) goto out_delete_session; } - if (have_tracepoints(&evsel_list)) + if (have_tracepoints(&evsel_list->entries)) perf_header__set_feat(&session->header, HEADER_TRACE_INFO); /* @@ -673,12 +613,7 @@ static int __cmd_record(int argc, const char **argv) close(child_ready_pipe[0]); } - if (!system_wide && no_inherit && !cpu_list) { - open_counters(-1); - } else { - for (i = 0; i < cpus->nr; i++) - open_counters(cpus->map[i]); - } + open_counters(evsel_list); perf_session__set_sample_type(session, sample_type); @@ -687,7 +622,8 @@ static int __cmd_record(int argc, const char **argv) if (err < 0) return err; } else if (file_new) { - err = perf_header__write(&session->header, output, false); + err = perf_header__write(&session->header, evsel_list, + output, false); if (err < 0) return err; } @@ -712,7 +648,7 @@ static int __cmd_record(int argc, const char **argv) return err; } - if (have_tracepoints(&evsel_list)) { + if (have_tracepoints(&evsel_list->entries)) { /* * FIXME err <= 0 here actually means that * there were no tracepoints so its not really @@ -721,7 +657,7 @@ static int __cmd_record(int argc, const char **argv) * return this more properly and also * propagate errors that now are calling die() */ - err = event__synthesize_tracing_data(output, &evsel_list, + err = event__synthesize_tracing_data(output, evsel_list, process_synthesized_event, session); if (err <= 0) { @@ -789,15 +725,15 @@ static int __cmd_record(int argc, const char **argv) if (hits == samples) { if (done) break; - err = poll(event_array, nr_poll, -1); + err = poll(evsel_list->pollfd, evsel_list->nr_fds, -1); waking++; } if (done) { - for (i = 0; i < nr_cpu; i++) { + for (i = 0; i < cpus->nr; i++) { struct perf_evsel *pos; - list_for_each_entry(pos, &evsel_list, node) { + list_for_each_entry(pos, &evsel_list->entries, node) { for (thread = 0; thread < threads->nr; thread++) @@ -838,10 +774,10 @@ static const char * const record_usage[] = { static bool force, append_file; const struct option record_options[] = { - OPT_CALLBACK('e', "event", NULL, "event", + OPT_CALLBACK('e', "event", &evsel_list, "event", "event selector. use 'perf list' to list available events", parse_events), - OPT_CALLBACK(0, "filter", NULL, "filter", + OPT_CALLBACK(0, "filter", &evsel_list, "filter", "event filter", parse_filter), OPT_INTEGER('p', "pid", &target_pid, "record events on existing process id"), @@ -892,6 +828,10 @@ int cmd_record(int argc, const char **argv, const char *prefix __used) int err = -ENOMEM; struct perf_evsel *pos; + evsel_list = perf_evlist__new(); + if (evsel_list == NULL) + return -ENOMEM; + argc = parse_options(argc, argv, record_options, record_usage, PARSE_OPT_STOP_AT_NON_OPTION); if (!argc && target_pid == -1 && target_tid == -1 && @@ -913,7 +853,8 @@ int cmd_record(int argc, const char **argv, const char *prefix __used) if (no_buildid_cache || no_buildid) disable_buildid_cache(); - if (list_empty(&evsel_list) && perf_evsel_list__create_default() < 0) { + if (evsel_list->nr_entries == 0 && + perf_evlist__add_default(evsel_list) < 0) { pr_err("Not enough memory for event selector list\n"); goto out_symbol_exit; } @@ -927,21 +868,22 @@ int cmd_record(int argc, const char **argv, const char *prefix __used) usage_with_options(record_usage, record_options); } - cpus = cpu_map__new(cpu_list); - if (cpus == NULL) { - perror("failed to parse CPUs map"); - return -1; - } + if (target_tid != -1) + cpus = cpu_map__dummy_new(); + else + cpus = cpu_map__new(cpu_list); - list_for_each_entry(pos, &evsel_list, node) { + if (cpus == NULL) + usage_with_options(record_usage, record_options); + + list_for_each_entry(pos, &evsel_list->entries, node) { if (perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0) goto out_free_fd; if (perf_header__push_event(pos->attr.config, event_name(pos))) goto out_free_fd; } - event_array = malloc((sizeof(struct pollfd) * MAX_NR_CPUS * - MAX_COUNTERS * threads->nr)); - if (!event_array) + + if (perf_evlist__alloc_pollfd(evsel_list, cpus->nr, threads->nr) < 0) goto out_free_fd; if (user_interval != ULLONG_MAX) @@ -959,13 +901,11 @@ int cmd_record(int argc, const char **argv, const char *prefix __used) } else { fprintf(stderr, "frequency and count are zero, aborting\n"); err = -EINVAL; - goto out_free_event_array; + goto out_free_fd; } err = __cmd_record(argc, argv); -out_free_event_array: - free(event_array); out_free_fd: thread_map__delete(threads); threads = NULL; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index c27e31f289e6..f6a43493d1d0 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -81,18 +81,17 @@ static int perf_session__add_hist_entry(struct perf_session *self, struct addr_location *al, struct sample_data *data) { - struct map_symbol *syms = NULL; struct symbol *parent = NULL; - int err = -ENOMEM; + int err = 0; struct hist_entry *he; struct hists *hists; struct perf_event_attr *attr; if ((sort__has_parent || symbol_conf.use_callchain) && data->callchain) { - syms = perf_session__resolve_callchain(self, al->thread, - data->callchain, &parent); - if (syms == NULL) - return -ENOMEM; + err = perf_session__resolve_callchain(self, al->thread, + data->callchain, &parent); + if (err) + return err; } attr = perf_header__find_attr(data->id, &self->header); @@ -101,16 +100,17 @@ static int perf_session__add_hist_entry(struct perf_session *self, else hists = perf_session__hists_findnew(self, data->id, 0, 0); if (hists == NULL) - goto out_free_syms; + return -ENOMEM; + he = __hists__add_entry(hists, al, parent, data->period); if (he == NULL) - goto out_free_syms; - err = 0; + return -ENOMEM; + if (symbol_conf.use_callchain) { - err = callchain_append(he->callchain, data->callchain, syms, + err = callchain_append(he->callchain, &self->callchain_cursor, data->period); if (err) - goto out_free_syms; + return err; } /* * Only in the newt browser we are doing integrated annotation, @@ -119,8 +119,7 @@ static int perf_session__add_hist_entry(struct perf_session *self, */ if (use_browser > 0) err = hist_entry__inc_addr_samples(he, al->addr); -out_free_syms: - free(syms); + return err; } @@ -222,7 +221,7 @@ static int perf_session__setup_sample_type(struct perf_session *self) } else if (!dont_use_callchains && callchain_param.mode != CHAIN_NONE && !symbol_conf.use_callchain) { symbol_conf.use_callchain = true; - if (register_callchain_param(&callchain_param) < 0) { + if (callchain_register_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain" " params\n"); return -EINVAL; @@ -424,7 +423,7 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, if (tok2) callchain_param.print_limit = strtod(tok2, &endptr); setup: - if (register_callchain_param(&callchain_param) < 0) { + if (callchain_register_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain params\n"); return -1; } diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index a482a191a0ca..8906adfdbd8e 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -43,11 +43,13 @@ #include "util/parse-options.h" #include "util/parse-events.h" #include "util/event.h" +#include "util/evlist.h" #include "util/evsel.h" #include "util/debug.h" #include "util/header.h" #include "util/cpumap.h" #include "util/thread.h" +#include "util/thread_map.h" #include <sys/prctl.h> #include <math.h> @@ -71,6 +73,8 @@ static struct perf_event_attr default_attrs[] = { }; +struct perf_evlist *evsel_list; + static bool system_wide = false; static struct cpu_map *cpus; static int run_idx = 0; @@ -166,7 +170,7 @@ static int create_perf_stat_counter(struct perf_evsel *evsel) PERF_FORMAT_TOTAL_TIME_RUNNING; if (system_wide) - return perf_evsel__open_per_cpu(evsel, cpus); + return perf_evsel__open_per_cpu(evsel, cpus, false, false); attr->inherit = !no_inherit; if (target_pid == -1 && target_tid == -1) { @@ -174,7 +178,7 @@ static int create_perf_stat_counter(struct perf_evsel *evsel) attr->enable_on_exec = 1; } - return perf_evsel__open_per_thread(evsel, threads); + return perf_evsel__open_per_thread(evsel, threads, false, false); } /* @@ -309,7 +313,7 @@ static int run_perf_stat(int argc __used, const char **argv) close(child_ready_pipe[0]); } - list_for_each_entry(counter, &evsel_list, node) { + list_for_each_entry(counter, &evsel_list->entries, node) { if (create_perf_stat_counter(counter) < 0) { if (errno == -EPERM || errno == -EACCES) { error("You may not have permission to collect %sstats.\n" @@ -347,12 +351,12 @@ static int run_perf_stat(int argc __used, const char **argv) update_stats(&walltime_nsecs_stats, t1 - t0); if (no_aggr) { - list_for_each_entry(counter, &evsel_list, node) { + list_for_each_entry(counter, &evsel_list->entries, node) { read_counter(counter); perf_evsel__close_fd(counter, cpus->nr, 1); } } else { - list_for_each_entry(counter, &evsel_list, node) { + list_for_each_entry(counter, &evsel_list->entries, node) { read_counter_aggr(counter); perf_evsel__close_fd(counter, cpus->nr, threads->nr); } @@ -555,10 +559,10 @@ static void print_stat(int argc, const char **argv) } if (no_aggr) { - list_for_each_entry(counter, &evsel_list, node) + list_for_each_entry(counter, &evsel_list->entries, node) print_counter(counter); } else { - list_for_each_entry(counter, &evsel_list, node) + list_for_each_entry(counter, &evsel_list->entries, node) print_counter_aggr(counter); } @@ -610,7 +614,7 @@ static int stat__set_big_num(const struct option *opt __used, } static const struct option options[] = { - OPT_CALLBACK('e', "event", NULL, "event", + OPT_CALLBACK('e', "event", &evsel_list, "event", "event selector. use 'perf list' to list available events", parse_events), OPT_BOOLEAN('i', "no-inherit", &no_inherit, @@ -648,6 +652,10 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) setlocale(LC_ALL, ""); + evsel_list = perf_evlist__new(); + if (evsel_list == NULL) + return -ENOMEM; + argc = parse_options(argc, argv, options, stat_usage, PARSE_OPT_STOP_AT_NON_OPTION); @@ -679,17 +687,14 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) usage_with_options(stat_usage, options); /* Set attrs and nr_counters if no event is selected and !null_run */ - if (!null_run && !nr_counters) { + if (!null_run && !evsel_list->nr_entries) { size_t c; - nr_counters = ARRAY_SIZE(default_attrs); - for (c = 0; c < ARRAY_SIZE(default_attrs); ++c) { - pos = perf_evsel__new(&default_attrs[c], - nr_counters); + pos = perf_evsel__new(&default_attrs[c], c); if (pos == NULL) goto out; - list_add(&pos->node, &evsel_list); + perf_evlist__add(evsel_list, pos); } } @@ -713,7 +718,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) return -1; } - list_for_each_entry(pos, &evsel_list, node) { + list_for_each_entry(pos, &evsel_list->entries, node) { if (perf_evsel__alloc_stat_priv(pos) < 0 || perf_evsel__alloc_counts(pos, cpus->nr) < 0 || perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0) @@ -741,9 +746,9 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) if (status != -1) print_stat(argc, argv); out_free_fd: - list_for_each_entry(pos, &evsel_list, node) + list_for_each_entry(pos, &evsel_list->entries, node) perf_evsel__free_stat_priv(pos); - perf_evsel_list__delete(); + perf_evlist__delete(evsel_list); out: thread_map__delete(threads); threads = NULL; diff --git a/tools/perf/builtin-test.c b/tools/perf/builtin-test.c index 5dcdba653d70..231e3e21810c 100644 --- a/tools/perf/builtin-test.c +++ b/tools/perf/builtin-test.c @@ -7,10 +7,11 @@ #include "util/cache.h" #include "util/debug.h" +#include "util/evlist.h" #include "util/parse-options.h" -#include "util/session.h" +#include "util/parse-events.h" #include "util/symbol.h" -#include "util/thread.h" +#include "util/thread_map.h" static long page_size; @@ -238,14 +239,14 @@ out: #include "util/evsel.h" #include <sys/types.h> -static int trace_event__id(const char *event_name) +static int trace_event__id(const char *evname) { char *filename; int err = -1, fd; if (asprintf(&filename, "/sys/kernel/debug/tracing/events/syscalls/%s/id", - event_name) < 0) + evname) < 0) return -1; fd = open(filename, O_RDONLY); @@ -289,7 +290,7 @@ static int test__open_syscall_event(void) goto out_thread_map_delete; } - if (perf_evsel__open_per_thread(evsel, threads) < 0) { + if (perf_evsel__open_per_thread(evsel, threads, false, false) < 0) { pr_debug("failed to open counter: %s, " "tweak /proc/sys/kernel/perf_event_paranoid?\n", strerror(errno)); @@ -347,9 +348,9 @@ static int test__open_syscall_event_on_all_cpus(void) } cpus = cpu_map__new(NULL); - if (threads == NULL) { - pr_debug("thread_map__new\n"); - return -1; + if (cpus == NULL) { + pr_debug("cpu_map__new\n"); + goto out_thread_map_delete; } @@ -364,7 +365,7 @@ static int test__open_syscall_event_on_all_cpus(void) goto out_thread_map_delete; } - if (perf_evsel__open(evsel, cpus, threads) < 0) { + if (perf_evsel__open(evsel, cpus, threads, false, false) < 0) { pr_debug("failed to open counter: %s, " "tweak /proc/sys/kernel/perf_event_paranoid?