aha/drivers/oprofile/cpu_buffer.h

121 lines
2.8 KiB
C
Raw Normal View History

/**
* @file cpu_buffer.h
*
* @remark Copyright 2002-2009 OProfile authors
* @remark Read the file COPYING
*
* @author John Levon <levon@movementarian.org>
* @author Robert Richter <robert.richter@amd.com>
*/
#ifndef OPROFILE_CPU_BUFFER_H
#define OPROFILE_CPU_BUFFER_H
#include <linux/types.h>
#include <linux/spinlock.h>
#include <linux/workqueue.h>
#include <linux/cache.h>
#include <linux/sched.h>
oprofile: port to the new ring_buffer This patch replaces the current oprofile cpu buffer implementation with the ring buffer provided by the tracing framework. The motivation here is to leave the pain of implementing ring buffers to others. Oh, no, there are more advantages. Main reason is the support of different sample sizes that could be stored in the buffer. Use cases for this are IBS and Cell spu profiling. Using the new ring buffer ensures valid and complete samples and allows copying the cpu buffer stateless without knowing its content. Second it will use generic kernel API and also reduce code size. And hopefully, there are less bugs. Since the new tracing ring buffer implementation uses spin locks to protect the buffer during read/write access, it is difficult to use the buffer in an NMI handler. In this case, writing to the buffer by the NMI handler (x86) could occur also during critical sections when reading the buffer. To avoid this, there are 2 buffers for independent read and write access. Read access is in process context only, write access only in the NMI handler. If the read buffer runs empty, both buffers are swapped atomically. There is potentially a small window during swapping where the buffers are disabled and samples could be lost. Using 2 buffers is a little bit overhead, but the solution is clear and does not require changes in the ring buffer implementation. It can be changed to a single buffer solution when the ring buffer access is implemented as non-locking atomic code. The new buffer requires more size to store the same amount of samples because each sample includes an u32 header. Also, there is more code to execute for buffer access. Nonetheless, the buffer implementation is proven in the ftrace environment and worth to use also in oprofile. Patches that changes the internal IBS buffer usage will follow. Cc: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Robert Richter <robert.richter@amd.com>
2008-12-09 00:21:32 +00:00
#include <linux/ring_buffer.h>
struct task_struct;
int alloc_cpu_buffers(void);
void free_cpu_buffers(void);
void start_cpu_work(void);
void end_cpu_work(void);
/* CPU buffer is composed of such entries (which are
* also used for context switch notes)
*/
struct op_sample {
unsigned long eip;
unsigned long event;
unsigned long data[0];
};
struct op_entry;
oprofile: port to the new ring_buffer This patch replaces the current oprofile cpu buffer implementation with the ring buffer provided by the tracing framework. The motivation here is to leave the pain of implementing ring buffers to others. Oh, no, there are more advantages. Main reason is the support of different sample sizes that could be stored in the buffer. Use cases for this are IBS and Cell spu profiling. Using the new ring buffer ensures valid and complete samples and allows copying the cpu buffer stateless without knowing its content. Second it will use generic kernel API and also reduce code size. And hopefully, there are less bugs. Since the new tracing ring buffer implementation uses spin locks to protect the buffer during read/write access, it is difficult to use the buffer in an NMI handler. In this case, writing to the buffer by the NMI handler (x86) could occur also during critical sections when reading the buffer. To avoid this, there are 2 buffers for independent read and write access. Read access is in process context only, write access only in the NMI handler. If the read buffer runs empty, both buffers are swapped atomically. There is potentially a small window during swapping where the buffers are disabled and samples could be lost. Using 2 buffers is a little bit overhead, but the solution is clear and does not require changes in the ring buffer implementation. It can be changed to a single buffer solution when the ring buffer access is implemented as non-locking atomic code. The new buffer requires more size to store the same amount of samples because each sample includes an u32 header. Also, there is more code to execute for buffer access. Nonetheless, the buffer implementation is proven in the ftrace environment and worth to use also in oprofile. Patches that changes the internal IBS buffer usage will follow. Cc: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Robert Richter <robert.richter@amd.com>
2008-12-09 00:21:32 +00:00
struct oprofile_cpu_buffer {
unsigned long buffer_size;
struct task_struct *last_task;
int last_is_kernel;
int tracing;
unsigned long sample_received;
unsigned long sample_lost_overflow;
unsigned long backtrace_aborted;
unsigned long sample_invalid_eip;
int cpu;
struct delayed_work work;
};
DECLARE_PER_CPU(struct oprofile_cpu_buffer, op_cpu_buffer);
/*
* Resets the cpu buffer to a sane state.
*
* reset these to invalid values; the next sample collected will
* populate the buffer with proper values to initialize the buffer
*/
static inline void op_cpu_buffer_reset(int cpu)
{
struct oprofile_cpu_buffer *cpu_buf = &per_cpu(op_cpu_buffer, cpu);
cpu_buf->last_is_kernel = -1;
cpu_buf->last_task = NULL;
}
/*
* op_cpu_buffer_add_data() and op_cpu_buffer_write_commit() may be
* called only if op_cpu_buffer_write_reserve() did not return NULL or
* entry->event != NULL, otherwise entry->size or entry->event will be
* used uninitialized.
*/
struct op_sample
*op_cpu_buffer_write_reserve(struct op_entry *entry, unsigned long size);
int op_cpu_buffer_write_commit(struct op_entry *entry);
struct op_sample *op_cpu_buffer_read_entry(struct op_entry *entry, int cpu);
unsigned long op_cpu_buffer_entries(int cpu);
/* returns the remaining free size of data in the entry */
static inline
int op_cpu_buffer_add_data(struct op_entry *entry, unsigned long val)
{
if (!entry->size)
return 0;
*entry->data = val;
entry->size--;
entry->data++;
return entry->size;
}
/* returns the size of data in the entry */
static inline
int op_cpu_buffer_get_size(struct op_entry *entry)
{
return entry->size;
}
/* returns 0 if empty or the size of data including the current value */
static inline
int op_cpu_buffer_get_data(struct op_entry *entry, unsigned long *val)
{
int size = entry->size;
if (!size)
return 0;
*val = *entry->data;
entry->size--;
entry->data++;
return size;
}
/* extra data flags */
#define KERNEL_CTX_SWITCH (1UL << 0)
#define IS_KERNEL (1UL << 1)
#define TRACE_BEGIN (1UL << 2)
#define USER_CTX_SWITCH (1UL << 3)
#endif /* OPROFILE_CPU_BUFFER_H */