\n", strerror(errno)); @@ -408,6 +409,8 @@ static int test__open_syscall_event_on_all_cpus(void) goto out_close_fd; } + err = 0; + for (cpu = 0; cpu < cpus->nr; ++cpu) { unsigned int expected; @@ -416,18 +419,18 @@ static int test__open_syscall_event_on_all_cpus(void) if (perf_evsel__read_on_cpu(evsel, cpu, 0) < 0) { pr_debug("perf_evsel__open_read_on_cpu\n"); - goto out_close_fd; + err = -1; + break; } expected = nr_open_calls + cpu; if (evsel->counts->cpu[cpu].val != expected) { pr_debug("perf_evsel__read_on_cpu: expected to intercept %d calls on cpu %d, got %" PRIu64 "\n", expected, cpus->map[cpu], evsel->counts->cpu[cpu].val); - goto out_close_fd; + err = -1; } } - err = 0; out_close_fd: perf_evsel__close_fd(evsel, 1, threads->nr); out_evsel_delete: @@ -437,6 +440,159 @@ out_thread_map_delete: return err; } +/* + * This test will generate random numbers of calls to some getpid syscalls, + * then establish an mmap for a group of events that are created to monitor + * the syscalls. + * + * It will receive the events, using mmap, use its PERF_SAMPLE_ID generated + * sample.id field to map back to its respective perf_evsel instance. + * + * Then it checks if the number of syscalls reported as perf events by + * the kernel corresponds to the number of syscalls made. + */ +static int test__basic_mmap(void) +{ + int err = -1; + event_t *event; + struct thread_map *threads; + struct cpu_map *cpus; + struct perf_evlist *evlist; + struct perf_event_attr attr = { + .type = PERF_TYPE_TRACEPOINT, + .read_format = PERF_FORMAT_ID, + .sample_type = PERF_SAMPLE_ID, + .watermark = 0, + }; + cpu_set_t cpu_set; + const char *syscall_names[] = { "getsid", "getppid", "getpgrp", + "getpgid", }; + pid_t (*syscalls[])(void) = { (void *)getsid, getppid, getpgrp, + (void*)getpgid }; +#define nsyscalls ARRAY_SIZE(syscall_names) + int ids[nsyscalls]; + unsigned int nr_events[nsyscalls], + expected_nr_events[nsyscalls], i, j; + struct perf_evsel *evsels[nsyscalls], *evsel; + + for (i = 0; i < nsyscalls; ++i) { + char name[64]; + + snprintf(name, sizeof(name), "sys_enter_%s", syscall_names[i]); + ids[i] = trace_event__id(name); + if (ids[i] < 0) { + pr_debug("Is debugfs mounted on /sys/kernel/debug?\n"); + return -1; + } + nr_events[i] = 0; + expected_nr_events[i] = random() % 257; + } + + threads = thread_map__new(-1, getpid()); + if (threads == NULL) { + pr_debug("thread_map__new\n"); + return -1; + } + + cpus = cpu_map__new(NULL); + if (threads == NULL) { + pr_debug("thread_map__new\n"); + goto out_free_threads; + } + + CPU_ZERO(&cpu_set); + CPU_SET(cpus->map[0], &cpu_set); + sched_setaffinity(0, sizeof(cpu_set), &cpu_set); + if (sched_setaffinity(0, sizeof(cpu_set), &cpu_set) < 0) { + pr_debug("sched_setaffinity() failed on CPU %d: %s ", + cpus->map[0], strerror(errno)); + goto out_free_cpus; + } + + evlist = perf_evlist__new(); + if (threads == NULL) { + pr_debug("perf_evlist__new\n"); + goto out_free_cpus; + } + + /* anonymous union fields, can't be initialized above */ + attr.wakeup_events = 1; + attr.sample_period = 1; + + for (i = 0; i < nsyscalls; ++i) { + attr.config = ids[i]; + evsels[i] = perf_evsel__new(&attr, i); + if (evsels[i] == NULL) { + pr_debug("perf_evsel__new\n"); + goto out_free_evlist; + } + + perf_evlist__add(evlist, evsels[i]); + + if (perf_evsel__open(evsels[i], cpus, threads, false, false) < 0) { + pr_debug("failed to open counter: %s, " + "tweak /proc/sys/kernel/perf_event_paranoid?\n", + strerror(errno)); + goto out_close_fd; + } + } + + if (perf_evlist__mmap(evlist, cpus, threads, 128, true) < 0) { + pr_debug("failed to mmap events: %d (%s)\n", errno, + strerror(errno)); + goto out_close_fd; + } + + for (i = 0; i < nsyscalls; ++i) + for (j = 0; j < expected_nr_events[i]; ++j) { + int foo = syscalls[i](); + ++foo; + } + + while ((event = perf_evlist__read_on_cpu(evlist, 0)) != NULL) { + struct sample_data sample; + + if (event->header.type != PERF_RECORD_SAMPLE) { + pr_debug("unexpected %s event\n", + event__get_event_name(event->header.type)); + goto out_munmap; + } + + event__parse_sample(event, attr.sample_type, false, &sample); + evsel = perf_evlist__id2evsel(evlist, sample.id); + if (evsel == NULL) { + pr_debug("event with id %" PRIu64 + " doesn't map to an evsel\n", sample.id); + goto out_munmap; + } + nr_events[evsel->idx]++; + } + + list_for_each_entry(evsel, &evlist->entries, node) { + if (nr_events[evsel->idx] != expected_nr_events[evsel->idx]) { + pr_debug("expected %d %s events, got %d\n", + expected_nr_events[evsel->idx], + event_name(evsel), nr_events[evsel->idx]); + goto out_munmap; + } + } + + err = 0; +out_munmap: + perf_evlist__munmap(evlist, 1); +out_close_fd: + for (i = 0; i < nsyscalls; ++i) + perf_evsel__close_fd(evsels[i], 1, threads->nr); +out_free_evlist: + perf_evlist__delete(evlist); +out_free_cpus: + cpu_map__delete(cpus); +out_free_threads: + thread_map__delete(threads); + return err; +#undef nsyscalls +} + static struct test { const char *desc; int (*func)(void); @@ -454,6 +610,10 @@ static struct test { .func = test__open_syscall_event_on_all_cpus, }, { + .desc = "read samples using the mmap interface", + .func = test__basic_mmap, + }, + { .func = NULL, }, }; diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index b6998e055767..ce2e50c891c7 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -21,10 +21,12 @@ #include "perf.h" #include "util/color.h" +#include "util/evlist.h" #include "util/evsel.h" #include "util/session.h" #include "util/symbol.h" #include "util/thread.h" +#include "util/thread_map.h" #include "util/util.h" #include <linux/rbtree.h> #include "util/parse-options.h" @@ -60,6 +62,8 @@ #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y)) +struct perf_evlist *evsel_list; + static bool system_wide = false; static int default_interval = 0; @@ -75,7 +79,7 @@ static struct cpu_map *cpus; static int realtime_prio = 0; static bool group = false; static unsigned int page_size; -static unsigned int mmap_pages = 16; +static unsigned int mmap_pages = 128; static int freq = 1000; /* 1 KHz */ static int delay_secs = 2; @@ -267,7 +271,7 @@ static void __zero_source_counters(struct sym_entry *syme) line = syme->src->lines; while (line) { - for (i = 0; i < nr_counters; i++) + for (i = 0; i < evsel_list->nr_entries; i++) line->count[i] = 0; line = line->next; } @@ -414,7 +418,7 @@ static double sym_weight(const struct sym_entry *sym) if (!display_weighted) return weight; - for (counter = 1; counter < nr_counters-1; counter++) + for (counter = 1; counter < evsel_list->nr_entries - 1; counter++) weight *= sym->count[counter]; weight /= (sym->count[counter] + 1); @@ -501,7 +505,7 @@ static void print_sym_table(void) rb_insert_active_sym(&tmp, syme); sum_ksamples += syme->snap_count; - for (j = 0; j < nr_counters; j++) + for (j = 0; j < evsel_list->nr_entries; j++) syme->count[j] = zero ? 0 : syme->count[j] * 7 / 8; } else list_remove_active_sym(syme); @@ -535,9 +539,9 @@ static void print_sym_table(void) esamples_percent); } - if (nr_counters == 1 || !display_weighted) { + if (evsel_list->nr_entries == 1 || !display_weighted) { struct perf_evsel *first; - first = list_entry(evsel_list.next, struct perf_evsel, node); + first = list_entry(evsel_list->entries.next, struct perf_evsel, node); printf("%" PRIu64, (uint64_t)first->attr.sample_period); if (freq) printf("Hz "); @@ -547,7 +551,7 @@ static void print_sym_table(void) if (!display_weighted) printf("%s", event_name(sym_evsel)); - else list_for_each_entry(counter, &evsel_list, node) { + else list_for_each_entry(counter, &evsel_list->entries, node) { if (counter->idx) printf("/"); @@ -606,7 +610,7 @@ static void print_sym_table(void) sym_width = winsize.ws_col - dso_width - 29; } putchar('\n'); - if (nr_counters == 1) + if (evsel_list->nr_entries == 1) printf(" samples pcnt"); else printf(" weight samples pcnt"); @@ -615,7 +619,7 @@ static void print_sym_table(void) printf(" RIP "); printf(" %-*.*s DSO\n", sym_width, sym_width, "function"); printf(" %s _______ _____", - nr_counters == 1 ? " " : "______"); + evsel_list->nr_entries == 1 ? " " : "______"); if (verbose) printf(" ________________"); printf(" %-*.*s", sym_width, sym_width, graph_line); @@ -634,7 +638,7 @@ static void print_sym_table(void) pcnt = 100.0 - (100.0 * ((sum_ksamples - syme->snap_count) / sum_ksamples)); - if (nr_counters == 1 || !display_weighted) + if (evsel_list->nr_entries == 1 || !display_weighted) printf("%20.2f ", syme->weight); else printf("%9.1f %10ld ", syme->weight, syme->snap_count); @@ -744,7 +748,7 @@ static void print_mapped_keys(void) fprintf(stdout, "\t[d] display refresh delay. \t(%d)\n", delay_secs); fprintf(stdout, "\t[e] display entries (lines). \t(%d)\n", print_entries); - if (nr_counters > 1) + if (evsel_list->nr_entries > 1) fprintf(stdout, "\t[E] active event counter. \t(%s)\n", event_name(sym_evsel)); fprintf(stdout, "\t[f] profile display filter (count). \t(%d)\n", count_filter); @@ -753,7 +757,7 @@ static void print_mapped_keys(void) fprintf(stdout, "\t[s] annotate symbol. \t(%s)\n", name?: "NULL"); fprintf(stdout, "\t[S] stop annotation.\n"); - if (nr_counters > 1) + if (evsel_list->nr_entries > 1) fprintf(stdout, "\t[w] toggle display weighted/count[E]r. \t(%d)\n", display_weighted ? 1 : 0); fprintf(stdout, @@ -783,7 +787,7 @@ static int key_mapped(int c) return 1; case 'E': case 'w': - return nr_counters > 1 ? 1 : 0; + return evsel_list->nr_entries > 1 ? 1 : 0; default: break; } @@ -831,22 +835,22 @@ static void handle_keypress(struct perf_session *session, int c) signal(SIGWINCH, SIG_DFL); break; case 'E': - if (nr_counters > 1) { + if (evsel_list->nr_entries > 1) { fprintf(stderr, "\nAvailable events:"); - list_for_each_entry(sym_evsel, &evsel_list, node) + list_for_each_entry(sym_evsel, &evsel_list->entries, node) fprintf(stderr, "\n\t%d %s", sym_evsel->idx, event_name(sym_evsel)); prompt_integer(&sym_counter, "Enter details event counter"); - if (sym_counter >= nr_counters) { - sym_evsel = list_entry(evsel_list.next, struct perf_evsel, node); + if (sym_counter >= evsel_list->nr_entries) { + sym_evsel = list_entry(evsel_list->entries.next, struct perf_evsel, node); sym_counter = 0; fprintf(stderr, "Sorry, no such event, using %s.\n", event_name(sym_evsel)); sleep(1); break; } - list_for_each_entry(sym_evsel, &evsel_list, node) + list_for_each_entry(sym_evsel, &evsel_list->entries, node) if (sym_evsel->idx == sym_counter) break; } else sym_counter = 0; @@ -930,6 +934,7 @@ repeat: /* Tag samples to be skipped. */ static const char *skip_symbols[] = { "default_idle", + "native_safe_halt", "cpu_idle", "enter_idle", "exit_idle", @@ -988,8 +993,7 @@ static int symbol_filter(struct map *map, struct symbol *sym) static void event__process_sample(const event_t *self, struct sample_data *sample, - struct perf_session *session, - struct perf_evsel *evsel) + struct perf_session *session) { u64 ip = self->ip.ip; struct sym_entry *syme; @@ -1082,8 +1086,12 @@ static void event__process_sample(const event_t *self, syme = symbol__priv(al.sym); if (!syme->skip) { - syme->count[evsel->idx]++; + struct perf_evsel *evsel; + syme->origin = origin; + evsel = perf_evlist__id2evsel(evsel_list, sample->id); + assert(evsel != NULL); + syme->count[evsel->idx]++; record_precise_ip(syme, evsel->idx, ip); pthread_mutex_lock(&active_symbols_lock); if (list_empty(&syme->node) || !syme->node.next) @@ -1092,156 +1100,52 @@ static void event__process_sample(const event_t *self, } } -struct mmap_data { - void *base; - int mask; - unsigned int prev; -}; - -static int perf_evsel__alloc_mmap_per_thread(struct perf_evsel *evsel, - int ncpus, int nthreads) -{ - evsel->priv = xyarray__new(ncpus, nthreads, sizeof(struct mmap_data)); - return evsel->priv != NULL ? 0 : -ENOMEM; -} - -static void perf_evsel__free_mmap(struct perf_evsel *evsel) -{ - xyarray__delete(evsel->priv); - evsel->priv = NULL; -} - -static unsigned int mmap_read_head(struct mmap_data *md) -{ - struct perf_event_mmap_page *pc = md->base; - int head; - - head = pc->data_head; - rmb(); - - return head; -} - -static void perf_session__mmap_read_counter(struct perf_session *self, - struct perf_evsel *evsel, - int cpu, int thread_idx) +static void perf_session__mmap_read_cpu(struct perf_session *self, int cpu) { - struct xyarray *mmap_array = evsel->priv; - struct mmap_data *md = xyarray__entry(mmap_array, cpu, thread_idx); - unsigned int head = mmap_read_head(md); - unsigned int old = md->prev; - unsigned char *data = md->base + page_size; struct sample_data sample; - int diff; - - /* - * If we're further behind than half the buffer, there's a chance - * the writer will bite our tail and mess up the samples under us. - * - * If we somehow ended up ahead of the head, we got messed up. - * - * In either case, truncate and restart at head. - */ - diff = head - old; - if (diff > md->mask / 2 || diff < 0) { - fprintf(stderr, "WARNING: failed to keep up with mmap data.\n"); - - /* - * head points to a known good entry, start there. - */ - old = head; - } - - for (; old != head;) { - event_t *event = (event_t *)&data[old & md->mask]; + event_t *event; - event_t event_copy; - - size_t size = event->header.size; - - /* - * Event straddles the mmap boundary -- header should always - * be inside due to u64 alignment of output. - */ - if ((old & md->mask) + size != ((old + size) & md->mask)) { - unsigned int offset = old; - unsigned int len = min(sizeof(*event), size), cpy; - void *dst = &event_copy; - - do { - cpy = min(md->mask + 1 - (offset & md->mask), len); - memcpy(dst, &data[offset & md->mask], cpy); - offset += cpy; - dst += cpy; - len -= cpy; - } while (len); - - event = &event_copy; - } + while ((event = perf_evlist__read_on_cpu(evsel_list, cpu)) != NULL) { + perf_session__parse_sample(self, event, &sample); - event__parse_sample(event, self, &sample); if (event->header.type == PERF_RECORD_SAMPLE) - event__process_sample(event, &sample, self, evsel); + event__process_sample(event, &sample, self); else event__process(event, &sample, self); - old += size; } - - md->prev = old; } -static struct pollfd *event_array; - static void perf_session__mmap_read(struct perf_session *self) { - struct perf_evsel *counter; - int i, thread_index; - - for (i = 0; i < cpus->nr; i++) { - list_for_each_entry(counter, &evsel_list, node) { - for (thread_index = 0; - thread_index < threads->nr; - thread_index++) { - perf_session__mmap_read_counter(self, - counter, i, thread_index); - } - } - } -} + int i; -int nr_poll; -int group_fd; + for (i = 0; i < cpus->nr; i++) + perf_session__mmap_read_cpu(self, i); +} -static void start_counter(int i, struct perf_evsel *evsel) +static void start_counters(struct perf_evlist *evlist) { - struct xyarray *mmap_array = evsel->priv; - struct mmap_data *mm; - struct perf_event_attr *attr; - int cpu = -1; - int thread_index; - - if (target_tid == -1) - cpu = cpus->map[i]; + struct perf_evsel *counter; - attr = &evsel->attr; + list_for_each_entry(counter, &evlist->entries, node) { + struct perf_event_attr *attr = &counter->attr; - attr->sample_type = PERF_SAMPLE_IP | PERF_SAMPLE_TID; + attr->sample_type = PERF_SAMPLE_IP | PERF_SAMPLE_TID; - if (freq) { - attr->sample_type |= PERF_SAMPLE_PERIOD; - attr->freq = 1; - attr->sample_freq = freq; - } + if (freq) { + attr->sample_type |= PERF_SAMPLE_PERIOD; + attr->freq = 1; + attr->sample_freq = freq; + } - attr->inherit = (cpu < 0) && inherit; - attr->mmap = 1; + if (evlist->nr_entries > 1) { + attr->sample_type |= PERF_SAMPLE_ID; + attr->read_format |= PERF_FORMAT_ID; + } - for (thread_index = 0; thread_index < threads->nr; thread_index++) { + attr->mmap = 1; try_again: - FD(evsel, i, thread_index) = sys_perf_event_open(attr, - threads->map[thread_index], cpu, group_fd, 0); - - if (FD(evsel, i, thread_index) < 0) { + if (perf_evsel__open(counter, cpus, threads, group, inherit) < 0) { int err = errno; if (err == EPERM || err == EACCES) @@ -1253,8 +1157,8 @@ try_again: * based cpu-clock-tick sw counter, which * is always available even if no PMU support: */ - if (attr->type == PERF_TYPE_HARDWARE - && attr->config == PERF_COUNT_HW_CPU_CYCLES) { + if (attr->type == PERF_TYPE_HARDWARE && + attr->config == PERF_COUNT_HW_CPU_CYCLES) { if (verbose) warning(" ... trying to fall back to cpu-clock-ticks\n"); @@ -1264,39 +1168,23 @@ try_again: goto try_again; } printf("\n"); - error("sys_perf_event_open() syscall returned with %d (%s). /bin/dmesg may provide additional information.\n", - FD(evsel, i, thread_index), strerror(err)); + error("sys_perf_event_open() syscall returned with %d " + "(%s). /bin/dmesg may provide additional information.\n", + err, strerror(err)); die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); exit(-1); } - assert(FD(evsel, i, thread_index) >= 0); - fcntl(FD(evsel, i, thread_index), F_SETFL, O_NONBLOCK); - - /* - * First counter acts as the group leader: - */ - if (group && group_fd == -1) - group_fd = FD(evsel, i, thread_index); - - event_array[nr_poll].fd = FD(evsel, i, thread_index); - event_array[nr_poll].events = POLLIN; - nr_poll++; - - mm = xyarray__entry(mmap_array, i, thread_index); - mm->prev = 0; - mm->mask = mmap_pages*page_size - 1; - mm->base = mmap(NULL, (mmap_pages+1)*page_size, - PROT_READ, MAP_SHARED, FD(evsel, i, thread_index), 0); - if (mm->base == MAP_FAILED) - die("failed to mmap with %d (%s)\n", errno, strerror(errno)); } + + if (perf_evlist__mmap(evlist, cpus, threads, mmap_pages, true) < 0) + die("failed to mmap with %d (%s)\n", errno, strerror(errno)); } static int __cmd_top(void) { pthread_t thread; - struct perf_evsel *counter; - int i, ret; + struct perf_evsel *first; + int ret; /* * FIXME: perf_session__new should allow passing a O_MMAP, so that all this * mmap reading, etc is encapsulated in it. Use O_WRONLY for now. @@ -1310,14 +1198,12 @@ static int __cmd_top(void) else event__synthesize_threads(event__process, session); - for (i = 0; i < cpus->nr; i++) { - group_fd = -1; - list_for_each_entry(counter, &evsel_list, node) - start_counter(i, counter); - } + start_counters(evsel_list); + first = list_entry(evsel_list->entries.next, struct perf_evsel, node); + perf_session__set_sample_type(session, first->attr.sample_type); /* Wait for a minimal set of events before starting the snapshot */ - poll(&event_array[0], nr_poll, 100); + poll(evsel_list->pollfd, evsel_list->nr_fds, 100); perf_session__mmap_read(session); @@ -1342,7 +1228,7 @@ static int __cmd_top(void) perf_session__mmap_read(session); if (hits == samples) - ret = poll(event_array, nr_poll, 100); + ret = poll(evsel_list->pollfd, evsel_list->nr_fds, 100); } return 0; @@ -1354,7 +1240,7 @@ static const char * const top_usage[] = { }; static const struct option options[] = { - OPT_CALLBACK('e', "event", NULL, "event", + OPT_CALLBACK('e', "event", &evsel_list, "event", "event selector. use 'perf list' to list available events", parse_events), OPT_INTEGER('c', "count", &default_interval, @@ -1404,6 +1290,10 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) struct perf_evsel *pos; int status = -ENOMEM; + evsel_list = perf_evlist__new(); + if (evsel_list == NULL) + return -ENOMEM; + page_size = sysconf(_SC_PAGE_SIZE); argc = parse_options(argc, argv, options, top_usage, 0); @@ -1419,11 +1309,6 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) usage_with_options(top_usage, options); } - event_array = malloc((sizeof(struct pollfd) * - MAX_NR_CPUS * MAX_COUNTERS * threads->nr)); - if (!event_array) - return -ENOMEM; - /* CPU and PID are mutually exclusive */ if (target_tid > 0 && cpu_list) { printf("WARNING: PID switch overriding CPU\n"); @@ -1431,7 +1316,8 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) cpu_list = NULL; } - if (!nr_counters && perf_evsel_list__create_default() < 0) { + if (!evsel_list->nr_entries && + perf_evlist__add_default(evsel_list) < 0) { pr_err("Not enough memory for event selector list\n"); return -ENOMEM; } @@ -1459,9 +1345,8 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) if (cpus == NULL) usage_with_options(top_usage, options); - list_for_each_entry(pos, &evsel_list, node) { - if (perf_evsel__alloc_mmap_per_thread(pos, cpus->nr, threads->nr) < 0 || - perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0) + list_for_each_entry(pos, &evsel_list->entries, node) { + if (perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0) goto out_free_fd; /* * Fill in the ones not specifically initialized via -c: @@ -1472,10 +1357,14 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) pos->attr.sample_period = default_interval; } - sym_evsel = list_entry(evsel_list.next, struct perf_evsel, node); + if (perf_evlist__alloc_pollfd(evsel_list, cpus->nr, threads->nr) < 0 || + perf_evlist__alloc_mmap(evsel_list, cpus->nr) < 0) + goto out_free_fd; + + sym_evsel = list_entry(evsel_list->entries.next, struct perf_evsel, node); symbol_conf.priv_size = (sizeof(struct sym_entry) + - (nr_counters + 1) * sizeof(unsigned long)); + (evsel_list->nr_entries + 1) * sizeof(unsigned long)); symbol_conf.try_vmlinux_path = (symbol_conf.vmlinux_name == NULL); if (symbol__init() < 0) @@ -1489,9 +1378,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) status = __cmd_top(); out_free_fd: - list_for_each_entry(pos, &evsel_list, node) - perf_evsel__free_mmap(pos); - perf_evsel_list__delete(); + perf_evlist__delete(evsel_list); return status; } diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 95aaf565c704..a5fc660c1f12 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -94,6 +94,32 @@ void get_term_dimensions(struct winsize *ws); #include "util/types.h" #include <stdbool.h> +struct perf_mmap { + void *base; + int mask; + unsigned int prev; +}; + +static inline unsigned int perf_mmap__read_head(struct perf_mmap *mm) +{ + struct perf_event_mmap_page *pc = mm->base; + int head = pc->data_head; + rmb(); + return head; +} + +static inline void perf_mmap__write_tail(struct perf_mmap *md, + unsigned long tail) +{ + struct perf_event_mmap_page *pc = md->base; + + /* + * ensure all reads are done before we write the tail out. + */ + /* mb(); */ + pc->data_tail = tail; +} + /* * prctl(PR_TASK_PERF_EVENTS_DISABLE) will (cheaply) disable all * counters in the current task. diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index e12d539417b2..f8c66d1435e0 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2009-2010, Frederic Weisbecker <fweisbec@gmail.com> + * Copyright (C) 2009-2011, Frederic Weisbecker <fweisbec@gmail.com> * * Handle the callchains from the stream in an ad-hoc radix tree and then * sort them in an rbtree. @@ -26,10 +26,10 @@ bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event) } #define chain_for_each_child(child, parent) \ - list_for_each_entry(child, &parent->children, brothers) + list_for_each_entry(child, &parent->children, siblings) #define chain_for_each_child_safe(child, next, parent) \ - list_for_each_entry_safe(child, next, &parent->children, brothers) + list_for_each_entry_safe(child, next, &parent->children, siblings) static void rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, @@ -38,14 +38,14 @@ rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; struct callchain_node *rnode; - u64 chain_cumul = cumul_hits(chain); + u64 chain_cumul = callchain_cumul_hits(chain); while (*p) { u64 rnode_cumul; parent = *p; rnode = rb_entry(parent, struct callchain_node, rb_node); - rnode_cumul = cumul_hits(rnode); + rnode_cumul = callchain_cumul_hits(rnode); switch (mode) { case CHAIN_FLAT: @@ -104,7 +104,7 @@ static void __sort_chain_graph_abs(struct callchain_node *node, chain_for_each_child(child, node) { __sort_chain_graph_abs(child, min_hit); - if (cumul_hits(child) >= min_hit) + if (callchain_cumul_hits(child) >= min_hit) rb_insert_callchain(&node->rb_root, child, CHAIN_GRAPH_ABS); } @@ -129,7 +129,7 @@ static void __sort_chain_graph_rel(struct callchain_node *node, chain_for_each_child(child, node) { __sort_chain_graph_rel(child, min_percent); - if (cumul_hits(child) >= min_hit) + if (callchain_cumul_hits(child) >= min_hit) rb_insert_callchain(&node->rb_root, child, CHAIN_GRAPH_REL); } @@ -143,7 +143,7 @@ sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_root *chain_root, rb_root->rb_node = chain_root->node.rb_root.rb_node; } -int register_callchain_param(struct callchain_param *param) +int callchain_register_param(struct callchain_param *param) { switch (param->mode) { case CHAIN_GRAPH_ABS: @@ -189,32 +189,27 @@ create_child(struct callchain_node *parent, bool inherit_children) chain_for_each_child(next, new) next->parent = new; } - list_add_tail(&new->brothers, &parent->children); + list_add_tail(&new->siblings, &parent->children); return new; } -struct resolved_ip { - u64 ip; - struct map_symbol ms; -}; - -struct resolved_chain { - u64 nr; - struct resolved_ip ips[0]; -}; - - /* * Fill the node with callchain values */ static void -fill_node(struct callchain_node *node, struct resolved_chain *chain, int start) +fill_node(struct callchain_node *node, struct callchain_cursor *cursor) { - unsigned int i; + struct callchain_cursor_node *cursor_node; + + node->val_nr = cursor->nr - cursor->pos; + if (!node->val_nr) + pr_warning("Warning: empty node in callchain tree\n"); - for (i = start; i < chain->nr; i++) { + cursor_node = callchain_cursor_current(cursor); + + while (cursor_node) { struct callchain_list *call; call = zalloc(sizeof(*call)); @@ -222,23 +217,25 @@ fill_node(struct callchain_node *node, struct resolved_chain *chain, int start) perror("not enough memory for the code path tree"); return; } - call->ip = chain->ips[i].ip; - call->ms = chain->ips[i].ms; + call->ip = cursor_node->ip; + call->ms.sym = cursor_node->sym; + call->ms.map = cursor_node->map; list_add_tail(&call->list, &node->val); + + callchain_cursor_advance(cursor); + cursor_node = callchain_cursor_current(cursor); } - node->val_nr = chain->nr - start; - if (!node->val_nr) - pr_warning("Warning: empty node in callchain tree\n"); } static void -add_child(struct callchain_node *parent, struct resolved_chain *chain, - int start, u64 period) +add_child(struct callchain_node *parent, + struct callchain_cursor *cursor, + u64 period) { struct callchain_node *new; new = create_child(parent, false); - fill_node(new, chain, start); + fill_node(new, cursor); new->children_hit = 0; new->hit = period; @@ -250,9 +247,10 @@ add_child(struct callchain_node *parent, struct resolved_chain *chain, * Then create another child to host the given callchain of new branch */ static void -split_add_child(struct callchain_node *parent, struct resolved_chain *chain, - struct callchain_list *to_split, int idx_parents, int idx_local, - u64 period) +split_add_child(struct callchain_node *parent, + struct callchain_cursor *cursor, + struct callchain_list *to_split, + u64 idx_parents, u64 idx_local, u64 period) { struct callchain_node *new; struct list_head *old_tail; @@ -272,14 +270,14 @@ split_add_child(struct callchain_node *parent, struct resolved_chain *chain, /* split the hits */ new->hit = parent->hit; new->children_hit = parent->children_hit; - parent->children_hit = cumul_hits(new); + parent->children_hit = callchain_cumul_hits(new); new->val_nr = parent->val_nr - idx_local; parent->val_nr = idx_local; /* create a new child for the new branch if any */ - if (idx_total < chain->nr) { + if (idx_total < cursor->nr) { parent->hit = 0; - add_child(parent, chain, idx_total, period); + add_child(parent, cursor, period); parent->children_hit += period; } else { parent->hit = period; @@ -287,36 +285,41 @@ split_add_child(struct callchain_node *parent, struct resolved_chain *chain, } static int -append_chain(struct callchain_node *root, struct resolved_chain *chain, - unsigned int start, u64 period); +append_chain(struct callchain_node *root, + struct callchain_cursor *cursor, + u64 period); static void -append_chain_children(struct callchain_node *root, struct resolved_chain *chain, - unsigned int start, u64 period) +append_chain_children(struct callchain_node *root, + struct callchain_cursor *cursor, + u64 period) { struct callchain_node *rnode; /* lookup in childrens */ chain_for_each_child(rnode, root) { - unsigned int ret = append_chain(rnode, chain, start, period); + unsigned int ret = append_chain(rnode, cursor, period); if (!ret) goto inc_children_hit; } /* nothing in children, add to the current node */ - add_child(root, chain, start, period); + add_child(root, cursor, period); inc_children_hit: root->children_hit += period; } static int -append_chain(struct callchain_node *root, struct resolved_chain *chain, - unsigned int start, u64 period) +append_chain(struct callchain_node *root, + struct callchain_cursor *cursor, + u64 period) { + struct callchain_cursor_node *curr_snap = cursor->curr; struct callchain_list *cnode; - unsigned int i = start; + u64 start = cursor->pos; bool found = false; + u64 matches; /* * Lookup in the current node @@ -324,141 +327,134 @@ append_chain(struct callchain_node *root, struct resolved_chain *chain, * anywhere inside a function. */ list_for_each_entry(cnode, &root->val, list) { + struct callchain_cursor_node *node; struct symbol *sym; - if (i == chain->nr) + node = callchain_cursor_current(cursor); + if (!node) break; - sym = chain->ips[i].ms.sym; + sym = node->sym; if (cnode->ms.sym && sym) { if (cnode->ms.sym->start != sym->start) break; - } else if (cnode->ip != chain->ips[i].ip) + } else if (cnode->ip != node->ip) break; if (!found) found = true; - i++; + + callchain_cursor_advance(cursor); } /* matches not, relay on the parent */ - if (!found) + if (!found) { + cursor->curr = curr_snap; + cursor->pos = start; return -1; + } + + matches = cursor->pos - start; /* we match only a part of the node. Split it and add the new chain */ - if (i - start < root->val_nr) { - split_add_child(root, chain, cnode, start, i - start, period); + if (matches < root->val_nr) { + split_add_child(root, cursor, cnode, start, matches, period); return 0; } /* we match 100% of the path, increment the hit */ - if (i - start == root->val_nr && i == chain->nr) { + if (matches == root->val_nr && cursor->pos == cursor->nr) { root->hit += period; return 0; } /* We match the node and still have a part remaining */ - append_chain_children(root, chain, i, period); + append_chain_children(root, cursor, period); return 0; } -static void filter_context(struct ip_callchain *old, struct resolved_chain *new, - struct map_symbol *syms) -{ - int i, j = 0; - - for (i = 0; i < (int)old->nr; i++) { - if (old->ips[i] >= PERF_CONTEXT_MAX) - continue; - - new->ips[j].ip = old->ips[i]; - new->ips[j].ms = syms[i]; - j++; - } - - new->nr = j; -} - - -int callchain_append(struct callchain_root *root, struct ip_callchain *chain, - struct map_symbol *syms, u64 period) +int callchain_append(struct callchain_root *root, + struct callchain_cursor *cursor, + u64 period) { - struct resolved_chain *filtered; - - if (!chain->nr) + if (!cursor->nr) return 0; - filtered = zalloc(sizeof(*filtered) + - chain->nr * sizeof(struct resolved_ip)); - if (!filtered) - return -ENOMEM; - - filter_context(chain, filtered, syms); - - if (!filtered->nr) - goto end; + callchain_cursor_commit(cursor); - append_chain_children(&root->node, filtered, 0, period); + append_chain_children(&root->node, cursor, period); - if (filtered->nr > root->max_depth) - root->max_depth = filtered->nr; -end: - free(filtered); + if (cursor->nr > root->max_depth) + root->max_depth = cursor->nr; return 0; } static int -merge_chain_branch(struct callchain_node *dst, struct callchain_node *src, - struct resolved_chain *chain) +merge_chain_branch(struct callchain_cursor *cursor, + struct callchain_node *dst, struct callchain_node *src) { + struct callchain_cursor_node **old_last = cursor->last; struct callchain_node *child, *next_child; struct callchain_list *list, *next_list; - int old_pos = chain->nr; + int old_pos = cursor->nr; int err = 0; list_for_each_entry_safe(list, next_list, &src->val, list) { - chain->ips[chain->nr].ip = list->ip; - chain->ips[chain->nr].ms = list->ms; - chain->nr++; + callchain_cursor_append(cursor, list->ip, + list->ms.map, list->ms.sym); list_del(&list->list); free(list); } - if (src->hit) - append_chain_children(dst, chain, 0, src->hit); + if (src->hit) { + callchain_cursor_commit(cursor); + append_chain_children(dst, cursor, src->hit); + } chain_for_each_child_safe(child, next_child, src) { - err = merge_chain_branch(dst, child, chain); + err = merge_chain_branch(cursor, dst, child); if (err) break; - list_del(&child->brothers); + list_del(&child->siblings); free(child); } - chain->nr = old_pos; + cursor->nr = old_pos; + cursor->last = old_last; return err; } -int callchain_merge(struct callchain_root *dst, struct callchain_root *src) +int callchain_merge(struct callchain_cursor *cursor, + struct callchain_root *dst, struct callchain_root *src) +{ + return merge_chain_branch(cursor, &dst->node, &src->node); +} + +int callchain_cursor_append(struct callchain_cursor *cursor, + u64 ip, struct map *map, struct symbol *sym) { - struct resolved_chain *chain; - int err; + struct callchain_cursor_node *node = *cursor->last; - chain = malloc(sizeof(*chain) + - src->max_depth * sizeof(struct resolved_ip)); - if (!chain) - return -ENOMEM; + if (!node) { + node = calloc(sizeof(*node), 1); + if (!node) + return -ENOMEM; - chain->nr = 0; + *cursor->last = node; + } - err = merge_chain_branch(&dst->node, &src->node, chain); + node->ip = ip; + node->map = map; + node->sym = sym; - free(chain); + cursor->nr++; - return err; + cursor->last = &node->next; + + return 0; } diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index c15fb8c24ad2..67137256a1cd 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -16,7 +16,7 @@ enum chain_mode { struct callchain_node { struct callchain_node *parent; - struct list_head brothers; + struct list_head siblings; struct list_head children; struct list_head val; struct rb_node rb_node; /* to sort nodes in an rbtree */ @@ -49,9 +49,30 @@ struct callchain_list { struct list_head list; }; +/* + * A callchain cursor is a single linked list that + * let one feed a callchain progressively. + * It keeps persitent allocated entries to minimize + * allocations. + */ +struct callchain_cursor_node { + u64 ip; + struct map *map; + struct symbol *sym; + struct callchain_cursor_node *next; +}; + +struct callchain_cursor { + u64 nr; + struct callchain_cursor_node *first; + struct callchain_cursor_node **last; + u64 pos; + struct callchain_cursor_node *curr; +}; + static inline void callchain_init(struct callchain_root *root) { - INIT_LIST_HEAD(&root->node.brothers); + INIT_LIST_HEAD(&root->node.siblings); INIT_LIST_HEAD(&root->node.children); INIT_LIST_HEAD(&root->node.val); @@ -61,15 +82,54 @@ static inline void callchain_init(struct callchain_root *root) root->max_depth = 0; } -static inline u64 cumul_hits(struct callchain_node *node) +static inline u64 callchain_cumul_hits(struct callchain_node *node) { return node->hit + node->children_hit; } -int register_callchain_param(struct callchain_param *param); -int callchain_append(struct callchain_root *root, struct ip_callchain *chain, - struct map_symbol *syms, u64 period); -int callchain_merge(struct callchain_root *dst, struct callchain_root *src); +int callchain_register_param(struct callchain_param *param); +int callchain_append(struct callchain_root *root, + struct callchain_cursor *cursor, + u64 period); + +int callchain_merge(struct callchain_cursor *cursor, + struct callchain_root *dst, struct callchain_root *src); bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event); + +/* + * Initialize a cursor before adding entries inside, but keep + * the previously allocated entries as a cache. + */ +static inline void callchain_cursor_reset(struct callchain_cursor *cursor) +{ + cursor->nr = 0; + cursor->last = &cursor->first; +} + +int callchain_cursor_append(struct callchain_cursor *cursor, u64 ip, + struct map *map, struct symbol *sym); + +/* Close a cursor writing session. Initialize for the reader */ +static inline void callchain_cursor_commit(struct callchain_cursor *cursor) +{ + cursor->curr = cursor->first; + cursor->pos = 0; +} + +/* Cursor reading iteration helpers */ +static inline struct callchain_cursor_node * +callchain_cursor_current(struct callchain_cursor *cursor) +{ + if (cursor->pos == cursor->nr) + return NULL; + + return cursor->curr; +} + +static inline void callchain_cursor_advance(struct callchain_cursor *cursor) +{ + cursor->curr = cursor->curr->next; + cursor->pos++; +} #endif /* __PERF_CALLCHAIN_H */ diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c index 3ccaa1043383..6893eec693ab 100644 --- a/tools/perf/util/cpumap.c +++ b/tools/perf/util/cpumap.c @@ -177,3 +177,8 @@ struct cpu_map *cpu_map__dummy_new(void) return cpus; } + +void cpu_map__delete(struct cpu_map *map) +{ + free(map); +} diff --git a/tools/perf/util/cpumap.h b/tools/perf/util/cpumap.h index f7a4f42f6307..072c0a374794 100644 --- a/tools/perf/util/cpumap.h +++ b/tools/perf/util/cpumap.h @@ -8,6 +8,6 @@ struct cpu_map { struct cpu_map *cpu_map__new(const char *cpu_list); struct cpu_map *cpu_map__dummy_new(void); -void *cpu_map__delete(struct cpu_map *map); +void cpu_map__delete(struct cpu_map *map); #endif /* __PERF_CPUMAP_H */ diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 1478ab4ee222..e4db8b888546 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -826,128 +826,3 @@ out_filtered: al->filtered = true; return 0; } - -static int event__parse_id_sample(const event_t *event, - struct perf_session *session, - struct sample_data *sample) -{ - const u64 *array; - u64 type; - - sample->cpu = sample->pid = sample->tid = -1; - sample->stream_id = sample->id = sample->time = -1ULL; - - if (!session->sample_id_all) - return 0; - - array = event->sample.array; - array += ((event->header.size - - sizeof(event->header)) / sizeof(u64)) - 1; - type = session->sample_type; - - if (type & PERF_SAMPLE_CPU) { - u32 *p = (u32 *)array; - sample->cpu = *p; - array--; - } - - if (type & PERF_SAMPLE_STREAM_ID) { - sample->stream_id = *array; - array--; - } - - if (type & PERF_SAMPLE_ID) { - sample->id = *array; - array--; - } - - if (type & PERF_SAMPLE_TIME) { - sample->time = *array; - array--; - } - - if (type & PERF_SAMPLE_TID) { - u32 *p = (u32 *)array; - sample->pid = p[0]; - sample->tid = p[1]; - } - - return 0; -} - -int event__parse_sample(const event_t *event, struct perf_session *session, - struct sample_data *data) -{ - const u64 *array; - u64 type; - - if (event->header.type != PERF_RECORD_SAMPLE) - return event__parse_id_sample(event, session, data); - - array = event->sample.array; - type = session->sample_type; - - if (type & PERF_SAMPLE_IP) { - data->ip = event->ip.ip; - array++; - } - - if (type & PERF_SAMPLE_TID) { - u32 *p = (u32 *)array; - data->pid = p[0]; - data->tid = p[1]; - array++; - } - - if (type & PERF_SAMPLE_TIME) { - data->time = *array; - array++; - } - - if (type & PERF_SAMPLE_ADDR) { - data->addr = *array; - array++; - } - - data->id = -1ULL; - if (type & PERF_SAMPLE_ID) { - data->id = *array; - array++; - } - - if (type & PERF_SAMPLE_STREAM_ID) { - data->stream_id = *array; - array++; - } - - if (type & PERF_SAMPLE_CPU) { - u32 *p = (u32 *)array; - data->cpu = *p; - array++; - } else - data->cpu = -1; - - if (type & PERF_SAMPLE_PERIOD) { - data->period = *array; - array++; - } - - if (type & PERF_SAMPLE_READ) { - pr_debug("PERF_SAMPLE_READ is unsuported for now\n"); - return -1; - } - - if (type & PERF_SAMPLE_CALLCHAIN) { - data->callchain = (struct ip_callchain *)array; - array += 1 + data->callchain->nr; - } - - if (type & PERF_SAMPLE_RAW) { - u32 *p = (u32 *)array; - data->raw_size = *p; - p++; - data->raw_data = p; - } - - return 0; -} diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 2b7e91902f10..d79e4edd82f9 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -169,9 +169,10 @@ struct addr_location; int event__preprocess_sample(const event_t *self, struct perf_session *session, struct addr_location *al, struct sample_data *data, symbol_filter_t filter); -int event__parse_sample(const event_t *event, struct perf_session *session, - struct sample_data *sample); const char *event__get_event_name(unsigned int id); +int event__parse_sample(const event_t *event, u64 type, bool sample_id_all, + struct sample_data *sample); + #endif /* __PERF_RECORD_H */ diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c new file mode 100644 index 000000000000..df0610e9c61b --- /dev/null +++ b/tools/perf/util/evlist.c @@ -0,0 +1,170 @@ +#include <poll.h> +#include "evlist.h" +#include "evsel.h" +#include "util.h" + +#include <linux/bitops.h> +#include <linux/hash.h> + +void perf_evlist__init(struct perf_evlist *evlist) +{ + int i; + + for (i = 0; i < PERF_EVLIST__HLIST_SIZE; ++i) + INIT_HLIST_HEAD(&evlist->heads[i]); + INIT_LIST_HEAD(&evlist->entries); +} + +struct perf_evlist *perf_evlist__new(void) +{ + struct perf_evlist *evlist = zalloc(sizeof(*evlist)); + + if (evlist != NULL) + perf_evlist__init(evlist); + + return evlist; +} + +static void perf_evlist__purge(struct perf_evlist *evlist) +{ + struct perf_evsel *pos, *n; + + list_for_each_entry_safe(pos, n, &evlist->entries, node) { + list_del_init(&pos->node); + perf_evsel__delete(pos); + } + + evlist->nr_entries = 0; +} + +void perf_evlist__exit(struct perf_evlist *evlist) +{ + free(evlist->mmap); + free(evlist->pollfd); + evlist->mmap = NULL; + evlist->pollfd = NULL; +} + +void perf_evlist__delete(struct perf_evlist *evlist) +{ + perf_evlist__purge(evlist); + perf_evlist__exit(evlist); + free(evlist); +} + +void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry) +{ + list_add_tail(&entry->node, &evlist->entries); + ++evlist->nr_entries; +} + +int perf_evlist__add_default(struct perf_evlist *evlist) +{ + struct perf_event_attr attr = { + .type = PERF_TYPE_HARDWARE, + .config = PERF_COUNT_HW_CPU_CYCLES, + }; + struct perf_evsel *evsel = perf_evsel__new(&attr, 0); + + if (evsel == NULL) + return -ENOMEM; + + perf_evlist__add(evlist, evsel); + return 0; +} + +int perf_evlist__alloc_pollfd(struct perf_evlist *evlist, int ncpus, int nthreads) +{ + int nfds = ncpus * nthreads * evlist->nr_entries; + evlist->pollfd = malloc(sizeof(struct pollfd) * nfds); + return evlist->pollfd != NULL ? 0 : -ENOMEM; +} + +void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd) +{ + fcntl(fd, F_SETFL, O_NONBLOCK); + evlist->pollfd[evlist->nr_fds].fd = fd; + evlist->pollfd[evlist->nr_fds].events = POLLIN; + evlist->nr_fds++; +} + +struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id) +{ + struct hlist_head *head; + struct hlist_node *pos; + struct perf_sample_id *sid; + int hash; + + if (evlist->nr_entries == 1) + return list_entry(evlist->entries.next, struct perf_evsel, node); + + hash = hash_64(id, PERF_EVLIST__HLIST_BITS); + head = &evlist->heads[hash]; + + hlist_for_each_entry(sid, pos, head, node) + if (sid->id == id) + return sid->evsel; + return NULL; +} + +event_t *perf_evlist__read_on_cpu(struct perf_evlist *evlist, int cpu) +{ + /* XXX Move this to perf.c, making it generally available */ + unsigned int page_size = sysconf(_SC_PAGE_SIZE); + struct perf_mmap *md = &evlist->mmap[cpu]; + unsigned int head = perf_mmap__read_head(md); + unsigned int old = md->prev; + unsigned char *data = md->base + page_size; + event_t *event = NULL; + int diff; + + /* + * If we're further behind than half the buffer, there's a chance + * the writer will bite our tail and mess up the samples under us. + * + * If we somehow ended up ahead of the head, we got messed up. + * + * In either case, truncate and restart at head. + */ + diff = head - old; + if (diff > md->mask / 2 || diff < 0) { + fprintf(stderr, "WARNING: failed to keep up with mmap data.\n"); + + /* + * head points to a known good entry, start there. + */ + old = head; + } + + if (old != head) { + size_t size; + + event = (event_t *)&data[old & md->mask]; + size = event->header.size; + + /* + * Event straddles the mmap boundary -- header should always + * be inside due to u64 alignment of output. + */ + if ((old & md->mask) + size != ((old + size) & md->mask)) { + unsigned int offset = old; + unsigned int len = min(sizeof(*event), size), cpy; + void *dst = &evlist->event_copy; + + do { + cpy = min(md->mask + 1 - (offset & md->mask), len); + memcpy(dst, &data[offset & md->mask], cpy); + offset += cpy; + dst += cpy; + len -= cpy; + } while (len); + + event = &evlist->event_copy; + } + + old += size; + } + + md->prev = old; + return event; +} diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h new file mode 100644 index 000000000000..acbe48eac608 --- /dev/null +++ b/tools/perf/util/evlist.h @@ -0,0 +1,41 @@ +#ifndef __PERF_EVLIST_H +#define __PERF_EVLIST_H 1 + +#include <linux/list.h> +#include "../perf.h" +#include "event.h" + +struct pollfd; + +#define PERF_EVLIST__HLIST_BITS 8 +#define PERF_EVLIST__HLIST_SIZE (1 << PERF_EVLIST__HLIST_BITS) + +struct perf_evlist { + struct list_head entries; + struct hlist_head heads[PERF_EVLIST__HLIST_SIZE]; + int nr_entries; + int nr_fds; + int mmap_len; + event_t event_copy; + struct perf_mmap *mmap; + struct pollfd *pollfd; +}; + +struct perf_evsel; + +struct perf_evlist *perf_evlist__new(void); +void perf_evlist__init(struct perf_evlist *evlist); +void perf_evlist__exit(struct perf_evlist *evlist); +void perf_evlist__delete(struct perf_evlist *evlist); + +void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry); +int perf_evlist__add_default(struct perf_evlist *evlist); + +int perf_evlist__alloc_pollfd(struct perf_evlist *evlist, int ncpus, int nthreads); +void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd); + +struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id); + +event_t *perf_evlist__read_on_cpu(struct perf_evlist *self, int cpu); + +#endif /* __PERF_EVLIST_H */ diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index f5cfed60af98..76ab553637d6 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -1,20 +1,33 @@ #include "evsel.h" +#include "evlist.h" #include "../perf.h" #include "util.h" #include "cpumap.h" -#include "thread.h" +#include "thread_map.h" + +#include <unistd.h> +#include <sys/mman.h> + +#include <linux/bitops.h> +#include <linux/hash.h> #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y)) +#define SID(e, x, y) xyarray__entry(e->id, x, y) + +void perf_evsel__init(struct perf_evsel *evsel, + struct perf_event_attr *attr, int idx) +{ + evsel->idx = idx; + evsel->attr = *attr; + INIT_LIST_HEAD(&evsel->node); +} struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx) { struct perf_evsel *evsel = zalloc(sizeof(*evsel)); - if (evsel != NULL) { - evsel->idx = idx; - evsel->attr = *attr; - INIT_LIST_HEAD(&evsel->node); - } + if (evsel != NULL) + perf_evsel__init(evsel, attr, idx); return evsel; } @@ -25,6 +38,12 @@ int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads) return evsel->fd != NULL ? 0 : -ENOMEM; } +int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads) +{ + evsel->id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id)); + return evsel->id != NULL ? 0 : -ENOMEM; +} + int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus) { evsel->counts = zalloc((sizeof(*evsel->counts) + @@ -38,6 +57,12 @@ void perf_evsel__free_fd(struct perf_evsel *evsel) evsel->fd = NULL; } +void perf_evsel__free_id(struct perf_evsel *evsel) +{ + xyarray__delete(evsel->id); + evsel->id = NULL; +} + void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads) { int cpu, thread; @@ -49,10 +74,34 @@ void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads) } } -void perf_evsel__delete(struct perf_evsel *evsel) +void perf_evlist__munmap(struct perf_evlist *evlist, int ncpus) +{ + int cpu; + + for (cpu = 0; cpu < ncpus; cpu++) { + if (evlist->mmap[cpu].base != NULL) { + munmap(evlist->mmap[cpu].base, evlist->mmap_len); + evlist->mmap[cpu].base = NULL; + } + } +} + +int perf_evlist__alloc_mmap(struct perf_evlist *evlist, int ncpus) +{ + evlist->mmap = zalloc(ncpus * sizeof(struct perf_mmap)); + return evlist->mmap != NULL ? 0 : -ENOMEM; +} + +void perf_evsel__exit(struct perf_evsel *evsel) { assert(list_empty(&evsel->node)); xyarray__delete(evsel->fd); + xyarray__delete(evsel->id); +} + +void perf_evsel__delete(struct perf_evsel *evsel) +{ + perf_evsel__exit(evsel); free(evsel); } @@ -128,7 +177,7 @@ int __perf_evsel__read(struct perf_evsel *evsel, } static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, - struct thread_map *threads) + struct thread_map *threads, bool group, bool inherit) { int cpu, thread; @@ -137,12 +186,20 @@ static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, return -1; for (cpu = 0; cpu < cpus->nr; cpu++) { + int group_fd = -1; + + evsel->attr.inherit = (cpus->map[cpu] < 0) && inherit; + for (thread = 0; thread < threads->nr; thread++) { FD(evsel, cpu, thread) = sys_perf_event_open(&evsel->attr, threads->map[thread], - cpus->map[cpu], -1, 0); + cpus->map[cpu], + group_fd, 0); if (FD(evsel, cpu, thread) < 0) goto out_close; + + if (group && group_fd == -1) + group_fd = FD(evsel, cpu, thread); } } @@ -175,10 +232,9 @@ static struct { .threads = { -1, }, }; -int perf_evsel__open(struct perf_evsel *evsel, - struct cpu_map *cpus, struct thread_map *threads) +int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, + struct thread_map *threads, bool group, bool inherit) { - if (cpus == NULL) { /* Work around old compiler warnings about strict aliasing */ cpus = &empty_cpu_map.map; @@ -187,15 +243,243 @@ int perf_evsel__open(struct perf_evsel *evsel, if (threads == NULL) threads = &empty_thread_map.map; - return __perf_evsel__open(evsel, cpus, threads); + return __perf_evsel__open(evsel, cpus, threads, group, inherit); +} + +int perf_evsel__open_per_cpu(struct perf_evsel *evsel, + struct cpu_map *cpus, bool group, bool inherit) +{ + return __perf_evsel__open(evsel, cpus, &empty_thread_map.map, group, inherit); +} + +int perf_evsel__open_per_thread(struct perf_evsel *evsel, + struct thread_map *threads, bool group, bool inherit) +{ + return __perf_evsel__open(evsel, &empty_cpu_map.map, threads, group, inherit); +} + +static int __perf_evlist__mmap(struct perf_evlist *evlist, int cpu, int prot, + int mask, int fd) +{ + evlist->mmap[cpu].prev = 0; + evlist->mmap[cpu].mask = mask; + evlist->mmap[cpu].base = mmap(NULL, evlist->mmap_len, prot, + MAP_SHARED, fd, 0); + if (evlist->mmap[cpu].base == MAP_FAILED) + return -1; + + perf_evlist__add_pollfd(evlist, fd); + return 0; +} + +static int perf_evlist__id_hash(struct perf_evlist *evlist, struct perf_evsel *evsel, + int cpu, int thread, int fd) +{ + struct perf_sample_id *sid; + u64 read_data[4] = { 0, }; + int hash, id_idx = 1; /* The first entry is the counter value */ + + if (!(evsel->attr.read_format & PERF_FORMAT_ID) || + read(fd, &read_data, sizeof(read_data)) == -1) + return -1; + + if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) + ++id_idx; + if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) + ++id_idx; + + sid = SID(evsel, cpu, thread); + sid->id = read_data[id_idx]; + sid->evsel = evsel; + hash = hash_64(sid->id, PERF_EVLIST__HLIST_BITS); + hlist_add_head(&sid->node, &evlist->heads[hash]); + return 0; +} + +/** perf_evlist__mmap - Create per cpu maps to receive events + * + * @evlist - list of events + * @cpus - cpu map being monitored + * @threads - threads map being monitored + * @pages - map length in pages + * @overwrite - overwrite older events? + * + * If overwrite is false the user needs to signal event consuption using: + * + * struct perf_mmap *m = &evlist->mmap[cpu]; + * unsigned int head = perf_mmap__read_head(m); + * + * perf_mmap__write_tail(m, head) + */ +int perf_evlist__mmap(struct perf_evlist *evlist, struct cpu_map *cpus, + struct thread_map *threads, int pages, bool overwrite) +{ + unsigned int page_size = sysconf(_SC_PAGE_SIZE); + int mask = pages * page_size - 1, cpu; + struct perf_evsel *first_evsel, *evsel; + int thread, prot = PROT_READ | (overwrite ? 0 : PROT_WRITE); + + if (evlist->mmap == NULL && + perf_evlist__alloc_mmap(evlist, cpus->nr) < 0) + return -ENOMEM; + + if (evlist->pollfd == NULL && + perf_evlist__alloc_pollfd(evlist, cpus->nr, threads->nr) < 0) + return -ENOMEM; + + evlist->mmap_len = (pages + 1) * page_size; + first_evsel = list_entry(evlist->entries.next, struct perf_evsel, node); + + list_for_each_entry(evsel, &evlist->entries, node) { + if ((evsel->attr.read_format & PERF_FORMAT_ID) && + evsel->id == NULL && + perf_evsel__alloc_id(evsel, cpus->nr, threads->nr) < 0) + return -ENOMEM; + + for (cpu = 0; cpu < cpus->nr; cpu++) { + for (thread = 0; thread < threads->nr; thread++) { + int fd = FD(evsel, cpu, thread); + + if (evsel->idx || thread) { + if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, + FD(first_evsel, cpu, 0)) != 0) + goto out_unmap; + } else if (__perf_evlist__mmap(evlist, cpu, prot, mask, fd) < 0) + goto out_unmap; + + if ((evsel->attr.read_format & PERF_FORMAT_ID) && + perf_evlist__id_hash(evlist, evsel, cpu, thread, fd) < 0) + goto out_unmap; + } + } + } + + return 0; + +out_unmap: + for (cpu = 0; cpu < cpus->nr; cpu++) { + if (evlist->mmap[cpu].base != NULL) { + munmap(evlist->mmap[cpu].base, evlist->mmap_len); + evlist->mmap[cpu].base = NULL; + } + } + return -1; } -int perf_evsel__open_per_cpu(struct perf_evsel *evsel, struct cpu_map *cpus) +static int event__parse_id_sample(const event_t *event, u64 type, + struct sample_data *sample) { - return __perf_evsel__open(evsel, cpus, &empty_thread_map.map); + const u64 *array = event->sample.array; + + array += ((event->header.size - + sizeof(event->header)) / sizeof(u64)) - 1; + + if (type & PERF_SAMPLE_CPU) { + u32 *p = (u32 *)array; + sample->cpu = *p; + array--; + } + + if (type & PERF_SAMPLE_STREAM_ID) { + sample->stream_id = *array; + array--; + } + + if (type & PERF_SAMPLE_ID) { + sample->id = *array; + array--; + } + + if (type & PERF_SAMPLE_TIME) { + sample->time = *array; + array--; + } + + if (type & PERF_SAMPLE_TID) { + u32 *p = (u32 *)array; + sample->pid = p[0]; + sample->tid = p[1]; + } + + return 0; } -int perf_evsel__open_per_thread(struct perf_evsel *evsel, struct thread_map *threads) +int event__parse_sample(const event_t *event, u64 type, bool sample_id_all, + struct sample_data *data) { - return __perf_evsel__open(evsel, &empty_cpu_map.map, threads); + const u64 *array; + + data->cpu = data->pid = data->tid = -1; + data->stream_id = data->id = data->time = -1ULL; + + if (event->header.type != PERF_RECORD_SAMPLE) { + if (!sample_id_all) + return 0; + return event__parse_id_sample(event, type, data); + } + + array = event->sample.array; + + if (type & PERF_SAMPLE_IP) { + data->ip = event->ip.ip; + array++; + } + + if (type & PERF_SAMPLE_TID) { + u32 *p = (u32 *)array; + data->pid = p[0]; + data->tid = p[1]; + array++; + } + + if (type & PERF_SAMPLE_TIME) { + data->time = *array; + array++; + } + + if (type & PERF_SAMPLE_ADDR) { + data->addr = *array; + array++; + } + + data->id = -1ULL; + if (type & PERF_SAMPLE_ID) { + data->id = *array; + array++; + } + + if (type & PERF_SAMPLE_STREAM_ID) { + data->stream_id = *array; + array++; + } + + if (type & PERF_SAMPLE_CPU) { + u32 *p = (u32 *)array; + data->cpu = *p; + array++; + } + + if (type & PERF_SAMPLE_PERIOD) { + data->period = *array; + array++; + } + + if (type & PERF_SAMPLE_READ) { + fprintf(stderr, "PERF_SAMPLE_READ is unsuported for now\n"); + return -1; + } + + if (type & PERF_SAMPLE_CALLCHAIN) { + data->callchain = (struct ip_callchain *)array; + array += 1 + data->callchain->nr; + } + + if (type & PERF_SAMPLE_RAW) { + u32 *p = (u32 *)array; + data->raw_size = *p; + p++; + data->raw_data = p; + } + + return 0; } diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index b2d755fe88a5..7962e7587dea 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -24,11 +24,24 @@ struct perf_counts { struct perf_counts_values cpu[]; }; +struct perf_evsel; + +/* + * Per fd, to map back from PERF_SAMPLE_ID to evsel, only used when there are + * more than one entry in the evlist. + */ +struct perf_sample_id { + struct hlist_node node; + u64 id; + struct perf_evsel *evsel; +}; + struct perf_evsel { struct list_head node; struct perf_event_attr attr; char *filter; struct xyarray *fd; + struct xyarray *id; struct perf_counts *counts; int idx; void *priv; @@ -36,19 +49,31 @@ struct perf_evsel { struct cpu_map; struct thread_map; +struct perf_evlist; struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx); +void perf_evsel__init(struct perf_evsel *evsel, + struct perf_event_attr *attr, int idx); +void perf_evsel__exit(struct perf_evsel *evsel); void perf_evsel__delete(struct perf_evsel *evsel); int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads); +int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads); int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus); +int perf_evlist__alloc_mmap(struct perf_evlist *evlist, int ncpus); void perf_evsel__free_fd(struct perf_evsel *evsel); +void perf_evsel__free_id(struct perf_evsel *evsel); void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads); -int perf_evsel__open_per_cpu(struct perf_evsel *evsel, struct cpu_map *cpus); -int perf_evsel__open_per_thread(struct perf_evsel *evsel, struct thread_map *threads); -int perf_evsel__open(struct perf_evsel *evsel, - struct cpu_map *cpus, struct thread_map *threads); +int perf_evsel__open_per_cpu(struct perf_evsel *evsel, + struct cpu_map *cpus, bool group, bool inherit); +int perf_evsel__open_per_thread(struct perf_evsel *evsel, + struct thread_map *threads, bool group, bool inherit); +int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus, + struct thread_map *threads, bool group, bool inherit); +int perf_evlist__mmap(struct perf_evlist *evlist, struct cpu_map *cpus, + struct thread_map *threads, int pages, bool overwrite); +void perf_evlist__munmap(struct perf_evlist *evlist, int ncpus); #define perf_evsel__match(evsel, t, c) \ (evsel->attr.type == PERF_TYPE_##t && \ diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index f6a929e74981..f0138d472339 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -8,6 +8,7 @@ #include <linux/list.h> #include <linux/kernel.h> +#include "evlist.h" #include "util.h" #include "header.h" #include "../perf.h" @@ -428,7 +429,8 @@ static bool perf_session__read_build_ids(struct perf_session *self, bool with_hi return ret; } -static int perf_header__adds_write(struct perf_header *self, int fd) +static int perf_header__adds_write(struct perf_header *self, + struct perf_evlist *evlist, int fd) { int nr_sections; struct perf_session *session; @@ -463,7 +465,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd) /* Write trace info */ trace_sec->offset = lseek(fd, 0, SEEK_CUR); - read_tracing_data(fd, &evsel_list); + read_tracing_data(fd, &evlist->entries); trace_sec->size = lseek(fd, 0, SEEK_CUR) - trace_sec->offset; } @@ -513,7 +515,8 @@ int perf_header__write_pipe(int fd) return 0; } -int perf_header__write(struct perf_header *self, int fd, bool at_exit) +int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, + int fd, bool at_exit) { struct perf_file_header f_header; struct perf_file_attr f_attr; @@ -566,7 +569,7 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit) self->data_offset = lseek(fd, 0, SEEK_CUR); if (at_exit) { - err = perf_header__adds_write(self, fd); + err = perf_header__adds_write(self, evlist, fd); if (err < 0) return err; } @@ -1133,7 +1136,7 @@ int event__process_event_type(event_t *self, return 0; } -int event__synthesize_tracing_data(int fd, struct list_head *pattrs, +int event__synthesize_tracing_data(int fd, struct perf_evlist *evlist, event__handler_t process, struct perf_session *session __unused) { @@ -1144,7 +1147,7 @@ int event__synthesize_tracing_data(int fd, struct list_head *pattrs, memset(&ev, 0, sizeof(ev)); ev.tracing_data.header.type = PERF_RECORD_HEADER_TRACING_DATA; - size = read_tracing_data_size(fd, pattrs); + size = read_tracing_data_size(fd, &evlist->entries); if (size <= 0) return size; aligned_size = ALIGN(size, sizeof(u64)); @@ -1154,7 +1157,7 @@ int event__synthesize_tracing_data(int fd, struct list_head *pattrs, process(&ev, NULL, session); - err = read_tracing_data(fd, pattrs); + err = read_tracing_data(fd, &evlist->entries); write_padded(fd, NULL, 0, padding); return aligned_size; diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index 33f16be7b72f..65afd7f74e0d 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -65,8 +65,11 @@ struct perf_header { int perf_header__init(struct perf_header *self); void perf_header__exit(struct perf_header *self); +struct perf_evlist; + int perf_header__read(struct perf_session *session, int fd); -int perf_header__write(struct perf_header *self, int fd, bool at_exit); +int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, + int fd, bool at_exit); int perf_header__write_pipe(int fd); int perf_header__add_attr(struct perf_header *self, @@ -113,7 +116,7 @@ int event__synthesize_event_types(event__handler_t process, int event__process_event_type(event_t *self, struct perf_session *session); -int event__synthesize_tracing_data(int fd, struct list_head *pattrs, +int event__synthesize_tracing_data(int fd, struct perf_evlist *evlist, event__handler_t process, struct perf_session *session); int event__process_tracing_data(event_t *self, diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 32f4f1f2f6e4..02ed318d7312 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -211,7 +211,9 @@ void hist_entry__free(struct hist_entry *he) * collapse the histogram */ -static bool collapse__insert_entry(struct rb_root *root, struct hist_entry *he) +static bool hists__collapse_insert_entry(struct hists *self, + struct rb_root *root, + struct hist_entry *he) { struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; @@ -226,8 +228,11 @@ static bool collapse__insert_entry(struct rb_root *root, struct hist_entry *he) if (!cmp) { iter->period += he->period; - if (symbol_conf.use_callchain) - callchain_merge(iter->callchain, he->callchain); + if (symbol_conf.use_callchain) { + callchain_cursor_reset(&self->callchain_cursor); + callchain_merge(&self->callchain_cursor, iter->callchain, + he->callchain); + } hist_entry__free(he); return false; } @@ -262,7 +267,7 @@ void hists__collapse_resort(struct hists *self) next = rb_next(&n->rb_node); rb_erase(&n->rb_node, &self->entries); - if (collapse__insert_entry(&tmp, n)) + if (hists__collapse_insert_entry(self, &tmp, n)) hists__inc_nr_entries(self, n); } @@ -425,7 +430,7 @@ static size_t __callchain__fprintf_graph(FILE *fp, struct callchain_node *self, u64 cumul; child = rb_entry(node, struct callchain_node, rb_node); - cumul = cumul_hits(child); + cumul = callchain_cumul_hits(child); remaining -= cumul; /* diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index ee789856a8c9..889559b86492 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -77,6 +77,8 @@ struct hists { u64 event_stream; u32 type; u16 col_len[HISTC_NR_COLS]; + /* Best would be to reuse the session callchain cursor */ + struct callchain_cursor callchain_cursor; }; struct hist_entry *__hists__add_entry(struct hists *self, diff --git a/tools/perf/util/include/linux/list.h b/tools/perf/util/include/linux/list.h index f5ca26e53fbb..356c7e467b83 100644 --- a/tools/perf/util/include/linux/list.h +++ b/tools/perf/util/include/linux/list.h @@ -1,3 +1,4 @@ +#include <linux/kernel.h> #include "../../../../include/linux/list.h" #ifndef PERF_LIST_H diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 135f69baf966..cf082daa43e3 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1,6 +1,7 @@ #include "../../../include/linux/hw_breakpoint.h" #include "util.h" #include "../perf.h" +#include "evlist.h" #include "evsel.h" #include "parse-options.h" #include "parse-events.h" @@ -11,10 +12,6 @@ #include "header.h" #include "debugfs.h" -int nr_counters; - -LIST_HEAD(evsel_list); - struct event_symbol { u8 type; u64 config; @@ -449,8 +446,8 @@ parse_single_tracepoint_event(char *sys_name, /* sys + ':' + event + ':' + flags*/ #define MAX_EVOPT_LEN (MAX_EVENT_LENGTH * 2 + 2 + 128) static enum event_result -parse_multiple_tracepoint_event(char *sys_name, const char *evt_exp, - char *flags) +parse_multiple_tracepoint_event(const struct option *opt, char *sys_name, + const char *evt_exp, char *flags) { char evt_path[MAXPATHLEN]; struct dirent *evt_ent; @@ -483,15 +480,16 @@ parse_multiple_tracepoint_event(char *sys_name, const char *evt_exp, if (len < 0) return EVT_FAILED; - if (parse_events(NULL, event_opt, 0)) + if (parse_events(opt, event_opt, 0)) return EVT_FAILED; } return EVT_HANDLED_ALL; } -static enum event_result parse_tracepoint_event(const char **strp, - struct perf_event_attr *attr) +static enum event_result +parse_tracepoint_event(const struct option *opt, const char **strp, + struct perf_event_attr *attr) { const char *evt_name; char *flags = NULL, *comma_loc; @@ -530,7 +528,7 @@ static enum event_result parse_tracepoint_event(const char **strp, return EVT_FAILED; if (strpbrk(evt_name, "*?")) { *strp += strlen(sys_name) + evt_length + 1; /* 1 == the ':' */ - return parse_multiple_tracepoint_event(sys_name, evt_name, + return parse_multiple_tracepoint_event(opt, sys_name, evt_name, flags); } else { return parse_single_tracepoint_event(sys_name, evt_name, @@ -740,11 +738,12 @@ parse_event_modifier(const char **strp, struct perf_event_attr *attr) * Symbolic names are (almost) exactly matched. */ static enum event_result -parse_event_symbols(const char **str, struct perf_event_attr *attr) +parse_event_symbols(const struct option *opt, const char **str, + struct perf_event_attr *attr) { enum event_result ret; - ret = parse_tracepoint_event(str, attr); + ret = parse_tracepoint_event(opt, str, attr); if (ret != EVT_FAILED) goto modifier; @@ -778,14 +777,15 @@ modifier: return ret; } -int parse_events(const struct option *opt __used, const char *str, int unset __used) +int parse_events(const struct option *opt, const char *str, int unset __used) { + struct perf_evlist *evlist = *(struct perf_evlist **)opt->value; struct perf_event_attr attr; enum event_result ret; for (;;) { memset(&attr, 0, sizeof(attr)); - ret = parse_event_symbols(&str, &attr); + ret = parse_event_symbols(opt, &str, &attr); if (ret == EVT_FAILED) return -1; @@ -794,12 +794,10 @@ int parse_events(const struct option *opt __used, const char *str, int unset __u if (ret != EVT_HANDLED_ALL) { struct perf_evsel *evsel; - evsel = perf_evsel__new(&attr, - nr_counters); + evsel = perf_evsel__new(&attr, evlist->nr_entries); if (evsel == NULL) return -1; - list_add_tail(&evsel->node, &evsel_list); - ++nr_counters; + perf_evlist__add(evlist, evsel); } if (*str == 0) @@ -813,13 +811,14 @@ int parse_events(const struct option *opt __used, const char *str, int unset __u return 0; } -int parse_filter(const struct option *opt __used, const char *str, +int parse_filter(const struct option *opt, const char *str, int unset __used) { + struct perf_evlist *evlist = *(struct perf_evlist **)opt->value; struct perf_evsel *last = NULL; - if (!list_empty(&evsel_list)) - last = list_entry(evsel_list.prev, struct perf_evsel, node); + if (evlist->nr_entries > 0) + last = list_entry(evlist->entries.prev, struct perf_evsel, node); if (last == NULL || last->attr.type != PERF_TYPE_TRACEPOINT) { fprintf(stderr, @@ -981,33 +980,3 @@ void print_events(void) exit(129); } - -int perf_evsel_list__create_default(void) -{ - struct perf_evsel *evsel; - struct perf_event_attr attr; - - memset(&attr, 0, sizeof(attr)); - attr.type = PERF_TYPE_HARDWARE; - attr.config = PERF_COUNT_HW_CPU_CYCLES; - - evsel = perf_evsel__new(&attr, 0); - - if (evsel == NULL) - return -ENOMEM; - - list_add(&evsel->node, &evsel_list); - ++nr_counters; - return 0; -} - -void perf_evsel_list__delete(void) -{ - struct perf_evsel *pos, *n; - - list_for_each_entry_safe(pos, n, &evsel_list, node) { - list_del_init(&pos->node); - perf_evsel__delete(pos); - } - nr_counters = 0; -} diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 458e3ecf17af..cf7e94abb676 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -9,11 +9,6 @@ struct list_head; struct perf_evsel; -extern struct list_head evsel_list; - -int perf_evsel_list__create_default(void); -void perf_evsel_list__delete(void); - struct option; struct tracepoint_path { @@ -25,8 +20,6 @@ struct tracepoint_path { extern struct tracepoint_path *tracepoint_id_to_path(u64 config); extern bool have_tracepoints(struct list_head *evlist); -extern int nr_counters; - const char *event_name(struct perf_evsel *event); extern const char *__event_name(int type, u64 config); diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 6e29d9c9dccc..859d377a3df3 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -31,6 +31,7 @@ #include <string.h> #include <stdarg.h> #include <limits.h> +#include <elf.h> #undef _GNU_SOURCE #include "util.h" @@ -111,7 +112,25 @@ static struct symbol *__find_kernel_function_by_name(const char *name, NULL); } -const char *kernel_get_module_path(const char *module) +static struct map *kernel_get_module_map(const char *module) +{ + struct rb_node *nd; + struct map_groups *grp = &machine.kmaps; + + if (!module) + module = "kernel"; + + for (nd = rb_first(&grp->maps[MAP__FUNCTION]); nd; nd = rb_next(nd)) { + struct map *pos = rb_entry(nd, struct map, rb_node); + if (strncmp(pos->dso->short_name + 1, module, + pos->dso->short_name_len - 2) == 0) { + return pos; + } + } + return NULL; +} + +static struct dso *kernel_get_module_dso(const char *module) { struct dso *dso; struct map *map; @@ -141,7 +160,13 @@ const char *kernel_get_module_path(const char *module) } } found: - return dso->long_name; + return dso; +} + +const char *kernel_get_module_path(const char *module) +{ + struct dso *dso = kernel_get_module_dso(module); + return (dso) ? dso->long_name : NULL; } #ifdef DWARF_SUPPORT @@ -1913,3 +1938,42 @@ int del_perf_probe_events(struct strlist *dellist) return ret; } +/* + * If a symbol corresponds to a function with global binding return 0. + * For all others return 1. + */ +static int filter_non_global_functions(struct map *map __unused, + struct symbol *sym) +{ + if (sym->binding != STB_GLOBAL) + return 1; + + return 0; +} + +int show_available_funcs(const char *module) +{ + struct map *map; + int ret; + + setup_pager(); + + ret = init_vmlinux(); + if (ret < 0) + return ret; + + map = kernel_get_module_map(module); + if (!map) { + pr_err("Failed to find %s map.\n", (module) ? : "kernel"); + return -EINVAL; + } + if (map__load(map, filter_non_global_functions)) { + pr_err("Failed to load map.\n"); + return -EINVAL; + } + if (!dso__sorted_by_name(map->dso, map->type)) + dso__sort_by_name(map->dso, map->type); + + dso__fprintf_symbols_by_name(map->dso, map->type, stdout); + return 0; +} diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h index 5accbedfea37..1fb4f18337d3 100644 --- a/tools/perf/util/probe-event.h +++ b/tools/perf/util/probe-event.h @@ -127,6 +127,7 @@ extern int show_line_range(struct line_range *lr, const char *module); extern int show_available_vars(struct perf_probe_event *pevs, int npevs, int max_probe_points, const char *module, bool externs); +extern int show_available_funcs(const char *module); /* Maximum index number of event-name postfix */ diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index ab83b6ac5d65..69215bff17e9 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -280,6 +280,19 @@ static bool die_compare_name(Dwarf_Die *dw_die, const char *tname) return name ? (strcmp(tname, name) == 0) : false; } +/* Get callsite line number of inline-function instance */ +static int die_get_call_lineno(Dwarf_Die *in_die) +{ + Dwarf_Attribute attr; + Dwarf_Word ret; + + if (!dwarf_attr(in_die, DW_AT_call_line, &attr)) + return -ENOENT; + + dwarf_formudata(&attr, &ret); + return (int)ret; +} + /* Get type die */ static Dwarf_Die *die_get_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem) { @@ -458,6 +471,151 @@ static Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, return die_find_child(sp_die, __die_find_inline_cb, &addr, die_mem); } +/* Walker on lines (Note: line number will not be sorted) */ +typedef int (* line_walk_handler_t) (const char *fname, int lineno, + Dwarf_Addr addr, void *data); + +struct __line_walk_param { + const char *fname; + line_walk_handler_t handler; + void *data; + int retval; +}; + +static int __die_walk_funclines_cb(Dwarf_Die *in_die, void *data) +{ + struct __line_walk_param *lw = data; + Dwarf_Addr addr; + int lineno; + + if (dwarf_tag(in_die) == DW_TAG_inlined_subroutine) { + lineno = die_get_call_lineno(in_die); + if (lineno > 0 && dwarf_entrypc(in_die, &addr) == 0) { + lw->retval = lw->handler(lw->fname, lineno, addr, + lw->data); + if (lw->retval != 0) + return DIE_FIND_CB_FOUND; + } + } + return DIE_FIND_CB_SIBLING; +} + +/* Walk on lines of blocks included in given DIE */ +static int __die_walk_funclines(Dwarf_Die *sp_die, + line_walk_handler_t handler, void *data) +{ + struct __line_walk_param lw = { + .handler = handler, + .data = data, + .retval = 0, + }; + Dwarf_Die die_mem; + Dwarf_Addr addr; + int lineno; + + /* Handle function declaration line */ + lw.fname = dwarf_decl_file(sp_die); + if (lw.fname && dwarf_decl_line(sp_die, &lineno) == 0 && + dwarf_entrypc(sp_die, &addr) == 0) { + lw.retval = handler(lw.fname, lineno, addr, data); + if (lw.retval != 0) + goto done; + } + die_find_child(sp_die, __die_walk_funclines_cb, &lw, &die_mem); +done: + return lw.retval; +} + +static int __die_walk_culines_cb(Dwarf_Die *sp_die, void *data) +{ + struct __line_walk_param *lw = data; + + lw->retval = __die_walk_funclines(sp_die, lw->handler, lw->data); + if (lw->retval != 0) + return DWARF_CB_ABORT; + + return DWARF_CB_OK; +} + +/* + * Walk on lines inside given PDIE. If the PDIE is subprogram, walk only on + * the lines inside the subprogram, otherwise PDIE must be a CU DIE. + */ +static int die_walk_lines(Dwarf_Die *pdie, line_walk_handler_t handler, + void *data) +{ + Dwarf_Lines *lines; + Dwarf_Line *line; + Dwarf_Addr addr; + const char *fname; + int lineno, ret = 0; + Dwarf_Die die_mem, *cu_die; + size_t nlines, i; + + /* Get the CU die */ + if (dwarf_tag(pdie) == DW_TAG_subprogram) + cu_die = dwarf_diecu(pdie, &die_mem, NULL, NULL); + else + cu_die = pdie; + if (!cu_die) { + pr_debug2("Failed to get CU from subprogram\n"); + return -EINVAL; + } + + /* Get lines list in the CU */ + if (dwarf_getsrclines(cu_die, &lines, &nlines) != 0) { + pr_debug2("Failed to get source lines on this CU.\n"); + return -ENOENT; + } + pr_debug2("Get %zd lines from this CU\n", nlines); + + /* Walk on the lines on lines list */ + for (i = 0; i < nlines; i++) { + line = dwarf_onesrcline(lines, i); + if (line == NULL || + dwarf_lineno(line, &lineno) != 0 || + dwarf_lineaddr(line, &addr) != 0) { + pr_debug2("Failed to get line info. " + "Possible error in debuginfo.\n"); + continue; + } + /* Filter lines based on address */ + if (pdie != cu_die) + /* + * Address filtering + * The line is included in given function, and + * no inline block includes it. + */ + if (!dwarf_haspc(pdie, addr) || + die_find_inlinefunc(pdie, addr, &die_mem)) + continue; + /* Get source line */ + fname = dwarf_linesrc(line, NULL, NULL); + + ret = handler(fname, lineno, addr, data); + if (ret != 0) + return ret; + } + + /* + * Dwarf lines doesn't include function declarations and inlined + * subroutines. We have to check functions list or given function. + */ + if (pdie != cu_die) + ret = __die_walk_funclines(pdie, handler, data); + else { + struct __line_walk_param param = { + .handler = handler, + .data = data, + .retval = 0, + }; + dwarf_getfuncs(cu_die, __die_walk_culines_cb, ¶m, 0); + ret = param.retval; + } + + return ret; +} + struct __find_variable_param { const char *name; Dwarf_Addr addr; @@ -1050,43 +1208,26 @@ static int call_probe_finder(Dwarf_Die *sp_die, struct probe_finder *pf) return ret; } -/* Find probe point from its line number */ -static int find_probe_point_by_line(struct probe_finder *pf) +static int probe_point_line_walker(const char *fname, int lineno, + Dwarf_Addr addr, void *data) { - Dwarf_Lines *lines; - Dwarf_Line *line; - size_t nlines, i; - Dwarf_Addr addr; - int lineno; - int ret = 0; - - if (dwarf_getsrclines(&pf->cu_die, &lines, &nlines) != 0) { - pr_warning("No source lines found.\n"); - return -ENOENT; - } + struct probe_finder *pf = data; + int ret; - for (i = 0; i < nlines && ret == 0; i++) { - line = dwarf_onesrcline(lines, i); - if (dwarf_lineno(line, &lineno) != 0 || - lineno != pf->lno) - continue; + if (lineno != pf->lno || strtailcmp(fname, pf->fname) != 0) + return 0; - /* TODO: Get fileno from line, but how? */ - if (strtailcmp(dwarf_linesrc(line, NULL, NULL), pf->fname) != 0) - continue; + pf->addr = addr; + ret = call_probe_finder(NULL, pf); - if (dwarf_lineaddr(line, &addr) != 0) { - pr_warning("Failed to get the address of the line.\n"); - return -ENOENT; - } - pr_debug("Probe line found: line[%d]:%d addr:0x%jx\n", - (int)i, lineno, (uintmax_t)addr); - pf->addr = addr; + /* Continue if no error, because the line will be in inline function */ + return ret < 0 ?: 0; +} - ret = call_probe_finder(NULL, pf); - /* Continuing, because target line might be inlined. */ - } - return ret; +/* Find probe point from its line number */ +static int find_probe_point_by_line(struct probe_finder *pf) +{ + return die_walk_lines(&pf->cu_die, probe_point_line_walker, pf); } /* Find lines which match lazy pattern */ @@ -1140,15 +1281,31 @@ out_close: return nlines; } +static int probe_point_lazy_walker(const char *fname, int lineno, + Dwarf_Addr addr, void *data) +{ + struct probe_finder *pf = data; + int ret; + + if (!line_list__has_line(&pf->lcache, lineno) || + strtailcmp(fname, pf->fname) != 0) + return 0; + + pr_debug("Probe line found: line:%d addr:0x%llx\n", + lineno, (unsigned long long)addr); + pf->addr = addr; + ret = call_probe_finder(NULL, pf); + + /* + * Continue if no error, because the lazy pattern will match + * to other lines + */ + return ret < 0 ?: 0; +} + /* Find probe points from lazy pattern */ static int find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) { - Dwarf_Lines *lines; - Dwarf_Line *line; - size_t nlines, i; - Dwarf_Addr addr; - Dwarf_Die die_mem; - int lineno; int ret = 0; if (list_empty(&pf->lcache)) { @@ -1162,45 +1319,7 @@ static int find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) return ret; } - if (dwarf_getsrclines(&pf->cu_die, &lines, &nlines) != 0) { - pr_warning("No source lines found.\n"); - return -ENOENT; - } - - for (i = 0; i < nlines && ret >= 0; i++) { - line = dwarf_onesrcline(lines, i); - - if (dwarf_lineno(line, &lineno) != 0 || - !line_list__has_line(&pf->lcache, lineno)) - continue; - - /* TODO: Get fileno from line, but how? */ - if (strtailcmp(dwarf_linesrc(line, NULL, NULL), pf->fname) != 0) - continue; - - if (dwarf_lineaddr(line, &addr) != 0) { - pr_debug("Failed to get the address of line %d.\n", - lineno); - continue; - } - if (sp_die) { - /* Address filtering 1: does sp_die include addr? */ - if (!dwarf_haspc(sp_die, addr)) - continue; - /* Address filtering 2: No child include addr? */ - if (die_find_inlinefunc(sp_die, addr, &die_mem)) - continue; - } - - pr_debug("Probe line found: line[%d]:%d addr:0x%llx\n", - (int)i, lineno, (unsigned long long)addr); - pf->addr = addr; - - ret = call_probe_finder(sp_die, pf); - /* Continuing, because target line might be inlined. */ - } - /* TODO: deallocate lines, but how? */ - return ret; + return die_walk_lines(sp_die, probe_point_lazy_walker, pf); } /* Callback parameter with return value */ @@ -1644,91 +1763,28 @@ static int line_range_add_line(const char *src, unsigned int lineno, return line_list__add_line(&lr->line_list, lineno); } -/* Search function declaration lines */ -static int line_range_funcdecl_cb(Dwarf_Die *sp_die, void *data) +static int line_range_walk_cb(const char *fname, int lineno, + Dwarf_Addr addr __used, + void *data) { - struct dwarf_callback_param *param = data; - struct line_finder *lf = param->data; - const char *src; - int lineno; + struct line_finder *lf = data; - src = dwarf_decl_file(sp_die); - if (src && strtailcmp(src, lf->fname) != 0) - return DWARF_CB_OK; - - if (dwarf_decl_line(sp_die, &lineno) != 0 || + if ((strtailcmp(fname, lf->fname) != 0) || (lf->lno_s > lineno || lf->lno_e < lineno)) - return DWARF_CB_OK; + return 0; - param->retval = line_range_add_line(src, lineno, lf->lr); - if (param->retval < 0) - return DWARF_CB_ABORT; - return DWARF_CB_OK; -} + if (line_range_add_line(fname, lineno, lf->lr) < 0) + return -EINVAL; -static int find_line_range_func_decl_lines(struct line_finder *lf) -{ - struct dwarf_callback_param param = {.data = (void *)lf, .retval = 0}; - dwarf_getfuncs(&lf->cu_die, line_range_funcdecl_cb, ¶m, 0); - return param.retval; + return 0; } /* Find line range from its line number */ static int find_line_range_by_line(Dwarf_Die *sp_die, struct line_finder *lf) { - Dwarf_Lines *lines; - Dwarf_Line *line; - size_t nlines, i; - Dwarf_Addr addr; - int lineno, ret = 0; - const char *src; - Dwarf_Die die_mem; - - line_list__init(&lf->lr->line_list); - if (dwarf_getsrclines(&lf->cu_die, &lines, &nlines) != 0) { - pr_warning("No source lines found.\n"); - return -ENOENT; - } - - /* Search probable lines on lines list */ - for (i = 0; i < nlines; i++) { - line = dwarf_onesrcline(lines, i); - if (dwarf_lineno(line, &lineno) != 0 || - (lf->lno_s > lineno || lf->lno_e < lineno)) - continue; - - if (sp_die) { - /* Address filtering 1: does sp_die include addr? */ - if (dwarf_lineaddr(line, &addr) != 0 || - !dwarf_haspc(sp_die, addr)) - continue; - - /* Address filtering 2: No child include addr? */ - if (die_find_inlinefunc(sp_die, addr, &die_mem)) - continue; - } - - /* TODO: Get fileno from line, but how? */ - src = dwarf_linesrc(line, NULL, NULL); - if (strtailcmp(src, lf->fname) != 0) - continue; - - ret = line_range_add_line(src, lineno, lf->lr); - if (ret < 0) - return ret; - } + int ret; - /* - * Dwarf lines doesn't include function declarations. We have to - * check functions list or given function. - */ - if (sp_die) { - src = dwarf_decl_file(sp_die); - if (src && dwarf_decl_line(sp_die, &lineno) == 0 && - (lf->lno_s <= lineno && lf->lno_e >= lineno)) - ret = line_range_add_line(src, lineno, lf->lr); - } else - ret = find_line_range_func_decl_lines(lf); + ret = die_walk_lines(sp_die ?: &lf->cu_die, line_range_walk_cb, lf); /* Update status */ if (ret >= 0) @@ -1758,9 +1814,6 @@ static int line_range_search_cb(Dwarf_Die *sp_die, void *data) struct line_finder *lf = param->data; struct line_range *lr = lf->lr; - pr_debug("find (%llx) %s\n", - (unsigned long long)dwarf_dieoffset(sp_die), - dwarf_diename(sp_die)); if (dwarf_tag(sp_die) == DW_TAG_subprogram && die_compare_name(sp_die, lr->function)) { lf->fname = dwarf_decl_file(sp_die); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 105f00bfd555..e6a07408669e 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -242,17 +242,16 @@ static bool symbol__match_parent_regex(struct symbol *sym) return 0; } -struct map_symbol *perf_session__resolve_callchain(struct perf_session *self, - struct thread *thread, - struct ip_callchain *chain, - struct symbol **parent) +int perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent) { u8 cpumode = PERF_RECORD_MISC_USER; unsigned int i; - struct map_symbol *syms = calloc(chain->nr, sizeof(*syms)); + int err; - if (!syms) - return NULL; + callchain_cursor_reset(&self->callchain_cursor); for (i = 0; i < chain->nr; i++) { u64 ip = chain->ips[i]; @@ -281,12 +280,15 @@ struct map_symbol *perf_session__resolve_callchain(struct perf_session *self, *parent = al.sym; if (!symbol_conf.use_callchain) break; - syms[i].map = al.map; - syms[i].sym = al.sym; } + + err = callchain_cursor_append(&self->callchain_cursor, + ip, al.map, al.sym); + if (err) + return err; } - return syms; + return 0; } static int process_event_synth_stub(event_t *event __used, @@ -494,7 +496,7 @@ static void flush_sample_queue(struct perf_session *s, if (iter->timestamp > limit) break; - event__parse_sample(iter->event, s, &sample); + perf_session__parse_sample(s, iter->event, &sample); perf_session_deliver_event(s, iter->event, &sample, ops, iter->file_offset); @@ -804,7 +806,7 @@ static int perf_session__process_event(struct perf_session *session, /* * For all kernel events we get the sample data */ - event__parse_sample(event, session, &sample); + perf_session__parse_sample(session, event, &sample); /* Preprocess sample records - precheck callchains */ if (perf_session__preprocess_sample(session, event, &sample)) diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index decd83f274fd..78239767011e 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -51,7 +51,8 @@ struct perf_session { int cwdlen; char *cwd; struct ordered_samples ordered_samples; - char filename[0]; + struct callchain_cursor callchain_cursor; + char filename[0]; }; struct perf_event_ops; @@ -94,10 +95,10 @@ int __perf_session__process_events(struct perf_session *self, int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); -struct map_symbol *perf_session__resolve_callchain(struct perf_session *self, - struct thread *thread, - struct ip_callchain *chain, - struct symbol **parent); +int perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent); bool perf_session__has_traces(struct perf_session *self, const char *msg); @@ -154,4 +155,13 @@ size_t perf_session__fprintf_nr_events(struct perf_session *self, FILE *fp) { return hists__fprintf_nr_events(&self->hists, fp); } + +static inline int perf_session__parse_sample(struct perf_session *session, + const event_t *event, + struct sample_data *sample) +{ + return event__parse_sample(event, session->sample_type, + session->sample_id_all, sample); +} + #endif /* __PERF_SESSION_H */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 00f4eade2e3e..d5d3b22250f3 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -7,61 +7,6 @@ #include "util.h" #include "debug.h" -/* Skip "." and ".." directories */ -static int filter(const struct dirent *dir) -{ - if (dir->d_name[0] == '.') - return 0; - else - return 1; -} - -struct thread_map *thread_map__new_by_pid(pid_t pid) -{ - struct thread_map *threads; - char name[256]; - int items; - struct dirent **namelist = NULL; - int i; - - sprintf(name, "/proc/%d/task", pid); - items = scandir(name, &namelist, filter, NULL); - if (items <= 0) - return NULL; - - threads = malloc(sizeof(*threads) + sizeof(pid_t) * items); - if (threads != NULL) { - for (i = 0; i < items; i++) - threads->map[i] = atoi(namelist[i]->d_name); - threads->nr = items; - } - - for (i=0; i<items; i++) - free(namelist[i]); - free(namelist); - - return threads; -} - -struct thread_map *thread_map__new_by_tid(pid_t tid) -{ - struct thread_map *threads = malloc(sizeof(*threads) + sizeof(pid_t)); - - if (threads != NULL) { - threads->map[0] = tid; - threads->nr = 1; - } - - return threads; -} - -struct thread_map *thread_map__new(pid_t pid, pid_t tid) -{ - if (pid != -1) - return thread_map__new_by_pid(pid); - return thread_map__new_by_tid(tid); -} - static struct thread *thread__new(pid_t pid) { struct thread *self = zalloc(sizeof(*self)); diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index d7574101054a..e5f2401c1b5e 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -18,24 +18,10 @@ struct thread { int comm_len; }; -struct thread_map { - int nr; - int map[]; -}; - struct perf_session; void thread__delete(struct thread *self); -struct thread_map *thread_map__new_by_pid(pid_t pid); -struct thread_map *thread_map__new_by_tid(pid_t tid); -struct thread_map *thread_map__new(pid_t pid, pid_t tid); - -static inline void thread_map__delete(struct thread_map *threads) -{ - free(threads); -} - int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); struct thread *perf_session__findnew(struct perf_session *self, pid_t pid); diff --git a/tools/perf/util/thread_map.c b/tools/perf/util/thread_map.c new file mode 100644 index 000000000000..a5df131b77c3 --- /dev/null +++ b/tools/perf/util/thread_map.c @@ -0,0 +1,64 @@ +#include <dirent.h> +#include <stdlib.h> +#include <stdio.h> +#include "thread_map.h" + +/* Skip "." and ".." directories */ +static int filter(const struct dirent *dir) +{ + if (dir->d_name[0] == '.') + return 0; + else + return 1; +} + +struct thread_map *thread_map__new_by_pid(pid_t pid) +{ + struct thread_map *threads; + char name[256]; + int items; + struct dirent **namelist = NULL; + int i; + + sprintf(name, "/proc/%d/task", pid); + items = scandir(name, &namelist, filter, NULL); + if (items <= 0) + return NULL; + + threads = malloc(sizeof(*threads) + sizeof(pid_t) * items); + if (threads != NULL) { + for (i = 0; i < items; i++) + threads->map[i] = atoi(namelist[i]->d_name); + threads->nr = items; + } + + for (i=0; i<items; i++) + free(namelist[i]); + free(namelist); + + return threads; +} + +struct thread_map *thread_map__new_by_tid(pid_t tid) +{ + struct thread_map *threads = malloc(sizeof(*threads) + sizeof(pid_t)); + + if (threads != NULL) { + threads->map[0] = tid; + threads->nr = 1; + } + + return threads; +} + +struct thread_map *thread_map__new(pid_t pid, pid_t tid) +{ + if (pid != -1) + return thread_map__new_by_pid(pid); + return thread_map__new_by_tid(tid); +} + +void thread_map__delete(struct thread_map *threads) +{ + free(threads); +} diff --git a/tools/perf/util/thread_map.h b/tools/perf/util/thread_map.h new file mode 100644 index 000000000000..3cb907311409 --- /dev/null +++ b/tools/perf/util/thread_map.h @@ -0,0 +1,15 @@ +#ifndef __PERF_THREAD_MAP_H +#define __PERF_THREAD_MAP_H + +#include <sys/types.h> + +struct thread_map { + int nr; + int map[]; +}; + +struct thread_map *thread_map__new_by_pid(pid_t pid); +struct thread_map *thread_map__new_by_tid(pid_t tid); +struct thread_map *thread_map__new(pid_t pid, pid_t tid); +void thread_map__delete(struct thread_map *threads); +#endif /* __PERF_THREAD_MAP_H */ diff --git a/tools/perf/util/ui/browsers/hists.c b/tools/perf/util/ui/browsers/hists.c index 60c463c16028..86428239fa65 100644 --- a/tools/perf/util/ui/browsers/hists.c +++ b/tools/perf/util/ui/browsers/hists.c @@ -377,7 +377,7 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *self, while (node) { struct callchain_node *child = rb_entry(node, struct callchain_node, rb_node); struct rb_node *next = rb_next(node); - u64 cumul = cumul_hits(child); + u64 cumul = callchain_cumul_hits(child); struct callchain_list *chain; char folded_sign = ' '; int first = true; |